Skip to content

Actions: pytorch/executorch

Actions

Test Metal Backend

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
metal.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
NXP backend: Unify testing infrastructure. (#19112) Test Metal Backend #11334: Commit eb19f24 pushed by MartinPavella
1h 33m 42s main
main
1h 33m 42s
lint Test Metal Backend #11333: Commit ef8aeb2 pushed by pytorch-bot Bot
1h 36m 39s ciflow/metal/19205
ciflow/metal/19205
1h 36m 39s
Enable Qwen3.5-MoE AOTI export on memory-constrained GPUs Test Metal Backend #11332: Pull request #19205 opened by Gasoonjia
1h 33m 30s cuda-aoti-low-mem-export
cuda-aoti-low-mem-export
1h 33m 30s
Merge branch 'main' into moe-no-float Test Metal Backend #11331: Commit 4e5c02d pushed by pytorch-bot Bot
1h 33m 34s ciflow/metal/19170
ciflow/metal/19170
1h 33m 34s
Make the output of MoE forward method have expected output in non cuda backends Test Metal Backend #11330: Pull request #19170 synchronize by Gasoonjia
1h 33m 33s moe-no-float
moe-no-float
1h 33m 33s
Qualcomm AI Engine Direct - Update Llama 1B Test PPL Score (#19202) Test Metal Backend #11329: Commit 4ac044b pushed by abhinaykukkadapu
22m 47s main
main
22m 47s
Add Triton INT4 dense kernels with dequant prefill path for Qwen3.5 MoE Test Metal Backend #11328: Pull request #19188 synchronize by digantdesai
29m 13s gh/digantdesai/51/head
gh/digantdesai/51/head
29m 13s
Add W4A8 INT8 activation kernels for batched MoE prefill Test Metal Backend #11327: Pull request #19187 synchronize by digantdesai
24m 6s gh/digantdesai/50/head
gh/digantdesai/50/head
24m 6s
Add structured stats reporting and GPU memory tracking to Qwen3.5 MoE runner Test Metal Backend #11326: Pull request #19190 synchronize by digantdesai
21m 38s gh/digantdesai/53/head
gh/digantdesai/53/head
21m 38s
Add libextension_memory_allocator.a to executorch_llm Apple framework… Test Metal Backend #11325: Commit 5a206ab pushed by rascani
25m 20s main
main
25m 20s
Factor Cortex-M FVP e2e job into a reusable workflow and add nightly … Test Metal Backend #11324: Commit 321c029 pushed by rascani
1h 46m 22s main
main
1h 46m 22s
lint Test Metal Backend #11323: Commit 1046aba pushed by pytorch-bot Bot
1h 45m 13s ciflow/metal/19170
ciflow/metal/19170
1h 45m 13s
Make the output of MoE forward method have expected output in non cuda backends Test Metal Backend #11322: Pull request #19170 synchronize by Gasoonjia
1h 38m 24s moe-no-float
moe-no-float
1h 38m 24s
Add W4A8 INT8 activation kernels for batched MoE prefill Test Metal Backend #11321: Pull request #19187 synchronize by digantdesai
1h 56m 6s gh/digantdesai/50/head
gh/digantdesai/50/head
1h 56m 6s
Add structured stats reporting and GPU memory tracking to Qwen3.5 MoE runner Test Metal Backend #11320: Pull request #19190 synchronize by digantdesai
1h 24m 4s gh/digantdesai/53/head
gh/digantdesai/53/head
1h 24m 4s
Add Triton INT4 dense kernels with dequant prefill path for Qwen3.5 MoE Test Metal Backend #11319: Pull request #19188 synchronize by digantdesai
1h 40m 48s gh/digantdesai/51/head
gh/digantdesai/51/head
1h 40m 48s
Qwen 3.5 MoE: bypass on-device sampler conditional in non-CUDA export Test Metal Backend #11318: Commit d8454cf pushed by pytorch-bot Bot
1h 21m 59s ciflow/metal/19170
ciflow/metal/19170
1h 21m 59s
Make the output of MoE forward method have expected output in non cuda backends Test Metal Backend #11317: Pull request #19170 synchronize by Gasoonjia
9m 51s moe-no-float
moe-no-float
9m 51s
Qwen 3.5 MoE: bypass on-device sampler conditional in non-CUDA export Test Metal Backend #11316: Commit b41d055 pushed by pytorch-bot Bot
2h 8m 35s ciflow/metal/19170
ciflow/metal/19170
2h 8m 35s
Make the output of MoE forward method have expected output in non cuda backends Test Metal Backend #11315: Pull request #19170 synchronize by Gasoonjia
5m 31s moe-no-float
moe-no-float
5m 31s
Add W4A8 INT8 activation kernels for batched MoE prefill Test Metal Backend #11314: Pull request #19187 synchronize by digantdesai
12m 20s gh/digantdesai/50/head
gh/digantdesai/50/head
12m 20s
Add structured stats reporting and GPU memory tracking to Qwen3.5 MoE runner Test Metal Backend #11313: Pull request #19190 synchronize by digantdesai
11m 57s gh/digantdesai/53/head
gh/digantdesai/53/head
11m 57s
Add Triton INT4 dense kernels with dequant prefill path for Qwen3.5 MoE Test Metal Backend #11312: Pull request #19188 synchronize by digantdesai
16m 38s gh/digantdesai/51/head
gh/digantdesai/51/head
16m 38s
Add structured stats reporting and GPU memory tracking to Qwen3.5 MoE runner Test Metal Backend #11311: Pull request #19190 synchronize by digantdesai
5m 24s gh/digantdesai/53/head
gh/digantdesai/53/head
5m 24s
Add Gemma 4 E2B/E4B multimodal model, export, and runner for ExecuTor… Test Metal Backend #11310: Commit 06d6943 pushed by meta-codesync Bot
2h 1m 41s main
main
2h 1m 41s