storytold/cloud-worker
2
0
Fork 0
mirror of https://github.com/storytold/cloud-worker.git synced 2026-10-18 07:33:13 +00:00
Code Issues Packages Projects Releases Wiki Activity

bench: survive server loss during polling (mark run failed, keep suite alive)

Browse Source 
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This commit is contained in:
Hanashi
2026-08-06 03:48:37 -04:00
parent 427f2878c4
commit 53b4a133e2
1 changed files with 13 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
bench/minimax_bench.py
+13 -1
View File
@@ -196,9 +196,21 @@ def run_one(api, cfg, poll=2.0, timeout=3600):
prompt_id = resp["prompt_id"] prompt_id = resp["prompt_id"]
exec_start = exec_end = None exec_start = exec_end = None
status_str = None status_str = None
poll_failures = 0
while time.time() - t0 < timeout: while time.time() - t0 < timeout:
time.sleep(poll) time.sleep(poll)
h = api.history(prompt_id) try:
h = api.history(prompt_id)
poll_failures = 0
except Exception as e:
poll_failures += 1
if poll_failures >= 10:
stop.set()
return {**cfg, "ok": False, "status": "server_lost",
"wall_s": round(time.time() - t0, 1), "exec_s": None,
"peak_vram_gb": round(peak["vram"] / 1e9, 1) if peak["vram"] else None,
"outputs": [], "error": f"server unreachable while polling: {e}"}
continue
if prompt_id in h: if prompt_id in h:
entry = h[prompt_id] entry = h[prompt_id]
status_str = entry.get("status", {}).get("status_str") status_str = entry.get("status", {}).get("status_str")