Yoo Jaehoon

Thanks.

I think you are right. I updated train.py. My code has a lot of out-dated coding style, and I recommend you use pytorch's lr-scheduler. I have a plan to update...

Would you introduce the specific code line that the error occurs? Also, would you translate your question into English, please? I cannot read Chinese. It would be better to attach...

I think your issue is related with https://github.com/Ugness/PiCANet-Implementation/issues/12#issuecomment-453928501. Please check your tensorboardX version. Thank you.

That problem happened from the **tqdm** module. I do not know the exact solution, but I recommend you to - google "tqdm printing to newline" - https://stackoverflow.com/questions/41707229/tqdm-printing-to-newline - update tqdm...

Please follow the steps in the Execution Guideline in README.md. https://github.com/Ugness/PiCANet-Implementation#execution-guide You can put the files anywhere, but you should give the directory as command-line options. If you meet any...

I think you do not have mask images under "[ROOT_DIR]/masks". - Please follow https://github.com/Ugness/PiCANet-Implementation#dataset if you were going to train the model. - if you want to inference, please follow...

1. If you train the network from the pretrained one, it would be possible that the loss didn't fall. 2. If you train the network from random initialization, please check...

Can you share your command line to execute the code?? for example, ```python train.py``` Make sure that you should give the path of your dataset. Here is an example. If...

We didn't measure FLOPS for training. In the case of 128-frame training, training on SkyTimelapse took 12hrs with 4 A100 GPUs, and training on Taichi and UCF-101 took about 2...