mlabonne/OrpoLlama-3-8B
TEXT GENERATIONConcurrency Cost:1Model Size:8BQuant:FP8Ctx Length:8kPublished:Apr 18, 2024License:otherArchitecture:Transformer0.1K Cold

OrpoLlama-3-8B is an 8 billion parameter language model developed by mlabonne, fine-tuned from Meta-Llama-3-8B using the ORPO method on the mlabonne/orpo-dpo-mix-40k dataset. This model utilizes a context window of 8192 tokens and follows the ChatML template. It demonstrates improved performance over Llama-3-8B-Instruct on specific benchmarks like GPT4All and TruthfulQA, making it suitable for general conversational AI and question-answering tasks.

Loading preview...