Commit 985ca2a

author

Stefan Knegt

authored

Update README.md

1 parent 4dfa923 commit 985ca2aCopy full SHA for 985ca2a

File tree

-1

lines changed

-1

lines changed

Lines changed: 5 additions & 1 deletion

Original file line number	Diff line number	Diff line change
`@@ -11,7 +11,7 @@ def _kl_independent_independent(p, q):`
`11`	`11`	`result = kl_divergence(p.base_dist, q.base_dist)`
`12`	`12`	`return _sum_rightmost(result, p.reinterpreted_batch_ndims)`
`13`	`13`	```
`14`		`-`
	`14`	`+`
`15`	`15`	`## Training`
`16`	`16`	`In order to train your own Probabilistic UNet in PyTorch, you should first write your own data loader. Then you can use the following code snippet to train the network`
`17`	`17`
`@@ -33,3 +33,7 @@ for epoch in range(epochs):`
`33`	`33`	`loss.backward()`
`34`	`34`	`optimizer.step()`
`35`	`35`	```
	`36`	`+`
	`37`	`+## Train on LIDC Dataset`
	`38`	`+One of the datasets used in the original paper is the LIDC dataset (https://wiki.cancerimagingarchive.net). I've preprocessed this data and stored them in 5 .pickle files which you can [download here](https://drive.google.com/drive/folders/1xKfKCQo8qa6SAr3u7qWNtQjIphIrvmd5?usp=sharing). After downloading the files you need to adjust the path in the data loader and you can start training your own Probabilistic UNet using the code snippet above.`
	`39`	`+`

Comments

(0)