stanley
stanley
I surface the same problem
For the regulation, I can't upload full logs. but I can provide the useful. 2022-08-16 04:07:43,711 - mmdet - INFO - Environment info: ------------------------------------------------------------ sys.platform: linux Python: 3.7.13 (default, Mar...
total_epochs = 24 find_unused_parameters = False runner = dict(type='EpochBasedRunner', max_epochs=24) gpu_ids = range(0, 1) device = 'cuda' 2022-08-16 04:07:46,570 - mmdet - INFO - Set random seed to 0, deterministic:...
I change the env as follow: sys.platform: linux Python: 3.8.10 (default, Jun 4 2021, 15:09:15) [GCC 7.5.0] CUDA available: True GPU 0: Quadro RTX 5000 CUDA_HOME: /usr/local/cuda NVCC: Build cuda_11.3.r11.3/compiler.29745058_0...
For my case, I use mmcv source to install and solve it.
I have finish the train. But when I go to test, a wired error happened. `bash tools/dist_test.sh projects/configs/petrv2/petrv2_BEVseg.py work_dirs/petrv2_BEVseg/latest.pth 1 --eval bbox ` The error like: `CHILD PROCESS FAILED WITH...
I find the problem is that my local pc is one gpu the original test command is unsuitable for me. I change dist_test.sh to `python3 -m torch.distributed.launch --nproc_per_node=$GPUS --master_port=$PORT \...
I also face the same problem. I install colmap for https://github.com/B1ueber2y/colmap.git. But the problem still exists.
> add an option, --imageSize 32 when you run it. why we must set the image size as 32, the paper said the imagesize is 128*128/64*64. How can I fix...
Test at (loop 49, iters: 25399) with mean_accu: 86.8898. I think the result is acceptable.