llama_cpp 0.16.0 → 0.16.1

Sign up to get free protection for your applications and to get access to all the features.
Files changed (134) hide show
  1. checksums.yaml +4 -4
  2. data/CHANGELOG.md +6 -0
  3. data/ext/llama_cpp/extconf.rb +2 -0
  4. data/ext/llama_cpp/llama_cpp.cpp +2 -0
  5. data/lib/llama_cpp/version.rb +2 -2
  6. data/sig/llama_cpp.rbs +2 -0
  7. data/vendor/tmp/llama.cpp/Makefile +110 -53
  8. data/vendor/tmp/llama.cpp/ggml-alloc.c +78 -22
  9. data/vendor/tmp/llama.cpp/ggml-backend-impl.h +20 -8
  10. data/vendor/tmp/llama.cpp/ggml-backend.c +178 -64
  11. data/vendor/tmp/llama.cpp/ggml-backend.h +3 -3
  12. data/vendor/tmp/llama.cpp/ggml-blas.cpp +363 -0
  13. data/vendor/tmp/llama.cpp/ggml-blas.h +23 -0
  14. data/vendor/tmp/llama.cpp/ggml-common.h +6 -0
  15. data/vendor/tmp/llama.cpp/ggml-cuda/argsort.cu +1 -0
  16. data/vendor/tmp/llama.cpp/ggml-cuda/dmmv.cu +21 -9
  17. data/vendor/tmp/llama.cpp/ggml-cuda/fattn-tile-f16.cu +1 -1
  18. data/vendor/tmp/llama.cpp/ggml-cuda/mmq.cu +15 -1491
  19. data/vendor/tmp/llama.cpp/ggml-cuda/mmvq.cu +76 -61
  20. data/vendor/tmp/llama.cpp/ggml-cuda/quantize.cu +77 -10
  21. data/vendor/tmp/llama.cpp/ggml-cuda/softmax.cu +1 -0
  22. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-f16-f16.cu +1 -1
  23. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-f16-q4_0.cu +1 -1
  24. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-f16-q4_1.cu +1 -1
  25. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-f16-q5_0.cu +1 -1
  26. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-f16-q5_1.cu +1 -1
  27. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-f16-q8_0.cu +1 -1
  28. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_0-f16.cu +1 -1
  29. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_0-q4_0.cu +1 -1
  30. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_0-q4_1.cu +1 -1
  31. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_0-q5_0.cu +1 -1
  32. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_0-q5_1.cu +1 -1
  33. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_0-q8_0.cu +1 -1
  34. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_1-f16.cu +1 -1
  35. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_1-q4_0.cu +1 -1
  36. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_1-q4_1.cu +1 -1
  37. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_1-q5_0.cu +1 -1
  38. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_1-q5_1.cu +1 -1
  39. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_1-q8_0.cu +1 -1
  40. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_0-f16.cu +1 -1
  41. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_0-q4_0.cu +1 -1
  42. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_0-q4_1.cu +1 -1
  43. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_0-q5_0.cu +1 -1
  44. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_0-q5_1.cu +1 -1
  45. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_0-q8_0.cu +1 -1
  46. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_1-f16.cu +1 -1
  47. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_1-q4_0.cu +1 -1
  48. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_1-q4_1.cu +1 -1
  49. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_1-q5_0.cu +1 -1
  50. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_1-q5_1.cu +1 -1
  51. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_1-q8_0.cu +1 -1
  52. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q8_0-f16.cu +1 -1
  53. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q8_0-q4_0.cu +1 -1
  54. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q8_0-q4_1.cu +1 -1
  55. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q8_0-q5_0.cu +1 -1
  56. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q8_0-q5_1.cu +1 -1
  57. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q8_0-q8_0.cu +1 -1
  58. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs256-f16-f16.cu +1 -1
  59. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs64-f16-f16.cu +1 -1
  60. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs64-f16-q4_0.cu +1 -1
  61. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs64-f16-q4_1.cu +1 -1
  62. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs64-f16-q5_0.cu +1 -1
  63. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs64-f16-q5_1.cu +1 -1
  64. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f16-instance-hs64-f16-q8_0.cu +1 -1
  65. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-f16-f16.cu +1 -1
  66. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-f16-q4_0.cu +1 -1
  67. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-f16-q4_1.cu +1 -1
  68. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-f16-q5_0.cu +1 -1
  69. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-f16-q5_1.cu +1 -1
  70. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-f16-q8_0.cu +1 -1
  71. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_0-f16.cu +1 -1
  72. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_0-q4_0.cu +1 -1
  73. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_0-q4_1.cu +1 -1
  74. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_0-q5_0.cu +1 -1
  75. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_0-q5_1.cu +1 -1
  76. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_0-q8_0.cu +1 -1
  77. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_1-f16.cu +1 -1
  78. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_1-q4_0.cu +1 -1
  79. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_1-q4_1.cu +1 -1
