npm - @fugood/llama.node - Versions diffs - 0.3.6 → 0.3.8 - Mend

@fugood/llama.node 0.3.6 → 0.3.8

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 (186) hide show

package/README.md +17 -2
package/bin/darwin/arm64/llama-node.node +0 -0
package/bin/darwin/x64/llama-node.node +0 -0
package/bin/linux/arm64/llama-node.node +0 -0
package/bin/linux/x64/llama-node.node +0 -0
package/bin/linux-cuda/arm64/llama-node.node +0 -0
package/bin/linux-cuda/x64/llama-node.node +0 -0
package/bin/linux-vulkan/arm64/llama-node.node +0 -0
package/bin/linux-vulkan/x64/llama-node.node +0 -0
package/bin/win32/arm64/llama-node.node +0 -0
package/bin/win32/arm64/node.lib +0 -0
package/bin/win32/x64/llama-node.node +0 -0
package/bin/win32/x64/node.lib +0 -0
package/bin/win32-vulkan/arm64/llama-node.node +0 -0
package/bin/win32-vulkan/arm64/node.lib +0 -0
package/bin/win32-vulkan/x64/llama-node.node +0 -0
package/bin/win32-vulkan/x64/node.lib +0 -0
package/lib/binding.ts +3 -1
package/lib/index.js +16 -1
package/lib/index.ts +16 -0
package/package.json +1 -1
package/src/EmbeddingWorker.cpp +4 -3
package/src/LlamaCompletionWorker.cpp +4 -2
package/src/LlamaContext.cpp +61 -6
package/src/LlamaContext.h +1 -0
package/src/common.hpp +6 -11
package/src/llama.cpp/.github/workflows/build.yml +19 -17
package/src/llama.cpp/.github/workflows/docker.yml +77 -30
package/src/llama.cpp/.github/workflows/editorconfig.yml +3 -1
package/src/llama.cpp/.github/workflows/server.yml +22 -3
package/src/llama.cpp/CMakeLists.txt +49 -24
package/src/llama.cpp/common/arg.cpp +82 -26
package/src/llama.cpp/common/arg.h +3 -0
package/src/llama.cpp/common/common.cpp +192 -72
package/src/llama.cpp/common/common.h +51 -18
package/src/llama.cpp/common/ngram-cache.cpp +12 -12
package/src/llama.cpp/common/ngram-cache.h +2 -2
package/src/llama.cpp/common/sampling.cpp +11 -6
package/src/llama.cpp/common/speculative.cpp +18 -15
package/src/llama.cpp/docs/build.md +2 -0
package/src/llama.cpp/examples/batched/batched.cpp +9 -7
package/src/llama.cpp/examples/batched-bench/batched-bench.cpp +3 -3
package/src/llama.cpp/examples/convert-llama2c-to-ggml/convert-llama2c-to-ggml.cpp +10 -8
package/src/llama.cpp/examples/cvector-generator/cvector-generator.cpp +11 -8
package/src/llama.cpp/examples/cvector-generator/mean.hpp +1 -1
package/src/llama.cpp/examples/cvector-generator/pca.hpp +1 -1
package/src/llama.cpp/examples/embedding/embedding.cpp +8 -7
package/src/llama.cpp/examples/eval-callback/eval-callback.cpp +7 -6
package/src/llama.cpp/examples/export-lora/export-lora.cpp +8 -7
package/src/llama.cpp/examples/gguf/gguf.cpp +10 -6
package/src/llama.cpp/examples/gguf-hash/gguf-hash.cpp +1 -0
package/src/llama.cpp/examples/gguf-split/gguf-split.cpp +8 -7
package/src/llama.cpp/examples/gritlm/gritlm.cpp +13 -10
package/src/llama.cpp/examples/imatrix/imatrix.cpp +13 -12
package/src/llama.cpp/examples/infill/infill.cpp +23 -24
package/src/llama.cpp/examples/llama-bench/llama-bench.cpp +44 -13
package/src/llama.cpp/examples/llama.android/llama/src/main/cpp/llama-android.cpp +11 -6
package/src/llama.cpp/examples/llava/clip.cpp +4 -2
package/src/llama.cpp/examples/llava/llava-cli.cpp +9 -6
package/src/llama.cpp/examples/llava/llava.cpp +2 -2
package/src/llama.cpp/examples/llava/minicpmv-cli.cpp +8 -4
package/src/llama.cpp/examples/llava/qwen2vl-cli.cpp +11 -8
package/src/llama.cpp/examples/lookahead/lookahead.cpp +6 -7
package/src/llama.cpp/examples/lookup/lookup-create.cpp +4 -9
