Hi, I wonder how to compute speedup on GPUs? Is it the total inference time of CPU divide by that on GPU? If so, I wonder how to calculate the total time on the CPU by mmdetection, do I need to put all the data on the CPU for retraining? Or just need to put the data on the CPU during inference?

Looking forward to reply！

Hi, I basically completed the pruning codes of faster-rcnn. I wonder the batch size and the number of gpu during pruning faster-rcnn, is it 16 on 8 2080ti which is same as all detection models? Thank you very much!

DOes your framework focus only on Resnet based backbone or it can be adapted for other models like yolov4 or yolov5.

Hi, thanks for this awesome work.
I'm wondering if there is an option to continue pruning, I tried setting pruning=True and deploy_from=checkpoint to achieve it, but it didn't work.
And then I modify the code in

Is there anything else I need to change?

Hi, I tried to complete the codes of pruning faster rcnn, but I don't know how to prune the FC layer in roi_head. I have finished the twostage_wrapper function as shown below

But I got error.

Could you help me see where the problem is? Thanks!

Hello, may I ask when will u release the code and models for classification task?
Thanks

Hi, the given finetuned ATSS model can't be tested, the error is as follows, maybe the file is incomplete. And I have tested finetuned models of RetinaNet and PAA, they are normal.

Hi, I tried to complete the codes of faster rcnn, but I don't know how to prune the FC layer in roi_head. I finished the twostage_wrapper function as shown below.

But I got the error.

Could you help me find out what went wrong? Thanks!

Hi, I basically completed the pruning codes of faster-rcnn. I wonder the batch size and the number of gpu during pruning faster-rcnn, is it 16 on 8 2080ti which is same as all detection models? Thank you very much!

Hi,
I have DepthWise Conv layer in my network. And It is not detected in self.conv2ancest.
Do the project support DepthWise convolution ?

Thanks.

Welcome update to OpenMMLab 2.0

I am Vansin, the technical operator of OpenMMLab. In September of last year, we announced the release of OpenMMLab 2.0 at the World Artificial Intelligence Conference in Shanghai. We invite you to upgrade your algorithm library to OpenMMLab 2.0 using MMEngine, which can be used for both research and commercial purposes. If you have any questions, please feel free to join us on the OpenMMLab Discord at https://discord.gg/amFNsyUBvm or add me on WeChat (van-sin) and I will invite you to the OpenMMLab WeChat group.

Here are the OpenMMLab 2.0 repos branches:

Attention: please create a new virtual environment for OpenMMLab 2.0.

Hello sir,

When I test with your finetuned RetinaNet model, I can't get the same mAP with you.

Could you please help to give me some advices?

Thanks.

Hi, I wonder if this line in Table 5 is wrong? It should be Faster-I

Hi, I'm very interested in your work. When I fintune the RetinaNet of your pruned model, the finetuning process will last for 2 days, which is very different from the 7 hours in the paper, as shown in the figure below.

One 1080ti is used and I didn't use slurm. The command for finetuning I used is python -u tools/train.py configs/retina/retina_finetune.py --work-dir ./ --gpu-ids 0

In PyTorch >= 1.8, this register_backward_hook can't be used directly, and register_full_backward_hook always return "using input grad before output grad bla...". Did you know how to replace register_backward_hook in higher PyTorch version?

Hi, when I finetune the faster rcnn ( flops 50%, act 64% ), FC layers of roi_head wasn't be pruned, I wonder is it normal?

In ResNeXt blocks, conv2 is a GConv and should be grouped with its child conv3.
However, if conv2 is not in any group while we assign a group for conv3, e.g. conv2 is behind conv3 in \mathbb{L}, conv3 will not be assigned to the same group with conv2. Instead, it will be assigned to a new group. The conv2 will not be assigned to this group since conv2 and conv3 have no shared parent.

When we modify the make_groups function according to this algorithm, it indeed fails to group the conv2&conv3 in ResNeXt50 blocks though the found ancestors are correct.

Could u please help check this problem?

Now in this repo, we tentatively change this function to

and it works.

Hi, I basically completed the pruning codes of faster-rcnn. I wonder the batch size and the number of gpu during pruning faster-rcnn, is it 16 on 8 2080ti which is same as all detection models? Thank you very much!

Hi, I'm trying to complete the pruning codes of FC layer. I wonder how to compute the acts of FC layer, is it n times oc where n is the batch size and oc is the channel numbers of output feature?

What's meaning of under code:
https://github.com/jshilong/FisherPruning/blob/456106c04eace7f1722577c2fce0d693eaead915/detection/tools/fisher_pruning.py#L501C1-L507C25
I think group which len(modules) == 1 should be a single group, is it right?

Is there any plan to release code for Classification? Or where else can I find the code.

Hi, I wonder if it is normal that the flops is too small during pruning RetinaNet? And does it mean I need to finetune when flops reduces about 50%?

Hi, The loss in Groups is computed based on bbox loss; what is the best way to integrate it with mask loss? Thanks,

I create a virtual environment, open-mmlab, with python=3.7, pytorch=1.3.0, cudatoolkit=10.0, mmcv=1.3.16, mmdet=2.13.0.
However, it still can not run.

File "tools/train.py", line 16, in
from mmdet.apis import set_random_seed, train_detector
ImportError: libcudart.so.10.1: cannot open shared object file: No such file or directory

Thanks for the great work, I have some questions about this code

1. In the paper you mentioned coupled channels, I understand what it means is that C2/C5 (you described in Figure 4) receive the same input. Is there any other meaning besides this?