-
On-Policy Transition Distillation
Paper ā¢ 2610.28048 ā¢ Published ā¢ 1 -
ModalityDance/OPTD-DeepSeek-R1-Distill-Qwen-1.5B-ThreeDomain
Text Generation ā¢ 2B ā¢ Updated ā¢ 122 -
ModalityDance/OPTD-DeepSeek-R1-Distill-Qwen-1.5B-Mid
Text Generation ā¢ 2B ā¢ Updated ā¢ 121 -
ModalityDance/OPTD-DeepSeek-R1-Distill-Qwen-1.5B-Late
Text Generation ā¢ 2B ā¢ Updated ā¢ 125
AI & ML interests
We focus on Natural Language Processing and Multimodal Learning, exploring generative AI across different modalities.
Recent Activity
View all activity
Papers
CoWorld Reasoner: Agentic Multimodal Reasoning with the Evolving Video World Model-Driven Harness
EngramEdit: Decoupled Knowledge Updates in LLMs through Conditional Memory
Checkpoints for the paper SLPO: Scaling Latent Reasoning via a Surrogate Policy (arXiv:2607.19691).
Checkpoints for the paper: "One Adapts to Any: Meta Reward Modeling for Personalized LLM Alignment".
Checkpoints and data for the paper AR-Omni: A Unified Autoregressive Model for Any-to-Any Generation
-
ModalityDance/AR-Omni-Pretrain-v0.1
Any-to-Any ā¢ Updated ā¢ 15 ā¢ 1 -
ModalityDance/AR-Omni-Chat-v0.1
Any-to-Any ā¢ Updated ā¢ 21 ā¢ 1 -
ModalityDance/AR-Omni-Instruct-v0.1
Viewer ā¢ Updated ā¢ 481k ā¢ 52 -
AR-Omni: A Unified Autoregressive Model for Any-to-Any Generation
Paper ā¢ 2601.17761 ā¢ Published ā¢ 14
checkpoints for the paper Parallel Test-Time Scaling for Latent Reasoning Models.
-
ModalityDance/latent-tts-rm
Token Classification ā¢ 0.1B ā¢ Updated ā¢ 37 -
ModalityDance/latent-tts-colar
Text Generation ā¢ 1B ā¢ Updated ā¢ 228 -
ModalityDance/latent-tts-coconut
Text Generation ā¢ 0.1B ā¢ Updated ā¢ 433 -
ModalityDance/latent-tts-codi
Text Generation ā¢ 0.1B ā¢ Updated ā¢ 264
"EngramEdit: Decoupled Knowledge Updates in LLMs through Conditional Memory"
-
EngramEdit: Decoupled Knowledge Updates in LLMs through Conditional Memory
Paper ā¢ 2610.10533 ā¢ Published ā¢ 9 -
ModalityDance/EngramEdit-LongCat-Flash-Lite-MCF-2K
Text Generation ā¢ Updated ā¢ 2 -
ModalityDance/EngramEdit-LongCat-Flash-Lite-zsRE-2K
Text Generation ā¢ Updated ā¢ 1 -
ModalityDance/EngramEdit-LongCat-Flash-Lite-MQuAKE-3K
Text Generation ā¢ Updated ā¢ 1
PhysTool-Bench is a benchmark that evaluates how well MLLMs perceive, select, and sequence PHYSICAL tools in real-world scenes.
checkpoints for the paper Reasoning in the Dark: Interleaved Vision-Text Reasoning in Latent Space
Checkpoints and data for the paper Omni-R1: Towards the Unified Generative Paradigm for Multimodal Reasoning.
