fool

Results 4 comments of fool

How to use Multi-Head Attention in ViT

dim(e.g., 1024) is the final output dimsion of multi attention module. dim_head = dim // head_num, when head_num = 1, dim_head is equal to dim. b: batch size n: num_patches...

import resample2d_cuda ImportError

# 3. Build pytorch Custom CUDA Extensions ./setup.sh

The pose landmarks and heatmaps of Market1501

Thank you very much

partial ReID datasets

sorry, I modified the code and converted python2 to python3. But the code still exists many issues, So I give up tring the code.