Machine learning

Current implementation of Go binding can not specify options.

GPUOptions struct is in internal package. And go generate doesn't work for protobuf directory. So we can't specify GPUOptions for NewSession.

Hello,

The code says that it will add compatibility for Postponed Evaluation of Annotations (PEP 563) when Python 3.9 is released (which already happened on 2020.10.5). Is there any plan to complete this?

https://github.com/huggingface/transformers/blob/2c2a31ffbcfe03339b1721348781aac4fc05bc5e/src/transformers/hf_argparser.py#L85-L90

🐛 Describe the bug

Floating point exception in native_group_norm when num_groups is 0.

Example to reproduce

import torch

input = torch.full((1, 1, 1, 1, 1, 1, 1, 1, 1, 1,), -1.5e+300, dtype=torch.float64, requires_grad=False)
weight = torch.full((1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,),
                    1.5e+300, dtype=torch.float64, requi

See in #22547

MatplotlibDeprecationWarning: Axes3D(fig) adding itself to the figure is deprecated since 3.4. Pass the keyword argument auto_add_to_figure=False and use fig.add_axes(ax) to suppress this warning. The default value of auto_add_to_figure will change to False in mpl3.5 and True values will no longer work in 3.6.  This is consistent with other Axes classes

We need to rep

Following the discussion in JuliaLang/julia#44168 (comment) (and JuliaLinearAlgebra/MKL.jl#106) I suggest to copy https://carstenbauer.github.io/ThreadPinning.jl/dev/explanations/blas/ into somewhere in the Julia manual if @carstenbauer is OK with it.

Currently many more Python projects like dask and optuna are using Python type hints. With the Python package of xgboost gaining more and more features, we should also adopt mypy as a safe guard against some type errors and for better code documentation.