As a user of prompt tuning, I want to be able to leverage multiple GPUs at train time!
Discussion
Extends https://github.com/caikit/caikit-nlp/issues/175 to leverage PyTorch FSDP; there may be some refactoring to consolidate with fine tuning FSDP configuration / dataset stuff, which may be part of this story, or left to a follow up. We will decide once we see how Peft / trainer integration goes.
Description
As a user of prompt tuning, I want to be able to leverage multiple GPUs at train time!
Discussion
Extends https://github.com/caikit/caikit-nlp/issues/175 to leverage PyTorch FSDP; there may be some refactoring to consolidate with fine tuning FSDP configuration / dataset stuff, which may be part of this story, or left to a follow up. We will decide once we see how Peft / trainer integration goes.
Acceptance Criteria