Jinlai Zhang

no, still cannot run

> Hi Do you need only hdf5 file?? Or all of the ShapeNet dataset?? Whole dataset is very big. 2019년 1월 25일 (금) 오후 4:35에 tangjilin 님이 작성: > […](#)...

```shell File "utils/egogesture_json.py", line 86, in val_csv_path, dst_json_path) File "utils/egogesture_json.py", line 53, in convert_egogesture_csv_to_activitynet_json train_database = convert_csv_to_dict(train_csv_path, 'training', labels) File "utils/egogesture_json.py", line 9, in convert_csv_to_dict data = pd.read_csv(csv_path, delimiter=' ',...

> Hi, > > I am running on ubuntu 16.04 with all prerequisites(libqt4-dev, qt4-qmake, libvtk5-dev, libpcl-1.7-all-dev, libsvm-dev, libsvm-tools) installed and I am able to run till cmake after that facing...

The top one is generated by evo kitti, the bottom one is plot by coodrdinates.

same problem

> Hi, I have encountered the same problem. Have you solved it? not yet.

> @callmeray > **I have already solved the problem above , the problem is caused by using multiple GPU.** > I add **os.environ['CUDA_VISIBLE_DEVICES']='0'** in the Test script ,then it runs...

succeed by replace `TH_INDEX_BASE` with 0

In most cases, the GPU used is RTX series, such as 2080ti, and the cuda used is cuda9.0. The correct one should install cuda10.0 and the corresponding Pytorch The final...