@novastera-oss/llamarn 0.2.7 → 0.3.0
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/cpp/include/llama.h +8 -3
- package/android/src/main/jniLibs/arm64-v8a/libggml-base.so +0 -0
- package/android/src/main/jniLibs/arm64-v8a/libggml-cpu.so +0 -0
- package/android/src/main/jniLibs/arm64-v8a/libggml.so +0 -0
- package/android/src/main/jniLibs/arm64-v8a/libllama.so +0 -0
- package/android/src/main/jniLibs/armeabi-v7a/libggml-base.so +0 -0
- package/android/src/main/jniLibs/armeabi-v7a/libggml-cpu.so +0 -0
- package/android/src/main/jniLibs/armeabi-v7a/libggml.so +0 -0
- package/android/src/main/jniLibs/armeabi-v7a/libllama.so +0 -0
- package/android/src/main/jniLibs/x86/libggml-base.so +0 -0
- package/android/src/main/jniLibs/x86/libggml-cpu.so +0 -0
- package/android/src/main/jniLibs/x86/libggml.so +0 -0
- package/android/src/main/jniLibs/x86/libllama.so +0 -0
- package/android/src/main/jniLibs/x86_64/libggml-base.so +0 -0
- package/android/src/main/jniLibs/x86_64/libggml-cpu.so +0 -0
- package/android/src/main/jniLibs/x86_64/libggml.so +0 -0
- package/android/src/main/jniLibs/x86_64/libllama.so +0 -0
- package/cpp/LlamaCppModel.cpp +56 -22
- package/cpp/build-info.cpp +2 -2
- package/cpp/llama.cpp/CMakeLists.txt +1 -2
- package/cpp/llama.cpp/README.md +4 -5
- package/cpp/llama.cpp/build-xcframework.sh +1 -1
- package/cpp/llama.cpp/common/CMakeLists.txt +4 -5
- package/cpp/llama.cpp/common/arg.cpp +24 -0
- package/cpp/llama.cpp/common/chat.cpp +37 -20
- package/cpp/llama.cpp/common/chat.h +2 -0
- package/cpp/llama.cpp/common/common.cpp +3 -0
- package/cpp/llama.cpp/common/common.h +5 -0
- package/cpp/llama.cpp/common/json-schema-to-grammar.cpp +3 -46
- package/cpp/llama.cpp/convert_hf_to_gguf.py +860 -23
- package/cpp/llama.cpp/convert_hf_to_gguf_update.py +9 -0
- package/cpp/llama.cpp/ggml/CMakeLists.txt +8 -2
- package/cpp/llama.cpp/ggml/include/ggml-backend.h +1 -1
- package/cpp/llama.cpp/ggml/include/ggml-cpu.h +2 -0
- package/cpp/llama.cpp/ggml/include/ggml.h +206 -10
- package/cpp/llama.cpp/ggml/src/CMakeLists.txt +17 -1
- package/cpp/llama.cpp/ggml/src/ggml-backend-reg.cpp +0 -8
- package/cpp/llama.cpp/ggml/src/ggml-backend.cpp +36 -18
- package/cpp/llama.cpp/ggml/src/ggml-cann/aclnn_ops.cpp +68 -5
- package/cpp/llama.cpp/ggml/src/ggml-cann/common.h +1 -1
- package/cpp/llama.cpp/ggml/src/ggml-cann/ggml-cann.cpp +16 -2
- package/cpp/llama.cpp/ggml/src/ggml-cpu/CMakeLists.txt +37 -3
- package/cpp/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp +10 -9
- package/cpp/llama.cpp/ggml/src/ggml-cpu/arch/arm/quants.c +109 -108
- package/cpp/llama.cpp/ggml/src/ggml-cpu/arch/arm/repack.cpp +1027 -1038
- package/cpp/llama.cpp/ggml/src/ggml-cpu/arch/loongarch/quants.c +53 -52
- package/cpp/llama.cpp/ggml/src/ggml-cpu/arch/powerpc/cpu-feats.cpp +82 -0
- package/cpp/llama.cpp/ggml/src/ggml-cpu/arch/powerpc/quants.c +56 -55
- package/cpp/llama.cpp/ggml/src/ggml-cpu/arch/riscv/quants.c +42 -41
- package/cpp/llama.cpp/ggml/src/ggml-cpu/arch/riscv/repack.cpp +24 -23
- package/cpp/llama.cpp/ggml/src/ggml-cpu/arch/s390/quants.c +29 -28
- package/cpp/llama.cpp/ggml/src/ggml-cpu/arch/wasm/quants.c +30 -29
- package/cpp/llama.cpp/ggml/src/ggml-cpu/arch/x86/quants.c +83 -82
- package/cpp/llama.cpp/ggml/src/ggml-cpu/arch/x86/repack.cpp +20 -19
