localcode 0.5.32__tar.gz → 0.5.34__tar.gz
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.
- {localcode-0.5.32/src/localcode.egg-info → localcode-0.5.34}/PKG-INFO +1 -1
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/jinja/lexer.cpp +20 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/jinja/value.cpp +6 -1
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-arch.cpp +1 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-arch.h +1 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-graph.cpp +48 -4
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-graph.h +10 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-kv-cache.cpp +4 -1
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-model.cpp +3 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-vocab.cpp +11 -1
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-vocab.h +1 -0
- localcode-0.5.34/llama-cpp-turboquant/src/models/flint.cpp +194 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/models.h +13 -0
- {localcode-0.5.32 → localcode-0.5.34}/pyproject.toml +1 -1
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/__init__.py +1 -1
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/bin/llama-server +0 -0
- localcode-0.5.34/src/localcode/bin/localcode-ui +4 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/models_catalog.py +68 -2
- localcode-0.5.34/src/localcode/ui/FORK_COMMIT +1 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/ui/plugin/server.ts +5 -4
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/ui/supervisor.py +32 -0
- {localcode-0.5.32 → localcode-0.5.34/src/localcode.egg-info}/PKG-INFO +1 -1
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode.egg-info/SOURCES.txt +1 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_comprehensive_models_matrix.py +1 -1
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_models_catalog.py +2 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_ui_server_lifecycle.py +42 -0
- localcode-0.5.32/src/localcode/bin/localcode-ui +0 -4
- localcode-0.5.32/src/localcode/ui/FORK_COMMIT +0 -1
- {localcode-0.5.32 → localcode-0.5.34}/LICENSE +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/MANIFEST.in +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/README.md +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/CMakePresets.json +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/LICENSE +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/cmake/arm64-apple-clang.cmake +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/cmake/arm64-linux-clang.cmake +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/cmake/arm64-windows-llvm.cmake +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/cmake/arm64-windows-msvc-cuda.cmake +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/cmake/build-info.cmake +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/cmake/common.cmake +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/cmake/download-models.cmake +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/cmake/git-vars.cmake +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/cmake/license.cmake +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/cmake/llama-config.cmake.in +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/cmake/llama.pc.in +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/cmake/riscv64-spacemit-linux-gnu-gcc.cmake +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/cmake/x64-windows-llvm.cmake +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/arg.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/arg.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/base64.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/build-info.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/chat-auto-parser-generator.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/chat-auto-parser-helpers.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/chat-auto-parser-helpers.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/chat-auto-parser.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/chat-diff-analyzer.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/chat-peg-parser.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/chat-peg-parser.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/chat.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/chat.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/common.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/common.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/console.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/console.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/debug.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/debug.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/download.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/download.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/fit.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/fit.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/hf-cache.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/hf-cache.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/http.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/imatrix-loader.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/imatrix-loader.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/jinja/caps.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/jinja/caps.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/jinja/lexer.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/jinja/parser.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/jinja/parser.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/jinja/runtime.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/jinja/runtime.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/jinja/string.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/jinja/string.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/jinja/utils.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/jinja/value.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/json-schema-to-grammar.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/json-schema-to-grammar.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/json-schema.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/json-schema.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/json.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/json.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/llguidance.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/log.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/log.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/ngram-cache.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/ngram-cache.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/ngram-map.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/ngram-map.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/ngram-mod.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/ngram-mod.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/parsers/cohere2moe.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/parsers/deepseek.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/parsers/functionary-v3-2.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/parsers/gemma4.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/parsers/gigachat-v3.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/parsers/gpt-oss.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/parsers/k2-horizon.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/parsers/kimi-k2.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/parsers/kimi-k3.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/parsers/lfm2.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/parsers/ling3.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/parsers/llm-jp-harmony.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/parsers/minicpm5.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/parsers/minimax-m3.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/parsers/ministral3.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/parsers/muse-glimmer.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/parsers/parsers.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/parsers/parsers.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/parsers/qwen3-coder.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/parsers/translate-gemma.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/peg-parser.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/peg-parser.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/preset.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/preset.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/reasoning-budget.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/reasoning-budget.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/sampling.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/sampling.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/speculative.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/speculative.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/subproc.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/subproc.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/trie.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/trie.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/unicode.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/common/unicode.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/include/ggml-alloc.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/include/ggml-backend.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/include/ggml-blas.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/include/ggml-cann.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/include/ggml-cpp.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/include/ggml-cpu.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/include/ggml-cuda.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/include/ggml-et.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/include/ggml-hexagon.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/include/ggml-metal.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/include/ggml-opencl.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/include/ggml-openvino.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/include/ggml-opt.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/include/ggml-rpc.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/include/ggml-sycl.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/include/ggml-virtgpu.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/include/ggml-vulkan.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/include/ggml-webgpu.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/include/ggml-zdnn.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/include/ggml-zendnn.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/include/ggml.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/include/gguf.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-alloc.c +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-backend-dl.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-backend-dl.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-backend-impl.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-backend-meta.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-backend-reg.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-backend.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-blas/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-blas/ggml-blas.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cann/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cann/acl_tensor.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cann/acl_tensor.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cann/aclnn_ops.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cann/aclnn_ops.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cann/common.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cann/ggml-cann.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-common.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/amx/amx.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/amx/amx.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/amx/common.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/amx/mmq.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/amx/mmq.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/arch/arm/cpu-feats.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/arch/arm/quants.c +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/arch/arm/repack.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/arch/loongarch/quants.c +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/arch/powerpc/cpu-feats.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/arch/powerpc/quants.c +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/arch/riscv/cpu-feats.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/arch/riscv/quants.c +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/arch/riscv/repack.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/arch/s390/cpu-feats.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/arch/s390/quants.c +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/arch/s390/repack.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/arch/wasm/quants.c +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/arch/x86/cpu-feats.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/arch/x86/quants.c +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/arch/x86/repack.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/arch-fallback.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/binary-ops.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/binary-ops.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/cmake/FindSIMD.cmake +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/cmake/FindSMTIME.cmake +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/common.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/ggml-cpu-impl.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/ggml-cpu.c +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/ggml-cpu.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/hbm.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/hbm.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/kleidiai/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/kleidiai/kernels.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/kleidiai/kernels.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/kleidiai/kleidiai.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/kleidiai/kleidiai.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/llamafile/sgemm.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/llamafile/sgemm.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/ops.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/ops.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/quants.c +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/quants.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/repack.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/repack.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/simd-gemm.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/simd-mappings.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/spacemit/ime.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/spacemit/ime.