Lightning-AI / Lightning-AI/lightning-thunder
test_grad.py assumes all ops in OpInfo database can be run on CPU
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 1.5k
- Forks
- 121
- PR merge metrics
- No merged PRs in 30d
Description
`_generate_supported_op_list` tries to run all ops in the opinfo db to filter ops for grad test.
However, `Tensor.cuda` requires CUDA runtime to be available. We have to add this op to the skip list (even when it should work when CUDA runtime is available).
See : https://github.com/Lightning-AI/lightning-thunder/pull/181#issuecomment-2057250507
cc @apaz-cli
Contributor guide
No contributing guide indexed for this repository
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 in test_grad.py at _generate_supported_op_list, which probes operations from the OpInfo database for gradient tests. Check the existing skip list and run the relevant grad test; done means Tensor.cuda is excluded from CPU probing while the remaining operations are still filtered correctly.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python, pytorch
- Domain
- testing-qa
- Issue type
- Bug
- Difficulty
- 2/5
- Estimated time
- 1-3 hours
- Activity status
- Stale
- Clarity
- Clearly specified
- Newbie friendliness
- 58/100