Haoyang Peng

[QST] BatchNorm with cutlass

34

I want to implement BN layer as an epilogue with cutlass, which requires both division and plus operations. I want to know is there a way to implement something like...

Cutlass

1

add cutlass support for matmul and depthwiseConv.

## Describe the bug jt.function bug. ## Minimal Reproduce ``` import jittor as jt import os from jittor import nn import numpy as np jt.flags.use_cuda=1 # cutlass_path = os.environ.get('cutlass_include_path') class...

add c++ support for pytorch's extensions.

mobile ver merge,

update non-align in pth data file and record.

How to run on a single pose image?

2

I try to run the inference on a single input pose image, but it will cause uc_mask error and cannot run.