@shenghuabi/llama 1.0.0

package/define/llama-server.define.d.ts

@@ -0,0 +1,2347 @@
+import * as v from 'valibot';
+export declare const LlamaServerDefine: v.ObjectSchema<{
+    readonly common: v.OptionalSchema<v.ObjectSchema<{
+        readonly 'verbose-prompt': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 在生成前打印详细提示（默认：false）\n- print a verbose prompt before generation (default: false)">]>;
+        readonly threads: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 生成期间使用的线程数（默认：-1）\n- number of threads to use during generation (default: -1)  (env: LLAMA_ARG_THREADS)">]>;
+        readonly 'threads-batch': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 批量和提示处理期间使用的线程数（默认：与--threads相同）\n- number of threads to use during batch and prompt processing (default: same as --threads)">]>;
+        readonly 'cpu-mask': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "M">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- CPU亲和力掩码：任意长度的十六进制数。补充cpu-range（默认：\"\"）\n- CPU affinity mask: arbitrarily long hex. Complements cpu-range (default: \"\")">]>;
+        readonly 'cpu-range': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "lo-hi">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- CPU范围用于亲和力。补充--cpu-mask\n- range of CPUs for affinity. Complements --cpu-mask">]>;
+        readonly 'cpu-strict': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.PicklistSchema<["0", "1"], undefined>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: ["0" | "1"] | undefined;
+        } | undefined, "- 使用严格CPU放置（默认：0）\n- use strict CPU placement (default: 0)">]>;
+        readonly prio: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.PicklistSchema<[0, 1, 2, 3], undefined>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [0 | 2 | 1 | 3] | undefined;
+        } | undefined, "- 设置进程/线程优先级：低(-1)、正常(0)、中等(1)、高(2)、实时(3)（默认值：0）\n- set process/thread priority : low(-1), normal(0), medium(1), high(2), realtime(3) (default: 0)">]>;
+        readonly poll: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "<0...100>">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 使用轮询级别等待工作 (0 - 无轮询，默认: 50)\n- use polling level to wait for work (0 - no polling, default: 50)">]>;
+        readonly 'cpu-mask-batch': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "M">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- CPU亲和力掩码：任意长度的十六进制数。补充cpu-range-batch（默认：与--cpu-mask相同）\n- CPU affinity mask: arbitrarily long hex. Complements cpu-range-batch (default: same as --cpu-mask)">]>;
+        readonly 'cpu-range-batch': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "lo-hi">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- CPU亲和力的范围。补充--cpu-mask-batch\n- ranges of CPUs for affinity. Complements --cpu-mask-batch">]>;
+        readonly 'cpu-strict-batch': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.PicklistSchema<["0", "1"], undefined>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: ["0" | "1"] | undefined;
+        } | undefined, "- 使用严格的CPU放置（默认：与--cpu-strict相同）\n- use strict CPU placement (default: same as --cpu-strict)">]>;
+        readonly 'prio-batch': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.PicklistSchema<[0, 1, 2, 3], undefined>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [0 | 2 | 1 | 3] | undefined;
+        } | undefined, "- 设置进程/线程优先级 : 0-正常, 1-中等, 2-高, 3-实时 (默认: 0)\n- set process/thread priority : 0-normal, 1-medium, 2-high, 3-realtime (default: 0)">]>;
+        readonly 'poll-batch': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.PicklistSchema<["0", "1"], undefined>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: ["0" | "1"] | undefined;
+        } | undefined, "- 使用轮询等待工作（默认：与--poll相同）\n- use polling to wait for work (default: same as --poll)">]>;
+        readonly 'ctx-size': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 提示上下文的大小（默认：4096，0 = 从模型加载）\n- size of the prompt context (default: 4096, 0 = loaded from model)  (env: LLAMA_ARG_CTX_SIZE)">]>;
+        readonly predict: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 预测的token数量（默认：-1，-1 = 无限）\n- number of tokens to predict (default: -1, -1 = infinity)  (env: LLAMA_ARG_N_PREDICT)">]>;
+        readonly 'batch-size': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 逻辑最大批处理大小（默认：2048）\n- logical maximum batch size (default: 2048)  (env: LLAMA_ARG_BATCH)">]>;
+        readonly 'ubatch-size': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 物理最大批处理大小（默认：512）\n- physical maximum batch size (default: 512)  (env: LLAMA_ARG_UBATCH)">]>;
+        readonly keep: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 保留初始提示中的令牌数量（默认：0，-1 = 所有）\n- number of tokens to keep from the initial prompt (default: 0, -1 = all)">]>;
+        readonly 'swa-full': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 使用全尺寸SWA缓存（默认：false） [更多信息](https://github.com/ggml-org/llama.cpp/pull/13194#issuecomment-2868343055)\n- use full-size SWA cache (default: false)  [(more info)](https://github.com/ggml-org/llama.cpp/pull/13194#issuecomment-2868343055)  (env: LLAMA_ARG_SWA_FULL)">]>;
+        readonly 'kv-unified': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 为所有序列的 KV 缓存使用一个统一的 KV 缓冲区（默认值为 `false`）：[更多信息](https://github.com/ggml-org/llama.cpp/pull/14363)（环境变量：`LLAMA_ARG_KV_SPLIT`）\n- use single unified KV buffer for the KV cache of all sequences (default: false)  [(more info)](https://github.com/ggml-org/llama.cpp/pull/14363)  (env: LLAMA_ARG_KV_SPLIT)">]>;
+        readonly 'flash-attn': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.PicklistSchema<["on", "off", "auto"], undefined>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: ["on" | "off" | "auto"] | undefined;
+        } | undefined, "- 启用 Flash 注意力（默认：禁用）\n- set Flash Attention use ('on', 'off', or 'auto', default: 'auto')  (env: LLAMA_ARG_FLASH_ATTN)">]>;
+        readonly 'no-perf': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 禁用内部libllama性能计时（默认：false）\n- disable internal libllama performance timings (default: false)  (env: LLAMA_ARG_NO_PERF)">]>;
+        readonly escape: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 处理转义序列（\\n, \\r, \\t, \\\", \\\\)（默认：true）\n- process escapes sequences (\\n, \\r, \\t, \\', \\\", \\\\) (default: true)">]>;
+        readonly 'no-escape': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 不要处理转义序列\n- do not process escape sequences">]>;
+        readonly 'rope-scaling': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.PicklistSchema<["none", "linear", "yarn"], undefined>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: ["none" | "linear" | "yarn"] | undefined;
+        } | undefined, "- RoPE频率缩放方法，除非模型指定，默认为线性\n- RoPE frequency scaling method, defaults to linear unless specified by the model  (env: LLAMA_ARG_ROPE_SCALING_TYPE)">]>;
+        readonly 'rope-scale': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- RoPE上下文缩放因子，将上下文扩展N倍\n- RoPE context scaling factor, expands context by a factor of N  (env: LLAMA_ARG_ROPE_SCALE)">]>;
+        readonly 'rope-freq-base': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- RoPE基频，用于NTK感知缩放（默认：从模型加载）\n- RoPE base frequency, used by NTK-aware scaling (default: loaded from model)  (env: LLAMA_ARG_ROPE_FREQ_BASE)">]>;
+        readonly 'rope-freq-scale': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- RoPE频率缩放因子，通过因子1/N扩展上下文\n- RoPE frequency scaling factor, expands context by a factor of 1/N  (env: LLAMA_ARG_ROPE_FREQ_SCALE)">]>;
+        readonly 'yarn-orig-ctx': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- YaRN: 原始上下文大小（默认：0 = 模型训练上下文大小）\n- YaRN: original context size of model (default: 0 = model training context size)  (env: LLAMA_ARG_YARN_ORIG_CTX)">]>;
+        readonly 'yarn-ext-factor': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- YaRN: 插值混合因子（默认：-1.0，0.0 = 完全插值）\n- YaRN: extrapolation mix factor (default: -1.0, 0.0 = full interpolation)  (env: LLAMA_ARG_YARN_EXT_FACTOR)">]>;
+        readonly 'yarn-attn-factor': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- YaRN: 缩放 sqrt(t) 或注意力幅度（默认：-1.0）\n- YaRN: scale sqrt(t) or attention magnitude (default: -1.0)  (env: LLAMA_ARG_YARN_ATTN_FACTOR)">]>;
+        readonly 'yarn-beta-slow': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- YaRN: 高校正维度或alpha（默认：-1.0）\n- YaRN: high correction dim or alpha (default: -1.0)  (env: LLAMA_ARG_YARN_BETA_SLOW)">]>;
+        readonly 'yarn-beta-fast': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- YaRN: 低校正维或beta（默认：-1）\n- YaRN: low correction dim or beta (default: -1.0)  (env: LLAMA_ARG_YARN_BETA_FAST)">]>;
+        readonly 'no-kv-offload': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 禁用KV卸载\n- disable KV offload  (env: LLAMA_ARG_NO_KV_OFFLOAD)">]>;
+        readonly 'no-repack': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 禁用重量重新打包功能\n- disable weight repacking  (env: LLAMA_ARG_NO_REPACK)">]>;
+        readonly 'cache-type-k': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.PicklistSchema<["f32", "f16", "bf16", "q8_0", "q4_0", "q4_1", "iq4_nl", "q5_0", "q5_1"], undefined>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: ["f32" | "f16" | "bf16" | "q8_0" | "q4_0" | "q4_1" | "iq4_nl" | "q5_0" | "q5_1"] | undefined;
+        } | undefined, "- KV缓存数据类型用于K 允许的值: f32, f16, bf16, q8_0, q4_0, q4_1, iq4_nl, q5_0, q5_1 （默认: f16）\n- KV cache data type for K  allowed values: f32, f16, bf16, q8_0, q4_0, q4_1, iq4_nl, q5_0, q5_1  (default: f16)  (env: LLAMA_ARG_CACHE_TYPE_K)">]>;
+        readonly 'cache-type-v': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.PicklistSchema<["f32", "f16", "bf16", "q8_0", "q4_0", "q4_1", "iq4_nl", "q5_0", "q5_1"], undefined>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: ["f32" | "f16" | "bf16" | "q8_0" | "q4_0" | "q4_1" | "iq4_nl" | "q5_0" | "q5_1"] | undefined;
+        } | undefined, "- KV缓存数据类型用于V，允许的值：f32, f16, bf16, q8_0, q4_0, q4_1, iq4_nl, q5_0, q5_1（默认：f16）\n- KV cache data type for V  allowed values: f32, f16, bf16, q8_0, q4_0, q4_1, iq4_nl, q5_0, q5_1  (default: f16)  (env: LLAMA_ARG_CACHE_TYPE_V)">]>;
+        readonly 'defrag-thold': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- KV缓存碎片化阈值（默认：0.1，<0-禁用）\n- KV cache defragmentation threshold (DEPRECATED)  (env: LLAMA_ARG_DEFRAG_THOLD)">]>;
+        readonly parallel: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 并行解码序列的数量（默认：1）\n- number of parallel sequences to decode (default: 1)  (env: LLAMA_ARG_N_PARALLEL)">]>;
+        readonly mlock: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 强制系统将模型保留在RAM中而不是交换或压缩\n- force system to keep model in RAM rather than swapping or compressing  (env: LLAMA_ARG_MLOCK)">]>;
+        readonly 'no-mmap': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 不要内存映射模型（加载速度较慢，但可能减少页面丢失，如果未使用mlock）\n- do not memory-map model (slower load but may reduce pageouts if not using mlock)  (env: LLAMA_ARG_NO_MMAP)">]>;
+        readonly numa: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.PicklistSchema<["distribute", "isolate", "numactl"], undefined>, v.MetadataAction<"distribute" | "isolate" | "numactl", {
+                readonly enumOptions: readonly [{
+                    readonly label: "distribute";
+                    readonly value: "distribute";
+                    readonly description: "将执行均匀分布在所有节点上";
+                }, {
+                    readonly label: "isolate";
+                    readonly value: "isolate";
+                    readonly description: "仅在执行开始的节点的CPU上生成线程";
+                }, {
+                    readonly label: "numactl";
+                    readonly value: "numactl";
+                    readonly description: "使用numactl提供的CPU映射";
+                }];
+            }>]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: ["distribute" | "isolate" | "numactl"] | undefined;
+        } | undefined, "- 尝试在某些NUMA系统上进行优化 - distribute：将执行均匀分布在所有节点上 - isolate：仅在执行开始的节点的CPU上生成线程 - numactl：使用numactl提供的CPU映射 如果之前未运行过此操作，请在使用此功能前删除系统页面缓存 详见 https://github.com/ggml-org/llama.cpp/issues/1437\n- attempt optimizations that help on some NUMA systems  - distribute: spread execution evenly over all nodes  - isolate: only spawn threads on CPUs on the node that execution started on  - numactl: use the CPU map provided by numactl  if run without this previously, it is recommended to drop the system page cache before using this  see https://github.com/ggml-org/llama.cpp/issues/1437  (env: LLAMA_ARG_NUMA)">]>;
+        readonly device: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "<dev1,dev2,..>">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 逗号分隔的设备列表，用于卸载（none = 不卸载）使用 --list-devices 查看可用设备列表\n- comma-separated list of devices to use for offloading (none = don't offload)  use --list-devices to see a list of available devices  (env: LLAMA_ARG_DEVICE)">]>;
+        readonly 'override-tensor': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "<tensor name pattern>=<buffer type>,...">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 将所有专家组合（Mixture of Experts, MoE）的权重数据存储在 CPU 中。\n- override tensor buffer type">]>;
+        readonly 'cpu-moe': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 将前 N 层的“专家混合”（Expert Mixing, MoE）权重数据保留在 CPU 中。\n- keep all Mixture of Experts (MoE) weights in the CPU  (env: LLAMA_ARG_CPU_MOE)">]>;
