cui-llama.rn 1.1.4 → 1.1.5
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.
- package/android/src/main/jni.cpp +2 -2
- package/cpp/common.cpp +35 -1946
- package/cpp/common.h +91 -128
- package/cpp/ggml-impl.h +32 -0
- package/cpp/ggml-metal.m +5 -6
- package/cpp/ggml-quants.c +242 -48
- package/cpp/ggml.c +89 -35
- package/cpp/ggml.h +25 -63
- package/cpp/llama-sampling.cpp +218 -94
- package/cpp/llama.cpp +80 -86
- package/cpp/llama.h +36 -11
- package/cpp/rn-llama.hpp +2 -1
- package/cpp/sampling.cpp +11 -4
- package/cpp/sampling.h +4 -56
- package/package.json +1 -1
package/android/src/main/jni.cpp
CHANGED
@@ -523,7 +523,7 @@ Java_com_rnllama_LlamaContext_doCompletion(
|
|
523
523
|
}
|
524
524
|
}
|
525
525
|
|
526
|
-
|
526
|
+
llama_perf_context_print(llama->ctx);
|
527
527
|
llama->is_predicting = false;
|
528
528
|
|
529
529
|
auto result = createWriteableMap(env);
|
@@ -636,7 +636,7 @@ Java_com_rnllama_LlamaContext_embedding(
|
|
636
636
|
llama->rewind();
|
637
637
|
|
638
638
|
// llama_reset_timings(llama->ctx);
|
639
|
-
|
639
|
+
llama_perf_context_reset(llama->ctx);
|
640
640
|
gpt_sampler_reset(llama->ctx_sampling);
|
641
641
|
|
642
642
|
|