Train speech language ID classification head #450
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
facebook-github-bot
added
the
CLA Signed
zrthxn
reviewed
May 10, 2024
src/seamless_communication/cli/m4t/classification_head/model.py
Outdated
Show resolved
Hide resolved
zrthxn
reviewed
May 10, 2024
|
|
||
| grad_scaler = torch.cuda.amp.GradScaler() | ||
| optimizer = AdamW( | ||
| params=frozen_model.parameters(), |
There was a problem hiding this comment.
We want to optimize the head not the frozen model
* Dataloader * Training loop fixes * One Hot encode language labels * Remove `dist_utils` * Remove duplicate option * remove label smoothing * Device class * Batching * Add float_dtype * Add padding mask * Add padding mask * Device * Model shape * Optimize head
* Small fixes * Argyment * Batch size * Model shape * unsqueeze * Long * Float tensor * Float tensor * DEvice * While set * src_lengths * Refactor
* Dataloader * Training loop fixes * One Hot encode language labels * Remove `dist_utils` * Remove duplicate option * remove label smoothing * Device class * Batching * Add float_dtype * Add padding mask * Add padding mask * Device * Model shape * Optimize head
* Small fixes * Argyment * Batch size * Model shape * unsqueeze * Long * Float tensor * Float tensor * DEvice * While set * src_lengths * Refactor
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Outdated suggestions cannot be applied.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.
Implements a new training script for training a classification head. The new nn.module in model.py is separate from other components of M4T and takes speech encoder output and maps it into language probabilities with softmax. All layers of M4T are frozen to prevent training of those components. A new dataset.py script is implemented for downloading common voice from hugging face.
Most recent results from training:
Samples in train dataset:
Samples in eval dataset:
Parameters:
I am collaborating with @zrthxn on this project