+        readonly 'n-cpu-moe': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 覆盖张量缓冲区类型\n- keep the Mixture of Experts (MoE) weights of the first N layers in the CPU  (env: LLAMA_ARG_N_CPU_MOE)">]>;
+        readonly 'gpu-layers': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 存储在VRAM中的图层数量\n- max. number of layers to store in VRAM (default: -1)  (env: LLAMA_ARG_N_GPU_LAYERS)">]>;
+        readonly 'split-mode': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.PicklistSchema<["none", "layer", "row"], undefined>, v.MetadataAction<"none" | "layer" | "row", {
+                readonly enumOptions: readonly [{
+                    readonly label: "none";
+                    readonly value: "none";
+                    readonly description: "仅使用一个GPU";
+                }, {
+                    readonly label: "layer";
+                    readonly value: "layer";
+                    readonly description: "分割层和KV跨GPU";
+                }, {
+                    readonly label: "row";
+                    readonly value: "row";
+                    readonly description: "分割行跨GPU";
+                }];
+            }>]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: ["none" | "layer" | "row"] | undefined;
+        } | undefined, "- 如何将模型分布在多个GPU上，可选：- none: 仅使用一个GPU - layer (默认): 分割层和KV跨GPU - row: 分割行跨GPU\n- how to split the model across multiple GPUs, one of:  - none: use one GPU only  - layer (default): split layers and KV across GPUs  - row: split rows across GPUs  (env: LLAMA_ARG_SPLIT_MODE)">]>;
+        readonly 'tensor-split': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "N0,N1,N2,...">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 将模型分片到每个GPU的分数，逗号分隔的比例列表，例如3,1\n- fraction of the model to offload to each GPU, comma-separated list of proportions, e.g. 3,1  (env: LLAMA_ARG_TENSOR_SPLIT)">]>;
+        readonly 'main-gpu': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "INDEX">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 用于模型（split-mode = none）或中间结果和KV（split-mode = row）的GPU（默认：0）\n- the GPU to use for the model (with split-mode = none), or for intermediate results and KV (with split-mode = row) (default: 0)  (env: LLAMA_ARG_MAIN_GPU)">]>;
+        readonly 'check-tensors': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 检查模型张量数据中的无效值（默认：false）\n- check model tensor data for invalid values (default: false)">]>;
+        readonly 'override-kv': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "KEY=TYPE:VALUE">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 用于通过键覆盖模型元数据的高级选项。可以多次指定。类型：int、float、bool、str。示例：--override-kv tokenizer.ggml.add_bos_token=bool:false\n- advanced option to override model metadata by key. may be specified multiple times.  types: int, float, bool, str. example: --override-kv tokenizer.ggml.add_bos_token=bool:false">]>;
+        readonly 'no-op-offload': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 禁用将主机张量操作卸载到设备（默认：false）\n- disable offloading host tensor operations to device (default: false)">]>;
+        readonly lora: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FNAME">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- LoRA适配器路径（可以重复使用多个适配器）\n- path to LoRA adapter (can be repeated to use multiple adapters)">]>;
+        readonly 'lora-scaled': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FNAME">]>, v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "SCALE">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string, number] | undefined;
+        } | undefined, "- LoRA适配器路径，带用户定义的缩放（可重复使用多个适配器）\n- path to LoRA adapter with user defined scaling (can be repeated to use multiple adapters)">]>;
+        readonly 'control-vector': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FNAME">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 添加控制向量  请注意：此参数可以重复以添加多个控制向量\n- add a control vector  note: this argument can be repeated to add multiple control vectors">]>;
+        readonly 'control-vector-scaled': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FNAME">]>, v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "SCALE">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string, number] | undefined;
+        } | undefined, "- 添加一个带用户定义缩放 SCALE 的控制向量  注意: 此参数可以重复以添加多个缩放的控制向量\n- add a control vector with user defined scaling SCALE  note: this argument can be repeated to add multiple scaled control vectors">]>;
+        readonly 'control-vector-layer-range': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "START">]>, v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "END">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string, string] | undefined;
+        } | undefined, "- 应用控制向量的图层范围，起始和结束值（包含）\n- layer range to apply the control vector(s) to, start and end inclusive">]>;
+        readonly model: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FNAME">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 模型路径 (默认: `models/$filename` 与 `--hf-file` 或 `--model-url` 的文件名，否则 models/7B/ggml-model-f16.gguf)\n- model path (default: `models/$filename` with filename from `--hf-file` or `--model-url` if set, otherwise models/7B/ggml-model-f16.gguf)  (env: LLAMA_ARG_MODEL)">]>;
+        readonly 'model-url': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "MODEL_URL">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- Docker Hub 模型仓库：该仓库是可选的，默认值为 `ai/`；`quant` 参数也是可选的，默认值为 `:latest`。例如：`gemma3`（默认值：未使用）。\n- model download url (default: unused)  (env: LLAMA_ARG_MODEL_URL)">]>;
+        readonly 'docker-repo': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.PicklistSchema<["<repo>/]<model>[:quant"], undefined>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: ["<repo>/]<model>[:quant"] | undefined;
+        } | undefined, "- 模型下载地址（默认：未使用）\n- Docker Hub model repository. repo is optional, default to ai/. quant is optional, default to :latest.  example: gemma3  (default: unused)  (env: LLAMA_ARG_DOCKER_REPO)">]>;
+        readonly 'hf-repo': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "<user>/<model>[:quant]">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- Hugging Face 模型仓库; quant 是可选的，不区分大小写，默认为 Q4_K_M，如果不存在 Q4_K_M 则回退到仓库中的第一个文件。如果可用还会自动下载 mmproj。要禁用，添加 --no-mmproj  示例：unsloth/phi-4-GGUF:q4_k_m  （默认：未使用）\n- Hugging Face model repository; quant is optional, case-insensitive, default to Q4_K_M, or falls back to the first file in the repo if Q4_K_M doesn't exist.  mmproj is also downloaded automatically if available. to disable, add --no-mmproj  example: unsloth/phi-4-GGUF:q4_k_m  (default: unused)  (env: LLAMA_ARG_HF_REPO)">]>;
+        readonly 'hf-repo-draft': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "<user>/<model>[:quant]">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 与--hf-repo相同，但用于草稿模型（默认：未使用）\n- Same as --hf-repo, but for the draft model (default: unused)  (env: LLAMA_ARG_HFD_REPO)">]>;
+        readonly 'hf-file': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FILE">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- Hugging Face模型文件。如果指定了，将覆盖--hf-repo中的量化（默认：未使用）\n- Hugging Face model file. If specified, it will override the quant in --hf-repo (default: unused)  (env: LLAMA_ARG_HF_FILE)">]>;
+        readonly 'hf-repo-v': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "<user>/<model>[:quant]">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- Hugging Face vocoder模型仓库（默认：未使用）\n- Hugging Face model repository for the vocoder model (default: unused)  (env: LLAMA_ARG_HF_REPO_V)">]>;
+        readonly 'hf-file-v': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FILE">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- Hugging Face模型文件用于声码器模型（默认：未使用）\n- Hugging Face model file for the vocoder model (default: unused)  (env: LLAMA_ARG_HF_FILE_V)">]>;
+        readonly 'hf-token': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "TOKEN">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- Hugging Face访问令牌（默认：来自HF_TOKEN环境变量的值）\n- Hugging Face access token (default: value from HF_TOKEN environment variable)  (env: HF_TOKEN)">]>;
+        readonly 'log-disable': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 日志禁用\n- Log disable">]>;
+        readonly 'log-file': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FNAME">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 日志记录到文件\n- Log to file">]>;
+        readonly 'log-colors': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.PicklistSchema<["on", "off", "auto"], undefined>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: ["on" | "off" | "auto"] | undefined;
+        } | undefined, "- 启用彩色日志记录\n- Set colored logging ('on', 'off', or 'auto', default: 'auto')  'auto' enables colors when output is to a terminal  (env: LLAMA_LOG_COLORS)">]>;
+        readonly verbose: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 设置 verbosity 级别为无穷大（即记录所有消息，用于调试）\n- Set verbosity level to infinity (i.e. log all messages, useful for debugging)">]>;
+        readonly offline: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 离线模式：强制使用缓存，阻止网络访问\n- Offline mode: forces use of cache, prevents network access  (env: LLAMA_OFFLINE)">]>;
+        readonly verbosity: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 设置日志的详细程度阈值。消息的详细程度越高，将被忽略。\n- Set the verbosity threshold. Messages with a higher verbosity will be ignored.  (env: LLAMA_LOG_VERBOSITY)">]>;
+        readonly 'log-prefix': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 启用日志消息中的前缀\n- Enable prefix in log messages  (env: LLAMA_LOG_PREFIX)">]>;
+        readonly 'log-timestamps': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 启用日志消息中的时间戳\n- Enable timestamps in log messages  (env: LLAMA_LOG_TIMESTAMPS)">]>;
+        readonly 'cache-type-k-draft': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "TYPE">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- KV缓存中用于草稿模型的K的数据类型，允许的值包括：f32、f16、bf16、q8_0、q4_0、q4_1、iq4_nl、q5_0、q5_1（默认值：f16）\n- KV cache data type for K for the draft model  allowed values: f32, f16, bf16, q8_0, q4_0, q4_1, iq4_nl, q5_0, q5_1  (default: f16)  (env: LLAMA_ARG_CACHE_TYPE_K_DRAFT)">]>;
+        readonly 'cache-type-v-draft': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "TYPE">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- KV缓存中用于草稿模型的V的数据类型，允许的值包括：f32、f16、bf16、q8_0、q4_0、q4_1、iq4_nl、q5_0、q5_1（默认值：f16）\n- KV cache data type for V for the draft model  allowed values: f32, f16, bf16, q8_0, q4_0, q4_1, iq4_nl, q5_0, q5_1  (default: f16)  (env: LLAMA_ARG_CACHE_TYPE_V_DRAFT)">]>;
+    }, undefined>, undefined>;
+    readonly exampleSpecific: v.OptionalSchema<v.ObjectSchema<{
+        readonly 'swa-checkpoints': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 每个时间槽内可创建的 SWA 检查点最大数量（默认值：3）：[更多信息](https://github.com/ggml-org/llama.cpp/pull/15293)\n- max number of SWA checkpoints per slot to create (default: 3)  [(more info)](https://github.com/ggml-org/llama.cpp/pull/15293)  (env: LLAMA_ARG_SWA_CHECKPOINTS)">]>;
+        readonly 'no-context-shift': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 禁用无限文本生成中的上下文转换（默认：禁用）\n- disables context shift on infinite text generation (default: enabled)  (env: LLAMA_ARG_NO_CONTEXT_SHIFT)">]>;
+        readonly 'context-shift': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 该功能允许在无限文本生成过程中实现上下文切换（默认设置为禁用状态）。\n- enables context shift on infinite text generation (default: disabled)  (env: LLAMA_ARG_CONTEXT_SHIFT)">]>;
+        readonly 'reverse-prompt': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "PROMPT">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 在 PROMPT 状态下停止程序的执行，然后以交互模式恢复程序的控制权。\n- halt generation at PROMPT, return control in interactive mode">]>;
+        readonly special: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 特殊标记输出启用（默认：false）\n- special tokens output enabled (default: false)">]>;
+        readonly 'no-warmup': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- skip warming up the model with an empty run\n- skip warming up the model with an empty run">]>;
+        readonly 'spm-infill': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 使用后缀/前缀/中间模式进行填充（代替前缀/后缀/中间）\n- use Suffix/Prefix/Middle pattern for infill (instead of Prefix/Suffix/Middle) as some models prefer this. (default: disabled)">]>;
+        readonly pooling: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.PicklistSchema<["none", "mean", "cls", "last", "rank"], undefined>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: ["none" | "mean" | "cls" | "last" | "rank"] | undefined;
+        } | undefined, "- 嵌入的池化类型，如果未指定则使用模型默认值\n- pooling type for embeddings, use model default if unspecified  (env: LLAMA_ARG_POOLING)">]>;
+        readonly 'cont-batching': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 启用连续批处理（又名动态批处理）（默认：启用）\n- enable continuous batching (a.k.a dynamic batching) (default: enabled)  (env: LLAMA_ARG_CONT_BATCHING)">]>;
+        readonly 'no-cont-batching': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- disable continuous batching\n- disable continuous batching  (env: LLAMA_ARG_NO_CONT_BATCHING)">]>;
