mlflow¶
Functions
Classes
Log using MLflow. |
MLflow Logger¶
- class lightning.pytorch.loggers.mlflow.MLFlowLogger(experiment_name='lightning_logs', run_name=None, tracking_uri=None, tags=None, save_dir='./mlruns', log_model=False, prefix='', artifact_location=None, run_id=None)[source]¶
Bases:
lightning.pytorch.loggers.logger.Logger
Log using MLflow.
Install it with pip:
pip install mlflow # or mlflow-skinny
from lightning.pytorch import Trainer from lightning.pytorch.loggers import MLFlowLogger mlf_logger = MLFlowLogger(experiment_name="lightning_logs", tracking_uri="file:./ml-runs") trainer = Trainer(logger=mlf_logger)
Use the logger anywhere in your
LightningModule
as follows:from lightning.pytorch import LightningModule class LitModel(LightningModule): def training_step(self, batch, batch_idx): # example self.logger.experiment.whatever_ml_flow_supports(...) def any_lightning_module_function_or_hook(self): self.logger.experiment.whatever_ml_flow_supports(...)
- Parameters
run_name¶ (
Optional
[str
]) – Name of the new run. The run_name is internally stored as amlflow.runName
tag. If themlflow.runName
tag has already been set in tags, the value is overridden by the run_name.tracking_uri¶ (
Optional
[str
]) – Address of local or remote tracking server. If not provided, defaults to MLFLOW_TRACKING_URI environment variable if set, otherwise it falls back to file:<save_dir>.tags¶ (
Optional
[Dict
[str
,Any
]]) – A dictionary tags for the experiment.save_dir¶ (
Optional
[str
]) – A path to a local directory where the MLflow runs get saved. Defaults to ./mlflow if tracking_uri is not provided. Has no effect if tracking_uri is provided.log_model¶ (
Literal
[True, False, ‘all’]) –Log checkpoints created by
ModelCheckpoint
as MLFlow artifacts.if
log_model == 'all'
, checkpoints are logged during training.if
log_model == True
, checkpoints are logged at the end of training, except whensave_top_k
== -1
which also logs every checkpoint during training.if
log_model == False
(default), no checkpoint is logged.
prefix¶ (
str
) – A string to put at the beginning of metric keys.artifact_location¶ (
Optional
[str
]) – The location to store run artifacts. If not provided, the server picks an appropriate default.run_id¶ (
Optional
[str
]) – The run identifier of the experiment. If not provided, a new run is started.
- Raises
ModuleNotFoundError – If required MLFlow package is not installed on the device.
- after_save_checkpoint(checkpoint_callback)[source]¶
Called after model checkpoint callback saves a new checkpoint.
- Parameters
checkpoint_callback¶ (
ModelCheckpoint
) – the model checkpoint callback instance- Return type
- log_metrics(metrics, step=None)[source]¶
Records metrics. This method logs metrics as soon as it received them.
- property experiment: None¶
Actual MLflow object. To use MLflow features in your
LightningModule
do the following.Example:
self.logger.experiment.some_mlflow_function()
- Return type
- property experiment_id: Optional[str]¶
Create the experiment if it does not exist to get the experiment id.
- lightning.pytorch.loggers.mlflow.resolve_tags(tags=None)[source]¶
- Parameters
tags¶ (
Optional
[Dict
]) – A dictionary of tags to override. If specified, tags passed in this argument will override those inferred from the context.
Returns: A dictionary of resolved tags.
Note
See
mlflow.tracking.context.registry
for more details.