@fugood/llama.node 1.3.0 → 1.3.2

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (143) hide show
  1. package/package.json +14 -14
  2. package/scripts/llama.cpp.patch +8 -8
  3. package/src/llama.cpp/common/CMakeLists.txt +2 -0
  4. package/src/llama.cpp/common/arg.cpp +44 -999
  5. package/src/llama.cpp/common/arg.h +2 -2
  6. package/src/llama.cpp/common/chat.cpp +17 -2
  7. package/src/llama.cpp/common/common.cpp +33 -0
  8. package/src/llama.cpp/common/common.h +15 -1
  9. package/src/llama.cpp/common/download.cpp +1054 -0
  10. package/src/llama.cpp/common/download.h +55 -0
  11. package/src/llama.cpp/ggml/CMakeLists.txt +1 -1
  12. package/src/llama.cpp/ggml/include/ggml.h +2 -0
  13. package/src/llama.cpp/ggml/src/CMakeLists.txt +6 -3
  14. package/src/llama.cpp/ggml/src/ggml-cpu/CMakeLists.txt +29 -11
  15. package/src/llama.cpp/ggml/src/ggml-cpu/arch/arm/quants.c +428 -26
  16. package/src/llama.cpp/ggml/src/ggml-cpu/arch/loongarch/quants.c +4 -5
  17. package/src/llama.cpp/ggml/src/ggml-cpu/arch/riscv/quants.c +108 -49
  18. package/src/llama.cpp/ggml/src/ggml-cpu/arch/s390/cpu-feats.cpp +50 -0
  19. package/src/llama.cpp/ggml/src/ggml-cpu/ggml-cpu-impl.h +3 -1
  20. package/src/llama.cpp/ggml/src/ggml-cpu/ggml-cpu.c +21 -21
  21. package/src/llama.cpp/ggml/src/ggml-cpu/ops.cpp +172 -75
  22. package/src/llama.cpp/ggml/src/ggml-cpu/ops.h +0 -4
  23. package/src/llama.cpp/ggml/src/ggml-cpu/repack.cpp +82 -21
  24. package/src/llama.cpp/ggml/src/ggml-cpu/simd-mappings.h +25 -25
  25. package/src/llama.cpp/include/llama.h +7 -3
  26. package/src/llama.cpp/src/CMakeLists.txt +95 -0
  27. package/src/llama.cpp/src/llama-arch.cpp +108 -0
  28. package/src/llama.cpp/src/llama-arch.h +11 -0
  29. package/src/llama.cpp/src/llama-batch.cpp +63 -31
  30. package/src/llama.cpp/src/llama-batch.h +12 -1
  31. package/src/llama.cpp/src/llama-chat.cpp +32 -0
  32. package/src/llama.cpp/src/llama-chat.h +1 -0
  33. package/src/llama.cpp/src/llama-context.cpp +36 -13
  34. package/src/llama.cpp/src/llama-context.h +5 -5
  35. package/src/llama.cpp/src/llama-cparams.h +1 -0
  36. package/src/llama.cpp/src/llama-graph.cpp +3 -3
  37. package/src/llama.cpp/src/llama-hparams.cpp +11 -1
  38. package/src/llama.cpp/src/llama-hparams.h +6 -0
  39. package/src/llama.cpp/src/llama-kv-cache-iswa.cpp +3 -1
  40. package/src/llama.cpp/src/llama-kv-cache.cpp +33 -1
  41. package/src/llama.cpp/src/llama-kv-cells.h +44 -2
  42. package/src/llama.cpp/src/llama-memory-recurrent.cpp +4 -3
  43. package/src/llama.cpp/src/llama-model.cpp +320 -13171
  44. package/src/llama.cpp/src/llama-model.h +8 -0
  45. package/src/llama.cpp/src/llama-quant.cpp +1 -1
  46. package/src/llama.cpp/src/llama-vocab.cpp +5 -0
  47. package/src/llama.cpp/src/llama-vocab.h +1 -0
  48. package/src/llama.cpp/src/models/apertus.cpp +125 -0
  49. package/src/llama.cpp/src/models/arcee.cpp +135 -0
  50. package/src/llama.cpp/src/models/arctic.cpp +138 -0
  51. package/src/llama.cpp/src/models/arwkv7.cpp +86 -0
  52. package/src/llama.cpp/src/models/baichuan.cpp +122 -0
  53. package/src/llama.cpp/src/models/bailingmoe.cpp +144 -0
  54. package/src/llama.cpp/src/models/bailingmoe2.cpp +135 -0
  55. package/src/llama.cpp/src/models/bert.cpp +176 -0
  56. package/src/llama.cpp/src/models/bitnet.cpp +160 -0
  57. package/src/llama.cpp/src/models/bloom.cpp +101 -0
