lucasjinreal / DCNv2_latest

DCNv2 supports decent pytorch such as torch 1.5+ (now 1.8+)
BSD 3-Clause "New" or "Revised" License
637 stars 124 forks source link

使用GPU训练时出现error in modulated_deformable_im2col_cuda: out of memory #62

Open creater-zq opened 2 years ago

creater-zq commented 2 years ago

您好,您的代码解决了我遇到的一些bugs。 但是在训练时我遇到一些错误。

跑着跑着,退出了。

  File "/DCNv2/dcn_v2.py",  in forward
    return dcn_v2_conv(input, offset, mask, self.weight, self.bias, self.stride, self.padding, self.dilation, self.deformable_groups,)
  File "DCNv2/dcn_v2.py",  in forward
    output = _backend.dcn_v2_forward(
RuntimeError: CUDA error: an illegal memory access was encountered
CUDA kernel errors might be asynchronously reported at some other API call,so the stacktrace below might be incorrect.
For debugging consider passing CUDA_LAUNCH_BLOCKING=1.
已放弃 (核心已转储)
SilyRab commented 10 months ago

我也是这样,后来解决了吗?我在1卡上跑,但是dcn有一部分东西在0卡上,然后数据或者网络参数不在一张卡上,就报和你一样的错误了

creater-zq commented 10 months ago

我也是这样,后来解决了吗?我在1卡上跑,但是dcn有一部分东西在0卡上,然后数据或者网络参数不在一张卡上,就报和你一样的错误了

我后来修改了一些代码,跑通了,不过我发现他这个不适合我的数据集,就没搞了,是好久之前了