+        readonly mmproj: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FILE">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 多模态投影文件路径。参见tools/mtmd/README.md 说明：如果使用-hf参数，可以省略此参数\n- path to a multimodal projector file. see tools/mtmd/README.md  note: if -hf is used, this argument can be omitted  (env: LLAMA_ARG_MMPROJ)">]>;
+        readonly 'mmproj-url': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "URL">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 多模态投影文件的URL。参见tools/mtmd/README.md\n- URL to a multimodal projector file. see tools/mtmd/README.md  (env: LLAMA_ARG_MMPROJ_URL)">]>;
+        readonly 'no-mmproj': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 显式禁用多模态投影器，适用于使用 -hf 时\n- explicitly disable multimodal projector, useful when using -hf  (env: LLAMA_ARG_NO_MMPROJ)">]>;
+        readonly 'no-mmproj-offload': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 不要将多模态投影器卸载到GPU\n- do not offload multimodal projector to GPU  (env: LLAMA_ARG_NO_MMPROJ_OFFLOAD)">]>;
+        readonly 'override-tensor-draft': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "<tensor name pattern>=<buffer type>,...">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 为草图模型覆盖张量缓冲区的类型\n- override tensor buffer type for draft model">]>;
+        readonly 'cpu-moe-draft': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 将所有专家组合（Mixture of Experts, MoE）的权重都存储在 CPU 中，以便用于模型的训练过程。\n- keep all Mixture of Experts (MoE) weights in the CPU for the draft model  (env: LLAMA_ARG_CPU_MOE_DRAFT)">]>;
+        readonly 'n-cpu-moe-draft': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 将前 N 层的专家组合（Expert Combination, MoE）权重存储在 CPU 中，以用于初步模型的训练。\n- keep the Mixture of Experts (MoE) weights of the first N layers in the CPU for the draft model  (env: LLAMA_ARG_N_CPU_MOE_DRAFT)">]>;
+        readonly alias: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "STRING">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 设置模型名称的别名（用于REST API）\n- set alias for model name (to be used by REST API)  (env: LLAMA_ARG_ALIAS)">]>;
+        readonly host: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "HOST">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 监听的IP地址，或以.sock结尾的UNIX套接字路径（默认：127.0.0.1）\n- ip address to listen, or bind to an UNIX socket if the address ends with .sock (default: 127.0.0.1)  (env: LLAMA_ARG_HOST)">]>;
+        readonly port: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "PORT">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 端口监听 (默认: 8080)\n- port to listen (default: 8080)  (env: LLAMA_ARG_PORT)">]>;
+        readonly path: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "PATH">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 路径以供静态文件服务（默认：）\n- path to serve static files from (default: )  (env: LLAMA_ARG_STATIC_PATH)">]>;
+        readonly 'api-prefix': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "PREFIX">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 服务器提供服务的路径前缀（不包括末尾的斜杠，默认:）。\n- prefix path the server serves from, without the trailing slash (default: )  (env: LLAMA_ARG_API_PREFIX)">]>;
+        readonly 'no-webui': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 禁用Web UI（默认：启用）\n- Disable the Web UI (default: enabled)  (env: LLAMA_ARG_NO_WEBUI)">]>;
+        readonly embedding: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 限制仅支持嵌入用例; 仅与专用嵌入模型一起使用 (默认: 禁用)\n- restrict to only support embedding use case; use only with dedicated embedding models (default: disabled)  (env: LLAMA_ARG_EMBEDDINGS)">]>;
+        readonly reranking: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 启用服务器上的重排序端点（默认：禁用）\n- enable reranking endpoint on server (default: disabled)  (env: LLAMA_ARG_RERANKING)">]>;
+        readonly 'api-key': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "KEY">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- API key to use for authentication (default: none)\n- API key to use for authentication (default: none)  (env: LLAMA_API_KEY)">]>;
+        readonly 'api-key-file': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FNAME">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 包含API密钥的文件路径（默认：无）\n- path to file containing API keys (default: none)">]>;
+        readonly 'ssl-key-file': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FNAME">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- PEM编码SSL私钥文件的路径\n- path to file a PEM-encoded SSL private key  (env: LLAMA_ARG_SSL_KEY_FILE)">]>;
+        readonly 'ssl-cert-file': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FNAME">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- PEM编码SSL证书文件路径\n- path to file a PEM-encoded SSL certificate  (env: LLAMA_ARG_SSL_CERT_FILE)">]>;
+        readonly 'chat-template-kwargs': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "STRING">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 为 JSON 模板解析器设置额外的参数\n- sets additional params for the json template parser  (env: LLAMA_CHAT_TEMPLATE_KWARGS)">]>;
+        readonly timeout: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 服务器读/写超时时间（秒），默认：600\n- server read/write timeout in seconds (default: 600)  (env: LLAMA_ARG_TIMEOUT)">]>;
+        readonly 'threads-http': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 用于处理HTTP请求的线程数（默认：-1）\n- number of threads used to process HTTP requests (default: -1)  (env: LLAMA_ARG_THREADS_HTTP)">]>;
+        readonly 'cache-reuse': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 最小块大小以尝试通过KV移动从缓存中重用 (默认: 0) [(card)](https://ggml.ai/f0.png)\n- min chunk size to attempt reusing from the cache via KV shifting (default: 0)  [(card)](https://ggml.ai/f0.png)  (env: LLAMA_ARG_CACHE_REUSE)">]>;
+        readonly metrics: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 启用 prometheus 兼容的 metrics 端点 (默认: 禁用)\n- enable prometheus compatible metrics endpoint (default: disabled)  (env: LLAMA_ARG_ENDPOINT_METRICS)">]>;
+        readonly props: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 启用通过POST /props更改全局属性（默认：禁用）\n- enable changing global properties via POST /props (default: disabled)  (env: LLAMA_ARG_ENDPOINT_PROPS)">]>;
+        readonly slots: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 启用插槽监控端点（默认：禁用）\n- enable slots monitoring endpoint (default: enabled)  (env: LLAMA_ARG_ENDPOINT_SLOTS)">]>;
+        readonly 'no-slots': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 禁用插槽监控端点\n- disables slots monitoring endpoint  (env: LLAMA_ARG_NO_ENDPOINT_SLOTS)">]>;
+        readonly 'slot-save-path': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "PATH">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 保存slot kv缓存的路径（默认：禁用）\n- path to save slot kv cache (default: disabled)">]>;
+        readonly jinja: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 使用jinja模板进行聊天（默认：禁用）\n- use jinja template for chat (default: disabled)  (env: LLAMA_ARG_JINJA)">]>;
+        readonly 'reasoning-format': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.PicklistSchema<["deepseek", "none"], undefined>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: ["none" | "deepseek"] | undefined;
+        } | undefined, "- 控制是否允许并/或从响应中提取思维标签，以及以何种格式返回；可选值包括：  \\n- none：将思维内容原样保留在 `message.content` 中  \\n- deepseek：将思维内容放入 `message.reasoning_content`（在流式模式下行为与 none 相同）  \\n（默认值：deepseek）\n- controls whether thought tags are allowed and/or extracted from the response, and in which format they're returned; one of:  - none: leaves thoughts unparsed in `message.content`  - deepseek: puts thoughts in `message.reasoning_content` (except in streaming mode, which behaves as `none`)  (default: auto)  (env: LLAMA_ARG_THINK)">]>;
+        readonly 'reasoning-budget': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 控制允许的思考量；目前只能选择以下选项之一：-1 表示无限制的思考预算，或 0 表示禁用思考（默认值：-1）\n- controls the amount of thinking allowed; currently only one of: -1 for unrestricted thinking budget, or 0 to disable thinking (default: -1)  (env: LLAMA_ARG_THINK_BUDGET)">]>;
+        readonly 'chat-template': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "JINJA_TEMPLATE">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 设置自定义的Jinja聊天模板（默认：从模型元数据中获取模板）如果指定了后缀/前缀，模板将被禁用 仅接受常用模板（除非在该标志之前设置了--jinja）：内置模板列表：bailing, chatglm3, chatglm4, chatml, command-r, deepseek, deepseek2, deepseek3, exaone3, falcon3, gemma, gigachat, glmedge, granite, llama2, llama2-sys, llama2-sys-bos, llama2-sys-strip, llama3, llama4, megrez, minicpm, mistral-v1, mistral-v3, mistral-v3-tekken, mistral-v7, mistral-v7-0, mistral-v7-tekken, monarch, openchat, orion, phi3, phi4, rwkv-world, smolvlm, vicuna, vicuna-orca, yandex, zephyr\n- set custom jinja chat template (default: template taken from model's metadata)  if suffix/prefix are specified, template will be disabled  only commonly used templates are accepted (unless --jinja is set before this flag):  list of built-in templates:  bailing, chatglm3, chatglm4, chatml, command-r, deepseek, deepseek2, deepseek3, exaone3, exaone4, falcon3, gemma, gigachat, glmedge, gpt-oss, granite, grok-2, hunyuan-dense, hunyuan-moe, kimi-k2, llama2, llama2-sys, llama2-sys-bos, llama2-sys-strip, llama3, llama4, megrez, minicpm, mistral-v1, mistral-v3, mistral-v3-tekken, mistral-v7, mistral-v7-tekken, monarch, openchat, orion, phi3, phi4, rwkv-world, seed_oss, smolvlm, vicuna, vicuna-orca, yandex, zephyr  (env: LLAMA_ARG_CHAT_TEMPLATE)">]>;
+        readonly 'chat-template-file': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "JINJA_TEMPLATE_FILE">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 设置自定义的Jinja聊天模板文件（默认：从模型元数据中获取模板）如果指定了后缀/前缀，模板将被禁用 仅接受常用模板（除非在该标志之前设置了--jinja）：内置模板列表：bailing, chatglm3, chatglm4, chatml, command-r, deepseek, deepseek2, deepseek3, exaone3, falcon3, gemma, gigachat, glmedge, granite, llama2, llama2-sys, llama2-sys-bos, llama2-sys-strip, llama3, llama4, megrez, minicpm, mistral-v1, mistral-v3, mistral-v3-tekken, mistral-v7, mistral-v7-0, mistral-v7-tekken, monarch, openchat, orion, phi3, phi4, rwkv-world, smolvlm, vicuna, vicuna-orca, yandex, zephyr\n- set custom jinja chat template file (default: template taken from model's metadata)  if suffix/prefix are specified, template will be disabled  only commonly used templates are accepted (unless --jinja is set before this flag):  list of built-in templates:  bailing, chatglm3, chatglm4, chatml, command-r, deepseek, deepseek2, deepseek3, exaone3, exaone4, falcon3, gemma, gigachat, glmedge, gpt-oss, granite, grok-2, hunyuan-dense, hunyuan-moe, kimi-k2, llama2, llama2-sys, llama2-sys-bos, llama2-sys-strip, llama3, llama4, megrez, minicpm, mistral-v1, mistral-v3, mistral-v3-tekken, mistral-v7, mistral-v7-tekken, monarch, openchat, orion, phi3, phi4, rwkv-world, seed_oss, smolvlm, vicuna, vicuna-orca, yandex, zephyr  (env: LLAMA_ARG_CHAT_TEMPLATE_FILE)">]>;
+        readonly 'no-prefill-assistant': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 是否在最后一条消息是助手消息时预填充助手的响应（默认：预填充启用）当设置此标志时，如果最后一条消息是助手消息，则将其视为完整消息且不预填充\n- whether to prefill the assistant's response if the last message is an assistant message (default: prefill enabled)  when this flag is set, if the last message is an assistant message then it will be treated as a full message and not prefilled    (env: LLAMA_ARG_NO_PREFILL_ASSISTANT)">]>;
+        readonly 'slot-prompt-similarity': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "SIMILARITY">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 提示必须与槽的提示匹配多少才能使用该槽（默认：0.50，0.0 = 禁用）\n- how much the prompt of a request must match the prompt of a slot in order to use that slot (default: 0.10, 0.0 = disabled)">]>;
+        readonly 'lora-init-without-apply': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 加载LoRA适配器而不应用它们（稍后通过POST /lora-adapters应用）（默认：禁用）\n- load LoRA adapters without applying them (apply later via POST /lora-adapters) (default: disabled)">]>;
+        readonly 'threads-draft': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 在生成过程中使用的线程数量（默认值：与 --threads 的值相同）\n- number of threads to use during generation (default: same as --threads)">]>;
+        readonly 'threads-batch-draft': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 在批量处理和即时处理过程中使用的线程数量（默认值：与 --threads-draft 的值相同）\n- number of threads to use during batch and prompt processing (default: same as --threads-draft)">]>;
+        readonly 'draft-max': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 用于推测解码的草稿令牌数量（默认：16）\n- number of tokens to draft for speculative decoding (default: 16)  (env: LLAMA_ARG_DRAFT_MAX)">]>;
+        readonly 'draft-min': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 最小的草稿标记数，用于推测解码（默认：0）\n- minimum number of draft tokens to use for speculative decoding (default: 0)  (env: LLAMA_ARG_DRAFT_MIN)">]>;
+        readonly 'draft-p-min': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "P">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 最小推测解码概率（贪心）（默认：0.8）\n- minimum speculative decoding probability (greedy) (default: 0.8)  (env: LLAMA_ARG_DRAFT_P_MIN)">]>;
+        readonly 'ctx-size-draft': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 提示上下文的大小用于草稿模型（默认：0，0 = 从模型加载）\n- size of the prompt context for the draft model (default: 0, 0 = loaded from model)  (env: LLAMA_ARG_CTX_SIZE_DRAFT)">]>;
