rhymes-ai/Aria
A multimodal native MoE language model with 25.3B total parameters and 64K token context window for vision-language understanding.
Not currently ranked — collecting fresh signals.
star history
Aria is an open multimodal native MoE model featuring 3.9B activated parameters per token and a 64K context window. It achieves state-of-the-art performance across language and multimodal tasks, with superior capabilities in video and document understanding. The codebase provides inference and fine-tuning capabilities through Hugging Face Transformers integration, supporting deployment on a single A100 GPU with bfloat16 precision.
Frequently asked
- What is rhymes-ai/Aria?
- A multimodal native MoE language model with 25.3B total parameters and 64K token context window for vision-language understanding.
- Is Aria open source?
- Yes — rhymes-ai/Aria is open source, released under the Apache-2.0 license.
- What language is Aria written in?
- rhymes-ai/Aria is primarily written in Jupyter Notebook.
- How popular is Aria?
- rhymes-ai/Aria has 1.1k stars on GitHub.
- Where can I find Aria?
- rhymes-ai/Aria is on GitHub at https://github.com/rhymes-ai/Aria.