deepspeedai / deepspeedai/DeepSpeed
[REQUEST] When training a FP16 model, the ability to set some of the layers to FP32
Open
Nobody has claimed this yet.
enhancement
- Dominant language
- Python
- Stars
- 43.1k
- Forks
- 5k
- Avg merge
- 4d 15h
- Merged PRs (30d)
- 112
Description
When training a FP16 model, I wonder if it's possible to set some of the layers to FP32.
I can add .to(torch.float16) and .to(torch.float32) to do the conversion between layers.
So the process will be like:
...
x = f(x) # x and f(x) are float16
x = x.to(torch.float32)
x = g(x) # x and g(x) are float32
x = x.to(torch.float16)
x = f(x) # x and f(x) are float16
...
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
No file, test, or entry point is named. Start by locating the FP16 training path and determine how selected layers could run in FP32 while surrounding tensors are converted as shown; done means a defined way to choose FP32 layers and verify the example behavior.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python, pytorch
- Domain
- machine-learning
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100