plug.directory
Add my server

Tibetan Speech Recognition

by drukaico.huggingface.drukai/tibetan-asr-wav2vec2-run8

How the score works

  • Remote endpoint
  • 2 tools
  • tibetan_asr_wav2vec2_run8_score Score a transcript against a reference with CER and syllable error rate. Both texts are NFC-normalised, the two tsheg code points are unified and shad punctuation is dropped. CER ignores whitespace; syllables are split on tsheg and whitespace. Returns: A Markdown summary of CER and SER, or an empty string if either text is missing.
  • tibetan_asr_wav2vec2_run8_transcribe Transcribe Tibetan speech to Tibetan (Uchen) script with openpecha/wav2vec2_run8. Long recordings are split into overlapping 10-second windows; audio over 10 minutes is truncated. If a user passes a file as an input, use the upload_file_to_gradio tool, if present, to upload the file to the gradio app and create a Gradio File Input. Then use the returned path as the input to the tool. Returns: The transcript, and a list of [start_seconds, end_seconds, text] segments.

As the server's own listing gives them. We never call a tool.