mil

mil copied to clipboard

I am trying to dump a model and its metrics to the disk using pickle. You can see the [Metrics](https://github.com/rosasalberto/mil/blob/master/mil/metrics/custom.py#L7) class here. Is there any fix for this issue?

nam1410

Resulted in large increase in performance. Also added distance metric configuration for application-specific performance options

AaronMcDaniel

used os.path.join to make referenced files work regardless of OS

AaronMcDaniel

I would like to install this package from source but there is no setup.py file. Is there a simple way to install this package from source for development?

AaronMcDaniel

``` (bags_train, y_train), (bags_test, y_test) = musk1.load() trainer = Trainer() metrics = [AUC, BinaryAccuracy] model = AttentionDeepPoolingMil(gated=False, threshold=0.4) pipeline = [('scale', StandarizerBagsList()), ('padding', Padding())] trainer.prepare(model, preprocess_pipeline=pipeline ,metrics=metrics) valid = KFold(n_splits=10,...

AaronMcDaniel

Error loading data when installed with pip

1

``` >>> from mil.data.datasets import musk1 >>> (bags_train, y_train), (bags_test, y_test) = musk1.load() Traceback (most recent call last): File "", line 1, in File "/usr/local/lib/python3.7/site-packages/mil/data/datasets/musk1.py", line 7, in load return...

jasonzutty

Hi, Great Job! I am looking for multi-classification MIL. Could you please guide me how to use your package for my case? thanks!

farnoush-shh

Hey there, I'm having some trouble with the library and sorting out the dependencies. According to the docs, I need to install these: ``` pip install numpy pip install scikit-learn...

GrunCrow

In other words, would pull requests get accepted or should the project get cloned in order to get updated instead?

johann-petrak