Closed linmi1 closed 4 months ago
when I run this code : all_imgs = torch.zeros(len(videos) - 1, 300, img_wh[-1] * img_wh[-2], 3)
will get:
Process finished with exit code 137
Hi. It may not be the GPU problem as the all_imgs = torch.zeros(len(videos) - 1, 300, img_wh[-1] * img_wh[-2], 3)
doesn't use GPUs. It may related to CPU memories. How much memory (RAM) are you using?
Hi. It may not be the GPU problem as the
all_imgs = torch.zeros(len(videos) - 1, 300, img_wh[-1] * img_wh[-2], 3)
doesn't use GPUs. It may related to CPU memories. How much memory (RAM) are you using?
yes ,we can add .to("cuda") to transfer it to cuda ,but it also need too much memory.
I tried a dynamic video on a 24G, 3090 graphics card, but it showed that the graphics memory was insufficient. If the input frame rate was reduced, the result would be very poor. Does anyone tried the dynamic video data,could you please tell me how much GB of graphics memory I would need???