@fugood/llama.node 1.0.0-beta.5 → 1.0.0-beta.7

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 (113) hide show
  1. package/lib/binding.ts +3 -1
  2. package/lib/index.js +2 -0
  3. package/lib/index.ts +3 -1
  4. package/package.json +14 -14
  5. package/scripts/llama.cpp.patch +27 -26
  6. package/src/EmbeddingWorker.cpp +1 -1
  7. package/src/LlamaCompletionWorker.cpp +28 -7
  8. package/src/LlamaCompletionWorker.h +4 -0
  9. package/src/LlamaContext.cpp +14 -17
  10. package/src/common.hpp +7 -6
  11. package/src/llama.cpp/CMakeLists.txt +15 -4
  12. package/src/llama.cpp/common/CMakeLists.txt +15 -24
  13. package/src/llama.cpp/common/arg.cpp +172 -110
  14. package/src/llama.cpp/common/chat-parser.cpp +385 -0
  15. package/src/llama.cpp/common/chat-parser.h +120 -0
  16. package/src/llama.cpp/common/chat.cpp +726 -596
  17. package/src/llama.cpp/common/chat.h +74 -8
  18. package/src/llama.cpp/common/common.cpp +56 -38
  19. package/src/llama.cpp/common/common.h +9 -3
  20. package/src/llama.cpp/common/json-partial.cpp +256 -0
  21. package/src/llama.cpp/common/json-partial.h +38 -0
  22. package/src/llama.cpp/common/json-schema-to-grammar.cpp +2 -1
  23. package/src/llama.cpp/common/json-schema-to-grammar.h +4 -4
  24. package/src/llama.cpp/common/sampling.cpp +7 -8
  25. package/src/llama.cpp/common/speculative.cpp +6 -4
  26. package/src/llama.cpp/ggml/CMakeLists.txt +48 -3
  27. package/src/llama.cpp/ggml/include/ggml.h +22 -3
  28. package/src/llama.cpp/ggml/src/CMakeLists.txt +81 -22
  29. package/src/llama.cpp/ggml/src/ggml-cpu/CMakeLists.txt +131 -49
  30. package/src/llama.cpp/ggml/src/ggml-cpu/amx/amx.cpp +1 -1
  31. package/src/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp +1 -1
  32. package/src/llama.cpp/ggml/src/ggml-cpu/arch/arm/cpu-feats.cpp +94 -0
  33. package/src/llama.cpp/ggml/src/ggml-cpu/arch/arm/quants.c +4113 -0
  34. package/src/llama.cpp/ggml/src/ggml-cpu/arch/arm/repack.cpp +2162 -0
  35. package/src/llama.cpp/ggml/src/ggml-cpu/arch/loongarch/quants.c +2638 -0
  36. package/src/llama.cpp/ggml/src/ggml-cpu/arch/powerpc/cpu-feats.cpp +82 -0
  37. package/src/llama.cpp/ggml/src/ggml-cpu/arch/powerpc/quants.c +2731 -0
  38. package/src/llama.cpp/ggml/src/ggml-cpu/arch/riscv/quants.c +2068 -0
  39. package/src/llama.cpp/ggml/src/ggml-cpu/arch/riscv/repack.cpp +396 -0
  40. package/src/llama.cpp/ggml/src/ggml-cpu/arch/s390/quants.c +1299 -0
  41. package/src/llama.cpp/ggml/src/ggml-cpu/arch/wasm/quants.c +1480 -0
  42. package/src/llama.cpp/ggml/src/ggml-cpu/arch/x86/quants.c +4310 -0
  43. package/src/llama.cpp/ggml/src/ggml-cpu/{ggml-cpu-aarch64.cpp → arch/x86/repack.cpp} +59 -3206
  44. package/src/llama.cpp/ggml/src/ggml-cpu/arch-fallback.h +184 -0
  45. package/src/llama.cpp/ggml/src/ggml-cpu/common.h +1 -1
  46. package/src/llama.cpp/ggml/src/ggml-cpu/ggml-cpu-impl.h +12 -13
  47. package/src/llama.cpp/ggml/src/ggml-cpu/ggml-cpu.c +64 -88
  48. package/src/llama.cpp/ggml/src/ggml-cpu/ggml-cpu.cpp +8 -8
  49. package/src/llama.cpp/ggml/src/ggml-cpu/{ggml-cpu-hbm.cpp → hbm.cpp} +1 -1
  50. package/src/llama.cpp/ggml/src/ggml-cpu/kleidiai/kleidiai.cpp +1 -1
  51. package/src/llama.cpp/ggml/src/ggml-cpu/llamafile/sgemm.cpp +56 -7
