ChaoticNeutrals/Eris_Floramix_DPO_7B

TEXT GENERATIONConcurrency Cost:1Model Size:7BQuant:FP8Ctx Length:8kPublished:Mar 7, 2024License:otherArchitecture:Transformer0.0K Cold

Eris Floramix DPO is a 7 billion parameter language model developed by ChaoticNeutrals, featuring an 8192-token context length. This model is a DPO fine-tune, combining elements of Eris Remix DPO and Flora DPO, with further fine-tuning on the Synthetic_Soul_1k dataset. It is specifically optimized for roleplay tasks, leveraging a DPO dataset focused on roleplay and Alpaca-style interactions.

Loading preview...

Eris Floramix DPO Overview

Eris Floramix DPO is a 7 billion parameter language model from ChaoticNeutrals, built upon a unique blend of existing models. It integrates the architectures and training methodologies of both Eris Remix DPO and Flora DPO. A key aspect of its development involved further fine-tuning on the Synthetic_Soul_1k dataset, enhancing its capabilities through diverse synthetic data.

Key Capabilities

  • DPO Fine-tuned: The model has undergone Direct Preference Optimization (DPO) using a specialized dataset.
  • Roleplay Optimization: It is specifically fine-tuned with a DPO dataset (athirdpath/DPO_Pairs-Roleplay-Alpaca-NSFW-v1-SHUFFLED) designed for roleplaying scenarios and Alpaca-style instruction following.
  • Context Length: Supports an 8192-token context window, allowing for more extensive and coherent interactions.

Good For

  • Roleplaying Applications: Excels in generating responses suitable for interactive roleplay scenarios.
  • Instruction Following: Benefits from Alpaca-style data, making it proficient in understanding and executing instructions.
  • Creative Text Generation: Its fine-tuning on diverse datasets suggests strong capabilities in generating varied and engaging text.

Popular Sampler Settings

Top 3 parameter combinations used by Featherless users for this model. Click a tab to see each config.

temperature
top_p
top_k
frequency_penalty
presence_penalty
repetition_penalty
min_p