Update Low Precision LayerNorm arguments #1994
Merged
Conversation
This file contains hidden or 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
mvpatel2000
reviewed
Feb 28, 2023
composer/algorithms/low_precision_layernorm/low_precision_layernorm.py
Outdated
Show resolved
Hide resolved
…rnorm.py Change assertion to if-Raise Co-authored-by: Mihir Patel <mihir.v.patel7@gmail.com>
mvpatel2000
approved these changes
Feb 28, 2023
dskhudia
approved these changes
Feb 28, 2023
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Outdated suggestions cannot be applied.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.
This PR changes the Low Precision LayerNorm (LPLN) arguments to match the torch.nn.LayerNorm arguments, so a user can manually replace a LayerNorm layer with an LPLN layer. Please note that using
apply_low_precision_layernorm(model)will replace all torch.nn.LayerNorms for you, so it is not necessary to use the LPLayerNorm class directly.
On exporting models with LPLN:
LowPrecision LayerNorm works with ONNX exporting, but it does not support Torchscript. torch.get_autocast_gpu_dtype() is not supported by Torchscript.
We change the test
test_gpu_huggingface_export_for_inference_onnxto use LPLN instead of Fused LayerNorm to verify export functionality.