Model Export to liteRT #2405
Model Export to liteRT #2405
Conversation
This reverts commit 62d2484.
This reverts commit de830b1.
Refactored exporter and registry logic for better type safety and error handling. Improved input signature methods in config classes by extracting sequence length logic. Enhanced LiteRT exporter with clearer verbose handling and stricter error reporting. Registry now conditionally registers LiteRT exporter and extends export method only if dependencies are available.
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Summary of Changes
Hello @pctablet505, I'm Gemini Code Assist1! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!
This pull request introduces a comprehensive and extensible framework for exporting Keras-Hub models to various formats, with an initial focus on LiteRT. The system is designed to seamlessly integrate with Keras-Hub's model architecture, particularly by addressing the unique challenge of handling dictionary-based model inputs during the export process. This enhancement significantly improves the deployability of Keras-Hub models by providing a standardized and robust export pipeline, alongside crucial compatibility fixes for TensorFlow's SavedModel/TFLite export mechanisms.
Highlights
- New Model Export Framework: Introduced a new, extensible framework for exporting Keras-Hub models, designed to support various formats and model types.
- LiteRT Export Support: Added specific support for exporting Keras-Hub models to the LiteRT format, verified for models like gemma3, llama3.2, and gpt2.
- Registry-Based Configuration: Implemented an
ExporterRegistryto manage and retrieve appropriate exporter configurations and exporters based on model type and target format. - Input Handling for Keras-Hub Models: Developed a
KerasHubModelWrapperto seamlessly convert Keras-Hub's dictionary-based inputs to the list-based inputs expected by the underlying Keras LiteRT exporter. - TensorFlow Export Compatibility: Added compatibility shims (
_get_save_specand_trackable_children) to Keras-HubBackbonemodels to ensure proper functioning with TensorFlow's SavedModel and TFLite export utilities. - Automated Export Method Extension: The
Taskclass in Keras-Hub models is now automatically extended with anexportmethod, simplifying the model export process for users.
Using Gemini Code Assist
The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.
Invoking Gemini
You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.
| Feature | Command | Description |
|---|---|---|
| Code Review | /gemini review |
Performs a code review for the current pull request in its current state. |
| Pull Request Summary | /gemini summary |
Provides a summary of the current pull request in its current state. |
| Comment | @gemini-code-assist | Responds in comments when explicitly tagged, both in pull request comments and review comments. |
| Help | /gemini help |
Displays a list of available commands. |
Customization
To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.
Limitations & Feedback
Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here.
You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.
Footnotes
-
Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩
There was a problem hiding this comment.
Code Review
This pull request introduces a significant new feature: model exporting to liteRT. The implementation is well-structured, using a modular and extensible registry pattern. However, there are several areas that require attention. The most critical issue is the complete absence of tests for the new export functionality, which is a direct violation of the repository's style guide stating that testing is non-negotiable. Additionally, I've identified a critical bug in the error handling logic within the lite_rt.py exporter that includes unreachable code. There are also several violations of the style guide regarding the use of type hints in function signatures across all new files. I've provided specific comments and suggestions to address these points, which should help improve the robustness, maintainability, and compliance of this new feature.
Introduces the keras_hub.api.export submodule and updates the main API to expose it. The new export module imports various exporter configs and functions from the internal export package, making them available through the public API.
Added ImageClassifierExporterConfig, ImageSegmenterExporterConfig, and ObjectDetectorExporterConfig to the export API. Improved input shape inference and dummy input generation for image-related exporter configs. Refactored LiteRTExporter to better handle model type checks and input signature logic, with improved error handling for input mapping.
Moved the 'import keras' statement to the top of the module and removed redundant local imports within class methods. This improves code clarity and avoids repeated imports.
Deleted the debug_object_detection.py script, which was used for testing object detection model outputs and export issues. This cleanup removes unused debugging code from the repository.
Renames all references of 'LiteRT' to 'Litert' across the codebase, including file names, class names, and function names. Updates exporter registry and API imports to use the new 'litert' naming. Also improves image model exporter configs to dynamically determine input dtype from the model, enhancing flexibility for different input types. Adds support for ImageSegmenter model type detection in the exporter registry.
Refactored InputSpec definitions in exporter configs for improved readability by placing each argument on a separate line. Updated import path in litert.py to import from keras.src.export.litert instead of keras.src.export.litert_exporter.
|
@pctablet505 can you update the Colab to use the changes from this PR? and keep the demo short where you load a model and export it, and reload and verify numerics? |
|
Also if you have addressed the Gemini comments, please respond on it. Is this PR dependent on the PR in Keras repo? - keras-team/keras#21674 |
Simplifies and unifies input signature and dummy input generation for text and image models by removing redundant helper methods and centralizing logic. Updates model building in KerasHubExporter to use input signatures and improves error handling. Refactors LiteRT exporter to use the new parameterized input signature and model building approach, reducing code duplication and improving maintainability.
