Sign inGet started

Silytics Discover

A live feed of what is new and trending in AI, pulled from public sources. Every entry links to its original source.

Latest research papers
  1. 1ChunkAct: Action Chunking for Agentic RoboticsBingxin Xu, Wenrui Bao, Xinxin Liu et al. · arXiv · 1d agocs.RO
  2. 2LoopWAM: Closed-Loop Generalist Policy Improvement from Deployment RolloutsWenxuan Guo, Yingping Liang, Zhaoguo Wang et al. · arXiv · 1d agocs.RO
  3. 3EgoExplorer: A Vision-Motion-Touch Observatory for Human Loco-Manipulation In SituYunsong Zhou, Xuekun Jiang, Haiqi Shen et al. · arXiv · 1d agocs.RO
  4. 4Dog-Breed Text-to-image Flow Matching from Pretraining to Fine-Tuning on 4 Consumer GPUsReyhaneh Esmailizadeh, Xingjian Leng, Zhanhao Liang et al. · arXiv · 1d agocs.CV
  5. 5Cepheid: Decoupling Scale from Structure for Cross-Family Motion CompletionChih-Hai Su, Tsung-Wei Ke, Yu-Lun Liu · arXiv · 1d agocs.CV
  6. 6What Matters in Quality-Aware VLA Pretraining: A Benchmark and Empirical StudyMingxuan Yan, Litian Gong, Dihong Huang et al. · arXiv · 1d agocs.RO
  7. 7Tev: Scaling Long-Horizon Agentic Environments with Dense Progress RewardsWenbo Hu, Shraman Pramanick, Krishna Kumar Singh et al. · arXiv · 1d agocs.CL
  8. 8PLAYGROUND: Progressive Layout Generation with Render-Grounded PlanningJaeho Lee, Junhwan Heo, Junghyun Park et al. · arXiv · 1d agocs.CV
  9. 9Data Free Training of Diffusion Models for Lattice Field TheoriesDennis Grinwald, Ankur Singha, Khaled Kahouli et al. · arXiv · 1d agohep-lat
  10. 10DoubtLess: Training-Free Contrastive Decoding for Efficient ReasoningMinghao Luo, Liang Chen · arXiv · 1d agocs.CL
Trending models
  1. 1Gembeddinggemma-2google · feature-extraction 2.3K 135.2K
  2. 2Jhumanizerjialinyyzz · text-generation 1.6K 65.6K
  3. 3QQwen-Image-2.1-TurboQwen · text-to-image 773 12.7K
  4. 4OMiniCPM-V-4.7-35B-A3Bopenbmb · image-text-to-text 615 4.2K
  5. 5AQwen-Image-2.1-Uncensored-Quantizedabenzerps · text-to-image 4.3K 2.3M
  6. 6AJEV-35Bautotrust · text-classification 718 28K
  7. 7CUnderdog-Saluki-27B-1.0ConwayResearch · text-generation 458 312K
  8. 8LLTX-2.5Lightricks · image-to-video 7.8K 1.8M
  9. 9AQwen-Image-2.1-Turbo-Uncensored-GGUFAtomicChat · text-to-image 451 258.3K
  10. 10QQwen3.8-27BQwen · image-text-to-text 17.7K 7.2M
Trending datasets
  1. 1Dtiktok-5.6B-videosdatasocial 298 17.2K
  2. 2Fopenai-mathFineEnvs 85 4.9K
  3. 3XMiMo-V2.6-RL-ossXiaomiMiMo 947 139.3K
  4. 4LEgoDemoLightwheelAI 208 59.5K
  5. 5LEgoProLightwheelAI 193 34.5K
  6. 6MFable-5.1-Max-Reasoning-Filtered-10000xMoreThought 332 5.6K
  7. 7Ltyped-decisionsLocalLLaMA 190 45.3K
  8. 8Sarxiv-completesecemp9 685 196.3K
  9. 9Bjabarti-llm-datasetbakrianoo 86 2.1K
  10. 10Nopus5-5-doctor-patient-conversations-all-human-diseasesnisten 318 3.8K
Popular repositories
  1. 1tensorflow/tensorflowAn Open Source Machine Learning Framework for EveryoneC++ 201K
  2. 2f/prompts.chatf.k.a. Awesome ChatGPT Prompts. Share, discover, and collect prompts from the community. Free and open source — self-host for your organization with complete privacy.HTML 173K
  3. 3huggingface/transformers🤗 Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training. Python 167.7K
  4. 4rasbt/LLMs-from-scratchImplement a ChatGPT-like LLM in PyTorch from scratch, step by stepJupyter Notebook 106.5K
  5. 5pytorch/pytorchTensors and Dynamic neural networks in Python with strong GPU accelerationPython 104.7K
  6. 6microsoft/ML-For-Beginners12 weeks, 26 lessons, 52 quizzes, classic Machine Learning for allJupyter Notebook 91.5K
  7. 7Developer-Y/cs-video-coursesList of Computer Science courses with video lectures. 83.7K
  8. 8mlabonne/llm-courseCourse to get into Large Language Models (LLMs) with roadmaps and Colab notebooks. 83.6K
  9. 9d2l-ai/d2l-zh《动手学深度学习》：面向中文读者、能运行、可讨论。中英文版被70多个国家的500多所大学用于教学。Python 81.7K
  10. 10netdata/netdataThe fastest path to AI-powered full stack observability, even for lean teams.Go 80.9K

Sources: arXiv, Hugging Face and GitHub public APIs. Refreshed hourly. Silytics only links to and summarizes metadata; rights remain with the original publishers.