/pytorch/aten/src/ATen/native/cuda/Loss.cu:102: operator(): block: [18,0,0], thread: [54,0,0] 断言 input_val >= zero && input_val <= one
失败。
/pytorch/aten/src/ATen/native/cuda/Loss.cu:102: operator(): block: [18,0,0], thread: [55,0,0] 断言 input_val >= zero && input_val <= one
失败。
/pytorch/aten/src/ATen/native/cuda/Loss.cu:102: operator(): block: [18,0,0], thread: [56,0,0] 断言 input_val >= zero && input_val <= one
失败。
/pytorch/aten/src/ATen/native/cuda/Loss.cu:102: operator(): block: [18,0,0], thread: [57,0,0] 断言 input_val >= zero && input_val <= one
失败。
/pytorch/aten/src/ATen/native/cuda/Loss.cu:102: operator(): block: [18,0,0], thread: [58,0,0] 断言input_val >= zero && input_val <= one
失败。
/pytorch/aten/src/ATen/native/cuda/Loss.cu:102: operator(): block: [18,0,0], thread: [59,0,0] 断言input_val >= zero && input_val <= one
失败。
追溯(最近的调用最先):
文件“run_toys.py”,第215行
loss = criterion(torch.reshape(out, [-1, dataset.out_dim]), torch.reshape(target, [-1, dataset.out_dim]))
文件“/usr/local/python3/lib/python3.6/site-packages/torch/nn/modules/module.py”,第727行
result = self.forward(*input, **kwargs)
文件“/usr/local/python3/lib/python3.6/site-packages/torch/nn/modules/loss.py”,第530行
return F.binary_cross_entropy(input, target, weight=self.weight, reduction=self.reduction)
文件“/usr/local/python3/lib/python3.6/site-packages/torch/nn/functional.py”,第2526行
binary_cross_entropy
input, target, weight, reduction_enum)
运行时错误:CUDA错误:设备端触发断言
代码
criterion = nn.CrossEntropyLoss()
loss = criterion(torch.reshape(out, [-1, dataset.out_dim]), torch.reshape(target, [-1, dataset.out_dim]))
loss = torch.mean(loss)
目标和输出的形状相同 # torch.Size([640, 32])
模型在我的CPU上运行良好,但在GPU上运行有问题。