Fix training data stats: remove indicvoices_r
Browse files
README.md
CHANGED
|
@@ -1,3 +1,4 @@
|
|
|
|
|
| 1 |
---
|
| 2 |
license: mit
|
| 3 |
language:
|
|
@@ -19,11 +20,10 @@ A fine-tuned version of [ResembleAI/chatterbox](https://huggingface.co/ResembleA
|
|
| 19 |
## Model Details
|
| 20 |
|
| 21 |
- **Base model**: ResembleAI/chatterbox (0.5B Llama-based TTS)
|
| 22 |
-
- **Fine-tuned on**: Bengali speech corpus (~
|
| 23 |
-
- ai4bharat/Shrutilipi (Bengali split) — 17,882 samples
|
| 24 |
-
- ai4bharat/Rasa (Bengali split) — 28,088 samples
|
| 25 |
-
- SPRINGLab/IndicTTS_Bengali — 12,852 samples
|
| 26 |
-
- ai4bharat/indicvoices_r (Bengali split) — 1,500 samples
|
| 27 |
- **Training steps**: 20,000
|
| 28 |
- **Epochs**: ~10.6
|
| 29 |
- **Architecture**: T3 (Text-to-Token Transformer) + HiFT-GAN vocoder
|
|
|
|
| 1 |
+
\
|
| 2 |
---
|
| 3 |
license: mit
|
| 4 |
language:
|
|
|
|
| 20 |
## Model Details
|
| 21 |
|
| 22 |
- **Base model**: ResembleAI/chatterbox (0.5B Llama-based TTS)
|
| 23 |
+
- **Fine-tuned on**: Bengali speech corpus (~99 hours, 58,820 samples)
|
| 24 |
+
- ai4bharat/Shrutilipi (Bengali split) — 17,882 samples, ~28h
|
| 25 |
+
- ai4bharat/Rasa (Bengali split) — 28,088 samples, ~51h
|
| 26 |
+
- SPRINGLab/IndicTTS_Bengali — 12,852 samples, ~20h
|
|
|
|
| 27 |
- **Training steps**: 20,000
|
| 28 |
- **Epochs**: ~10.6
|
| 29 |
- **Architecture**: T3 (Text-to-Token Transformer) + HiFT-GAN vocoder
|