search

uber-research / UPSNet

UPSNet: A Unified Panoptic Segmentation Network
Other
645 stars 120 forks source link

ImportError: cannot import name 'deform_conv_cuda' #105

Closed StoicGilgamesh closed 4 years ago

StoicGilgamesh commented 4 years ago

First of all thanks for this great research and the repository! Very helpful.

I have followed the instructions but when I try to train or test using pre-trained weights I get the following error. Could you please help?

upsnet/../upsnet/config/config.py:180: YAMLLoadWarning: calling yaml.load() without Loader=... is deprecated, as the default Loader is unsafe. Please read https://msg.pyyaml.org/load for full details. exp_config = edict(yaml.load(f)) Traceback (most recent call last): File "upsnet/upsnet_end2end_test.py", line 44, in from upsnet.models import * File "upsnet/../upsnet/models/init.py", line 1, in from .resnet_upsnet import resnet_50_upsnet, resnet_101_upsnet File "upsnet/../upsnet/models/resnet_upsnet.py", line 22, in from upsnet.models.resnet import get_params, resnet_rcnn, ResNetBackbone File "upsnet/../upsnet/models/resnet.py", line 21, in from upsnet.operators.modules.deform_conv import DeformConv File "upsnet/../upsnet/operators/modules/deform_conv.py", line 22, in from upsnet.operators.functions.deform_conv import DeformConvFunction File "upsnet/../upsnet/operators/functions/deform_conv.py", line 21, in from .._ext.deform_conv import deform_conv_cuda ImportError: cannot import name 'deform_conv_cuda' (/scratch_net/biwidl204/hgupta/anaconda/conda_envs/py36

SunNYNO1 commented 4 years ago

I meet the same problem，do you solve this problem？ could you help me？

StoicGilgamesh commented 4 years ago

For me, the problem was with my Cuda settings and PyTorch version on the cluster. PyTorch was not able to recognize GPUS.

SunNYNO1 commented 4 years ago

thank you very much, but I find the github dont have the deform_conv_cuda.py under the path, do you know why, thanks again ------------------ 原始邮件 ------------------ 发件人: "Hrishikesh Gupta"notifications@github.com 发送时间: 2019年11月19日(星期二) 晚上6:18 收件人: "uber-research/UPSNet"UPSNet@noreply.github.com; 抄送: "sun"1246680611@qq.com;"Comment"comment@noreply.github.com; 主题: Re: [uber-research/UPSNet] ImportError: cannot import name'deform_conv_cuda' (#105)

For me, the problem was with my Cuda settings and PyTorch version on the cluster. PyTorch was not able to recognize GPUS.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

moshuilanting commented 2 years ago

init.sh 编译的时候应该就存在问题，进到operators重新编译build_deform_conv.py

cd ../operators python build_deform_conv.py build_ext --inplace

我会报这个错误 "error: ‘THCState_getCurrentStream’ was not declared in this scope" 找到对应行把“THCState_getCurrentStream(state)”都替换为c10::cuda::getCurrentCUDAStream()，这个问题就没了

Sujan-PhdWork commented 4 months ago

this configuration worked for me conda install pytorch==1.1.0 torchvision==0.3.0 cudatoolkit=9.0 -c pytorch