Closed Njasa2k closed 1 year ago
Hi, we managed to train FreestyleNet on a single RTX 3090 GPU (24G VRAM) with batch size = 1. For the inference, 10G VRAM is fine in our case.
Hi, we managed to train FreestyleNet on a single RTX 3090 GPU (24G VRAM) with batch size = 1. For the inference, 10G VRAM is fine in our case.