Work Progress Report on Garchen Rinpoche's STT Model

🧩 SIG ལས་གཞི། Garchen Rinpoche Speech SIG
1

1. Model Evaluation

Conducted two consecutive fine-tuning sessions of Garchen Rinpoche’s STT model using:

  1. 5 hours of His Eminence Garchen Rinpoche’s audio transcription (Click here for the data).

  2. 10 hours of His Eminence Garchen Rinpoche’s audio transcription (Click here for the data)..

2. Transcription Speed Analysis

3. Manual Transcription Work

  • Successfully completed 23 hours 30 minutes of manual transcription by transcribers.

  • Manual transcription work is ongoing, with progress moving toward the target of completing 31.45 hours of audio processed.

Summary:
The evaluation, speed analysis, and completed manual transcription mark significant progress in the development of the STT Model for His Eminence Garchen Rinpoche’s teachings. Ongoing transcription efforts will further enhance training data for improved model performance.

For model, visit the following link:

Smart Speech-to-Text with VAD and Subtitles (Multi-Model)