deepspeedai / deepspeedai/DeepSpeed
hidden_dim constraint in transformer cuda kernel
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 43.1k
- Forks
- 5k
- Avg merge
- 4d 15h
- Merged PRs (30d)
- 112
Description
I found that there is constraint on the dimensionality when we use the transformer cuda kernel: https://github.com/microsoft/DeepSpeed/blob/d720fdb6857f4b71d922ca1e8efbe5271b5fb7c2/csrc/transformer/normalize_kernels.cu#L232-L250
I wonder what is the reason behind it? Is there any plan to support arbitrary dimensionality? Or, If I want to use hidden_dim=4096 or 8192, what do I need to do to make it work? Thanks.
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 inspecting csrc/transformer/normalize_kernels.cu at lines 232-250, which the issue identifies as the source of the hidden_dim constraint. Determine the reason for the restriction and what would be needed to support hidden_dim values of 4096 or 8192; the issue currently has no defined patch, test, or completion criterion.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- cpp
- Domain
- machine-learning, performance
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 20/100