ml-cvnets
ml-cvnets copied to clipboard
apple
How to solve this problem:ModuleNotFoundError: No module named 'cvnets.models.classification.'
Hi@sacmehta @mchorton
When I comment out "self.check_model()" in "./ml-cvnets/cvnets/models/classification/mobilevit.py",and run command as figure1,I got a error as figure2.
I uncomment "self.check_model()",run command again,but also got this error.
How to solve this problem?
Thank you very much!
@acrlife it looks like python cannot find the module - is CVNets missing from your pythonpath? Assuming you're running from the root directory in the CVNets repository, try adding PYTHONPATH=. before your command.
@mchorton I have found reason of this problem: When I installed cvnets ,command cvnets-train can works, but it does't work after I add my own model file named "denseformer.py" to "ml-cvnets/cvnets/models/classification". I have try to reinstall cvnets by pip install --editable ., but it also does't work.
Figure 1 shows the situation after just installing CVnets, and Figure 2 shows the situation after adding my own model file.
I really want to know how to train my own model using CVNets,thank you!
The error implies you're trying to import cvnets.models.classification.config somewhere. Since the config folder isn't a module, you're getting an import error. I think you meant to remove the config from your import statement.
@mchorton I think CVNets has built a fully functional framework for the main tasks in the field of computer vision. I would like to know how to use CVNets to train my own model. Can you provide a tutorial? Thank you very much!
@acrlife we have documentation here: https://apple.github.io/ml-cvnets/getting_started.html
I believe your issue involved importing the wrong directory, as mentioned above. If you are still having trouble, try this:
- Revert all your code changes, so that your branch matches the
mainbranch. - Add your model definition to an existing .py file in which some other model is defined. That way, you can rule out an import error.