- package/cpp/llama.cpp/ggml/src/ggml-cpu/common.h +3 -2
- package/cpp/llama.cpp/ggml/src/ggml-cpu/ggml-cpu-impl.h +9 -3
- package/cpp/llama.cpp/ggml/src/ggml-cpu/ggml-cpu.c +111 -103
- package/cpp/llama.cpp/ggml/src/ggml-cpu/ggml-cpu.cpp +4 -0
- package/cpp/llama.cpp/ggml/src/ggml-cpu/llamafile/sgemm.cpp +3 -2
- package/cpp/llama.cpp/ggml/src/ggml-cpu/ops.cpp +1405 -240
- package/cpp/llama.cpp/ggml/src/ggml-cpu/ops.h +8 -0
- package/cpp/llama.cpp/ggml/src/ggml-cpu/quants.c +25 -24
- package/cpp/llama.cpp/ggml/src/ggml-cpu/repack.cpp +56 -40
- package/cpp/llama.cpp/ggml/src/ggml-cpu/simd-mappings.h +212 -34
- package/cpp/llama.cpp/ggml/src/ggml-cpu/vec.cpp +35 -11
- package/cpp/llama.cpp/ggml/src/ggml-cpu/vec.h +187 -54
- package/cpp/llama.cpp/ggml/src/ggml-cuda/common.cuh +71 -29
- package/cpp/llama.cpp/ggml/src/ggml-cuda/conv2d-dw.cu +161 -0
- package/cpp/llama.cpp/ggml/src/ggml-cuda/conv2d-dw.cuh +5 -0
- package/cpp/llama.cpp/ggml/src/ggml-cuda/conv2d-transpose.cu +91 -0
- package/cpp/llama.cpp/ggml/src/ggml-cuda/conv2d-transpose.cuh +4 -0
- package/cpp/llama.cpp/ggml/src/ggml-cuda/convert.cu +22 -0
- package/cpp/llama.cpp/ggml/src/ggml-cuda/convert.cuh +5 -0
- package/cpp/llama.cpp/ggml/src/ggml-cuda/cross-entropy-loss.cu +2 -14
- package/cpp/llama.cpp/ggml/src/ggml-cuda/fattn-common.cuh +4 -1
- package/cpp/llama.cpp/ggml/src/ggml-cuda/fattn-mma-f16.cuh +8 -4
- package/cpp/llama.cpp/ggml/src/ggml-cuda/fattn-tile-f16.cu +6 -4
- package/cpp/llama.cpp/ggml/src/ggml-cuda/fattn-tile-f32.cu +14 -12
- package/cpp/llama.cpp/ggml/src/ggml-cuda/fattn-vec-f16.cuh +5 -3
- package/cpp/llama.cpp/ggml/src/ggml-cuda/fattn-vec-f32.cuh +15 -10
- package/cpp/llama.cpp/ggml/src/ggml-cuda/fattn-wmma-f16.cu +12 -6
- package/cpp/llama.cpp/ggml/src/ggml-cuda/getrows.cu +8 -0
- package/cpp/llama.cpp/ggml/src/ggml-cuda/ggml-cuda.cu +269 -110
- package/cpp/llama.cpp/ggml/src/ggml-cuda/mean.cu +19 -0
- package/cpp/llama.cpp/ggml/src/ggml-cuda/mean.cuh +3 -0
- package/cpp/llama.cpp/ggml/src/ggml-cuda/mmq.cuh +2 -8
- package/cpp/llama.cpp/ggml/src/ggml-cuda/mmv.cu +257 -87
- package/cpp/llama.cpp/ggml/src/ggml-cuda/mmv.cuh +2 -3
- package/cpp/llama.cpp/ggml/src/ggml-cuda/rope.cu +21 -27
- package/cpp/llama.cpp/ggml/src/ggml-cuda/scale.cu +8 -6
- package/cpp/llama.cpp/ggml/src/ggml-cuda/softmax.cu +119 -58
- package/cpp/llama.cpp/ggml/src/ggml-cuda/ssm-conv.cu +10 -2
- package/cpp/llama.cpp/ggml/src/ggml-cuda/ssm-scan.cu +192 -52
- package/cpp/llama.cpp/ggml/src/ggml-cuda/sumrows.cu +5 -18
- package/cpp/llama.cpp/ggml/src/ggml-cuda/sumrows.cuh +0 -1
- package/cpp/llama.cpp/ggml/src/ggml-cuda/unary.cu +97 -0
- package/cpp/llama.cpp/ggml/src/ggml-cuda/unary.cuh +11 -0
- package/cpp/llama.cpp/ggml/src/ggml-cuda/upscale.cu +92 -6
- package/cpp/llama.cpp/ggml/src/ggml-cuda/vendors/hip.h +14 -5
- package/cpp/llama.cpp/ggml/src/ggml-impl.h +125 -183
- package/cpp/llama.cpp/ggml/src/ggml-metal/CMakeLists.txt +4 -2
- package/cpp/llama.cpp/ggml/src/ggml-metal/ggml-metal-impl.h +51 -9
- package/cpp/llama.cpp/ggml/src/ggml-metal/ggml-metal.m +394 -80
- package/cpp/llama.cpp/ggml/src/ggml-metal/ggml-metal.metal +616 -239
- package/cpp/llama.cpp/ggml/src/ggml-musa/mudnn.cuh +2 -2
- package/cpp/llama.cpp/ggml/src/ggml-opencl/CMakeLists.txt +3 -0
