This model offers the best performance/calc-cost for near real-time purposes involving Korean-related tasks. I converted and quantised the model to MLX for private use.

Downloads last month
19
Safetensors
Model size
7.66B params
Tensor type
BF16
·
U32
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for woung717/EVEE-Korean-Instruct-7B-v2.0-Preview-MLX-6bits

Quantized
(4)
this model