good first issue
仓库指标
- 星标
- (1 个星标)
- PR 合并指标
- (PR 指标待抓取)
描述
Summary
Add a unit test to run the trainer with a very small config using a tiny model, ensuring that model loading, dataset preprocessing, and trainer setup succeed.
Acceptance Criteria
- A new test file
tests/test_trainer_small_config.pythat:- Uses
sshleifer/tiny-gpt2 - Builds a minimal JSON config
- Calls
CodeExplainerTrainer.load_model(),load_dataset(None),preprocess_dataset(), andsetup_trainer()
- Uses
- Test should run quickly (< 60s) and not perform full training.
References
- Existing tests in
tests/test_trainer.py - Tiny models used in CI