- package/cpp/llama.cpp/ggml/src/ggml-opencl/ggml-opencl.cpp +741 -571
- package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/gelu.cl +27 -0
- package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/glu.cl +337 -0
- package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/mul_mat_f16_f32.cl +130 -0
- package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/scale.cl +3 -2
- package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/set_rows.cl +95 -0
- package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/softmax_4_f16.cl +24 -11
- package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/softmax_4_f32.cl +24 -11
- package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/softmax_f16.cl +24 -11
- package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/softmax_f32.cl +24 -11
- package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/upscale.cl +2 -3
- package/cpp/llama.cpp/ggml/src/ggml-quants.c +6 -6
- package/cpp/llama.cpp/ggml/src/ggml-sycl/backend.hpp +1 -0
- package/cpp/llama.cpp/ggml/src/ggml-sycl/binbcast.cpp +5 -6
- package/cpp/llama.cpp/ggml/src/ggml-sycl/common.hpp +1 -24
- package/cpp/llama.cpp/ggml/src/ggml-sycl/concat.cpp +28 -41
- package/cpp/llama.cpp/ggml/src/ggml-sycl/conv.cpp +4 -10
- package/cpp/llama.cpp/ggml/src/ggml-sycl/convert.cpp +99 -166
- package/cpp/llama.cpp/ggml/src/ggml-sycl/cpy.cpp +94 -72
- package/cpp/llama.cpp/ggml/src/ggml-sycl/dmmv.cpp +49 -67
- package/cpp/llama.cpp/ggml/src/ggml-sycl/dpct/helper.hpp +31 -1
- package/cpp/llama.cpp/ggml/src/ggml-sycl/element_wise.cpp +697 -1098
- package/cpp/llama.cpp/ggml/src/ggml-sycl/element_wise.hpp +18 -9
- package/cpp/llama.cpp/ggml/src/ggml-sycl/getrows.cpp +6 -9
- package/cpp/llama.cpp/ggml/src/ggml-sycl/ggml-sycl.cpp +104 -62
- package/cpp/llama.cpp/ggml/src/ggml-sycl/gla.cpp +2 -2
- package/cpp/llama.cpp/ggml/src/ggml-sycl/im2col.cpp +1 -1
- package/cpp/llama.cpp/ggml/src/ggml-sycl/mmq.cpp +60 -80
- package/cpp/llama.cpp/ggml/src/ggml-sycl/mmvq.cpp +132 -201
- package/cpp/llama.cpp/ggml/src/ggml-sycl/norm.cpp +55 -74
- package/cpp/llama.cpp/ggml/src/ggml-sycl/rope.cpp +39 -38
- package/cpp/llama.cpp/ggml/src/ggml-sycl/set_rows.cpp +131 -0
- package/cpp/llama.cpp/ggml/src/ggml-sycl/set_rows.hpp +8 -0
- package/cpp/llama.cpp/ggml/src/ggml-sycl/softmax.cpp +3 -3
- package/cpp/llama.cpp/ggml/src/ggml-sycl/sycl_hw.cpp +3 -1
- package/cpp/llama.cpp/ggml/src/ggml-sycl/sycl_hw.hpp +3 -0
- package/cpp/llama.cpp/ggml/src/ggml-sycl/tsembd.cpp +3 -8
- package/cpp/llama.cpp/ggml/src/ggml-sycl/wkv.cpp +12 -16
- package/cpp/llama.cpp/ggml/src/ggml-vulkan/CMakeLists.txt +12 -1
- package/cpp/llama.cpp/ggml/src/ggml-vulkan/ggml-vulkan.cpp +767 -292
- package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/CMakeLists.txt +4 -0
- package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/copy_to_quant.comp +58 -7
- package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn.comp +28 -23
- package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_base.comp +14 -9
- package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_cm1.comp +38 -32
- package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_cm2.comp +32 -27
- package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_split_k_reduce.comp +44 -12
