Antoine Chauvin

Antoine Chauvin

Platform Engineer

Paris, France

Joined April 8, 2024

Karma 34

Antoine Chauvin

1

Posted by antoine

@bhagya @abdullah can you send your submission to antoine@aqora.io and attach files at:
  • $HOME/.local/state/aqora/aqora.log* if you are on Linux,
  • $HOME/Library/Caches/aqora/aqora.log* if you are on macOS,
  • %USERPROFILE%\AppData\Local\aqora\aqora.log* if you are on Windows.
Antoine Chauvin

2

Posted by antoine

Hello @riya !
You can submit on behalf of the team using your personal account or by creating an organisation and adding members to it here:
When you run aqora template in a terminal, you should be prompted to choose an account to submit a submission on.
You can also attach a submission to an organisation you are part of by adding a configuration inside pyproject.toml. Just add a key entity = "username_here" under the [tool.aqora] group!
Don’t hesitate to let me know if I can be of further help,
Enjoy 👋
Antoine Chauvin

3

Posted by antoine

how do you train the model with the data ?
I must say that it really depends on the ML library you are likely to use. Which one are you using?
i am not sure where to put the dataset_type = "train"
dataset = FIDataset(dataset_type,".aqora/data/data")
You should use the dataset variable here as the input for your ML implementation.
I also recommend that your read this thread if you are having some troubles with FIDataset.
Let me know if you need further assistance!
Antoine Chauvin

1

Posted by antoine

How about this one?
dataset = FIDataset(dataset_type, "../.aqora/data/data")
Antoine Chauvin

1

Posted by antoine

Could you try this line instead?
dataset = FIDataset(dataset_type, ".aqora/data/data")
Antoine Chauvin

1

Posted by antoine

Hello Pascal!
Which issue exactly are you facing, please?
Antoine Chauvin

2

Posted by antoine

Hello Priyanshi!
The competition will continue remotely from January, 24th (8:30AM CET) to March, 28th (12:00PM CET), allowing participants from across the globe to collaborate and innovate.
Everyone is welcome to participate!
Antoine Chauvin

2

Posted by antoine

Hello Aarshita!
I updated the use-case to relax dependency requirements. I downgraded torch @ 2.2.2 and numpy @ 1.26.4, which should enable you to run your submission on a macOS Intel machine. To update, you will have to:
  1. Downgrade your dependency requirements too in your file pyproject.toml,
  2. Run aqora clean,
  3. Run aqora install.
Please let me know how it goes for you, I'll be happy to further help you sort things out!
Antoine Chauvin

3

Posted by antoine

Thank you for helping us spot this bug! Here is how you can load the FI dataset:
dataset = FIDataset(dataset_type, ".aqora/data/data")
We will update soon the use-case sources so it can be more easily used in the future!
Antoine Chauvin

2

Posted by antoine

Hello @peterys!
It seems that torch v2.4.1 supports Python 3.8-3.12 on Windows (amd64 only), Linux (both amd64 and arm64), and macOS (arm64 only).
You can find your Python version and platform using these commands:
$ python3 -V
Python 3.12.8
$ python3 -m platform
macOS-15.2-arm64-arm-64bit
I can attest this install works well with the quantum-signals-lob use-case.
You may have a too recent Python installed, fortunately you can easily install whichever Python of your choosing using uv python install X.Y.X.
Please let me know if you need further help!