Has filled their profile page
Hi! Does someone faced this error “RuntimeError: Not compiled with GPU support” using detectron2? I tried docker images from pytorch and nvidia, but in each case have the same error. Also in image build stage after running “python -c ‘import torch; from torch.utils.cpp_extension import CUDA_HOME; print(torch.cuda.is_available(), CUDA_HOME)’”
output : No CUDA runtime is found, using CUDA_HOME=’/usr/local/cuda’
It so weird, everything worked well before
I can’t find new dataset) there are links only to old versions
Hi! Can you say please what’s wrong with my submissions v17 and v16?
Ok, thanks! What GPU memory limit for evaluation?
And also what is “Evaluate Predictions: No more retries left”?
Hi! My submission failed with error “Generate Predictions On Validation Data: Max duration limit exceeded”. What does it mean? Does evaluation has time limit? I haven’t found any information about it.
Hi. I have a problem with submissions, but two days ago it was ok. I try to submit from two of my competition repos - so I commit, and add a new tag with increasing number. But my commits doesn’t start a new evaluation issue. Can you help me please?