- package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/geglu.comp +13 -0
- package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/geglu_erf.comp +27 -0
- package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/geglu_quick.comp +11 -0
- package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/gelu_erf.comp +39 -0
- package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/glu_head.comp +15 -0
- package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/glu_main.comp +29 -0
- package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mm.comp +128 -72
- package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mm_cm2.comp +38 -9
- package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/reglu.comp +9 -0
- package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/rms_norm.comp +12 -3
- package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/roll.comp +46 -0
- package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/rope_multi.comp +7 -9
- package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/rope_neox.comp +7 -9
- package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/rope_norm.comp +7 -9
- package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/scale.comp +1 -1
- package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/soft_max.comp +20 -4
- package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/swiglu.comp +9 -0
- package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/upscale.comp +69 -5
- package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp +23 -3
- package/cpp/llama.cpp/ggml/src/ggml.c +449 -72
- package/cpp/llama.cpp/ggml/src/gguf.cpp +13 -2
- package/cpp/llama.cpp/gguf-py/gguf/constants.py +285 -0
- package/cpp/llama.cpp/gguf-py/gguf/gguf_writer.py +27 -0
- package/cpp/llama.cpp/gguf-py/gguf/tensor_mapping.py +137 -21
- package/cpp/llama.cpp/gguf-py/gguf/vocab.py +109 -7
- package/cpp/llama.cpp/gguf-py/pyproject.toml +2 -2
- package/cpp/llama.cpp/include/llama.h +8 -43
- package/cpp/llama.cpp/models/templates/Mistral-Small-3.2-24B-Instruct-2506.jinja +124 -0
- package/cpp/llama.cpp/src/llama-arch.cpp +265 -3
- package/cpp/llama.cpp/src/llama-arch.h +36 -1
- package/cpp/llama.cpp/src/llama-batch.cpp +596 -359
- package/cpp/llama.cpp/src/llama-batch.h +105 -70
- package/cpp/llama.cpp/src/llama-chat.cpp +26 -6
- package/cpp/llama.cpp/src/llama-chat.h +1 -0
- package/cpp/llama.cpp/src/llama-context.cpp +101 -107
- package/cpp/llama.cpp/src/llama-context.h +13 -13
- package/cpp/llama.cpp/src/llama-graph.cpp +286 -404
- package/cpp/llama.cpp/src/llama-graph.h +78 -79
- package/cpp/llama.cpp/src/llama-hparams.cpp +11 -1
- package/cpp/llama.cpp/src/llama-hparams.h +11 -0
- package/cpp/llama.cpp/src/llama-kv-cache-unified-iswa.cpp +74 -66
- package/cpp/llama.cpp/src/llama-kv-cache-unified-iswa.h +23 -26
- package/cpp/llama.cpp/src/llama-kv-cache-unified.cpp +312 -157
- package/cpp/llama.cpp/src/llama-kv-cache-unified.h +79 -46
- package/cpp/llama.cpp/src/llama-kv-cells.h +97 -21
- package/cpp/llama.cpp/src/llama-memory-hybrid.cpp +73 -69
- package/cpp/llama.cpp/src/llama-memory-hybrid.h +19 -22
- package/cpp/llama.cpp/src/llama-memory-recurrent.cpp +88 -77
- package/cpp/llama.cpp/src/llama-memory-recurrent.h +15 -20
- package/cpp/llama.cpp/src/llama-memory.cpp +17 -0