  80. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_1-q5_0.cu +1 -1
  81. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_1-q5_1.cu +1 -1
  82. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_1-q8_0.cu +1 -1
  83. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_0-f16.cu +1 -1
  84. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_0-q4_0.cu +1 -1
  85. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_0-q4_1.cu +1 -1
  86. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_0-q5_0.cu +1 -1
  87. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_0-q5_1.cu +1 -1
  88. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_0-q8_0.cu +1 -1
  89. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_1-f16.cu +1 -1
  90. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_1-q4_0.cu +1 -1
  91. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_1-q4_1.cu +1 -1
  92. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_1-q5_0.cu +1 -1
  93. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_1-q5_1.cu +1 -1
  94. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_1-q8_0.cu +1 -1
  95. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q8_0-f16.cu +1 -1
  96. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q8_0-q4_0.cu +1 -1
  97. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q8_0-q4_1.cu +1 -1
  98. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q8_0-q5_0.cu +1 -1
  99. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q8_0-q5_1.cu +1 -1
  100. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q8_0-q8_0.cu +1 -1
  101. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs256-f16-f16.cu +1 -1
  102. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs64-f16-f16.cu +1 -1
  103. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs64-f16-q4_0.cu +1 -1
  104. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs64-f16-q4_1.cu +1 -1
  105. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs64-f16-q5_0.cu +1 -1
  106. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs64-f16-q5_1.cu +1 -1
  107. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-vec-f32-instance-hs64-f16-q8_0.cu +1 -1
  108. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-wmma-f16-instance-kqfloat-cpb16.cu +1 -1
  109. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-wmma-f16-instance-kqfloat-cpb32.cu +1 -1
  110. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-wmma-f16-instance-kqhalf-cpb16.cu +1 -1
  111. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-wmma-f16-instance-kqhalf-cpb32.cu +1 -1
  112. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/fattn-wmma-f16-instance-kqhalf-cpb8.cu +1 -1
  113. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/mmq-instance-q2_k.cu +5 -0
  114. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/mmq-instance-q3_k.cu +5 -0
  115. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/mmq-instance-q4_0.cu +5 -0
  116. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/mmq-instance-q4_1.cu +5 -0
  117. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/mmq-instance-q4_k.cu +5 -0
  118. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/mmq-instance-q5_0.cu +5 -0
  119. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/mmq-instance-q5_1.cu +5 -0
  120. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/mmq-instance-q5_k.cu +5 -0
  121. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/mmq-instance-q6_k.cu +5 -0
  122. data/vendor/tmp/llama.cpp/ggml-cuda/template-instances/mmq-instance-q8_0.cu +5 -0
  123. data/vendor/tmp/llama.cpp/ggml-cuda/unary.cu +20 -0
  124. data/vendor/tmp/llama.cpp/ggml-cuda.cu +95 -129
  125. data/vendor/tmp/llama.cpp/ggml-kompute.cpp +8 -7
  126. data/vendor/tmp/llama.cpp/ggml-metal.m +11 -9
  127. data/vendor/tmp/llama.cpp/ggml-rpc.cpp +13 -12
  128. data/vendor/tmp/llama.cpp/ggml-sycl.cpp +19 -23
  129. data/vendor/tmp/llama.cpp/ggml-vulkan-shaders.hpp +1230 -1129
  130. data/vendor/tmp/llama.cpp/ggml-vulkan.cpp +181 -148
  131. data/vendor/tmp/llama.cpp/ggml.c +102 -275
  132. data/vendor/tmp/llama.cpp/llama.cpp +103 -47
  133. data/vendor/tmp/llama.cpp/llama.h +4 -0
  134. metadata +15 -3
checksums.yaml CHANGED
@@ -1,7 +1,7 @@
1
1
  ---
2
2
  SHA256:
3
- metadata.gz: 5b79658bc49026edcbd896cac4a1d904060622f2311876afbdba773021399ad1
4
- data.tar.gz: 064fa60e433863e6919f0c0acbd238cf5d5712058cb834a139a5e5cf798d095e
3
+ metadata.gz: '0408c7d579262d0ba34013671a446a76513f6f4564270ef8ba471203fba75c59'
4
+ data.tar.gz: a8085d9678999bb79ba788f7ce856c3f4fe1c6e131af569eaa54aa71fc9ae773
5
5
  SHA512:
6
- metadata.gz: 3248ba69cd0eefcc8b36bdcb03fe13a86da826f4a97a4c61bc62632c2f646647dfaac2b906dd2cb672740c30046e9f588d8e9687b6b8e4bc0a5fc03134d62ec5
7
- data.tar.gz: 91164427363b01f805ae3be98a8f44d7aba0e7c437db7daa2b396bf3329398189613036ac4cb4f5d471194edb02485e32529ca1b9c140144332a0e34107d3666
6
+ metadata.gz: 32e8f294a7f88db05abba3d1e11c951a38f366cac83712f89aa68ed95c581d8eaa4df3d5473f1af3cee965d7a66ea2bf5ccb00222337c59d97fca32ba5e9cade
7
+ data.tar.gz: 2c5f66e2902eb1d72e45261e80f1a2599534e02d0e54e83ff432414d355fab67f6bda9eee095f17904a4fb2d3bf5cb4a1509a346fadff13d999c2f057db972c6
data/CHANGELOG.md CHANGED
@@ -1,3 +1,9 @@
1
+ ## [[0.16.1](https://github.com/yoshoku/llama_cpp.rb/compare/v0.16.0...v0.16.1)] - 2024-06-15
2
+
3
+ - Bump llama.cpp from b3091 to b3151.
4
+ - Add `--with-openblas641` and `--with-no-llamafile` config options.
5
+ - Add `LLAMA_VOCAB_PRE_TYPE_PORO` and `LLAMA_GRETYPE_CHAR_ANY` constants.
6
+
1
7
  ## [[0.16.0](https://github.com/yoshoku/llama_cpp.rb/compare/v0.15.4...v0.16.0)] - 2024-06-08
2
8
 
