JieQin
JieQin
sys.platform: linux Python: 3.7.3 (default, Jan 22 2021, 20:04:44) [GCC 8.3.0] CUDA available: True GPU 0,1,2,3,4,5,6,7: Tesla V100-SXM2-32GB CUDA_HOME: /usr/local/cuda NVCC: Cuda compilation tools, release 11.3, V11.3.109 GCC: x86_64-linux-gnu-gcc (Debian...
1, running command: bash scripts/dist_train.sh 0,1,3,4,5,6,7 configs/base_nir/deeplabv3plus_r101.py 2, training config is too complex.
sys.platform: linux Python: 3.7.3 (default, Jan 22 2021, 20:04:44) [GCC 8.3.0] CUDA available: True GPU 0,1,2,3,4,5,6,7: Tesla V100-SXM2-32GB CUDA_HOME: /usr/local/cuda NVCC: Cuda compilation tools, release 11.3, V11.3.109 GCC: x86_64-linux-gnu-gcc (Debian...
Yes, I start it with tools/dist_train.sh and the number of gpus is 8.
Logically, there will be no problem of predefined size errors. You can provide the specific code location where the error is reported to facilitate locating the problem.
You should check whether the input of the ''args.train_list'' is correct. The normal input should be ''voc12/train.txt" or "voc12/val.txt".
You can get the model ckpt for demo in [model](https://drive.google.com/file/d/1X0oWfcpZo5bDkyFw7xiGBk_Yqx5gxhj_/view?usp=sharing).
You should get cam images by making the "make_cam_pass" as True.
I have updated the code. Please retry it.
Thanks for following our work, the segmentation code refers to [DeepLab v2](https://github.com/kazuto1011/deeplab-pytorch).