metadata
base_model:
- openai/whisper-medium
metrics:
- wer
pipeline_tag: automatic-speech-recognition
tags:
- upper_sorbian
Model Description
This model was fine-tuned of transcribed upper sorbian speech. This is intended to contribute to the preservation of the Upper Sorbian language and can be freely used for language preservation projects in accordance with the license.
Note: This model is a snapshot from an ongoing fine-tuning session. Further (improved) versions of this will follow shortly.
Model Details
- Model Name: DL-whisper-medium-hsb
- Model Version: 1.0.0
- Model Date: 2024-11-18
- Publisher:
- Architecture: Whisper Medium
- License:
Training Details
Parameters:
- Learning rate: 6.25 x 10-⁶, linear decay
- Warmup: 1000 steps
- Additional: BF16 training
Performance
- Word Error Rate: 6.48