3
9
  **Breaking Changes**
data/ext/llama_cpp/extconf.rb CHANGED
@@ -14,12 +14,14 @@ make_envs << ' LLAMA_DEBUG=1' if with_config('debug')
14
14
  make_envs << ' LLAMA_QKK_64=1' if with_config('qkk-64')
15
15
  make_envs << ' LLAMA_NO_ACCELERATE=1' if with_config('no-accelerate')
16
16
  make_envs << ' LLAMA_OPENBLAS=1' if with_config('openblas')
17
+ make_envs << ' LLAMA_OPENBLAS64=1' if with_config('openblas64')
17
18
  make_envs << ' LLAMA_BLIS=1' if with_config('blis')
18
19
  make_envs << ' LLAMA_CUBLAS=1' if with_config('cublas') # Deprecated, use --with-cuda instead
19
20
  make_envs << ' LLAMA_CUDA=1' if with_config('cuda')
20
21
  make_envs << ' LLAMA_HIPBLAS=1' if with_config('hipblas')
21
22
  make_envs << ' LLAMA_VULKAN=1' if with_config('vulkan')
22
23
  make_envs << ' LLAMA_NO_OPENMP=1' if with_config('no-openmp')
24
+ make_envs << ' LLAMA_NO_LLAMAFILE=1' if with_config('no-llamafile')
23
25
 
24
26
  make_envs << ' LLAMA_METAL_EMBED_LIBRARY=1' if RUBY_PLATFORM.match?(/darwin/)
25
27
 
data/ext/llama_cpp/llama_cpp.cpp CHANGED
@@ -3494,6 +3494,7 @@ extern "C" void Init_llama_cpp(void) {
3494
3494
  rb_define_const(rb_mLLaMACpp, "LLAMA_VOCAB_PRE_TYPE_OLMO", INT2NUM(LLAMA_VOCAB_PRE_TYPE_OLMO));
3495
3495
  rb_define_const(rb_mLLaMACpp, "LLAMA_VOCAB_PRE_TYPE_DBRX", INT2NUM(LLAMA_VOCAB_PRE_TYPE_DBRX));
3496
3496
  rb_define_const(rb_mLLaMACpp, "LLAMA_VOCAB_PRE_TYPE_SMAUG", INT2NUM(LLAMA_VOCAB_PRE_TYPE_SMAUG));
3497
+ rb_define_const(rb_mLLaMACpp, "LLAMA_VOCAB_PRE_TYPE_PORO", INT2NUM(LLAMA_VOCAB_PRE_TYPE_PORO));
3497
3498
 
3498
3499
  rb_define_const(rb_mLLaMACpp, "LLAMA_TOKEN_TYPE_UNDEFINED", INT2NUM(LLAMA_TOKEN_TYPE_UNDEFINED));
3499
3500
  rb_define_const(rb_mLLaMACpp, "LLAMA_TOKEN_TYPE_NORMAL", INT2NUM(LLAMA_TOKEN_TYPE_NORMAL));
@@ -3559,6 +3560,7 @@ extern "C" void Init_llama_cpp(void) {
3559
3560
  rb_define_const(rb_mLLaMACpp, "LLAMA_GRETYPE_CHAR_NOT", INT2NUM(LLAMA_GRETYPE_CHAR_NOT));
3560
3561
  rb_define_const(rb_mLLaMACpp, "LLAMA_GRETYPE_CHAR_RNG_UPPER", INT2NUM(LLAMA_GRETYPE_CHAR_RNG_UPPER));
3561
3562
  rb_define_const(rb_mLLaMACpp, "LLAMA_GRETYPE_CHAR_ALT", INT2NUM(LLAMA_GRETYPE_CHAR_ALT));
3563
+ rb_define_const(rb_mLLaMACpp, "LLAMA_GRETYPE_CHAR_ANY", INT2NUM(LLAMA_GRETYPE_CHAR_ANY));
3562
3564
 
