add pretrained models

Author: wpingnet

README.md

@@ -62,11 +62,22 @@ python inference_e2e.py \
 --output_dir /path/to/your/output_wav
 ```
 
+## Pretrained Models
+We provide the [pretrained models](https://drive.google.com/drive/folders/1e9wdM29d-t3EHUpBb8T4dcHrkYGAXTgq).
+One can download the checkpoints of generator (e.g., g_05000000) and discriminator (e.g., do_05000000) within the listed folders.
 
-## TODO
+|Folder Name|Sampling Rate|Mel band|Params.|Dataset|Fine-Tuned|
+|------|---|---|---|------|---|
+|bigvgan_24khz_100band|24 kHz|100|112M|LibriTTS|No|
+|bigvgan_base_24khz_100band|24 kHz|100|14M|LibriTTS|No|
+|bigvgan_22khz_80band|22 kHz|80|112M|LibriTTS + VCTK + LJSpeech|No|
+|bigvgan_base_22khz_80band|22 kHz|80|14M|LibriTTS + VCTK + LJSpeech|No|
+
+The paper results are based on 24kHz BigVGAN models trained on LibriTTS dataset.
+We also provide 22kHz BigVGAN models for TTS applications.
 
-Pretrained weights are expected to be released soon.
 
+## TODO
 
 Current codebase only provides a plain PyTorch implementation for the filtered nonlinearity. We are working on a fast CUDA kernel implementation, which will be released in the future.