Removed redundant registry initialization in export_model and clarified model building comments in KerasHubExporter. Switched to using warnings.warn for import errors in models/__init__.py instead of print statements for better error reporting.
|
@gemini-code-assist please review it again |
There was a problem hiding this comment.
Code Review
This is a significant and valuable addition to KerasHub, enabling model exportation to LiteRT. The overall structure is well-designed, following a modular and extensible pattern similar to Hugging Face's Optimum.
I've reviewed this pull request, focusing on correctness, maintainability, and adherence to the repository's style guide. The style guide was instrumental in this review, particularly the sections on Testing and Validation (lines 401-496) and Docstrings and Type Hinting (lines 364-398).
My main findings are:
-
Critical Issue: Missing Tests
- The pull request introduces substantial new functionality for model exporting across several new files (
base.py,configs.py,litert.py,registry.py), but it lacks corresponding tests. - The repository style guide is explicit that "Testing is a non-negotiable part of every contribution" (line 403) and "Every .py file containing logic...must have a corresponding
_test.pyfile" (line 406). - Please add comprehensive unit tests for the new export logic, covering different model types, configurations, and edge cases. This is crucial to ensure the robustness and correctness of this feature.
- The pull request introduces substantial new functionality for model exporting across several new files (
-
Other Findings
- I've also left several inline comments regarding a bug in model type detection, incorrect dtype handling, and violations of the docstring style guide. Please address these to improve code quality and consistency.
Refined dtype extraction logic in image and object model exporter configs to better handle different dtype representations. Updated LiteRT exporter to use Keras io_utils for progress messages and improved verbose flag handling. Added ObjectDetector and ImageSegmenter to export registry model type checks. Enhanced docstrings for clarity and consistency in base exporter classes.
Changed exporter registry and config registration to use model classes instead of string type names for improved type safety and clarity. Updated input signature methods to use isinstance checks and standardized padding_mask dtype to int32. Enhanced LiteRTExporter to dynamically determine input signature parameters based on model type and preprocessor attributes.
Replaces the try-except block for importing keras with a direct import, assuming keras is always available. Simplifies the code and removes the KERAS_AVAILABLE flag.
Introduce new test modules for export base classes, configuration classes, LiteRT export functionality, registry logic, and production model export verification. Also update TensorFlow CUDA requirements to include ai-edge-litert for LiteRT export support.
Replaces the previous wrapper with type-specific adapter classes for text and image models in the LiteRT exporter, improving input conversion logic and maintainability. Also updates docstrings and return type annotations for consistency across exporter config classes.
There was a problem hiding this comment.
Code Review
This pull request introduces a robust and well-structured framework for exporting Keras Hub models to the LiteRT format. The design, which utilizes a registry, exporter configurations, and model-specific adapters, is commendable for its extensibility and clear separation of concerns. The addition of comprehensive unit and integration tests, including numerical verification for production models, significantly increases confidence in this new feature.
My review focuses on improving maintainability by reducing code duplication in the configuration classes, enhancing the flexibility of the exporter registry, and fixing a minor bug in the tests. I've also suggested opportunities to simplify some of the implementation details. Overall, this is a high-quality contribution that adds significant value to Keras Hub.
| def get_input_signature(self, image_size=None): | ||
| """Get input signature for image classifier models. | ||
| Args: | ||
| image_size: Optional image size. If None, inferred from model. | ||
| Returns: | ||
| `dict`. Dictionary mapping input names to their specifications | ||
| """ | ||
| if image_size is None: | ||
| # Get from preprocessor | ||
| if hasattr(self.model, "preprocessor") and self.model.preprocessor: | ||
| if hasattr(self.model.preprocessor, "image_size"): | ||
| image_size = self.model.preprocessor.image_size | ||
|
|
||
| # Try to infer from model inputs | ||
| if ( | ||
| image_size is None | ||
| and hasattr(self.model, "inputs") | ||
| and self.model.inputs | ||
| ): | ||
| input_shape = self.model.inputs[0].shape | ||
| if ( | ||
| len(input_shape) == 4 | ||
| and input_shape[1] is not None | ||
| and input_shape[2] is not None | ||
| ): | ||
| image_size = (input_shape[1], input_shape[2]) | ||
|
|
||
| if image_size is None: | ||
| raise ValueError( | ||
| "Could not determine image size from model. " | ||
| "Model should have a preprocessor with image_size " | ||
| "attribute, or model inputs should have concrete shapes." | ||
| ) | ||
|
|
||
| if isinstance(image_size, int): | ||
| image_size = (image_size, image_size) | ||
|
|
||
| # Get input dtype | ||
| dtype = "float32" | ||
| if hasattr(self.model, "inputs") and self.model.inputs: | ||
| model_dtype = self.model.inputs[0].dtype | ||
| dtype = ( | ||
| model_dtype.name | ||
| if hasattr(model_dtype, "name") | ||
| else model_dtype | ||
| ) | ||
|
|
||
| return { | ||
| "images": keras.layers.InputSpec( | ||
| shape=(None, *image_size, 3), | ||
| dtype=dtype, | ||
| name="images", | ||
| ), | ||
| } |
There was a problem hiding this comment.
