Tempo: Small Vision-Language Models are Smart Compressors for Long Video Understanding

🦞 Just talk to your agent — it learns and EVOLVES 🧬.

Learning to See by Looking at Noise

Pytorch implementation of Transfusion, "Predict the Next Token and Diffuse Images with One Multi-Modal Model", from MetaAI

Procedural Image Programs for Representation Learning - NeurIPS 2022

[CVPR 2026] Mesh4D: 4D Mesh Reconstruction and Tracking from Monocular Video

A data generation pipeline for creating semi-realistic synthetic multi-object videos with rich annotations such as instance segmentation masks, depth maps, and optical flow.

A generative world for general-purpose robotics & embodied AI learning.

A Diagnostic Dataset for Compositional Language and Elementary Visual Reasoning

[CVPR 2026] 👋 Dataset and Benchmark code for EgoEdit

This repository contains the official code and data for CogIP-Bench (Cognition Image Property Benchmark) and the associated alignment methods described in the paper "From Pixels to Feelings: Aligni…

Implementation of Reinforcement Pre-Training (RPT) for Language Models - ArXiv:2506.08007

State-of-the-art Image & Video CLIP, Multimodal Large Language Models, and More!

Fully Open Framework for Democratized Multimodal Training

Official PyTorch Implementation of "Diffusion Transformers with Representation Autoencoders"

NeurIPS 2025 Spotlight; ICLR2024 Spotlight; CVPR 2024; EMNLP 2024

Qwen3-VL is the multimodal large language model series developed by Qwen team, Alibaba Cloud.

Baking Gaussian Splatting into Diffusion Denoiser for Fast and Scalable Single-stage Image-to-3D Generation and Reconstruction (ICCV 2025)

Qwen3 is the large language model series developed by Qwen team, Alibaba Cloud.

verl/HybridFlow: A Flexible and Efficient RL Post-Training Framework

[ICLR 2026] Official PyTorch Implementation of RLP: Reinforcement as a Pretraining Objective

[CVPR 2025] Fast3R: Towards 3D Reconstruction of 1000+ Images in One Forward Pass

MapAnything: Universal Feed-Forward Metric 3D Reconstruction

(Accepted by IJCV) Liquid: Language Models are Scalable and Unified Multi-modal Generators

🔥🔥🔥 Latest Papers, Codes and Datasets on Video-LMM Post-Training

Code for Words That Make Language Models Perceive

[CVPR 2025] 🔥 Official impl. of "TokenFlow: Unified Image Tokenizer for Multimodal Understanding and Generation".