Arij
December 12, 2022, 8:56am
1
@muellerzr
I am using this script to train some models.
checkpointing_steps = args.checkpointing_steps
if checkpointing_steps is not None and checkpointing_steps.isdigit():
checkpointing_steps = int(checkpointing_steps)
# We need to initialize the trackers we use, and also store our configuration.
# The trackers initializes automatically on the main process.
if args.with_tracking:
experiment_config = vars(args)
# TensorBoard cannot log Enums, need the raw value
experiment_config["lr_scheduler_type"] = experiment_config["lr_scheduler_type"].value
accelerator.init_trackers("summarization_no_trainer", experiment_config)
# Metric
metric = evaluate.load("rouge")
# Train!
total_batch_size = args.per_device_train_batch_size * accelerator.num_processes * args.gradient_accumulation_steps
logger.info("***** Running training *****")
logger.info(f" Num examples = {len(train_dataset)}")
logger.info(f" Num Epochs = {args.num_train_epochs}")
I saw this answe for using wandb by accelerator.
The question is as follow for the script hyperparameters I used the following paramters for tracking
–with_tracking --report_to=‘wandb’ --checkpointing_steps=‘epoch’
still I can not see the results anywhere or track them. I think I missed something. Also If I want to use all how can I track the results?
prcisely I am getting warning ``log_with=wandb was passed but no supported trackers are currently installed. accelerate
Arij
December 13, 2022, 1:37pm
2
for anyone stuck with this issue you need to install wandb using
pip install wandb