Leheng Li

How to use mixed-precision training?

1

Hi, thanks for sharing your work! I want to save GPU memory during training. I follow this [link](https://zhuanlan.zhihu.com/p/375224982) to set up FP16 setting. I simply add `fp16 = dict(loss_scale='dynamic') `...

Question about KITTI raw dataset

6

Hi, Thanks for sharing your work! I am wondering when will you release the dataset pipeline for KITTI raw and other datasets. By the way, how to evaluate the network...

About onnx and pytorch version

3

When I run test_fcos_onnx_model.py.I encounted this: ` from fcos_core import _C ImportError: /home/dut-616/FCOS-master/fcos_core/_C.cpython-37m-x86_64-linux-gnu.so: undefined symbol: _ZN6caffe26detail37_typeMetaDataInstance_preallocated_32E ` I am using pytorch 1.4.0 and torchvision 0.5.0.dev20200107 py37_cu100. It seems like the...

Hi Xiangyu, Firstly, thanks for your great work on domain adaptation. Secondly, after reading the paper, i come up 2 question(if i have wrong understanding please point it out): 1.The...

Hi, thanks for open-sourcing your great work! I'm really impressed by the mesh result that you show in Figure 13. I'm interested in reproducing this result in driving scenes. Would...