how to run on cpu enviroment

[peichenzhong]

bash ./scripts/shrec/train.sh
THCudaCheck FAIL file=/opt/conda/conda-bld/pytorch_1549630534704/work/torch/csrc/cuda/Module.cpp line=34 error=35 : CUDA driver version is insufficient for CUDA runtime version
Traceback (most recent call last):
File "train.py", line 9, in
opt = TrainOptions().parse()
File "/home/peichen/python_test/MeshCNN/options/base_options.py", line 54, in parse
torch.cuda.set_device(self.opt.gpu_ids[0])
File "/home/peichen/anaconda3/envs/meshcnn/lib/python3.6/site-packages/torch/cuda/init.py", line 264, in set_device
torch._C._cuda_setDevice(device)
RuntimeError: cuda runtime error (35) : CUDA driver version is insufficient for CUDA runtime version at /opt/conda/conda-bld/pytorch_1549630534704/work/torch/csrc/cuda/Module.cpp:34

[ankitpatnala]

add an argument --gpu_ids -1/

[peichenzhong]

and python test.py --gpu_ids -1 --dataroot datasets
Running Test
loaded mean / std from cache
Traceback (most recent call last):
File "test.py", line 25, in
run_test()
File "test.py", line 12, in run_test
model = create_model(opt)
File "/home/peichen/python_test/MeshCNN/models/init.py", line 3, in create_model
model = ClassifierModel(opt)
File "/home/peichen/python_test/MeshCNN/models/mesh_classifier.py", line 33, in init
self.gpu_ids, opt.arch, opt.init_type, opt.init_gain)
File "/home/peichen/python_test/MeshCNN/models/networks.py", line 102, in define_classifier
opt.resblocks)
File "/home/peichen/python_test/MeshCNN/models/networks.py", line 132, in init
norm_args = get_norm_args(norm_layer, self.k[1:])
File "/home/peichen/python_test/MeshCNN/models/networks.py", line 38, in get_norm_args
raise NotImplementedError('normalization layer [%s] is not found' % norm_layer.func.name)
NotImplementedError: normalization layer [BatchNorm2d] is not found

how to deal with it ??thank

[peichenzhong]

torch_mask = torch.from_numpy(edges_mask.copy())
TypeError: can't convert np.ndarray of type numpy.bool_. The only supported types are: double, float, float16, int64, int32, and uint8.

[ankitpatnala]

you can convert the type to uint8
I had the same issue as i created the environment from the repo's environment.yml file which has pytorch 1.0 but the current code's dependency is pytorch 1.2.0. So the author recommended me to use the old checkout of this branch which somehow works.

[peichenzhong]

so i should updata my pytorch into 1.2.0 ??? i have a try

[ankitpatnala]

i dont know the author suggested me to use the old checkout

please follow this link:
#30 (comment)

but author has changed the environment file in her latest commit which was 6 hours ago. May be you can update your branch and rerun the scripts.

[peichenzhong]

thank , this project is running ,but i have a problem how to watch the layer or models on the tensorboardx???

[peichenzhong]

i want to watch the graph ,what should i do

[ankitpatnala]

I never used tensorboard but you can find it's documentation .you can save the checkpoints on a file and try to open the tensboard logdir on other terminal upon which it will give you a local host IP which you can run in your browser. Before that install tensorboard packages as it is not present in the environment.yml file

[peichenzhong]

i hava some problem that how could i compute the input edge feature which is a 5-dimensional vector every edge??? Anyway , i am confused about the edges value.

[peichenzhong]

i hava some problem that how could i compute the input edge feature which is a 5-dimensional vector every edge??? Anyway , i am confused about the edges value.

[ranahanocka]

hi @peichenzhong ,

i hava some problem that how could i compute the input edge feature which is a 5-dimensional vector every edge??? Anyway , i am confused about the edges value.

I don't understand your question. Please go through the examples in this Github and follow the instructions in the README, I think it will clarify most of your problems. Also, you can read through the Github Issues, which might also answer some questions.

Thanks & Good luck!
-Rana