Pavlo Molchanov
Pavlo Molchanov
Answering your questions: 1. For these experiments we generated 1000 batches of batch size 256 in total. DI/ADI generate the entire batch of data at once. Most likely hyper parameters...
We have restrictions on providing trained models and this complicates full code release. Most likely we will release experiment for 2. Data-free Knowledge Transfer soon and will try to do...
Hi, the temperature used was set to the default value of 3 and never changed. We will take into consideration the need to share details on KD training for Cifar10...