Dear challenge organizer, Thank you for providing the instructions for the test phase of the challenge. I have a couple of questions regarding the testing environment: 1. Will the Docker image be tested in an environment with a CPU or GPU on your end? If it needs to run on a GPU, could you please provide the specifications, such as the CUDA and cuDNN versions, GPU model, available memory, and compute capability? 2. After submitting a Docker image, will there be any confirmation to inform us whether the Docker image can successfully run in your environment? If the image does not run successfully, will we have an opportunity to fix it ? Thank you for your assistance. I look forward to your response.

Created by Hang Min HollieMin19
Thanks for the question; we will try our best to facilitate this; please specify the desired CUDA or other run-time version you need in email after your submission.
Thank you very much for the information. Could you please also let us know the NVIDIA driver version (nvidia-smi) to confirm whether the driver supports higher versions of CUDA (e.g., CUDA >= 11.7) and PyTorch >= 2.0 in Docker? Thank you for your assistance. We look forward to your response.
Thank you for your inquiry. The Docker image will be tested in an environment with a GPU. The specifications for the GPU environment are as follows: ? CUDA Version: 11.3 ? cuDNN Version: 8.2 ? GPU Model: NVIDIA GTX 3090 ? Available Memory: 45 GB After submitting your Docker image, you will receive a confirmation informing you whether the image successfully runs in our environment. If the image does not run successfully, you will indeed have an opportunity to fix it and resubmit. Please let us know if you need any additional information.

Clarification on Test Phase Environment for Docker Image Submission page is loading…