feat: support gpu & support invoke speechscore via code

[dujl]

Accelerate distill_mos and dnsmos speech scoring using GPU

[gemini-code-assist]

Summary of Changes

Hello @dujl, 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 significantly enhances the speechscore library by introducing GPU acceleration for key models, specifically DISTILL_MOS and DNSMOS, to improve performance. Concurrently, it refactors the model weight loading mechanism across several modules to ensure more robust and portable path resolution, which is crucial for seamless programmatic integration and execution of the speech scoring functionalities.

Highlights

• GPU Acceleration: Enabled GPU support for the DISTILL_MOS model by moving the model and its input tensors to the detected CUDA device if available, otherwise defaulting to CPU.
• ONNX GPU Inference: Configured DNSMOS ONNX inference sessions to prioritize CUDAExecutionProvider, allowing for GPU-accelerated scoring when a compatible GPU is present.
• Robust Model Path Resolution: Improved the loading of model weights for DISTILL_MOS, DNSMOS, and NISQA by making their paths relative to their respective module files, enhancing portability and programmatic invocation.

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

1. 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. ↩

[gemini-code-assist]

Code Review

This pull request introduces GPU support for various scoring models and improves path handling to make invoking speechscore from code more robust. The changes for DISTILL_MOS and DNSMOS correctly add GPU support and fix model path resolution. The path resolution has also been improved for NISQA.

My review focuses on ensuring the GPU support is consistently applied. I've found one place where GPU support was missed. I also have a minor suggestion regarding file formatting conventions.

[gemini-code-assist]

This model is hardcoded to use the CPU, which is inconsistent with the goal of this pull request to add GPU support. You should dynamically select the device (GPU if available, otherwise CPU), similar to the implementation for DISTILL_MOS.

You will need to add import torch at the top of the file to use torch.device.

Suggested change

	self.model = load_nisqa_model(os.path.join(os.path.dirname(os.path.abspath(__file__)), "weights/nisqa.tar"), device='cpu')
	device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
	self.model = load_nisqa_model(os.path.join(os.path.dirname(os.path.abspath(__file__)), "weights/nisqa.tar"), device=device)

[gemini-code-assist]

It's a good practice to end files with a single newline character. This is recommended by style guides like PEP 8 and can prevent issues with some command-line tools and diff viewers that expect a final newline.