Upload README.md with huggingface_hub
Browse files
README.md
ADDED
|
@@ -0,0 +1,66 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language:
|
| 3 |
+
- ar
|
| 4 |
+
- be
|
| 5 |
+
- bg
|
| 6 |
+
- bn
|
| 7 |
+
- cs
|
| 8 |
+
- cy
|
| 9 |
+
- da
|
| 10 |
+
- de
|
| 11 |
+
- el
|
| 12 |
+
- en
|
| 13 |
+
- es
|
| 14 |
+
- et
|
| 15 |
+
- fa
|
| 16 |
+
- fi
|
| 17 |
+
- fr
|
| 18 |
+
- gl
|
| 19 |
+
- hi
|
| 20 |
+
- hu
|
| 21 |
+
- it
|
| 22 |
+
- ja
|
| 23 |
+
- ka
|
| 24 |
+
- lt
|
| 25 |
+
- lv
|
| 26 |
+
- mk
|
| 27 |
+
- mr
|
| 28 |
+
- nl
|
| 29 |
+
- pl
|
| 30 |
+
- pt
|
| 31 |
+
- ro
|
| 32 |
+
- ru
|
| 33 |
+
- sk
|
| 34 |
+
- sl
|
| 35 |
+
- sr
|
| 36 |
+
- sv
|
| 37 |
+
- sw
|
| 38 |
+
- ta
|
| 39 |
+
- th
|
| 40 |
+
- tr
|
| 41 |
+
- uk
|
| 42 |
+
- ur
|
| 43 |
+
- vi
|
| 44 |
+
- zh
|
| 45 |
+
library_name: transformers
|
| 46 |
+
license: mit
|
| 47 |
+
metrics:
|
| 48 |
+
- bleu
|
| 49 |
+
pipeline_tag: audio-text-to-text
|
| 50 |
+
---
|
| 51 |
+
|
| 52 |
+
Test ultravox model. More coming soon... I hope so.
|
| 53 |
+
|
| 54 |
+
```python
|
| 55 |
+
import transformers
|
| 56 |
+
import numpy as np
|
| 57 |
+
import librosa
|
| 58 |
+
|
| 59 |
+
pipe = transformers.pipeline(model='AtAndDev/UVOX-40k-Llama-3.2-3B-Instruct', trust_remote_code=True, device="cuda")
|
| 60 |
+
|
| 61 |
+
path = "voice_input.mp3"
|
| 62 |
+
audio, sr = librosa.load(path, sr=16000)
|
| 63 |
+
|
| 64 |
+
turns = []
|
| 65 |
+
pipe({'audio': audio, 'turns': turns, 'sampling_rate': sr}, max_new_tokens=100)
|
| 66 |
+
```
|