🗂️ Navigation

RedPajama-INCITE

A family of open-source models trained on the RedPajama dataset.

Visit Website →

Overview

The RedPajama-INCITE family of models is a set of open-source large language models developed by Together AI and a consortium of other organizations. These models are trained on the RedPajama dataset, a 1.2 trillion token, fully open-source reproduction of the data used to train the LLaMA models. The RedPajama-INCITE family includes base models, instruction-tuned models, and chat models of various sizes.

✨ Key Features

  • Trained on the open-source RedPajama dataset
  • Includes base, instruction-tuned, and chat models
  • Available in various sizes (e.g., 3B, 7B)
  • Aims to replicate the LLaMA recipe with open data and models

🎯 Key Differentiators

  • Based on a fully open-source reproduction of the LLaMA training data
  • Provides a complete suite of base, instruction-tuned, and chat models

Unique Value: Provides a fully open-source and reproducible alternative to the LLaMA family of models.

🎯 Use Cases (4)

Natural language processing research Chatbots and conversational AI Instruction-following tasks As a base for further fine-tuning

🏆 Alternatives

OpenLLaMA LLaMA

Offers a complete ecosystem of data and models for building on the LLaMA architecture in a fully open manner.

💻 Platforms

Self-hosted

✅ Offline Mode Available

🔌 Integrations

Hugging Face Transformers

💰 Pricing

Contact for pricing
Free Tier Available

Free tier: Free to use.

Visit RedPajama-INCITE Website →