Ping

Dear authors, Since the paper has been accepted, can the code be released?

> 好，那么moco v3用这个来训练对吗，谢谢 https://github.com/facebookresearch/moco-v3 你好！ 可以分享一下训练moco-v3的经验吗？我用小数据训练总是timeout，查了很久也不知道怎么处理。我用torch 1.11, torchvision 0.12, 双卡训练的 @gzhuinjune

就是submit job之后就不动了，一直是下面这个状态： => creating model 'vit_small' => creating model 'vit_small' 过了那个时间之后就提示NCCL错误： Watchdog caught collective operation timeout: Some NCCL operations have failed or timed out. Due to the asynchronous nature of...

这个moco-v3 repo 说只支持多卡，我用单卡会报错：In this repo, only multi-gpu, DistributedDataParallel training is supported; single-gpu or DataParallel training is not supported. This code is improved to better suit the multi-node setting, and by...

official repo推荐的是CUDA 10.2 PyTorch 1.9.0， 但是需要设置master address, 您可以告诉我怎么设置吗？ 我用的学校的服务器，不知道这个dist-url，我就用的torch 1.11 torchrun ```python main_moco.py \ -a vit_base \ --optimizer=adamw --lr=1.5e-4 --weight-decay=.1 \ --epochs=300 --warmup-epochs=40 \ --stop-grad-conv1 --moco-m-cos --moco-t=.2 \ --dist-url 'tcp://[your...

谢谢回复！ 我刚发现用旧版本跑成功了，需要设置正确的master address和port。