Integrations
MarkovML provides integrations with the following machine-learning frameworks to track experiments automatically. Please review the details below.
Keras
When auto_record
is used for Keras
models, the following information is recorded:
Hyper-parameters:
Information provided by
keras.model.optimizer.get_config()
Metrics:
Epoch-time vs. epoch
Loss vs. epoch
Accuracy vs. epoch
Learning rate vs. epoch (if using adaptive learning rate using
LearningRateScheduler
callback)
Pytorch-Lightning
When auto_record
is used for Pytorch-Lightning
models, the following information is recorded:
Hyper-parameters:
Information provided in
trainer.optimizer.defaults
.
Metrics
Epoch-time vs. epoch
Average running loss vs. epoch
Learning rate vs. epoch
Custom metrics logged by the user using
self.log
in the PyTorch-lightning model.
XGBoost
When auto_record
is used for XGBoost
models, the following information is recorded:
Hyper-parameters:
Parameters returned by
xgboost.Booster.save_config()
. More information for the same can be found in XGBoost documentation here.
Metrics
Epoch-time vs. epoch
Custom metrics that the user logs. If the user does not specify any metrics, XGBoost, by default, logs
logloss
on the validation sets specified.
Scikit-learn
When auto_record
is used for Sklearn
models, the following information is recorded:
Hyper-parameters:
Parameters returned by
sklearn.base.BaseEstimator.get_params()
. More information is in the scikit-learn documentation here.
Metrics are not recorded for sklearn
Last updated
Was this helpful?