Julian de Wit

Results 68 comments of Julian de Wit

Hello I just saw them as different labels. During training it would take the average.. Bigger problem might be nodules that were labeled by some doctors and not by others....

hello, I noticed that newer keras versions have different parameters for the "fit" function. Where I gave the number of training examples as parameter, I believe it's now the number...

It's best to take a look ate the sourcecode.

Kaggle dropped te data. But you can still train the model.. You just cannot predict.

using scikit learn

Hello, I think tensorflow sees no way to distribute the network over multiple GPU's. Although in theory it should be smart enough to split the batch in 2 parts en...

Indeed I also had to look twice after this time. The level 1 models are combined into level2 folder. The models in level2 are combined into the submission folder.

step2_train_mass_segmenter.py Also has a predict phase. This one will generate this file. You can also leave it out. It will not change the score very much.

Hello, I believe all in one map. Regards, Julian.