Deploying large language models (LLMs) in production often runs into a hard trade-off: reduce model size and latency through quantization,…
YOLOv9: Train-from-Scratch Object Detection That Beats Pretrained Models with Programmable Gradient Information 9391
YOLOv9 marks a significant leap forward in real-time object detection by directly confronting a long-standing but often overlooked problem in…
Mulberry: Step-by-Step Multimodal Reasoning with o1-Like Reflection for Trustworthy AI Decisions 1217
Traditional multimodal large language models (MLLMs) often produce answers without revealing how they got there—especially when dealing with complex questions…
ELF: Train Real-Time Strategy AI Bots 10x Faster with a Lightweight, Flexible RL Platform 2094
Reinforcement learning (RL) for real-time strategy (RTS) games has long been bottlenecked by slow simulation, rigid environment interfaces, and high…
AdaNet: Automate High-Quality Model Ensembling with Minimal Effort in TensorFlow 3462
In modern machine learning workflows, teams often face a tough trade-off: spend days or weeks manually tuning architectures and hyperparameters,…
GhostNet: High-Accuracy Vision Models with Minimal Compute for Edge Deployment 4355
Overview Deploying powerful computer vision models on resource-constrained devices—such as smartphones, IoT sensors, or drones—has long been a major engineering…
EfficientViT-SAM: Real-Time, High-Accuracy Image Segmentation Without Compromise 3102
If you’ve worked with Meta’s Segment Anything Model (SAM), you know its power—and its pain points. While SAM delivers state-of-the-art…
Bitnet.cpp: Run 1.58-Bit LLMs at the Edge with Lossless Speed and Efficiency 24456
Large language models (LLMs) are becoming increasingly central to real-world applications—but their computational demands remain a major barrier for edge…
SWE-Lancer: Benchmark Real-World Freelance Coding Tasks to Measure LLMs’ True Engineering Value 1438
Evaluating large language models (LLMs) on synthetic coding benchmarks often fails to reflect their real-world utility. Enter SWE-Lancer—a rigorously constructed…
Open-Sora Plan: Open-Source High-Quality Long Video Generation for Real-World Applications 12044
Open-Sora Plan is an open-source initiative designed to democratize access to state-of-the-art video generation capabilities. Inspired by the promise of…