package/src/llama.cpp/examples/lookup/lookup-stats.cpp +3 -7
package/src/llama.cpp/examples/lookup/lookup.cpp +5 -6
package/src/llama.cpp/examples/main/main.cpp +51 -29
package/src/llama.cpp/examples/parallel/parallel.cpp +5 -6
package/src/llama.cpp/examples/passkey/passkey.cpp +7 -5
package/src/llama.cpp/examples/perplexity/perplexity.cpp +37 -23
package/src/llama.cpp/examples/quantize-stats/quantize-stats.cpp +12 -14
package/src/llama.cpp/examples/retrieval/retrieval.cpp +8 -8
package/src/llama.cpp/examples/rpc/rpc-server.cpp +12 -0
package/src/llama.cpp/examples/run/CMakeLists.txt +1 -1
package/src/llama.cpp/examples/run/linenoise.cpp/linenoise.cpp +1351 -0
package/src/llama.cpp/examples/run/linenoise.cpp/linenoise.h +114 -0
package/src/llama.cpp/examples/run/run.cpp +175 -61
package/src/llama.cpp/examples/save-load-state/save-load-state.cpp +4 -25
package/src/llama.cpp/examples/server/CMakeLists.txt +1 -0
package/src/llama.cpp/examples/server/httplib.h +1295 -409
package/src/llama.cpp/examples/server/server.cpp +387 -181
package/src/llama.cpp/examples/server/tests/requirements.txt +1 -0
package/src/llama.cpp/examples/server/utils.hpp +170 -58
package/src/llama.cpp/examples/simple/simple.cpp +9 -8
package/src/llama.cpp/examples/simple-chat/simple-chat.cpp +16 -12
package/src/llama.cpp/examples/speculative/speculative.cpp +22 -23
package/src/llama.cpp/examples/speculative-simple/speculative-simple.cpp +8 -12
package/src/llama.cpp/examples/tokenize/tokenize.cpp +17 -5
package/src/llama.cpp/examples/tts/tts.cpp +64 -23
package/src/llama.cpp/ggml/CMakeLists.txt +5 -21
package/src/llama.cpp/ggml/include/ggml-backend.h +2 -0
package/src/llama.cpp/ggml/include/ggml-cpp.h +1 -0
package/src/llama.cpp/ggml/include/ggml.h +36 -145
package/src/llama.cpp/ggml/include/gguf.h +202 -0
package/src/llama.cpp/ggml/src/CMakeLists.txt +6 -3
package/src/llama.cpp/ggml/src/ggml-alloc.c +5 -0
package/src/llama.cpp/ggml/src/ggml-backend-impl.h +0 -1
package/src/llama.cpp/ggml/src/ggml-backend-reg.cpp +79 -49
package/src/llama.cpp/ggml/src/ggml-backend.cpp +5 -2
package/src/llama.cpp/ggml/src/ggml-cpu/CMakeLists.txt +33 -23
package/src/llama.cpp/ggml/src/ggml-cpu/ggml-cpu-aarch64.cpp +57 -72
package/src/llama.cpp/ggml/src/ggml-cpu/ggml-cpu-quants.c +87 -2
package/src/llama.cpp/ggml/src/ggml-cpu/ggml-cpu.c +335 -66
package/src/llama.cpp/ggml/src/ggml-cpu/ggml-cpu.cpp +10 -2
package/src/llama.cpp/ggml/src/ggml-cpu/llamafile/sgemm.cpp +1090 -378
package/src/llama.cpp/ggml/src/ggml-cpu/llamafile/sgemm.h +2 -2
package/src/llama.cpp/ggml/src/ggml-cuda/vendors/cuda.h +1 -0
package/src/llama.cpp/ggml/src/ggml-cuda/vendors/hip.h +3 -0
package/src/llama.cpp/ggml/src/ggml-cuda/vendors/musa.h +3 -0
package/src/llama.cpp/ggml/src/ggml-hip/CMakeLists.txt +3 -1
package/src/llama.cpp/ggml/src/ggml-impl.h +11 -16
package/src/llama.cpp/ggml/src/ggml-metal/CMakeLists.txt +16 -0
package/src/llama.cpp/ggml/src/ggml-opencl/ggml-opencl.cpp +6 -6
package/src/llama.cpp/ggml/src/ggml-rpc/ggml-rpc.cpp +154 -35
package/src/llama.cpp/ggml/src/ggml-sycl/backend.hpp +1 -0
package/src/llama.cpp/ggml/src/ggml-sycl/common.cpp +9 -3
package/src/llama.cpp/ggml/src/ggml-sycl/common.hpp +18 -0
package/src/llama.cpp/ggml/src/ggml-sycl/concat.cpp +3 -2
package/src/llama.cpp/ggml/src/ggml-sycl/concat.hpp +1 -2
package/src/llama.cpp/ggml/src/ggml-sycl/conv.cpp +3 -2
package/src/llama.cpp/ggml/src/ggml-sycl/conv.hpp +1 -2
package/src/llama.cpp/ggml/src/ggml-sycl/dpct/helper.hpp +40 -95
