Fetching metadata from the HF Docker repository...
Nov 19, 2025 08:16 Fix audio loading for API: use torchaudio directly Obiang +1: Claude Bypass SpeechBrain's load_audio which uses fetch() that
misinterprets file paths as HuggingFace repo IDs.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com> 07:58 Use gr.Textbox for JSON API to receive raw file paths 07:49 Use gr.File for JSON API to fix file path handling 07:43 Fix API file path handling for predict_json endpoint 07:24 Add API Documentation tab with examples and usage guide 07:15 Add JSON API endpoint for programmatic access 07:03 Remove _docs folder from repository Oct 20, 2025 11:17 Update title to emphasize explainability and prototype-based approach 11:04 Enable API access and add comprehensive API documentation 10:58 Add interactive audio examples in the UI 10:46 Update to Pro-TeVA branding and add architecture diagram 10:39 Add dual F0 visualization: extracted vs predicted 10:31 Add dual F0 visualization: extracted vs predicted 10:12 Fix CTC decode: blank_index as positional argument 10:01 Fix module access: use self.hparams instead of self.modules 09:55 Add transformers dependency for HuBERT model 09:46 Fix README short_description for HF Spaces 09:44 Track WAV files with Git LFS 09:31 first commit 08:18 init push 04:46 initial commit 
