Current documentation in README explains how to install the toolkit and how to run examples. However, I don't think this is enough for users who want to make some changes to the existing recipes or make their own new recipe. In that case, one needs to understand what run.sh does step by step, but I think docs for that are missing at the moment. It would be great if we provide documentation for:

During self-play phase we usually collect different examples for the same board states. Should we preprocess such examples before optimizing the NNet? In the current implementation, we don't preprocess them so we train NNet and expect different output from the same input values. I think this may be wrong.

Organization Name:

Short summary about the issue/question:
Job

The function mask_head_loss_pre takes gt_bboxes as the 4th positional argument, but the docstring does not contain what this argument is.
https://github.com/chainer/chainercv/blob/master/chainercv/links/model/fpn/mask_head.py#L171

2019-09-17 15:58:06.381228 STDOUT 2012] | chainerrl/tests/links_tests/test_stateless_recurrent_sequential.py .F... [100%]
-- | --
2019-09-17 15:58:06.381228 STDOUT 2012] |  
2019-09-17 15:58:06.381229 STDOUT 2012] | =================================== FAILURES ===================================
2019-09-17 15:58:06.381230 STDOUT 2012] | ___________ TestStatelessRecurrentSequential.test_n_

Great work!

I discovered in https://github.com/arogozhnikov/einops/blob/master/einops/einops.py#L199 that you also support ellipsis. Its an important feature so you may want to add it to the documentation.

When I run text_recognition_demo.py with parameters gpu=-1(that is cpu), it can return the results.
But if I set gpu=0, there is something wrong.
The error is
ValueError：numpy and cupy must not be used together type(W):<class 'numpy.ndarray'>, type(x):<class 'cupy.core.core.ndarray'>,type(b):<class 'numpy.ndarray'>
I know nothing about Chainer, what should I do to solve it?

I noticed in example code, some argument are inconsistent in the commented parts, such as lines in examples/molnet/train_molnet.py:
https://github.com/chainer/chainer-chemistry/blob/56e83dedb5de9dc9eb08ebf292be9ba76a4883ba/examples/molnet/train_molnet.py#L267-L274

args.gpu and concat_mols are not consistent with other codes and will throw errors. They should be modified to device and

term1 = 4 * sq_tan_alpha + matmul(f + f_p, transpose(f_p))
is not term1 = 4 * sq_tan_alpha * matmul(f + f_p, transpose(f_p)) ?
The paper's formula：
fa;p;n = 4 tan2 α(xa + xp)T xn − 2(1 + tan2 α)xT a xp