Zhenwei
Zhenwei
I am running the reproducing experiment of lxmert pretraining. I wish that you can provide a log file, so I can find problems at early epochs.
It is too slow to do the face detection and alignment when dataset is large if process images one by one. So I wander if it is possible to change...
At line 215-219 of `deformable_attn.py`, the original code is ``` Python # B*M, H*W, C_v feat = torch.einsum('nlds, nls -> nld', scale_features, A) # B, H, W, C feat =...
I wrote following code to do a words sort task ```Python #!/usr/bin/python # -*- coding: utf-8 -*- from mrjob.job import MRJob import re class MRwordCount(MRJob): def mapper(self, in_key, in_value): bins...