TheBlackBox

cannot run training on FFHQ with batchsize 64 and image size 48

3

I have 8 V100 gpus and run the training code without error with image size 32 and batch size 64 but cuda out of memory occurs when I set image...

No module named cython_bbox

3

I haven't encountered error running **sh make.sh** but when I run test_net.py I still got this error below: 'VIS': False, 'VIS_TH': 0.9} loading annotations into memory... Done (t=0.66s) creating index......

train_step1 all pixels are classified as backgound(black)

2

I did some small change and now the train step 1 can run but all pixels are classified as backgound(black). Anyone has this same issue? Initially the build_loss function is...

Question about custom dataset preparing, tensorflow.python.framework.errors_impl.OutOfRangeError: 2 root error(s) found.

3

Hello, thanks for your great job. I have encountered a problem when trying to do training with my own dataset. I created tfrecord with my own dataset(with jpg files only)....

AssertionError: Attribute 'thing_classes' in the metadata of 'objects365_train' cannot be set to a different value!

4

## 📚 Documentation I arrange the dataset as guided in the documentation, but get this error when trying to train split detectors. * Links to the relevant documentation/comment:

@CharlesShang

Details about training inpainting model

5

Hi, this is an awesome work. I am looking into the code but cannot find the entrance to train the inpainting model. Can you elaborate it a little bit?

psroialign onnx support

1

As far as I know, there is no onnx support for psroialign yet. Does anybody know there is any support for this?

I tested the classificaiton accuracy of the pretrained model and get Top1 0.082 and Top5 0.444. This is weired. Why is it so low? Anyone who tests it as well?

The coco dataset person_keypoint_val_2017.json contains 5000 images but only 2693 are returned fro the CocoDataGenerator. Anyone has the same issue or am I wrong about this?