Yup works perfectly, Thank youu
Thanks @jeremy for the tip about torch and torchvision.
I’m working on a Google Cloud VM with code that requires an older fast.ai version (1.0.54) due to a call we use changing in 1.0.55+. This might be useful for people requiring a repeatable install.
To meet all of the requirements for v1 I’ve found the following work - note explicitly specifying the torch version due to changing support within fastai.
conda install -c pytorch -c fastai fastai pytorch=1.1.0 pip install fastai torch==1.1.0 torchvision git clone https://github.com/fastai/fastai cd fastai git checkout release-1.0.54 # to specify version as noted above tools/run-after-git-clone pip install -e ".[dev]"
Able to happily run dogs_cats.ipynb after this (as well as my code).