- package/cpp/llama.cpp/src/llama-memory.h +21 -22
- package/cpp/llama.cpp/src/llama-model-saver.cpp +1 -0
- package/cpp/llama.cpp/src/llama-model.cpp +5301 -2922
- package/cpp/llama.cpp/src/llama-model.h +40 -0
- package/cpp/llama.cpp/src/llama-quant.cpp +88 -5
- package/cpp/llama.cpp/src/llama-vocab.cpp +37 -3
- package/cpp/llama.cpp/src/llama-vocab.h +42 -0
- package/cpp/rn-utils.h +3 -0
- package/ios/include/chat.h +2 -0
- package/ios/include/common.h +5 -0
- package/ios/include/llama.h +8 -43
- package/ios/libs/llama.xcframework/Info.plist +19 -19
- package/ios/libs/llama.xcframework/ios-arm64/dSYMs/llama.dSYM/Contents/Resources/DWARF/llama +0 -0
- package/ios/libs/llama.xcframework/ios-arm64/dSYMs/llama.dSYM/Contents/Resources/Relocations/aarch64/llama.yml +5059 -4863
- package/ios/libs/llama.xcframework/ios-arm64/llama.framework/Headers/ggml-backend.h +1 -1
- package/ios/libs/llama.xcframework/ios-arm64/llama.framework/Headers/ggml-cpu.h +2 -0
- package/ios/libs/llama.xcframework/ios-arm64/llama.framework/Headers/ggml.h +206 -10
- package/ios/libs/llama.xcframework/ios-arm64/llama.framework/Headers/llama.h +8 -43
- package/ios/libs/llama.xcframework/ios-arm64/llama.framework/llama +0 -0
- package/ios/libs/llama.xcframework/ios-arm64_x86_64-simulator/dSYMs/llama.dSYM/Contents/Resources/DWARF/llama +0 -0
- package/ios/libs/llama.xcframework/ios-arm64_x86_64-simulator/dSYMs/llama.dSYM/Contents/Resources/Relocations/aarch64/llama.yml +5030 -4834
- package/ios/libs/llama.xcframework/ios-arm64_x86_64-simulator/dSYMs/llama.dSYM/Contents/Resources/Relocations/x86_64/llama.yml +3889 -3742
- package/ios/libs/llama.xcframework/ios-arm64_x86_64-simulator/llama.framework/Headers/ggml-backend.h +1 -1
- package/ios/libs/llama.xcframework/ios-arm64_x86_64-simulator/llama.framework/Headers/ggml-cpu.h +2 -0
- package/ios/libs/llama.xcframework/ios-arm64_x86_64-simulator/llama.framework/Headers/ggml.h +206 -10
- package/ios/libs/llama.xcframework/ios-arm64_x86_64-simulator/llama.framework/Headers/llama.h +8 -43
- package/ios/libs/llama.xcframework/ios-arm64_x86_64-simulator/llama.framework/llama +0 -0
- package/ios/libs/llama.xcframework/macos-arm64_x86_64/dSYMs/llama.dSYM/Contents/Resources/DWARF/llama +0 -0
- package/ios/libs/llama.xcframework/macos-arm64_x86_64/dSYMs/llama.dSYM/Contents/Resources/Relocations/aarch64/llama.yml +5030 -4834
- package/ios/libs/llama.xcframework/macos-arm64_x86_64/dSYMs/llama.dSYM/Contents/Resources/Relocations/x86_64/llama.yml +3891 -3744
- package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Headers/ggml-backend.h +1 -1
- package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Headers/ggml-cpu.h +2 -0
- package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Headers/ggml.h +206 -10
- package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Headers/llama.h +8 -43
- package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/A/Headers/ggml-backend.h +1 -1
- package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/A/Headers/ggml-cpu.h +2 -0
- package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/A/Headers/ggml.h +206 -10
- package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/A/Headers/llama.h +8 -43
- package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/A/llama +0 -0
