metadata
language:
- en
metrics:
- wer
base_model:
- openai/whisper-small
tags:
- whisper
- stt
- speech-to-text
- speech
- automatic-speech-recognition
- fine-tuned
Pilotgpt Jack 5H V4
Fine-tuned Whisper model based on openai/whisper-small.
Training Results
| Metric | Base Model | Fine-tuned |
|---|---|---|
| WER | 59.62% | 14.56% |
Improvement: 45.06% WER reduction (lower is better)
Training Details
- Base Model: openai/whisper-small
- Training Dataset: Trelis/pilotgpt-jack-5h-sentence-rewritten-v2
- Train Loss: 0.7019
- Training Time: 19.2 minutes
Inference
from transformers import pipeline
asr = pipeline("automatic-speech-recognition", model="Trelis/pilotgpt-jack-5h-v4")
result = asr("path/to/audio.wav")
print(result["text"])
Training Logs
Full training logs are available in training_log.txt.
Fine-tuned using Trelis Studio