yolov3-pytorch训练报错

2020-11-11 04:07:31 3.7k 阅读 AI 写同款 · GEO 优化AI 选题数据监测一键分发 ›

⭐⭐⭐全网最详细YOLOv8教程，从环境配置到自己训练，再到数据集的制作，包的各位⭐⭐⭐ 问题原因：当你“pip install ultralytics” 时，torch和torchvision会作为它的依赖一起安装了，当执行“pip install torch torchvision torchaudio --index-url。”时，会卸载torch，然后再安装torch+cuda版本，但是不会自动卸载torchvision已有版本，然后再安装torchvision+cuda版本。🎉🎉🎉恭喜各位道友，已经可以训练模型了，现在就是制作自己的数据集，所谓申神通秘境，随心所欲，万物皆可识别也阅读详情

/pytorch/aten/src/ATen/native/cuda/IndexKernel.cu:60: lambda [](int)->auto::operator()(int)->auto: block: [0,0,0], thread: [41,0,0] Assertion `index >= -sizes[i] && index < sizes[i] && "index out of bounds"` failed.

/pytorch/aten/src/ATen/native/cuda/IndexKernel.cu:60: lambda [](int)->auto::operator()(int)->auto: block: [0,0,0], thread: [45,0,0] Assertion `index >= -sizes[i] && index < sizes[i] && "index out of bounds"` failed.

/pytorch/aten/src/ATen/native/cuda/IndexKernel.cu:60: lambda [](int)->auto::operator()(int)->auto: block: [0,0,0], thread: [47,0,0] Assertion `index >= -sizes[i] && index < sizes[i] && "index out of bounds"` failed.

Traceback (most recent call last):

  File "train.py", line 105, in <module>

    loss, outputs = model(imgs, targets)

  File "/home/ducong/anaconda3/envs/yolov3-pytorch/lib/python3.6/site-packages/torch/nn/modules/module.py", line 547, in __call__

    result = self.forward(*input, **kwargs)

  File "/home/ducong/ducong/YOLOv3/PyTorch-YOLOv3/models.py", line 260, in forward

    x, layer_loss = module[0](x, targets, img_dim)

  File "/home/ducong/anaconda3/envs/yolov3-pytorch/lib/python3.6/site-packages/torch/nn/modules/module.py", line 547, in __call__

    result = self.forward(*input, **kwargs)

  File "/home/ducong/ducong/YOLOv3/PyTorch-YOLOv3/models.py", line 188, in forward

    ignore_thres=self.ignore_thres,

  File "/home/ducong/ducong/YOLOv3/PyTorch-YOLOv3/utils/utils.py", line 306, in build_targets

    noobj_mask[b[i], anchor_ious > ignore_thres, gj[i], gi[i]] = 0

RuntimeError: CUDA error: device-side assert triggered

这是咋了啊，不太懂，有知道的吗

...全文

PyTorch YOLOv3-train.py 按照 https://gitee.com/sycret/PyTorch-YOLOv3/tree/master 上的方法去train自己的dataset报错了准备看一下为什么顺便加深下对yolov3代码的理解。 train $python3 train.py --model_def config/yolov3-custom.cfg --data_config config/custom.data error 看起来是个数组越界的错，去看看哪里错了。(还顺便发现classes.names上传错文件了已阅读详情

OngoingDC

CUDA高性能计算讨论 357 成员 615 内容

+ 关注 + 加入社区

yolov3-pytorch训练报错

相关推荐