Marco-Cheung commited on
Commit
5099504
1 Parent(s): ce5168c

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +10 -1
app.py CHANGED
@@ -1,7 +1,7 @@
1
  import gradio as gr
2
  import numpy as np
3
  import torch
4
- from transformers import AutoProcessor, pipeline, BarkModel
5
 
6
  ASR_MODEL_NAME = "bofenghuang/whisper-large-v2-cv11-german"
7
  TTS_MODEL_NAME = "suno/bark-small"
@@ -10,11 +10,20 @@ voices = {
10
  "male" : "v2/de_speaker_0",
11
  "female" : "v2/de_speaker_3"
12
  }
 
13
  device = "cuda:0" if torch.cuda.is_available() else "cpu"
14
 
15
  # load speech translation checkpoint
16
  asr_pipe = pipeline("automatic-speech-recognition", model=ASR_MODEL_NAME, chunk_length_s=10,device=device)
17
 
 
 
 
 
 
 
 
 
18
  # load text-to-speech checkpoint
19
  processor = AutoProcessor.from_pretrained("suno/bark-small")
20
  model = BarkModel.from_pretrained("suno/bark-small").to(device)
 
1
  import gradio as gr
2
  import numpy as np
3
  import torch
4
+ from transformers import AutoProcessor, pipeline, BarkModel, GenerationConfig
5
 
6
  ASR_MODEL_NAME = "bofenghuang/whisper-large-v2-cv11-german"
7
  TTS_MODEL_NAME = "suno/bark-small"
 
10
  "male" : "v2/de_speaker_0",
11
  "female" : "v2/de_speaker_3"
12
  }
13
+
14
  device = "cuda:0" if torch.cuda.is_available() else "cpu"
15
 
16
  # load speech translation checkpoint
17
  asr_pipe = pipeline("automatic-speech-recognition", model=ASR_MODEL_NAME, chunk_length_s=10,device=device)
18
 
19
+ # update the generation config
20
+ MULTILINGUAL = True # set True for multilingual models, False for English-only
21
+ generation_config = GenerationConfig.from_pretrained("openai/whisper-large-v2")
22
+ #else:
23
+ #generation_config = GenerationConfig.from_pretrained("openai/whisper-medium.en")
24
+ generation_config.push_to_hub(ASR_MODEL_NAME)
25
+
26
+
27
  # load text-to-speech checkpoint
28
  processor = AutoProcessor.from_pretrained("suno/bark-small")
29
  model = BarkModel.from_pretrained("suno/bark-small").to(device)