
Hello @avt_nyanja and @Zindi
Now that this competition is nearing the end, I Would LIke to ask for some clarification.
It was stated clearly that you need models that are NOT GPU dependent and that can run on CPU.
So where are you guys going to test these models. I presume it's in the laptop specs that you provided not google collab right? and you can correct me if I am wrong.
So if you are going to perform the evaluations on the laptop specs you mentioned can we provide scripts instead of notebooks as part of our final submissions? At the end these platforms like llama.cpp and ollama both run on the terminal no?
Or Docker YAML files that when built will spin docker containers that'll make it easy to work with the solutions.
Could You Share Documention of What you did to Achieve Score after comptition ends For eductional Purpose?
I will
Hi Kolesh, yes scripts and/or notebooks are acceptable for any competition.
At Zindi our aim for all competitions is for your solutions to be of value to the client. This varies from client to client based on their needs and their data journey.
Some clients find a notebook easier to engage with and learn from as they can itterate through it step wise. Notebooks also allow for easy visualisation of charts. We have found that all our clients find value in the EDA.
Other clients who are further in their data journey and are ready to implement from the close of a competition would prefer scripts but they have the experience to take what is needed from a notebook and turn it into scripts as needed.
We will try indicate on competitions which we think would be better for a client but at the moment it is a judgement call, what is easier for you and what will push you to be a better coder.