-
ModalityDance/Omni-R1-Zero
Any-to-Any ā¢ 7B ā¢ Updated ā¢ 18 -
ModalityDance/Omni-R1
Any-to-Any ā¢ 7B ā¢ Updated ā¢ 37 -
ModalityDance/Omni-Bench
Viewer ā¢ Updated ā¢ 799 ā¢ 32 ā¢ 1 -
Omni-R1: Towards the Unified Generative Paradigm for Multimodal Reasoning
Paper ā¢ 2601.09536 ā¢ Published ā¢ 5
-
On-Policy Transition Distillation
Paper ā¢ 2610.28048 ā¢ Published ā¢ 1 -
ModalityDance/OPTD-DeepSeek-R1-Distill-Qwen-1.5B-ThreeDomain
Text Generation ā¢ 2B ā¢ Updated ā¢ 122 -
ModalityDance/OPTD-DeepSeek-R1-Distill-Qwen-1.5B-Mid
Text Generation ā¢ 2B ā¢ Updated ā¢ 121 -
ModalityDance/OPTD-DeepSeek-R1-Distill-Qwen-1.5B-Late
Text Generation ā¢ 2B ā¢ Updated ā¢ 125
"EngramEdit: Decoupled Knowledge Updates in LLMs through Conditional Memory"
-
EngramEdit: Decoupled Knowledge Updates in LLMs through Conditional Memory
Paper ā¢ 2610.10533 ā¢ Published ā¢ 9 -
ModalityDance/EngramEdit-LongCat-Flash-Lite-MCF-2K
Text Generation ā¢ Updated ā¢ 2 -
ModalityDance/EngramEdit-LongCat-Flash-Lite-zsRE-2K
Text Generation ā¢ Updated ā¢ 1 -
ModalityDance/EngramEdit-LongCat-Flash-Lite-MQuAKE-3K
Text Generation ā¢ Updated ā¢ 1
Checkpoints for the paper SLPO: Scaling Latent Reasoning via a Surrogate Policy (arXiv:2607.19691).
PhysTool-Bench is a benchmark that evaluates how well MLLMs perceive, select, and sequence PHYSICAL tools in real-world scenes.
Checkpoints for the paper: "One Adapts to Any: Meta Reward Modeling for Personalized LLM Alignment".
checkpoints for the paper Reasoning in the Dark: Interleaved Vision-Text Reasoning in Latent Space
Checkpoints and data for the paper AR-Omni: A Unified Autoregressive Model for Any-to-Any Generation
-
ModalityDance/AR-Omni-Pretrain-v0.1
Any-to-Any ā¢ Updated ā¢ 15 ā¢ 1 -
ModalityDance/AR-Omni-Chat-v0.1
Any-to-Any ā¢ Updated ā¢ 21 ā¢ 1 -
ModalityDance/AR-Omni-Instruct-v0.1
Viewer ā¢ Updated ā¢ 481k ā¢ 52 -
AR-Omni: A Unified Autoregressive Model for Any-to-Any Generation
Paper ā¢ 2601.17761 ā¢ Published ā¢ 14
Checkpoints and data for the paper Omni-R1: Towards the Unified Generative Paradigm for Multimodal Reasoning.
-
ModalityDance/Omni-R1-Zero
Any-to-Any ā¢ 7B ā¢ Updated ā¢ 18 -
ModalityDance/Omni-R1
Any-to-Any ā¢ 7B ā¢ Updated ā¢ 37 -
ModalityDance/Omni-Bench
Viewer ā¢ Updated ā¢ 799 ā¢ 32 ā¢ 1 -
Omni-R1: Towards the Unified Generative Paradigm for Multimodal Reasoning
Paper ā¢ 2601.09536 ā¢ Published ā¢ 5
checkpoints for the paper Parallel Test-Time Scaling for Latent Reasoning Models.
-
ModalityDance/latent-tts-rm
Token Classification ā¢ 0.1B ā¢ Updated ā¢ 37 -
ModalityDance/latent-tts-colar
Text Generation ā¢ 1B ā¢ Updated ā¢ 228 -
ModalityDance/latent-tts-coconut
Text Generation ā¢ 0.1B ā¢ Updated ā¢ 433 -
ModalityDance/latent-tts-codi
Text Generation ā¢ 0.1B ā¢ Updated ā¢ 264