Is there anything that would get me on the right direction to make a submission as it is supposed to be done? Even though on my local computer the .json file is created after running the program with setting the input and output paths with os.environ[“AICROWD_TEST_IMAGES_PATH”] = “…” and os.environ[“AICROWD_PREDICTIONS_OUTPUT_PATH”] = “…”, I have no clue on what and how to put on gitlab so that everything is as it is supposed to be to have a runable submission.
I have the random submission successfully uploaded and that went smooth, but I want to use pytorch, I want to upload .pth model but after adding with LFS and pushing I don’t see it in the repository.
Can someone please help me to get onto the right track? I am a little frustrated as I’ve worked on it some hours for a few days and submission is the hardest problem that one have to solve for participation.
Help, please?
You can start with this baseline submission that’s based on MMDetection, which itself is based on PyTorch:
Clone the above repository, change your username in the aicrowd.json file, and then follow the instructions here (from # Add AICrowd git remote endpoint forward):
This way, you should be able to upload a valid submission and get on the leaderboard with the baseline result (AP=0.478 and AR=0.687).
After that, you can simply swap the included .pth model with your own - just change the run.sh script with the correct model filename. Keep in mind that the baseline submission also has scripts for training & evaluation, so you can use these tools to your advantage.
I ran into some problems while trying to use my own models with this baseline submission, but after cloning the repo again and changing only the model file, it works once more (the repo was updated recently).
I also don’t see my repo updated after the latest submissions (it did update for the first couple of submissions in February, and after that it stopped updating with every submission), but everything is successfully uploaded nonetheless. So just be mindful that you upload your model to LFS and push it, and you should be fine.
Best of luck!
1 Like
Thank you very much for this, I hope that everything is in here that I need to know to make a legit submission.
Thank you!
1 Like
I’m not quite sure what errors you ran into while changing the model, I’ve made a notebook that does exactly that and makes a submission directly on colab. You can check it out here Train using mmdetection and submit via Colab!
Regards
Shraddhaa
1 Like