AI & ML interests
None defined yet.
Recent Activity
Papers
Benchmarking Visual State Tracking in Multimodal Video Understanding
PaintBench: Deterministic Evaluation of Precise Visual Editing
Improved Baselines with Representation Autoencoders
Model weights for Solaris: Building a Multiplayer Video World Model in Minecraft
Data used during Cambrian-S's 4-stage training
Training and evaluation datasets collected for Solaris: Building a Multiplayer Video World Model in Minecraft
Collection for "Scaling Text-to-Image Diffusion Transformers with Representation Autoencoders"
-
nyu-visionx/Scale-RAE-Qwen1.5B_DiT2.4B
Text Generation • 4B • Updated • 887 -
nyu-visionx/Scale-RAE-Qwen1.5B_DiT2.4B-64ep
Text Generation • 4B • Updated • 39 -
nyu-visionx/Scale-RAE-Qwen7B_DiT9.8B
Text Generation • 17B • Updated • 14 • 1 -
nyu-visionx/Scale-RAE-Qwen7B_DiT9.8B-64ep
Text Generation • 17B • Updated • 12
VSI-SUPER benchmark proposed in Cambrian-S
Collection for Diffusion Transformers with Representation Autoencoders
-
Cambrian-1: A Fully Open, Vision-Centric Exploration of Multimodal LLMs
Paper • 2406.16860 • Published • 63 -
nyu-visionx/cambrian-8b
Text Generation • 8B • Updated • 393 • 64 -
nyu-visionx/cambrian-13b
Text Generation • 13B • Updated • 29 • 19 -
nyu-visionx/cambrian-34b
Text Generation • 35B • Updated • 20 • 27
Improved Baselines with Representation Autoencoders
Training and evaluation datasets collected for Solaris: Building a Multiplayer Video World Model in Minecraft
Model weights for Solaris: Building a Multiplayer Video World Model in Minecraft
Collection for "Scaling Text-to-Image Diffusion Transformers with Representation Autoencoders"
-
nyu-visionx/Scale-RAE-Qwen1.5B_DiT2.4B
Text Generation • 4B • Updated • 887 -
nyu-visionx/Scale-RAE-Qwen1.5B_DiT2.4B-64ep
Text Generation • 4B • Updated • 39 -
nyu-visionx/Scale-RAE-Qwen7B_DiT9.8B
Text Generation • 17B • Updated • 14 • 1 -
nyu-visionx/Scale-RAE-Qwen7B_DiT9.8B-64ep
Text Generation • 17B • Updated • 12
Data used during Cambrian-S's 4-stage training
VSI-SUPER benchmark proposed in Cambrian-S
Collection for Diffusion Transformers with Representation Autoencoders
-
Cambrian-1: A Fully Open, Vision-Centric Exploration of Multimodal LLMs
Paper • 2406.16860 • Published • 63 -
nyu-visionx/cambrian-8b
Text Generation • 8B • Updated • 393 • 64 -
nyu-visionx/cambrian-13b
Text Generation • 13B • Updated • 29 • 19 -
nyu-visionx/cambrian-34b
Text Generation • 35B • Updated • 20 • 27