🏥 This Week on Zindi: Hardware Restriction Querry

Malawi Public Health Systems LLM Challenge

Helping Malawi

$2 000 USD

Challenge completed over 1 year ago

Skills you will learn

Questioning and Answering

Generative AI

407 joined

74 active

Info Data Chat Leaderboard

Start

Jan 24, 24

Mar 03, 24

Reveal

Mar 03, 24

Nayal_17

Hardware Restriction Querry

Notebooks · 27 Jan 2024, 04:52 · 6

@zindi

Does this "You are limited to the following or equivalent hardware specifications: Multi-core processor (e.g., Intel Core i7/i9, AMD Ryzen 7/9) for efficient processing." means that we can't use GPUs and TPUs?

Discussion 6 answers

avt_nyanja

The competition restricts the task to open source LLMs that do not require GPU's and training must be compatible with multi core processer capped at Intel Core i7/i9, AMD Ryzen 7/9.

29 Jan 2024, 06:54

Upvotes 1

Nayal_17

@avt_nyanja can we use colab or kaggle cpus.

Kaggle:  model_name : AMD EPYC 7B12
Colab: model_name : Intel(R) Xeon(R) CPU @ 2.20GHz

replied to avt_nyanja29 Jan 2024, 09:56

Upvotes 0

duongkstn

Do that mean we CAN NOT use GPU to train model ?

replied to avt_nyanja31 Jan 2024, 09:45

Upvotes 0

Nayal_17

@duongkstn You don't have to train the model. Check some blogpost to start with. As there is no GPU allowed you can't quantize the model but still you can do question answering.

replied to duongkstn31 Jan 2024, 11:30

Upvotes 0

duongkstn

you are wrong ! "The trained LLM should be able to run locally.". So you can use your trained model

replied to Nayal_1731 Jan 2024, 12:02

Upvotes 0

Mutisyaboy

Could we get the official statement from Zindi on the matter of GPU usage?

12 Feb 2024, 06:32

Upvotes 0

Join the largest network for
data scientists and AI builders

About FAQs

Status