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…
Just launched
Yesterday
- Datasetandrebadini/catalogo-teses-capes
Catálogo de Teses e Dissertações (CAPES e BDTD) English summary. This dataset contains 1,728,035 records from Brazilian theses and dissertations catalogues (1987–2026). R
- Model
- Model
- Modelvivekchahaman600/Mermaid
A Beautiful Mermaid Character Lora. Keep the Lora Strength to 0.4 or as required.
- PaperExGrad: Beyond Suppression - Explicit Clean-Gradient Subtraction in Adversarial Training
Adversarial training derives updates from surrogate losses, but source-loss descent does not directly characterize an update's effect on adversarial margins. We introduce
- Modelhinablue/[Boogu Image] Asian Mix Lora
v1.0 使用 Krea2 小資料集訓練，基於 Boogu Image Base 訓練，但也可以直接在 Turbo 上使用。但由於 Turbo 模型生成圖片會有點塑膠的感覺，所以請自行評估使用。 Trained on a small Krea2 dataset and based on the Boogu Image Base, but
- Model
- DatasetSOTAagi2030/Civic-Seed-Germination-Release
Civic Seed Germination Release This dataset publishes 2 seed batches selected from controlled germination trials dated 2026-04-08. Selection required a germination rate o
- DatasetiMVRlab/SurgVLM-SurgeryReportGeneration-Eval
Surgical Report Generation - Eval Task Private dataset. See the accompanying code package (build_surgery_report_eval.py + README, sent separately) for the full task descr
- Toolpi-openai-toolkit
OpenAI toolkit for Pi: Codex Remote Context windows, remote compaction v2, routed Web Search, image gen.
- Datasetlyca28/indonesia-poi-openStreetMaps
Indonesia POI: OpenStreetMap Detail Semua objek OpenStreetMap yang memiliki tag di wilayah Indonesia, yaitu node, way, dan relation, diekstrak dari indonesia-latest.osm.p
- Datasetlyca28/indonesia-poi-openStreetMaps-detail
Indonesia POI: OpenStreetMap Detail Semua objek OpenStreetMap yang memiliki tag di wilayah Indonesia, yaitu node, way, dan relation, diekstrak dari indonesia-latest.osm.p
- Model
- Model
- PaperGuideEvolve: Self-Evolving Agent Framework with Expert Guidance and 3Ts Mechanisms
This study introduces GuideEvolve, a self-evolving agent framework incorporating expert guidance and three autonomous regulatory mechanisms, collectively termed the 3Ts:
- Toolbrightspace-mcp-server
MCP server for Brightspace (D2L). Check grades, due dates, assignments, announcements, syllabus, rosters and more via Claude, ChatGPT, Cursor, Windsurf, or any MCP client
- Datasetliuganghuggingface/moldata-polymer_gas_permeability
Polymer gas permeability (CH4, CO2, H2, N2, O2) Gas permeability of 786 polymers for five gases, released with torch-molecule. Processed copy of polymer_gas_permeability
- Model
- Model
- Datasetliuganghuggingface/moldata-zinc250k
ZINC250k About 250,000 drug-like molecules sampled from ZINC with computed logP, QED, and SA score, released with the chemical variational autoencoder. Processed copy of
- Modelartoftomworks/Prompt Studio 1.19 — Give Your AI Creations a Memory
# Prompt Studio 1.19 — Give Your AI Creations a Memory You made a great image or video. Weeks later, can you still find its prompt, model, LoRAs, seed, and settings — and
- Datasetliuganghuggingface/moldata-admet_ai
TDC ADMET multitask table (ADMET-AI recipe) The 41 TDC ADMET benchmark endpoints merged into one table, one row per molecule, following ADMET-AI's merge recipe; released
- PaperCode Comprehensor: Hierarchical Documentation and Cross-Modal Retrieval for Large Repositories
Understanding large, complex, and rapidly evolving code bases remains a core challenge for software engineering at scale, hindering onboarding, maintenance, refactoring,
- Model
- Datasetliuganghuggingface/moldata-broad6k
Broad6K cell-based assay panel About 6,000 molecules with Broad Institute cell-based assay readouts, curated for InfoAlign. Processed copy of broad6k from the molecular d
- Datasetliuganghuggingface/moldata-chembl2k
ChEMBL2K multi-target activity panel About 2,000 molecules with activity labels on 41 protein targets, curated from ChEMBL for InfoAlign. Processed copy of chembl2k from
- Datasetobries39/ClimbInst
ClimbInst: Climbing Hold Instance Segmentation Dataset ClimbInst is an instance segmentation dataset for climbing holds and volumes on indoor climbing walls. It was devel
- Launch
- ModelFelldude/Moss Image FinalCut (5GB - 6Step)
Moss Image FinalCut FinalCut was trained to increase performance at low steps with reduced weights. 4 Step is possible however less anatomical errors appear at 6 steps CF
- PaperAgent Behavior Maps: Locating Agents by What They Would Do, Without Running Them
LLM agents with similar success rates can follow different action sequences. An Agent Behavior Map represents LLM-harness combinations using centered action log-likelihoo
- PaperClunoGS: Compact Language 3D Gaussian Splatting via Unified Optimization
Language 3D Gaussian Splatting enables open-vocabulary scene understanding by associating Gaussian primitives with language features. These semantic attributes increase m
- Dataset
- Datasetzchn/robo_test_green_marker_v0
robo_test_green_marker_v0 Scripted SO-101 follower demos: pick a green cube and place it on an ArUco marker (ids 0/2/3), with leader views V1/V2. Converted from Pi POST /
- DatasetMockingbird315/piper-drawpattern-100
Piper DrawPattern: 10 Demos and 100 Episodes This collection contains 10 dual-camera reference demonstrations. Each demonstration is paired with exactly 10 robot executio
- Model
- Model
- PaperMitigating Model Collapse through Dreaming Learning
Training generative models on synthetic data causes loss of diversity and a decrease in performance, a phenomenon known as model collapse. We study this degeneration spec
- Model
- PaperMulticlass Boundary Refinement with Neural Cellular Automata for Remote Sensing Land Cover Segmentation
Semantic segmentation of remote sensing imagery often places class boundaries a few pixels away from where they truly lie, and this position error is concentrated in the
- PaperBuilding Auditable AI Text Detectors with Sparse Autoencoders
AI-generated text detectors are increasingly deployed in consequential settings---peer review, education, admissions, journalism, and content moderation---where their out
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.