npm - @fugood/llama.node - Versions diffs - 0.3.11 → 0.3.13 - Mend

@fugood/llama.node 0.3.11 → 0.3.13

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 (77) hide show

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 +1 -0
package/lib/index.js +26 -20
package/lib/index.ts +32 -28
package/package.json +1 -1
package/src/LlamaCompletionWorker.cpp +14 -0
package/src/LlamaContext.cpp +13 -4
package/src/llama.cpp/.github/workflows/build.yml +35 -3
package/src/llama.cpp/.github/workflows/docker.yml +2 -0
package/src/llama.cpp/.github/workflows/labeler.yml +1 -1
package/src/llama.cpp/common/CMakeLists.txt +20 -3
package/src/llama.cpp/common/arg.cpp +180 -3
package/src/llama.cpp/common/chat-template.hpp +21 -7
package/src/llama.cpp/common/chat.cpp +220 -101
package/src/llama.cpp/common/chat.hpp +3 -0
package/src/llama.cpp/common/common.h +15 -7
package/src/llama.cpp/common/llguidance.cpp +3 -3
package/src/llama.cpp/common/log.cpp +1 -0
package/src/llama.cpp/common/log.h +2 -1
package/src/llama.cpp/common/minja.hpp +24 -9
package/src/llama.cpp/common/sampling.cpp +52 -46
package/src/llama.cpp/common/speculative.h +1 -1
package/src/llama.cpp/docs/build.md +2 -2
package/src/llama.cpp/examples/imatrix/imatrix.cpp +2 -1
package/src/llama.cpp/examples/llama-bench/llama-bench.cpp +6 -5
package/src/llama.cpp/examples/llama.android/llama/src/main/cpp/CMakeLists.txt +1 -1
package/src/llama.cpp/examples/perplexity/perplexity.cpp +1 -0
package/src/llama.cpp/examples/run/run.cpp +5 -12
package/src/llama.cpp/examples/server/CMakeLists.txt +1 -1
package/src/llama.cpp/examples/server/httplib.h +381 -292
package/src/llama.cpp/examples/server/server.cpp +58 -47
package/src/llama.cpp/examples/server/utils.hpp +7 -5
package/src/llama.cpp/ggml/include/ggml-cpu.h +1 -1
package/src/llama.cpp/ggml/include/ggml-metal.h +1 -1
package/src/llama.cpp/ggml/include/ggml-vulkan.h +0 -2
package/src/llama.cpp/ggml/include/ggml.h +1 -1
package/src/llama.cpp/ggml/src/ggml-common.h +0 -2
package/src/llama.cpp/ggml/src/ggml-cpu/ggml-cpu-impl.h +6 -12
package/src/llama.cpp/ggml/src/ggml-cpu/ggml-cpu-quants.c +852 -268
package/src/llama.cpp/ggml/src/ggml-cpu/ggml-cpu.c +200 -107
package/src/llama.cpp/ggml/src/ggml-cpu/ggml-cpu.cpp +2 -5
package/src/llama.cpp/ggml/src/ggml-cpu/llamafile/sgemm.cpp +9 -8
package/src/llama.cpp/ggml/src/ggml-cuda/CMakeLists.txt +2 -2
package/src/llama.cpp/ggml/src/ggml-opencl/ggml-opencl.cpp +26 -4
package/src/llama.cpp/ggml/src/ggml-sycl/ggml-sycl.cpp +6 -7
package/src/llama.cpp/ggml/src/ggml-vulkan/ggml-vulkan.cpp +812 -569
package/src/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp +25 -1
package/src/llama.cpp/ggml/src/ggml.c +1 -1
package/src/llama.cpp/include/llama.h +14 -10
package/src/llama.cpp/src/llama-grammar.cpp +1 -1
package/src/llama.cpp/src/llama-grammar.h +1 -1
package/src/llama.cpp/src/llama-impl.h +6 -6
package/src/llama.cpp/src/llama-kv-cache.h +1 -1
package/src/llama.cpp/src/llama-mmap.h +1 -0
package/src/llama.cpp/src/llama-model.cpp +1 -1
package/src/llama.cpp/src/llama-sampling.cpp +131 -57
package/src/llama.cpp/src/llama.cpp +7 -5
package/src/llama.cpp/src/unicode.cpp +9 -2
package/src/llama.cpp/tests/test-backend-ops.cpp +5 -5
package/src/llama.cpp/tests/test-chat.cpp +237 -69
package/src/llama.cpp/tests/test-gguf.cpp +4 -4
package/src/llama.cpp/tests/test-sampling.cpp +15 -0

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

@@ -181,6 +181,17 @@ static void test_dry(
     tester.check();
 }
+static void test_top_n_sigma(const std::vector<float> & probs, const std::vector<float> & probs_expected, int n) {
+    sampler_tester tester(probs, probs_expected);
+    DUMP(&tester.cur_p);
+    tester.apply(llama_sampler_init_top_n_sigma(n));
+    tester.apply(llama_sampler_init_dist (0));
+    DUMP(&tester.cur_p);
+    tester.check();
+}
 static void test_sampler_queue(const size_t n_vocab, const std::string & samplers_sequence, const int top_k, const float top_p, const float min_p
 ) {
     sampler_tester tester(n_vocab);
@@ -348,6 +359,10 @@ int main(void) {
     test_dry({0.2f, 0.2f, 0.2f, 0.2f, 0.2f}, {0, 1, 2, 0, 1}, {0.241818f, 0.241818f, 0.241818f, 0.241818f, 0.032727f}, 2.0f, 1.1f, 2, 5, {});
     test_dry({0.2f, 0.2f, 0.2f, 0.2f, 0.2f}, {0, 1, 2, 3, 4, 0, 1}, {0.2f, 0.2f, 0.2f, 0.2f, 0.2f}, 1.0f, 1.1f, 4, 7, {});
+    test_top_n_sigma({0.1f, 0.2f, 0.3f, 0.4f}, {0.571429f, 0.428571f, 0.0f, 0.0f}, 1.00f);
+    test_top_n_sigma({0.1f, 0.2f, 0.3f, 0.4f}, {1.0f, 0.0f, 0.0f, 0.0f}, 0.00f);
+    test_top_n_sigma({0.1f, 0.2f, 0.3f, 0.4f}, {0.4f, 0.3f, 0.2f, 0.1f}, 3.00f);
     test_sampler_queue(10000, "k", 10000, 1.0f, 1.0f);
     test_sampler_queue(10000, "k",     1, 1.0f, 1.0f);
     test_sampler_queue(10000, "p", 10000, 1.0f, 1.0f);