openbmb/UltraLM-13b
TEXT GENERATIONConcurrency Cost:1Model Size:13BQuant:FP8Ctx Length:4kPublished:Jun 26, 2023Architecture:Transformer0.1K Cold
UltraLM-13b is a 13 billion parameter chat language model developed by openbmb, fine-tuned from LLaMA-13b. It is specifically trained on the UltraChat dataset for multi-turn conversational AI. This model excels at generating coherent and contextually relevant responses in chat-based interactions, making it suitable for dialogue systems and interactive applications.
Loading preview...
UltraLM-13b Overview
UltraLM-13b is a 13 billion parameter language model developed by openbmb, specifically designed for multi-turn chat applications. It is fine-tuned from the LLaMA-13b base model using the extensive UltraChat dataset, which focuses on high-quality conversational data.
Key Capabilities
- Multi-turn Chat: Optimized for engaging in extended, coherent conversations.
- Contextual Understanding: Leverages a 4096-token context window to maintain dialogue flow and relevance.
- LLaMA-based Architecture: Benefits from the robust architecture of LLaMA-13b.
Good For
- Dialogue Systems: Ideal for building chatbots, virtual assistants, and interactive agents.
- Conversational AI: Generating human-like responses in various chat scenarios.
- Research: Exploring multi-turn conversation modeling based on the UltraChat dataset and LLaMA architecture. Users should note that this model is distributed as delta weights and requires recovery from the original LLaMA-13b model.