AI Dataset2026
Provides 16M+ instruction–response samples and ~81 GB (7,090 compressed GitHub repos) distilled from 68 open-source sources, organized into 8 categories for SFT, coding agents and reasoning research. Model-generated content; released as a curated MIT-licensed collection.
AI Video Papers2026
Autoregressively synthesizes long-horizon, playable video worlds conditioned on current state and user actions for real-time interaction. Ships as an open-source, full-stack framework covering data preparation, model architectures, training, inference acceleration, and deployment for interactive generative worlds.
AI Video Papers2026
Pretrains a DiT-based Mixture-of-Experts video foundation model for embodied intelligence by augmenting internet videos with robot-centric footage and using a multi-dimensional reward system to prioritize physical realism and task completion while scaling MoE for better capacity vs. inference trade-offs.
AI Video2026
Generates videos from text and image+text prompts using a 30B Mixture-of-Experts model tuned for embodied intelligence; includes a refiner and structured prompt rewriter, and supports diffusers/SGLang runtimes with multi-GPU inference.
AI Dataset2026
Provides live Codex-CLI agent run traces from GPT-5.6 Sol capturing coding, debugging, security reviews, and harness/seed workflows in cumulative next-action prefixes — suitable for supervised fine-tuning and analysis of tool-using coding agents.
Enables RL post-training with million-token prompts under a fixed GPU budget by evaluating shared prompt state without autograd, retaining only minimal model state, and replaying short response branches; instantiated as GRPO and demonstrated on Qwen3.6-27B and GLM-5.2 up to multi-million token execution.
Expands a Transformer’s residual stream into many parallel streams and introduces xHC to scale Hyper-Connections beyond N=4. Combines temporal feature augmentation with sparse residual updates (update k=4 of N=16) and xHC-Flash memory optimizations to raise downstream scores while cutting effective compute and memory traffic.
AI Dataset2026
Provides intermediate pretraining checkpoints for the Aether-7B-5Attn base model to enable reproducible training-dynamics research. Includes three raw checkpoints (110k, 115k, 162k steps) packaged with model.safetensors, config, and tokenizer; uses a custom aether_v2_7way architecture requiring the aether_pkg loader.
AI Dataset2026
Provides 2,000 hours of synchronized, high‑fidelity robot‑free bimanual manipulation demonstrations with multi‑view video, calibrated end‑effector trajectories, gripper states, and language annotations. Curated from a 20,000+ hour corpus; features 6 camera views, ~3 mm pose accuracy, <40 µs cross‑sensor sync, and LeRobot v3‑style Parquet+MP4 export under CC BY 4.0.
AI Model2026
Generates and reasons about multimodal physical-world content—text, images, video and action trajectories—conditioned on text, images, video and robot/vehicle action inputs. An edge-sized (4B) Mixture‑of‑Transformers omni-model optimized for single‑GPU inference and Physical AI tasks (image→video, action prediction, robot policy).
Studies train-time knowledge injection via hypernetworks that generate fixed LoRA adapters from large fact corpora, empirically characterizing power-law scaling across hypernetwork depth, width, and target model size and reporting improved OOD generalization.
A PyTorch-native training framework for agentic reinforcement learning research that keeps researcher-facing code compact and editable. Uses an asynchronous loop to train multimodal and mixture-of-experts policies while never training on tokens the agent didn't generate; matches Megatron-style stacks under a comparable protocol and ships recipes and containers on GitHub.