When attempting to download cityscapes_2048x1024 I got: ./download-models.sh: line 721: download_fcn_resnet18_cityscapes_2048x512: command not found
It looks like there was a typo, and line 721 needs to be changed from:
download_fcn_resnet18_cityscapes_2048x512 to download_fcn_resnet18_cityscapes_2048x1024

Thanks for the amazing repo!

I trained > 100k, the loss looks fine, the bounding box looks too much, but the mask looks very wrong. Is it the visualization bug or we need more training?

In file binary segmentation (camvid).ipynb, block 5, there is:

# Lets look at data we have
dataset = Dataset(x_train_dir, y_train_dir, classes=['car', 'pedestrian'])
image, mask = dataset[5] # get some sample
visualize(
    image=image, 
    cars_mask=mask[..., 0].squeeze(),
    sky_mask=mask[..., 1].squeeze(),
    background_mask=mask[..., 2].squeeze(),
)

here, sky_mask

(I will compile a list and hopefully open a PR if needed)

Describe the bug
This behavior is present in a plethora of catalyst's callbacks and losses. It's consistent, but it's definitely confusing for many new users.

To Reproduce
Steps to reproduce the behavior:
Use these functions/classes:
Callbacks

• [MeterMetricsCallback](https://github.com/catalyst-team/catalyst/blob/mas

In 'pointnet2/scannet/preprocessing/collect_scannet_scenes.py' line #17, we refer to 'scannet_all.txt' file. It would be helpful if you guide where to get or how to generate this file?

Pytorch v1.1 (which we now use) comes with an official TensorBoard integration.

We should remove our custom logging and chart plotting and move to TensorBoard.

• https://pytorch.org/docs/stable/tensorboard.html
• https://github.com/pytorch/pytorch/releases/tag/v1.1.0

I am not clear about what the iteration means. Could you explain more about it? Also, how is it related to epoch?

Thanks.

1. 31页，第一行
   
   应改成
   

2. 35 页， 正定矩阵的定义 （公式 2-15）
   标准正定矩阵的定义，应该没有 “对称矩阵” 这一条件吧？ 只是您在本书大多讨论的是“对称”正定矩阵，所以加上了 “对称” 这一条件？

3. 38页， 奇异值分解定义（公式 2-16）
   ![default](https://