Introduction to dots.llm1
dots.llm1 is the first big language model open-sourced by Little Rednote, using the Mixed Expertise (MoE) architecture. The model is hosted on the Hugging Face platform and developed by the rednote-hilab team.
Core features
- parameter scale: has 142 billion parameters, but only 14 billion parameters are activated during inference, greatly reducing computational cost
- Training data: Trained using 11.2 trillion non-synthetic high-quality corpus to ensure output quality
- performance: Average score of 91.3 in Chinese tests, outperforming several mainstream open source models
- contextual support: Extremely long context processing capability of 32,768 tokens supported
- Deployment flexibility: Provides multiple deployment options, including Docker and vLLM
Applicable Scenarios
dots.llm1 is particularly well suited for tasks such as text generation, dialog systems, and content creation, while excelling in Chinese language processing.
This answer comes from the articledots.llm1: the first MoE large language model open-sourced by Little Red BookThe
May not be reproduced without permission:AI productivity tools " What is dots.llm1 and what are its main features?
English 
简体中文 
日本語 
Deutsch 
Português do Brasil