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/spacemit/ime1_kernels.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/spacemit/ime2_kernels.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/spacemit/ime_env.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/spacemit/ime_env.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/spacemit/ime_kernels.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/spacemit/repack.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/spacemit/repack.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/spacemit/rvv_kernels.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/spacemit/rvv_kernels.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/spacemit/spine_barrier.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/spacemit/spine_mem_pool.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/spacemit/spine_mem_pool.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/spacemit/spine_tcm.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/tiled/tiled-kernel.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/tiled/tiled-kernel.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/tiled/tiled.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/tiled/tiled.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/traits.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/traits.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/unary-ops.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/unary-ops.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/vec.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cpu/vec.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cuda/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cuda/vendors/cuda.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cuda/vendors/hip.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-cuda/vendors/musa.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-et/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-et/cmake/embed_one_kernel.cmake +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-et/et-kernels/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-et/et-kernels/src/block_ops.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-et/et-kernels/src/ggml_tensor.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-et/et-kernels/src/math_fp.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-et/et-kernels/src/platform.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-et/et-kernels/src/quants.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-et/et-kernels/src/tensor.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-et/ggml-et-common.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-et/ggml-et-cpu-compare.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-et/ggml-et-cpu-compare.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-et/ggml-et-kernels.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-et/ggml-et-kernels.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-et/ggml-et-memops.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-et/ggml-et-memops.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-et/ggml-et-ops.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-et/ggml-et-ops.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-et/ggml-et-uberkernel-common.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-et/ggml-et.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-feats.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/ggml-hexagon.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/act-ops.c +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/allreduce-ops.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/argsort-ops.c +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/argsort-ops.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/binary-ops.c +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/binary-ops.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/cmake-toolchain.cmake +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/concat-ops.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/cpy-ops.c +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/cpy-ops.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/cumsum-ops.c +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/dma-copy.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/dma-queue.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/flash-attn-ops.c +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/flash-attn-ops.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/gated-delta-net-ops.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/gather-rows.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/get-rows-ops.c +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/get-rows-ops.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hex-bitmap.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hex-common.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hex-dump.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hex-fastdiv.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hex-profile.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hex-utils.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hmx-fa-kernels.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hmx-mm-kernels-tiled.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hmx-queue.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hmx-utils.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/htp-ctx.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/htp-fence.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/htp-ops.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/htp-tensor.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/htp-vtcm.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hvx-arith.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hvx-base.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hvx-copy.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hvx-div.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hvx-dump.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hvx-erf.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hvx-exp.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hvx-fa-kernels.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hvx-flash-attn.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hvx-floor.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hvx-inverse.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hvx-log.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hvx-mm-kernels-float.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hvx-mm-kernels-tiled.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hvx-norm.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hvx-pow.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hvx-quant.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hvx-reduce.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hvx-repl.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hvx-scale.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hvx-sigmoid.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hvx-sin-cos.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hvx-sqrt.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hvx-types.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/hvx-utils.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/im2col-ops.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/main.c +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/matmul-ops.c +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/matmul-ops.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/pool-ops.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/repeat-ops.c +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/rope-ops.c +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/rope-ops.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/set-rows-ops.c +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/set-rows-ops.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/softmax-ops.c +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/softmax-ops.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/ssm-conv.c +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/ssm-conv.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/sum-rows-ops.c +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/unary-ops.c +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/unary-ops.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp/work-queue.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp-drv.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp-drv.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/htp-opnode.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hexagon/libdl.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-hip/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-impl.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/ggml-metal-common.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/ggml-metal-common.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/ggml-metal-context.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/ggml-metal-context.m +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/ggml-metal-device.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/ggml-metal-device.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/ggml-metal-device.m +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/ggml-metal-fusion.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/ggml-metal-fusion.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/ggml-metal-impl.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/ggml-metal-ops.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/ggml-metal-ops.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/ggml-metal-tuning.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/ggml-metal-tuning.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/ggml-metal.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/argsort.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/binbcast.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/common.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/conv.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/dequantize.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/fa_aux.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/fa_common.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/fa_f16.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/fa_f32.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/fa_q4_0.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/fa_q4_1.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/fa_q5_0.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/fa_q5_1.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/fa_q8_0.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/fa_turbo.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/fa_vec_common.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/fa_vec_f16.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/fa_vec_f32.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/fa_vec_q4_0.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/fa_vec_q4_1.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/fa_vec_q5_0.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/fa_vec_q5_1.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/fa_vec_q8_0.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/fa_vec_turbo.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/gated_delta_net.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/misc.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/mul_mm.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/mul_mv.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/mul_mv_mma.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/norm.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/pool.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/quantize.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/quantize.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/reduce.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/rope.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/softmax.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/solve_tri.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/ssm.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/tri.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/turbo.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/unary.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/upscale.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/kernels/wkv.metal +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/turbo-matrices.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-metal/turbo-wht.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-musa/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-opencl/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-opencl/cl-program-cache.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-opencl/cl-program-cache.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-opencl/fa_tune.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-opencl/ggml-opencl.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-opencl/libdl.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/ggml-decoder.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/ggml-decoder.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/ggml-openvino-extra.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/ggml-openvino-extra.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/ggml-openvino.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/ggml-quants.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/ggml-quants.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/model-cache.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/model-cache.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/decoder.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/frontend.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/frontend.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/input_model.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/input_model.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/node_context.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/add.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/add_id.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/argsort.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/clamp.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/concat.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/cont.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/conv.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/cpy.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/cumsum.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/diag.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/div.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/fill.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/flash_attn_ext.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/gated_delta_net.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/gated_delta_net.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/gather_matmul.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/get_rows.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/glu_geglu.