@ggjy Thank you for your hard work,

• Will the architecture search code and documentation be released?
• How to predict a single image using your trained model?

WORKDIR = './ work_dirs/hitdet_1x/ '作者您好，这个文件目录在哪个地方呢，我没有找到，或者这个目录的目的是什么

Is there a plan for uploading the NAS code ?
I only find the searched network code in this repo.

Hi, I met with some errors while reporting the evaluation metrics. It won't be able to find the results after 6 epochs. And the validation accuracy is decreasing during training.

2020-06-27 06:10:10,721 - INFO - Epoch [1][29317/29317] lr: 0.02000, bbox_mAP: 0.1380, bbox_mAP_50: 0.2780, bbox_mAP_75: 0.1230, bbox_mAP_s: 0.0660, bbox_mAP_m: 0.1570, bbox_mAP_l: 0.1900, bbox_mAP_copypaste: 0.138 0.278 0.123 0.066 0.157 0.190

2020-06-27 06:10:10,721 - INFO - Epoch [1][29317/29317] lr: 0.02000, bbox_mAP: 0.1380, bbox_mAP_50: 0.2780, bbox_mAP_75: 0.1230, bbox_mAP_s: 0.0660, bbox_mAP_m: 0.1570, bbox_mAP_l: 0.1900, bbox_mAP_copypaste: 0.138 0.278 0.123 0.066 0.157 0.190

2020-06-27 11:37:35,337 - INFO - Epoch [2][29317/29317] lr: 0.02000, bbox_mAP: 0.0670, bbox_mAP_50: 0.1270, bbox_mAP_75: 0.0660, bbox_mAP_s: 0.0370, bbox_mAP_m: 0.0850, bbox_mAP_l: 0.0750, bbox_mAP_copypaste: 0.067 0.127 0.066 0.037 0.085 0.075

2020-06-27 17:05:47,249 - INFO - Epoch [3][29317/29317] lr: 0.02000, bbox_mAP: 0.0160, bbox_mAP_50: 0.0280, bbox_mAP_75: 0.0180, bbox_mAP_s: 0.0120, bbox_mAP_m: 0.0300, bbox_mAP_l: 0.0100, bbox_mAP_copypaste: 0.016 0.028 0.018 0.012 0.030 0.010

2020-06-27 22:33:56,026 - INFO - Epoch [4][29317/29317] lr: 0.02000, bbox_mAP: 0.0030, bbox_mAP_50: 0.0050, bbox_mAP_75: 0.0030, bbox_mAP_s: 0.0030, bbox_mAP_m: 0.0040, bbox_mAP_l: 0.0000, bbox_mAP_copypaste: 0.003 0.005 0.003 0.003 0.004 0.000

2020-06-28 04:01:11,995 - INFO - Epoch [5][29317/29317] lr: 0.02000, bbox_mAP: 0.0000, bbox_mAP_50: 0.0000, bbox_mAP_75: 0.0000, bbox_mAP_s: 0.0000, bbox_mAP_m: 0.0000, bbox_mAP_l: 0.0000, bbox_mAP_copypaste: 0.000 0.000 0.000 0.000 0.000 0.000

2020-06-28 09:28:33,976 - INFO - Epoch [6][29317/29317] lr: 0.02000, bbox_mAP: 0.0000, bbox_mAP_50: 0.0000, bbox_mAP_75: 0.0000, bbox_mAP_s: 0.0000, bbox_mAP_m: 0.0000, bbox_mAP_l: 0.0000, bbox_mAP_copypaste: 0.000 0.000 0.000 0.000 0.000 0.000

2020-06-28 14:55:26,538 - ERROR - The testing results of the whole dataset is empty.

could you explain what the f is in equ 4 in your paper?
tanks a lot

Hello, thanks for sharing your nice work. And could I ask when do you plan to release the code?

Thanks~

Hi there, thanks for sharing your work! It's very inspiring. I have a question for searching time cost, which is not specified in your paper. Would you please tell me an approximate number, for example, how many GPU days? Thanks a lot!

Thanks for your excellent work!

I'm trying to reproduce the paper, but I have some confusion about the search phase.

Could you please further describe the training details of the search phase, such as batch size, image size and so on?

Looking forward to your reply.

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.

Thanks for this repository

I was wondering whether anyone has managed to get this running on colab, on a custom dataset