3563
3565
  rb_define_const(rb_mLLaMACpp, "LLAMA_ROPE_SCALING_TYPE_UNSPECIFIED", INT2NUM(LLAMA_ROPE_SCALING_TYPE_UNSPECIFIED));
3564
3566
  rb_define_const(rb_mLLaMACpp, "LLAMA_ROPE_SCALING_TYPE_NONE", INT2NUM(LLAMA_ROPE_SCALING_TYPE_NONE));
data/lib/llama_cpp/version.rb CHANGED
@@ -3,8 +3,8 @@
3
3
  # llama_cpp.rb provides Ruby bindings for the llama.cpp.
4
4
  module LLaMACpp
5
5
  # The version of llama_cpp.rb you install.
6
- VERSION = '0.16.0'
6
+ VERSION = '0.16.1'
7
7
 
8
8
  # The version of llama.cpp bundled with llama_cpp.rb.
9
- LLAMA_CPP_VERSION = 'b3091'
9
+ LLAMA_CPP_VERSION = 'b3151'
10
10
  end
data/sig/llama_cpp.rbs CHANGED
@@ -31,6 +31,7 @@ module LLaMACpp
31
31
  LLAMA_VOCAB_PRE_TYPE_OLMO: Integer
32
32
  LLAMA_VOCAB_PRE_TYPE_DBRX: Integer
33
33
  LLAMA_VOCAB_PRE_TYPE_SMAUG: Integer
34
+ LLAMA_VOCAB_PRE_TYPE_PORO: Integer
34
35
 
35
36
  LLAMA_TOKEN_ATTR_UNDEFINED: Integer
36
37
  LLAMA_TOKEN_ATTR_UNKNOWN: Integer
@@ -86,6 +87,7 @@ module LLaMACpp
86
87
  LLAMA_GRETYPE_CHAR_NOT: Integer
87
88
  LLAMA_GRETYPE_CHAR_RNG_UPPER: Integer
88
89
  LLAMA_GRETYPE_CHAR_ALT: Integer
90
+ LLAMA_GRETYPE_CHAR_ANY: Integer
89
91
 
90
92
  LLAMA_ROPE_SCALING_TYPE_UNSPECIFIED: Integer
91
93
  LLAMA_ROPE_SCALING_TYPE_NONE: Integer
data/vendor/tmp/llama.cpp/Makefile CHANGED
@@ -1,8 +1,44 @@
1
1
  # Define the default target now so that it is always the first target
2
2
  BUILD_TARGETS = \
3
- main quantize quantize-stats perplexity imatrix embedding vdot q8dot train-text-from-scratch convert-llama2c-to-ggml \
4
- simple batched batched-bench save-load-state server gguf gguf-split eval-callback llama-bench libllava.a llava-cli baby-llama \
5
- retrieval speculative infill tokenize benchmark-matmult parallel finetune export-lora lookahead lookup passkey gritlm tests/test-c.o
3
+ libllava.a \
4
+ llama-baby-llama \
5
+ llama-batched \
6
+ llama-batched-bench \
7
+ llama-bench \
8
+ llama-benchmark-matmult \
9
+ llama-cli \
10
+ llama-convert-llama2c-to-ggml \
11
+ llama-embedding \
12
+ llama-eval-callback \
13
+ llama-export-lora \
14
+ llama-finetune \
15
+ llama-gbnf-validator \
16
+ llama-gguf \
17
+ llama-gguf-split \
18
+ llama-gritlm \
19
+ llama-imatrix \
20
+ llama-infill \
21
+ llama-llava-cli \
22
+ llama-lookahead \
23
+ llama-lookup \
24
+ llama-lookup-create \
25
+ llama-lookup-merge \
26
+ llama-lookup-stats \
27
+ llama-parallel \
28
+ llama-passkey \
29
+ llama-perplexity \
30
+ llama-q8dot \
31
+ llama-quantize \
32
+ llama-quantize-stats \
33
+ llama-retrieval \
34
+ llama-save-load-state \
35
+ llama-server \
36
+ llama-simple \
37
+ llama-speculative \
38
+ llama-tokenize \
39
+ llama-train-text-from-scratch \
40
+ llama-vdot \
41
+ tests/test-c.o
6
42
 
7
43
  # Binaries only useful for tests
8
44
  TEST_TARGETS = \
@@ -406,10 +442,11 @@ ifndef LLAMA_NO_ACCELERATE
406
442
  # Mac OS - include Accelerate framework.
407
443
  # `-framework Accelerate` works both with Apple Silicon and Mac Intel
408
444
  ifeq ($(UNAME_S),Darwin)
409
- MK_CPPFLAGS += -DGGML_USE_ACCELERATE
445
+ MK_CPPFLAGS += -DGGML_USE_ACCELERATE -DGGML_USE_BLAS
410
446
  MK_CPPFLAGS += -DACCELERATE_NEW_LAPACK
411
447
  MK_CPPFLAGS += -DACCELERATE_LAPACK_ILP64
412
448
  MK_LDFLAGS += -framework Accelerate
449
+ OBJS += ggml-blas.o
413
450
  endif
414
451
  endif # LLAMA_NO_ACCELERATE
415
452
 
