Spaces:
Sleeping
Sleeping
primepake
commited on
Commit
·
24941fa
1
Parent(s):
3595b78
add fsq training
Browse files
README.md
CHANGED
|
@@ -11,6 +11,7 @@ This repository provides an implementation of the MiniMax-Speech model, featurin
|
|
| 11 |
## Key Features
|
| 12 |
|
| 13 |
- [ ] **24kHz Audio Support**: High-quality audio generation at 24kHz sampling rate
|
|
|
|
| 14 |
- [ ] **Two-Stage Architecture**: Optimized training pipeline with discrete and continuous representations
|
| 15 |
- [ ] **Modular Design**: Separate components for audio codec and variational autoencoder
|
| 16 |
- [ ] **CosyVoice2 Decoder**: Leverages proven components from the CosyVoice2's Decoder framework
|
|
|
|
| 11 |
## Key Features
|
| 12 |
|
| 13 |
- [ ] **24kHz Audio Support**: High-quality audio generation at 24kHz sampling rate
|
| 14 |
+
- [ ] **FSQ tokenizer training**: Training FSQ from scratch
|
| 15 |
- [ ] **Two-Stage Architecture**: Optimized training pipeline with discrete and continuous representations
|
| 16 |
- [ ] **Modular Design**: Separate components for audio codec and variational autoencoder
|
| 17 |
- [ ] **CosyVoice2 Decoder**: Leverages proven components from the CosyVoice2's Decoder framework
|