search

alankbi / detecto

Build fully-functioning computer vision models with PyTorch
https://detecto.readthedocs.io/
MIT License
612 stars 107 forks source link

Confusion with workaround for using a non-locally trained pth file for model_weights #126

Open pranjshrivastava opened 1 year ago

pranjshrivastava commented 1 year ago

Hi Alan, I have used your Detecto framework code frequently on Google Colaboratory and have always successfully trained and used it for detection applications Colab. Recently I tried to train it on Google Colab and downloaded the model_weights.pth file. I used this file to run the detection locally on a Jetson Nano by Nvidia.

I faced the following Runtime error: Error(s) in loading state_dict for FasterRCNN

I read through the previous issues and the workaround that you provided for buggy load function. However, you seem to have given model_name for a different model.

I'll be grateful if you could help me fix my problem and what model_name i can use for the workaround of load function.

SutirthaChakraborty commented 1 year ago

Did you find any solution ? @pranjshrivastava