@@ -420,21 +457,30 @@ ifndef LLAMA_NO_OPENMP
420
457
  endif # LLAMA_NO_OPENMP
421
458
 
422
459
  ifdef LLAMA_OPENBLAS
423
- MK_CPPFLAGS += -DGGML_USE_OPENBLAS $(shell pkg-config --cflags-only-I openblas)
460
+ MK_CPPFLAGS += -DGGML_USE_BLAS $(shell pkg-config --cflags-only-I openblas)
424
461
  MK_CFLAGS += $(shell pkg-config --cflags-only-other openblas)
425
462
  MK_LDFLAGS += $(shell pkg-config --libs openblas)
463
+ OBJS += ggml-blas.o
426
464
  endif # LLAMA_OPENBLAS
427
465
 
428
- ifndef LLAMA_NO_LLAMAFILE
429
- MK_CPPFLAGS += -DGGML_USE_LLAMAFILE
430
- OBJS += sgemm.o
431
- endif
466
+ ifdef LLAMA_OPENBLAS64
467
+ MK_CPPFLAGS += -DGGML_USE_BLAS $(shell pkg-config --cflags-only-I openblas64)
468
+ MK_CFLAGS += $(shell pkg-config --cflags-only-other openblas64)
469
+ MK_LDFLAGS += $(shell pkg-config --libs openblas64)
470
+ OBJS += ggml-blas.o
471
+ endif # LLAMA_OPENBLAS64
432
472
 
433
473
  ifdef LLAMA_BLIS
434
- MK_CPPFLAGS += -DGGML_USE_OPENBLAS -I/usr/local/include/blis -I/usr/include/blis
474
+ MK_CPPFLAGS += -DGGML_USE_BLAS -I/usr/local/include/blis -I/usr/include/blis
435
475
  MK_LDFLAGS += -lblis -L/usr/local/lib
476
+ OBJS += ggml-blas.o
436
477
  endif # LLAMA_BLIS
437
478
 
479
+ ifndef LLAMA_NO_LLAMAFILE
480
+ MK_CPPFLAGS += -DGGML_USE_LLAMAFILE
481
+ OBJS += sgemm.o
482
+ endif
483
+
438
484
  ifdef LLAMA_RPC
439
485
  MK_CPPFLAGS += -DGGML_USE_RPC
440
486
  OBJS += ggml-rpc.o
@@ -446,6 +492,7 @@ ifdef LLAMA_CUBLAS
446
492
  endif
447
493
 
448
494
  OBJS_CUDA_TEMP_INST = $(patsubst %.cu,%.o,$(wildcard ggml-cuda/template-instances/fattn-wmma*.cu))
495
+ OBJS_CUDA_TEMP_INST += $(patsubst %.cu,%.o,$(wildcard ggml-cuda/template-instances/mmq*.cu))
449
496
  ifdef LLAMA_CUDA_FA_ALL_QUANTS
450
497
  OBJS_CUDA_TEMP_INST += $(patsubst %.cu,%.o,$(wildcard ggml-cuda/template-instances/fattn-vec*.cu))
451
498
  else
@@ -741,6 +788,9 @@ ggml-backend.o: ggml-backend.c ggml.h ggml-backend.h
741
788
  ggml-quants.o: ggml-quants.c ggml.h ggml-quants.h ggml-common.h
742
789
  $(CC) $(CFLAGS) -c $< -o $@
743
790
 
791
+ ggml-blas.o: ggml-blas.cpp ggml-blas.h
792
+ $(CXX) $(CXXFLAGS) -c $< -o $@
793
+
744
794
  unicode.o: unicode.cpp unicode.h
745
795
  $(CXX) $(CXXFLAGS) -c $< -o $@
746
796
 
@@ -777,12 +827,13 @@ libllama.so: llama.o ggml.o $(OBJS)
777
827
  libllama.a: llama.o ggml.o $(OBJS) $(COMMON_DEPS)
778
828
  ar rcs libllama.a llama.o ggml.o $(OBJS) $(COMMON_DEPS)
779
829
 
830
+
780
831
  lib: llama.o ggml.o $(OBJS)
781
832
  $(CXX) $(CXXFLAGS) -shared -fPIC -o libllama$(DSO_EXT) $^ $(LDFLAGS)
782
833
  ar rcs libllama.a $^
783
834
 
