LLM inference in C/C++
C++ ggml
Commits on master
metal : fix graph capture and handle empty graphs (#29390)
84e76d8 Georgi Gerganov committed 1 month ago
llama-grammar: fix numeric truncation for token_id parsing (#29382)
5cf3a35 Daniel Kuts committed 1 month ago
ui : fix missing svg use and animation elements in preview and download (#28962)
945064f Nandan Vallamdasu committed 1 month ago
ci : use hf-jobs-cpu-performance, disable pytest workers (#29369)
70596c4 Georgi Gerganov committed 1 month ago
vulkan: int8 coopmat1 matmul implementation for AMD RDNA3 and RDNA4 (#27952)
70c4e15 Ruben Ortlam committed 1 month ago
vulkan: handle misalignment in conv_2d and conv_3d (#29365)
6b790a9 Ruben Ortlam committed 1 month ago
vulkan: tune KHR cooperative matrix support for Adreno GPUs (#29328)
3423f94 Raman Shinde committed 1 month ago
server,common : fix the GCC 12 stringop-overread false positive (again) (#29325)
2b70583 Adrien Gallouët committed 1 month ago
test-save-load-state : print a per-model results table in --models mode (#29316)
4c5957c Georgi Gerganov committed 1 month ago
hexagon: reject MUL_MAT_ID when src1 precision is F32 (#29348)
9710a32 Jhen-Jie Hong committed 1 month ago
scripts : make-release-desc - link previous release in changelog title (#29336)
013b31c Georgi Gerganov committed 1 month ago
convert : allow vision target for DFlash/Dspark (#29339)
bd4f514 Tarek Dakhran committed 1 month ago
tests: add `-b/--backend` option to test-llama-archs for testing a specific backend (#27372)
d2e5458 Masashi Yoshimura committed 1 month ago
ci : use hf-jobs-cpu-xl runner in server sanitize workflow (#29297)
6e60f35 Georgi Gerganov committed 1 month ago