Environment Variables
The proxiML job execution environment presets a variety of environment variables to make the development of worker training scripts easier and more flexible.
ML_JOB_NAME
: The name of the job with spaces converted to underscores.ML_JOB_ID
: A unique ID specific to this job.ML_WORKER_NUM
: The number of the worker in a multi-worker training job.ML_WORKER_ID
: A unique ID specific to this job worker.ML_DATA_PATH
: The local directory of the data loaded through theDataset
option. This directory is read-only.ML_OUTPUT_PATH
: The local directory that will be uploaded to theOutput storage path
after job completion.ML_MODEL_PATH
: The local directory containing the model code. This is the default working directory when a worker starts.ML_CHECKPOINT_PATH
: The local directory of the checkpoints loaded through theCheckpoint
option. This directory is read-only.ML_CLIENT_IP
: The IP address a worker can use to communicate back to the customer running the connection utility. This will only be accessible when the customer is connected.ML_DEVICE_NAME
: (CloudBender Devices Only) The friendly name of the device the endpoint is running on.ML_DEVICE_ID
: (CloudBender Devices Only) The unique ID of the device the endpoint is running on.