ptrblck

@seryilmaz it seems your recent change needs a guard for older cublas versions e.g. in https://github.com/NVIDIA/apex/blob/ae1cdd64314e598b935a8138b3532d4b652a8f12/csrc/fused_dense_cuda.cu#L687

I've merged https://github.com/NVIDIA/apex/pull/1162. Could you pull the latest master and retry the build again, please?

@USTClj Could you post a code snippet to reproduce this issue? I tried to reproduce it using `opt_level='O1'`, but it seems to work on my machine using: ```python optimizer =...

I'm not sure to understand your use case completely. Even if I freeze the affine parameters in all batchnorm layers, my code snippet still works. Could you have a look...

Hi @Msabih, do you have a small code snippet to reproduce this issue? I've posted some code in the last post, which is working fine for me. If you don't...

@DannyDannyDanny If no GPU is detected on the system, you won't be able to use `apex`. We should improve the error message on importing `apex` and raise an Exception, if...

Hi @xysong1201, thanks for the PR! You are right, the architecture differs a bit in comparison to the original paper implementation. I'm currently traveling and will have a look at...

CC @nWEIdia for the corresponding `pytorch/builder` PR.

@johnnynunez It's not available as described in the review: https://github.com/pytorch/pytorch/pull/121684#discussion_r1521773989

@Caenorst Did you verify all install commands? E.g. running this in a new and clean environment: ``` conda install pytorch==1.11.0 torchvision==0.12.0 torchaudio==0.11.0 cudatoolkit=11.0 -c conda-forge ``` I get: ``` PackagesNotFoundError:...