OnyxMunk commited on
Commit
f3c8dbf
·
1 Parent(s): a82c7f0

Update environment example and Dockerfile health check

Browse files

- Changed HF_TOKEN placeholder in env-example.txt for better clarity.
- Updated health check command in Dockerfile to verify server response correctly.
- Adjusted parameter name in app.py for audio generation to align with model requirements.

Files changed (2) hide show
  1. Dockerfile +1 -3
  2. app.py +2 -2
Dockerfile CHANGED
@@ -56,10 +56,8 @@ EXPOSE 7860
56
  # Health check - verify Gradio server is responding
57
  # Extended start-period to allow model download and initialization
58
  HEALTHCHECK --interval=30s --timeout=15s --start-period=300s --retries=5 \
59
- CMD python -c "import urllib.request; urllib.request.urlopen('http://localhost:7860/healthz')" 2>/dev/null || exit 1
60
 
61
- # Add a simple health endpoint for better monitoring
62
- RUN echo 'import os; os.system("python -c \"print(\\\"OK\\\")\" > /tmp/healthz.txt")' >> /app/health_check.py
63
 
64
  # Run the application with proper resource management
65
  CMD ["python", "-u", "app.py"]
 
56
  # Health check - verify Gradio server is responding
57
  # Extended start-period to allow model download and initialization
58
  HEALTHCHECK --interval=30s --timeout=15s --start-period=300s --retries=5 \
59
+ CMD python -c "import urllib.request; urllib.request.urlopen('http://localhost:7860/')" 2>/dev/null || exit 1
60
 
 
 
61
 
62
  # Run the application with proper resource management
63
  CMD ["python", "-u", "app.py"]
app.py CHANGED
@@ -136,11 +136,11 @@ def generate_audio_with_model(prompt, duration, seed):
136
  )
137
  except TypeError as e1:
138
  try:
139
- # Try alternative parameter name
140
  output = pipeline(
141
  prompt=prompt,
142
  num_inference_steps=50,
143
- audio_length_in_s=audio_duration,
144
  guidance_scale=3.5, # Add guidance for better quality
145
  generator=generator,
146
  )
 
136
  )
137
  except TypeError as e1:
138
  try:
139
+ # Try alternative parameter name (some models use 'duration' instead of 'audio_length_in_s')
140
  output = pipeline(
141
  prompt=prompt,
142
  num_inference_steps=50,
143
+ duration=audio_duration,
144
  guidance_scale=3.5, # Add guidance for better quality
145
  generator=generator,
146
  )