齐晨旭 comments

Results 5 comments of


                                            齐晨旭

训练中途loss变nan

> 将工作人员信息更改为 > > samples_per_gpu=5， workers_per_gpu=2， > > 我一直在使用 fp16 > > 我遇到了另一种类型的错误， `ERROR:torch.distributed.elastic.agent.server.local_elastic_agent:[default] Worker group failed` 我仍然很困惑到底是什么问题。 > > 我将此培训与 2 个 gpu 一起使用 `bash tools/dist_train_partially_faster_rcnn50_att_90_4classes.sh semi 1 10...

有关结果与论文不一致，还有数据集问题

您好，我想咨询一下FCOS*(P2)是指对FCOS进行了什么特殊处理？

[Err]: RuntimeError: Default process group has not been initialized, please make sure to call init_process_group.

modify the config RESNETS: NORM: "SyncBN"->''BN''

Do you think you'll ever migrate your code to the latest stable versions of mmcv and mmdetection?

Hi, I used the torch1.12.1 cuda11.6 but got the erro: Traceback (most recent call last): File "tools/zero_shot_utils.py", line 3, in from mmdet.datasets import DATASETS, build_dataloader File "/home/dell/exp/qcx/ZSD-SC-Resolver-main/mmdetection/mmdet/datasets/__init__.py", line 2, in...

Do you think you'll ever migrate your code to the latest stable versions of mmcv and mmdetection?

My environment settings are as follows：nvcc: NVIDIA (R) Cuda compiler driver Copyright (c) 2005-2021 NVIDIA Corporation Built on Thu_Nov_18_09:45:30_PST_2021 Cuda compilation tools, release 11.5, V11.5.119 Build cuda_11.5.r11.5/compiler.30672275_0 # packages in...