AI Video2025
Real-time, streaming world model for interactive long-horizon video and scene generation — provides persistent memory across interactions, streaming inference for live scenarios, and example demos for building interactive environments.
AI Train2025
Forecasts financial candlesticks (OHLCV K-lines) with a decoder-only transformer pre-trained on 12B+ records from 45 exchanges. A tokenizer turns market data into discrete tokens, enabling price/volatility forecasting and synthetic K-line generation.
AI API2025
Runs named-entity recognition, text classification, structured-JSON parsing and relation extraction from one 205M-parameter encoder in a single CPU forward pass, using schemas with per-field regex validators. A larger 1B model is available via API.
AI Image2025
Detects, segments, and tracks every instance of an open-vocabulary concept in images and video from a text phrase or visual exemplar, not just one object per prompt. An 848M-param model reaching ~75-80% of human accuracy across 270K concepts.
AI Agent2025
Unifies agentic tasks, reasoning, and coding in a single MoE model with 355B total / 32B active parameters and a switchable thinking mode. A lighter 106B-param Air variant trades scale for efficiency; both ship MIT-licensed.
AI Image2025
A 20B-parameter MMDiT diffusion model that generates and edits images with accurate embedded text, including dense Chinese and English typography. Handles complex multi-line layouts and identity-preserving edits while keeping text legible.
AI Image2025
Self-supervised vision foundation model producing dense, patch-level features that transfer to classification, segmentation, depth, and detection with a frozen backbone. Spans ViT-S (21M) to ViT-7B (6.7B params), plus ConvNeXt and satellite variants.
AI Dataset2025
A large multi-config collection of query–document pairs assembled to reproduce and extend the mGTE/LateOn data recipe for pre-training text embedding models. Data come in source-specific configs and include per-row drop/duplicate flags and guidance for using cleaned subsets for training.
AI Audio2025
Isolates any single sound from a complex audio mixture using a text description, a visual cue from a video frame, or a time span, returning both the isolated target and the residual. Released in small, base, and large sizes plus visual-prompt variants.
AI Video2025
Generates explorable, 3D-consistent virtual worlds from a single image or short video. Includes official implementations of Lyra‑1 (feed‑forward 3D/4D scene generation via video-diffusion self-distillation) and Lyra‑2 (long-horizon, explorable generative 3D worlds). Best for research and creative prototyping; requires substantial GPU compute.
Compares standard human psychometric questionnaires (PVQ, BFI) with generation‑based profiling to test whether questionnaires predict real LLM responses. Finds big divergences: questionnaires exploit lexical cues and elicit alignment‑consistent answers, mischaracterizing LLM behavior on everyday queries.
AI Others2025
Automatically removes safety alignment from transformer LLMs via directional ablation, with Optuna's TPE optimizer tuning the parameters — no retraining or model-internals expertise needed; hit 3/100 refusals at 0.16 KL on Gemma-3-12b.