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.

Related to #5142, AlbertTokenizer (which uses SentencePiece) doesn't decode special tokens (like [CLS], [MASK]) properly. This issue was discovered when adding the Nystromformer model (#14659), which uses this tokenizer.

To reproduce (Transformers v4.15 or below):

!pip install -q transformers sentencepiece

from transformers import AlbertTokenizer

tokenizer = AlbertTokenizer.from

🐛 Describe the bug

Usage of RRefContext::handleException in torch/csrc/distributed/rpc/rref_context.cpp is wrong when the future has an error.

RRefContext::handleException uses TORCH_CHECK which throws.

Callers of RRefContext::handleException don't expect that and run code after it without any guarding.

Versions

master

cc @pietern @mrshenli @pritamdamania87

Describe the issue linked to the documentation

Many legitimate notebook style examples have been broken, and specifically by the following PR
scikit-learn/scikit-learn#9061

Here are all the examples that use patterns like # ####### (found by ag -l '# #####*\s#' examples | sort, note there may be false positives ... for example examples/impute/plot_missing_val

julia> throw(ArgumentError("the quick brown fox jumps over the lazy dog"))
ERROR: ArgumentError: the quick brown fox jumps over the lazy dog
Stacktrace:
 [1] top-level scope
   @ REPL[13]:1

julia> throw(DimensionMismatch("the quick brown fox jumps over the lazy dog"))
ERROR: DimensionMismatch("the quick brown fox jumps over the lazy dog")

I believe it should be printing l

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.