RuntimeError: lines.type().is_cuda()INTERNAL ASSERT FAILED at ““, please report a bug to PyTorch.

Posted 2022-12-23 沉迷单车的追风少年

项目场景：

复现sketch R2CNN代码，neuralline rasterize操作。

---

问题描述

完整报错：

Traceback (most recent call last):
  File "attention_map.py", line 120, in <module>
    output = model(x)
  File "/root/anaconda3/envs/torch1/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1102, in _call_impl
    return forward_call(*input, **kwargs)
  File "attention_map.py", line 108, in forward
    images = RasterIntensityFunc.apply(points, intensities, self.img_size, self.thickness, self.eps, self.device)
  File "/root/sketchtext/sketch-diffusion-main/neuralline/rasterize.py", line 35, in forward
    line_weight_map, locks, img_size, thickness, eps)
RuntimeError: lines.type().is_cuda()INTERNAL ASSERT FAILED at "/root/sketchtext/sketch-diffusion-main/neuralline/rasterize_cuda.cpp":43, please report a bug to PyTorch. lines must be a CUDA tensor