Skip to content

Actions: pytorch/executorch

Actions

cuda-perf

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
cuda-perf.yml
1,882 workflow runs
1,882 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
[ExecuTorch][llm] Wire --use_moe_quantized_op through export pipeline… cuda-perf #1885: Commit 9789eea pushed by digantdesai
42s main
main
42s
[ExecuTorch][llm] Add MoE source transformation with INT4 quantized p… cuda-perf #1884: Commit afffe4f pushed by digantdesai
34s main
main
34s
[ExecuTorch][llm] Add Meta kernel and registration probe for quantize… cuda-perf #1883: Commit 1a6b1c7 pushed by digantdesai
42s main
main
42s
[ExecuTorch][llm] Add quantized_moe_ffn custom op with C++ kernel (#2… cuda-perf #1882: Commit b623687 pushed by digantdesai
1h 43m 30s main
main
1h 43m 30s
backends/mlx: runtime MoE expert-sort for decode (issue #20554) (#20685) cuda-perf #1881: Commit fdb88c3 pushed by metascroy
39s main
main
39s
Bump torchao to 03ca489dc and nightly wheel to dev20260729 (#21476) cuda-perf #1880: Commit 097b9aa pushed by rascani
40s main
main
40s
[MLX] Grow the KV-cache pool on demand (#21525) cuda-perf #1879: Commit 849f9d8 pushed by kiymetakdemir
41s main
main
41s
[ET-VK][ops] Add buffer unfold support cuda-perf #1878: Commit e4f6748 pushed by SS-JIA
1h 20m 24s main
main
1h 20m 24s
[MLX] Add off-graph KV-cache flat runtime (#21501) cuda-perf #1877: Commit bc13a7b pushed by kiymetakdemir
1m 30s main
main
1m 30s
migrate backends/vulkan/tools/gpuinfo (TARGETS+BUCK with similar rule… cuda-perf #1876: Commit e0c610a pushed by meta-codesync Bot
4h 5m 50s main
main
4h 5m 50s
migrate backends/vulkan/test (fbcode python_unittest + xplat fb_nativ… cuda-perf #1875: Commit 644990e pushed by meta-codesync Bot
49s main
main
49s
drop empty TARGETS in backends/mediatek/runtime (#21430) cuda-perf #1874: Commit cdc5d28 pushed by meta-codesync Bot
50s main
main
50s
[ExecuTorch][Core AI][0/x] Dev tooling: test runner + code-writing sk… cuda-perf #1873: Commit f516236 pushed by metascroy
1m 9s main
main
1m 9s
migrate backends/apple/coreml (TARGETS+BUCK both with rules) (#21429) cuda-perf #1872: Commit 9c3dfd9 pushed by meta-codesync Bot
1h 14m 40s main
main
1h 14m 40s
[executorch][cuda] Optimize short-query INT5 matvec kernels (#21505) cuda-perf #1871: Commit 250020b pushed by Gasoonjia
41s main
main
41s
NXP backend: Enable Hardswish with new Neutron flow (#21500) cuda-perf #1870: Commit 86c353e pushed by roman-janik-nxp
1m 36s main
main
1m 36s
Move targets.bzl to proper location (#21495) cuda-perf #1869: Commit d8c6be2 pushed by bigfootjon
42s main
main
42s
NXP backend: added support for aten.rsqrt using new Neutron MLIR fl… cuda-perf #1868: Commit 6683757 pushed by novak-vaclav
1h 19m 41s main
main
1h 19m 41s
Add pass to replace some ops with channels last variants. (#20788) cuda-perf #1867: Commit 30f9775 pushed by MartinPavella
40s main
main
40s
Arm backend: Give SiLU its own output quantization params (#21437) cuda-perf #1866: Commit 761c64c pushed by Erik-Lundell
41s main
main
41s
Support bf16 delegation for fully-connected (#21493) cuda-perf #1865: Commit d29706a pushed by meta-codesync Bot
40s main
main
40s
Support grouped state heads in GDN attention (#21402) cuda-perf #1864: Commit c2221eb pushed by meta-codesync Bot
1h 29m 49s main
main
1h 29m 49s
Fix XNNPACK pooling crashing on default stride and single-element ker… cuda-perf #1863: Commit f287863 pushed by JakeStevens
40s main
main
40s
fix(mlx): forward boundary-partitioned getitem instead of indexing a … cuda-perf #1862: Commit 43cb2b2 pushed by metascroy
36s main
main
36s
[executorch][cuda] Optimize short-query INT6 matvec kernels (#21504) cuda-perf #1861: Commit 6898c99 pushed by Gasoonjia
4m 6s main
main
4m 6s