My flask app just stops when it starts loading my torch model

1

Hello,
I have saved my Pytorch model as a state_dict, and am now trying to load it into flask for inference.
However when I run my flask API, it just stopped. And using my debugger got me here.

Screenshot from 2021-04-27 15-53-22
Screenshot from 2021-04-27 15-53-22727×230 22.8 KB

The same model loads in my notebook. Why is this happening? Any help would be great!

1 Like
2

you can check this Deploying PyTorch in Python via a REST API with Flask — PyTorch Tutorials 1.8.1+cu102 documentation

3

Hey, It doesn’t seem to be a code problem. The flask app used to work on my old system, but doesn’t on my new one. Same code in both places.

4

please provide more code in you utils.py and pytorch_model.py

5

Sure
this is from pytorch_model.py

from torch import nn 
import torch.nn.functional as F



class Model(nn.Module):
    def __init__(self):
        super(Model, self).__init__()
        self.linearA = nn.Linear(13,100)
        self.linearB = nn.Linear(100, 50)
        self.linearC = nn.Linear(50, 1)

    def forward(self, x):
        yA = F.relu(self.linearA(x))
        yB = F.relu(self.linearB(yA))
        return self.linearC(yB)

and this is the part from utils.py

data.model = Model()
data.model.load_state_dict(torch.load(data.model_path))