+        readonly 'device-draft': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "<dev1,dev2,..>">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 逗号分隔的设备列表，用于卸载草稿模型（none = 不卸载）使用--list-devices查看可用设备列表\n- comma-separated list of devices to use for offloading the draft model (none = don't offload)  use --list-devices to see a list of available devices">]>;
+        readonly 'gpu-layers-draft': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 存储在VRAM中的图层数量用于草稿模型\n- number of layers to store in VRAM for the draft model  (env: LLAMA_ARG_N_GPU_LAYERS_DRAFT)">]>;
+        readonly 'model-draft': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FNAME">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 用于推测解码的草稿模型（默认：未使用）\n- draft model for speculative decoding (default: unused)  (env: LLAMA_ARG_MODEL_DRAFT)">]>;
+        readonly 'spec-replace': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "TARGET">]>, v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "DRAFT">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string, string] | undefined;
+        } | undefined, "- 如果草稿模型（Draft Model）与主模型（Main Model）不兼容，那么需要将目标字符串（Target String）转换为草稿格式（Draft Format）。\n- translate the string in TARGET into DRAFT if the draft model and main model are not compatible">]>;
+        readonly 'model-vocoder': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FNAME">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 语音合成模型用于音频生成（默认：未使用）\n- vocoder model for audio generation (default: unused)">]>;
+        readonly 'tts-use-guide-tokens': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 使用引导标记以提高TTS单词回忆\n- Use guide tokens to improve TTS word recall">]>;
+        readonly 'embd-bge-small-en-default': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 使用默认的bge-small-en-v1.5模型（注意：可以从互联网下载权重）\n- use default bge-small-en-v1.5 model (note: can download weights from the internet)">]>;
+        readonly 'embd-e5-small-en-default': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 使用默认的e5-small-v2模型（注意：可以从互联网下载权重）\n- use default e5-small-v2 model (note: can download weights from the internet)">]>;
+        readonly 'embd-gte-small-default': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 使用默认的gte-small模型（注意：可以从互联网下载权重）\n- use default gte-small model (note: can download weights from the internet)">]>;
+        readonly 'fim-qwen-1.5b-default': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 使用默认的Qwen 2.5 Coder 1.5B\n- use default Qwen 2.5 Coder 1.5B (note: can download weights from the internet)">]>;
+        readonly 'fim-qwen-3b-default': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 使用默认的Qwen 2.5 Coder 3B（注意：可以从互联网下载权重）\n- use default Qwen 2.5 Coder 3B (note: can download weights from the internet)">]>;
+        readonly 'fim-qwen-7b-default': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 使用默认的Qwen 2.5 Coder 7B（注意：可以从互联网下载权重）\n- use default Qwen 2.5 Coder 7B (note: can download weights from the internet)">]>;
+        readonly 'fim-qwen-7b-spec': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 使用Qwen 2.5 Coder 7B + 0.5B draft进行推测解码（注意：可以从互联网下载权重）\n- use Qwen 2.5 Coder 7B + 0.5B draft for speculative decoding (note: can download weights from the internet)">]>;
+        readonly 'fim-qwen-14b-spec': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 使用Qwen 2.5 Coder 14B + 0.5B draft进行推测解码（注意：可以从互联网下载权重）\n- use Qwen 2.5 Coder 14B + 0.5B draft for speculative decoding (note: can download weights from the internet)">]>;
+        readonly 'fim-qwen-30b-default': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 使用默认的 Qwen 3 Coder 30B A3B Instruct；请注意：可以从互联网上下载相应的权重数据。\n- use default Qwen 3 Coder 30B A3B Instruct (note: can download weights from the internet)">]>;
+    }, undefined>, undefined>;
+    readonly sampling: v.OptionalSchema<v.ObjectSchema<{
+        readonly samplers: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "SAMPLERS">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- samplers that will be used for generation in the order, separated by ';'  (default: penalties;dry;top_n_sigma;top_k;typ_p;top_p;min_p;xtc;temperature)\n- samplers that will be used for generation in the order, separated by ';'  (default: penalties;dry;top_n_sigma;top_k;typ_p;top_p;min_p;xtc;temperature)">]>;
+        readonly seed: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "SEED">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- RNG种子（默认：-1，使用随机种子）\n- RNG seed (default: -1, use random seed for -1)">]>;
+        readonly 'sampling-seq': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "SEQUENCE">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 简化的采样序列，用于采样器（默认：edskypmxt）\n- simplified sequence for samplers that will be used (default: edskypmxt)">]>;
+        readonly 'ignore-eos': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+        } | undefined, "- 忽略流结束标记并继续生成（隐含--logit-bias EOS-inf）\n- ignore end of stream token and continue generating (implies --logit-bias EOS-inf)">]>;
+        readonly temp: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- temperature (default: 0.8)\n- temperature (default: 0.8)">]>;
+        readonly 'top-k': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- top-k 采样 (默认: 40, 0 = 禁用)\n- top-k sampling (default: 40, 0 = disabled)">]>;
+        readonly 'top-p': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- top-p 采样（默认：0.9，1.0 = 禁用）\n- top-p sampling (default: 0.9, 1.0 = disabled)">]>;
+        readonly 'min-p': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- min-p 采样 (默认: 0.1, 0.0 = 禁用)\n- min-p sampling (default: 0.1, 0.0 = disabled)">]>;
+        readonly 'top-nsigma': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- Top-N-Sigma抽样方法（默认值：-1.0；当两个参数都为-1.0时，该方法被禁用）\n- top-n-sigma sampling (default: -1.0, -1.0 = disabled)">]>;
+        readonly 'xtc-probability': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- xtc probability (默认: 0.0, 0.0 = 禁用)\n- xtc probability (default: 0.0, 0.0 = disabled)">]>;
+        readonly 'xtc-threshold': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- xtc threshold (默认: 0.1, 1.0 = 禁用)\n- xtc threshold (default: 0.1, 1.0 = disabled)">]>;
+        readonly typical: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 局部典型抽样，参数 p（默认：1.0，1.0 = 禁用）\n- locally typical sampling, parameter p (default: 1.0, 1.0 = disabled)">]>;
+        readonly 'repeat-last-n': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 用于惩罚的最后n个token（默认：64，0 = 禁用，-1 = 上下文大小）\n- last n tokens to consider for penalize (default: 64, 0 = disabled, -1 = ctx_size)">]>;
+        readonly 'repeat-penalty': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 惩罚重复的token序列（默认：1.0，1.0 = 禁用）\n- penalize repeat sequence of tokens (default: 1.0, 1.0 = disabled)">]>;
+        readonly 'presence-penalty': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 重复 alpha 存在惩罚（默认：0.0，0.0 = 禁用）\n- repeat alpha presence penalty (default: 0.0, 0.0 = disabled)">]>;
+        readonly 'frequency-penalty': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 重复 alpha 频率惩罚（默认：0.0，0.0 = 禁用）\n- repeat alpha frequency penalty (default: 0.0, 0.0 = disabled)">]>;
+        readonly 'dry-multiplier': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 设置 DRY 采样乘数（默认：0.0，0.0 = 禁用）\n- set DRY sampling multiplier (default: 0.0, 0.0 = disabled)">]>;
+        readonly 'dry-base': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 设置 DRY 采样基础值 (默认: 1.75)\n- set DRY sampling base value (default: 1.75)">]>;
+        readonly 'dry-allowed-length': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 设置DRY采样的允许长度（默认：2）\n- set allowed length for DRY sampling (default: 2)">]>;
+        readonly 'dry-penalty-last-n': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 设置最后n个token的DRY惩罚（默认：-1，0 = 禁用，-1 = 上下文大小）\n- set DRY penalty for the last n tokens (default: -1, 0 = disable, -1 = context size)">]>;
+        readonly 'dry-sequence-breaker': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "STRING">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 添加序列破环器用于DRY抽样，清除默认的破环器（'\\n', ':', '\"', '*'）；使用 \"none\" 不使用任何序列破环器\n- add sequence breaker for DRY sampling, clearing out default breakers ('\\n', ':', '\"', '*') in the process; use \"none\" to not use any sequence breakers">]>;
+        readonly 'dynatemp-range': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 动态温度范围 (默认: 0.0, 0.0 = 禁用)\n- dynamic temperature range (default: 0.0, 0.0 = disabled)">]>;
+        readonly 'dynatemp-exp': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 动态温度指数（默认：1.0）\n- dynamic temperature exponent (default: 1.0)">]>;
+        readonly mirostat: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- 使用Mirostat抽样。如果使用了Top K、Nucleus和Locally Typical抽样器，则会被忽略。（默认：0，0=禁用，1=Mirostat，2=Mirostat 2.0）\n- use Mirostat sampling.  Top K, Nucleus and Locally Typical samplers are ignored if used.  (default: 0, 0 = disabled, 1 = Mirostat, 2 = Mirostat 2.0)">]>;
+        readonly 'mirostat-lr': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- Mirostat学习率，参数eta（默认：0.1）\n- Mirostat learning rate, parameter eta (default: 0.1)">]>;
+        readonly 'mirostat-ent': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [number] | undefined;
+        } | undefined, "- Mirostat目标熵，参数tau（默认：5.0）\n- Mirostat target entropy, parameter tau (default: 5.0)">]>;
+        readonly 'logit-bias': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "TOKEN_ID(+/-)BIAS">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 修改标记出现在完成中的可能性，例如：--logit-bias 15043+1 增加标记 ' Hello' 的可能性，或 --logit 15043-1 减少标记 ' Hello' 的可能性\n- modifies the likelihood of token appearing in the completion,  i.e. `--logit-bias 15043+1` to increase likelihood of token ' Hello',  or `--logit-bias 15043-1` to decrease likelihood of token ' Hello'">]>;
+        readonly grammar: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "GRAMMAR">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- BNF类似语法约束生成（参见grammars目录中的示例）（默认：''）\n- BNF-like grammar to constrain generations (see samples in grammars/ dir) (default: '')">]>;
+        readonly 'grammar-file': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FNAME">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 从文件读取语法\n- file to read grammar from">]>;
+        readonly 'json-schema': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "SCHEMA">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- JSON schema 用于约束生成（https://json-schema.org/），例如 `{}` 表示任何 JSON 对象。对于包含外部 $refs 的 schema，使用 --grammar + example/json_schema_to_grammar.py 代替\n- JSON schema to constrain generations (https://json-schema.org/), e.g. `{}` for any JSON object  For schemas w/ external $refs, use --grammar + example/json_schema_to_grammar.py instead">]>;
+        readonly 'json-schema-file': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+            readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FILE">]>], undefined>, undefined>]>;
+        }, undefined>, undefined>, v.DescriptionAction<{
+            enable: boolean;
+            value?: [string] | undefined;
+        } | undefined, "- 包含用于约束生成的JSON架构的文件（https://json-schema.org/），例如{}表示任何JSON对象  对于具有外部$refs的架构，使用--grammar + example/json_schema_to_grammar.py代替\n- File containing a JSON schema to constrain generations (https://json-schema.org/), e.g. `{}` for any JSON object  For schemas w/ external $refs, use --grammar + example/json_schema_to_grammar.py instead">]>;
+    }, undefined>, undefined>;
+}, undefined>;
+export type LlamaServerType = v.InferOutput<typeof LlamaServerDefine>;
+export declare const ExecDefine: v.ObjectSchema<{
+    readonly version: v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TitleAction<string, "llama.cpp 版本号">, v.DescriptionAction<string, "llama.cpp的版本号,可以从 https://github.com/ggerganov/llama.cpp/releases 获得">]>;
+    readonly device: v.SchemaWithPipe<readonly [v.PicklistSchema<["cpu", "cuda12.4", "hip-radeon", "sycl", "vulkan"], undefined>, v.TitleAction<"cpu" | "cuda12.4" | "hip-radeon" | "sycl" | "vulkan", "llama.cpp 使用设备">]>;
+}, undefined>;
+export declare const LlamaServerSwapItemDefine: v.ObjectSchema<{
+    readonly config: v.OptionalSchema<v.ObjectSchema<{
+        readonly common: v.OptionalSchema<v.ObjectSchema<{
+            readonly 'verbose-prompt': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 在生成前打印详细提示（默认：false）\n- print a verbose prompt before generation (default: false)">]>;
+            readonly threads: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 生成期间使用的线程数（默认：-1）\n- number of threads to use during generation (default: -1)  (env: LLAMA_ARG_THREADS)">]>;
+            readonly 'threads-batch': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 批量和提示处理期间使用的线程数（默认：与--threads相同）\n- number of threads to use during batch and prompt processing (default: same as --threads)">]>;
+            readonly 'cpu-mask': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "M">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- CPU亲和力掩码：任意长度的十六进制数。补充cpu-range（默认：\"\"）\n- CPU affinity mask: arbitrarily long hex. Complements cpu-range (default: \"\")">]>;