  58. package/src/llama.cpp/src/models/chameleon.cpp +178 -0
  59. package/src/llama.cpp/src/models/chatglm.cpp +132 -0
  60. package/src/llama.cpp/src/models/codeshell.cpp +111 -0
  61. package/src/llama.cpp/src/models/cogvlm.cpp +100 -0
  62. package/src/llama.cpp/src/models/cohere2-iswa.cpp +131 -0
  63. package/src/llama.cpp/src/models/command-r.cpp +122 -0
  64. package/src/llama.cpp/src/models/dbrx.cpp +123 -0
  65. package/src/llama.cpp/src/models/deci.cpp +135 -0
  66. package/src/llama.cpp/src/models/deepseek.cpp +144 -0
  67. package/src/llama.cpp/src/models/deepseek2.cpp +236 -0
  68. package/src/llama.cpp/src/models/dots1.cpp +134 -0
  69. package/src/llama.cpp/src/models/dream.cpp +105 -0
  70. package/src/llama.cpp/src/models/ernie4-5-moe.cpp +150 -0
  71. package/src/llama.cpp/src/models/ernie4-5.cpp +110 -0
  72. package/src/llama.cpp/src/models/exaone.cpp +114 -0
  73. package/src/llama.cpp/src/models/exaone4.cpp +123 -0
  74. package/src/llama.cpp/src/models/falcon-h1.cpp +113 -0
  75. package/src/llama.cpp/src/models/falcon.cpp +120 -0
  76. package/src/llama.cpp/src/models/gemma-embedding.cpp +120 -0
  77. package/src/llama.cpp/src/models/gemma.cpp +112 -0
  78. package/src/llama.cpp/src/models/gemma2-iswa.cpp +125 -0
  79. package/src/llama.cpp/src/models/gemma3-iswa.cpp +131 -0
  80. package/src/llama.cpp/src/models/gemma3n-iswa.cpp +377 -0
  81. package/src/llama.cpp/src/models/glm4-moe.cpp +153 -0
  82. package/src/llama.cpp/src/models/glm4.cpp +127 -0
  83. package/src/llama.cpp/src/models/gpt2.cpp +105 -0
  84. package/src/llama.cpp/src/models/gptneox.cpp +144 -0
  85. package/src/llama.cpp/src/models/granite-hybrid.cpp +196 -0
  86. package/src/llama.cpp/src/models/granite.cpp +211 -0
  87. package/src/llama.cpp/src/models/graph-context-mamba.cpp +283 -0
  88. package/src/llama.cpp/src/models/grok.cpp +159 -0
  89. package/src/llama.cpp/src/models/grovemoe.cpp +141 -0
  90. package/src/llama.cpp/src/models/hunyuan-dense.cpp +132 -0
  91. package/src/llama.cpp/src/models/hunyuan-moe.cpp +154 -0
  92. package/src/llama.cpp/src/models/internlm2.cpp +120 -0
  93. package/src/llama.cpp/src/models/jais.cpp +86 -0
  94. package/src/llama.cpp/src/models/jamba.cpp +106 -0
  95. package/src/llama.cpp/src/models/lfm2.cpp +173 -0
  96. package/src/llama.cpp/src/models/llada-moe.cpp +122 -0
  97. package/src/llama.cpp/src/models/llada.cpp +99 -0
  98. package/src/llama.cpp/src/models/llama-iswa.cpp +174 -0
  99. package/src/llama.cpp/src/models/llama.cpp +155 -0
  100. package/src/llama.cpp/src/models/mamba.cpp +55 -0
  101. package/src/llama.cpp/src/models/minicpm3.cpp +199 -0
  102. package/src/llama.cpp/src/models/minimax-m2.cpp +124 -0
  103. package/src/llama.cpp/src/models/models.h +481 -0
  104. package/src/llama.cpp/src/models/mpt.cpp +126 -0
  105. package/src/llama.cpp/src/models/nemotron-h.cpp +121 -0
  106. package/src/llama.cpp/src/models/nemotron.cpp +122 -0
  107. package/src/llama.cpp/src/models/neo-bert.cpp +104 -0
  108. package/src/llama.cpp/src/models/olmo.cpp +121 -0
  109. package/src/llama.cpp/src/models/olmo2.cpp +150 -0
  110. package/src/llama.cpp/src/models/olmoe.cpp +124 -0
  111. package/src/llama.cpp/src/models/openai-moe-iswa.cpp +124 -0
  112. package/src/llama.cpp/src/models/openelm.cpp +124 -0
  113. package/src/llama.cpp/src/models/orion.cpp +123 -0
  114. package/src/llama.cpp/src/models/pangu-embedded.cpp +121 -0
