flake8
SpeedTorch
ohmyzsh
pypy
skorch
missingno
gensim
numexpr
bottleneck
tpot
wandb
FBLearner Flow
mlflow
Neptune.ai
nyaggle
fastai有协同过滤
torch-optimizer 拓展pytorch优化器
test tube
neptune-notebooks
comet_ml
tensorwatchnotebook里可运行
swifter加速pandas
modin(加载数据和查询数据更快，统计方法pandas更快)
numba
cupy(1000万以上数据才更快)
Cython
hyperparameter_hunter

DeepSpeed 微软分布式训练工具
dvc
hiddenlayer
syncthing
ignite
pytorch-lightning
fastai
prefetch_generator
torchsummary
apex
dali
optuna
pip-review
autogluon
arthas
greys-anatomy
implicit
eli5
icecream
stackprinter
dlrm
buffalo
cupy
acptum
featuretools
boruta_py
Categorical-encoding
Tsfresh
Scikit-Optimize
Hyperopt
ray
ENAS-pytorch
MLBox
h2o
AutoKeras
TPOT
Auto-Sklearn