deepspeedai / deepspeedai/DeepSpeed
What does partition_activations mean?
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 43.1k
- Forks
- 5k
- Avg merge
- 4d 15h
- Merged PRs (30d)
- 112
Description
I am confused what deepspeed will do with 'partition_activations: true' in the 'activation_checkpointing' section of the config file.
Also, from the source code it seems that deepspeed will flatten all the input tensors except the last one when partition_activations is enabled. Does it mean that modules should manually resize the input tensors in the forward function?
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 with the activation_checkpointing configuration and the source code handling partition_activations. Trace how input tensors are flattened and passed through module forward functions, then document what the option does and whether callers must resize inputs.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python, pytorch
- Domain
- distributed-systems, documentation, machine-learning
- Issue type
- Documentation
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 30/100