- package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/Current/Headers/ggml-backend.h +1 -1
- package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/Current/Headers/ggml-cpu.h +2 -0
- package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/Current/Headers/ggml.h +206 -10
- package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/Current/Headers/llama.h +8 -43
- package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/Current/llama +0 -0
- package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/llama +0 -0
- package/ios/libs/llama.xcframework/tvos-arm64/dSYMs/llama.dSYM/Contents/Resources/DWARF/llama +0 -0
- package/ios/libs/llama.xcframework/tvos-arm64/dSYMs/llama.dSYM/Contents/Resources/Relocations/aarch64/llama.yml +5059 -4863
- package/ios/libs/llama.xcframework/tvos-arm64/llama.framework/Headers/ggml-backend.h +1 -1
- package/ios/libs/llama.xcframework/tvos-arm64/llama.framework/Headers/ggml-cpu.h +2 -0
- package/ios/libs/llama.xcframework/tvos-arm64/llama.framework/Headers/ggml.h +206 -10
- package/ios/libs/llama.xcframework/tvos-arm64/llama.framework/Headers/llama.h +8 -43
- package/ios/libs/llama.xcframework/tvos-arm64/llama.framework/llama +0 -0
- package/ios/libs/llama.xcframework/tvos-arm64_x86_64-simulator/dSYMs/llama.dSYM/Contents/Resources/DWARF/llama +0 -0
- package/ios/libs/llama.xcframework/tvos-arm64_x86_64-simulator/dSYMs/llama.dSYM/Contents/Resources/Relocations/aarch64/llama.yml +5030 -4834
- package/ios/libs/llama.xcframework/tvos-arm64_x86_64-simulator/dSYMs/llama.dSYM/Contents/Resources/Relocations/x86_64/llama.yml +3889 -3742
- package/ios/libs/llama.xcframework/tvos-arm64_x86_64-simulator/llama.framework/Headers/ggml-backend.h +1 -1
- package/ios/libs/llama.xcframework/tvos-arm64_x86_64-simulator/llama.framework/Headers/ggml-cpu.h +2 -0
- package/ios/libs/llama.xcframework/tvos-arm64_x86_64-simulator/llama.framework/Headers/ggml.h +206 -10
- package/ios/libs/llama.xcframework/tvos-arm64_x86_64-simulator/llama.framework/Headers/llama.h +8 -43
- package/ios/libs/llama.xcframework/tvos-arm64_x86_64-simulator/llama.framework/llama +0 -0
- package/ios/libs/llama.xcframework/xros-arm64/dSYMs/llama.dSYM/Contents/Resources/DWARF/llama +0 -0
- package/ios/libs/llama.xcframework/xros-arm64/dSYMs/llama.dSYM/Contents/Resources/Relocations/aarch64/llama.yml +5095 -4900
- package/ios/libs/llama.xcframework/xros-arm64/llama.framework/Headers/ggml-backend.h +1 -1
- package/ios/libs/llama.xcframework/xros-arm64/llama.framework/Headers/ggml-cpu.h +2 -0
- package/ios/libs/llama.xcframework/xros-arm64/llama.framework/Headers/ggml.h +206 -10
- package/ios/libs/llama.xcframework/xros-arm64/llama.framework/Headers/llama.h +8 -43
- package/ios/libs/llama.xcframework/xros-arm64/llama.framework/llama +0 -0
- package/ios/libs/llama.xcframework/xros-arm64_x86_64-simulator/dSYMs/llama.dSYM/Contents/Resources/DWARF/llama +0 -0
- package/ios/libs/llama.xcframework/xros-arm64_x86_64-simulator/dSYMs/llama.dSYM/Contents/Resources/Relocations/aarch64/llama.yml +5066 -4871
- package/ios/libs/llama.xcframework/xros-arm64_x86_64-simulator/dSYMs/llama.dSYM/Contents/Resources/Relocations/x86_64/llama.yml +3919 -3773
- package/ios/libs/llama.xcframework/xros-arm64_x86_64-simulator/llama.framework/Headers/ggml-backend.h +1 -1
