search

ej0cl6 / deep-active-learning

Deep Active Learning
MIT License
783 stars 182 forks source link

Mapping between paper names to python class #11

Closed kobybibas closed 2 years ago

kobybibas commented 2 years ago

Hey, Thanks for this awesome repository. A quick question, is there a mapping between the paper names to the python code classes? Specifically I'm interested in ''ACTIVE LEARNING FOR CONVOLUTIONAL NEURAL NETWORKS: A CORE-SET APPROACH`` which supposes to be one of these: "RandomSampling", "LeastConfidence", "MarginSampling", "EntropySampling", "LeastConfidenceDropout", "MarginSamplingDropout", "EntropySamplingDropout", "KMeansSampling", "KCenterGreedy", "BALDDropout", "AdversarialBIM", "AdversarialDeepFool",

ej0cl6 commented 2 years ago

Hi, thanks for your interest. Currently we have approximated version of CoreSet (k-means and k-centers). The main reason for the original CoreSet is that it requires an external solver which is not free and it's hard to combine the external solver with the current code smoothly (many people asked about the usage). So I decided to remove it from the package. However, you can still find the implementation by the original authors here: https://github.com/ozansener/active_learning_coreset Or you can check the old commit in this repository. I believe it's till there, but the overall training framework now is a bit different.

kobybibas commented 2 years ago

Ok, thanks a lot!