Heinrich Dinkel commited on
Commit ·
f8e3b40
1
Parent(s): 3f1e105
Added README
Browse files
README.md
CHANGED
|
@@ -12,6 +12,13 @@ license: apache-2.0
|
|
| 12 |
DashengTokenizer is a high-performance continious audio tokenizer designed for audio understanding and generation tasks.
|
| 13 |
Compared to previous works, our framework simply trains a single linear layer to enable audio generation for semantically strong encoders.
|
| 14 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 15 |
|
| 16 |
|
| 17 |
|
|
|
|
| 12 |
DashengTokenizer is a high-performance continious audio tokenizer designed for audio understanding and generation tasks.
|
| 13 |
Compared to previous works, our framework simply trains a single linear layer to enable audio generation for semantically strong encoders.
|
| 14 |
|
| 15 |
+
Achievements:
|
| 16 |
+
|
| 17 |
+
* State-of-the-Art Audio Understanding: DashengTokenizer consistently outperforms most previous self-supervised and supervised audio encoders.
|
| 18 |
+
* High-Fidelity Signal Reconstruction: Maintains exceptional signal integrity, ensuring that audio remains crisp and accurate after processing.
|
| 19 |
+
* Accelerated Audio Generation Training: Achieves optimal performance significantly faster than standard VAE models, reducing training time and costs.
|
| 20 |
+
* Superior Speech Enhancement: Provides a more robust encoding foundation for isolating and clarifying speech in noisy environments.
|
| 21 |
+
|
| 22 |
|
| 23 |
|
| 24 |
|