EVA Qwen2.5 14B
eva-unit-01/eva-qwen-2.5-14b
Created Sep 30, 202432,768 context
A model specializing in RP and creative writing, this model is based on Qwen2.5-14B, fine-tuned with a mixture of synthetic and natural data.
It is trained on 1.5M tokens of role-play data, and fine-tuned on 1.5M tokens of synthetic data.