BlueRaccoon
commited on
Commit
•
8d18b6c
1
Parent(s):
4c00fe0
i promise this is the last update for now
Browse files
README.md
CHANGED
@@ -45,7 +45,8 @@ model-index:
|
|
45 |
|
46 |
# Whisper Small Uzbek
|
47 |
|
48 |
-
This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) trained on the mozilla-foundation/common_voice_11_0 uz and google/fleurs uz_uz datasets
|
|
|
49 |
It achieves the following results on the common_voice_11_0 evaluation set:
|
50 |
- Loss: 0.3872
|
51 |
- Wer: 23.6509
|
@@ -56,8 +57,11 @@ It achieves the following results on the FLEURS evaluation set:
|
|
56 |
## Model description
|
57 |
|
58 |
This model was created as part of the Whisper fine-tune sprint event.
|
|
|
59 |
Based on eval, this model achieves a WER of 23.6509 against the Common Voice 11 dataset and 47.15 against the FLEURS dataset.
|
60 |
-
|
|
|
|
|
61 |
![A part of Table 13 from the paper "Robust Speech Recognition via Large-Scale Weak Supervision", which shows the WER achieved by the Whisper model under the FLEURS dataset. Highlighted is the best score it achieved under for the Uzbek language, which was 90.2.](https://huggingface.co/BlueRaccoon/whisper-small-uz/resolve/main/uzbektable13.png)
|
62 |
|
63 |
## Intended uses & limitations
|
|
|
45 |
|
46 |
# Whisper Small Uzbek
|
47 |
|
48 |
+
This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) trained and evaluated on the mozilla-foundation/common_voice_11_0 uz and google/fleurs uz_uz datasets.
|
49 |
+
|
50 |
It achieves the following results on the common_voice_11_0 evaluation set:
|
51 |
- Loss: 0.3872
|
52 |
- Wer: 23.6509
|
|
|
57 |
## Model description
|
58 |
|
59 |
This model was created as part of the Whisper fine-tune sprint event.
|
60 |
+
|
61 |
Based on eval, this model achieves a WER of 23.6509 against the Common Voice 11 dataset and 47.15 against the FLEURS dataset.
|
62 |
+
|
63 |
+
This is a significant improvement over the smallest reported WER of 90.2 for the Uzbek language recorded on the [Whisper article](https://cdn.openai.com/papers/whisper.pdf):
|
64 |
+
|
65 |
![A part of Table 13 from the paper "Robust Speech Recognition via Large-Scale Weak Supervision", which shows the WER achieved by the Whisper model under the FLEURS dataset. Highlighted is the best score it achieved under for the Uzbek language, which was 90.2.](https://huggingface.co/BlueRaccoon/whisper-small-uz/resolve/main/uzbektable13.png)
|
66 |
|
67 |
## Intended uses & limitations
|