Brucethemoose

Yi-34B-200K-RPMerge

brucethemoose

54

The Yi-34B-200K-RPMerge model is a merge of several 34B parameter Yi models created by maintainer brucethemoose. The goal of this merge is to produce a model with a 40K+ context length and enhanced storytelling and instruction-following capabilities. It combines models like DrNicefellow/ChatAllInOne-Yi-34B-200K-V1, migtissera/Tess-34B-v1.5b, and cgato/Thespis-34b-v0.7 which excel at instruction following and roleplaying, along with some "undertrained" Yi models like migtissera/Tess-M-Creative-v1.0 for enhanced completion performance. Model inputs and outputs The Yi-34B-200K-RPMerge model is a text-to-text model, taking in text prompts and generating text outputs. Inputs Text prompts for the model to continue or respond to Outputs Generated text continuations or responses to the input prompts Capabilities The Yi-34B-200K-RPMerge model demonstrates strong instruction-following and storytelling capabilities, with the ability to engage in coherent, multi-turn roleplaying scenarios. It combines the instruction-following prowess of models like ChatAllInOne-Yi-34B-200K-V1 with the creative flair of models like Tess-M-Creative-v1.0, allowing it to produce engaging narratives and responses. What can I use it for? The Yi-34B-200K-RPMerge model would be well-suited for applications requiring extended context, narrative generation, and instruction-following, such as interactive fiction, creative writing assistants, and open-ended conversational AI. Its roleplaying and storytelling abilities make it a compelling choice for building engaging chatbots or virtual characters. Things to try Experiment with the model's prompt templates, as the maintainer suggests using the "Orca-Vicuna" format for best results. Additionally, try providing the model with detailed system prompts or instructions to see how it responds and tailors its output to the given scenario or persona.

Read more