784
835
  clean:
785
- rm -vrf *.o tests/*.o *.so *.a *.dll *.dylib benchmark-matmult lookup-create lookup-merge lookup-stats common/build-info.cpp *.dot $(COV_TARGETS) $(BUILD_TARGETS) $(TEST_TARGETS)
836
+ rm -vrf *.o tests/*.o *.so *.a *.dll *.dylib common/build-info.cpp *.dot $(COV_TARGETS) $(BUILD_TARGETS) $(TEST_TARGETS)
786
837
  rm -vrf ggml-cuda/*.o
787
838
  rm -vrf ggml-cuda/template-instances/*.o
788
839
 
@@ -797,62 +848,62 @@ clean:
797
848
  # Helper function that replaces .c, .cpp, and .cu file endings with .o:
798
849
  GET_OBJ_FILE = $(patsubst %.c,%.o,$(patsubst %.cpp,%.o,$(patsubst %.cu,%.o,$(1))))
799
850
 
800
- main: examples/main/main.cpp ggml.o llama.o $(COMMON_DEPS) console.o grammar-parser.o $(OBJS)
851
+ llama-cli: examples/main/main.cpp ggml.o llama.o $(COMMON_DEPS) console.o grammar-parser.o $(OBJS)
801
852
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
802
853
  $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
803
854
  @echo
804
- @echo '==== Run ./main -h for help. ===='
855
+ @echo '==== Run ./llama-cli -h for help. ===='
805
856
  @echo
806
857
 
807
- infill: examples/infill/infill.cpp ggml.o llama.o $(COMMON_DEPS) console.o grammar-parser.o $(OBJS)
858
+ llama-infill: examples/infill/infill.cpp ggml.o llama.o $(COMMON_DEPS) console.o grammar-parser.o $(OBJS)
808
859
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
809
860
  $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
810
861
 
811
- simple: examples/simple/simple.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
862
+ llama-simple: examples/simple/simple.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
812
863
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
813
864
  $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
814
865
 
815
- tokenize: examples/tokenize/tokenize.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
866
+ llama-tokenize: examples/tokenize/tokenize.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
816
867
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
817
868
  $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
818
869
 
819
- batched: examples/batched/batched.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
870
+ llama-batched: examples/batched/batched.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
820
871
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
821
872
  $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
822
873
 
823
- batched-bench: examples/batched-bench/batched-bench.cpp build-info.o ggml.o llama.o $(COMMON_DEPS) $(OBJS)
874
+ llama-batched-bench: examples/batched-bench/batched-bench.cpp build-info.o ggml.o llama.o $(COMMON_DEPS) $(OBJS)
824
875
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
825
876
  $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
826
877
 
827
- quantize: examples/quantize/quantize.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
878
+ llama-quantize: examples/quantize/quantize.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
828
879
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
829
880
  $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
830
881
 
831
- quantize-stats: examples/quantize-stats/quantize-stats.cpp build-info.o ggml.o llama.o $(OBJS)
882
+ llama-quantize-stats: examples/quantize-stats/quantize-stats.cpp build-info.o ggml.o llama.o $(OBJS)
832
883
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
833
884
  $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
834
885
 
835
- perplexity: examples/perplexity/perplexity.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
886
+ llama-perplexity: examples/perplexity/perplexity.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
836
887
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
837
888
  $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
838
889
 
839
- imatrix: examples/imatrix/imatrix.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
890
+ llama-imatrix: examples/imatrix/imatrix.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
840
891
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
841
892
  $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
842
893
 
843
- embedding: examples/embedding/embedding.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
894
+ llama-embedding: examples/embedding/embedding.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
844
895
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
845
896
  $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
846
897
 
847
- gritlm: examples/gritlm/gritlm.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
898
+ llama-gritlm: examples/gritlm/gritlm.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
848
899
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
849
900
  $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
850
901
 
851
- save-load-state: examples/save-load-state/save-load-state.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
902
+ llama-save-load-state: examples/save-load-state/save-load-state.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
852
903
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
853
904
  $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
854
905
 
855
- server: examples/server/server.cpp examples/server/utils.hpp examples/server/httplib.h common/json.hpp examples/server/colorthemes.css.hpp examples/server/style.css.hpp examples/server/theme-beeninorder.css.hpp examples/server/theme-ketivah.css.hpp examples/server/theme-mangotango.css.hpp examples/server/theme-playground.css.hpp examples/server/theme-polarnight.css.hpp examples/server/theme-snowstorm.css.hpp examples/server/index.html.hpp examples/server/index-new.html.hpp examples/server/index.js.hpp examples/server/completion.js.hpp examples/server/system-prompts.js.hpp examples/server/prompt-formats.js.hpp examples/server/json-schema-to-grammar.mjs.hpp common/stb_image.h ggml.o llama.o $(COMMON_DEPS) grammar-parser.o $(OBJS)
906
+ llama-server: examples/server/server.cpp examples/server/utils.hpp examples/server/httplib.h common/json.hpp examples/server/colorthemes.css.hpp examples/server/style.css.hpp examples/server/theme-beeninorder.css.hpp examples/server/theme-ketivah.css.hpp examples/server/theme-mangotango.css.hpp examples/server/theme-playground.css.hpp examples/server/theme-polarnight.css.hpp examples/server/theme-snowstorm.css.hpp examples/server/index.html.hpp examples/server/index-new.html.hpp examples/server/index.js.hpp examples/server/completion.js.hpp examples/server/system-prompts.js.hpp examples/server/prompt-formats.js.hpp examples/server/json-schema-to-grammar.mjs.hpp common/stb_image.h ggml.o llama.o $(COMMON_DEPS) grammar-parser.o $(OBJS)
856
907
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
857
908
  $(CXX) $(CXXFLAGS) $(filter-out %.h %.hpp $<,$^) -Iexamples/server $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS) $(LWINSOCK2)
858
909
 