+            readonly 'cpu-range': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "lo-hi">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- CPU范围用于亲和力。补充--cpu-mask\n- range of CPUs for affinity. Complements --cpu-mask">]>;
+            readonly 'cpu-strict': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.PicklistSchema<["0", "1"], undefined>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: ["0" | "1"] | undefined;
+            } | undefined, "- 使用严格CPU放置（默认：0）\n- use strict CPU placement (default: 0)">]>;
+            readonly prio: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.PicklistSchema<[0, 1, 2, 3], undefined>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [0 | 2 | 1 | 3] | undefined;
+            } | undefined, "- 设置进程/线程优先级：低(-1)、正常(0)、中等(1)、高(2)、实时(3)（默认值：0）\n- set process/thread priority : low(-1), normal(0), medium(1), high(2), realtime(3) (default: 0)">]>;
+            readonly poll: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "<0...100>">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 使用轮询级别等待工作 (0 - 无轮询，默认: 50)\n- use polling level to wait for work (0 - no polling, default: 50)">]>;
+            readonly 'cpu-mask-batch': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "M">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- CPU亲和力掩码：任意长度的十六进制数。补充cpu-range-batch（默认：与--cpu-mask相同）\n- CPU affinity mask: arbitrarily long hex. Complements cpu-range-batch (default: same as --cpu-mask)">]>;
+            readonly 'cpu-range-batch': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "lo-hi">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- CPU亲和力的范围。补充--cpu-mask-batch\n- ranges of CPUs for affinity. Complements --cpu-mask-batch">]>;
+            readonly 'cpu-strict-batch': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.PicklistSchema<["0", "1"], undefined>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: ["0" | "1"] | undefined;
+            } | undefined, "- 使用严格的CPU放置（默认：与--cpu-strict相同）\n- use strict CPU placement (default: same as --cpu-strict)">]>;
+            readonly 'prio-batch': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.PicklistSchema<[0, 1, 2, 3], undefined>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [0 | 2 | 1 | 3] | undefined;
+            } | undefined, "- 设置进程/线程优先级 : 0-正常, 1-中等, 2-高, 3-实时 (默认: 0)\n- set process/thread priority : 0-normal, 1-medium, 2-high, 3-realtime (default: 0)">]>;
+            readonly 'poll-batch': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.PicklistSchema<["0", "1"], undefined>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: ["0" | "1"] | undefined;
+            } | undefined, "- 使用轮询等待工作（默认：与--poll相同）\n- use polling to wait for work (default: same as --poll)">]>;
+            readonly 'ctx-size': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 提示上下文的大小（默认：4096，0 = 从模型加载）\n- size of the prompt context (default: 4096, 0 = loaded from model)  (env: LLAMA_ARG_CTX_SIZE)">]>;
+            readonly predict: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 预测的token数量（默认：-1，-1 = 无限）\n- number of tokens to predict (default: -1, -1 = infinity)  (env: LLAMA_ARG_N_PREDICT)">]>;
+            readonly 'batch-size': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 逻辑最大批处理大小（默认：2048）\n- logical maximum batch size (default: 2048)  (env: LLAMA_ARG_BATCH)">]>;
+            readonly 'ubatch-size': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 物理最大批处理大小（默认：512）\n- physical maximum batch size (default: 512)  (env: LLAMA_ARG_UBATCH)">]>;
+            readonly keep: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 保留初始提示中的令牌数量（默认：0，-1 = 所有）\n- number of tokens to keep from the initial prompt (default: 0, -1 = all)">]>;
+            readonly 'swa-full': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 使用全尺寸SWA缓存（默认：false） [更多信息](https://github.com/ggml-org/llama.cpp/pull/13194#issuecomment-2868343055)\n- use full-size SWA cache (default: false)  [(more info)](https://github.com/ggml-org/llama.cpp/pull/13194#issuecomment-2868343055)  (env: LLAMA_ARG_SWA_FULL)">]>;
+            readonly 'kv-unified': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 为所有序列的 KV 缓存使用一个统一的 KV 缓冲区（默认值为 `false`）：[更多信息](https://github.com/ggml-org/llama.cpp/pull/14363)（环境变量：`LLAMA_ARG_KV_SPLIT`）\n- use single unified KV buffer for the KV cache of all sequences (default: false)  [(more info)](https://github.com/ggml-org/llama.cpp/pull/14363)  (env: LLAMA_ARG_KV_SPLIT)">]>;
+            readonly 'flash-attn': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.PicklistSchema<["on", "off", "auto"], undefined>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: ["on" | "off" | "auto"] | undefined;
+            } | undefined, "- 启用 Flash 注意力（默认：禁用）\n- set Flash Attention use ('on', 'off', or 'auto', default: 'auto')  (env: LLAMA_ARG_FLASH_ATTN)">]>;
+            readonly 'no-perf': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 禁用内部libllama性能计时（默认：false）\n- disable internal libllama performance timings (default: false)  (env: LLAMA_ARG_NO_PERF)">]>;
+            readonly escape: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 处理转义序列（\\n, \\r, \\t, \\\", \\\\)（默认：true）\n- process escapes sequences (\\n, \\r, \\t, \\', \\\", \\\\) (default: true)">]>;
+            readonly 'no-escape': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 不要处理转义序列\n- do not process escape sequences">]>;
+            readonly 'rope-scaling': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.PicklistSchema<["none", "linear", "yarn"], undefined>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: ["none" | "linear" | "yarn"] | undefined;
+            } | undefined, "- RoPE频率缩放方法，除非模型指定，默认为线性\n- RoPE frequency scaling method, defaults to linear unless specified by the model  (env: LLAMA_ARG_ROPE_SCALING_TYPE)">]>;
+            readonly 'rope-scale': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- RoPE上下文缩放因子，将上下文扩展N倍\n- RoPE context scaling factor, expands context by a factor of N  (env: LLAMA_ARG_ROPE_SCALE)">]>;
+            readonly 'rope-freq-base': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- RoPE基频，用于NTK感知缩放（默认：从模型加载）\n- RoPE base frequency, used by NTK-aware scaling (default: loaded from model)  (env: LLAMA_ARG_ROPE_FREQ_BASE)">]>;
+            readonly 'rope-freq-scale': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- RoPE频率缩放因子，通过因子1/N扩展上下文\n- RoPE frequency scaling factor, expands context by a factor of 1/N  (env: LLAMA_ARG_ROPE_FREQ_SCALE)">]>;
+            readonly 'yarn-orig-ctx': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- YaRN: 原始上下文大小（默认：0 = 模型训练上下文大小）\n- YaRN: original context size of model (default: 0 = model training context size)  (env: LLAMA_ARG_YARN_ORIG_CTX)">]>;
+            readonly 'yarn-ext-factor': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- YaRN: 插值混合因子（默认：-1.0，0.0 = 完全插值）\n- YaRN: extrapolation mix factor (default: -1.0, 0.0 = full interpolation)  (env: LLAMA_ARG_YARN_EXT_FACTOR)">]>;
+            readonly 'yarn-attn-factor': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- YaRN: 缩放 sqrt(t) 或注意力幅度（默认：-1.0）\n- YaRN: scale sqrt(t) or attention magnitude (default: -1.0)  (env: LLAMA_ARG_YARN_ATTN_FACTOR)">]>;
+            readonly 'yarn-beta-slow': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- YaRN: 高校正维度或alpha（默认：-1.0）\n- YaRN: high correction dim or alpha (default: -1.0)  (env: LLAMA_ARG_YARN_BETA_SLOW)">]>;
+            readonly 'yarn-beta-fast': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- YaRN: 低校正维或beta（默认：-1）\n- YaRN: low correction dim or beta (default: -1.0)  (env: LLAMA_ARG_YARN_BETA_FAST)">]>;
+            readonly 'no-kv-offload': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 禁用KV卸载\n- disable KV offload  (env: LLAMA_ARG_NO_KV_OFFLOAD)">]>;
+            readonly 'no-repack': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 禁用重量重新打包功能\n- disable weight repacking  (env: LLAMA_ARG_NO_REPACK)">]>;
+            readonly 'cache-type-k': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.PicklistSchema<["f32", "f16", "bf16", "q8_0", "q4_0", "q4_1", "iq4_nl", "q5_0", "q5_1"], undefined>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: ["f32" | "f16" | "bf16" | "q8_0" | "q4_0" | "q4_1" | "iq4_nl" | "q5_0" | "q5_1"] | undefined;
+            } | undefined, "- KV缓存数据类型用于K 允许的值: f32, f16, bf16, q8_0, q4_0, q4_1, iq4_nl, q5_0, q5_1 （默认: f16）\n- KV cache data type for K  allowed values: f32, f16, bf16, q8_0, q4_0, q4_1, iq4_nl, q5_0, q5_1  (default: f16)  (env: LLAMA_ARG_CACHE_TYPE_K)">]>;
+            readonly 'cache-type-v': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.PicklistSchema<["f32", "f16", "bf16", "q8_0", "q4_0", "q4_1", "iq4_nl", "q5_0", "q5_1"], undefined>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: ["f32" | "f16" | "bf16" | "q8_0" | "q4_0" | "q4_1" | "iq4_nl" | "q5_0" | "q5_1"] | undefined;
+            } | undefined, "- KV缓存数据类型用于V，允许的值：f32, f16, bf16, q8_0, q4_0, q4_1, iq4_nl, q5_0, q5_1（默认：f16）\n- KV cache data type for V  allowed values: f32, f16, bf16, q8_0, q4_0, q4_1, iq4_nl, q5_0, q5_1  (default: f16)  (env: LLAMA_ARG_CACHE_TYPE_V)">]>;
+            readonly 'defrag-thold': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- KV缓存碎片化阈值（默认：0.1，<0-禁用）\n- KV cache defragmentation threshold (DEPRECATED)  (env: LLAMA_ARG_DEFRAG_THOLD)">]>;
+            readonly parallel: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 并行解码序列的数量（默认：1）\n- number of parallel sequences to decode (default: 1)  (env: LLAMA_ARG_N_PARALLEL)">]>;
+            readonly mlock: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 强制系统将模型保留在RAM中而不是交换或压缩\n- force system to keep model in RAM rather than swapping or compressing  (env: LLAMA_ARG_MLOCK)">]>;
+            readonly 'no-mmap': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 不要内存映射模型（加载速度较慢，但可能减少页面丢失，如果未使用mlock）\n- do not memory-map model (slower load but may reduce pageouts if not using mlock)  (env: LLAMA_ARG_NO_MMAP)">]>;
+            readonly numa: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.PicklistSchema<["distribute", "isolate", "numactl"], undefined>, v.MetadataAction<"distribute" | "isolate" | "numactl", {
+                    readonly enumOptions: readonly [{
+                        readonly label: "distribute";
+                        readonly value: "distribute";
+                        readonly description: "将执行均匀分布在所有节点上";
+                    }, {
+                        readonly label: "isolate";
+                        readonly value: "isolate";
+                        readonly description: "仅在执行开始的节点的CPU上生成线程";
+                    }, {
+                        readonly label: "numactl";
+                        readonly value: "numactl";
+                        readonly description: "使用numactl提供的CPU映射";
+                    }];
+                }>]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: ["distribute" | "isolate" | "numactl"] | undefined;
+            } | undefined, "- 尝试在某些NUMA系统上进行优化 - distribute：将执行均匀分布在所有节点上 - isolate：仅在执行开始的节点的CPU上生成线程 - numactl：使用numactl提供的CPU映射 如果之前未运行过此操作，请在使用此功能前删除系统页面缓存 详见 https://github.com/ggml-org/llama.cpp/issues/1437\n- attempt optimizations that help on some NUMA systems  - distribute: spread execution evenly over all nodes  - isolate: only spawn threads on CPUs on the node that execution started on  - numactl: use the CPU map provided by numactl  if run without this previously, it is recommended to drop the system page cache before using this  see https://github.com/ggml-org/llama.cpp/issues/1437  (env: LLAMA_ARG_NUMA)">]>;
+            readonly device: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "<dev1,dev2,..>">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 逗号分隔的设备列表，用于卸载（none = 不卸载）使用 --list-devices 查看可用设备列表\n- comma-separated list of devices to use for offloading (none = don't offload)  use --list-devices to see a list of available devices  (env: LLAMA_ARG_DEVICE)">]>;
+            readonly 'override-tensor': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "<tensor name pattern>=<buffer type>,...">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 将所有专家组合（Mixture of Experts, MoE）的权重数据存储在 CPU 中。\n- override tensor buffer type">]>;
+            readonly 'cpu-moe': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 将前 N 层的“专家混合”（Expert Mixing, MoE）权重数据保留在 CPU 中。\n- keep all Mixture of Experts (MoE) weights in the CPU  (env: LLAMA_ARG_CPU_MOE)">]>;
+            readonly 'n-cpu-moe': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 覆盖张量缓冲区类型\n- keep the Mixture of Experts (MoE) weights of the first N layers in the CPU  (env: LLAMA_ARG_N_CPU_MOE)">]>;
+            readonly 'gpu-layers': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 存储在VRAM中的图层数量\n- max. number of layers to store in VRAM (default: -1)  (env: LLAMA_ARG_N_GPU_LAYERS)">]>;
+            readonly 'split-mode': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.PicklistSchema<["none", "layer", "row"], undefined>, v.MetadataAction<"none" | "layer" | "row", {
+                    readonly enumOptions: readonly [{
+                        readonly label: "none";
+                        readonly value: "none";