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/glu_geglu_quick.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/glu_swiglu.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/im2col.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/l2_norm.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/mean.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/moe_compressed.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/mul_mat_id.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/mulmat.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/norm.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/pad.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/permute.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/pool_2d.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/repeat.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/reshape.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/rms_norm.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/roll.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/rope.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/scale.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/set.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/set_rows.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/softmax.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/solve_tri.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/sqr.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/ssm_conv.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/sum.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/sum_rows.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/transpose.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/tri.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/unary.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/upscale.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op/view.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op_table.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/op_table.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/pass/align_eltwise_ranks.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/pass/align_eltwise_ranks.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/pass/fuse_argsort_topk.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/pass/fuse_argsort_topk.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/pass/fuse_moe_compressed.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/pass/fuse_moe_compressed.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/pass/fuse_moe_router.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/pass/fuse_moe_router.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/pass/fuse_to_conv.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/pass/fuse_to_conv.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/pass/fuse_to_sdpa.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/pass/fuse_to_sdpa.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/pass/kv_state_seq_axis.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/pass/kv_state_seq_axis.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/pass/mark_decompression_convert_constant_folding.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/pass/mark_dequantization_subgraph.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/pass/squeeze_matmul.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/pass/squeeze_matmul.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/rt_info/weightless_caching_attributes.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/translate_session.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/translate_session.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/utils.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/openvino/utils.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/utils.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-openvino/utils.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-opt.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-quants.c +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-quants.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-rpc/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-rpc/ggml-rpc.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-rpc/log.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-rpc/log.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-rpc/transport-apple.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-rpc/transport-apple.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-rpc/transport.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-rpc/transport.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/add-id.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/add-id.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/backend.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/base.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/binbcast.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/binbcast.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/col2im-1d.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/col2im-1d.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/common.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/common.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/concat.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/concat.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/conv.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/conv.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/conv2d-dw.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/conv2d-dw.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/conv2d-transpose.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/conv2d-transpose.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/conv2d.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/conv2d.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/conv3d.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/conv3d.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/convert.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/convert.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/count-equal.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/count-equal.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/cpy.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/cpy.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/cross_entropy_loss.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/cross_entropy_loss.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/cumsum.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/cumsum.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/dequantize.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/diag.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/diag.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/dmmv.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/dmmv.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/dpct/helper.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/dsv4-hc.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/dsv4-hc.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/element_wise.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/element_wise.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/esimd.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/fattn-common.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/fattn-mkl.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/fattn-onednn.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/fattn-onednn.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/fattn-sparse.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/fattn-sparse.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/fattn-tile.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/fattn-tile.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/fattn-vec.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/fattn.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/fattn.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/fill.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/fill.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/fused-gemm.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/fused-gemm.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/fusion.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/fusion.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/fwht.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/fwht.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/gated_delta_net.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/gated_delta_net.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/gemm.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/getrows.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/getrows.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/ggml-sycl.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/gla.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/gla.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/im2col.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/im2col.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/lightning-indexer.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/lightning-indexer.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/mem.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/mem.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/memtrace.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/memtrace.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/mmq.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/mmq.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/mmvq.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/mmvq.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/norm.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/norm.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/opt-step.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/opt-step.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/outprod.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/outprod.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/pad.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/pad.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/pad_reflect_1d.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/pad_reflect_1d.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/pool.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/pool.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/presets.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/quantize.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/quants.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/repeat_back.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/repeat_back.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/roll.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/roll.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/rope.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/rope.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/set.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/set.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/set_rows.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/set_rows.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/softmax.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/softmax.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/solve_tri.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/solve_tri.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/ssm_conv.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/ssm_conv.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/ssm_scan.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/ssm_scan.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/sycl_hw.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/sycl_hw.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-tile-instance-dkq112-dv112.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-tile-instance-dkq128-dv128.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-tile-instance-dkq256-dv256.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-tile-instance-dkq40-dv40.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-tile-instance-dkq512-dv512.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-tile-instance-dkq576-dv512.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-tile-instance-dkq64-dv64.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-tile-instance-dkq72-dv72.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-tile-instance-dkq80-dv80.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-tile-instance-dkq96-dv96.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-f16-f16.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-f16-q4_0.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-f16-q4_1.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-f16-q5_0.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-f16-q5_1.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-f16-q8_0.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_0-f16.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_0-q4_0.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_0-q4_1.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_0-q5_0.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_0-q5_1.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_0-q8_0.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_1-f16.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_1-q4_0.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_1-q4_1.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_1-q5_0.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_1-q5_1.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_1-q8_0.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_0-f16.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_0-q4_0.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_0-q4_1.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_0-q5_0.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_0-q5_1.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_0-q8_0.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_1-f16.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_1-q4_0.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_1-q4_1.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_1-q5_0.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_1-q5_1.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_1-q8_0.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q8_0-f16.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q8_0-q4_0.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q8_0-q4_1.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q8_0-q5_0.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q8_0-q5_1.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q8_0-q8_0.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/topk-moe.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/topk-moe.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/topk-radix.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/topk-radix.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/tsembd.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/tsembd.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/type.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/upscale.