MERaLiON/LLaMA-3-MERaLiON-8B-Instruct
TEXT GENERATIONConcurrency Cost:1Model Size:8BQuant:FP8Ctx Length:8kPublished:Dec 12, 2024License:meralion-public-licenseArchitecture:Transformer0.0K Cold

MERaLiON/LLaMA-3-MERaLiON-8B-Instruct is an 8 billion parameter multilingual large language model developed by I²R, A*STAR, built upon the Llama-3-8B architecture with an 8192-token context length. It is enhanced through continued pretraining on over 120 billion tokens, with a focus on English, Chinese, and Indonesian, and optimized for instruction-following via model merging. This model excels in multilingual understanding and generation, particularly in Southeast Asian language contexts, demonstrating improved performance on benchmarks like Cross-MMLU and IndoMMLU.

Loading preview...