Hey All - I was working on my image classifier and wanted to experiment with different models rather than just resnet34/50. I did some investigation and wanted to see if how Googlenet would perform.
The resnet34/50 models we are using are coming directly from torchvision and the fast.ai docs clearly state * On top of the models offered by torchvision, the fastai library has implementations for the following models:*
So I tried to implement the Googlenet model in the same way I would use the resnet34 model learn = cnn_learner(data, models.resnet34, metrics=error_rate)
and I get the error 'module module ‘fastai.vision.models’ has no attribute ‘googlenet’
I took at the look at the docs for the models module and I noticed that the googlenet is not imported from torchvision https://github.com/fastai/fastai/blob/master/fastai/vision/models/init.py
I know i could just import the googlenet myself or make an update to the models module to add in the googlenet but I am curious as to WHY it’s not included? Are LeNets dated and will never perform as well as a restnet?
Also - please let me know if there is a better way to have this kind of conversation, if there is a place to chat about the fast.ai library design I would love to jump in on that conversation.
Thanks!