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/upscale.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/vecdotq.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/wkv.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-sycl/wkv.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-threading.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-threading.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-turbo-quant.c +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/apir_cs_ggml-rpc-front.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/backend/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/backend/apir_cs_ggml-rpc-back.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/backend/backend-convert.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/backend/backend-dispatched-backend.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/backend/backend-dispatched-buffer-type.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/backend/backend-dispatched-buffer.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/backend/backend-dispatched-device.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/backend/backend-dispatched.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/backend/backend-dispatched.gen.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/backend/backend-dispatched.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/backend/backend-virgl-apir.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/backend/backend.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/backend/shared/api_remoting.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/backend/shared/apir_backend.gen.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/backend/shared/apir_backend.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/backend/shared/apir_cs.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/backend/shared/apir_cs_ggml.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/backend/shared/apir_cs_rpc.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/ggml-backend-buffer-type.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/ggml-backend-buffer.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/ggml-backend-device.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/ggml-backend-reg.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/ggml-backend.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/ggml-remoting.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/include/apir_hw.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/virtgpu-apir.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/virtgpu-forward-backend.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/virtgpu-forward-buffer-type.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/virtgpu-forward-buffer.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/virtgpu-forward-device.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/virtgpu-forward-impl.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/virtgpu-forward.gen.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/virtgpu-shm.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/virtgpu-shm.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/virtgpu-utils.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/virtgpu-utils.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/virtgpu.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-virtgpu/virtgpu.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-vulkan/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-vulkan/ggml-vulkan-buffers.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-vulkan/ggml-vulkan-common.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-vulkan/ggml-vulkan-debug.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-vulkan/ggml-vulkan-push-constants.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-vulkan/ggml-vulkan-types.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-vulkan/ggml-vulkan.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-vulkan/vulkan-shaders/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-webgpu/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-webgpu/ggml-webgpu-shader-lib.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-webgpu/ggml-webgpu.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-webgpu/pre_wgsl.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-zdnn/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-zdnn/common.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-zdnn/ggml-zdnn.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-zdnn/mmf.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-zdnn/mmf.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-zdnn/mmq.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-zdnn/mmq.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-zdnn/quantize.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-zdnn/quantize.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-zdnn/utils.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-zdnn/utils.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-zendnn/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml-zendnn/ggml-zendnn.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml.c +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/ggml.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/ggml/src/gguf.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/include/llama-cpp.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/include/llama.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/licenses/LICENSE-jsonhpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-adapter.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-adapter.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-batch.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-batch.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-chat.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-chat.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-context.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-context.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-cparams.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-cparams.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-ext.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-grammar.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-grammar.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-hparams.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-hparams.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-impl.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-impl.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-io.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-io.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-kv-cache-dsa-iswa.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-kv-cache-dsa-iswa.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-kv-cache-dsa.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-kv-cache-dsa.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-kv-cache-dsv4.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-kv-cache-dsv4.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-kv-cache-iswa.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-kv-cache-iswa.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-kv-cache-msa.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-kv-cache-msa.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-kv-cache.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-kv-cells.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-memory-hybrid-idx.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-memory-hybrid-idx.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-memory-hybrid-iswa.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-memory-hybrid-iswa.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-memory-hybrid.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-memory-hybrid.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-memory-recurrent.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-memory-recurrent.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-memory.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-memory.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-mmap.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-mmap.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-model-loader.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-model-loader.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-model-saver.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-model-saver.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-model.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-moe-cache.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-moe-cache.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-quant.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-quant.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-sampler.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama-sampler.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/llama.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/afmoe.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/apertus.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/arcee.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/arctic.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/arwkv7.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/baichuan.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/bailingmoe.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/bailingmoe2.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/bailingmoe3.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/bert.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/bitnet.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/bloom.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/chameleon.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/chatglm.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/clef.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/clip.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/codeshell.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/cogvlm.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/cohere2.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/cohere2moe.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/command-r.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/dbrx.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/deci.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/deepseek.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/deepseek2.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/deepseek2ocr.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/deepseek32.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/deepseek4.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/delta-net-base.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/dflash.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/diffusion-gemma.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/dots1.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/dots3note.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/dream.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/eagle3.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/ernie4-5-moe.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/ernie4-5.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/eurobert.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/exaone-moe.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/exaone.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/exaone4.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/falcon-h1.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/falcon.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/gemma-embedding.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/gemma-embedding2.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/gemma.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/gemma2.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/gemma3.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/gemma3n.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/gemma4-assistant.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/gemma4-common.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/gemma4.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/glm-dsa.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/glm4-moe.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/glm4.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/glm5-next.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/gpt2.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/gptneox.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/granite-hybrid.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/granite-moe.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/granite-swa.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/granite-switch.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/granite.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/grok.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/grovemoe.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/hrm-text.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/hunyuan-dense.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/hunyuan-moe.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/hunyuan-vl.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/hy-v3.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/hy-v4.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/internlm2.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/jais.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/jais2.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/jamba.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/jina-bert-v2.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/jina-bert-v3.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/k2-horizon.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/kimi-k3.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/kimi-linear.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/kolibri1.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/laguna.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/lfm2.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/lfm2moe.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/llada-moe.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/llada.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/llama-embed.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/llama.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/llama4.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/maincoder.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/mamba-base.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/mamba.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/mamba2.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/maple.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/mellum.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/mimo2.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/minicpm.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/minicpm3.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/minimax-01.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/minimax-m2.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/minimax-m3.