Dear organizers, We submitted the docker image after successfully testing it locally and received the system mail for our submission. 1: We built the docker image using: ``` --platform=linux/amd64 ``` Can you confirm that the docker images will be evaluated on AMD64 hardware? 2: As mentioned on the instructions page, I would kindly request that a more powerful GPU like the suggested NVIDIA A100 is used for the inference of our method. Our team name is "imi-graz". 3: Will we receive a confirmation after our method was successfully evaluated? Conversely, will we be informed in case there are any issues with running the docker image? Thank you

Created by Franz Thaler thaler
Dear Paula, thank you for the clarifications. :) Since you mentioned your limitations regarding the evaluation on an A100 GPU, I can give you more context. We tested our docker image on an RTX 3090 Ti with 24 GB of memory on which our computed average runtime is approximately 25 minutes. However, the runtime of our method depends on the size of the mask and can therefore last longer for individual cases. The other GPUs you have available should be sufficient to do inference, however, I am mostly worried about the inference time limit of 1 hour per sample as I cannot test these GPUs and don't know the worst case in the test set. So, assuming that we will not be disqualified in case the inference time for individual cases exceeds the 1 hour limit, you can of course also run our docker image on one of the other GPUs. Best, Franz
@thaler Dear Franz, thank you for your question! We can confirm that your docker will be evaluated on AMD64 hardware, so I expect no problems in that regard. :) I have noted down that you (team imi-graz) have requested an A100 for running it. Currently we see no reason why it wouldn't be possible to run it on the A100. But I'll notify you in case there are any unexpected updates. (The only issue is that the cluster that we use to get access to the A100 has a scheduling system and if there are many jobs it can cost us time before we can actually test it. Additionally we do not have unlimited resources regarding computation on that cluster, that is why we try to avoid it if possible). In case your docker does not run on our end, we will definitely notify you. Strictly speaking, each team can only submit one docker image, but if you contact us and show effort then we will most likely grant you a second attempt. However, please note that no third, fourth, ... attempt can be granted! It is each team's responsibility to make sure that the container runs. I hope this answered your questions. Let me know if there's anything else you would like to know! :) Best Paula 2026 BraTS Team

Task 4 Docker submission: Inference hardware and verification page is loading…