OLMOE is a mixture of experts model from AllenAI.
7B model - 1.3b active parameters. Performs as well as Gemma2 3b but using open/free data and 1/3 of the active parameters.
OLMOE is a mixture of experts model from AllenAI.
7B model - 1.3b active parameters. Performs as well as Gemma2 3b but using open/free data and 1/3 of the active parameters.