使用分布式训练报错

[zhn6818]

将distributed = True RuntimeError: Output 0 of SliceBackward0 is a view and is being modified inplace. This view is the output of a function that returns multiple views. Such functions do not allow the output views to be modified inplace. You should replace the inplace operation by an out-of-place one. 这个问题如何解决，望指教

[bubbliiiing]

运行指令是？ 截图看看

[zhn6818]

CUDA_VISIBLE_DEVICES=4,5,6,7 python -m torch.distributed.launch --nproc_per_node=4 train.py

[bubbliiiing]

有截图吗 现有的看不出来，感觉可能是环境问题，0 0，你是什么环境

[Jamesgender]

这个问题主要是版本的问题，然后如果想修改的话，需要将报错行 （估计在yolo training105行） 更改为：

output[..., :2] = (output[..., :2] + grid.type_as(output)) * stride

    temp_output = output.clone()
    temp_output[..., :2] = (output[..., :2] + grid.type_as(output)) * stride
    output = temp_output
    return output, grid

[bubbliiiing]

为啥捏0 0有理由吗

[Jamesgender]

RuntimeError: Output 0 of SliceBackward0 is a view and is being modified inplace.

https://github.com/SHI-Labs/FcF-Inpainting/issues/26

[bubbliiiing]

6啊baby

[bubbliiiing]

我在想要不要改一下

[zhn6818]

给大佬鞠躬

[RykerYang]

我也给大佬磕一个