Ming Flash Omni 2.0
Ming-flash-omni 2.0 is an open sparse MoE omni-modal model that unifies text, image, video and audio understanding and generation, using a Ling-2.0 Mixture-of-Experts backbone with 100B parameters and about 6B active per token.
- Type
- Multimodal
- Weights
- Closed
- Downloads
- 4.2k
- Likes
- 274
- Pipeline
- any to any
- License
- mit
Model hub
inclusionAI/Ming-flash-omni-2.0
- Repo
- inclusionAI/Ming-flash-omni-2.0
- Library
- diffusers
- Access
- Public
Technical tags and languages
Papers
2
More from InclusionAI
11
-
16k DL
Ling 3.0 Flash
Text
Ling-3.0-flash is a 124B-parameter Mixture-of-Experts language model with about 5.1B parameters activated per…
-
Ling 3.0 Flash Sante
Text
A health and medicine focused mixture of experts language model with 124B total parameters and 5.1B active pe…
-
23k DL
Ling 3.0 Tiny
Text
Ling 3.0 Tiny is InclusionAI’s compact Mixture-of-Experts language model for responsive agents, instruction f…
-
LLaDA2 X
Text
LLaDA2.X is InclusionAI’s diffusion language model family, scaling to 100B parameters and using parallel deco…
-
219k DL
LLaDA2.0-mini
Text
text generation · License apache-2.0 · 219,109 downloads · by inclusionAI
-
119k DL
LLaDA2.1-mini
Text
text generation · License apache-2.0 · 119,146 downloads · by inclusionAI
-
19k DL
Ming Omni
Multimodal
Ming-Omni is an end-to-end multimodal assistant that reads text, images, audio, and video, then replies with…
-
3.9k DL
Ring 2.5 1T
Text
Trillion-parameter “thinking” model described as using a hybrid linear attention architecture for long-horizo…
About InclusionAI
1
Inclusion‑AI is a UK-based nonprofit organisation that researches and promotes inclusive and equitable AI systems, focussing on how machine learning tools can better serve under-represented communities and reduce bias.