RuntimeError: Numpy is not available when running Streamlit + PyTorch + Torchvision app on Streamlit Cloud

Question

`I’m deploying a Streamlit app on Streamlit Cloud that uses PyTorch and Torchvision for image segmentation.

Locally, everything works fine, but when I deploy on Streamlit Cloud I get this error:

File "/home/adminuser/venv/lib/python3.9/site-packages/torchvision/transforms/functional.py", line 166, in to_tensor
    img = torch.from_numpy(np.array(pic, mode_to_nptype.get(pic.mode, np.uint8), copy=True))
RuntimeError: Numpy is not available

Repository link : https://github.com/Khushi-51/AI_SpillGuard_OSD-Khushi

Minimal example that fails on Streamlit Cloud:

import torch
import numpy as np

print("Torch sees NumPy:", torch.tensor(\[1.0, 2.0\]).numpy())

What I tried

My requirements.txt originally looked like this:

torch==2.0.1
torchvision>=0.15.2
numpy>=1.21.0
pillow>=8.3.0
gdown>=4.7.1
opencv-python>=4.5.0
...

I already added NumPy to requirements.txt.
Tried reordering so NumPy is installed before torch.
Cleared .venv and cached builds in Streamlit Cloud.
Tried Rebuild Environment from Streamlit Cloud settings.
Works perfectly locally with streamlit run app.py.
Still, the deployed app throws RuntimeError: Numpy is not available`

Santhosh S · Accepted Answer · 2025-10-01 10:57:02Z

0

Place numpy at top with specific version
Then don't use == for numpy and torchvision
PyTorch 2.0.1 works with Numpy 1.21.x - 1.24.x

Now, requirement.txt should look like

numpy==1.24.0
torch==2.0.1
torchvision==0.15.2
....

answered Oct 1 at 10:57

Santhosh S

17 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

RuntimeError: Numpy is not available when running Streamlit + PyTorch + Torchvision app on Streamlit Cloud

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related