12月1月2月3月4月5月6月7月8月9月10月11月
周一周三周五
一年内 265370 次贡献
更少更多
- frostbyte_neo 推送了仓库 github-featured/ggml-org--ggml 的 gg/cmake-rm-unused 分支比较 10 提交 »615a37ef42 cmake : remove unused filead04d94462 sync : whisper.cppaa96b947b5 cuda : fix compile warnings (whisper/0)f7cb4b731a sync : llama.cpp00d7d43e4c add tensor type checking as part of cuda graph properties (llama/19186)
- frostbyte_neo 推送了仓库 github-featured/ggml-org--ggml 的 experiments/blocking 分支比较 10 提交 »3afb833f84 wip : unsuccessful attempts speeding mul_mat using blocking67ac34fcfa sync : whisper.cppe2f39f4b52 whisper : sync with whisper.cpp8e3c634b27 whisper : various improvements8ca553add4 whisper : add C-style API
- frostbyte_neo 推送了仓库 github-featured/ggml-org--ggml 的 feature/parallel-decoding-gpt2-example 分支比较 10 提交 »2024e42a1e Remove not needed exit checkc205875285 Add check for end token and mark the stream as finished63ab3d6169 minor : readme993d226faf Fix not generating n_predict tokens and fix warn898718c0cc gpt-2-batched : simplify kv cache stuff (#574)
- frostbyte_neo 推送了仓库 github-featured/ggml-org--ggml 的 fix-mul-mat 分支比较 10 提交 »c5a21cd21b ggml : fix mul_mat src1 indexing when src1 is not contiguousf6365c0605 ggml : apply mul_mat broadcast fix (sync llama.cpp)56216523fa cuda : sync llama.cppc2a5a35879 ggml : broadcast mul_mat + conv batch support (#325)b21406426d ggml : pool naming changes + coding style fixes
- frostbyte_neo 推送了仓库 github-featured/ggml-org--ggml 的 copilot/investigate-failing-tests 分支比较 10 提交 »85837de93e Remove accidentally committed build artifacts3e7fb1d55d Add _codeql_build_dir to gitignoreacc9044246 Complete fix for test-conv2d-dw on ARM SVEa7aa13dab2 Fix test-conv2d-dw failure on ARM SVE by using runtime vector lengthab76d0c7a9 Initial plan
- frostbyte_neo 推送了仓库 github-featured/ggml-org--ggml 的 cerebras-wip 分支比较 10 提交 »23cc0e4425 wip wip wipc55f50a18e ggml : make it work on Windows (#46)24e23612d2 tests : add test-blas068debc017e Fix CMake indentation434b8eaf66 add OpenBLAS detection and modify tests codes (#40)
- frostbyte_neo 推送了仓库 github-featured/ggml-org--ggml 的 copilot/fix-cpu-backend-deadlock 分支比较 10 提交 »aeb817eb34 Fix CPU backend deadlock in RWKV_WKV6 and GLA with 33+ threads2f8bceb70f Initial plana04eea0761 ggml : bump version to 0.9.9 (#1449)15547928f2 sync : whisper.cpp404fcb9d7c sync : llama.cpp
- frostbyte_neo 推送了仓库 github-featured/ggml-org--ggml 的 ci-test 分支比较 10 提交 »c5d62d8c4e ci : test push into a different brancha1caff0f1a ci : fix335f98d0db ci : move out the README.md header generation185ce86ab2 ci : utilize gg_printf9fa436999a ci : fix header
- frostbyte_neo 推送了仓库 github-featured/ggml-org--ggml 的 4bit 分支比较 10 提交 »baeb88b858 tests : add 4-bit Clover-based quantization3b3ad42906 tests : add SVD experimentsa6acb3318a sync : latest whisper.cpp (scratch buffers in ggml)47b297224e Update README.md0467385010 cmake : configure CMAKE_C_FLAGS and target_link_libraries for MSVC (#15)