Doyup Lee

tf codes link: https://github.com/LeeDoYup/bottom-up-attention-tf

https://github.com/LeeDoYup/bottom-up-attention-tf

Reconstruction loss

1

In the paper, the authors describe that the recon loss is replaced with perceptual loss. However, in the code, the actual recon loss is L1 (not L2) + perceptual loss....

I tried to use MS-COCO dataset for training. After i serialize the MS-COCO image using ``` img = tf.gfile.FastGFile(img_path, 'rb').read() or img = open(img_path, 'rb').read() ``` and save tfrecords, i...

I can i train glow model with my custom dataset (actually, i want to train the model on MS-COCO). Is anybody who has tried to use her/his own dataset?

Reproduce of FixMatch

1

I wonder the reason for the large gaps between the error rates of the original paper (FixMatch) and this implementation.

The information about training data

5

Hello, thanks for the awesome project ! I love the results and respect for the authors' efforts. By the way, I cannot find the information of training data. I wonder...

Fix random seed bug of DDP in the ImageNet example

1

the random seed has to be set in the main_worker, not in the `def main()`. I found that although the seed is set in `def main()`, each process in distributed...

When the codes are updated? I want to do experiment for my paper. I have difficulty to refer your codes and i want to show completed code. Thank you.

Hello. I have a question. When i train a model with the ImageNet dataset on "GPU", how can i use the multi-node training? I think the scripts in READ.md do...