mmcv版本问题 about remote-sensing-rvsa HOT 11 OPEN

@libingDY 您好，当前这套代码是完完全全的老版本mmcv，mmdet，mmseg等，老版本上完全可以跑通。由于新老版本改动较大，我没有精力，也没有必要将老版本全部更新为新版本，建议您采用老版本来运行。据我所知，新版本可能已经支持mmcv_custom中的部分功能，您也可以找找看，这样可能就不需要mmcv_custom了。在之后的工作中，我们会采用全套新版本，谢谢关注！

from remote-sensing-rvsa.

感谢您的恢复，我用低版本的mmcv\mmdet进行了代码运行，但是我又遇到了以下问题：
Traceback (most recent call last):
File "tools/train.py", line 153, in
main()
File "tools/train.py", line 142, in main
train_detector(
File "/data4/libing/bisai/OBBDetection/mmdet/apis/train.py", line 133, in train_detector
runner.run(data_loaders, cfg.workflow, cfg.total_epochs)
File "/data4/libing/anaconda3/lib/python3.8/site-packages/mmcv_full-1.7.1-py3.8-linux-x86_64.egg/mmcv/runner/epoch_based_runner.py", line 136, in run
epoch_runner(data_loaders[i], **kwargs)
File "/data4/libing/anaconda3/lib/python3.8/site-packages/mmcv_full-1.7.1-py3.8-linux-x86_64.egg/mmcv/runner/epoch_based_runner.py", line 53, in train
self.run_iter(data_batch, train_mode=True, **kwargs)
File "/data4/libing/anaconda3/lib/python3.8/site-packages/mmcv_full-1.7.1-py3.8-linux-x86_64.egg/mmcv/runner/epoch_based_runner.py", line 31, in run_iter
outputs = self.model.train_step(data_batch, self.optimizer,
File "/data4/libing/anaconda3/lib/python3.8/site-packages/mmcv_full-1.7.1-py3.8-linux-x86_64.egg/mmcv/parallel/data_parallel.py", line 77, in train_step
return self.module.train_step(*inputs[0], **kwargs[0])
File "/data4/libing/bisai/OBBDetection/mmdet/models/detectors/base.py", line 237, in train_step
losses = self(**data)
File "/data4/libing/anaconda3/lib/python3.8/site-packages/torch/nn/modules/module.py", line 889, in _call_impl
result = self.forward(*input, **kwargs)
File "/data4/libing/bisai/OBBDetection/mmdet/core/fp16/decorators.py", line 51, in new_func
return old_func(*args, **kwargs)
File "/data4/libing/bisai/OBBDetection/mmdet/models/detectors/base.py", line 172, in forward
return self.forward_train(img, img_metas, **kwargs)
File "/data4/libing/bisai/OBBDetection/mmdet/models/detectors/obb/obb_two_stage.py", line 154, in forward_train
x = self.extract_feat(img)
File "/data4/libing/bisai/OBBDetection/mmdet/models/detectors/obb/obb_two_stage.py", line 84, in extract_feat
x = self.backbone(img)
File "/data4/libing/anaconda3/lib/python3.8/site-packages/torch/nn/modules/module.py", line 889, in _call_impl
result = self.forward(*input, **kwargs)
File "/data4/libing/bisai/OBBDetection/mmdet/models/backbones/vitae_nc_win_rvsa_wsz7.py", line 777, in forward
x = self.forward_features(x)
File "/data4/libing/bisai/OBBDetection/mmdet/models/backbones/vitae_nc_win_rvsa_wsz7.py", line 764, in forward_features
x = blk(x, Hp, Wp)
File "/data4/libing/anaconda3/lib/python3.8/site-packages/torch/nn/modules/module.py", line 889, in _call_impl
result = self.forward(*input, **kwargs)
File "/data4/libing/bisai/OBBDetection/mmdet/models/backbones/vitae_nc_win_rvsa_wsz7.py", line 540, in forward
convX = self.drop_path(self.PCM(x_2d).permute(0, 2, 3, 1).contiguous().view(b, n, c))
File "/data4/libing/anaconda3/lib/python3.8/site-packages/torch/nn/modules/module.py", line 889, in _call_impl
result = self.forward(*input, **kwargs)
File "/data4/libing/anaconda3/lib/python3.8/site-packages/torch/nn/modules/container.py", line 119, in forward
input = module(input)
File "/data4/libing/anaconda3/lib/python3.8/site-packages/torch/nn/modules/module.py", line 889, in _call_impl
result = self.forward(*input, **kwargs)
File "/data4/libing/anaconda3/lib/python3.8/site-packages/torch/nn/modules/batchnorm.py", line 532, in forward
world_size = torch.distributed.get_world_size(process_group)
File "/data4/libing/anaconda3/lib/python3.8/site-packages/torch/distributed/distributed_c10d.py", line 711, in get_world_size
return _get_group_size(group)
File "/data4/libing/anaconda3/lib/python3.8/site-packages/torch/distributed/distributed_c10d.py", line 263, in _get_group_size
default_pg = _get_default_group()
File "/data4/libing/anaconda3/lib/python3.8/site-packages/torch/distributed/distributed_c10d.py", line 347, in _get_default_group
raise RuntimeError("Default process group has not been initialized, "
RuntimeError: Default process group has not been initialized, please make sure to call init_process_group.
您遇到过相似的问题吗？

from remote-sensing-rvsa.

@libingDY 您好，我没有碰到过，你这是DDP方面的问题吧，我没有改OBBDetection，可能是你命令不对

from remote-sensing-rvsa.

好的，非常感谢您的回复

from remote-sensing-rvsa.

