Replies: 1 comment 1 reply
-
|
Figured out a way yet? |
Beta Was this translation helpful? Give feedback.
1 reply
-
|
Not yet @shruts37 |
Beta Was this translation helpful? Give feedback.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
I have a Detectron2 pre-trained model resnet50.pkl that was originally trained on the COCO 2017 dataset. Now, I want to make it a custom pre-trained model by further training it on my own dataset.What i did till now is i trained as a proof of concept trained model on colab on balloon dataset exported model .pth format convert it to .pkl using convert-torchvision-to-d2.py however i got an error so replace the lines
newmodel[k] = obj.pop(old_k).detach().numpy() with newmodel[k] = obj.pop(old_k)however it could not train a model using this. Any advice or suggestions would be highly appreciated!Beta Was this translation helpful? Give feedback.
All reactions