Skip to content

fix(eval): retry empty checkpoint rows instead of resuming them as done #65

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
ronaldtse merged 1 commit into from
Aug 31, 2026
Merged

fix(eval): retry empty checkpoint rows instead of resuming them as done #65

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 9 additions & 0 deletions eval_sadeed_glm.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -111,6 +111,15 @@ def main() -> None:
for line in CKPT.read_text(encoding="utf-8").splitlines():
row = json.loads(line)
done[row["idx"]] = row["pred"]
# exhausted retries checkpoint as empty strings; resuming them
# as "done" silently scores them as all-error rows (the 2026-
# 08-31 GLM-5.3-Flash run read 15.96 DER with 140 such rows)
empties = [i for i, pred in done.items() if not pred.strip()]
for i in empties:
del done[i]
if empties:
print(f"[glm] dropping {len(empties)} empty checkpoint rows "
"for retry", flush=True)
todo = [i for i in range(len(inputs)) if i not in done]
print(f"[glm] model={MODEL} effort={EFFORT or 'thinking-disabled'} "
f"done={len(done)} todo={len(todo)}", flush=True)
Expand Down