[Feature] Support Llama Maverick tuning and evaluation #1667
Description
Feature request
We currently have configs for Llama 4 Scout. We'd also like to have examples for Maverick, which will be trickier given its size.
Motivation / references
https://github.com/oumi-ai/oumi/tree/main/configs/recipes/llama4
https://huggingface.co/meta-llama/Llama-4-Maverick-17B-128E-Instruct
Your contribution
Can advise