jambroz/sixtyoneeighty-7b-dpo

TEXT GENERATIONConcurrency Cost:1Model Size:7BQuant:FP8Ctx Length:8kPublished:Apr 5, 2024License:apache-2.0Architecture:Transformer Open Weights Cold

The jambroz/sixtyoneeighty-7b-dpo is a 7 billion parameter language model developed by sixtyoneeighty, fine-tuned for chat applications. This model utilizes a DPO (Direct Preference Optimization) finetuning approach, specifically trained on the Intel/orca_dpo_pairs dataset. It is designed for conversational tasks, building upon the jambroz/sixtyoneeighty-7b base model.

Loading preview...

Model Overview

The jambroz/sixtyoneeighty-7b-dpo is a 7 billion parameter language model developed by sixtyoneeighty. It is a chat finetune, meaning it has been optimized for conversational interactions. The model's training specifically leveraged the Intel/orca_dpo_pairs dataset, indicating a focus on instruction-following and preference alignment through Direct Preference Optimization (DPO).

Key Capabilities

  • Chat Finetune: Optimized for engaging in conversational dialogues.
  • DPO Training: Benefits from Direct Preference Optimization using the Intel/orca_dpo_pairs dataset, enhancing its ability to follow instructions and generate preferred responses.
  • Apache 2.0 License: Offers flexibility for use and distribution under an open-source license.

Good For

  • Conversational AI: Suitable for building chatbots, virtual assistants, and other dialogue-based applications.
  • Instruction Following: Its DPO training suggests proficiency in understanding and executing user instructions effectively.
  • Research and Development: Provides a base for further experimentation and finetuning on specific chat-oriented tasks.

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