@@ -865,23 +916,23 @@ examples/server/%.hpp: examples/server/public/% Makefile
865
916
  echo "unsigned int $${NAME}_len = $(shell cat $< | wc -c );" \
866
917
  ) > $@
867
918
 
868
- gguf: examples/gguf/gguf.cpp ggml.o $(OBJS)
919
+ llama-gguf: examples/gguf/gguf.cpp ggml.o $(OBJS)
869
920
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
870
921
  $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
871
922
 
872
- gguf-split: examples/gguf-split/gguf-split.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
923
+ llama-gguf-split: examples/gguf-split/gguf-split.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
873
924
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
874
925
  $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
875
926
 
876
- eval-callback: examples/eval-callback/eval-callback.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
927
+ llama-eval-callback: examples/eval-callback/eval-callback.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
877
928
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
878
929
  $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
879
930
 
880
- train-text-from-scratch: examples/train-text-from-scratch/train-text-from-scratch.cpp ggml.o llama.o $(COMMON_DEPS) train.o $(OBJS)
931
+ llama-train-text-from-scratch: examples/train-text-from-scratch/train-text-from-scratch.cpp ggml.o llama.o $(COMMON_DEPS) train.o $(OBJS)
881
932
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
882
933
  $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
883
934
 
884
- convert-llama2c-to-ggml: examples/convert-llama2c-to-ggml/convert-llama2c-to-ggml.cpp ggml.o llama.o $(OBJS)
935
+ llama-convert-llama2c-to-ggml: examples/convert-llama2c-to-ggml/convert-llama2c-to-ggml.cpp ggml.o llama.o $(OBJS)
885
936
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
886
937
  $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
887
938
 
@@ -892,55 +943,61 @@ llama-bench: examples/llama-bench/llama-bench.cpp ggml.o llama.o $(COMMON_DEPS)
892
943
  libllava.a: examples/llava/llava.cpp examples/llava/llava.h examples/llava/clip.cpp examples/llava/clip.h common/stb_image.h common/base64.hpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
893
944
  $(CXX) $(CXXFLAGS) -static -fPIC -c $< -o $@ -Wno-cast-qual
894
945
 
895
- llava-cli: examples/llava/llava-cli.cpp examples/llava/clip.h examples/llava/clip.cpp examples/llava/llava.h examples/llava/llava.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
946
+ llama-llava-cli: examples/llava/llava-cli.cpp examples/llava/clip.h examples/llava/clip.cpp examples/llava/llava.h examples/llava/llava.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
896
947
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
897
948
  $(CXX) $(CXXFLAGS) -c examples/llava/clip.cpp -o $(call GET_OBJ_FILE, examples/llava/clip.cpp) -Wno-cast-qual
898
949
  $(CXX) $(CXXFLAGS) -c examples/llava/llava.cpp -o $(call GET_OBJ_FILE, examples/llava/llava.cpp)
899
950
  $(CXX) $(CXXFLAGS) $(filter-out %.h $< examples/llava/clip.cpp examples/llava/llava.cpp,$^) $(call GET_OBJ_FILE, $<) $(call GET_OBJ_FILE, examples/llava/clip.cpp) $(call GET_OBJ_FILE, examples/llava/llava.cpp) -o $@ $(LDFLAGS)
900
951
 
901
- baby-llama: examples/baby-llama/baby-llama.cpp ggml.o llama.o $(COMMON_DEPS) train.o $(OBJS)
952
+ llama-baby-llama: examples/baby-llama/baby-llama.cpp ggml.o llama.o $(COMMON_DEPS) train.o $(OBJS)
953
+ $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
954
+ $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
955
+
956
+ llama-finetune: examples/finetune/finetune.cpp ggml.o llama.o $(COMMON_DEPS) train.o $(OBJS)
902
957
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
903
958
  $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
904
959
 
905
- finetune: examples/finetune/finetune.cpp ggml.o llama.o $(COMMON_DEPS) train.o $(OBJS)
960
+ llama-export-lora: examples/export-lora/export-lora.cpp ggml.o common/common.h $(OBJS)
906
961
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
907
962
  $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
908
963
 
