mskov commited on
Commit
13cee50
β€’
1 Parent(s): d7388cd

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +5 -2
app.py CHANGED
@@ -1,8 +1,7 @@
1
  import os
2
  import sys
3
  os.system("pip install transformers==4.27.0")
4
- from transformers import pipeline, WhisperModel
5
- import gradio as gr
6
  os.system("pip install evaluate")
7
  os.system("pip install datasets")
8
  os.system("pip install llvmlite")
@@ -21,9 +20,12 @@ disable_caching()
21
 
22
  p = pipeline("automatic-speech-recognition")
23
 
 
 
24
  huggingface_token = os.environ["huggingface_token"]
25
 
26
  whisper_miso=WhisperModel.from_pretrained("mskov/whisper_miso", use_auth_token=huggingface_token)
 
27
 
28
 
29
  task_evaluator = evaluator("automatic-speech-recognition")
@@ -35,6 +37,7 @@ results = task_evaluator.compute(
35
  model_or_pipeline=whisper_miso,
36
  #model_or_pipeline="mskov/whisper-small.en",
37
  data=dataset,
 
38
  input_column="audio",
39
  label_column="audio",
40
  # device=None,
 
1
  import os
2
  import sys
3
  os.system("pip install transformers==4.27.0")
4
+ from transformers import pipeline, WhisperModel, WhisperTokenizer
 
5
  os.system("pip install evaluate")
6
  os.system("pip install datasets")
7
  os.system("pip install llvmlite")
 
20
 
21
  p = pipeline("automatic-speech-recognition")
22
 
23
+ #config = AutoConfig.from_pretrained('whisper-small')
24
+
25
  huggingface_token = os.environ["huggingface_token"]
26
 
27
  whisper_miso=WhisperModel.from_pretrained("mskov/whisper_miso", use_auth_token=huggingface_token)
28
+ miso_tokenizer = WhisperTokenizer.from_pretrained("mskov/whisper_miso/tokenizer_config.json", use_auth_token=huggingface_token)
29
 
30
 
31
  task_evaluator = evaluator("automatic-speech-recognition")
 
37
  model_or_pipeline=whisper_miso,
38
  #model_or_pipeline="mskov/whisper-small.en",
39
  data=dataset,
40
+ tokenizer=miso_tokenizer,
41
  input_column="audio",
42
  label_column="audio",
43
  # device=None,