package/src/llama.cpp/ggml/src/ggml-sycl/element_wise.cpp +48 -48
package/src/llama.cpp/ggml/src/ggml-sycl/element_wise.hpp +24 -24
package/src/llama.cpp/ggml/src/ggml-sycl/ggml-sycl.cpp +238 -164
package/src/llama.cpp/ggml/src/ggml-sycl/gla.cpp +105 -0
package/src/llama.cpp/ggml/src/ggml-sycl/gla.hpp +8 -0
package/src/llama.cpp/ggml/src/ggml-sycl/outprod.cpp +3 -3
package/src/llama.cpp/ggml/src/ggml-sycl/outprod.hpp +1 -2
package/src/llama.cpp/ggml/src/ggml-sycl/tsembd.cpp +3 -2
package/src/llama.cpp/ggml/src/ggml-sycl/tsembd.hpp +1 -2
package/src/llama.cpp/ggml/src/ggml-sycl/wkv6.cpp +7 -5
package/src/llama.cpp/ggml/src/ggml-sycl/wkv6.hpp +1 -2
package/src/llama.cpp/ggml/src/ggml-vulkan/CMakeLists.txt +74 -4
package/src/llama.cpp/ggml/src/ggml-vulkan/ggml-vulkan.cpp +314 -116
package/src/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/CMakeLists.txt +4 -2
package/src/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp +9 -3
package/src/llama.cpp/ggml/src/ggml.c +117 -1327
package/src/llama.cpp/ggml/src/gguf.cpp +1329 -0
package/src/llama.cpp/include/llama-cpp.h +6 -1
package/src/llama.cpp/include/llama.h +138 -75
package/src/llama.cpp/src/CMakeLists.txt +13 -1
package/src/llama.cpp/src/llama-adapter.cpp +347 -0
package/src/llama.cpp/src/llama-adapter.h +74 -0
package/src/llama.cpp/src/llama-arch.cpp +1487 -0
package/src/llama.cpp/src/llama-arch.h +400 -0
package/src/llama.cpp/src/llama-batch.cpp +368 -0
package/src/llama.cpp/src/llama-batch.h +88 -0
package/src/llama.cpp/src/llama-chat.cpp +578 -0
package/src/llama.cpp/src/llama-chat.h +52 -0
package/src/llama.cpp/src/llama-context.cpp +1775 -0
package/src/llama.cpp/src/llama-context.h +128 -0
package/src/llama.cpp/src/llama-cparams.cpp +1 -0
package/src/llama.cpp/src/llama-cparams.h +37 -0
package/src/llama.cpp/src/llama-grammar.cpp +5 -4
package/src/llama.cpp/src/llama-grammar.h +3 -1
package/src/llama.cpp/src/llama-hparams.cpp +71 -0
package/src/llama.cpp/src/llama-hparams.h +139 -0
package/src/llama.cpp/src/llama-impl.cpp +167 -0
package/src/llama.cpp/src/llama-impl.h +16 -136
package/src/llama.cpp/src/llama-kv-cache.cpp +718 -0
package/src/llama.cpp/src/llama-kv-cache.h +218 -0
package/src/llama.cpp/src/llama-mmap.cpp +589 -0
package/src/llama.cpp/src/llama-mmap.h +67 -0
package/src/llama.cpp/src/llama-model-loader.cpp +1124 -0
package/src/llama.cpp/src/llama-model-loader.h +167 -0
package/src/llama.cpp/src/llama-model.cpp +3953 -0
package/src/llama.cpp/src/llama-model.h +370 -0
package/src/llama.cpp/src/llama-quant.cpp +934 -0
package/src/llama.cpp/src/llama-quant.h +1 -0
package/src/llama.cpp/src/llama-sampling.cpp +147 -32
package/src/llama.cpp/src/llama-sampling.h +3 -19
package/src/llama.cpp/src/llama-vocab.cpp +1832 -575
package/src/llama.cpp/src/llama-vocab.h +97 -142
package/src/llama.cpp/src/llama.cpp +7160 -20314
package/src/llama.cpp/src/unicode.cpp +8 -3
package/src/llama.cpp/tests/CMakeLists.txt +2 -0
package/src/llama.cpp/tests/test-autorelease.cpp +3 -3
package/src/llama.cpp/tests/test-backend-ops.cpp +370 -59
package/src/llama.cpp/tests/test-chat-template.cpp +162 -125
package/src/llama.cpp/tests/test-gguf.cpp +222 -187
package/src/llama.cpp/tests/test-model-load-cancel.cpp +1 -1
package/src/llama.cpp/tests/test-sampling.cpp +0 -1
package/src/llama.cpp/tests/test-tokenizer-0.cpp +4 -4
package/src/llama.cpp/tests/test-tokenizer-1-bpe.cpp +9 -7
package/src/llama.cpp/tests/test-tokenizer-1-spm.cpp +8 -6

