Motocle commited on
Commit
08fc52f
1 Parent(s): 496449d

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +20 -26
app.py CHANGED
@@ -5,8 +5,8 @@ import pytube as pt
5
  from transformers import pipeline
6
  from huggingface_hub import model_info
7
 
8
- MODEL_NAME = "openai/whisper-small" #this always needs to stay in line 8 :D sorry for the hackiness
9
- lang = "en"
10
 
11
  device = 0 if torch.cuda.is_available() else "cpu"
12
  pipe = pipeline(
@@ -61,37 +61,31 @@ demo = gr.Blocks()
61
  mf_transcribe = gr.Interface(
62
  fn=transcribe,
63
  inputs=[
64
- gr.inputs.Audio(source="microphone", type="filepath", optional=True),
65
- gr.inputs.Audio(source="upload", type="filepath", optional=True),
66
  ],
67
- outputs="text",
68
- layout="horizontal",
69
- theme="huggingface",
70
- title="Whisper Demo: Transcribe Audio",
71
- description=(
72
- "Transcribe long-form microphone or audio inputs with the click of a button! Demo uses the the fine-tuned"
73
- f" checkpoint [{MODEL_NAME}](https://huggingface.co/{MODEL_NAME}) and 🤗 Transformers to transcribe audio files"
74
- " of arbitrary length."
75
- ),
76
- allow_flagging="never",
77
  )
78
 
79
  yt_transcribe = gr.Interface(
80
  fn=yt_transcribe,
81
- inputs=[gr.inputs.Textbox(lines=1, placeholder="Paste the URL to a YouTube video here", label="YouTube URL")],
82
- outputs=["html", "text"],
83
- layout="horizontal",
84
- theme="huggingface",
85
- title="Whisper Demo: Transcribe YouTube",
86
- description=(
87
- "Transcribe long-form YouTube videos with the click of a button! Demo uses the the fine-tuned checkpoint:"
88
- f" [{MODEL_NAME}](https://huggingface.co/{MODEL_NAME}) and 🤗 Transformers to transcribe audio files of"
89
- " arbitrary length."
90
- ),
91
- allow_flagging="never",
92
  )
93
 
94
  with demo:
95
  gr.TabbedInterface([mf_transcribe, yt_transcribe], ["Transcribe Audio", "Transcribe YouTube"])
96
 
97
- demo.launch(enable_queue=True)
 
5
  from transformers import pipeline
6
  from huggingface_hub import model_info
7
 
8
+ MODEL_NAME = "openai/whisper-large-v3" #this always needs to stay in line 8 :D sorry for the hackiness
9
+ lang = "ja"
10
 
11
  device = 0 if torch.cuda.is_available() else "cpu"
12
  pipe = pipeline(
 
61
  mf_transcribe = gr.Interface(
62
  fn=transcribe,
63
  inputs=[
64
+ gr.Audio(type="filepath"),
65
+ gr.Audio(type="filepath"),
66
  ],
67
+ outputs="text"
68
+ )
69
+ mf_transcribe.title = "Whisper Demo: Transcribe Audio"
70
+ mf_transcribe.description = (
71
+ "Transcribe long-form microphone or audio inputs with the click of a button! Demo uses the the fine-tuned"
72
+ f" checkpoint [{MODEL_NAME}](https://huggingface.co/{MODEL_NAME}) and 🤗 Transformers to transcribe audio files"
73
+ " of arbitrary length."
 
 
 
74
  )
75
 
76
  yt_transcribe = gr.Interface(
77
  fn=yt_transcribe,
78
+ inputs=[gr.Textbox(lines=1, placeholder="Paste the URL to a YouTube video here", label="YouTube URL")],
79
+ outputs=["html", "text"]
80
+ )
81
+ yt_transcribe.title = "Whisper Demo: Transcribe YouTube"
82
+ yt_transcribe.description = (
83
+ "Transcribe long-form YouTube videos with the click of a button! Demo uses the the fine-tuned checkpoint:"
84
+ f" [{MODEL_NAME}](https://huggingface.co/{MODEL_NAME}) and 🤗 Transformers to transcribe audio files of"
85
+ " arbitrary length."
 
 
 
86
  )
87
 
88
  with demo:
89
  gr.TabbedInterface([mf_transcribe, yt_transcribe], ["Transcribe Audio", "Transcribe YouTube"])
90
 
91
+ demo.launch() #'enable_queue=True' を削除