  115. package/src/llama.cpp/src/models/phi2.cpp +121 -0
  116. package/src/llama.cpp/src/models/phi3.cpp +152 -0
  117. package/src/llama.cpp/src/models/plamo.cpp +110 -0
  118. package/src/llama.cpp/src/models/plamo2.cpp +316 -0
  119. package/src/llama.cpp/src/models/plm.cpp +168 -0
  120. package/src/llama.cpp/src/models/qwen.cpp +108 -0
  121. package/src/llama.cpp/src/models/qwen2.cpp +117 -0
  122. package/src/llama.cpp/src/models/qwen2moe.cpp +151 -0
  123. package/src/llama.cpp/src/models/qwen2vl.cpp +117 -0
  124. package/src/llama.cpp/src/models/qwen3.cpp +117 -0
  125. package/src/llama.cpp/src/models/qwen3moe.cpp +124 -0
  126. package/src/llama.cpp/src/models/qwen3vl-moe.cpp +149 -0
  127. package/src/llama.cpp/src/models/qwen3vl.cpp +141 -0
  128. package/src/llama.cpp/src/models/refact.cpp +94 -0
  129. package/src/llama.cpp/src/models/rwkv6-base.cpp +162 -0
  130. package/src/llama.cpp/src/models/rwkv6.cpp +94 -0
  131. package/src/llama.cpp/src/models/rwkv6qwen2.cpp +86 -0
  132. package/src/llama.cpp/src/models/rwkv7-base.cpp +135 -0
  133. package/src/llama.cpp/src/models/rwkv7.cpp +90 -0
  134. package/src/llama.cpp/src/models/seed-oss.cpp +124 -0
  135. package/src/llama.cpp/src/models/smallthinker.cpp +120 -0
  136. package/src/llama.cpp/src/models/smollm3.cpp +128 -0
  137. package/src/llama.cpp/src/models/stablelm.cpp +146 -0
  138. package/src/llama.cpp/src/models/starcoder.cpp +100 -0
  139. package/src/llama.cpp/src/models/starcoder2.cpp +121 -0
  140. package/src/llama.cpp/src/models/t5-dec.cpp +166 -0
  141. package/src/llama.cpp/src/models/t5-enc.cpp +96 -0
  142. package/src/llama.cpp/src/models/wavtokenizer-dec.cpp +149 -0
  143. package/src/llama.cpp/src/models/xverse.cpp +108 -0
package/package.json CHANGED
@@ -1,7 +1,7 @@
1
1
  {
2
2
  "name": "@fugood/llama.node",
3
3
  "access": "public",
4
- "version": "1.3.0",
4
+ "version": "1.3.2",
5
5
  "description": "An another Node binding of llama.cpp",
6
6
  "main": "lib/index.js",
7
7
  "scripts": {
@@ -72,19 +72,19 @@
72
72
  "CMakeLists.txt"
73
73
  ],
74
74
  "optionalDependencies": {
75
- "@fugood/node-llama-linux-x64": "1.3.0",
76
- "@fugood/node-llama-linux-x64-vulkan": "1.3.0",
77
- "@fugood/node-llama-linux-x64-cuda": "1.3.0",
78
- "@fugood/node-llama-linux-arm64": "1.3.0",
79
- "@fugood/node-llama-linux-arm64-vulkan": "1.3.0",
80
- "@fugood/node-llama-linux-arm64-cuda": "1.3.0",
81
- "@fugood/node-llama-win32-x64": "1.3.0",
82
- "@fugood/node-llama-win32-x64-vulkan": "1.3.0",
83
- "@fugood/node-llama-win32-x64-cuda": "1.3.0",
84
- "@fugood/node-llama-win32-arm64": "1.3.0",
85
- "@fugood/node-llama-win32-arm64-vulkan": "1.3.0",
86
- "@fugood/node-llama-darwin-x64": "1.3.0",
87
- "@fugood/node-llama-darwin-arm64": "1.3.0"
75
+ "@fugood/node-llama-linux-x64": "1.3.2",
76
+ "@fugood/node-llama-linux-x64-vulkan": "1.3.2",
77
+ "@fugood/node-llama-linux-x64-cuda": "1.3.2",
78
+ "@fugood/node-llama-linux-arm64": "1.3.2",
79
+ "@fugood/node-llama-linux-arm64-vulkan": "1.3.2",
80
+ "@fugood/node-llama-linux-arm64-cuda": "1.3.2",
81
+ "@fugood/node-llama-win32-x64": "1.3.2",
82
+ "@fugood/node-llama-win32-x64-vulkan": "1.3.2",
83
+ "@fugood/node-llama-win32-x64-cuda": "1.3.2",
84
+ "@fugood/node-llama-win32-arm64": "1.3.2",
85
+ "@fugood/node-llama-win32-arm64-vulkan": "1.3.2",
86
+ "@fugood/node-llama-darwin-x64": "1.3.2",
