minke.yu/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 12 Packages Projects Releases Wiki Activity

[CI] Fold duplicate-server suites and prune the retract matrix on 1-gpu-5090 (#33745)

Browse Source
This commit is contained in:
Liangsheng Yin
2026-08-05 12:41:51 -07:00
committed by GitHub
parent 36853b8ffc
commit de34dd11e9
11 changed files with 215 additions and 418 deletions
Download Patch File Download Diff File Expand all files Collapse all files
test/registered/model_loading/test_weight_cache_daemon.py
+1 -1
View File
@@ -27,7 +27,7 @@ DEFAULT_MODEL = "Qwen/Qwen3-0.6B"
# file per suite, TestWeightCacheDaemonTP2 self-skips when fewer than 2 GPUs are
# visible (i.e. on the 1-gpu runner).
register_cuda_ci(est_time=100, stage="extra-a", runner_config="2-gpu-large")
register_cuda_ci(est_time=100, stage="base-b", runner_config="1-gpu-small")
register_cuda_ci(est_time=45, stage="base-b", runner_config="1-gpu-small")
# Capture the client server's logs so test_loaded_via_ipc can assert the IPC
# load path actually ran (and did not silently fall back to disk).