Update README.md
Browse files
README.md
CHANGED
|
@@ -25,7 +25,7 @@ model-index:
|
|
| 25 |
---
|
| 26 |
|
| 27 |
This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on the NST dataset.
|
| 28 |
-
Aborted after 6000 steps / 0.4 epochs as it wasen't promising when manualy evaluated on an SVT broadcast. The punctation, capitalization and entities like Norge seems worse than original so probably need to fix dataset before more training.
|
| 29 |
|
| 30 |
### Training results
|
| 31 |
| Step | Wer |
|
|
|
|
| 25 |
---
|
| 26 |
|
| 27 |
This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on the NST dataset.
|
| 28 |
+
Aborted after 6000 steps / 0.4 epochs as it wasen't promising when manualy evaluated on an SVT broadcast. The punctation, capitalization and entities like Norge seems worse than original so probably need to fix dataset before more training. Re-split the test dataset to contain a thousand samples so evaluate didn't take hours.
|
| 29 |
|
| 30 |
### Training results
|
| 31 |
| Step | Wer |
|