- package/ios/libs/llama.xcframework/xros-arm64_x86_64-simulator/llama.framework/Headers/ggml-cpu.h +2 -0
- package/ios/libs/llama.xcframework/xros-arm64_x86_64-simulator/llama.framework/Headers/ggml.h +206 -10
- package/ios/libs/llama.xcframework/xros-arm64_x86_64-simulator/llama.framework/Headers/llama.h +8 -43
- package/ios/libs/llama.xcframework/xros-arm64_x86_64-simulator/llama.framework/llama +0 -0
- package/package.json +1 -1
- package/cpp/llama.cpp/ggml/include/ggml-kompute.h +0 -50
- package/cpp/llama.cpp/ggml/src/ggml-kompute/CMakeLists.txt +0 -166
- package/cpp/llama.cpp/ggml/src/ggml-kompute/ggml-kompute.cpp +0 -2251
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/common.comp +0 -112
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_add.comp +0 -58
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_addrow.comp +0 -25
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_cpy_f16_f16.comp +0 -52
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_cpy_f16_f32.comp +0 -52
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_cpy_f32_f16.comp +0 -52
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_cpy_f32_f32.comp +0 -52
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_diagmask.comp +0 -30
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_gelu.comp +0 -22
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_getrows.comp +0 -17
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_getrows_f16.comp +0 -31
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_getrows_f32.comp +0 -31
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_getrows_q4_0.comp +0 -38
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_getrows_q4_1.comp +0 -39
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_getrows_q6_k.comp +0 -44
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_mul.comp +0 -52
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_f16.comp +0 -69
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_mat_f32.comp +0 -51
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_q4_0.comp +0 -33
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_q4_1.comp +0 -35
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_q4_k.comp +0 -140
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_q6_k.comp +0 -106
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_q8_0.comp +0 -73
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_mul_mv_q_n.comp +0 -52
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_mul_mv_q_n_pre.comp +0 -28
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_norm.comp +0 -84
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_relu.comp +0 -21
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_rmsnorm.comp +0 -53
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_rope_neox_f16.comp +0 -52
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_rope_neox_f32.comp +0 -52
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_rope_norm_f16.comp +0 -52
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_rope_norm_f32.comp +0 -52
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_scale.comp +0 -19
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_scale_8.comp +0 -23
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_silu.comp +0 -22
