R
RunPod•2w ago
selectedpixel

Error loading ASGI app. Could not import module "main".

I'm trying to run a Fast API app on a Pod. I've followed this tutorial https://www.runpod.io/articles/guides/deploy-fastapi-applications-gpu-cloud My Docker Image runs locally fine. The container gets created fine on RunPod. However I gett an error message in the Container Logs. 
Error loading ASGI app. Could not import module "main".
Error loading ASGI app. Could not import module "main".
If I attempt to override the start CMD then I just get error messages saying a file can't be found, even if I enter the exact same CMD that is in the docker file. 
5/29/2025, 6:43:14 AM
[FATAL tini (31)] exec CMD failed: No such file or directory
5/29/2025, 6:43:14 AM
[FATAL tini (31)] exec CMD failed: No such file or directory
This is my Docker FIle 
# Use an official PyTorch image with CUDA support as base (contains Python, CUDA, PyTorch)
FROM pytorch/pytorch:2.0.1-cuda11.7-cudnn8-runtime

RUN pip install --upgrade pip

# Install the Python requirements
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt

# Copy application code
WORKDIR /app
COPY app/ /app/


ENV HF_HOME="/tmp/huggingface/hub"
ENV TOKENIZERS_PARALLELISM=false
ENV TORCH_HOME="app/torch"

# Expose port 8000 (the port our FastAPI app will run on)
EXPOSE 8000

# Specify the entrypoint to run the app with Uvicorn (listening on all interfaces at port 8000)
CMD ["uvicorn", "api:app", "--host", "0.0.0.0", "--port", "8000"]
# Use an official PyTorch image with CUDA support as base (contains Python, CUDA, PyTorch)
FROM pytorch/pytorch:2.0.1-cuda11.7-cudnn8-runtime

RUN pip install --upgrade pip

# Install the Python requirements
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt

# Copy application code
WORKDIR /app
COPY app/ /app/


ENV HF_HOME="/tmp/huggingface/hub"
ENV TOKENIZERS_PARALLELISM=false
ENV TORCH_HOME="app/torch"

# Expose port 8000 (the port our FastAPI app will run on)
EXPOSE 8000

# Specify the entrypoint to run the app with Uvicorn (listening on all interfaces at port 8000)
CMD ["uvicorn", "api:app", "--host", "0.0.0.0", "--port", "8000"]
How to Deploy FastAPI Applications with GPU Access in the Cloud
Deploying a FastAPI application that uses GPU acceleration (e.g. with PyTorch) can be achieved by containerizing the app and running it on a cloud GPU...
4 Replies
Jason
Jason•2w ago
maybe its the way you import things in python, where is that error coming from? which file
selectedpixel
selectedpixelOP•2w ago
Actually I think the problem was that my changes weren't pushed to Docker Hub correctly.
Jason
Jason•2w ago
Did you tried to use same tags? Try other tag
selectedpixel
selectedpixelOP•2w ago
The problem was just that I wasn't retagging the local container. It's working now. Thank you

Did you find this page helpful?