npm - cui-llama.rn - Versions diffs - 1.4.0 → 1.4.1 - Mend

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

package/android/src/main/jni.cpp +9 -9
package/cpp/common.cpp +163 -60
package/cpp/common.h +43 -12
package/cpp/ggml-alloc.c +1042 -1037
package/cpp/ggml-backend-impl.h +255 -256
package/cpp/ggml-backend-reg.cpp +582 -582
package/cpp/ggml-backend.cpp +2002 -2002
package/cpp/ggml-backend.h +354 -352
package/cpp/ggml-common.h +1853 -1853
package/cpp/ggml-cpp.h +39 -39
package/cpp/ggml-cpu-aarch64.cpp +4247 -4247
package/cpp/ggml-cpu-aarch64.h +8 -8
package/cpp/ggml-cpu-impl.h +386 -386
package/cpp/ggml-cpu-quants.c +10920 -10839
package/cpp/ggml-cpu-traits.cpp +36 -36
package/cpp/ggml-cpu-traits.h +38 -38
package/cpp/ggml-cpu.c +329 -60
package/cpp/ggml-cpu.cpp +10 -2
package/cpp/ggml-cpu.h +135 -135
package/cpp/ggml-impl.h +567 -567
package/cpp/ggml-metal-impl.h +17 -17
package/cpp/ggml-metal.m +4884 -4884
package/cpp/ggml-quants.c +5238 -5238
package/cpp/ggml-threading.h +14 -14
package/cpp/ggml.c +6514 -6448
package/cpp/ggml.h +2194 -2163
package/cpp/gguf.cpp +1329 -1325
package/cpp/gguf.h +202 -202
package/cpp/json-schema-to-grammar.cpp +1045 -1045
package/cpp/json-schema-to-grammar.h +8 -8
package/cpp/json.hpp +24766 -24766
package/cpp/llama-adapter.cpp +347 -346
package/cpp/llama-adapter.h +74 -73
package/cpp/llama-arch.cpp +1487 -1434
package/cpp/llama-arch.h +400 -395
package/cpp/llama-batch.cpp +368 -368
package/cpp/llama-batch.h +88 -88
package/cpp/llama-chat.cpp +578 -567
package/cpp/llama-chat.h +52 -51
package/cpp/llama-context.cpp +1775 -1771
package/cpp/llama-context.h +128 -128
package/cpp/llama-cparams.cpp +1 -1
package/cpp/llama-cparams.h +37 -37
package/cpp/llama-cpp.h +30 -30
package/cpp/llama-grammar.cpp +1139 -1139
package/cpp/llama-grammar.h +143 -143
package/cpp/llama-hparams.cpp +71 -71
package/cpp/llama-hparams.h +139 -140
package/cpp/llama-impl.cpp +167 -167
package/cpp/llama-impl.h +61 -61
package/cpp/llama-kv-cache.cpp +718 -718
package/cpp/llama-kv-cache.h +218 -218
package/cpp/llama-mmap.cpp +2 -1
package/cpp/llama-mmap.h +67 -67
package/cpp/llama-model-loader.cpp +1124 -1011
package/cpp/llama-model-loader.h +167 -158
package/cpp/llama-model.cpp +3997 -2202
package/cpp/llama-model.h +370 -391
package/cpp/llama-sampling.cpp +2408 -2406
package/cpp/llama-sampling.h +32 -48
package/cpp/llama-vocab.cpp +3247 -1982
package/cpp/llama-vocab.h +125 -182
package/cpp/llama.cpp +416 -2886
package/cpp/llama.h +1323 -1285
package/cpp/log.cpp +401 -401
package/cpp/log.h +121 -121
package/cpp/rn-llama.hpp +18 -12
package/cpp/sampling.cpp +505 -500
package/cpp/sgemm.cpp +2597 -2597
package/cpp/speculative.cpp +277 -274
package/cpp/speculative.h +28 -28
package/cpp/unicode.cpp +2 -3
package/package.json +1 -1