+                        readonly description: "仅使用一个GPU";
+                    }, {
+                        readonly label: "layer";
+                        readonly value: "layer";
+                        readonly description: "分割层和KV跨GPU";
+                    }, {
+                        readonly label: "row";
+                        readonly value: "row";
+                        readonly description: "分割行跨GPU";
+                    }];
+                }>]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: ["none" | "layer" | "row"] | undefined;
+            } | undefined, "- 如何将模型分布在多个GPU上，可选：- none: 仅使用一个GPU - layer (默认): 分割层和KV跨GPU - row: 分割行跨GPU\n- how to split the model across multiple GPUs, one of:  - none: use one GPU only  - layer (default): split layers and KV across GPUs  - row: split rows across GPUs  (env: LLAMA_ARG_SPLIT_MODE)">]>;
+            readonly 'tensor-split': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "N0,N1,N2,...">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 将模型分片到每个GPU的分数，逗号分隔的比例列表，例如3,1\n- fraction of the model to offload to each GPU, comma-separated list of proportions, e.g. 3,1  (env: LLAMA_ARG_TENSOR_SPLIT)">]>;
+            readonly 'main-gpu': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "INDEX">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 用于模型（split-mode = none）或中间结果和KV（split-mode = row）的GPU（默认：0）\n- the GPU to use for the model (with split-mode = none), or for intermediate results and KV (with split-mode = row) (default: 0)  (env: LLAMA_ARG_MAIN_GPU)">]>;
+            readonly 'check-tensors': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 检查模型张量数据中的无效值（默认：false）\n- check model tensor data for invalid values (default: false)">]>;
+            readonly 'override-kv': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "KEY=TYPE:VALUE">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 用于通过键覆盖模型元数据的高级选项。可以多次指定。类型：int、float、bool、str。示例：--override-kv tokenizer.ggml.add_bos_token=bool:false\n- advanced option to override model metadata by key. may be specified multiple times.  types: int, float, bool, str. example: --override-kv tokenizer.ggml.add_bos_token=bool:false">]>;
+            readonly 'no-op-offload': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 禁用将主机张量操作卸载到设备（默认：false）\n- disable offloading host tensor operations to device (default: false)">]>;
+            readonly lora: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FNAME">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- LoRA适配器路径（可以重复使用多个适配器）\n- path to LoRA adapter (can be repeated to use multiple adapters)">]>;
+            readonly 'lora-scaled': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FNAME">]>, v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "SCALE">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string, number] | undefined;
+            } | undefined, "- LoRA适配器路径，带用户定义的缩放（可重复使用多个适配器）\n- path to LoRA adapter with user defined scaling (can be repeated to use multiple adapters)">]>;
+            readonly 'control-vector': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FNAME">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 添加控制向量  请注意：此参数可以重复以添加多个控制向量\n- add a control vector  note: this argument can be repeated to add multiple control vectors">]>;
+            readonly 'control-vector-scaled': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FNAME">]>, v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "SCALE">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string, number] | undefined;
+            } | undefined, "- 添加一个带用户定义缩放 SCALE 的控制向量  注意: 此参数可以重复以添加多个缩放的控制向量\n- add a control vector with user defined scaling SCALE  note: this argument can be repeated to add multiple scaled control vectors">]>;
+            readonly 'control-vector-layer-range': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "START">]>, v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "END">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string, string] | undefined;
+            } | undefined, "- 应用控制向量的图层范围，起始和结束值（包含）\n- layer range to apply the control vector(s) to, start and end inclusive">]>;
+            readonly model: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FNAME">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 模型路径 (默认: `models/$filename` 与 `--hf-file` 或 `--model-url` 的文件名，否则 models/7B/ggml-model-f16.gguf)\n- model path (default: `models/$filename` with filename from `--hf-file` or `--model-url` if set, otherwise models/7B/ggml-model-f16.gguf)  (env: LLAMA_ARG_MODEL)">]>;
+            readonly 'model-url': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "MODEL_URL">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- Docker Hub 模型仓库：该仓库是可选的，默认值为 `ai/`；`quant` 参数也是可选的，默认值为 `:latest`。例如：`gemma3`（默认值：未使用）。\n- model download url (default: unused)  (env: LLAMA_ARG_MODEL_URL)">]>;
+            readonly 'docker-repo': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.PicklistSchema<["<repo>/]<model>[:quant"], undefined>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: ["<repo>/]<model>[:quant"] | undefined;
+            } | undefined, "- 模型下载地址（默认：未使用）\n- Docker Hub model repository. repo is optional, default to ai/. quant is optional, default to :latest.  example: gemma3  (default: unused)  (env: LLAMA_ARG_DOCKER_REPO)">]>;
+            readonly 'hf-repo': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "<user>/<model>[:quant]">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- Hugging Face 模型仓库; quant 是可选的，不区分大小写，默认为 Q4_K_M，如果不存在 Q4_K_M 则回退到仓库中的第一个文件。如果可用还会自动下载 mmproj。要禁用，添加 --no-mmproj  示例：unsloth/phi-4-GGUF:q4_k_m  （默认：未使用）\n- Hugging Face model repository; quant is optional, case-insensitive, default to Q4_K_M, or falls back to the first file in the repo if Q4_K_M doesn't exist.  mmproj is also downloaded automatically if available. to disable, add --no-mmproj  example: unsloth/phi-4-GGUF:q4_k_m  (default: unused)  (env: LLAMA_ARG_HF_REPO)">]>;
+            readonly 'hf-repo-draft': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "<user>/<model>[:quant]">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 与--hf-repo相同，但用于草稿模型（默认：未使用）\n- Same as --hf-repo, but for the draft model (default: unused)  (env: LLAMA_ARG_HFD_REPO)">]>;
+            readonly 'hf-file': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FILE">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- Hugging Face模型文件。如果指定了，将覆盖--hf-repo中的量化（默认：未使用）\n- Hugging Face model file. If specified, it will override the quant in --hf-repo (default: unused)  (env: LLAMA_ARG_HF_FILE)">]>;
+            readonly 'hf-repo-v': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "<user>/<model>[:quant]">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- Hugging Face vocoder模型仓库（默认：未使用）\n- Hugging Face model repository for the vocoder model (default: unused)  (env: LLAMA_ARG_HF_REPO_V)">]>;
+            readonly 'hf-file-v': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FILE">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- Hugging Face模型文件用于声码器模型（默认：未使用）\n- Hugging Face model file for the vocoder model (default: unused)  (env: LLAMA_ARG_HF_FILE_V)">]>;
+            readonly 'hf-token': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "TOKEN">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- Hugging Face访问令牌（默认：来自HF_TOKEN环境变量的值）\n- Hugging Face access token (default: value from HF_TOKEN environment variable)  (env: HF_TOKEN)">]>;
+            readonly 'log-disable': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 日志禁用\n- Log disable">]>;
+            readonly 'log-file': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FNAME">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 日志记录到文件\n- Log to file">]>;
+            readonly 'log-colors': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.PicklistSchema<["on", "off", "auto"], undefined>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: ["on" | "off" | "auto"] | undefined;
+            } | undefined, "- 启用彩色日志记录\n- Set colored logging ('on', 'off', or 'auto', default: 'auto')  'auto' enables colors when output is to a terminal  (env: LLAMA_LOG_COLORS)">]>;
+            readonly verbose: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 设置 verbosity 级别为无穷大（即记录所有消息，用于调试）\n- Set verbosity level to infinity (i.e. log all messages, useful for debugging)">]>;
+            readonly offline: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 离线模式：强制使用缓存，阻止网络访问\n- Offline mode: forces use of cache, prevents network access  (env: LLAMA_OFFLINE)">]>;
+            readonly verbosity: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 设置日志的详细程度阈值。消息的详细程度越高，将被忽略。\n- Set the verbosity threshold. Messages with a higher verbosity will be ignored.  (env: LLAMA_LOG_VERBOSITY)">]>;
+            readonly 'log-prefix': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 启用日志消息中的前缀\n- Enable prefix in log messages  (env: LLAMA_LOG_PREFIX)">]>;
+            readonly 'log-timestamps': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 启用日志消息中的时间戳\n- Enable timestamps in log messages  (env: LLAMA_LOG_TIMESTAMPS)">]>;
+            readonly 'cache-type-k-draft': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "TYPE">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- KV缓存中用于草稿模型的K的数据类型，允许的值包括：f32、f16、bf16、q8_0、q4_0、q4_1、iq4_nl、q5_0、q5_1（默认值：f16）\n- KV cache data type for K for the draft model  allowed values: f32, f16, bf16, q8_0, q4_0, q4_1, iq4_nl, q5_0, q5_1  (default: f16)  (env: LLAMA_ARG_CACHE_TYPE_K_DRAFT)">]>;
+            readonly 'cache-type-v-draft': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "TYPE">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- KV缓存中用于草稿模型的V的数据类型，允许的值包括：f32、f16、bf16、q8_0、q4_0、q4_1、iq4_nl、q5_0、q5_1（默认值：f16）\n- KV cache data type for V for the draft model  allowed values: f32, f16, bf16, q8_0, q4_0, q4_1, iq4_nl, q5_0, q5_1  (default: f16)  (env: LLAMA_ARG_CACHE_TYPE_V_DRAFT)">]>;
+        }, undefined>, undefined>;
+        readonly exampleSpecific: v.OptionalSchema<v.ObjectSchema<{
+            readonly 'swa-checkpoints': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 每个时间槽内可创建的 SWA 检查点最大数量（默认值：3）：[更多信息](https://github.com/ggml-org/llama.cpp/pull/15293)\n- max number of SWA checkpoints per slot to create (default: 3)  [(more info)](https://github.com/ggml-org/llama.cpp/pull/15293)  (env: LLAMA_ARG_SWA_CHECKPOINTS)">]>;
+            readonly 'no-context-shift': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 禁用无限文本生成中的上下文转换（默认：禁用）\n- disables context shift on infinite text generation (default: enabled)  (env: LLAMA_ARG_NO_CONTEXT_SHIFT)">]>;
+            readonly 'context-shift': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 该功能允许在无限文本生成过程中实现上下文切换（默认设置为禁用状态）。\n- enables context shift on infinite text generation (default: disabled)  (env: LLAMA_ARG_CONTEXT_SHIFT)">]>;
+            readonly 'reverse-prompt': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "PROMPT">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 在 PROMPT 状态下停止程序的执行，然后以交互模式恢复程序的控制权。\n- halt generation at PROMPT, return control in interactive mode">]>;
+            readonly special: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 特殊标记输出启用（默认：false）\n- special tokens output enabled (default: false)">]>;
+            readonly 'no-warmup': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- skip warming up the model with an empty run\n- skip warming up the model with an empty run">]>;
+            readonly 'spm-infill': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 使用后缀/前缀/中间模式进行填充（代替前缀/后缀/中间）\n- use Suffix/Prefix/Middle pattern for infill (instead of Prefix/Suffix/Middle) as some models prefer this. (default: disabled)">]>;
+            readonly pooling: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.PicklistSchema<["none", "mean", "cls", "last", "rank"], undefined>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: ["none" | "mean" | "cls" | "last" | "rank"] | undefined;
+            } | undefined, "- 嵌入的池化类型，如果未指定则使用模型默认值\n- pooling type for embeddings, use model default if unspecified  (env: LLAMA_ARG_POOLING)">]>;
+            readonly 'cont-batching': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 启用连续批处理（又名动态批处理）（默认：启用）\n- enable continuous batching (a.k.a dynamic batching) (default: enabled)  (env: LLAMA_ARG_CONT_BATCHING)">]>;
+            readonly 'no-cont-batching': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- disable continuous batching\n- disable continuous batching  (env: LLAMA_ARG_NO_CONT_BATCHING)">]>;
+            readonly mmproj: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FILE">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 多模态投影文件路径。参见tools/mtmd/README.md 说明：如果使用-hf参数，可以省略此参数\n- path to a multimodal projector file. see tools/mtmd/README.md  note: if -hf is used, this argument can be omitted  (env: LLAMA_ARG_MMPROJ)">]>;
+            readonly 'mmproj-url': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "URL">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 多模态投影文件的URL。参见tools/mtmd/README.md\n- URL to a multimodal projector file. see tools/mtmd/README.md  (env: LLAMA_ARG_MMPROJ_URL)">]>;