The logic for get_input_signature is very similar across the image-based exporter configurations (ImageClassifierExporterConfig, ObjectDetectorExporterConfig, ImageSegmenterExporterConfig), especially the parts that determine image_size and dtype. To avoid code duplication, you could extract this common logic into a base class like ImageExporterConfig. The child classes would then only need to define their specific EXPECTED_INPUTS and construct the final signature dictionary.
|
@gemini-code-assist review the code, and find the remaining issues. |
There was a problem hiding this comment.
Code Review
This pull request introduces a comprehensive model export functionality to liteRT, which is a significant and valuable addition. The implementation is well-structured, following a registry pattern with specific configurations for different model types, which is great for extensibility. The code is accompanied by a good set of unit and integration tests.
My review focuses on improving maintainability by reducing code duplication, fixing a bug in the model type detection logic, and correcting a few issues in the tests. These changes will make the new export feature more robust and easier to maintain in the future.
| def get_input_signature(self, image_size=None): | ||
| """Get input signature for image classifier models. | ||
| Args: | ||
| image_size: Optional image size. If None, inferred from model. | ||
| Returns: | ||
| `dict`. Dictionary mapping input names to their specifications | ||
| """ | ||
| if image_size is None: | ||
| # Get from preprocessor | ||
| if hasattr(self.model, "preprocessor") and self.model.preprocessor: | ||
| if hasattr(self.model.preprocessor, "image_size"): | ||
| image_size = self.model.preprocessor.image_size | ||
|
|
||
| # Try to infer from model inputs | ||
| if ( | ||
| image_size is None | ||
| and hasattr(self.model, "inputs") | ||
| and self.model.inputs | ||
| ): | ||
| input_shape = self.model.inputs[0].shape | ||
| if ( | ||
| len(input_shape) == 4 | ||
| and input_shape[1] is not None | ||
| and input_shape[2] is not None | ||
| ): | ||
| image_size = (input_shape[1], input_shape[2]) | ||
|
|
||
| if image_size is None: | ||
| raise ValueError( | ||
| "Could not determine image size from model. " | ||
| "Model should have a preprocessor with image_size " | ||
| "attribute, or model inputs should have concrete shapes." | ||
| ) | ||
|
|
||
| if isinstance(image_size, int): | ||
| image_size = (image_size, image_size) | ||
|
|
||
| # Get input dtype | ||
| dtype = "float32" | ||
| if hasattr(self.model, "inputs") and self.model.inputs: | ||
| model_dtype = self.model.inputs[0].dtype | ||
| dtype = ( | ||
| model_dtype.name | ||
| if hasattr(model_dtype, "name") | ||
| else model_dtype | ||
| ) | ||
|
|
||
| return { | ||
| "images": keras.layers.InputSpec( | ||
| shape=(None, *image_size, 3), | ||
| dtype=dtype, | ||
| name="images", | ||
| ), | ||
| } |
There was a problem hiding this comment.
The logic within this get_input_signature method for determining image_size and input dtype is duplicated in ObjectDetectorExporterConfig and ImageSegmenterExporterConfig.
To improve maintainability and reduce redundancy, consider extracting this shared logic into a common base class (e.g., ImageModelExporterConfig). The individual image model configuration classes (ImageClassifierExporterConfig, ObjectDetectorExporterConfig, ImageSegmenterExporterConfig) could then inherit from this new base class.
This would centralize the logic, making future changes easier.
Updated docstrings in base.py, configs.py, and litert.py to specify explicit type annotations for function arguments and return values. This improves code readability and helps developers understand expected input types for exporter configuration and usage.
Ensure exported model filepath ends with '.tflite' and update verbose message to reflect correct path. Remove unused 'mask' argument from model calls in LiteRTExporter adapters. Update test to use input dtypes from interpreter for test inputs.
Simplifies calls to keras_hub_model in TextModelAdapter and ImageModelAdapter by removing unnecessary line breaks and grouping arguments. Also updates string quotes for consistency and improves formatting in litert_test.py for readability.
|
@gemini-code-assist |
|
@pctablet505 I can provide a new code review for the current state of the pull request, but I cannot directly check previous review comments or their resolution status. To initiate a new review, please use the |
|
@gemini-code-assist review |
There was a problem hiding this comment.
