CI on dataset loading
Open
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 1.2k
- Forks
- 349
- Avg merge
- 1d 21h
- Merged PRs (30d)
- 318
Description
For all Huggingface datasets, nightly tests for downloading them and ensuring they work. i.e. commands like this
config_paths="resources_servers/math_with_judge/configs/math_with_judge.yaml,\
responses_api_models/openai_model/configs/openai_model.yaml"
ng_prepare_data "+config_paths=[$config_paths]" \
+policy_model_name=dummy \
+policy_api_key=dummy \
+policy_base_url=dummy \
+output_dirpath=data/math_with_judge \
+mode=train_preparation \
+should_download=true \
+data_source=huggingface \
+hf_token=...
does not error for all configs registered with HF datasets.
One complexity is that there are some configs that need additional variables. we should just use pythong directly rather than a top level ng_prepare_data
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by reviewing the registered Hugging Face dataset configs and the Python entry point behind ng_prepare_data, then inspect existing CI or data-loading tests. Done means nightly tests cover all registered Hugging Face datasets, including configs requiring additional variables, and downloading them completes without errors.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- huggingface, python
- Domain
- ci-cd, data-engineering, testing
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100