+            readonly 'no-mmproj': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 显式禁用多模态投影器，适用于使用 -hf 时\n- explicitly disable multimodal projector, useful when using -hf  (env: LLAMA_ARG_NO_MMPROJ)">]>;
+            readonly 'no-mmproj-offload': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 不要将多模态投影器卸载到GPU\n- do not offload multimodal projector to GPU  (env: LLAMA_ARG_NO_MMPROJ_OFFLOAD)">]>;
+            readonly 'override-tensor-draft': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "<tensor name pattern>=<buffer type>,...">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 为草图模型覆盖张量缓冲区的类型\n- override tensor buffer type for draft model">]>;
+            readonly 'cpu-moe-draft': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 将所有专家组合（Mixture of Experts, MoE）的权重都存储在 CPU 中，以便用于模型的训练过程。\n- keep all Mixture of Experts (MoE) weights in the CPU for the draft model  (env: LLAMA_ARG_CPU_MOE_DRAFT)">]>;
+            readonly 'n-cpu-moe-draft': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 将前 N 层的专家组合（Expert Combination, MoE）权重存储在 CPU 中，以用于初步模型的训练。\n- keep the Mixture of Experts (MoE) weights of the first N layers in the CPU for the draft model  (env: LLAMA_ARG_N_CPU_MOE_DRAFT)">]>;
+            readonly alias: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "STRING">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 设置模型名称的别名（用于REST API）\n- set alias for model name (to be used by REST API)  (env: LLAMA_ARG_ALIAS)">]>;
+            readonly host: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "HOST">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 监听的IP地址，或以.sock结尾的UNIX套接字路径（默认：127.0.0.1）\n- ip address to listen, or bind to an UNIX socket if the address ends with .sock (default: 127.0.0.1)  (env: LLAMA_ARG_HOST)">]>;
+            readonly port: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "PORT">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 端口监听 (默认: 8080)\n- port to listen (default: 8080)  (env: LLAMA_ARG_PORT)">]>;
+            readonly path: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "PATH">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 路径以供静态文件服务（默认：）\n- path to serve static files from (default: )  (env: LLAMA_ARG_STATIC_PATH)">]>;
+            readonly 'api-prefix': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "PREFIX">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 服务器提供服务的路径前缀（不包括末尾的斜杠，默认:）。\n- prefix path the server serves from, without the trailing slash (default: )  (env: LLAMA_ARG_API_PREFIX)">]>;
+            readonly 'no-webui': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 禁用Web UI（默认：启用）\n- Disable the Web UI (default: enabled)  (env: LLAMA_ARG_NO_WEBUI)">]>;
+            readonly embedding: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 限制仅支持嵌入用例; 仅与专用嵌入模型一起使用 (默认: 禁用)\n- restrict to only support embedding use case; use only with dedicated embedding models (default: disabled)  (env: LLAMA_ARG_EMBEDDINGS)">]>;
+            readonly reranking: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 启用服务器上的重排序端点（默认：禁用）\n- enable reranking endpoint on server (default: disabled)  (env: LLAMA_ARG_RERANKING)">]>;
+            readonly 'api-key': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "KEY">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- API key to use for authentication (default: none)\n- API key to use for authentication (default: none)  (env: LLAMA_API_KEY)">]>;
+            readonly 'api-key-file': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FNAME">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 包含API密钥的文件路径（默认：无）\n- path to file containing API keys (default: none)">]>;
+            readonly 'ssl-key-file': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FNAME">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- PEM编码SSL私钥文件的路径\n- path to file a PEM-encoded SSL private key  (env: LLAMA_ARG_SSL_KEY_FILE)">]>;
+            readonly 'ssl-cert-file': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FNAME">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- PEM编码SSL证书文件路径\n- path to file a PEM-encoded SSL certificate  (env: LLAMA_ARG_SSL_CERT_FILE)">]>;
+            readonly 'chat-template-kwargs': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "STRING">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 为 JSON 模板解析器设置额外的参数\n- sets additional params for the json template parser  (env: LLAMA_CHAT_TEMPLATE_KWARGS)">]>;
+            readonly timeout: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 服务器读/写超时时间（秒），默认：600\n- server read/write timeout in seconds (default: 600)  (env: LLAMA_ARG_TIMEOUT)">]>;
+            readonly 'threads-http': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 用于处理HTTP请求的线程数（默认：-1）\n- number of threads used to process HTTP requests (default: -1)  (env: LLAMA_ARG_THREADS_HTTP)">]>;
+            readonly 'cache-reuse': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 最小块大小以尝试通过KV移动从缓存中重用 (默认: 0) [(card)](https://ggml.ai/f0.png)\n- min chunk size to attempt reusing from the cache via KV shifting (default: 0)  [(card)](https://ggml.ai/f0.png)  (env: LLAMA_ARG_CACHE_REUSE)">]>;
+            readonly metrics: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 启用 prometheus 兼容的 metrics 端点 (默认: 禁用)\n- enable prometheus compatible metrics endpoint (default: disabled)  (env: LLAMA_ARG_ENDPOINT_METRICS)">]>;
+            readonly props: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 启用通过POST /props更改全局属性（默认：禁用）\n- enable changing global properties via POST /props (default: disabled)  (env: LLAMA_ARG_ENDPOINT_PROPS)">]>;
+            readonly slots: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 启用插槽监控端点（默认：禁用）\n- enable slots monitoring endpoint (default: enabled)  (env: LLAMA_ARG_ENDPOINT_SLOTS)">]>;
+            readonly 'no-slots': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 禁用插槽监控端点\n- disables slots monitoring endpoint  (env: LLAMA_ARG_NO_ENDPOINT_SLOTS)">]>;
+            readonly 'slot-save-path': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "PATH">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 保存slot kv缓存的路径（默认：禁用）\n- path to save slot kv cache (default: disabled)">]>;
+            readonly jinja: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 使用jinja模板进行聊天（默认：禁用）\n- use jinja template for chat (default: disabled)  (env: LLAMA_ARG_JINJA)">]>;
+            readonly 'reasoning-format': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.PicklistSchema<["deepseek", "none"], undefined>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: ["none" | "deepseek"] | undefined;
+            } | undefined, "- 控制是否允许并/或从响应中提取思维标签，以及以何种格式返回；可选值包括：  \\n- none：将思维内容原样保留在 `message.content` 中  \\n- deepseek：将思维内容放入 `message.reasoning_content`（在流式模式下行为与 none 相同）  \\n（默认值：deepseek）\n- controls whether thought tags are allowed and/or extracted from the response, and in which format they're returned; one of:  - none: leaves thoughts unparsed in `message.content`  - deepseek: puts thoughts in `message.reasoning_content` (except in streaming mode, which behaves as `none`)  (default: auto)  (env: LLAMA_ARG_THINK)">]>;
+            readonly 'reasoning-budget': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 控制允许的思考量；目前只能选择以下选项之一：-1 表示无限制的思考预算，或 0 表示禁用思考（默认值：-1）\n- controls the amount of thinking allowed; currently only one of: -1 for unrestricted thinking budget, or 0 to disable thinking (default: -1)  (env: LLAMA_ARG_THINK_BUDGET)">]>;
+            readonly 'chat-template': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "JINJA_TEMPLATE">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 设置自定义的Jinja聊天模板（默认：从模型元数据中获取模板）如果指定了后缀/前缀，模板将被禁用 仅接受常用模板（除非在该标志之前设置了--jinja）：内置模板列表：bailing, chatglm3, chatglm4, chatml, command-r, deepseek, deepseek2, deepseek3, exaone3, falcon3, gemma, gigachat, glmedge, granite, llama2, llama2-sys, llama2-sys-bos, llama2-sys-strip, llama3, llama4, megrez, minicpm, mistral-v1, mistral-v3, mistral-v3-tekken, mistral-v7, mistral-v7-0, mistral-v7-tekken, monarch, openchat, orion, phi3, phi4, rwkv-world, smolvlm, vicuna, vicuna-orca, yandex, zephyr\n- set custom jinja chat template (default: template taken from model's metadata)  if suffix/prefix are specified, template will be disabled  only commonly used templates are accepted (unless --jinja is set before this flag):  list of built-in templates:  bailing, chatglm3, chatglm4, chatml, command-r, deepseek, deepseek2, deepseek3, exaone3, exaone4, falcon3, gemma, gigachat, glmedge, gpt-oss, granite, grok-2, hunyuan-dense, hunyuan-moe, kimi-k2, llama2, llama2-sys, llama2-sys-bos, llama2-sys-strip, llama3, llama4, megrez, minicpm, mistral-v1, mistral-v3, mistral-v3-tekken, mistral-v7, mistral-v7-tekken, monarch, openchat, orion, phi3, phi4, rwkv-world, seed_oss, smolvlm, vicuna, vicuna-orca, yandex, zephyr  (env: LLAMA_ARG_CHAT_TEMPLATE)">]>;
+            readonly 'chat-template-file': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "JINJA_TEMPLATE_FILE">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 设置自定义的Jinja聊天模板文件（默认：从模型元数据中获取模板）如果指定了后缀/前缀，模板将被禁用 仅接受常用模板（除非在该标志之前设置了--jinja）：内置模板列表：bailing, chatglm3, chatglm4, chatml, command-r, deepseek, deepseek2, deepseek3, exaone3, falcon3, gemma, gigachat, glmedge, granite, llama2, llama2-sys, llama2-sys-bos, llama2-sys-strip, llama3, llama4, megrez, minicpm, mistral-v1, mistral-v3, mistral-v3-tekken, mistral-v7, mistral-v7-0, mistral-v7-tekken, monarch, openchat, orion, phi3, phi4, rwkv-world, smolvlm, vicuna, vicuna-orca, yandex, zephyr\n- set custom jinja chat template file (default: template taken from model's metadata)  if suffix/prefix are specified, template will be disabled  only commonly used templates are accepted (unless --jinja is set before this flag):  list of built-in templates:  bailing, chatglm3, chatglm4, chatml, command-r, deepseek, deepseek2, deepseek3, exaone3, exaone4, falcon3, gemma, gigachat, glmedge, gpt-oss, granite, grok-2, hunyuan-dense, hunyuan-moe, kimi-k2, llama2, llama2-sys, llama2-sys-bos, llama2-sys-strip, llama3, llama4, megrez, minicpm, mistral-v1, mistral-v3, mistral-v3-tekken, mistral-v7, mistral-v7-tekken, monarch, openchat, orion, phi3, phi4, rwkv-world, seed_oss, smolvlm, vicuna, vicuna-orca, yandex, zephyr  (env: LLAMA_ARG_CHAT_TEMPLATE_FILE)">]>;
+            readonly 'no-prefill-assistant': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 是否在最后一条消息是助手消息时预填充助手的响应（默认：预填充启用）当设置此标志时，如果最后一条消息是助手消息，则将其视为完整消息且不预填充\n- whether to prefill the assistant's response if the last message is an assistant message (default: prefill enabled)  when this flag is set, if the last message is an assistant message then it will be treated as a full message and not prefilled    (env: LLAMA_ARG_NO_PREFILL_ASSISTANT)">]>;
+            readonly 'slot-prompt-similarity': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "SIMILARITY">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 提示必须与槽的提示匹配多少才能使用该槽（默认：0.50，0.0 = 禁用）\n- how much the prompt of a request must match the prompt of a slot in order to use that slot (default: 0.10, 0.0 = disabled)">]>;
+            readonly 'lora-init-without-apply': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 加载LoRA适配器而不应用它们（稍后通过POST /lora-adapters应用）（默认：禁用）\n- load LoRA adapters without applying them (apply later via POST /lora-adapters) (default: disabled)">]>;
+            readonly 'threads-draft': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 在生成过程中使用的线程数量（默认值：与 --threads 的值相同）\n- number of threads to use during generation (default: same as --threads)">]>;
+            readonly 'threads-batch-draft': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 在批量处理和即时处理过程中使用的线程数量（默认值：与 --threads-draft 的值相同）\n- number of threads to use during batch and prompt processing (default: same as --threads-draft)">]>;
+            readonly 'draft-max': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 用于推测解码的草稿令牌数量（默认：16）\n- number of tokens to draft for speculative decoding (default: 16)  (env: LLAMA_ARG_DRAFT_MAX)">]>;
+            readonly 'draft-min': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 最小的草稿标记数，用于推测解码（默认：0）\n- minimum number of draft tokens to use for speculative decoding (default: 0)  (env: LLAMA_ARG_DRAFT_MIN)">]>;
+            readonly 'draft-p-min': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "P">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 最小推测解码概率（贪心）（默认：0.8）\n- minimum speculative decoding probability (greedy) (default: 0.8)  (env: LLAMA_ARG_DRAFT_P_MIN)">]>;
+            readonly 'ctx-size-draft': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 提示上下文的大小用于草稿模型（默认：0，0 = 从模型加载）\n- size of the prompt context for the draft model (default: 0, 0 = loaded from model)  (env: LLAMA_ARG_CTX_SIZE_DRAFT)">]>;
+            readonly 'device-draft': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "<dev1,dev2,..>">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 逗号分隔的设备列表，用于卸载草稿模型（none = 不卸载）使用--list-devices查看可用设备列表\n- comma-separated list of devices to use for offloading the draft model (none = don't offload)  use --list-devices to see a list of available devices">]>;
