Skip to content

[GFX1250] MiniMax-M3 gfx1250 enabling #3377

[GFX1250] MiniMax-M3 gfx1250 enabling

[GFX1250] MiniMax-M3 gfx1250 enabling #3377

Triggered via pull request June 30, 2026 11:13
Status Failure
Total duration 3h 56m 12s
Artifacts 4

atom-sglang-test.yaml

on: pull_request
Check Pre Checkin Signal
1m 36s
Check Pre Checkin Signal
Download aiter wheel
57s
Download aiter wheel
Matrix: atom-sglang-accuracy
Fit to window
Zoom out
Zoom in

Annotations

1 error
Accuracy (DeepSeek-V4-Pro TP8)
The self-hosted runner lost communication with the server. Verify the machine is running and has a healthy network connection. Anything in your workflow that terminates the runner process, starves it for CPU/Memory, or blocks its network access can cause this error.

Artifacts

Produced during runtime
Name Size Digest
aiter-whl Expired
557 MB
sha256:c7f2d9435a57ab6080f610600647a7c28b1f6c4974e1dd58a5c5246223f2dbc6
sglang-DeepSeek-R1-FP4 TP4-artifacts Expired
83.5 KB
sha256:72ccd328ab0313071044e39c4302dcf2b350eed772a3a2854204f814657c9a9b
sglang-DeepSeek-R1-FP8 TP4-artifacts Expired
65.8 KB
sha256:c0537c00f333c10fa005508e95980c56d0139e2270f63ae54c38b4466cea7e5f
sglang-Qwen3.5-35B-A3B-FP8 TP2-artifacts Expired
67.9 KB
sha256:c5a407c13f7ea14807486a7575152377693fae99ca7783ebc47ed17480a10e28