  52. package/src/llama.cpp/ggml/src/ggml-cpu/llamafile/sgemm.h +5 -0
  53. package/src/llama.cpp/ggml/src/ggml-cpu/ops.cpp +282 -100
  54. package/src/llama.cpp/ggml/src/ggml-cpu/ops.h +1 -0
  55. package/src/llama.cpp/ggml/src/ggml-cpu/quants.c +1157 -0
  56. package/src/llama.cpp/ggml/src/ggml-cpu/{ggml-cpu-quants.h → quants.h} +26 -0
  57. package/src/llama.cpp/ggml/src/ggml-cpu/repack.cpp +1570 -0
  58. package/src/llama.cpp/ggml/src/ggml-cpu/repack.h +98 -0
  59. package/src/llama.cpp/ggml/src/ggml-cpu/simd-mappings.h +119 -5
  60. package/src/llama.cpp/ggml/src/ggml-cpu/{ggml-cpu-traits.cpp → traits.cpp} +1 -1
  61. package/src/llama.cpp/ggml/src/ggml-cpu/vec.cpp +85 -16
  62. package/src/llama.cpp/ggml/src/ggml-cpu/vec.h +204 -49
  63. package/src/llama.cpp/include/llama.h +145 -40
  64. package/src/llama.cpp/src/CMakeLists.txt +5 -1
  65. package/src/llama.cpp/src/llama-arch.cpp +99 -3
  66. package/src/llama.cpp/src/llama-arch.h +10 -1
  67. package/src/llama.cpp/src/llama-batch.cpp +728 -272
  68. package/src/llama.cpp/src/llama-batch.h +112 -54
  69. package/src/llama.cpp/src/llama-chat.cpp +19 -2
  70. package/src/llama.cpp/src/llama-chat.h +1 -0
  71. package/src/llama.cpp/src/llama-context.cpp +525 -339
  72. package/src/llama.cpp/src/llama-context.h +38 -17
  73. package/src/llama.cpp/src/llama-cparams.cpp +4 -0
  74. package/src/llama.cpp/src/llama-cparams.h +2 -0
  75. package/src/llama.cpp/src/llama-grammar.cpp +12 -2
  76. package/src/llama.cpp/src/llama-graph.cpp +413 -353
  77. package/src/llama.cpp/src/llama-graph.h +112 -56
  78. package/src/llama.cpp/src/llama-hparams.cpp +10 -2
  79. package/src/llama.cpp/src/llama-hparams.h +13 -2
  80. package/src/llama.cpp/src/llama-kv-cache-unified-iswa.cpp +279 -0
  81. package/src/llama.cpp/src/llama-kv-cache-unified-iswa.h +128 -0
  82. package/src/llama.cpp/src/llama-kv-cache-unified.cpp +1815 -0
  83. package/src/llama.cpp/src/llama-kv-cache-unified.h +303 -0
  84. package/src/llama.cpp/src/llama-kv-cells.h +415 -0
  85. package/src/llama.cpp/src/llama-memory-hybrid.cpp +246 -0
  86. package/src/llama.cpp/src/llama-memory-hybrid.h +138 -0
  87. package/src/llama.cpp/src/llama-memory-recurrent.cpp +1112 -0
  88. package/src/llama.cpp/src/llama-memory-recurrent.h +183 -0
  89. package/src/llama.cpp/src/llama-memory.cpp +41 -0
  90. package/src/llama.cpp/src/llama-memory.h +86 -5
  91. package/src/llama.cpp/src/llama-mmap.cpp +1 -1
  92. package/src/llama.cpp/src/llama-model-loader.cpp +42 -17
  93. package/src/llama.cpp/src/llama-model-saver.cpp +1 -0
  94. package/src/llama.cpp/src/llama-model.cpp +1137 -528
  95. package/src/llama.cpp/src/llama-model.h +4 -0
  96. package/src/llama.cpp/src/llama-quant.cpp +2 -1
  97. package/src/llama.cpp/src/llama-sampling.cpp +2 -2
  98. package/src/llama.cpp/src/llama-vocab.cpp +69 -32
  99. package/src/llama.cpp/src/llama-vocab.h +1 -0
  100. package/src/llama.cpp/src/llama.cpp +11 -7
  101. package/src/llama.cpp/src/unicode.cpp +5 -0
  102. package/src/tts_utils.h +1 -1
  103. package/src/llama.cpp/common/json.hpp +0 -24766
  104. package/src/llama.cpp/common/minja/chat-template.hpp +0 -541
  105. package/src/llama.cpp/common/minja/minja.hpp +0 -2974
