siddqamar commited on
Commit
48103d0
·
verified ·
1 Parent(s): d09d492

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +28 -71
app.py CHANGED
@@ -1,86 +1,43 @@
1
  import gradio as gr
 
2
  import os
3
- from transformers import WhisperProcessor, WhisperForConditionalGeneration
4
- import numpy as np
5
- import librosa
6
 
7
- # Initialize Whisper model
8
- processor = WhisperProcessor.from_pretrained("openai/whisper-base")
9
- model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base")
10
-
11
- # Set light green theme
12
- theme = gr.themes.Base(
13
- primary_hue="emerald",
14
- secondary_hue="emerald",
15
- neutral_hue="gray",
16
- )
17
-
18
- def validate_file(file_path):
19
- # Check if file exists
20
- if not file_path or not os.path.exists(file_path):
21
- return False, "No file uploaded or file not found."
22
-
23
- # Check file size (25 MB limit)
24
- file_size_mb = os.path.getsize(file_path) / (1024 * 1024)
25
- if file_size_mb > 25:
26
- return False, f"File size is {file_size_mb:.2f} MB. Please upload a file smaller than 25 MB."
27
-
28
- # Check file extension
29
- file_extension = os.path.splitext(file_path)[1].lower()
30
- if file_extension not in ['.mp3', '.wav']:
31
- return False, "Only .mp3 and .wav formats are supported."
32
-
33
- return True, "File is valid."
34
 
35
  def transcribe_audio(audio_file):
36
- # Check if audio_file is None
37
  if audio_file is None:
38
- return "Please upload an audio file."
39
 
40
- # Validate the file first
41
- is_valid, message = validate_file(audio_file)
42
- if not is_valid:
43
- return message
 
 
44
 
45
  try:
46
- # Load audio file
47
- speech_array, sampling_rate = librosa.load(audio_file, sr=16000)
48
-
49
- # Process the audio file
50
- input_features = processor(speech_array, sampling_rate=16000, return_tensors="pt").input_features
51
 
52
- # Generate token ids
53
- predicted_ids = model.generate(input_features)
54
-
55
- # Decode token ids to text
56
- transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True)[0]
57
-
58
- return transcription
59
 
 
 
60
  except Exception as e:
61
- return f"An error occurred during transcription: {str(e)}"
62
 
63
- # Create Gradio interface
64
- with gr.Blocks(theme=theme) as demo:
65
- gr.Markdown("# Audio Transcription with Whisper")
66
- gr.Markdown("Upload an audio file (.mp3 or .wav) of maximum 25MB to get the transcription.")
67
-
68
- with gr.Row():
69
- with gr.Column():
70
- # Fixed: Use sources parameter instead of type
71
- audio_input = gr.Audio(sources=["upload"], label="Upload Audio File")
72
- submit_btn = gr.Button("Transcribe", variant="primary")
73
-
74
- with gr.Column():
75
- output = gr.Textbox(label="Transcription Result", lines=10)
76
-
77
- submit_btn.click(fn=transcribe_audio, inputs=audio_input, outputs=output)
78
-
79
- gr.Markdown("### Limitations")
80
- gr.Markdown("- Maximum file size: 25 MB")
81
- gr.Markdown("- Supported formats: .mp3 and .wav")
82
- gr.Markdown("- Uses the Whisper base model which works best with clear audio")
83
 
84
- # Launch the app
85
  if __name__ == "__main__":
86
- demo.launch()
 
1
  import gradio as gr
2
+ import whisper
3
  import os
 
 
 
4
 
5
+ model = whisper.load_model("base")
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
6
 
7
  def transcribe_audio(audio_file):
8
+ # Check if file is uploaded
9
  if audio_file is None:
10
+ return "Error: Please upload an audio file.", None
11
 
12
+ # Get the file path - in newer Gradio versions, audio_file might be a string path directly
13
+ file_path = audio_file if isinstance(audio_file, str) else audio_file.name
14
+
15
+ # Check file size (25MB limit)
16
+ if os.path.getsize(file_path) > 25 * 1024 * 1024:
17
+ return "Error: File size exceeds 25MB limit.", None
18
 
19
  try:
20
+ result = model.transcribe(file_path)
21
+ output_filename = os.path.splitext(os.path.basename(file_path))[0] + ".txt"
 
 
 
22
 
23
+ with open(output_filename, "w") as text_file:
24
+ text_file.write(result["text"])
 
 
 
 
 
25
 
26
+ return result["text"], output_filename
27
+
28
  except Exception as e:
29
+ return f"Error during transcription: {str(e)}", None
30
 
31
+ iface = gr.Interface(
32
+ fn=transcribe_audio,
33
+ inputs=gr.File(label="Upload Audio File (Max 25MB)", file_types=["audio"]),
34
+ outputs=[
35
+ gr.Textbox(label="Transcription"),
36
+ gr.File(label="Download Transcript")
37
+ ],
38
+ title="Free Transcript Maker",
39
+ description="Upload an audio file (WAV, MP3, etc.) up to 25MB to get its transcription. The transcript will be displayed and available for download. Please use responsibly."
40
+ )
 
 
 
 
 
 
 
 
 
 
41
 
 
42
  if __name__ == "__main__":
43
+ iface.launch()