cui-llama.rn 1.4.0 → 1.4.1

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 (73) hide show
  1. package/android/src/main/jni.cpp +9 -9
  2. package/cpp/common.cpp +163 -60
  3. package/cpp/common.h +43 -12
  4. package/cpp/ggml-alloc.c +1042 -1037
  5. package/cpp/ggml-backend-impl.h +255 -256
  6. package/cpp/ggml-backend-reg.cpp +582 -582
  7. package/cpp/ggml-backend.cpp +2002 -2002
  8. package/cpp/ggml-backend.h +354 -352
  9. package/cpp/ggml-common.h +1853 -1853
  10. package/cpp/ggml-cpp.h +39 -39
  11. package/cpp/ggml-cpu-aarch64.cpp +4247 -4247
  12. package/cpp/ggml-cpu-aarch64.h +8 -8
  13. package/cpp/ggml-cpu-impl.h +386 -386
  14. package/cpp/ggml-cpu-quants.c +10920 -10839
  15. package/cpp/ggml-cpu-traits.cpp +36 -36
  16. package/cpp/ggml-cpu-traits.h +38 -38
  17. package/cpp/ggml-cpu.c +329 -60
  18. package/cpp/ggml-cpu.cpp +10 -2
  19. package/cpp/ggml-cpu.h +135 -135
  20. package/cpp/ggml-impl.h +567 -567
  21. package/cpp/ggml-metal-impl.h +17 -17
  22. package/cpp/ggml-metal.m +4884 -4884
  23. package/cpp/ggml-quants.c +5238 -5238
  24. package/cpp/ggml-threading.h +14 -14
  25. package/cpp/ggml.c +6514 -6448
  26. package/cpp/ggml.h +2194 -2163
  27. package/cpp/gguf.cpp +1329 -1325
  28. package/cpp/gguf.h +202 -202
  29. package/cpp/json-schema-to-grammar.cpp +1045 -1045
  30. package/cpp/json-schema-to-grammar.h +8 -8
  31. package/cpp/json.hpp +24766 -24766
  32. package/cpp/llama-adapter.cpp +347 -346
  33. package/cpp/llama-adapter.h +74 -73
  34. package/cpp/llama-arch.cpp +1487 -1434
  35. package/cpp/llama-arch.h +400 -395
  36. package/cpp/llama-batch.cpp +368 -368
  37. package/cpp/llama-batch.h +88 -88
  38. package/cpp/llama-chat.cpp +578 -567
  39. package/cpp/llama-chat.h +52 -51
  40. package/cpp/llama-context.cpp +1775 -1771
  41. package/cpp/llama-context.h +128 -128
  42. package/cpp/llama-cparams.cpp +1 -1
  43. package/cpp/llama-cparams.h +37 -37
  44. package/cpp/llama-cpp.h +30 -30
  45. package/cpp/llama-grammar.cpp +1139 -1139
  46. package/cpp/llama-grammar.h +143 -143
  47. package/cpp/llama-hparams.cpp +71 -71
  48. package/cpp/llama-hparams.h +139 -140
  49. package/cpp/llama-impl.cpp +167 -167
  50. package/cpp/llama-impl.h +61 -61
  51. package/cpp/llama-kv-cache.cpp +718 -718
  52. package/cpp/llama-kv-cache.h +218 -218
  53. package/cpp/llama-mmap.cpp +2 -1
  54. package/cpp/llama-mmap.h +67 -67
  55. package/cpp/llama-model-loader.cpp +1124 -1011
  56. package/cpp/llama-model-loader.h +167 -158
  57. package/cpp/llama-model.cpp +3997 -2202
  58. package/cpp/llama-model.h +370 -391
  59. package/cpp/llama-sampling.cpp +2408 -2406
  60. package/cpp/llama-sampling.h +32 -48
  61. package/cpp/llama-vocab.cpp +3247 -1982
  62. package/cpp/llama-vocab.h +125 -182
  63. package/cpp/llama.cpp +416 -2886
  64. package/cpp/llama.h +1323 -1285
  65. package/cpp/log.cpp +401 -401
  66. package/cpp/log.h +121 -121
  67. package/cpp/rn-llama.hpp +18 -12
  68. package/cpp/sampling.cpp +505 -500
  69. package/cpp/sgemm.cpp +2597 -2597
  70. package/cpp/speculative.cpp +277 -274
  71. package/cpp/speculative.h +28 -28
  72. package/cpp/unicode.cpp +2 -3
  73. package/package.json +1 -1
package/cpp/llama-sampling.h CHANGED
@@ -1,48 +1,32 @@
1
- #pragma once
2
-
3
- // TODO: rename llama-sampling.h/.cpp to llama-sampler.h/.cpp ?
4
-
5
- #include "llama-grammar.h"
6
-
7
- struct llama_vocab;
8
- struct llama_grammar;
9
-
10
- // sampler chain
11
-
12
- struct llama_sampler_chain {
13
- llama_sampler_chain_params params;
14
-
15
- std::vector<struct llama_sampler *> samplers;
16
-
17
- // timing
18
-
19
- mutable int64_t t_sample_us;
20
-
21
- mutable int32_t n_sample;
22
- };
23
-
24
- struct llama_sampler * llama_sampler_init_grammar_impl(
25
- const struct llama_vocab & vocab,
26
- const char * grammar_str,
27
- const char * grammar_root);
28
-
29
- struct llama_sampler * llama_sampler_init_infill_impl(
30
- const struct llama_vocab & vocab);
31
-
32
- struct llama_sampler * llama_sampler_init_dry_impl(
33
- const struct llama_vocab & vocab,
34
- int32_t context_size,
35
- float dry_multiplier,
36
- float dry_base,
37
- int32_t dry_allowed_length,
38
- int32_t dry_penalty_last_n,
39
- const char ** seq_breakers,
40
- size_t num_breakers);
41
-
42
- struct llama_sampler * llama_sampler_init_dry_testing(
43
- int32_t context_size,
44
- float dry_multiplier,
45
- float dry_base,
46
- int32_t dry_allowed_length,
47
- int32_t dry_penalty_last_n,
48
- const std::vector<std::vector<llama_token>>& seq_breakers);
1
+ #pragma once
2
+
3
+ // TODO: rename llama-sampling.h/.cpp to llama-sampler.h/.cpp ?
4
+
5
+ #include "llama.h"
6
+
7
+ #include <vector>
8
+
9
+ struct llama_vocab;
10
+ struct llama_grammar;
11
+
12
+ // sampler chain
13
+
14
+ struct llama_sampler_chain {
15
+ llama_sampler_chain_params params;
16
+
17
+ std::vector<struct llama_sampler *> samplers;
18
+
19
+ // timing
20
+
21
+ mutable int64_t t_sample_us;
22
+
23
+ mutable int32_t n_sample;
24
+ };
25
+
26
+ struct llama_sampler * llama_sampler_init_dry_testing(
27
+ int32_t context_size,
28
+ float dry_multiplier,
29
+ float dry_base,
30
+ int32_t dry_allowed_length,
31
+ int32_t dry_penalty_last_n,
32
+ const std::vector<std::vector<llama_token>>& seq_breakers);