[Question] How to run on collab

[Bhushan20-oss]

Hi Maintainers,

Thank you for the excellent work on this project — the demo results look very promising.

I am trying to run the demo using Google Colab, but I am facing dependency and environment issues that make it difficult to reproduce the results shown in the documentation.

I would really appreciate step-by-step guidance specifically for Google Colab, covering the complete workflow:

What I’m Looking For

1. Environment setup in Colab

   • Required Python version
   • GPU/CUDA requirements
   • Any system packages that need installation

2. Dependency installation

   • Exact versions of PyTorch, TorchVision, Torchaudio
   • Any libraries that commonly cause conflicts
   • Recommended installation order (if important)

3. Model setup

   • How to download model weights
   • Where to place them in Colab environment
   • Any authentication required (e.g., Hugging Face)

4. Running the demo

   • Minimal working example for inference
   • Sample audio input format requirements
   • Expected runtime and GPU memory needs

5. Common pitfalls

   • Known version conflicts
   • Troubleshooting tips for CUDA / Torch mismatches

---

Issue Faced

While installing dependencies, version conflicts between PyTorch and related libraries caused runtime errors (e.g., missing TorchVision operators). A Colab-specific installation guide would help avoid such issues.

---

If possible, a Colab notebook similar to the demo environment would be extremely helpful for users trying to evaluate the project quickly.

Thank you for your support and for maintaining this project!

Best regards.