909
- export-lora: examples/export-lora/export-lora.cpp ggml.o common/common.h $(OBJS)
964
+ llama-retrieval: examples/retrieval/retrieval.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
910
965
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
911
966
  $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
912
967
 
913
- retrieval: examples/retrieval/retrieval.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
968
+ llama-speculative: examples/speculative/speculative.cpp ggml.o llama.o $(COMMON_DEPS) grammar-parser.o $(OBJS)
914
969
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
915
970
  $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
916
971
 
917
- speculative: examples/speculative/speculative.cpp ggml.o llama.o $(COMMON_DEPS) grammar-parser.o $(OBJS)
972
+ llama-parallel: examples/parallel/parallel.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
918
973
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
919
974
  $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
920
975
 
921
- parallel: examples/parallel/parallel.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
976
+ llama-lookahead: examples/lookahead/lookahead.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
922
977
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
923
978
  $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
924
979
 
925
- lookahead: examples/lookahead/lookahead.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
980
+ llama-lookup: examples/lookup/lookup.cpp ggml.o llama.o ngram-cache.o $(COMMON_DEPS) $(OBJS)
926
981
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
927
982
  $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
928
983
 
929
- lookup: examples/lookup/lookup.cpp ggml.o llama.o ngram-cache.o $(COMMON_DEPS) $(OBJS)
984
+ llama-lookup-create: examples/lookup/lookup-create.cpp ggml.o llama.o ngram-cache.o $(COMMON_DEPS) $(OBJS)
930
985
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
931
986
  $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
932
- $(CXX) $(CXXFLAGS) -c examples/lookup/lookup-create.cpp -o $(call GET_OBJ_FILE, examples/lookup/lookup-create.cpp)
933
- $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, examples/lookup/lookup-create.cpp) -o lookup-create $(LDFLAGS)
934
- $(CXX) $(CXXFLAGS) -c examples/lookup/lookup-merge.cpp -o $(call GET_OBJ_FILE, examples/lookup/lookup-merge.cpp)
935
- $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, examples/lookup/lookup-merge.cpp) -o lookup-merge $(LDFLAGS)
936
- $(CXX) $(CXXFLAGS) -c examples/lookup/lookup-stats.cpp -o $(call GET_OBJ_FILE, examples/lookup/lookup-stats.cpp)
937
- $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, examples/lookup/lookup-stats.cpp) -o lookup-stats $(LDFLAGS)
938
-
939
- passkey: examples/passkey/passkey.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
987
+
988
+ llama-lookup-merge: examples/lookup/lookup-merge.cpp ggml.o llama.o ngram-cache.o $(COMMON_DEPS) $(OBJS)
989
+ $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
990
+ $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
991
+
992
+ llama-lookup-stats: examples/lookup/lookup-stats.cpp ggml.o llama.o ngram-cache.o $(COMMON_DEPS) $(OBJS)
993
+ $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
994
+ $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
995
+
996
+ llama-passkey: examples/passkey/passkey.cpp ggml.o llama.o $(COMMON_DEPS) $(OBJS)
940
997
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
941
998
  $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
942
999
 
943
- gbnf-validator: examples/gbnf-validator/gbnf-validator.cpp ggml.o llama.o $(COMMON_DEPS) grammar-parser.o $(OBJS)
1000
+ llama-gbnf-validator: examples/gbnf-validator/gbnf-validator.cpp ggml.o llama.o $(COMMON_DEPS) grammar-parser.o $(OBJS)
944
1001
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
945
1002
  $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
946
1003
 
@@ -966,20 +1023,20 @@ build-info.o: common/build-info.cpp
966
1023
 
967
1024
  tests: $(TEST_TARGETS)
968
1025
 
969
- benchmark-matmult: examples/benchmark/benchmark-matmult.cpp build-info.o ggml.o $(OBJS)
1026
+ llama-benchmark-matmult: examples/benchmark/benchmark-matmult.cpp build-info.o ggml.o $(OBJS)
970
1027
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
971
1028
  $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
972
1029
 
973
- run-benchmark-matmult: benchmark-matmult
1030
+ run-benchmark-matmult: llama-benchmark-matmult
974
1031
  ./$@
975
1032
 
976
1033
  .PHONY: run-benchmark-matmult swift
977
1034
 
978
- vdot: pocs/vdot/vdot.cpp ggml.o $(OBJS)
1035
+ llama-vdot: pocs/vdot/vdot.cpp ggml.o $(OBJS)
979
1036
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
980
1037
  $(CXX) $(CXXFLAGS) $(filter-out $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
981
1038
 
982
- q8dot: pocs/vdot/q8dot.cpp ggml.o $(OBJS)
1039
+ llama-q8dot: pocs/vdot/q8dot.cpp ggml.o $(OBJS)
983
1040
  $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
984
1041
  $(CXX) $(CXXFLAGS) $(filter-out $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
985
1042