CultriX/OmniTrixAI

TEXT GENERATIONConcurrency Cost:1Model Size:7BQuant:FP8Ctx Length:4kPublished:Jan 25, 2024License:apache-2.0Architecture:Transformer Open Weights Cold

CultriX/OmniTrixAI is a 7 billion parameter language model created by CultriX, formed by merging mlabonne/NeuralBeagle14-7B, FelixChao/WestSeverus-7B-DPO-v2, and CultriX/MergeTrix-7B-v2 using the DARE TIES merge method. This model is designed for general-purpose text generation tasks, leveraging the combined strengths of its constituent models. It offers a 4096-token context length, making it suitable for a variety of conversational and content creation applications.

Loading preview...

OmniTrixAI: A Merged 7B Language Model

OmniTrixAI is a 7 billion parameter language model developed by CultriX, created through a strategic merge of several high-performing models. This model leverages the DARE TIES merge method to combine the capabilities of:

  • mlabonne/NeuralBeagle14-7B
  • FelixChao/WestSeverus-7B-DPO-v2
  • CultriX/MergeTrix-7B-v2

The base model for this merge is senseable/WestLake-7B-v2. The merging process involved specific density and weight parameters for each contributing model, aiming to synthesize their strengths into a cohesive and capable new model. OmniTrixAI is configured with float16 precision and includes int8_mask for optimized performance.

Key Capabilities

  • General-purpose text generation: Designed to handle a wide array of conversational and content creation tasks.
  • Leverages diverse training: Benefits from the varied datasets and fine-tuning approaches of its constituent models.
  • Optimized for deployment: Configured with float16 and int8_mask for efficient inference.

Good For

  • Developers seeking a robust 7B model built from a combination of established models.
  • Applications requiring a balance of performance and resource efficiency.
  • Experimentation with merged model architectures for improved generalization.