  106. package/src/llama.cpp/common/stb_image.h +0 -7988
  107. package/src/llama.cpp/ggml/src/ggml-cpu/ggml-cpu-aarch64.h +0 -8
  108. package/src/llama.cpp/ggml/src/ggml-cpu/ggml-cpu-quants.c +0 -13326
  109. package/src/llama.cpp/src/llama-kv-cache.cpp +0 -2827
  110. package/src/llama.cpp/src/llama-kv-cache.h +0 -515
  111. /package/src/llama.cpp/ggml/src/ggml-cpu/{cpu-feats-x86.cpp → arch/x86/cpu-feats.cpp} +0 -0
  112. /package/src/llama.cpp/ggml/src/ggml-cpu/{ggml-cpu-hbm.h → hbm.h} +0 -0
  113. /package/src/llama.cpp/ggml/src/ggml-cpu/{ggml-cpu-traits.h → traits.h} +0 -0
package/src/llama.cpp/ggml/src/ggml-cpu/{ggml-cpu-quants.h → quants.h} RENAMED
@@ -58,6 +58,32 @@ void ggml_vec_dot_iq4_nl_q8_0 (int n, float * GGML_RESTRICT s, size_t bs, const
58
58
  void ggml_vec_dot_iq4_xs_q8_K (int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
59
59
  void ggml_vec_dot_iq3_s_q8_K (int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
60
60
 
61
+ // Generic implementation
62
+ void quantize_row_q8_0_generic(const float * GGML_RESTRICT x, void * GGML_RESTRICT vy, int64_t k);
63
+ void quantize_row_q8_1_generic(const float * GGML_RESTRICT x, void * GGML_RESTRICT vy, int64_t k);
64
+ void quantize_row_q8_K_generic(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k);
65
+ void ggml_vec_dot_q4_0_q8_0_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
66
+ void ggml_vec_dot_q4_1_q8_1_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
67
+ void ggml_vec_dot_q5_0_q8_0_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
68
+ void ggml_vec_dot_q5_1_q8_1_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
69
+ void ggml_vec_dot_q8_0_q8_0_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
70
+ void ggml_vec_dot_tq1_0_q8_K_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
71
+ void ggml_vec_dot_tq2_0_q8_K_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
72
+ void ggml_vec_dot_q2_K_q8_K_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
73
+ void ggml_vec_dot_q3_K_q8_K_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
74
+ void ggml_vec_dot_q4_K_q8_K_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
75
+ void ggml_vec_dot_q5_K_q8_K_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
76
+ void ggml_vec_dot_q6_K_q8_K_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
77
+ void ggml_vec_dot_iq2_xxs_q8_K_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
78
+ void ggml_vec_dot_iq2_xs_q8_K_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
79
+ void ggml_vec_dot_iq2_s_q8_K_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
80
+ void ggml_vec_dot_iq3_xxs_q8_K_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
81
+ void ggml_vec_dot_iq3_s_q8_K_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
82
+ void ggml_vec_dot_iq1_s_q8_K_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
83
+ void ggml_vec_dot_iq1_m_q8_K_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
84
+ void ggml_vec_dot_iq4_nl_q8_0_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
85
+ void ggml_vec_dot_iq4_xs_q8_K_generic(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
86
+
61
87
  #ifdef __cplusplus
62
88
  }
63
89
  #endif