Skip to content

Set the logging mode#

Depending on how you want the data to be stored and synchronized, you can use Neptune in five different connection modes.

For example, if you're debugging your code, you can use Neptune in such a way that no data is uploaded or stored at all.

Mode Parameter value Description
Asynchronous (default) async Tracked data is stored locally and synchronized separately in the background.
Debug debug No data is stored locally or sent to Neptune.
Offline offline Data is not sent to Neptune, but stored locally. Can later be synced manually.
Read only read-only Used to only fetch data from existing objects. Ensures that no data is added or changed.
Synchronous sync Neptune waits for other calls to reach the server before executing a tracking call.

Setting the connection mode#

You can set the mode through either an environment variable or an initialization parameter.

Environment variable#

The environment variable is used if the mode parameter is omitted from the init function.

export NEPTUNE_MODE='debug'

export NEPTUNE_MODE='debug'
Current session only
set NEPTUNE_MODE='debug'
System-wide
setx NEPTUNE_MODE 'debug'

Tip

This can be a handy way to debug or "switch off" Neptune without modifying your code.

Argument of init() function#

To set the connection mode for the logging, provide it to the mode argument of the initialization function:

import neptune

run = neptune.init_run(mode="sync")

To only enable querying and no logging of new data, you can initialize an object in read only mode:

model_version = neptune.init_model_version(
    with_id="NER-PRETRAINED-8",
    mode="read-only",
)

Related