Refactor codegen model to use @capture_outputs and @can_return_tuple decorators#44046
Closed
ManasVardhan wants to merge 1 commit into
Closed
Refactor codegen model to use @capture_outputs and @can_return_tuple decorators#44046ManasVardhan wants to merge 1 commit into
ManasVardhan wants to merge 1 commit into
Conversation
…decorators Replace manual output_attentions/output_hidden_states/return_dict handling in CodeGen model with the @capture_outputs and @can_return_tuple decorators, following the pattern established in llama and other refactored models. Changes: - Add _can_record_outputs dict to CodeGenPreTrainedModel mapping output types to their source modules (hidden_states -> CodeGenBlock, attentions -> CodeGenAttention) - Add @capture_outputs decorator to CodeGenModel.forward - Add @can_return_tuple decorator to CodeGenForCausalLM.forward - Remove manual output_attentions, output_hidden_states, return_dict parameter handling from forward methods - Simplify CodeGenBlock.forward to return only hidden_states (tensor) - Remove output_attentions parameter from CodeGenAttention.forward and CodeGenBlock.forward signatures Part of huggingface#43979
Contributor
|
[For maintainers] Suggested jobs to run (before merge) run-slow: codegen |
Author
|
Superseded by #44094 |
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What does this PR do?
Refactors the
codegenmodel to use the@capture_outputsand@can_return_tupledecorators, replacing manualoutput_attentions/output_hidden_states/return_dicthandling.Changes:
_can_record_outputsdict toCodeGenPreTrainedModelmapping output types to their source modules (hidden_states→CodeGenBlock,attentions→CodeGenAttention)@capture_outputsdecorator toCodeGenModel.forward@can_return_tupledecorator toCodeGenForCausalLM.forwardoutput_attentions,output_hidden_states,return_dictparameter handling from forward methodsCodeGenBlock.forwardto return onlyhidden_states(tensor)output_attentionsparameter fromCodeGenAttention.forwardandCodeGenBlock.forwardsignaturesAll 107 codegen model tests pass.
Fixes part of #43979