New
Everything new in AI as it appears: launches announced by the labs, brand-new models, datasets, papers and open-source tools. Checked every minute or two. Launches feed · Trending · News
Live now
- Connecting…
Saturday 24 October
- DatasetETHRC/v1-demo
This dataset was created using LeRobot. Dataset Structure meta/info.json: { "codebase_version": "v3.0", "fps": 50, "features": { "observation.state": { "dtype": "float32"
- Datasetkangjiezheng/MUSS-data
MUSS data Data and reference files for MUSS: A Multi-scale and Sequence-based Model for Single-cell Gene Regulation (NeurIPS 2026). Code and reproduction guide · Model ch
- Dataset
- DatasetThoandanh/pico-scribble-300-v1
Pico Scribble 300 (v1) Bộ dữ liệu Pico Scribble 300 là phiên bản nghiên cứu được thiết kế để huấn luyện và đánh giá các mô hình chỉnh sửa ảnh tương tác có điều kiện (Scri
- Dataset
- Dataset
- Datasetrichardbaxterseo/yubhub-live-jobs
YubHub Live Jobs Dataset A snapshot of 34,770 enriched job postings from 1,190 employers, read directly from employers' own careers pages by YubHub and enriched with sala
- DatasetRafaelJaime/smartseed-grasp-merged
This dataset was created using LeRobot. Dataset Structure meta/info.json: { "codebase_version": "v3.0", "fps": 30, "features": { "action": { "dtype": "float32", "names":
- Datasetgonzalobenegas/biotasks-notebook-pilot-issue-19-pymc
Hierarchical rat-tumor inference: autonomous notebook-derived task One computational reproduction task covering both terminal analyses in the pinned PyMC notebook: a norm
- Dataset
- DatasetTheFinAI/FinDendro
FinDendro: Temporal Self-Knowledge of LLMs on Dated Financial Facts FinDendro measures two dates for a language model: where its knowledge of dated financial facts actual
- Datasetjuiceb0xc0de/qwen3.5-2b-cybersec-cgs-atlas
juiceb0xc0de/qwen3.5-2b-cybersec-cgs-atlas A brain atlas for reaperdoesntknow/Qwen3.5-2B-CyberSec, a 2.3B-parameter Qwen3_5ForConditionalGeneration with 24 layers, fine-t
- Datasetsw-voice/swamiji-tts-eval-samples
Swamiji TTS: old vs new (English) old_audio = current prod voice. new_audio = new Stage 2 model (loudness-matched). transcript = the text both were given.
- Dataset
- Dataset
- Dataset
- Dataset
- DatasetethanCSL/eval_openarm_pringles_lerobot_sim_10_generated_500_smolvla_v00
This dataset was created using LeRobot. Dataset Structure meta/info.json: { "codebase_version": "v3.0", "fps": 30, "features": { "action": { "dtype": "float32", "shape":
- Dataset
- DatasetSqls1/dota2-particles
Dota 2 Particle Effects Every particle system of Dota 2 (abilities, cosmetics, units, buildings, maps, the interface, events) as JSON, with the textures they draw and the
- Datasethirundo-io/tool-injection-answered-forget-set
Tool-output prompt-injection forget set (answered targets) A forget set for hardening openai/gpt-oss-20b against indirect prompt injection, meaning attacker instructions
- Dataset
- DatasetLunita02/Cubos_150
This dataset was created using LeRobot. Dataset Structure meta/info.json: { "codebase_version": "v3.0", "fps": 30, "features": { "action": { "dtype": "float32", "names":
- Dataset
- Dataset
- DatasetLunita02/hackathon_dataset_20261024_160132_sin42
This dataset was created using LeRobot. Dataset Structure meta/info.json: { "codebase_version": "v3.0", "fps": 30, "features": { "action": { "dtype": "float32", "names":
- Datasettoners-amatory-9h/Annoy-PyEdu-Rs-Raw
Annoy: This should be a paper Title 📑 Paper | 🌐 Project Page | 💾 Released Resources | 📦
- Datasettoners-amatory-9h/Annoy-PyEdu-Rs
Annoy: This should be a paper Title 📑 Paper | 🌐 Project Page | 💾 Released Resources | 📦
- DatasetLunita02/hackathon_dataset_20261024_172813
This dataset was created using LeRobot. Dataset Structure meta/info.json: { "codebase_version": "v3.0", "fps": 30, "features": { "action": { "dtype": "float32", "names":
- Dataset
- DatasetMikelRubi/up-down_20261024_172048
This dataset was created using LeRobot. Dataset Structure meta/info.json: { "codebase_version": "v3.0", "fps": 30, "features": { "action": { "dtype": "float32", "names":
- Datasetfesefsde12/CQU-BPDD
The Chongqing University Bituminous Pavement Disease Detection Dataset (CQU-BPDD) The CQU-BPDD consists of 60,056 bituminous pavement images, which were automatically cap
- DatasetPureOne/Qeira-II-Carry-Holonomy-Rank-Law
Qeïra II — The Carry–Holonomy Rank Law An exact multivariate first-carry rank theorem and a correction to an odd-prime matrix-digit conjecture. Research author: Artificia
- DatasetMikelRubi/up-down-bad_20261024_172018
This dataset was created using LeRobot. Dataset Structure meta/info.json: { "codebase_version": "v3.0", "fps": 30, "features": { "action": { "dtype": "float32", "names":
- DatasetAbhay07042005/business-risk-prediction-dataset
Yelp Business Economic Indicators Dataset ⚠️ Important Version Notice (Please Read) Version 3 (v3) should be used instead of v1 or v2. v1 (~1k rows) and v2 (~10k rows) us
- DatasetMikelRubi/so101-pick-cube_20261024_171946
This dataset was created using LeRobot. Dataset Structure meta/info.json: { "codebase_version": "v3.0", "fps": 30, "features": { "action": { "dtype": "float32", "names":
- Datasetdaddy696969696969/ASVspoof2021_LA
ASVspoof 2021 LA Benchmark-ready packaging of the Logical Access (LA) evaluation partition from ASVspoof 2021 for speech anti-spoofing and synthetic / deepfake voice dete
- Datasetvdanilov/video-bokeh-library
Video Bokeh Library Cut-out objects and background photos, each with a disparity map, for composing training video for depth-aware bokeh. The library holds 8,586 foregrou
- Datasetvdanilov/video-bokeh
Video Bokeh Synthetic video for training and evaluating depth-aware bokeh. Every frame is composited from known layers, so its disparity, the soft matte of each object, i
- Datasetjvzhu/algorithmic-governance-comparative-education
Algorithmic Governance and Comparative Education in Times of Polycrisis Digital Monitoring, Institutional Compliance, and Educational Systems Companion materials for a bo
Headlines and links come from the labs' own blogs, Hacker News and the GDELT news index, and from the sources our index already follows. We show a short excerpt and always link to the original. A launch reported by a news site rather than the lab is marked “Reported” until the lab itself announces it.