package/src/llama.cpp/tests/test-model-load-cancel.cpp CHANGED Viewed

@@ -21,7 +21,7 @@ int main(int argc, char *argv[] ) {
         (void) ctx;
         return progress > 0.50;
     };
-    auto * model = llama_load_model_from_file(model_path, params);
+    auto * model = llama_model_load_from_file(model_path, params);
     llama_backend_free();
     return model == nullptr ? EXIT_SUCCESS : EXIT_FAILURE;
 }

package/src/llama.cpp/tests/test-sampling.cpp CHANGED Viewed

@@ -144,7 +144,6 @@ static void test_penalties(
     sampler_tester tester(probs, probs_expected);
-    const size_t n_vocab = probs.size();
     auto * sampler = llama_sampler_init_penalties(last_tokens.size(), repeat_penalty, alpha_frequency, alpha_presence);
     for (size_t i = 0; i < last_tokens.size(); i++) {

package/src/llama.cpp/tests/test-tokenizer-0.cpp CHANGED Viewed

@@ -152,7 +152,7 @@ int main(int argc, char **argv) {
         mparams.vocab_only = true;
-        model = llama_load_model_from_file(fname.c_str(), mparams);
+        model = llama_model_load_from_file(fname.c_str(), mparams);
         if (model == NULL) {
             fprintf(stderr, "%s: error: failed to load vocab '%s'\n", __func__, fname.c_str());
@@ -161,11 +161,11 @@ int main(int argc, char **argv) {
         auto cparams = llama_context_default_params();
-        ctx = llama_new_context_with_model(model, cparams);
+        ctx = llama_init_from_model(model, cparams);
         if (ctx == NULL) {
             fprintf(stderr, "%s: error: failed to load vocab '%s'\n", __func__, fname.c_str());
-            llama_free_model(model);
+            llama_model_free(model);
             return 1;
         }
     }
@@ -300,7 +300,7 @@ int main(int argc, char **argv) {
         fprintf(stderr, "%s : tokens written to '%s'\n", __func__, (fname_text + ".tokcpp").c_str());
     }
-    llama_free_model(model);
+    llama_model_free(model);
     llama_free(ctx);
     llama_backend_free();

