Update README.md
Browse files
README.md
CHANGED
|
@@ -18,25 +18,16 @@ tags:
|
|
| 18 |
pipeline_tag: automatic-speech-recognition
|
| 19 |
---
|
| 20 |
|
|
|
|
| 21 |
|
| 22 |
-
|
| 23 |
-
|
| 24 |
-
This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the common_voice_16_1 dataset. It is finetuned for the Armenian language.
|
| 25 |
It achieves the following results on the evaluation set:
|
| 26 |
- Loss: 0.2853
|
| 27 |
- Wer: 38.1160
|
| 28 |
|
| 29 |
-
##
|
| 30 |
-
|
| 31 |
-
More information needed
|
| 32 |
-
|
| 33 |
-
## Intended uses & limitations
|
| 34 |
-
|
| 35 |
-
More information needed
|
| 36 |
-
|
| 37 |
-
## Training and evaluation data
|
| 38 |
|
| 39 |
-
|
| 40 |
|
| 41 |
## Training procedure
|
| 42 |
|
|
|
|
| 18 |
pipeline_tag: automatic-speech-recognition
|
| 19 |
---
|
| 20 |
|
| 21 |
+
## Model description
|
| 22 |
|
| 23 |
+
Chillarmo/whisper-small-hy-AM is an AI model designed for speech-to-text conversion specifically tailored for the Armenian language. Leveraging the power of fine-tuning, this model, named whisper-small-hy-AM, is based on [openai/whisper-small](https://huggingface.co/openai/whisper-small) and trained on the common_voice_16_1 dataset.
|
|
|
|
|
|
|
| 24 |
It achieves the following results on the evaluation set:
|
| 25 |
- Loss: 0.2853
|
| 26 |
- Wer: 38.1160
|
| 27 |
|
| 28 |
+
## Training Data and Future Enhancements
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 29 |
|
| 30 |
+
The training data consists of Mozilla Common Voice version 16.1. Plans for future improvements include continuing the training process and integrating an additional 10 hours of data from datasets such as google/fleurs and possibly google/xtreme_s. Despite its current performance, efforts are underway to further reduce the WER.
|
| 31 |
|
| 32 |
## Training procedure
|
| 33 |
|