Turns a flow-matching image generator's self-exploration into dense, per-step supervision without a pretrained teacher; it branches the student's next-state into stochastic SDE candidates, scores them against a deterministic self-reference, and applies an advantage-weighted pull–push velocity regression with reward-level fusion for multi-objective alignment.
Develops a vision-language foundation model for autonomous driving that unifies 3D BEV perception, visual question answering, and motion planning without changing the pretrained VLM architecture. Key elements include an external BEV perception head for 3D detection and occupancy, a Planning Expert using flow-matching for trajectory prediction, and a staged training recipe combining driving and general VLM data.
AI Audio2026
Ruibin Yuan, Hanfeng Lin +56Multimodal Art Projection (MAP), Hong Kong University of Science and Technology (HKUST)
Generates full songs from lyrics and a style prompt, producing vocals and accompaniment and exporting editable symbolic scores for melody/chord control. Uses symbolic planning with agentic multi-turn editing and runs local 48 kHz inference on a 24GB GPU.
Generates long-form, text-controlled music with explicit arrangement and planning. Uses a 50 Hz single-codebook tokenizer, a flow-matching diffusion Transformer to predict VAE latents, and an MoE autoregressor with ABC‑CoT planning to produce 48 kHz audio up to 5m30s.
AI Audio2026
Encodes real audio into YuE2 semantic tokens and ships a matching NAR-branch LoRA so the YuE2 decoder renders realistic latents—enables tokenizing recordings, training artist LoRAs, and generating songs or covers. Includes an 8-layer tokenizer head and a rank-32 decoder LoRA; requires the YuE2 base models and GPU resources.
Jiahao Lu, Minghao Yin +6The Hong Kong University of Science and Technology, ARC Lab, Tencent IEG +2
Compresses a geometry-foundation model’s multi-level features into a compact latent that decodes jointly to RGB, depth, cameras and point maps — enabling a conditional flow to generate 3D-consistent video and novel views with measurably improved coherence.
AI Image2026
A distilled LoRA adapter for Qwen-Image-2.1 that runs text-to-image generation and instruction-driven image editing in a few transformer passes (shipped as a 6-step r256 LoRA). Samples with a fixed sigma schedule, no classifier-free guidance; non-commercial research license.
AI Video2026
Enables native joint video-and-audio generation and training up to 2K using a dynamic block-sparse attention scheme that cuts attention cost while preserving audio–visual coupling; provides preview checkpoints, training and inference code. High-resolution training requires large multi-GPU (80GB) clusters and CUDA ≥ 12.4.
Xingyu Miao, Zizun Li +46Shanghai AI Laboratory: Affiliation: Project page:https://internrobotics.github.io/InternW0-Delta/
Learns joint predictive visual dynamics and action generation for generalist robot manipulation, translating future-relevant visual representations into actions. Integrates a Mixture-of-Transformers coupling a video expert and action expert, a frozen vision-language model for semantics, 4D distillation, and Causal Imprint; pretrained on a 20K+ hour heterogeneous corpus.
AI Model2026
Converts Turkish text to 48 kHz speech offline using a compact 5.6M DiT acoustic model plus a 3M vocoder (8.6M parameters, ~34 MB). Streams audio with very low latency (~3.86 ms first audio on an RTX 4090), supports fast batching and runs under an Apache‑2.0 license.
AI Model2026
Generates high-resolution images directly in pixel space and can be fine-tuned, without architectural changes, for monocular depth estimation and image restoration/upscaling. Built as a 3B-parameter diffusion transformer (no VAE), with shipped depth and 4× restoration fine-tunes; default outputs ~1024×1024.
Predicts identity-preserving dense pixel correspondences between image pairs that violate spatio-temporal priors (e.g., edits and reference-guided generation). Fuses generative (FLUX2) and semantic (DINOv3) foundation representations with heterogeneous supervision and teacher-guided iterative refinement to generalize beyond classical optical-flow assumptions.