package/src/llama.cpp/tests/test-tokenizer-1-bpe.cpp CHANGED Viewed

@@ -46,7 +46,7 @@ int main(int argc, char **argv) {
         mparams.vocab_only = true;
-        model = llama_load_model_from_file(fname.c_str(), mparams);
+        model = llama_model_load_from_file(fname.c_str(), mparams);
         if (model == NULL) {
             fprintf(stderr, "%s: error: failed to load vocab '%s'\n", __func__, fname.c_str());
@@ -55,17 +55,19 @@ int main(int argc, char **argv) {
         auto cparams = llama_context_default_params();
-        ctx = llama_new_context_with_model(model, cparams);
+        ctx = llama_init_from_model(model, cparams);
         if (ctx == NULL) {
             fprintf(stderr, "%s: error: failed to load vocab '%s'\n", __func__, fname.c_str());
-            llama_free_model(model);
+            llama_model_free(model);
             return 1;
         }
     }
-    //GGML_ASSERT(llama_vocab_type(model) == LLAMA_VOCAB_TYPE_BPE);
-    if (llama_vocab_type(model) != LLAMA_VOCAB_TYPE_BPE) {
+    const llama_vocab * vocab = llama_model_get_vocab(model);
+    //GGML_ASSERT(llama_vocab_type(vocab) == LLAMA_VOCAB_TYPE_BPE);
+    if (llama_vocab_type(vocab) != LLAMA_VOCAB_TYPE_BPE) {
         return 99;
     }
@@ -75,7 +77,7 @@ int main(int argc, char **argv) {
     atexit([]() { console::cleanup(); });
 #endif
-    const int n_vocab = llama_n_vocab(model);
+    const int n_vocab = llama_vocab_n_tokens(vocab);
     for (int i = 0; i < n_vocab; ++i) {
         std::string str = common_detokenize(ctx, std::vector<int>(1, i));
@@ -143,7 +145,7 @@ int main(int argc, char **argv) {
         }
     }
-    llama_free_model(model);
+    llama_model_free(model);
     llama_free(ctx);
     llama_backend_free();

package/src/llama.cpp/tests/test-tokenizer-1-spm.cpp CHANGED Viewed

@@ -34,7 +34,7 @@ int main(int argc, char ** argv) {
         mparams.vocab_only = true;
-        model = llama_load_model_from_file(fname.c_str(), mparams);
+        model = llama_model_load_from_file(fname.c_str(), mparams);
         if (model == NULL) {
             fprintf(stderr, "%s: error: failed to load vocab '%s'\n", __func__, fname.c_str());
@@ -43,17 +43,19 @@ int main(int argc, char ** argv) {
         auto cparams = llama_context_default_params();
-        ctx = llama_new_context_with_model(model, cparams);
+        ctx = llama_init_from_model(model, cparams);
         if (ctx == NULL) {
             fprintf(stderr, "%s: error: failed to load vocab '%s'\n", __func__, fname.c_str());
-            llama_free_model(model);
+            llama_model_free(model);
             return 1;
         }
     }
+    const llama_vocab * vocab = llama_model_get_vocab(model);
     //GGML_ASSERT(llama_vocab_type(model) == LLAMA_VOCAB_TYPE_SPM);
-    if (llama_vocab_type(model) != LLAMA_VOCAB_TYPE_SPM) {
+    if (llama_vocab_type(vocab) != LLAMA_VOCAB_TYPE_SPM) {
         return 99;
     }
@@ -63,7 +65,7 @@ int main(int argc, char ** argv) {
     atexit([]() { console::cleanup(); });
 #endif
-    const int n_vocab = llama_n_vocab(model);
+    const int n_vocab = llama_vocab_n_tokens(vocab);
     for (int i = 0; i < n_vocab; ++i) {
         std::string str = common_detokenize(ctx, std::vector<int>(1, i), true);
@@ -113,7 +115,7 @@ int main(int argc, char ** argv) {
         }
     }
-    llama_free_model(model);
+    llama_model_free(model);
     llama_free(ctx);
     llama_backend_free();