search

VITA-Group / FasterSeg

[ICLR 2020] "FasterSeg: Searching for Faster Real-time Semantic Segmentation" by Wuyang Chen, Xinyu Gong, Xianming Liu, Qian Zhang, Yuan Li, Zhangyang Wang
MIT License
525 stars 107 forks source link

There are deprecations in torch 1.1.0 #31

Closed 18022443868 closed 4 years ago

18022443868 commented 4 years ago

Traceback (most recent call last): File "E:/vSLAM/FasterSeg-master/train/train.py", line 28, in from eval import SegEvaluator File "E:\vSLAM\FasterSeg-master\train\eval.py", line 9, in from engine.evaluator import Evaluator File "E:\vSLAM\FasterSeg-master\tools\engine\evaluator.py", line 11, in from utils.pyt_utils import load_model, link_file, ensure_dir File "E:\vSLAM\FasterSeg-master\tools\utils\pyt_utils.py", line 23, in def reduce_tensor(tensor, dst=0, op=dist.ReduceOp.SUM, world_size=1): AttributeError: module 'torch.distributed' has no attribute 'ReduceOp'

pytorh=1.1.0 windows10

egedib commented 4 years ago

If you want to evaluate/test only, you can comment those lines (23-37) out.

It doesn't solve your problem if you want to train though.

18022443868 commented 4 years ago

How to train on WIN10 single GPU?

chenwydj commented 4 years ago

Hi @18022443868!

Thank you for your interest in our work!

As mentioned in our Readme, we only test our code in Ubuntu environment. We did not test on Windows.