+            readonly 'gpu-layers-draft': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 存储在VRAM中的图层数量用于草稿模型\n- number of layers to store in VRAM for the draft model  (env: LLAMA_ARG_N_GPU_LAYERS_DRAFT)">]>;
+            readonly 'model-draft': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FNAME">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 用于推测解码的草稿模型（默认：未使用）\n- draft model for speculative decoding (default: unused)  (env: LLAMA_ARG_MODEL_DRAFT)">]>;
+            readonly 'spec-replace': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "TARGET">]>, v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "DRAFT">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string, string] | undefined;
+            } | undefined, "- 如果草稿模型（Draft Model）与主模型（Main Model）不兼容，那么需要将目标字符串（Target String）转换为草稿格式（Draft Format）。\n- translate the string in TARGET into DRAFT if the draft model and main model are not compatible">]>;
+            readonly 'model-vocoder': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FNAME">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 语音合成模型用于音频生成（默认：未使用）\n- vocoder model for audio generation (default: unused)">]>;
+            readonly 'tts-use-guide-tokens': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 使用引导标记以提高TTS单词回忆\n- Use guide tokens to improve TTS word recall">]>;
+            readonly 'embd-bge-small-en-default': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 使用默认的bge-small-en-v1.5模型（注意：可以从互联网下载权重）\n- use default bge-small-en-v1.5 model (note: can download weights from the internet)">]>;
+            readonly 'embd-e5-small-en-default': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 使用默认的e5-small-v2模型（注意：可以从互联网下载权重）\n- use default e5-small-v2 model (note: can download weights from the internet)">]>;
+            readonly 'embd-gte-small-default': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 使用默认的gte-small模型（注意：可以从互联网下载权重）\n- use default gte-small model (note: can download weights from the internet)">]>;
+            readonly 'fim-qwen-1.5b-default': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 使用默认的Qwen 2.5 Coder 1.5B\n- use default Qwen 2.5 Coder 1.5B (note: can download weights from the internet)">]>;
+            readonly 'fim-qwen-3b-default': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 使用默认的Qwen 2.5 Coder 3B（注意：可以从互联网下载权重）\n- use default Qwen 2.5 Coder 3B (note: can download weights from the internet)">]>;
+            readonly 'fim-qwen-7b-default': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 使用默认的Qwen 2.5 Coder 7B（注意：可以从互联网下载权重）\n- use default Qwen 2.5 Coder 7B (note: can download weights from the internet)">]>;
+            readonly 'fim-qwen-7b-spec': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 使用Qwen 2.5 Coder 7B + 0.5B draft进行推测解码（注意：可以从互联网下载权重）\n- use Qwen 2.5 Coder 7B + 0.5B draft for speculative decoding (note: can download weights from the internet)">]>;
+            readonly 'fim-qwen-14b-spec': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 使用Qwen 2.5 Coder 14B + 0.5B draft进行推测解码（注意：可以从互联网下载权重）\n- use Qwen 2.5 Coder 14B + 0.5B draft for speculative decoding (note: can download weights from the internet)">]>;
+            readonly 'fim-qwen-30b-default': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 使用默认的 Qwen 3 Coder 30B A3B Instruct；请注意：可以从互联网上下载相应的权重数据。\n- use default Qwen 3 Coder 30B A3B Instruct (note: can download weights from the internet)">]>;
+        }, undefined>, undefined>;
+        readonly sampling: v.OptionalSchema<v.ObjectSchema<{
+            readonly samplers: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "SAMPLERS">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- samplers that will be used for generation in the order, separated by ';'  (default: penalties;dry;top_n_sigma;top_k;typ_p;top_p;min_p;xtc;temperature)\n- samplers that will be used for generation in the order, separated by ';'  (default: penalties;dry;top_n_sigma;top_k;typ_p;top_p;min_p;xtc;temperature)">]>;
+            readonly seed: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "SEED">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- RNG种子（默认：-1，使用随机种子）\n- RNG seed (default: -1, use random seed for -1)">]>;
+            readonly 'sampling-seq': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "SEQUENCE">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 简化的采样序列，用于采样器（默认：edskypmxt）\n- simplified sequence for samplers that will be used (default: edskypmxt)">]>;
+            readonly 'ignore-eos': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+            } | undefined, "- 忽略流结束标记并继续生成（隐含--logit-bias EOS-inf）\n- ignore end of stream token and continue generating (implies --logit-bias EOS-inf)">]>;
+            readonly temp: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- temperature (default: 0.8)\n- temperature (default: 0.8)">]>;
+            readonly 'top-k': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- top-k 采样 (默认: 40, 0 = 禁用)\n- top-k sampling (default: 40, 0 = disabled)">]>;
+            readonly 'top-p': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- top-p 采样（默认：0.9，1.0 = 禁用）\n- top-p sampling (default: 0.9, 1.0 = disabled)">]>;
+            readonly 'min-p': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- min-p 采样 (默认: 0.1, 0.0 = 禁用)\n- min-p sampling (default: 0.1, 0.0 = disabled)">]>;
+            readonly 'top-nsigma': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- Top-N-Sigma抽样方法（默认值：-1.0；当两个参数都为-1.0时，该方法被禁用）\n- top-n-sigma sampling (default: -1.0, -1.0 = disabled)">]>;
+            readonly 'xtc-probability': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- xtc probability (默认: 0.0, 0.0 = 禁用)\n- xtc probability (default: 0.0, 0.0 = disabled)">]>;
+            readonly 'xtc-threshold': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- xtc threshold (默认: 0.1, 1.0 = 禁用)\n- xtc threshold (default: 0.1, 1.0 = disabled)">]>;
+            readonly typical: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 局部典型抽样，参数 p（默认：1.0，1.0 = 禁用）\n- locally typical sampling, parameter p (default: 1.0, 1.0 = disabled)">]>;
+            readonly 'repeat-last-n': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 用于惩罚的最后n个token（默认：64，0 = 禁用，-1 = 上下文大小）\n- last n tokens to consider for penalize (default: 64, 0 = disabled, -1 = ctx_size)">]>;
+            readonly 'repeat-penalty': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 惩罚重复的token序列（默认：1.0，1.0 = 禁用）\n- penalize repeat sequence of tokens (default: 1.0, 1.0 = disabled)">]>;
+            readonly 'presence-penalty': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 重复 alpha 存在惩罚（默认：0.0，0.0 = 禁用）\n- repeat alpha presence penalty (default: 0.0, 0.0 = disabled)">]>;
+            readonly 'frequency-penalty': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 重复 alpha 频率惩罚（默认：0.0，0.0 = 禁用）\n- repeat alpha frequency penalty (default: 0.0, 0.0 = disabled)">]>;
+            readonly 'dry-multiplier': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 设置 DRY 采样乘数（默认：0.0，0.0 = 禁用）\n- set DRY sampling multiplier (default: 0.0, 0.0 = disabled)">]>;
+            readonly 'dry-base': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 设置 DRY 采样基础值 (默认: 1.75)\n- set DRY sampling base value (default: 1.75)">]>;
+            readonly 'dry-allowed-length': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 设置DRY采样的允许长度（默认：2）\n- set allowed length for DRY sampling (default: 2)">]>;
+            readonly 'dry-penalty-last-n': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 设置最后n个token的DRY惩罚（默认：-1，0 = 禁用，-1 = 上下文大小）\n- set DRY penalty for the last n tokens (default: -1, 0 = disable, -1 = context size)">]>;
+            readonly 'dry-sequence-breaker': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "STRING">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 添加序列破环器用于DRY抽样，清除默认的破环器（'\\n', ':', '\"', '*'）；使用 \"none\" 不使用任何序列破环器\n- add sequence breaker for DRY sampling, clearing out default breakers ('\\n', ':', '\"', '*') in the process; use \"none\" to not use any sequence breakers">]>;
+            readonly 'dynatemp-range': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 动态温度范围 (默认: 0.0, 0.0 = 禁用)\n- dynamic temperature range (default: 0.0, 0.0 = disabled)">]>;
+            readonly 'dynatemp-exp': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 动态温度指数（默认：1.0）\n- dynamic temperature exponent (default: 1.0)">]>;
+            readonly mirostat: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- 使用Mirostat抽样。如果使用了Top K、Nucleus和Locally Typical抽样器，则会被忽略。（默认：0，0=禁用，1=Mirostat，2=Mirostat 2.0）\n- use Mirostat sampling.  Top K, Nucleus and Locally Typical samplers are ignored if used.  (default: 0, 0 = disabled, 1 = Mirostat, 2 = Mirostat 2.0)">]>;
+            readonly 'mirostat-lr': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- Mirostat学习率，参数eta（默认：0.1）\n- Mirostat learning rate, parameter eta (default: 0.1)">]>;
+            readonly 'mirostat-ent': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.NumberSchema<undefined>, v.DescriptionAction<number, "N">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [number] | undefined;
+            } | undefined, "- Mirostat目标熵，参数tau（默认：5.0）\n- Mirostat target entropy, parameter tau (default: 5.0)">]>;
+            readonly 'logit-bias': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "TOKEN_ID(+/-)BIAS">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 修改标记出现在完成中的可能性，例如：--logit-bias 15043+1 增加标记 ' Hello' 的可能性，或 --logit 15043-1 减少标记 ' Hello' 的可能性\n- modifies the likelihood of token appearing in the completion,  i.e. `--logit-bias 15043+1` to increase likelihood of token ' Hello',  or `--logit-bias 15043-1` to decrease likelihood of token ' Hello'">]>;
+            readonly grammar: v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "GRAMMAR">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- BNF类似语法约束生成（参见grammars目录中的示例）（默认：''）\n- BNF-like grammar to constrain generations (see samples in grammars/ dir) (default: '')">]>;
+            readonly 'grammar-file': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FNAME">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 从文件读取语法\n- file to read grammar from">]>;
+            readonly 'json-schema': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "SCHEMA">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- JSON schema 用于约束生成（https://json-schema.org/），例如 `{}` 表示任何 JSON 对象。对于包含外部 $refs 的 schema，使用 --grammar + example/json_schema_to_grammar.py 代替\n- JSON schema to constrain generations (https://json-schema.org/), e.g. `{}` for any JSON object  For schemas w/ external $refs, use --grammar + example/json_schema_to_grammar.py instead">]>;
+            readonly 'json-schema-file': v.SchemaWithPipe<readonly [v.OptionalSchema<v.ObjectSchema<{
+                readonly enable: v.OptionalSchema<v.BooleanSchema<undefined>, false>;
+                readonly value: v.SchemaWithPipe<readonly [v.OptionalSchema<v.TupleSchema<[v.SchemaWithPipe<readonly [v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TrimAction]>, v.DescriptionAction<string, "FILE">]>], undefined>, undefined>]>;
+            }, undefined>, undefined>, v.DescriptionAction<{
+                enable: boolean;
+                value?: [string] | undefined;
+            } | undefined, "- 包含用于约束生成的JSON架构的文件（https://json-schema.org/），例如{}表示任何JSON对象  对于具有外部$refs的架构，使用--grammar + example/json_schema_to_grammar.py代替\n- File containing a JSON schema to constrain generations (https://json-schema.org/), e.g. `{}` for any JSON object  For schemas w/ external $refs, use --grammar + example/json_schema_to_grammar.py instead">]>;
+        }, undefined>, undefined>;
+    }, undefined>, undefined>;
+    readonly aliases: v.OptionalSchema<v.ArraySchema<v.StringSchema<undefined>, undefined>, undefined>;
+    readonly env: v.OptionalSchema<v.RecordSchema<v.StringSchema<undefined>, v.AnySchema, undefined>, undefined>;
+    readonly ttl: v.OptionalSchema<v.NumberSchema<undefined>, undefined>;
+    readonly checkEndpoint: v.OptionalSchema<v.StringSchema<undefined>, undefined>;
+    readonly useModelName: v.OptionalSchema<v.StringSchema<undefined>, undefined>;
+    readonly exec: v.OptionalSchema<v.UnionSchema<[v.StringSchema<undefined>, v.ObjectSchema<{
+        readonly version: v.SchemaWithPipe<readonly [v.StringSchema<undefined>, v.TitleAction<string, "llama.cpp 版本号">, v.DescriptionAction<string, "llama.cpp的版本号,可以从 https://github.com/ggerganov/llama.cpp/releases 获得">]>;
+        readonly device: v.SchemaWithPipe<readonly [v.PicklistSchema<["cpu", "cuda12.4", "hip-radeon", "sycl", "vulkan"], undefined>, v.TitleAction<"cpu" | "cuda12.4" | "hip-radeon" | "sycl" | "vulkan", "llama.cpp 使用设备">]>;
+    }, undefined>], undefined>, undefined>;
+    readonly model: v.OptionalSchema<v.StringSchema<undefined>, undefined>;
+    readonly proxy: v.SchemaWithPipe<readonly [v.OptionalSchema<v.StringSchema<undefined>, undefined>, v.DescriptionAction<string | undefined, "正常情况下无需配置会自动生成">]>;
+}, undefined>;
+export type LlamaServerSwapItemType = v.InferOutput<typeof LlamaServerSwapItemDefine>;
+export type LlamaServerSwapItemInputType = v.InferInput<typeof LlamaServerSwapItemDefine>;