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/mistral3.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/mistral4.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/modern-bert.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/mpt.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/muse-glimmer.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/nanbeige.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/nemotron-h-moe.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/nemotron-h.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/nemotron.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/neo-bert.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/nomic-bert-moe.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/nomic-bert.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/olmo.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/olmo2.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/olmoe.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/openai-moe.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/openelm.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/orion.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/paddleocr.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/pangu-embed.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/phi2.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/phi3.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/phimoe.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/plamo.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/plamo2.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/plamo3.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/plm.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/pockettts.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/qwen.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/qwen2.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/qwen2moe.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/qwen2vl.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/qwen3.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/qwen35.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/qwen35moe.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/qwen3moe.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/qwen3next.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/qwen3tts.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/qwen3vl.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/qwen3vlmoe.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/qwen4exp.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/refact.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/rnd1.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/rwkv6-base.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/rwkv6.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/rwkv6qwen2.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/rwkv7-base.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/rwkv7.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/seed-oss.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/smallthinker.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/smollm3.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/spark2-5.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/stablelm.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/starcoder.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/starcoder2.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/step35.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/t5.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/t5encoder.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/talkie.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/vllm-decision2.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/wavtokenizer-dec.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/models/xverse.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/turbo-rotation-data-32.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/turbo-rotation-data.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/unicode-data.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/unicode-data.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/unicode.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/src/unicode.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/batched-bench/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/batched-bench/batched-bench.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/batched-bench/main.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/cli/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/cli/cli-client.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/cli/cli-client.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/cli/cli-context.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/cli/cli-context.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/cli/cli-server.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/cli/cli-ui.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/cli/cli.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/cli/main.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/completion/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/completion/completion.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/completion/main.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/cvector-generator/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/cvector-generator/cvector-generator.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/cvector-generator/mean.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/cvector-generator/pca.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/diffusion/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/diffusion/diffusion-cli.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/diffusion/diffusion.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/diffusion/diffusion.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/diffusion-gemma-server/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/diffusion-gemma-server/diffusion-gemma-server.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/diffusion-gemma-server/diffusion-gemma-visual-server.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/export-lora/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/export-lora/export-lora.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/fit-params/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/fit-params/fit-params.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/fit-params/main.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/gguf-split/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/gguf-split/gguf-split.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/imatrix/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/imatrix/imatrix.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/llama-bench/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/llama-bench/llama-bench.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/llama-bench/main.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/clip-graph.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/clip-impl.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/clip-model.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/clip.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/clip.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/debug/mtmd-debug.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/debug/mtmd-debug.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/deprecation-warning.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/cogvlm.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/conformer.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/deepseek4v.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/deepseekocr.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/deepseekocr2.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/dots3note.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/dotsocr.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/exaone4_5.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/gemma4a.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/gemma4ua.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/gemma4uv.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/gemma4v.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/glm4v.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/granite-speech.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/granite4-vision.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/hunyuanocr.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/hunyuanvl.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/internvl.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/kimik25.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/kimivl.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/ling3vl.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/llama4.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/llava.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/mimo-audio.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/mimovl.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/minicpmv.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/minimax-m3.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/mobilenetv5.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/models.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/muse-glimmer.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/nemotron-v2-vl.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/paddleocr.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/parakeet.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/pixtral.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/pockettts-gen.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/pockettts-seanet.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/pockettts-spkenc.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/qwen2vl.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/qwen3a.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/qwen3tts-gen.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/qwen3tts-spkenc.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/qwen3vl.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/siglip.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/step3vl.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/whisper-enc.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/yasa2.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/models/youtuvl.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/mtmd-audio.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/mtmd-audio.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/mtmd-cli.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/mtmd-helper-common.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/mtmd-helper-gen.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/mtmd-helper.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/mtmd-helper.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/mtmd-image.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/mtmd-image.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/mtmd-internal.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/mtmd.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/mtmd/mtmd.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/perplexity/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/perplexity/main.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/perplexity/perplexity.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/quantize/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/quantize/main.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/quantize/quantize.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/results/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/results/results.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/rpc/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/rpc/rpc-server.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/server/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/server/main.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/server/server-chat.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/server/server-chat.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/server/server-common.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/server/server-common.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/server/server-context.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/server/server-context.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/server/server-cors-proxy.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/server/server-decision.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/server/server-decision.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/server/server-http.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/server/server-http.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/server/server-mcp.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/server/server-mcp.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/server/server-models.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/server/server-models.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/server/server-queue.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/server/server-queue.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/server/server-schema.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/server/server-schema.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/server/server-stream.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/server/server-stream.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/server/server-task.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/server/server-task.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/server/server-tools.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/server/server-tools.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/server/server.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/tokenize/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/tokenize/tokenize.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/tts/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/tts/tts.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/tuning/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/tuning/bench.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/tuning/bench.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/tuning/fa-vec.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/tuning/fa-vec.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/tuning/main.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/tools/ui/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/vendor/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/vendor/cpp-httplib/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/vendor/cpp-httplib/httplib.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/vendor/cpp-httplib/httplib.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/vendor/hash/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/vendor/hash/hash.cpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/vendor/hash/hash.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/vendor/hash/rotate-bits/rotate-bits.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/vendor/hash/sha1/sha1.