Fallback to CPU to test build

Author: exlab-code

Dockerfile

@@ -1,8 +1,11 @@
-FROM runpod/pytorch:2.2.0-py3.10-cuda12.1.1-devel-ubuntu22.04
+FROM runpod/pytorch:2.1.0-py3.10-cuda11.8.0-devel-ubuntu22.04
 
 WORKDIR /
 
-# The base image already includes CUDA 12.1 and cuDNN - no additional installs needed
+# Install system dependencies
+RUN apt-get update && apt-get install -y --no-install-recommends \
+ wget \
+ && rm -rf /var/lib/apt/lists/*
 
 # Install dependencies
 COPY requirements.txt .
@@ -11,7 +14,7 @@ RUN pip install --no-cache-dir -r requirements.txt
 # Copy handler
 COPY rp_handler.py .
 
-# cuDNN and CUDA libraries are already configured in the base image
+# Use CPU for now to test if build works at all
 
 # Start the handler
 CMD python3 -u rp_handler.py

rp_handler.py

@@ -66,16 +66,16 @@ def load_whisper_model():
  model_size = os.getenv("WHISPER_MODEL", "medium")
  compute_type = os.getenv("WHISPER_COMPUTE_TYPE", "float16")
 
- logger.info(f"🚀 Loading faster-whisper model: {model_size} with {compute_type} precision on GPU")
+ logger.info(f"🚀 Loading faster-whisper model: {model_size} with CPU for build testing")
 
- # GPU-only for performance testing - cuDNN libraries should now be available via LD_LIBRARY_PATH
+ # Try CPU first to test if build works at all
  whisper_model = WhisperModel(
  model_size, 
- device="cuda",
- compute_type=compute_type,
- cpu_threads=1
+ device="cpu",
+ compute_type="int8",
+ cpu_threads=4
  )
- logger.info("✅ GPU model loaded successfully - REAL GPU PERFORMANCE! 🚀")
+ logger.info("✅ CPU model loaded successfully - testing build")
 
  except Exception as e:
  logger.error(f"❌ Failed to load whisper model: {e}")
@@ -184,7 +184,7 @@ def handler(job):
  "real_time_factor": rtf,
  "model": os.getenv("WHISPER_MODEL", "medium"),
  "compute_type": os.getenv("WHISPER_COMPUTE_TYPE", "float16"),
- "device": "cuda",
+ "device": "cpu",
  "speakers_detected": len(set(seg["speaker"] for seg in cleaned_segments)),
  "segments_count": len(cleaned_segments),
  "serverless": True