@shenghuabi/llama 1.0.0

package/define/common.define.d.ts

@@ -0,0 +1,534 @@
+import * as v from 'valibot';
+export declare const CommonDefine: 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>;