11/26/2024 08:17:24 - INFO - __main__ - Distributed environment: MULTI_GPU  Backend: nccl
Num processes: 3
Process index: 0
Local process index: 0
Device: cuda:0

Mixed precision type: no

11/26/2024 08:17:38 - INFO - __main__ - Initializing the InstructPix2Pix UNet from the pretrained UNet.
11/26/2024 08:18:08 - INFO - __main__ - ***** Running training *****
11/26/2024 08:18:08 - INFO - __main__ -   Num examples = 108093
11/26/2024 08:18:08 - INFO - __main__ -   Num Epochs = 1
11/26/2024 08:18:08 - INFO - __main__ -   Instantaneous batch size per device = 8
11/26/2024 08:18:08 - INFO - __main__ -   Total train batch size (w. parallel, distributed & accumulation) = 48
11/26/2024 08:18:08 - INFO - __main__ -   Gradient Accumulation steps = 2
11/26/2024 08:18:08 - INFO - __main__ -   Total optimization steps = 100