allganize/Llama-3-Alpha-Ko-8B-Evo
TEXT GENERATIONConcurrency Cost:1Model Size:8BQuant:FP8Ctx Length:8kPublished:May 23, 2024License:llama3Architecture:Transformer0.0K Warm

allganize/Llama-3-Alpha-Ko-8B-Evo is an 8 billion parameter language model developed by allganize, built upon the Meta-Llama-3-8B architecture. This 'Evo' model is designed as a base for fine-tuning, excelling in complex language tasks and logical reasoning in both Korean and English through Evolutionary Model Merging. It demonstrates strong performance on benchmarks like LogicKor, making it suitable for diverse and demanding bilingual applications.

Loading preview...