Skip to content

Actions: ggml-org/llama.cpp

Actions

EditorConfig Checker

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
editorconfig.yml
2,500+ workflow runs
2,500+ workflow runs

Filter by Event

⌥ + click/return to exclude

Filter by Status

⌥ + click/return to exclude

Filter by Branch

⌥ + click/return to exclude

Filter by Actor

⌥ + click/return to exclude
cuda: update uncoalesced memory reads in pool2d (#29425) EditorConfig Checker #61348: Commit 42c787e pushed by JohannesGaessler
45s master
master
45s
context : simplify setters EditorConfig Checker #61347: Pull request #29730 synchronize by nikwen
54s nikwen:nikwen/simplify-context
nikwen:nikwen/simplify-context
54s
server : preserve context checkpoints across slot save/restore EditorConfig Checker #61346: Pull request #26004 synchronize by Tough-Respawn
Action required Tough-Respawn:fix/kv-restore-checkpoint
Tough-Respawn:fix/kv-restore-checkpoint
Action required
sycl: use ExternalProject to let this backend be built with SYCL compiler while everything else could use another EditorConfig Checker #61345: Pull request #29506 synchronize by a1batross
Action required a1batross:mix-rocm-sycl-build
a1batross:mix-rocm-sycl-build
Action required
server : accumulate generated text and tokens as parse input (#29876) EditorConfig Checker #61344: Commit 18b5f8b pushed by aldehir
3m 56s master
master
3m 56s
model : use exact GELU for ModernBERT encoders EditorConfig Checker #61343: Pull request #30108 opened by boshjerns
Action required boshjerns:modernbert-encoder-gelu-erf
boshjerns:modernbert-encoder-gelu-erf
Action required
vulkan: use 4 rows for NVIDIA MUL_MAT_ID MMVQ except pre-Turing EditorConfig Checker #61342: Pull request #29274 synchronize by SG-Amadeus
Action required SG-Amadeus:sg-amadeus/vulkan-mmid-mmvq-rows4
SG-Amadeus:sg-amadeus/vulkan-mmid-mmvq-rows4
Action required
chat : name tool and argument parser rules by index EditorConfig Checker #61341: Pull request #30088 synchronize by aldehir
56s Frost-54:master
Frost-54:master
56s
llama : add a GPU cache for MoE experts kept in host memory EditorConfig Checker #61340: Pull request #29887 synchronize by am17an
1m 27s am17an:aman/moe-cache
am17an:aman/moe-cache
1m 27s
vulkan: use 4 rows for NVIDIA MUL_MAT_ID MMVQ except pre-Turing EditorConfig Checker #61339: Pull request #29274 synchronize by SG-Amadeus
Action required SG-Amadeus:sg-amadeus/vulkan-mmid-mmvq-rows4
SG-Amadeus:sg-amadeus/vulkan-mmid-mmvq-rows4
Action required
model-loader: add --reclaim-mmap-source to drop dormant mmap pages (Fixes #16761) EditorConfig Checker #61338: Pull request #24156 synchronize by markkobo
Action required markkobo:feature/cpu-repack-reclaim
markkobo:feature/cpu-repack-reclaim
Action required
vulkan : fix TOP_K for +inf/NaN inputs and k = 1 on negative values EditorConfig Checker #61337: Pull request #30107 opened by gianni-cor
Action required gianni-cor:vulkan-topk-nonfinite
gianni-cor:vulkan-topk-nonfinite
Action required
vulkan : one warp tile per subgroup in the small matmul tiles EditorConfig Checker #61336: Pull request #30106 opened by ianloic
Action required ianloic:vulkan-small-warptile-subgroups
ianloic:vulkan-small-warptile-subgroups
Action required
Formalizing scheduler and async backend behavior through tests EditorConfig Checker #61335: Pull request #27258 synchronize by aendk
48s aendk:akieslinger/sched-backend-test-coverage-master
aendk:akieslinger/sched-backend-test-coverage-master
48s
cpu: add F16 SWIGLU_OAI reference and activation-op test coverage EditorConfig Checker #61334: Pull request #29200 synchronize by cqderek
Action required qualcomm:upstream-f16-swiglu-reference
qualcomm:upstream-f16-swiglu-reference
Action required
vulkan : honor FP32 source precision in MUL_MAT_ID EditorConfig Checker #61333: Pull request #30105 opened by linuxid10t
Action required linuxid10t:fix/mistral4-vulkan-offload
linuxid10t:fix/mistral4-vulkan-offload
Action required
hexagon: improved GELU accuracy EditorConfig Checker #61332: Pull request #30104 opened by kurquhar
Action required qualcomm:kurquhar_htp_gelu_accuracy
qualcomm:kurquhar_htp_gelu_accuracy
Action required
CUDA/HIP: fix race in flash_attn_ext_f16_process_tile when nbatch_combine != DKQ/2 EditorConfig Checker #61331: Pull request #30103 opened by IMbackK
56s IMbackK:fattn_fix
IMbackK:fattn_fix
56s
vulkan: use density gate for MUL_MAT_VEC_ID path EditorConfig Checker #61330: Pull request #27332 synchronize by theycallmeloki
Action required theycallmeloki:vulkan-moe-density-gate
theycallmeloki:vulkan-moe-density-gate
Action required
CUDA : fix PAD for more than 65535 rows or slices EditorConfig Checker #61329: Pull request #30102 opened by pskrunner14
Action required pskrunner14:cuda-pad-large-grid
pskrunner14:cuda-pad-large-grid
Action required
llama: share the nextn tensor flags between models (#30097) EditorConfig Checker #61328: Commit 448147d pushed by ServeurpersoCom
8m 38s master
master
8m 38s
WIP: Persist checkpoints when saving hybrid model slots to disk EditorConfig Checker #61327: Pull request #30101 opened by rain-sk
-1s matarbot:pr/ckpt-persist
matarbot:pr/ckpt-persist
-1s
Add TML Inkling architecture EditorConfig Checker #61326: Pull request #25731 synchronize by danielhanchen
1m 58s danielhanchen:add-inkling
danielhanchen:add-inkling
1m 58s
mtmd: add cohere2 vision support EditorConfig Checker #61325: Pull request #30062 synchronize by Terrencezzj
1m 5s Terrencezzj:cohere2-vision
Terrencezzj:cohere2-vision
1m 5s
metal : few-row MMA mat-mul for the remaining src0 types (#30065) EditorConfig Checker #61324: Commit 9881906 pushed by ggerganov
3m 47s master
master
3m 47s