Makes RetrievalModelV2 support item tower with transforms (e.g. pre-trained embeddings) #1198
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
Documentation preview |
…atch_predict() support Loader with transforms for pre-trained embeddings in item tower
…upport Loader with transforms (which include pre-trained embeddings)
gabrielspmoreira
force-pushed
the
tf/pretrained_emb_item_tower
branch
from
ea1db75 to
d4dc945
Compare
…y and item embeddings
|
Check out this pull request on See visual diffs & provide feedback on Jupyter Notebooks. Powered by ReviewNB |
|
@oliverholworthy hello. do you think you can review this PR? this is an important feature that'd be needed by users and us. thanks. |
sararb
approved these changes
Jul 20, 2023
| @@ -1553,7 +1553,7 @@ def predict( | |||
| return out | |||
|
|
|||
| def batch_predict( | |||
| self, dataset: merlin.io.Dataset, batch_size: int, **kwargs | |||
| self, dataset: Union[merlin.io.Dataset, Loader], batch_size: int, **kwargs | |||
There was a problem hiding this comment.
Can you update the docstrings to expose to the user that we are supporting Loader as a possible input dataset
oliverholworthy
approved these changes
Jul 20, 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Outdated suggestions cannot be applied.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.
Goals ⚽
batch_predict()support only aDatasetas input. That prevents performingbatch_predict()using aLoader, which might contain transforms that are important to the modelRetrievalModelV2, when among item features there are pre-trained embeddings loaded withEmbeddingOperatortransform.Loaderwithtransformsto allbatch_predict()based methods (e.g.to_top_k_encoder().Implementation Details 🚧
batch_predict()method and associated methods were changed to acceptLoaderbesidesDataset. When aLoaderis passed, it cascades any transforms (like theEmbeddingOperator) to theLoaderthat is created inside thedata_iterator_func.Testing Details 🔍
test_two_tower_model_topk_evaluation_with_pretrained_embto ensure that it is possible to have pre-trained embeddings in the item tower and convert theretrieval_model.to_top_k_encoder().