Bilal Khan

sampling using a single top_k_top_p kernel for all cases

5

Hi, we want to be able to pass in different values of topk/topp for each element in the batch at runtime for sampling, and it would be great if we...

Hi, I didn't see #1 before submitting my pull request #11 , so I'm opening this issue now that I finished my implementation of the squeeze-and-excitation block and the code...

Hi, I added the squeeze and excitation block from the Squeeze-and-Excitation Networks [paper](https://arxiv.org/abs/1709.01507). The code for the SE block is in ` torchcontrib/nn/modules/linear.py` and the tests for it are in...