File size: 814 Bytes
7ee1956
 
 
 
 
43f2182
 
d8d5cc3
7ee1956
 
43f2182
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
---
license: mit
tags:
- model_hub_mixin
- pytorch_model_hub_mixin
base_model:
- Alethia/BigCodec
pipeline_tag: feature-extraction
---

# BigCodec

Python library for bigcodec. This is something that the author of bigcodec didn't do, so I'm doing it for him.

## Setup

```
pip install bigcodec
```

## Usage

```
import torch.accelerator
from bigcodec import BigCodec
import torchaudio

codec = BigCodec.from_pretrained("intexcp/bigcodec")

wav = torchaudio.load("enc.wav")[0]
wav = wav.unsqueeze(0)
device = torch.accelerator.current_accelerator().type if torch.accelerator.is_available() else "cpu"
codec = codec.to(device)
wav = wav.to(device)
with torch.no_grad():
    enc = codec.encode(wav)
    dec = codec.decode(enc)
print(enc)
torchaudio.save("dec.wav", dec.squeeze(0).cpu(), 16000, encoding="PCM_F")
```