Loggers¶

Lightning supports the most popular logging frameworks (TensorBoard, Comet, Neptune, etc…). TensorBoard is used by default, but you can pass to the Trainer any combination of the following loggers.

Note

All loggers log by default to os.getcwd(). To change the path without creating a logger set Trainer(default_root_dir=’/your/path/to/save/checkpoints’)

Comet.ml¶

Comet.ml is a third-party logger. To use CometLogger as your logger do the following. First, install the package:

pip install comet-ml

Then configure the logger and pass it to the Trainer:

import os
from pytorch_lightning.loggers import CometLogger

comet_logger = CometLogger(
    api_key=os.environ.get("COMET_API_KEY"),
    workspace=os.environ.get("COMET_WORKSPACE"),  # Optional
    save_dir=".",  # Optional
    project_name="default_project",  # Optional
    rest_api_key=os.environ.get("COMET_REST_API_KEY"),  # Optional
    experiment_name="lightning_logs",  # Optional
)
trainer = Trainer(logger=comet_logger)

The CometLogger is available anywhere except __init__ in your LightningModule.

class MyModule(LightningModule):
    def any_lightning_module_function_or_hook(self):
        some_img = fake_image()
        self.logger.experiment.add_image("generated_images", some_img, 0)

See also

CometLogger docs.

MLflow¶

MLflow is a third-party logger. To use MLFlowLogger as your logger do the following. First, install the package:

pip install mlflow

Then configure the logger and pass it to the Trainer:

from pytorch_lightning.loggers import MLFlowLogger

mlf_logger = MLFlowLogger(experiment_name="lightning_logs", tracking_uri="file:./ml-runs")
trainer = Trainer(logger=mlf_logger)

See also

MLFlowLogger docs.

Neptune.ai¶

Neptune.ai is a third-party logger. To use NeptuneLogger as your logger do the following. First, install the package:

pip install neptune-client

or with conda:

conda install -c conda-forge neptune-client

Then configure the logger and pass it to the Trainer:

from pytorch_lightning.loggers import NeptuneLogger

neptune_logger = NeptuneLogger(
    api_key="ANONYMOUS",  # replace with your own
    project="common/pytorch-lightning-integration",  # format "<WORKSPACE/PROJECT>"
    tags=["training", "resnet"],  # optional
)
trainer = Trainer(logger=neptune_logger)

The NeptuneLogger is available anywhere except __init__ in your LightningModule.

class MyModule(LightningModule):
    def any_lightning_module_function_or_hook(self):
        # generic recipe for logging custom metadata (neptune specific)
        metadata = ...
        self.logger.experiment["your/metadata/structure"].log(metadata)

Note that syntax: self.logger.experiment["your/metadata/structure"].log(metadata) is specific to Neptune and it extends logger capabilities. Specifically, it allows you to log various types of metadata like scores, files, images, interactive visuals, CSVs, etc. Refer to the Neptune docs for more detailed explanations.

You can always use regular logger methods: log_metrics() and log_hyperparams() as these are also supported.

See also

NeptuneLogger docs.

Logger user guide.

Tensorboard¶

To use TensorBoard as your logger do the following.

from pytorch_lightning.loggers import TensorBoardLogger

logger = TensorBoardLogger("tb_logs", name="my_model")
trainer = Trainer(logger=logger)

The TensorBoardLogger is available anywhere except __init__ in your LightningModule.

class MyModule(LightningModule):
    def any_lightning_module_function_or_hook(self):
        some_img = fake_image()
        self.logger.experiment.add_image("generated_images", some_img, 0)

To see your logs, run the following command in the terminal:

tensorboard --logdir=<logging_folder>

To visualize tensorboard in a jupyter notebook environment, run the following command in a jupyter cell:

%reload_ext tensorboard
%tensorboard --logdir=<logging_folder>

See also

TensorBoardLogger docs.

Weights and Biases¶

Weights and Biases is a third-party logger. To use WandbLogger as your logger do the following. First, install the package:

pip install wandb

Then configure the logger and pass it to the Trainer:

from pytorch_lightning.loggers import WandbLogger

# instrument experiment with W&B
wandb_logger = WandbLogger(project="MNIST", log_model="all")
trainer = Trainer(logger=wandb_logger)

# log gradients and model topology
wandb_logger.watch(model)

The WandbLogger is available anywhere except __init__ in your LightningModule.

class MyModule(LightningModule):
    def any_lightning_module_function_or_hook(self):
        some_img = fake_image()
        # Option 1
        self.logger.experiment.log({"generated_images": [wandb.Image(some_img, caption="...")]})
        # Option 2 for specifically logging images
        self.logger.log_image(key="generated_images", images=[some_img])

To visualize using wandb in a jupyter notebook environment use the following magic line command:

%%wandb

# Your training loop here

To display any existing dashboards, sweeps or reports directly in your notebook using the %wandb magic:

# Display a project workspace
%wandb USERNAME/PROJECT

More information is available here.

See also

WandbLogger docs.
W&B Documentation
Demo in Google Colab with hyperparameter search and model logging

Multiple Loggers¶

Lightning supports the use of multiple loggers, just pass a list to the Trainer.

from pytorch_lightning.loggers import TensorBoardLogger, WandbLogger

logger1 = TensorBoardLogger(save_dir="tb_logs", name="my_model")
logger2 = WandbLogger(save_dir="tb_logs", name="my_model")
trainer = Trainer(logger=[logger1, logger2])

The loggers are available as a list anywhere except __init__ in your LightningModule.

class MyModule(LightningModule):
    def any_lightning_module_function_or_hook(self):
        some_img = fake_image()
        # Option 1
        self.logger.experiment[0].add_image("generated_images", some_img, 0)
        # Option 2
        self.logger[0].experiment.add_image("generated_images", some_img, 0)