Don’t like telling you, but you posted results are nothing worth. Your code trains on the training and testing data! This is why you get those increadible results. You should revise you code to only use the train data. Updating the following line, should do the job
dls = dblock.dataloaders(os.path.join(path.__str__(),”train”), bs=128)
Your accuracy should drop dramatically.
Don’t like telling you, but you posted results are nothing worth. Your code trains on the training and testing data! This is why you get those increadible results. You should revise you code to only use the train data. Updating the following line, should do the job
dls = dblock.dataloaders(os.path.join(path.__str__(),”train”), bs=128)
Your accuracy should drop dramatically.
Whoops! Fixed it. Thanks for that.