Mobile AI Workshop General Organization Questions

[Andrey Ignatov]

Any questions related to general MAI Workshop and Challenges organization can be asked in this thread.

 

[Pu Lu]

When will these challenges be ready？

 

[Raymond]

Will those challenges be held this year? It's already May now and the challenges are not released.

 

[Andrey Ignatov]

Hi @Pu Lu and @Raymond,

The challenges are already open: https://ai-benchmark.com/workshops/mai/2022/

 

[dbwp031]

Hi, In http://www.wikicfp.com/cfp/servlet/event.showcfp?eventid=171915&copyownerid=90126, the 2023 challenges should be opened on 21 March. But the challenges are not released until now. When these challenges will be start?

 

[jin23]

Dear orginizers,
When does the challenge start?
The main challenge website says it starts in March 2023, but it's May 10th, far from March.
I'd like to know the competition will be held this year or not.
Would you please announce any update?

 

[nickarcos]

Could you please share the articles published in the workshops from 2023 and 2024? It's a shame that they are no longer recorded so that people on the other side of the world like me can have this precious knowledge.

 

[morty0229]

I am writing to seek clarification regarding the dataset used for the Mobile AI 2025 sRGB Image Enhancement Challenge. In the competition data description section, it is mentioned that the data consists of "iPhone-canon" images. However, upon reviewing the provided dataset link, I noticed that the number of images available is 160K, which does not align with the stated 162K in the description.

Whether the 162K figure refers to the "sony-canon" dataset instead?

 

[Johon Reese]

I am participating in the Mobile AI 2025 sRGB Image Enhancement Challenge and found some errors in Final test phase submission guidelines.
(https://codalab.lisn.upsaclay.fr/competitions/21564#participate)
Since the task involves 3-channel RGB-to-RGB processing, I suspect the input shape’s 4th channel (currently 4) might be causing a mismatch.
Could you confirm the exact input image height and width required for the challenge?
Is the input expected to be strictly 3-channel RGB (e.g., [H, W, 3]), or does the 4th channel serve a specific purpose?
My model is designed for RGB enhancement, so the 4-channel input might be a configuration oversight. I appreciate your guidance to resolve this discrepancy.

 

[cblks]

CodaLab - Competition

codalab.lisn.upsaclay.fr

 

[cblks]

您好，我目前正在参加移动 AI 研讨会 @ CVPR 2025，实时逼真散景渲染挑战赛，下载此数据集时，显示 404NO FOUND。我该如何解决这个问题？

 

[caillen]

我正在参加图像增强赛道，但是一直在排队没有通过，这是为什么？

 

[caillen]

您好，我目前正在参加移动 AI 研讨会 @ CVPR 2025，实时逼真散景渲染挑战赛，下载此数据集时，显示 404NO FOUND。我该如何解决这个问题？

打扰一下，你的赛道的注册已经通过了吗

 

[moonmina]

I would like to ask whether there are any plans to postpone the schedule for this MAI 2025 challenge.
As the submission deadline for the model file and factsheet is approaching, we have yet to receive updates regarding the inference speed and performance on the NPU.

 

[Alex_Fall]

I am participating in the Mobile AI 2025 sRGB Image Enhancement Challenge and found some errors in Final test phase submission guidelines.
(https://codalab.lisn.upsaclay.fr/competitions/21564#participate)
Since the task involves 3-channel RGB-to-RGB processing, I suspect the input shape’s 4th channel (currently 4) might be causing a mismatch.
Could you confirm the exact input image height and width required for the challenge?
Is the input expected to be strictly 3-channel RGB (e.g., [H, W, 3]), or does the 4th channel serve a specific purpose?
My model is designed for RGB enhancement, so the 4-channel input might be a configuration oversight. I appreciate your guidance to resolve this discrepancy.

Hi, I have also encountered this problem.May I ask how you solved it?

 

[Andrey Ignatov]

Hi, I have also encountered this problem.May I ask how you solved it?

Since the task involves 3-channel RGB-to-RGB processing, I suspect the input shape’s 4th channel (currently 4) might be causing a mismatch.
Could you confirm the exact input image height and width required for the challenge?
Is the input expected to be strictly 3-channel RGB (e.g., [H, W, 3]), or does the 4th channel serve a specific purpose?

You are right, that was an incorrect info, we've updated it.