Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

import MultiScaleDeformableAttention as MSDA #1

Open
Guanerhua opened this issue Aug 10, 2021 · 13 comments
Open

import MultiScaleDeformableAttention as MSDA #1

Guanerhua opened this issue Aug 10, 2021 · 13 comments

Comments

@Guanerhua
Copy link

I can not install MultiScaleDeformableAttention can you help me
@HITerStudy
Copy link

I can not install MultiScaleDeformableAttention can you help me

Hello, I also meet this problem, but I found the module in https://github.com/fundamentalvision/Deformable-DETR, you can try it!

@volgachen
Copy link
Collaborator

I can not install MultiScaleDeformableAttention can you help me

Hello, I also meet this problem, but I found the module in https://github.com/fundamentalvision/Deformable-DETR, you can try it!

Yeah, this module is directly borrowed from https://github.com/fundamentalvision/Deformable-DETR. You can try follow the original instruction.

@Guanerhua Could you provide your detailed enrionment information and error report? Maybe we can help you solve it. So far as I can tell, make sure that your pytorch & cudatoolkit version (in your conda) keep consistent with your system cuda version (usually /usr/local/cuda), this is a common problem.

@Guanerhua
Copy link
Author

I can not install MultiScaleDeformableAttention can you help me

Hello, I also meet this problem, but I found the module in https://github.com/fundamentalvision/Deformable-DETR, you can try it!

Thanks，I solve my problem

@Guanerhua
Copy link
Author

I can not install MultiScaleDeformableAttention can you help me

Hello, I also meet this problem, but I found the module in https://github.com/fundamentalvision/Deformable-DETR, you can try it!

Yeah, this module is directly borrowed from https://github.com/fundamentalvision/Deformable-DETR. You can try follow the original instruction.

@Guanerhua Could you provide your detailed enrionment information and error report? Maybe we can help you solve it. So far as I can tell, make sure that your pytorch & cudatoolkit version (in your conda) keep consistent with your system cuda version (usually /usr/local/cuda), this is a common problem.

Thanks，I solve my problem

@Guanerhua
Copy link
Author

I can not install MultiScaleDeformableAttention can you help me

Hello, I also meet this problem, but I found the module in https://github.com/fundamentalvision/Deformable-DETR, you can try it!

Hi,bro. I meet another problem 'Sample_Classifier in depatch_embed is error' .Can you help me ?
`class Simple_DePatch(Simple_Patch):
def init(self, box_coder, show_dim=4, use_auxiliary=-1, **kwargs):
super().init(show_dim, **kwargs)
self.box_coder = box_coder
self.register_buffer("value_spatial_shapes", torch.as_tensor([[self.img_size, self.img_size]], dtype=torch.long))
self.register_buffer("value_level_start_index", torch.as_tensor([0], dtype=torch.long))
self.output_proj = nn.Linear(self.in_chans * self.patch_pixel * self.patch_pixel, self.embed_dim)
self.num_sample_points = self.patch_pixel * self.patch_pixel * self.patch_count * self.patch_count
self.use_auxiliary = use_auxiliary > -1
微信图片_20210816095902

    if self.use_auxiliary:
        self.classifier = Sample_Classifier(use_auxiliary, kwargs["in_chans"])`

@zhixiaoqiang-haha
Copy link

I can not install MultiScaleDeformableAttention can you help me

Hello, I also meet this problem, but I found the module in https://github.com/fundamentalvision/Deformable-DETR, you can try it!

Yeah, this module is directly borrowed from https://github.com/fundamentalvision/Deformable-DETR. You can try follow the original instruction.
@Guanerhua Could you provide your detailed enrionment information and error report? Maybe we can help you solve it. So far as I can tell, make sure that your pytorch & cudatoolkit version (in your conda) keep consistent with your system cuda version (usually /usr/local/cuda), this is a common problem.

Thanks，I solve my problem

@Guanerhua Hello, I would like to ask how you solved this problem about MultiScaleDeformableAttention

@eeric
Copy link

eeric commented Aug 17, 2021

solution:
#3

@Guanerhua
Copy link
Author

I can not install MultiScaleDeformableAttention can you help me

Hello, I also meet this problem, but I found the module in https://github.com/fundamentalvision/Deformable-DETR, you can try it!

