Wenhe Jia

@GuoxingYan I just add **DecodeBBoxOp** and **DIstributeFpnRpnProposalsOp** on Detectron to train MaskRCNN, the results showed above are not reliable, because a lot of details are not implemented, such as **cls_agnostic_bbox_reg**...

Thank you very much! Looking forward to your later release @zhaoweicai I have added weighted loss and learning rate for different RCNN stages. But compared to the open source of...

@zhaoweicai I noticed that the AP of my first stage are much lower than those of baseline, and the AP of low IOU threshold stay low alone with stage2 and...

I trained the resnet101 model, following the solver, max iterationis 120000, and the mAP is just 70.21%. Which model have you trained?

> try smaller batching > as an example, i've trained 512x512 model on 8gb gpu with batch_size=1 I also have GPU memory limitation, on a GPU with 12GB memory, I...

I cut the the network after proposal_opr processing to calculate nms time cost, it turns out that when processing 1000 proposals, timg cost of tf nms is 0.00061s and the...

Have you merged "moving_mean" and "moving_variance" params into "gamma" and "beta"?

ok, fine, i'll check it out soon, thx

@Zehaos I took your advice and tried to resume training from this step, and it goes well by now. Is this problem occurred during your training progress? What is that...

@zpp13 @chenmyzju I just commented out the code for training RPN1, and run `bash scripts/train_alternate.sh` to resume training from generating RPN detection. And you should kill the progress on your...