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/vendor/hash/sha256/sha256.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/vendor/hash/xxhash/xxhash.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/vendor/miniaudio/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/vendor/miniaudio/miniaudio.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/vendor/nlohmann/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/vendor/nlohmann/json.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/vendor/nlohmann/json_fwd.hpp +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/vendor/sheredom/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/vendor/sheredom/subprocess.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/vendor/stb/CMakeLists.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/llama-cpp-turboquant/vendor/stb/stb_image.h +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/setup.cfg +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/setup.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/__main__.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/_subproc_env.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/bin/__init__.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/bootstrap.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/config.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/decision.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/diagnostics.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/entrypoint.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/errors.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/events.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/hf_quants.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/memory_guard.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/model_config.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/model_families.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/model_files.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/models.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/paths.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/performance.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/reasoning_capabilities.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/recommendations.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/redaction.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/runtime.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/server_manager.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/skills/debug.md +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/skills/edit-verified.md +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/skills/explain.md +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/skills/finish-verified.md +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/skills/git-commit-safely.md +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/skills/locate.md +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/skills/plan-task.md +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/skills/review.md +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/skills/run-tests.md +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/thermal.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/thinking.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/transfer.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/ui/UPSTREAM_COMMIT +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/ui/__init__.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/ui/api.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/ui/auth.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/ui/context_budget.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/ui/diagnostics.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/ui/gateway.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/ui/http.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/ui/launch.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/ui/plugin/context.ts +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/ui/plugin/localcode.ts +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/ui/plugin/modes.ts +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/ui/plugin/rpc.ts +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/ui/plugin/tui.ts +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/ui/ports.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/ui/process_guard.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/ui/server_cmd.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/ui/warmup.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/voice.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode/voice_live.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode.egg-info/dependency_links.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode.egg-info/entry_points.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode.egg-info/requires.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/src/localcode.egg-info/top_level.txt +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_architecture.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_bootstrap.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_build_ui_cleanup.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_cancel_download.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_compatibility.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_comprehensive_bg_download.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_comprehensive_binary.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_comprehensive_cli.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_comprehensive_decisions.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_comprehensive_diffusion.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_comprehensive_download.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_comprehensive_install.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_comprehensive_machines.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_comprehensive_model_gate.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_comprehensive_project_trust.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_comprehensive_upstream_gates.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_comprehensive_voice.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_comprehensive_whisper.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_config.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_context_safety.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_download.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_eval_gates.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_eval_performance.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_eval_suite.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_existing_model_files.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_fresh_install.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_kv_cap.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_metal_kv_selection.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_model_arg_resolution.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_model_config.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_model_verification.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_ram_ctx_ladder.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_real_models.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_reasoning_capabilities.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_redaction.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_release_identity.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_release_intent.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_resumable_transfer.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_runtime.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_server_flags_accepted.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_service_diagnostics.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_speed_quality.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_subproc_env.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_thermal.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_ui_api.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_ui_auth.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_ui_context_budget.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_ui_gateway_live.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_ui_launch.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_ui_local_http.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_ui_picker_nonblocking.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_ui_process_guard.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_ui_server.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_ui_shutdown.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_ui_warmup.py +0 -0
- {localcode-0.5.32 → localcode-0.5.34}/tests/test_voice_live.py +0 -0
|
@@ -1,7 +1,9 @@
|
|
|
1
1
|
#include "lexer.h"
|
|
2
2
|
#include "runtime.h"
|
|
3
|
+
#include "unicode.h"
|
|
3
4
|
|
|
4
5
|
#include <cctype>
|
|
6
|
+
#include <charconv>
|
|
5
7
|
#include <functional>
|
|
6
8
|
#include <map>
|
|
7
9
|
#include <string>
|
|
@@ -75,6 +77,24 @@ lexer_result lexer::tokenize(const std::string & source) {
|
|
|
75
77
|
}
|
|
76
78
|
// add escaped char
|
|
77
79
|
char escaped_char = src[pos++];
|
|
80
|
+
if (escaped_char == 'x') {
|
|
81
|
+
if (pos + 1 >= src.size()) {
|
|
82
|
+
throw lexer_exception("incomplete hexadecimal escape", source, pos);
|
|
83
|
+
}
|
|
84
|
+
|
|
85
|
+
unsigned int value = 0;
|
|
86
|
+
const char * first = &src[pos];
|
|
87
|
+
const char * last = first + 2;
|
|
88
|
+
|
|
89
|
+
auto [ptr, ec] = std::from_chars(first, last, value, 16);
|
|
90
|
+
if (ec != std::errc{} || ptr != last) {
|
|
91
|
+
throw lexer_exception("invalid hexadecimal escape", source, pos);
|
|
92
|
+
}
|
|
93
|
+
|
|
94
|
+
str += common_unicode_cpt_to_utf8(value);
|
|
95
|
+
pos += 2;
|
|
96
|
+
continue;
|
|
97
|
+
}
|
|
78
98
|
if (escape_chars.find(escaped_char) == escape_chars.end()) {
|
|
79
99
|
throw lexer_exception(std::string("unknown escape character \\") + escaped_char, source, pos);
|
|
80
100
|
}
|
|
@@ -423,7 +423,12 @@ const func_builtins & global_builtins() {
|
|
|
423
423
|
throw raised_exception("Jinja Exception: " + msg);
|
|
424
424
|
}},
|
|
425
425
|
{"dict", toobject},
|
|
426
|
-
{"namespace",
|
|
426
|
+
{"namespace", [](const func_args & args) -> value {
|
|
427
|
+
// unlike a dict, a namespace has no methods: ns.items is the attribute, as in jinja2
|
|
428
|
+
value out = toobject(args);
|
|
429
|
+
cast_val<value_object>(out)->has_builtins = false;
|
|
430
|
+
return out;
|
|
431
|
+
}},
|
|
427
432
|
{"strftime_now", [](const func_args & args) -> value {
|
|
428
433
|
args.ensure_vals<value_string>();
|
|
429
434
|
std::string format = args.get_pos(0)->as_string().str();
|
|
@@ -34,6 +34,7 @@ static const std::map<llm_arch, const char *> LLM_ARCH_NAMES = {
|
|
|
34
34
|
{ LLM_ARCH_QWEN2MOE, "qwen2moe" },
|
|
35
35
|
{ LLM_ARCH_QWEN2VL, "qwen2vl" },
|
|
36
36
|
{ LLM_ARCH_QWEN3, "qwen3" },
|
|
37
|
+
{ LLM_ARCH_FLINT, "flint" },
|
|
37
38
|
{ LLM_ARCH_QWEN3MOE, "qwen3moe" },
|
|
38
39
|
{ LLM_ARCH_QWEN3NEXT, "qwen3next" },
|
|
39
40
|
{ LLM_ARCH_QWEN3VL, "qwen3vl" },
|
|
@@ -1652,6 +1652,43 @@ ggml_tensor * llm_graph_context::build_lora_mm(
|
|
|
1652
1652
|
return res;
|
|
1653
1653
|
}
|
|
1654
1654
|
|
|
1655
|
+
llm_graph_amax_src1 llm_graph_context::build_amax_src1_f16(
|
|
1656
|
+
ggml_tensor * cur) const {
|
|
1657
|
+
llm_graph_amax_src1 result = {
|
|
1658
|
+
/*.res =*/ cur,
|
|
1659
|
+
/*.scale_s =*/ nullptr,
|
|
1660
|
+
};
|
|
1661
|
+
|
|
1662
|
+
const int64_t n_embd = cur->ne[0];
|
|
1663
|
+
|
|
1664
|
+
// amax = max|cur| along the rows for each [expert, token] -> [1, n_expert, n_tokens]
|
|
1665
|
+
ggml_tensor * ab = ggml_abs(ctx0, cur);
|
|
1666
|
+
ggml_tensor * amax = ggml_pool_1d(ctx0, ab, GGML_OP_POOL_MAX, n_embd, n_embd, 0);
|
|
1667
|
+
|
|
1668
|
+
// ref: https://github.com/ggml-org/llama.cpp/pull/26223
|
|
1669
|
+
constexpr float k_log2_e = 1.4426950408889634f; // 1 / ln(2)
|
|
1670
|
+
constexpr float k_ln_2 = 0.6931471805599453f;
|
|
1671
|
+
|
|
1672
|
+
ggml_tensor * e;
|
|
1673
|
+
// note: this can technically be:
|
|
1674
|
+
// e = ggml_scale(ctx0, amax, 1.0f/32768.0f);
|
|
1675
|
+
// but we leave x32 headroom for the F16 sums of 32 values in Q8 activation blocks
|
|
1676
|
+
e = ggml_scale(ctx0, amax, 1.0f/1024.0f);
|
|
1677
|
+
e = ggml_log (ctx0, e);
|
|
1678
|
+
e = ggml_scale(ctx0, e, k_log2_e);
|
|
1679
|
+
e = ggml_ceil (ctx0, e);
|
|
1680
|
+
e = ggml_clamp(ctx0, e, 0.0f, 127.0f);
|
|
1681
|
+
e = ggml_scale(ctx0, e, k_ln_2);
|
|
1682
|
+
e = ggml_exp (ctx0, e);
|
|
1683
|
+
|
|
1684
|
+
result = {
|
|
1685
|
+
/*.res =*/ ggml_div(ctx0, cur, e),
|
|
1686
|
+
/*.scale_s =*/ e,
|
|
1687
|
+
};
|
|
1688
|
+
|
|
1689
|
+
return result;
|
|
1690
|
+
}
|
|
1691
|
+
|
|
1655
1692
|
ggml_tensor * llm_graph_context::build_lora_mm_id(
|
|
1656
1693
|
ggml_tensor * w, // ggml_tensor * as
|
|
1657
1694
|
ggml_tensor * cur, // ggml_tensor * b
|
|
@@ -2440,10 +2477,17 @@ ggml_tensor * llm_graph_context::build_moe_ffn(
|
|
|
2440
2477
|
GGML_ABORT("fatal error");
|
|
2441
2478
|
}
|
|
2442
2479
|
|
|
2443
|
-
|
|
2444
|
-
|
|
2445
|
-
|
|
2446
|
-
|
|
2480
|
+
if (arch == LLM_ARCH_FLINT) {
|
|
2481
|
+
// Flint's expert activations exceed F16 range; rescale per expert token (upstream #30861, #30928)
|
|
2482
|
+
llm_graph_amax_src1 amax_src1 = build_amax_src1_f16(cur);
|
|
2483
|
+
experts = build_lora_mm_id(down_exps, amax_src1.res, selected_experts, down_exps_s, slots); // [n_embd, n_expert_used, n_tokens]
|
|
2484
|
+
experts = ggml_mul(ctx0, experts, amax_src1.scale_s);
|
|
2485
|
+
} else {
|
|
2486
|
+
experts = build_lora_mm_id(down_exps, cur, selected_experts, down_exps_s, slots); // [n_embd, n_expert_used, n_tokens]
|
|
2487
|
+
if (arch == LLM_ARCH_MISTRAL4) {
|
|
2488
|
+
// src1 can exceed F16 range
|
|
2489
|
+
ggml_prec_set_src(experts, GGML_PREC_F32, 1);
|
|
2490
|
+
}
|
|
2447
2491
|
}
|
|
2448
2492
|
cb(experts, "ffn_moe_down", il);
|
|
2449
2493
|
|
|
@@ -1008,6 +1008,11 @@ struct llm_graph_qkv {
|
|
|
1008
1008
|
ggml_tensor * v; // [n_embd_head, n_head_kv, n_tokens]
|
|
1009
1009
|
};
|
|
1010
1010
|
|
|
1011
|
+
struct llm_graph_amax_src1 {
|
|
1012
|
+
ggml_tensor * res; // cur rescaled
|
|
1013
|
+
ggml_tensor * scale_s; // scale to undo on the matmul result, or nullptr
|
|
1014
|
+
};
|
|
1015
|
+
|
|
1011
1016
|
struct llm_graph_context {
|
|
1012
1017
|
const llm_arch arch;
|
|
1013
1018
|
|
|
@@ -1111,6 +1116,11 @@ struct llm_graph_context {
|
|
|
1111
1116
|
ggml_tensor * w_s = nullptr,
|
|
1112
1117
|
ggml_tensor * slots = nullptr) const;
|
|
1113
1118
|
|
|
1119
|
+
// rescale `cur` (the src1 of a matmul) by a power of two, computed from its maximum absolute
|
|
1120
|
+
// keep both individual values and Q8 activation block sums within F16 range
|
|
1121
|
+
llm_graph_amax_src1 build_amax_src1_f16(
|
|
1122
|
+
ggml_tensor * cur) const;
|
|
1123
|
+
|
|
1114
1124
|
ggml_tensor * build_norm(
|
|
1115
1125
|
ggml_tensor * cur,
|
|
1116
1126
|
ggml_tensor * mw,
|
|
@@ -1072,7 +1072,10 @@ bool llama_kv_cache::update(llama_context * lctx, bool do_shift, const stream_co
|
|
|
1072
