Allow users to choose devices for models #264
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
This addresses a problem that pops up when you've got several GPUs in play. When you set the device-map to 'auto', the model loads onto multiple GPUs just fine. But then, when you try to do 'model.to(device)', it tries to cram everything onto just one GPU. |
|
I've updated this branch with some formatting fixes and test/docstring changes from @rlouf. |
| @@ -114,7 +114,7 @@ def test_transformers_integration_choice(): | |||
|
|
|||
| def test_transformers_integration_with_pad_token(): | |||
| model_name = "hf-internal-testing/tiny-random-XLMRobertaXLForCausalLM" | |||
| model = models.transformers(model_name, device="cpu") | |||
| model = models.transformers(model_name, device="meta") | |||
There was a problem hiding this comment.
This was necessary in order to get around some "meta device" issues related to accelerate.
brandonwillard
changed the title
brandonwillard
added
enhancement
transformers
brandonwillard
approved these changes
Sep 7, 2023
|
Thank you @BramVanroy ! |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
Currently, there are some issues when loading a transformers model because it explicitly calls
.to(device). The default is "cpu" (which is already the default in transformers to begin with). So I recommend to removedeviceas an argument completely and let the user decide how they want to load the model. That allows things like device_map and loading inkbit. Internally inside theTransformersclass nothing changes, but we setself.devicetomodel.deviceto ensure that everything stays compatible in the rest of the library.closes #238