package/cpp/ggml-cpp.h CHANGED Viewed

@@ -1,39 +1,39 @@
-#pragma once
-#ifndef __cplusplus
-#error "This header is for C++ only"
-#endif
-#include "ggml.h"
-#include "ggml-alloc.h"
-#include "ggml-backend.h"
-#include "gguf.h"
-#include <memory>
-// Smart pointers for ggml types
-// ggml
-struct lm_ggml_context_deleter { void operator()(lm_ggml_context * ctx) { lm_ggml_free(ctx); } };
-struct lm_gguf_context_deleter { void operator()(lm_gguf_context * ctx) { lm_gguf_free(ctx); } };
-typedef std::unique_ptr<lm_ggml_context, lm_ggml_context_deleter> lm_ggml_context_ptr;
-typedef std::unique_ptr<lm_gguf_context, lm_gguf_context_deleter> lm_gguf_context_ptr;
-// ggml-alloc
-struct lm_ggml_gallocr_deleter { void operator()(lm_ggml_gallocr_t galloc) { lm_ggml_gallocr_free(galloc); } };
-typedef std::unique_ptr<lm_ggml_gallocr_t, lm_ggml_gallocr_deleter> lm_ggml_gallocr_ptr;
-// ggml-backend
-struct lm_ggml_backend_deleter        { void operator()(lm_ggml_backend_t backend)       { lm_ggml_backend_free(backend); } };
-struct lm_ggml_backend_buffer_deleter { void operator()(lm_ggml_backend_buffer_t buffer) { lm_ggml_backend_buffer_free(buffer); } };
-struct lm_ggml_backend_event_deleter  { void operator()(lm_ggml_backend_event_t event)   { lm_ggml_backend_event_free(event); } };
-struct lm_ggml_backend_sched_deleter  { void operator()(lm_ggml_backend_sched_t sched)   { lm_ggml_backend_sched_free(sched); } };
-typedef std::unique_ptr<lm_ggml_backend,        lm_ggml_backend_deleter>        lm_ggml_backend_ptr;
-typedef std::unique_ptr<lm_ggml_backend_buffer, lm_ggml_backend_buffer_deleter> lm_ggml_backend_buffer_ptr;
-typedef std::unique_ptr<lm_ggml_backend_event,  lm_ggml_backend_event_deleter>  lm_ggml_backend_event_ptr;
-typedef std::unique_ptr<lm_ggml_backend_sched,  lm_ggml_backend_sched_deleter>  lm_ggml_backend_sched_ptr;
+#pragma once
+#ifndef __cplusplus
+#error "This header is for C++ only"
+#endif
+#include "ggml.h"
+#include "ggml-alloc.h"
+#include "ggml-backend.h"
+#include "gguf.h"
+#include <memory>
+// Smart pointers for ggml types
+// ggml
+struct lm_ggml_context_deleter { void operator()(lm_ggml_context * ctx) { lm_ggml_free(ctx); } };
+struct lm_gguf_context_deleter { void operator()(lm_gguf_context * ctx) { lm_gguf_free(ctx); } };
+typedef std::unique_ptr<lm_ggml_context, lm_ggml_context_deleter> lm_ggml_context_ptr;
+typedef std::unique_ptr<lm_gguf_context, lm_gguf_context_deleter> lm_gguf_context_ptr;
+// ggml-alloc
+struct lm_ggml_gallocr_deleter { void operator()(lm_ggml_gallocr_t galloc) { lm_ggml_gallocr_free(galloc); } };
+typedef std::unique_ptr<lm_ggml_gallocr_t, lm_ggml_gallocr_deleter> lm_ggml_gallocr_ptr;
+// ggml-backend
+struct lm_ggml_backend_deleter        { void operator()(lm_ggml_backend_t backend)       { lm_ggml_backend_free(backend); } };
+struct lm_ggml_backend_buffer_deleter { void operator()(lm_ggml_backend_buffer_t buffer) { lm_ggml_backend_buffer_free(buffer); } };
+struct lm_ggml_backend_event_deleter  { void operator()(lm_ggml_backend_event_t event)   { lm_ggml_backend_event_free(event); } };
+struct lm_ggml_backend_sched_deleter  { void operator()(lm_ggml_backend_sched_t sched)   { lm_ggml_backend_sched_free(sched); } };
+typedef std::unique_ptr<lm_ggml_backend,        lm_ggml_backend_deleter>        lm_ggml_backend_ptr;
+typedef std::unique_ptr<lm_ggml_backend_buffer, lm_ggml_backend_buffer_deleter> lm_ggml_backend_buffer_ptr;
+typedef std::unique_ptr<lm_ggml_backend_event,  lm_ggml_backend_event_deleter>  lm_ggml_backend_event_ptr;
+typedef std::unique_ptr<lm_ggml_backend_sched,  lm_ggml_backend_sched_deleter>  lm_ggml_backend_sched_ptr;