I noticed a specific code snippet in your project and it has sparked my interest.
trainer = Trainer(
enable_checkpointing=not args.finetune_skip_es,
devices=1,
...
precision=int(args.precision)
if args.precision in ["16", "32"]
else args.precision,
callbacks=callbacks,
accumulate_grad_batches=args.gradient_accumulation_steps,
)
I would like to inquire if your project supports multi-GPU training. Could you please provide information on whether there are built-in features available for utilizing multiple GPUs in the training process?
Hello,
I noticed a specific code snippet in your project and it has sparked my interest.
I would like to inquire if your project supports multi-GPU training. Could you please provide information on whether there are built-in features available for utilizing multiple GPUs in the training process?
Thank you for your assistance!
Best regards