aifeifei798/llama3-8B-DarkIdol-2.3-Uncensored-32K is an 8 billion parameter Llama 3-based language model, fine-tuned for uncensored responses and specialized in diverse role-playing scenarios. This merged model, built upon winglian/Llama-3-8b-64k-PoSE, supports a theoretical 64K context length (tested up to 32K) and is optimized for quick, scholarly responses. It is designed for text generation and code output, with specific quantizations for efficient CPU-only operation and optional vision capabilities.
No reviews yet. Be the first to review!