Code Review
This pull request introduces a comprehensive model export functionality for liteRT, which is a significant and valuable addition to keras-hub. The implementation is well-structured, following a modular design with base classes, configurations, and a registry. The inclusion of extensive unit and integration tests, including numerical verification for production models, is commendable and ensures the reliability of the new feature.
I have provided a few suggestions aimed at improving maintainability by reducing code duplication and simplifying some of the logic. These are minor refactorings that can enhance the clarity and robustness of the codebase. Overall, this is an excellent contribution.
| # For list-like _DictWrapper (e.g., transformer_layers) | ||
| if hasattr(child, "_data") and isinstance( | ||
| child._data, list | ||
| ): | ||
| # Create a clean list of the trackable items | ||
| clean_list = [] | ||
| for item in child._data: | ||
| if hasattr(item, "_trackable_children"): | ||
| clean_list.append(item) | ||
| if clean_list: | ||
| clean_children[name] = clean_list | ||
| # For dict-like _DictWrapper | ||
| elif hasattr(child, "_data") and isinstance( | ||
| child._data, dict | ||
| ): | ||
| clean_dict = {} | ||
| for k, v in child._data.items(): | ||
| if hasattr(v, "_trackable_children"): | ||
| clean_dict[k] = v | ||
| if clean_dict: | ||
| clean_children[name] = clean_dict | ||
| # Skip if we can't unwrap safely |
There was a problem hiding this comment.
The logic for unwrapping _DictWrapper objects can be made more concise and Pythonic by using list and dictionary comprehensions. This would improve readability.
# For list-like _DictWrapper (e.g., transformer_layers)
if hasattr(child, "_data") and isinstance(
child._data, list
):
# Create a clean list of the trackable items
clean_list = [
item
for item in child._data
if hasattr(item, "_trackable_children")
]
if clean_list:
clean_children[name] = clean_list
# For dict-like _DictWrapper
elif hasattr(child, "_data") and isinstance(
child._data, dict
):
clean_dict = {
k: v
for k, v in child._data.items()
if hasattr(v, "_trackable_children")
}
if clean_dict:
clean_children[name] = clean_dict
# Skip if we can't unwrap safelyAdded a warning comment about using the private _DictWrapper API from tensorflow.python.trackable.data_structures in Backbone. This highlights potential instability and suggests considering alternatives or stricter TensorFlow version pinning if issues arise.
Removed the unused TextModelExporterConfig and its imports. Refactored sequence length and image dtype inference into helper functions for reuse and clarity. Updated LiteRTExporter to pass parameters to the export wrapper and simplified model type checks. Cleaned up Keras-Hub model detection logic in registry.
Replaced explicit for-loops with list and dict comprehensions for filtering trackable children in lists and dicts. This improves code readability and conciseness in the Backbone model.
Replaces hardcoded model class list with iteration over registered configs in ExporterRegistry. This improves maintainability and extensibility by removing direct imports and manual class checks.
Removed unnecessary try/except blocks and test skipping logic from multiple test files, making test failures more explicit and improving readability. Updated registry initialization to ensure Seq2SeqLM is registered before CausalLM. Simplified _is_keras_hub_model logic to use isinstance(Task) for more robust model type detection.
|
@gemini-code-assist review |
There was a problem hiding this comment.
Code Review
This pull request introduces a significant new feature: support for exporting Keras-Hub models to the LiteRT format. The implementation is well-designed and modular, with base classes for exporters and configurations, a registry for discoverability, and a specific LiteRTExporter. The feature is nicely integrated by extending the export method on Task models, providing a seamless user experience. The accompanying tests are thorough, covering unit tests with mocks and crucial integration tests with production models that include numerical verification. I have identified a couple of minor issues: one related to redundant code in an __init__.py file and another concerning a misleading docstring. These are straightforward to address. Overall, this is a high-quality contribution that adds valuable functionality.
Corrected the docstring in _is_model_compatible to refer to text classifier instead of image classifier.
| @@ -0,0 +1,1561 @@ | |||
| # LiteRT Model Export Design Document | |||
There was a problem hiding this comment.
remove thsi file. Adding a guide on Keras io is okay. Keep a separate design doc
| @@ -0,0 +1,536 @@ | |||
| """Tests for LiteRT export with specific production models. | |||
There was a problem hiding this comment.
better way is to add a standardized test like this - https://github.com/keras-team/keras-hub/blob/master/keras_hub/src/tests/test_case.py#L414
enable it for all models
There was a problem hiding this comment.
we can not run the test for all models on keras-hub, some are in multiple gigabytes, and that would require too much memory that won't be available. the system with 50GB crashes for 7b models.
2 participants
Added support for model export to keras-hub models.
This PR requires keras-team/keras#21674 as prerequisite, the export feature in keras.
Then it is built on top of that.
Simple Demo
Complete Numeric verification tests multiple models for numeric verifications.
Verified models: