* Refactor model and artifact structures with version compatibility filtering
- Introduced `Artifact` struct and nested it within the `Model` struct for improved clarity and organization.
- Updated enums, logic, and parsing to align with the new struct definitions.
- Implemented version compatibility filtering for model bundles using the `is_bundle_version_compatible` helper.
- Enhanced artifact download handling by adding checks for missing URIs, better error management, and improved logging.
- Adjusted model fetching to point to the latest endpoint (`v3`).
* Make linter happy
* Make linter happy
* Refactor model data parsing to improve readability.
Replaced kwargs-based data extraction with explicit parameter passing for clarity. This enhances code readability and reduces ambiguities in method calls, making the parsing logic more maintainable and straightforward.
* Refactor error handling in active model bundle retrieval.
Wrapped the logic to fetch the active model bundle in a try-except block to prevent unhandled exceptions. This ensures more robust error handling and avoids potential crashes when retrieving or processing model data.
* Refactor exception handling in get_active_model_bundle
Replace bare except with Exception to improve specificity and clarity. This ensures better debugging practices and aligns with recommended coding standards. Other minor whitespace adjustments were made for improved readability.
* Update model path to use artifact fileName property
Replaced `fileName` with `artifact.fileName` in the custom model path construction. This ensures compatibility with updated drive model structures and avoids potential file resolution issues.
* Refactor model runner methods for improved abstraction.
Moved slicing logic to a private `_slice_outputs` method and decoupled `_run_model` for clearer subclass implementation. Removed redundant `output` attribute in `ModelState` to streamline data handling.
* Add output parsing to model_runner and remove duplicate logic
Integrates an output parser directly into `model_runner` for streamlined inference and parsing. Removes redundant parser initialization from `modeld` to avoid duplication and enhance maintainability.
* Reordering
* linter
* linter
Enhance action to fail on non-successful workflow runs
Add logic to check the conclusion of the watched workflow run. If the run ends with a non-successful status, the action now exits with an error to improve error handling and ensure reliability.
Co-authored-by: Jason Wen <haibin.wen3@gmail.com>
* Remove supercombo model ONNX file.
Deleted the large `supercombo.onnx` model file from the repository. This cleanup reduces repository size and dependency on unused or outdated files for this version.
* Disable tinygrad model compilation on macos temporarily
* Remove unused dmonitoring model file.
Deleted the ONNX model for dmonitoring as it is no longer required. This eliminates unnecessary assets and reduces repository size.
* Removing the model also from the snpe build, we have them, prebuilt
infiniteCable2