kodonho/Solar-OrcaDPO-Solar-Instruct-SLERP

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:10.7BQuant:FP8Ctx Length:4kPublished:Jan 12, 2024License:cc-by-nc-4.0Architecture:Transformer Open Weights Warm

The kodonho/Solar-OrcaDPO-Solar-Instruct-SLERP is a 10.7 billion parameter English instruction-tuned language model. It is a merged model based on upstage/SOLAR-10.7B-Instruct-v1.0 and bhavinjawade/SOLAR-10B-OrcaDPO-Jawade, utilizing a gradient slerp merging technique. This model achieves an average score of 74.3, making it suitable for general-purpose conversational AI and instruction-following tasks.

Loading preview...

Model Overview

The kodonho/Solar-OrcaDPO-Solar-Instruct-SLERP is a 10.7 billion parameter English instruction-tuned language model. It is a composite model created by merging two existing models:

  • [upstage/SOLAR-10.7B-Instruct-v1.0]
  • [bhavinjawade/SOLAR-10B-OrcaDPO-Jawade]

This merge was performed using a gradient slerp technique, aiming to combine the strengths of its base models. The model is designed for general instruction-following and conversational tasks in English.

Key Characteristics

  • Parameter Count: 10.7 billion parameters.
  • Base Models: Built upon the SOLAR-10.7B-Instruct-v1.0 and SOLAR-10B-OrcaDPO-Jawade architectures.
  • Merging Method: Utilizes a gradient slerp approach for model combination.
  • Performance: Achieves an average score of 74.3, indicating solid performance across various benchmarks.
  • Context Length: Supports a context length of 4096 tokens.

Use Cases

This model is well-suited for applications requiring a capable instruction-following LLM, such as:

  • General-purpose chatbots and conversational agents.
  • Text generation based on specific instructions.
  • Question answering and information retrieval.
  • Tasks benefiting from a merged model's combined knowledge and capabilities.