Fix dataset loading schema validation issue in CI#304

Open

juanmichelini wants to merge 32 commits intomainfrom

fix-dataset-schema-validation

Commits on Jan 13, 2026

Fix dataset loading schema validation issue in CI
juanmichelini
committed
Add trust_remote_code=True to dataset loading
juanmichelini
committed
Add manual dataset loading fallback for schema validation issues
juanmichelini
committed
Fix UnboundLocalError for os import
juanmichelini
committed
fix: correct exception handling in dataset loading retry logic

juanmichelini
and
openhands-agent
committed
fix: load parquet files directly to bypass schema validation issues

juanmichelini
and
openhands-agent
committed
fix: use streaming dataset to bypass schema validation

juanmichelini
and
openhands-agent
committed
fix: fallback to 'train' split when requested split doesn't exist

juanmichelini
and
openhands-agent
committed
fix: use 'number' field instead of 'version' in Multi-SWE-Bench dataset

juanmichelini
and
openhands-agent
committed
fix: lowercase Docker repository names for Multi-SWE-Bench

juanmichelini
and
openhands-agent
committed
Allow up to 5 image build failures in multiswebench build workflow
juanmichelini
committed
fix: specify --output-dir builds in multiswebench build workflow
juanmichelini
committed
feat: add real-time build progress logging
juanmichelini
committed
fix: respect n-limit parameter in multiswebench build_images.py
juanmichelini
committed
fix: use get_dataset() utility in multiswebench run_infer.py
juanmichelini
committed

Commits on Jan 14, 2026

Commits on Jan 15, 2026

Commits on Jan 16, 2026

Commits on Jan 17, 2026

Fix: Reset repo to base_commit to prevent including PR changes in patch
juanmichelini
committed

Commits on Jan 18, 2026

Commits on Jan 19, 2026