Effective batch size
`per_device_batch_size × gradient_accumulation_steps` — the real batch size after accumulation.
per_device_batch_size × gradient_accumulation_steps — the real batch size after accumulation. (M03)
`per_device_batch_size × gradient_accumulation_steps` — the real batch size after accumulation.
per_device_batch_size × gradient_accumulation_steps — the real batch size after accumulation. (M03)