Arlind Kadra
## Describe the bug After training the UIS-RNN algorithm, I call the predict() method with the test_sequences. In this case, the test sequence is a 2-d numpy array, not a...
As the title says, BOHB seems to be counting constants that are given in the Configuration Space for the moment it switches from using random search to using a model....
I was a bit surprised, but I actually did not see anywhere the implementation for SGD with decoupled weight decay. If it was, you can just reject this PR.
https://github.com/automl/Auto-PyTorch/blob/0cc6b4696abcd2e912ebd4fc9a6c283ef5607b2e/autoPyTorch/pipeline/components/training/trainer/base_trainer.py#L229 The class documentation should be updated. As I checked, the module also is in need of a documentation update.
Add unit tests where we restrict the search space as follows: - Skip Connections are disabled. - Snapshot Ensembling is disabled. - Weight Decay is disabled. - Dropout is disabled....
I have a question regarding dropout and how it is applied. In the previous non-refactored version of AutoPyTorch there was a `dropout_shape` hyperparameter, which does not seem to be there...
Currently, there are a few problems that crash the library when working with imbalanced datasets. - There are Null values and one unique value in the train split and Null...
There is currently an extra layer at the end of the architecture. This does not figure on the `master `branch or here https://github.com/pytorch/vision/blob/master/torchvision/models/resnet.py
https://github.com/automl/Auto-PyTorch/blob/a4e08e297dc298b53d5962f0f4b4e55b5e9ded8c/autoPyTorch/api/base_task.py#L430-L435 The second if check at Line 433 is not needed. We can just continue with the statements under the first.
I am calling the function `detach_from_suite ` to remove a task (**211981**) from a benchmark suite (**240**) that was created by me. The server returns a status code of 200,...