Recording

Storing data

xmean,xstd = 0.28, 0.35
@inplace
def transformi(b): b['image'] = [(TF.to_tensor(o)-xmean)/xstd for o in b['image']]

_dataset = load_dataset('fashion_mnist').with_transform(transformi)

dls = DataLoaders.from_dataset_dict(_dataset, 256, num_workers=4)

---

source

CoreCBs

 CoreCBs (device='cpu', module_filter=<function noop>, **metrics)

Initialize self. See help(type(self)) for accurate signature.

trainer = Trainer(dls,
                  nn.CrossEntropyLoss(), 
                  torch.optim.Adam, 
                  get_model_conv(),
                  callbacks=[CoreCBs(Accuracy=MulticlassAccuracy()),OneBatchCB()])

trainer.fit()

	train	valid	Accuracy
0	2.3023	2.293682	0.203776

	train	valid	Accuracy
1	2.288657	2.277659	0.448568

	train	valid	Accuracy
2	2.265343	2.246487	0.443359