Add low cpu mem support for LLM instantiation #3460
Closed
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Reduces memory overhead when LLMs are loaded onto CPU memory but enabling
low_cpu_mem_usage
that loads the model using ~1x model size CPU memory. See https://huggingface.co/docs/transformers/main_classes/model#transformers.PreTrainedModel.from_pretrained for more info.This also enables half-precision loading by default, which seems like a reasonable thing to do. It uses float16 since that always works, unlike bf16 which has hardware restrictions.