1072
|
LLAMA_LOG_DEBUG("%s: applying K-shift\n", __func__);
|
|
1073
1073
|
|
|
1074
1074
|
// apply K-shift if needed
|
|
1075
|
-
if (hparams.rope_type != LLAMA_ROPE_TYPE_NONE
|
|
1075
|
+
if (hparams.rope_type != LLAMA_ROPE_TYPE_NONE &&
|
|
1076
|
+
std::any_of(layers.begin(), layers.end(), [&](const auto & layer) {
|
|
1077
|
+
return hparams.has_rope(layer.il);
|
|
1078
|
+
})) {
|
|
1076
1079
|
ggml_backend_sched_reset(sched);
|
|
1077
1080
|
|
|
1078
1081
|
auto * res = lctx->get_gf_res_reserve();
|
|
@@ -110,6 +110,8 @@ static llama_model * llama_model_mapping(llm_arch arch, const llama_model_params
|
|
|
110
110
|
return new llama_model_qwen2moe(params);
|
|
111
111
|
case LLM_ARCH_QWEN3:
|
|
112
112
|
return new llama_model_qwen3(params);
|
|
113
|
+
case LLM_ARCH_FLINT:
|
|
114
|
+
return new llama_model_flint(params);
|
|
113
115
|
case LLM_ARCH_QWEN3MOE:
|
|
114
116
|
return new llama_model_qwen3moe(params);
|
|
115
117
|
case LLM_ARCH_QWEN3VL:
|
|
@@ -3546,6 +3548,7 @@ llama_rope_type llama_model_rope_type(const llama_model * model) {
|
|
|
3546
3548
|
case LLM_ARCH_QWEN2MOE:
|
|
3547
3549
|
case LLM_ARCH_QWEN3:
|
|
3548
3550
|
case LLM_ARCH_VLLM_DECISION2:
|
|
3551
|
+
case LLM_ARCH_FLINT:
|
|
3549
3552
|
case LLM_ARCH_QWEN3MOE:
|
|
3550
3553
|
case LLM_ARCH_LLADA_MOE:
|
|
3551
3554
|
case LLM_ARCH_RND1:
|
|
@@ -281,6 +281,13 @@ struct llm_tokenizer_bpe : llm_tokenizer {
|
|
|
281
281
|
llm_tokenizer_bpe(const llama_vocab & vocab) {
|
|
282
282
|
GGML_ASSERT(vocab.get_type() == LLAMA_VOCAB_TYPE_BPE);
|
|
283
283
|
switch (vocab.get_pre_type()) {
|
|
284
|
+
case LLAMA_VOCAB_PRE_TYPE_FLINT:
|
|
285
|
+
regex_exprs = {
|
|
286
|
+
"\\p{N}{1,3}",
|
|
287
|
+
"[一-龥-ゟ゠-ヿ]+",
|
|
288
|
+
"[^\\r\\n\\p{L}\\p{N}]?[\\p{Lu}\\p{Lt}\\p{Lm}\\p{Lo}\\p{M}]*[\\p{Ll}\\p{Lm}\\p{Lo}\\p{M}]+(?:'[sS]|'[tT]|'[rR][eE]|'[vV][eE]|'[mM]|'[lL][lL]|'[dD])?|[^\\r\\n\\p{L}\\p{N}]?[\\p{Lu}\\p{Lt}\\p{Lm}\\p{Lo}\\p{M}]+[\\p{Ll}\\p{Lm}\\p{Lo}\\p{M}]*(?:'[sS]|'[tT]|'[rR][eE]|'[vV][eE]|'[mM]|'[lL][lL]|'[dD])?| ?[\\p{P}\\p{S}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+",
|
|
289
|
+
};
|
|
290
|
+
break;
|
|
284
291
|
case LLAMA_VOCAB_PRE_TYPE_LLAMA3:
|
|
285
292
|
regex_exprs = {
|
|
286
293
|
// original regex from tokenizer.json
|
|
@@ -2291,6 +2298,9 @@ void llama_vocab::impl::load(llama_model_loader & ml, const LLM_KV & kv) {
|
|
|
2291
2298
|
LLAMA_LOG_WARN("%s: ************************************ \n", __func__);
|
|
2292
2299
|
LLAMA_LOG_WARN("%s: \n", __func__);
|
|
2293
2300
|
pre_type = LLAMA_VOCAB_PRE_TYPE_DEFAULT;
|
|
2301
|
+
} else if (tokenizer_pre == "flint") {
|
|
2302
|
+
pre_type = LLAMA_VOCAB_PRE_TYPE_FLINT;
|
|
2303
|
+
clean_spaces = false;
|
|
2294
2304
|
} else if (tokenizer_pre == "default") {
|
|
2295
2305
|
pre_type = LLAMA_VOCAB_PRE_TYPE_DEFAULT;
|
|
2296
2306
|
} else if (tokenizer_pre == "minicpm5") {
|
|
@@ -3012,7 +3022,6 @@ void llama_vocab::impl::load(llama_model_loader & ml, const LLM_KV & kv) {
|
|
|
3012
3022
|
// this is currently determined based on the token text, which is obviously not ideal
|
|
3013
3023
|
// ref: https://github.com/ggml-org/llama.cpp/issues/9606
|
|
3014
3024
|
special_eog_ids.clear();
|
|
3015
|
-
|
|
3016
3025
|
if (special_fim_pad_id != LLAMA_TOKEN_NULL && special_eog_ids.count(special_fim_pad_id) == 0) {
|
|
3017
3026
|
special_eog_ids.insert(special_fim_pad_id);
|
|
3018
3027
|
}
|
|
@@ -3045,6 +3054,7 @@ void llama_vocab::impl::load(llama_model_loader & ml, const LLM_KV & kv) {
|
|
|
3045
3054
|
|| t.first == "[EOT]" // Kimi-K2
|
|
3046
3055
|
|| t.first == "[EOS]" // Kimi-K2
|
|
3047
3056
|
|| t.first == "<|end_of_text|>"
|
|
3057
|
+
|| (pre_type == LLAMA_VOCAB_PRE_TYPE_FLINT && (t.first == "<|user|>" || t.first == "<|observation|>"))
|
|
3048
3058
|
|| t.first == "<end_of_utterance>" // smoldocling
|
|
3049
3059
|
|| t.first == "<eos>" // gemma4
|
|
3050
3060
|
|| t.first == "<turn|>" // gemma4
|
|
@@ -0,0 +1,194 @@
|
|
|
1
|
+
#include "models.h"
|
|
2
|
+
|
|
3
|
+
void llama_model_flint::load_arch_hparams(llama_model_loader & ml) {
|
|
4
|
+
ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
|
|
5
|
+
ml.get_key(LLM_KV_LEADING_DENSE_BLOCK_COUNT, hparams.n_layer_dense_lead);
|
|
6
|
+
ml.get_key_or_arr(LLM_KV_EXPERT_FEED_FORWARD_LENGTH, hparams.n_ff_exp_arr, hparams.n_layer_all);
|
|
7
|
+
ml.get_key(LLM_KV_EXPERT_SHARED_COUNT, hparams.n_expert_shared);
|
|
8
|
+
ml.get_key(LLM_KV_EXPERT_WEIGHTS_SCALE, hparams.expert_weights_scale, false);
|
|
9
|
+
ml.get_key(LLM_KV_EXPERT_WEIGHTS_NORM, hparams.expert_weights_norm);
|
|
10
|
+
ml.get_key(LLM_KV_EXPERT_GATING_FUNC, hparams.expert_gating_func);
|
|
11
|
+
|
|
12
|
+
ml.get_key(LLM_KV_EXPERT_SHARED_FEED_FORWARD_LENGTH, hparams.n_ff_shexp);
|
|
13
|
+
ml.get_arr(LLM_KV_ATTENTION_SLIDING_WINDOW_PATTERN, hparams.is_swa_impl);
|
|
14
|
+
ml.get_arr(LLM_KV_ATTENTION_ROPE_PATTERN, hparams.rope_pattern);
|
|
15
|
+
ml.get_key(LLM_KV_ATTENTION_SLIDING_WINDOW, hparams.n_swa);
|
|
16
|
+
hparams.swa_type = LLAMA_SWA_TYPE_STANDARD;
|
|
17
|
+
type = LLM_TYPE_UNKNOWN;
|
|
18
|
+
}
|
|
19
|
+
|
|
20
|
+
void llama_model_flint::load_arch_tensors(llama_model_loader &) {
|
|
21
|
+
LLAMA_LOAD_LOCALS;
|
|
22
|
+
|
|
23
|
+
const int64_t n_ff_exp = hparams.n_ff_exp();
|
|
24
|
+
|
|
25
|
+
tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
|
|
26
|
+
|
|
27
|
+
output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
|
|
28
|
+
output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, 0);
|
|
29
|
+
|
|
30
|
+
for (int i = 0; i < n_layer; ++i) {
|
|
31
|
+
auto & layer = layers[i];
|
|
32
|
+
|
|
33
|
+
layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
|
|
34
|
+
|
|
35
|
+
create_tensor_qkv(layer, i, n_embd, n_embd_head_k * n_head, n_embd_k_gqa, n_embd_v_gqa, 0);
|
|
36
|
+
layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd_head_k * n_head, n_embd}, 0);
|
|
37
|
+
|
|
38
|
+
layer.attn_k_norm = create_tensor(tn(LLM_TENSOR_ATTN_K_NORM, "weight", i), {n_embd_head_k}, 0);
|
|
39
|
+
layer.attn_q_norm = create_tensor(tn(LLM_TENSOR_ATTN_Q_NORM, "weight", i), {n_embd_head_k}, 0);
|
|
40
|
+
|
|
41
|
+
layer.attn_sinks = create_tensor(tn(LLM_TENSOR_ATTN_SINKS, "weight", i), {n_head}, 0);
|
|
42
|
+
|
|
43
|
+
layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
|
|
44
|
+
|
|
45
|
+
if (i < (int) hparams.n_layer_dense_lead) {
|
|
46
|
+
layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0);
|
|
47
|
+
layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd}, 0);
|
|
48
|
+
layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0);
|
|
49
|
+
} else {
|
|
50
|
+
layer.ffn_gate_inp = create_tensor(tn(LLM_TENSOR_FFN_GATE_INP, "weight", i), {n_embd, n_expert}, 0);
|
|
51
|
+
layer.ffn_exp_probs_b = create_tensor(tn(LLM_TENSOR_FFN_EXP_PROBS_B, "bias", i), {n_expert}, TENSOR_NOT_REQUIRED);
|
|
52
|
+
|
|
53
|
+
if (n_expert == 0) {
|
|
54
|
+
throw std::runtime_error("n_expert must be > 0");
|
|
55
|
+
}
|
|
56
|
+
if (n_expert_used == 0) {
|
|
57
|
+
throw std::runtime_error("n_expert_used must be > 0");
|
|
58
|
+
}
|
|
59
|
+
|
|
60
|
+
// MoE branch
|
|
61
|
+
layer.ffn_gate_exps = create_tensor(tn(LLM_TENSOR_FFN_GATE_EXPS, "weight", i), { n_embd, n_ff_exp, n_expert}, 0);
|
|
62
|
+
layer.ffn_down_exps = create_tensor(tn(LLM_TENSOR_FFN_DOWN_EXPS, "weight", i), {n_ff_exp, n_embd, n_expert}, 0);
|
|
63
|
+
layer.ffn_up_exps = create_tensor(tn(LLM_TENSOR_FFN_UP_EXPS, "weight", i), { n_embd, n_ff_exp, n_expert}, 0);
|
|
64
|
+
|
|
65
|
+
// Shared expert branch
|
|
66
|
+
layer.ffn_gate_shexp = create_tensor(tn(LLM_TENSOR_FFN_GATE_SHEXP, "weight", i), {n_embd, hparams.n_ff_shexp}, 0);
|
|
67
|
+
layer.ffn_down_shexp = create_tensor(tn(LLM_TENSOR_FFN_DOWN_SHEXP, "weight", i), { hparams.n_ff_shexp, n_embd}, 0);
|
|
68
|
+
layer.ffn_up_shexp = create_tensor(tn(LLM_TENSOR_FFN_UP_SHEXP, "weight", i), {n_embd, hparams.n_ff_shexp}, 0);
|
|
69
|
+
}
|
|
70
|
+
}
|
|
71
|
+
}
|
|
72
|
+
|
|
73
|
+
std::unique_ptr<llm_graph_context> llama_model_flint::build_arch_graph(const llm_graph_params & params) const {
|
|
74
|
+
return std::make_unique<graph>(*this, params);
|
|
75
|
+
}
|
|
76
|
+
|
|
77
|
+
llama_model_flint::graph::graph(const llama_model & model, const llm_graph_params & params) :
|
|
78
|
+
llm_graph_context(params) {
|
|
79
|
+
const int64_t n_embd_head = hparams.n_embd_head_v();
|
|
80
|
+
|
|
81
|
+
GGML_ASSERT(n_embd_head == hparams.n_embd_head_k());
|
|
82
|
+
GGML_ASSERT(n_embd_head == n_rot);
|
|
83
|
+
|
|
84
|
+
ggml_tensor * cur;
|
|
85
|
+
ggml_tensor * inpL;
|
|
86
|
+
|
|
87
|
+
inpL = build_inp_embd(model.tok_embd);
|
|
88
|
+
|
|
89
|
+
// inp_pos - contains the positions
|
|
90
|
+
ggml_tensor * inp_pos = build_inp_pos();
|
|
91
|
+
|
|
92
|
+
auto * inp_attn = build_attn_inp_kv_iswa();
|
|
93
|
+
|
|
94
|
+
ggml_tensor * inp_out_ids = build_inp_out_ids();
|
|
95
|
+
|
|
96
|
+
for (int il = 0; il < n_layer; ++il) {
|
|
97
|
+
res->t_layer_inp[il] = inpL;
|
|
98
|
+
ggml_tensor * inpSA = inpL;
|
|
99
|
+
|
|
100
|
+
// norm
|
|
101
|
+
cur = build_norm(inpL, model.layers[il].attn_norm, NULL, LLM_NORM_RMS, il);
|
|
102
|
+
cb(cur, "attn_norm", il);
|
|
103
|
+
|
|
104
|
+
// self_attention
|
|
105
|
+
{
|
|
106
|
+
// compute Q and K and RoPE them
|
|
107
|
+
auto [Qcur, Kcur, Vcur] = build_qkv(model.layers[il], cur,
|
|
108
|
+
n_embd_head, n_head, n_head_kv, il);
|
|
109
|
+
|
|
110
|
+
Qcur = build_norm(Qcur, model.layers[il].attn_q_norm, nullptr, LLM_NORM_RMS, il);
|
|
111
|
+
Kcur = build_norm(Kcur, model.layers[il].attn_k_norm, nullptr, LLM_NORM_RMS, il);
|
|
112
|
+
if (hparams.has_rope(il)) {
|
|
113
|
+
Qcur = ggml_rope_ext(ctx0, Qcur, inp_pos, nullptr, n_rot, rope_type, n_ctx_orig,
|
|
114
|
+
freq_base, freq_scale, ext_factor, attn_factor, beta_fast, beta_slow);
|
|
115
|
+
Kcur = ggml_rope_ext(ctx0, Kcur, inp_pos, nullptr, n_rot, rope_type, n_ctx_orig,
|
|
116
|
+
freq_base, freq_scale, ext_factor, attn_factor, beta_fast, beta_slow);
|
|
117
|
+
}
|
|
118
|
+
cb(Qcur, "Qcur", il);
|
|
119
|
+
cb(Kcur, "Kcur", il);
|
|
120
|
+
cb(Vcur, "Vcur", il);
|
|
121
|
+
|
|
122
|
+
cur = build_attn(inp_attn,
|
|
123
|
+
model.layers[il].wo, model.layers[il].wo_b, model.layers[il].wo_s,
|
|
124
|
+
Qcur, Kcur, Vcur, nullptr, model.layers[il].attn_sinks, nullptr, 1.0f / sqrtf(float(n_embd_head)), il);
|
|
125
|
+
}
|
|
126
|
+
if (il == n_layer - 1 && inp_out_ids) {
|
|
127
|
+
cur = ggml_get_rows(ctx0, cur, inp_out_ids);
|
|
128
|
+
inpSA = ggml_get_rows(ctx0, inpSA, inp_out_ids);
|
|
129
|
+
}
|
|
130
|
+
ggml_tensor * ffn_inp = ggml_add(ctx0, cur, inpSA);
|
|
131
|
+
cb(ffn_inp, "ffn_inp", il);
|
|
132
|
+
|
|
133
|
+
// MoE branch
|
|
134
|
+
cur = build_norm(ffn_inp, model.layers[il].ffn_norm, NULL, LLM_NORM_RMS, il);
|
|
135
|
+
cb(cur, "ffn_norm", il);
|
|
136
|
+
|
|
137
|
+
if ((uint32_t) il < hparams.n_layer_dense_lead) {
|
|
138
|
+
cur = build_ffn(cur,
|
|
139
|
+
model.layers[il].ffn_up, NULL, model.layers[il].ffn_up_s,
|
|
140
|
+
model.layers[il].ffn_gate, NULL, model.layers[il].ffn_gate_s,
|
|
141
|
+
model.layers[il].ffn_down, NULL, model.layers[il].ffn_down_s,
|
|
142
|
+
NULL, LLM_FFN_SILU, LLM_FFN_PAR, il);
|
|
143
|
+
cb(cur, "ffn_out", il);
|
|
144
|
+
} else {
|
|
145
|
+
ggml_tensor * moe_out = build_moe_ffn(cur,
|
|
146
|
+
model.layers[il].ffn_gate_inp,
|
|
147
|
+
model.layers[il].ffn_up_exps,
|
|
148
|
+
model.layers[il].ffn_gate_exps,
|
|
149
|
+
model.layers[il].ffn_down_exps,
|
|
150
|
+
model.layers[il].ffn_exp_probs_b,
|
|
151
|
+
n_expert, n_expert_used,
|
|
152
|
+
LLM_FFN_SILU, hparams.expert_weights_norm,
|
|
153
|
+
hparams.expert_weights_scale,
|
|
154
|
+
(llama_expert_gating_func_type) hparams.expert_gating_func,
|
|
155
|
+
il, nullptr, nullptr,
|
|
156
|
+
model.layers[il].ffn_up_exps_s,
|
|
157
|
+
model.layers[il].ffn_gate_exps_s,
|
|
158
|
+
model.layers[il].ffn_down_exps_s);
|
|
159
|
+
cb(moe_out, "ffn_moe_out", il);
|
|
160
|
+
|
|
161
|
+
{
|
|
162
|
+
ggml_tensor * ffn_shexp =
|
|
163
|
+
build_ffn(cur,
|
|
164
|
+
model.layers[il].ffn_up_shexp, NULL, model.layers[il].ffn_up_shexp_s,
|
|
165
|
+
model.layers[il].ffn_gate_shexp, NULL, model.layers[il].ffn_gate_shexp_s,
|
|
166
|
+
model.layers[il].ffn_down_shexp, NULL, model.layers[il].ffn_down_shexp_s,
|
|
167
|
+
NULL, LLM_FFN_SILU, LLM_FFN_PAR, il);
|
|
168
|
+
cb(ffn_shexp, "ffn_shexp", il);
|
|
169
|
+
|
|
170
|
+
cur = ggml_add(ctx0, moe_out, ffn_shexp);
|
|
171
|
+
cb(cur, "ffn_out", il);
|
|
172
|
+
}
|
|
173
|
+
}
|
|
174
|
+
cur = ggml_add(ctx0, cur, ffn_inp);
|
|
175
|
+
|
|
176
|
+
cur = build_cvec(cur, il);
|
|
177
|
+
cb(cur, "l_out", il);
|
|
178
|
+
|
|
179
|
+
// input for next layer
|
|
180
|
+
inpL = cur;
|
|
181
|
+
}
|
|
182
|
+
cur = inpL;
|
|
183
|
+
|
|
184
|
+
cur = build_norm(cur, model.output_norm, NULL, LLM_NORM_RMS, -1);
|
|
185
|
+
|
|
186
|
+
cb(cur, "result_norm", -1);
|
|
187
|
+
res->t_embd = cur;
|
|
188
|
+
|
|
189
|
+
// lm_head
|
|
190
|
+
cur = build_lora_mm(model.output, cur, model.output_s);
|
|
191
|
+
|
|
192
|
+
cb(cur, "result_output", -1);
|
|
193
|
+
res->t_logits = cur;
|
|
194
|
+
}
|
|
@@ -640,6 +640,19 @@ struct llama_model_qwen3moe : public llama_model_base {
|
|
|
640
640
|
};
|
|
641
641
|
|
|
642
642
|
|
|
643
|
+
struct llama_model_flint : public llama_model_base {
|
|
644
|
+
llama_model_flint(const struct llama_model_params & params) : llama_model_base(params) {}
|
|
645
|
+
void load_arch_hparams(llama_model_loader & ml) override;
|
|
646
|
+
void load_arch_tensors(llama_model_loader & ml) override;
|
|
647
|
+
|
|
648
|
+
struct graph : public llm_graph_context {
|
|
649
|
+
graph(const llama_model & model, const llm_graph_params & params);
|
|
650
|
+
};
|
|
651
|
+
|
|
652
|
+
std::unique_ptr<llm_graph_context> build_arch_graph(const llm_graph_params & params) const override;
|
|
653
|
+
};
|
|
654
|
+
|
|
655
|
+
|
|
643
656
|
struct llama_model_qwen3vl : public llama_model_base {
|
|
644
657
|
llama_model_qwen3vl(const struct llama_model_params & params) : llama_model_base(params) {}
|
|
645
658
|
void load_arch_hparams(llama_model_loader & ml) override;
|
|
Binary file
|
|
@@ -575,6 +575,56 @@ CHOICES: list[ModelChoice] = [
|
|
|
575
575
|
mmproj_size_gb=0.6,
|
|
576
576
|
mmproj_hf_filename="mmproj-OpenJev-Q8_0.gguf",
|
|
577
577
|
),
|
|
578
|
+
ModelChoice(
|
|
579
|
+
key="flint",
|
|
580
|
+
revision="20468edfc38ab43d6914cc88b08cbecb05d0fc62",
|
|
581
|
+
sha256="63c2ecc461ca545134b670c660e76fa20f47240f8d39b6510bfd21c4dd6e072f",
|
|
582
|
+
size_bytes=25054688704,
|
|
583
|
+
name="Flint 40B-A2.5B (Q4)",
|
|
584
|
+
hf_repo="unsloth/flint-GGUF",
|
|
585
|
+
filename="flint-UD-Q4_K_XL.gguf",
|
|
586
|
+
size_gb=25.1,
|
|
587
|
+
active_params="2.5B active (40B total MoE, top-6 of 256 experts + 1 shared)",
|
|
588
|
+
architecture="flint",
|
|
589
|
+
license="Apache 2.0",
|
|
590
|
+
humaneval_pass_at_1=None,
|
|
591
|
+
notes=(
|
|
592
|
+
"Recursive's coding-agent MoE (Oct 2026): 39.8B total, ~2.5B active "
|
|
593
|
+
"per token, sliding-window / full attention interleave, 256K context. "
|
|
594
|
+
"Trained for investigating codebases, fixing bugs and tool use. "
|
|
595
|
+
"UD-Q4_K_XL keeps sensitive layers above 4-bit. ~25 GB weights: "
|
|
596
|
+
"needs 48 GB unified memory, comfortable on 64 GB. Text-only. "
|
|
597
|
+
"It reasons on every step and cannot be told not to, so agent "
|
|
598
|
+
"tasks run slower than on Qwen 3.6. Opt-in, never auto-recommended."