Yeah, this module is directly borrowed from https://github.com/fundamentalvision/Deformable-DETR. You can try follow the original instruction.
@Guanerhua Could you provide your detailed enrionment information and error report? Maybe we can help you solve it. So far as I can tell, make sure that your pytorch & cudatoolkit version (in your conda) keep consistent with your system cuda version (usually /usr/local/cuda), this is a common problem.

Thanks，I solve my problem

@Guanerhua Hello, I would like to ask how you solved this problem about MultiScaleDeformableAttention

You can try it!
image

@zhixiaoqiang-haha
Copy link

zhixiaoqiang-haha commented Aug 18, 2021 via email

@volgachen
Copy link
Collaborator

I can not install MultiScaleDeformableAttention can you help me

Hello, I also meet this problem, but I found the module in https://github.com/fundamentalvision/Deformable-DETR, you can try it!

Hi,bro. I meet another problem 'Sample_Classifier in depatch_embed is error' .Can you help me ?
`class Simple_DePatch(Simple_Patch):
def init(self, box_coder, show_dim=4, use_auxiliary=-1, **kwargs):
super().init(show_dim, **kwargs)
self.box_coder = box_coder
self.register_buffer("value_spatial_shapes", torch.as_tensor([[self.img_size, self.img_size]], dtype=torch.long))
self.register_buffer("value_level_start_index", torch.as_tensor([0], dtype=torch.long))
self.output_proj = nn.Linear(self.in_chans * self.patch_pixel * self.patch_pixel, self.embed_dim)
self.num_sample_points = self.patch_pixel * self.patch_pixel * self.patch_count * self.patch_count
self.use_auxiliary = use_auxiliary > -1
微信图片_20210816095902

    if self.use_auxiliary:
        self.classifier = Sample_Classifier(use_auxiliary, kwargs["in_chans"])`

I am sorry. Content related to use_auxiliary is useless and deprecated. I forget to delete it here. You can simply delete it. (self.use_auxiliary is always False.)

@huchunjun425
Copy link

I can not install MultiScaleDeformableAttention can you help me

Hello, I also meet this problem, but I found the module in https://github.com/fundamentalvision/Deformable-DETR, you can try it!

Yeah, this module is directly borrowed from https://github.com/fundamentalvision/Deformable-DETR. You can try follow the original instruction.

@Guanerhua Could you provide your detailed enrionment information and error report? Maybe we can help you solve it. So far as I can tell, make sure that your pytorch & cudatoolkit version (in your conda) keep consistent with your system cuda version (usually /usr/local/cuda), this is a common problem.

i also meet this problem,could you please tell this details how to solve this issue? thanks.

@RonghaoXXXu
Copy link

RonghaoXXXu commented Dec 3, 2022 •
edited
Loading

Hi,bro. I followed the origin introduction form https://github.com/fundamentalvision/Deformable-DETR. But there is a very strange problems, which are as followed with this. Can you help me? platform:Windows 10 

vision.obj : error LNK2001: 无法解析的外部符号 __imp___tls_offset_?init@?1??lazy_init_num_threads@internal@at@@YAXXZ@4_NA
vision.obj : error LNK2001: 无法解析的外部符号 __imp___tls_index_?init@?1??lazy_init_num_threads@internal@at@@YAXXZ@4_NA
build\lib.win-amd64-cpython-38\MultiScaleDeformableAttention.cp38-win_amd64.pyd : fatal error LNK1120: 2 个无法解析的外部命令
error: command 'D:\\VS\\2022\\BuildTools\\VC\\Tools\\MSVC\\14.34.31933\\bin\\HostX86\\x64\\link.exe' failed with exit code 1120

@UsamaHasan007
Copy link

Any Idea I can build it on Ampere Architecture which supports CUDA 11.0 or greater

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
9 participants
@eeric @HITerStudy @volgachen @RonghaoXXXu @zhixiaoqiang-haha @Guanerhua @huchunjun425 @UsamaHasan007 and others