87
+ "@fugood/node-llama-darwin-arm64": "1.3.2"
88
88
  },
89
89
  "devDependencies": {
90
90
  "@babel/preset-env": "^7.24.4",
package/scripts/llama.cpp.patch CHANGED
@@ -1,8 +1,8 @@
1
1
  diff --git a/src/llama.cpp/common/CMakeLists.txt b/src/llama.cpp/common/CMakeLists.txt
2
- index fe290bf8f..d377e29b9 100644
2
+ index 7086d08e5..9a727bcf8 100644
3
3
  --- a/src/llama.cpp/common/CMakeLists.txt
4
4
  +++ b/src/llama.cpp/common/CMakeLists.txt
5
- @@ -170,9 +170,16 @@ if (LLAMA_LLGUIDANCE)
5
+ @@ -172,9 +172,16 @@ if (LLAMA_LLGUIDANCE)
6
6
  set(LLAMA_COMMON_EXTRA_LIBS ${LLAMA_COMMON_EXTRA_LIBS} llguidance ${LLGUIDANCE_PLATFORM_LIBS})
7
7
  endif ()
8
8
 
@@ -21,7 +21,7 @@ index fe290bf8f..d377e29b9 100644
21
21
 
22
22
  #
23
23
  diff --git a/src/llama.cpp/common/chat.cpp b/src/llama.cpp/common/chat.cpp
24
- index 63583fb22..f8be20148 100644
24
+ index 938872e82..6364f173f 100644
25
25
  --- a/src/llama.cpp/common/chat.cpp
26
26
  +++ b/src/llama.cpp/common/chat.cpp
27
27
  @@ -6,9 +6,6 @@
@@ -51,7 +51,7 @@ index 63583fb22..f8be20148 100644
51
51
  struct templates_params {
52
52
  json messages;
53
53
  json tools;
54
- @@ -812,7 +799,7 @@ static std::string apply(
54
+ @@ -811,7 +798,7 @@ static std::string apply(
55
55
  tmpl_inputs.extra_context.merge_patch(*additional_context);
56
56
  }
57
57
  // TODO: add flag to control date/time, if only for testing purposes.
@@ -85,10 +85,10 @@ index 50efb0d4e..f471a84c7 100644
85
85
  struct common_chat_tool_call {
86
86
  std::string name;
87
87
  diff --git a/src/llama.cpp/common/common.cpp b/src/llama.cpp/common/common.cpp
88
- index b0591e84b..93759f884 100644
88
+ index a8d709ab1..d8aed9c7e 100644
89
89
  --- a/src/llama.cpp/common/common.cpp
90
90
  +++ b/src/llama.cpp/common/common.cpp
91
- @@ -1126,6 +1126,7 @@ struct llama_model_params common_model_params_to_llama(common_params & params) {
91
+ @@ -1159,6 +1159,7 @@ struct llama_model_params common_model_params_to_llama(common_params & params) {
92
92
  mparams.n_gpu_layers = params.n_gpu_layers;
93
93
  }
94
94
 
@@ -97,7 +97,7 @@ index b0591e84b..93759f884 100644
97
97
  mparams.split_mode = params.split_mode;
98
98
  mparams.tensor_split = params.tensor_split;
99
99
  diff --git a/src/llama.cpp/common/common.h b/src/llama.cpp/common/common.h
100
- index a8cb630ea..0919ec5d3 100644
100
+ index f42c083fa..c573cc812 100644
101
101
  --- a/src/llama.cpp/common/common.h
102
102
  +++ b/src/llama.cpp/common/common.h
103
103
  @@ -274,6 +274,7 @@ struct lr_opt {
@@ -109,7 +109,7 @@ index a8cb630ea..0919ec5d3 100644
109
109
  int32_t n_ctx = 4096; // context size
110
110
  int32_t n_batch = 2048; // logical batch size for prompt processing (must be >=32 to use BLAS)
111
111
  diff --git a/src/llama.cpp/ggml/src/ggml-cpu/CMakeLists.txt b/src/llama.cpp/ggml/src/ggml-cpu/CMakeLists.txt
112
- index 34323afa0..1a6924db0 100644
112
+ index a55191aed..53e318c62 100644
113
113
  --- a/src/llama.cpp/ggml/src/ggml-cpu/CMakeLists.txt
114
114
  +++ b/src/llama.cpp/ggml/src/ggml-cpu/CMakeLists.txt
115
115
  @@ -106,7 +106,7 @@ function(ggml_add_cpu_backend_variant_impl tag_name)
package/src/llama.cpp/common/CMakeLists.txt CHANGED
@@ -56,6 +56,8 @@ add_library(${TARGET} STATIC
56
56
  common.h
57
57
  console.cpp
58
58
  console.h
59
+ download.cpp
60
+ download.h
59
61
  http.h
60
62
  json-partial.cpp
61
63
  json-partial.h