Skip to content

Activity

Showing most recent first

Vulkan: Print coopmat shapes, then exit

0cc4mcreated 0cc4m/vulkan-print-coopmat-shapes • 87dae2f • 
5 hours ago

Deleted branch

ggerganovdeleted gg/update-authors • 
yesterday

authors : update (#12271)

Pull request merge
ggerganovpushed 1 commit to master • 6fefc05…0fd7ca7 • 
yesterday

authors : update

ggerganovcreated gg/update-authors • 9eff3b3 • 
yesterday

ggml-backend : make path_str compatible with C++20 (#12269)

Pull request merge
slarenpushed 1 commit to master • 7ab3643…6fefc05 • 
yesterday

Vulkan: Add device architecture enum and logic to recognize AMD gener…

0cc4mcreated 0cc4m/vulkan-device-architecture • 2584074 • 
yesterday

server : infill gen ends on new line (#12254)

Pull request merge
ggerganovpushed 1 commit to master • 7c7f3b7…7ab3643 • 
yesterday

server : infill gen ends on new line

ggerganovcreated gg/server-infill-end-on-nl • c75753a • 
2 days ago

ggml : skip intermediate .air file when compiling .metallib (#12247)

Pull request merge
danbevpushed 1 commit to master • 102ac18…7c7f3b7 • 
2 days ago

Deleted branch

ggerganovdeleted sync-ggml-25-03-07 • 
2 days ago

sync : ggml

Pull request merge
ggerganovpushed 2 commits to master • 68d0027…102ac18 • 
2 days ago

sync : ggml

ggerganovcreated sync-ggml-25-03-07 • 7d4cd42 • 
2 days ago

ggml-cpu: faster AVX2 variant for IQ1_M (#12216)

Pull request merge
ggerganovpushed 1 commit to master • ea00281…68d0027 • 
2 days ago

llama : remove redundant keywords (struct, enum)

Force push
ggerganovforce pushed to gg/llama-kv-cache-v2 • 766edbf…62ba774 • 
2 days ago

llama : remove redundant keywords (struct, enum)

Force push
ggerganovforce pushed to gg/llama-kv-cache-v2 • f85d0b3…766edbf • 
2 days ago

llama : remove redundant keywords (struct, enum)

ggerganovpushed 1 commit to gg/llama-kv-cache-v2 • 4dbbde7…f85d0b3 • 
2 days ago

ci : fix save-load test invocations (#12245)

Pull request merge
ggerganovpushed 1 commit to master • 8fad3c7…ea00281 • 
2 days ago

ci : fix save-load test invokations

ggerganovcreated gg/ci-fix-save-load • aefa65e • 
2 days ago

server : Log original chat template parsing error (#12233)

Pull request merge
ngxsonpushed 1 commit to master • 7cf64f6…8fad3c7 • 
2 days ago

graph : clean-up

ggerganovpushed 1 commit to gg/llama-kv-cache-v2 • 7ac655b…4dbbde7 • 
2 days ago

clang-tidy : disable bugprone-branch-clone

ggerganovcreated gg/clang-tidy-disable-bugprone • aae2903 • 
2 days ago

sync: minja - support QwQ-32B (#12235)

Pull request merge
ochafikpushed 1 commit to master • 5e2d57b…7cf64f6 • 
2 days ago

context : clean-up

ggerganovpushed 1 commit to gg/llama-kv-cache-v2 • bfef2e0…7ac655b • 
2 days ago

metal : simplify kernel arguments using a struct (#3229) (#12194)

Pull request merge
danbevpushed 1 commit to master • f1648e9…5e2d57b • 
2 days ago

HIP: fix rocWMMA build flags under Windows (#12230)

Pull request merge
JohannesGaesslerpushed 1 commit to master • d6c95b0…f1648e9 • 
2 days ago

metal : fix default.metallib build (#12224)

Pull request merge
danbevpushed 1 commit to master • d76a86d…d6c95b0 • 
2 days ago

opencl: Noncontiguous norm, rms_norm, disable fp16 for some ops (…

Pull request merge
ericcurtinpushed 1 commit to master • 776f9e5…d76a86d • 
2 days ago

cmake : fix undefined reference errors for std::filesystem in ggml (#…

Pull request merge
ericcurtinpushed 1 commit to master • 3d652bf…776f9e5 • 
2 days ago

CUDA: determine FA parallel blocks at runtime

JohannesGaesslercreated jg/cuda-fa-np-runtime • 8ca8cfb • 
2 days ago

readme : update bindings (#12229)

Pull request merge
ggerganovpushed 1 commit to master • 5220a16…3d652bf • 
2 days ago