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/op_softmax.comp +0 -72
- package/cpp/llama.cpp/ggml/src/ggml-kompute/kompute-shaders/rope_common.comp +0 -71
|
@@ -229,7 +229,11 @@ typedef struct {
|
|
|
229
229
|
uint64_t nb21;
|
|
230
230
|
uint64_t nb22;
|
|
231
231
|
uint64_t nb23;
|
|
232
|
+
int32_t ne32;
|
|
233
|
+
int32_t ne33;
|
|
232
234
|
uint64_t nb31;
|
|
235
|
+
uint64_t nb32;
|
|
236
|
+
uint64_t nb33;
|
|
233
237
|
int32_t ne1;
|
|
234
238
|
int32_t ne2;
|
|
235
239
|
float scale;
|
|
@@ -422,6 +426,17 @@ typedef struct {
|
|
|
422
426
|
int32_t KHW; // KH * KW, pre-computed on CPU to save GPU resources
|
|
423
427
|
} ggml_metal_kargs_im2col;
|
|
424
428
|
|
|
429
|
+
typedef struct{
|
|
430
|
+
int32_t ne00;
|
|
431
|
+
uint64_t nb01;
|
|
432
|
+
int32_t ne10;
|
|
433
|
+
uint64_t nb11;
|
|
434
|
+
int32_t ne0;
|
|
435
|
+
uint64_t nb1;
|
|
436
|
+
int32_t i00;
|
|
437
|
+
int32_t i10;
|
|
438
|
+
} ggml_metal_kargs_glu;
|
|
439
|
+
|
|
425
440
|
typedef struct {
|
|
426
441
|
int64_t ne00;
|
|
427
442
|
int64_t ne01;
|
|
@@ -450,9 +465,21 @@ typedef struct {
|
|
|
450
465
|
} ggml_metal_kargs_sum_rows;
|
|
451
466
|
|
|
452
467
|
typedef struct {
|
|
453
|
-
|
|
454
|
-
|
|
455
|
-
|
|
468
|
+
int32_t ne00;
|
|
469
|
+
int32_t ne01;
|
|
470
|
+
int32_t ne02;
|
|
471
|
+
uint64_t nb01;
|
|
472
|
+
uint64_t nb02;
|
|
473
|
+
uint64_t nb03;
|
|
474
|
+
int32_t ne11;
|
|
475
|
+
int32_t ne12;
|
|
476
|
+
int32_t ne13;
|
|
477
|
+
uint64_t nb11;
|
|
478
|
+
uint64_t nb12;
|
|
479
|
+
uint64_t nb13;
|
|
480
|
+
uint64_t nb1;
|
|
481
|
+
uint64_t nb2;
|
|
482
|
+
uint64_t nb3;
|
|
456
483
|
float scale;
|
|
457
484
|
float max_bias;
|
|
458
485
|
float m0;
|
|
@@ -488,26 +515,25 @@ typedef struct {
|
|
|
488
515
|
typedef struct {
|
|
489
516
|
int64_t d_state;
|
|
490
517
|
int64_t d_inner;
|
|
518
|
+
int64_t n_head;
|
|
519
|
+
int64_t n_group;
|
|
491
520
|
int64_t n_seq_tokens;
|
|
492
521
|
int64_t n_seqs;
|
|
493
|
-
uint64_t nb00;
|
|
494
522
|
uint64_t nb01;
|
|
495
523
|
uint64_t nb02;
|
|
496
|
-
uint64_t
|
|
524
|
+
uint64_t nb03;
|
|
497
525
|
uint64_t nb11;
|
|
498
526
|
uint64_t nb12;
|
|
499
527
|
uint64_t nb13;
|
|
500
|
-
uint64_t nb20;
|
|
501
528
|
uint64_t nb21;
|
|
502
529
|
uint64_t nb22;
|
|
503
|
-
uint64_t nb30;
|
|
504
530
|
uint64_t nb31;
|
|
505
|
-
uint64_t nb40;
|
|
506
531
|
uint64_t nb41;
|
|
507
532
|
uint64_t nb42;
|
|
508
|
-
uint64_t
|
|
533
|
+
uint64_t nb43;
|
|
509
534
|
uint64_t nb51;
|
|
510
535
|
uint64_t nb52;
|
|
536
|
+
uint64_t nb53;
|
|
511
537
|
} ggml_metal_kargs_ssm_scan;
|
|
512
538
|
|
|
513
539
|
typedef struct {
|
|
@@ -521,6 +547,22 @@ typedef struct {
|
|
|
521
547
|
uint64_t nb2;
|
|
522
548
|
} ggml_metal_kargs_get_rows;
|
|
523
549
|
|
|
550
|
+
typedef struct {
|
|
551
|
+
int32_t nk0;
|
|
552
|
+
int32_t ne01;
|
|
553
|
+
uint64_t nb01;
|
|
554
|
+
uint64_t nb02;
|
|
555
|
+
uint64_t nb03;
|
|
556
|
+
int32_t ne11;
|
|
557
|
+
int32_t ne12;
|
|
558
|
+
uint64_t nb10;
|
|
559
|
+
uint64_t nb11;
|
|
560
|
+
uint64_t nb12;
|
|
561
|
+
uint64_t nb1;
|
|
562
|
+
uint64_t nb2;
|
|
563
|
+
uint64_t nb3;
|
|
564
|
+
} ggml_metal_kargs_set_rows;
|
|
565
|
+
|
|
524
566
|
typedef struct {
|
|
525
567
|
int64_t ne00;
|
|
526
568
|
int64_t ne01;
|