Detectron.pytorch

Detectron.pytorch copied to clipboard

torch.utils.ffi is deprecated. Please use cpp extensions instead.

1

Torch 1.0 not support

engineer1109

pytorch0.4.1 cuda9+cudnn7 After saving last step checkpoint, detectron just got stucked. I manually killed it, the error infos shows it may be the deadlock in dataloader, can anyone help me?

greathope

### Expected results Will there be a Mask-X-RCNN structure for training in the future which is introduced in "Learning to Segment Everything"

timho102003

Compile error: 'cuda.h'

6

Not really an issue, just want to share my experience. If you are using the code in some clusters, cuda might not be installed under `/usr/local/cuda/`. In this case, in...

Yuliang-Zou

I want to use imgaug package to do the extra data aug ,but I do not know where to add , can anyone give me some suggestions?

yuyijie1995

Grabbing Features for an SVM

8

Hey everyone, I am trying to get the features used for the masks. I have looked in roi_feature_transform and mask_net in model_builder.py However, I don't understand the output of the...

B2Gdevs

Why freeze all bn (affine) layers?

1

lilichu

when I use your detectron ,I find my mask_loss didn't decrease? what happen ,and why?

lingtengqiu

Loss keep increasing

2

Hello! I am trying to train on my own dataset, and I got a loss like below: > [Dec21-15-15-54_user-SYS-7048GR-TR_step][FPN_SE_ARP.yml][Step 201 / 150000] loss: 0.041566, lr: 0.002250 time: 1.166229, eta: 2...

SingL3

Can this implementation handle images without objects?

2

Hello! I want to train a model with some samples that has no objects ( just background ). Can this implementation handle that?

SingL3