Lightning-AI / Lightning-AI/lightning-thunder
Consider if functions like full_like and expand_as should be symbols (or not)
- Dominant language
- Python
- Stars
- 1.5k
- Forks
- 121
- PR merge metrics
- No merged PRs in 30d
Description
Functions like `full_like` and `expand_as` are symbols that accept tensors, only to dump the tensor's metadata and call another option. We should consider making these non-symbols to simplify transform logic, clearly indicating that the original tensor doesn't need to be kept alive for the call to these operations.
cc @apaz-cli
Contributor guide
No contributing guide indexed for this repository
Research direction
Start by locating the definitions of `full_like` and `expand_as` and the transform logic that handles symbols. Determine whether these operations can be non-symbols while preserving their metadata behavior and clarifying that the source tensor need not remain alive. Done means the design decision is resolved and the affected transform behavior is updated consistently.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python, pytorch
- Domain
- compilers
- Issue type
- Refactor
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 30/100