Setsugesuka
commited on
Commit
•
4ef3163
1
Parent(s):
992c286
Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,58 @@
|
|
1 |
---
|
2 |
license: mit
|
3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
license: mit
|
3 |
---
|
4 |
+
|
5 |
+
# Amphion Vocoder Pretrained Models
|
6 |
+
|
7 |
+
We provide a [HiFi-GAN](https://github.com/open-mmlab/Amphion/tree/main/egs/vocoder/gan/tfr_enhanced_hifigan) pretrained checkpoint for speech, which is trained on 685 hours of speech data.
|
8 |
+
|
9 |
+
## Quick Start
|
10 |
+
|
11 |
+
To utilize these pretrained vocoders, just run the following commands:
|
12 |
+
|
13 |
+
### Step1: Download the checkpoint
|
14 |
+
```bash
|
15 |
+
git lfs install
|
16 |
+
git clone https://huggingface.co/amphion/vocoder
|
17 |
+
```
|
18 |
+
|
19 |
+
### Step2: Clone the Amphion's Source Code of GitHub
|
20 |
+
```bash
|
21 |
+
git clone https://github.com/open-mmlab/Amphion.git
|
22 |
+
```
|
23 |
+
|
24 |
+
### Step3: Specify the checkpoint's path
|
25 |
+
Use the soft link to specify the downloaded checkpoint in the first step:
|
26 |
+
|
27 |
+
```bash
|
28 |
+
cd Amphion
|
29 |
+
mkdir ckpts/vocoder
|
30 |
+
cd ckpts/vocoder
|
31 |
+
ln -s ../vocoder/hifigan_speech ckpts/vocoder/hifigan_speech
|
32 |
+
```
|
33 |
+
|
34 |
+
### Step4: Inference
|
35 |
+
For analysis synthesis on the processed dataset, raw waveform, or predicted mel spectrograms, you can follow the inference part of [this recipe](https://github.com/open-mmlab/Amphion/blob/main/egs/vocoder/gan/tfr_enhanced_hifigan/README.md).
|
36 |
+
|
37 |
+
```bash
|
38 |
+
sh egs/vocoder/gan/tfr_enhanced_hifigan/run.sh --stage 3 \
|
39 |
+
--infer_mode [Your chosen inference mode] \
|
40 |
+
--infer_datasets [Datasets you want to inference, needed when infer_from_dataset] \
|
41 |
+
--infer_feature_dir [Your path to your predicted acoustic features, needed when infer_from_feature] \
|
42 |
+
--infer_audio_dir [Your path to your audio files, needed when infer_form_audio] \
|
43 |
+
--infer_expt_dir Amphion/ckpts/vocoder/[YourExptName] \
|
44 |
+
--infer_output_dir Amphion/ckpts/vocoder/[YourExptName]/result \
|
45 |
+
```
|
46 |
+
|
47 |
+
## Citaions
|
48 |
+
|
49 |
+
```bibtex
|
50 |
+
@misc{gu2023cqt,
|
51 |
+
title={Multi-Scale Sub-Band Constant-Q Transform Discriminator for High-Fidelity Vocoder},
|
52 |
+
author={Yicheng Gu and Xueyao Zhang and Liumeng Xue and Zhizheng Wu},
|
53 |
+
year={2023},
|
54 |
+
eprint={2311.14957},
|
55 |
+
archivePrefix={arXiv},
|
56 |
+
primaryClass={cs.SD}
|
57 |
+
}
|
58 |
+
```
|