您好，能提供您老版本的安装包版本吗？感谢

from remote-sensing-rvsa.

@zhongyas 如果你说mmcv-full的话，安装的时候可以指定版本，我这里现在没有旧的了，如果你说的是obbdetection和mmsegmentation，完整框架在RSP仓库里，RVSA仓库只是提供相关的backbone和config等文件

from remote-sensing-rvsa.

十分感谢您的回复

from remote-sensing-rvsa.

你好，我想问一下，这个backbone在初始化时候，使用的norm_cfg是SyncBN吗

from remote-sensing-rvsa.

@hhb442 如果在多卡上finetune ViTAE-RVSA的话是，readme有写

from remote-sensing-rvsa.

请问能否提供一下你的conda环境的压缩包呢，我想进行克隆以避免不用的版本错误

from remote-sensing-rvsa.

感谢您的恢复，我用低版本的mmcv\mmdet进行了代码运行，但是我又遇到了以下问题： Traceback (most recent call last): File "tools/train.py", line 153, in main() File "tools/train.py", line 142, in main train_detector( File "/data4/libing/bisai/OBBDetection/mmdet/apis/train.py", line 133, in train_detector runner.run(data_loaders, cfg.workflow, cfg.total_epochs) File "/data4/libing/anaconda3/lib/python3.8/site-packages/mmcv_full-1.7.1-py3.8-linux-x86_64.egg/mmcv/runner/epoch_based_runner.py", line 136, in run epoch_runner(data_loaders[i], **kwargs) File "/data4/libing/anaconda3/lib/python3.8/site-packages/mmcv_full-1.7.1-py3.8-linux-x86_64.egg/mmcv/runner/epoch_based_runner.py", line 53, in train self.run_iter(data_batch, train_mode=True, **kwargs) File "/data4/libing/anaconda3/lib/python3.8/site-packages/mmcv_full-1.7.1-py3.8-linux-x86_64.egg/mmcv/runner/epoch_based_runner.py", line 31, in run_iter outputs = self.model.train_step(data_batch, self.optimizer, File "/data4/libing/anaconda3/lib/python3.8/site-packages/mmcv_full-1.7.1-py3.8-linux-x86_64.egg/mmcv/parallel/data_parallel.py", line 77, in train_step return self.module.train_step(*inputs[0], **kwargs[0]) File "/data4/libing/bisai/OBBDetection/mmdet/models/detectors/base.py", line 237, in train_step losses = self(**data) File "/data4/libing/anaconda3/lib/python3.8/site-packages/torch/nn/modules/module.py", line 889, in _call_impl result = self.forward(*input, **kwargs) File "/data4/libing/bisai/OBBDetection/mmdet/core/fp16/decorators.py", line 51, in new_func return old_func(*args, **kwargs) File "/data4/libing/bisai/OBBDetection/mmdet/models/detectors/base.py", line 172, in forward return self.forward_train(img, img_metas, **kwargs) File "/data4/libing/bisai/OBBDetection/mmdet/models/detectors/obb/obb_two_stage.py", line 154, in forward_train x = self.extract_feat(img) File "/data4/libing/bisai/OBBDetection/mmdet/models/detectors/obb/obb_two_stage.py", line 84, in extract_feat x = self.backbone(img) File "/data4/libing/anaconda3/lib/python3.8/site-packages/torch/nn/modules/module.py", line 889, in _call_impl result = self.forward(*input, **kwargs) File "/data4/libing/bisai/OBBDetection/mmdet/models/backbones/vitae_nc_win_rvsa_wsz7.py", line 777, in forward x = self.forward_features(x) File "/data4/libing/bisai/OBBDetection/mmdet/models/backbones/vitae_nc_win_rvsa_wsz7.py", line 764, in forward_features x = blk(x, Hp, Wp) File "/data4/libing/anaconda3/lib/python3.8/site-packages/torch/nn/modules/module.py", line 889, in _call_impl result = self.forward(*input, **kwargs) File "/data4/libing/bisai/OBBDetection/mmdet/models/backbones/vitae_nc_win_rvsa_wsz7.py", line 540, in forward convX = self.drop_path(self.PCM(x_2d).permute(0, 2, 3, 1).contiguous().view(b, n, c)) File "/data4/libing/anaconda3/lib/python3.8/site-packages/torch/nn/modules/module.py", line 889, in _call_impl result = self.forward(*input, **kwargs) File "/data4/libing/anaconda3/lib/python3.8/site-packages/torch/nn/modules/container.py", line 119, in forward input = module(input) File "/data4/libing/anaconda3/lib/python3.8/site-packages/torch/nn/modules/module.py", line 889, in _call_impl result = self.forward(*input, **kwargs) File "/data4/libing/anaconda3/lib/python3.8/site-packages/torch/nn/modules/batchnorm.py", line 532, in forward world_size = torch.distributed.get_world_size(process_group) File "/data4/libing/anaconda3/lib/python3.8/site-packages/torch/distributed/distributed_c10d.py", line 711, in get_world_size return _get_group_size(group) File "/data4/libing/anaconda3/lib/python3.8/site-packages/torch/distributed/distributed_c10d.py", line 263, in _get_group_size default_pg = _get_default_group() File "/data4/libing/anaconda3/lib/python3.8/site-packages/torch/distributed/distributed_c10d.py", line 347, in _get_default_group raise RuntimeError("Default process group has not been initialized, " RuntimeError: Default process group has not been initialized, please make sure to call init_process_group. 您遇到过相似的问题吗？

如果是单卡训练的话，应该是nn.SyncBatchNorm引起的，换成nn.BatchNorm2d就行了

from remote-sensing-rvsa.