Update README.md
Browse files
README.md
CHANGED
|
@@ -45,7 +45,7 @@ Finetuned version of Facebooks [VoxPopuli-sv large](https://huggingface.co/faceb
|
|
| 45 |
When using this model, make sure that your speech input is sampled at 16kHz.
|
| 46 |
|
| 47 |
## Training
|
| 48 |
-
This model has been fine-tuned for 80000 updates on NST + CommonVoice and then for an additional 20000 steps on only CommonVoice.
|
| 49 |
|
| 50 |
## Usage
|
| 51 |
The model can be used directly (without a language model) as follows:
|
|
|
|
| 45 |
When using this model, make sure that your speech input is sampled at 16kHz.
|
| 46 |
|
| 47 |
## Training
|
| 48 |
+
This model has been fine-tuned for 80000 updates on NST + CommonVoice and then for an additional 20000 steps on only CommonVoice. The additional fine-tuning on CommonVoce hurts performance on the NST+CommonVoice test set somewhat and, unsurprisingly, improves it on the CommonVoice test set. It seems to perform generally better though [citation needed].
|
| 49 |
|
| 50 |
## Usage
|
| 51 |
The model can be used directly (without a language model) as follows:
|