[FEATURE] Grouped convolutions for Conv1x1, WaveNet groups_1x1 hyperparameter
#184
+695
−50
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
- Added support for grouped convolutions in the Conv1x1 class by introducing a groups parameter in the constructor. - Implemented validation checks to ensure in_channels and out_channels are divisible by groups. - Updated weight initialization and processing methods to handle grouped convolution logic. - Enhanced the set_weights_ and process methods to accommodate grouped weight organization and processing. - Added necessary error handling for invalid group configurations. This update enhances the flexibility of the Conv1x1 layer for various convolutional architectures.
- Included new test cases for the Conv1x1 layer in run_tests.cpp, covering various construction and processing scenarios. - Tests validate functionality such as construction with groups, input/output channel validation, and processing with and without bias. - Enhanced test coverage for grouped convolutions, ensuring comprehensive validation of the Conv1x1 layer's capabilities.
- Refactored test cases in test_conv_1x1.cpp to use named constants for input/output channels and bias parameters, improving readability. - Ensured all assertions for output dimensions are based on the defined constants, enhancing maintainability. - Updated tests to validate behavior for both basic and grouped convolution scenarios, ensuring comprehensive coverage of the Conv1x1 functionality.
…rouping parameters - Updated the WaveNet _Layer and _LayerArray constructors to accept new parameters: groups_condition and groups_1x1, enhancing flexibility for grouped convolutions. - Modified relevant test cases in test_layer, test_layer_array, and test_full to incorporate the new parameters, ensuring comprehensive testing of the updated functionality. - Ensured backward compatibility by maintaining existing parameter structures while adding new ones for improved configurability.
… grouping parameters - Removed the groups_condition parameter from the constructors of _Layer and _LayerArray, streamlining the interface for grouped convolutions. - Updated relevant method calls and test cases to reflect this change, ensuring consistency and clarity in the codebase. - Maintained backward compatibility by defaulting groups_1x1 to 1 where applicable.
…ps_condition parameter - Removed the groups_condition parameter from various test cases in test_layer, test_layer_array, test_full, and test_real_time_safe. - Updated constructor calls for _Layer and _LayerArray to reflect this change, ensuring consistency across tests. - Enhanced code clarity by streamlining the parameter list in relevant test functions.
- Increased groups_1x1 from 1 to 2 in test_real_time_safe.cpp to support grouped processing. - Enhanced comments to clarify weight initialization for grouped convolutions, detailing the identity matrix setup for each group. - Ensured consistency in the test setup for grouped convolution scenarios, improving clarity and maintainability.
sdatkinson
changed the title
Conv1x1Conv1x1, WaveNet groups_1x1 hyperparameter
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.
Resolves #172
Also add
groups_1x1for WaveNet layers