mega.pytorch
mega.pytorch copied to clipboard
Scalsol
Memory Enhanced Global-Local Aggregation for Video Object Detection, CVPR2020
Hello there, How to plot the P-R or loss curve during training or from the log.txt file? Without that it appears hard to understand how the training is going. appreciate...
python demo/demo.py mega configs/MEGA/vid_R_50_C4_MEGA_1x.yaml configs/MEGA/MEGA_R_50.pth --suffix ".jpg" --visualize-path datasets/pic/uav0000086_00000_v --output-folder visualization/uav86_MEGA 29%|██████████▊ | 4/14 [00:04
I konw that these config files are uesd as follows on the condition that gpu is 1: BASE_RCNN_1gpu.yaml and vid_R_101_C4_MEGA_1x.yaml. But where can i modify the parameter , NUM_WORKERS of...
你好，非常感谢您开源的代码。 我也是刚刚做视频检测，有一些关于VID数据集的信息想要请教您一下。 当训练的时候，对于key frame(去检测目标的帧)，采用了多少张support frame(丰富key frame的支持帧)啊。我自己是采用了前后一定范围内各取一张 support frame 去丰富当前的 key frame，但是效果一直不是很好，mAP50一直卡在67, 68左右上不去。
Hi, I have an issue in running the demo, there's no error and the video is outputed in vizualisations folder but without any bbox. Could you give me some cue?...
Hello, I want to train on the ICDAR2013 dataset. The dataset consists of video files as opposed to image sequences in ImageNet. How do I convert it to the required...
There is a question when using the demo.py to inference my own video: I have use the script that in your readme: python demo/demo.py base configs/vid_R_101_C4_1x.yaml models/R_101.pth --video --visualize-path datasets/test.mp4...
Thanks for your impressive codebase. I try to reproduce the results of RDN, but the best result is 80.9 which is lower than 81.7 obviously. Shoud I modify any parameters...
https://github.com/Scalsol/mega.pytorch/blob/0f7f2cad2c494f8215c41aa2c2b82d3ac0046925/mega_core/data/datasets/vid_mega.py#L119 测试图片被global给覆盖了
