I’m trying to deploy a tabular model, I’m exporting the model to be used on new data.
Saving the model and weights using:
learn.export('export.pkl')
Loading a learner using:
learn = load_learner(fname='export.pkl', cpu=False)
The problem is that load_learner returns a ‘collections.OrderedDict’ instead of a tabular_learner
Am I using the correct learn.export load_learner?
can you link a notebook for further debugging?
I am unable to reproduce the error in Google Colab, however it still persists in the local environment:
Also all the versions of fastai, fastcore, torch and torchvision are the same
EDIT: I created a seperate post for this.
Hi there
It seems that I’m not allowed to create posts so I’ll hijack this post since the issue is similar.
I created a model, exported it and now i’m trying to load it in again.
learn.export(fname=“model.pkl”)
learn = load_learner(“model.pkl”)
However this throws an error:
Does anyone know what the problem could be here?
Thanks in advance!
Here is my installation, I couldnt include it in my previous post since I’m not allowed to include more than one image.
