BigHugger
GH Repository · om-ai-lab

OmAgent

[EMNLP-2024] Build multimodal language agents for fast prototype and production

stars
2,667
30-day movement
+10/day
Related entries
60
Connections
1
agent-appPythongeminiopenaillavallamagradiogptlarge-language-modelsragmultimodalagentvision-and-languagelanguage-agentvlmmultimodal-agentworkflowchatbotpythonllmsmart-hardwaregpt4

OmAgent is a Python framework from om-ai-lab for building multimodal language agents, published alongside an EMNLP-2024 paper. Its topics indicate support for workflows, RAG, vision-language models, and multiple LLM backends (OpenAI/GPT, Gemini, LLaMA, LLaVA), with a Gradio interface.

You want to prototype multimodal agents quickly and move them toward production under an Apache-2.0 license.

Use it to

  • Prototype multimodal language agents
  • Build agent workflows with RAG
  • Experiment with vision-language models
  • Build chatbots over multiple LLM backends
  • Explore smart-hardware agent scenarios

For Researchers and developers building multimodal LLM agents

Role
agent-app
Language
Python
Licence
Apache-2.0
Forks
292
Open issues
7
Last push
2025-03-19
Latest release
v0.1.0 · 2024-10-17
topicsmultimodallanguage-agentsllmragworkflowpython