|
|
599
|
+
),
|
|
600
|
+
# Flint reasons on every turn and has no off switch: with the server's
|
|
601
|
+
# reasoning flags it still opens <think>, and the thought leaked into the
|
|
602
|
+
# reply. "always" leaves the parser on so it lands in reasoning_content.
|
|
603
|
+
reasoning_control="always",
|
|
604
|
+
reasoning_budget_tokens=0,
|
|
605
|
+
),
|
|
606
|
+
ModelChoice(
|
|
607
|
+
key="flint-q8",
|
|
608
|
+
revision="20468edfc38ab43d6914cc88b08cbecb05d0fc62",
|
|
609
|
+
sha256="5df42b56859c7b5142228005d7d63c327ec9e34c99c72548016d805309a0d284",
|
|
610
|
+
size_bytes=42301076928,
|
|
611
|
+
name="Flint 40B-A2.5B (Q8)",
|
|
612
|
+
hf_repo="unsloth/flint-GGUF",
|
|
613
|
+
filename="flint-Q8_0.gguf",
|
|
614
|
+
size_gb=42.3,
|
|
615
|
+
active_params="2.5B active (40B total MoE, top-6 of 256 experts + 1 shared)",
|
|
616
|
+
architecture="flint",
|
|
617
|
+
license="Apache 2.0",
|
|
618
|
+
humaneval_pass_at_1=None,
|
|
619
|
+
notes=(
|
|
620
|
+
"Same Flint MoE as the Q4 entry at near-lossless Q8_0. Same 2.5B "
|
|
621
|
+
"active parameters, so decode speed stays in the same class. "
|
|
622
|
+
"~42 GB weights: the high-RAM pick, for 96-128 GB unified memory. "
|
|
623
|
+
"Text-only."
|
|
624
|
+
),
|
|
625
|
+
reasoning_control="always",
|
|
626
|
+
reasoning_budget_tokens=0,
|
|
627
|
+
),
|
|
578
628
|
]
|
|
579
629
|
|
|
580
630
|
|
|
@@ -636,8 +686,9 @@ _NO_AUTO_RECOMMEND_ARCHS = {"diffusion_gemma", "kolibri1"}
|
|
|
636
686
|
# Entries excluded by key, for models that share an architecture with a
|
|
637
687
|
# recommended one: OpenJev is a qwen35 like Qwen 3.8, but it is a decision
|
|
638
688
|
# model and must never be a first-run coding default. Saluki remains opt-in
|
|
639
|
-
# until coding quality has been measured on this stack.
|
|
640
|
-
|
|
689
|
+
# until coding quality has been measured on this stack. Flint reasons on every
|
|
690
|
+
# step (no off switch) and timed out where Qwen 3.6 finished on the 48 GB tier.
|
|
691
|
+
_NO_AUTO_RECOMMEND_KEYS = {"openjev", "saluki", "flint", "flint-q8"}
|
|
641
692
|
|
|
642
693
|
|
|
643
694
|
def _auto_recommendable(c: "ModelChoice") -> bool:
|
|
@@ -959,6 +1010,21 @@ MODEL_GROUPS: list[ModelGroup] = [
|
|
|
959
1010
|
"Text-only."
|
|
960
1011
|
),
|
|
961
1012
|
),
|
|
1013
|
+
ModelGroup(
|
|
1014
|
+
key="flint",
|
|
1015
|
+
display_name="Flint 40B-A2.5B",
|
|
1016
|
+
maker="Recursive",
|
|
1017
|
+
hf_repo="unsloth/flint-GGUF",
|
|
1018
|
+
family="flint",
|
|
1019
|
+
architecture="flint",
|
|
1020
|
+
license="Apache 2.0",
|
|
1021
|
+
notes=(
|
|
1022
|
+
"Recursive's coding-agent sparse MoE — 40B total, ~2.5B active. "
|
|
1023
|
+
"Sliding-window / full attention interleave, 256K context, "
|
|
1024
|
+
"reasoning and multi-turn tool use. Unsloth Dynamic quants. "
|
|
1025
|
+
"Text-only."
|
|
1026
|
+
),
|
|
1027
|
+
),
|
|
962
1028
|
ModelGroup(
|
|
963
1029
|
key="openjev",
|
|
964
1030
|
supports_systemone=True,
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
67d27fc5b80eb01836554d49fb76957d528b0ef0
|
|
@@ -262,11 +262,12 @@ export default {
|
|
|
262
262
|
const output = { context: [] as string[] };
|
|
263
263
|
await hooks["experimental.session.compacting"]({}, output);
|
|
264
264
|
if (output.context.length) {
|
|
265
|
+
// Summarizer guidance rides as its own trailing message, so the cached
|
|
266
|
+
// conversation prefix is reused. Appended to the user's last turn it read as
|
|
267
|
+
// something the user said: summaries recorded it as user instructions, and a
|
|
268
|
+
// weaker model summarized the checklist instead of the work.
|
|
265
269
|
const text = output.context.join("\n\n");
|
|
266
|
-
|
|
267
|
-
messages = last?.role === "user" && Array.isArray(last.content)
|
|
268
|
-
? [...messages.slice(0, -1), Object.assign(Object.create(Object.getPrototypeOf(last)), last, { content: [...last.content, { type: "text", text }] })]
|
|
269
|
-
: [...messages, { role: "user", content: [{ type: "text", text }] }];
|
|
270
|
+
messages = [...messages, { role: "user", content: [{ type: "text", text: `[Summary guidance from LocalCode, not from the user] ${text}` }] }];
|
|
270
271
|
}
|
|
271
272
|
}
|
|
272
273
|
event.messages = messages;
|
|
@@ -467,6 +467,37 @@ class Supervisor:
|
|
|
467
467
|
except Exception: # noqa: BLE001
|
|
468
468
|
return False
|
|
469
469
|
|
|
470
|
+
def crash_watchdog(self, interval: float = 1.0, cooldown: float = 120.0) -> None:
|
|
471
|
+
"""Restart the loaded model once after an unexpected exit.
|
|
472
|
+
|
|
473
|
+
A crash mid-reply otherwise leaves every open chat retrying a dead server. One
|
|
474
|
+
automatic restart lets the runtime's retry succeed; a second crash within the
|
|
475
|
+
cooldown stays an error so a model that keeps crashing does not loop.
|
|
476
|
+
"""
|
|
477
|
+
last_restart = float("-inf")
|
|
478
|
+
while not self._closing.is_set():
|
|
479
|
+
time.sleep(interval)
|
|
480
|
+
failure = self._server_failure()
|
|
481
|
+
alias = getattr(self, "current", None)
|
|
482
|
+
if not failure or not alias or time.monotonic() - last_restart < cooldown:
|
|
483
|
+
continue
|
|
484
|
+
if not self.lock.acquire(blocking=False):
|
|
485
|
+
continue
|
|
486
|
+
try:
|
|
487
|
+
if not self._server_failure():
|
|
488
|
+
continue
|
|
489
|
+
last_restart = time.monotonic()
|
|
490
|
+
log(f"supervisor: {failure['detail'].split('. ')[0]}; restarting {alias} once")
|
|
491
|
+
self.state = {"state": "loading", "model": alias, "detail": "restarting after a crash", "pct": None}
|
|
492
|
+
if self.start(alias):
|
|
493
|
+
self.state = {"state": "ready", "model": alias, "detail": "", "pct": None}
|
|
494
|
+
log(f"supervisor: {alias} restarted after a crash")
|
|
495
|
+
else:
|
|
496
|
+
self.state = {"state": "error", "model": alias, "pct": None,
|
|
497
|
+
"detail": "Model server crashed and could not restart. Open /server to retry or choose another model."}
|
|
498
|
+
finally:
|
|
499
|
+
self.lock.release()
|
|
500
|
+
|
|
470
501
|
def server_status(self) -> dict:
|
|
471
502
|
"""Report current liveness, not just the last successful model load."""
|
|
472
503
|
state = dict(self.state)
|
|
@@ -1355,6 +1386,7 @@ def main() -> int:
|
|
|
1355
1386
|
# First-run journey: the TUI opens first and its /models picker loads a model.
|
|
1356
1387
|
sup.state = {"state": "idle", "model": None, "detail": "no model loaded", "pct": None}
|
|
1357
1388
|
log("ready")
|
|
1389
|
+
threading.Thread(target=sup.crash_watchdog, daemon=True).start()
|
|
1358
1390
|
# NOT signal.pause(): it returns on ANY signal, including the SIGCHLD from
|
|
1359
1391
|
# a llama-server we just stopped, which made the supervisor exit mid-switch.
|
|
1360
1392
|
threading.Event().wait()
|