Harness^2: Recursive Agent Harnessing for an Open World

Harness² improves agent execution harnesses through recursive task-level refinement and continual harness-level learning, boosting performance across benchmarks without ground-truth verifiers.

https://arxiv.org/abs/2610.14224

https://hjfy.top/arxiv/2610.14224

https://github.com/google-research/harness2 (22028)

FreeToken: Efficient Edge-Native MoE Serving with Bandwidth-Adaptive Execution

FreeToken is an edge-native Mixture-of-Experts serving system that dynamically maps computation and model state onto heterogeneous local hardware to run large open-weight models on personal machines.

https://arxiv.org/abs/2608.16157

https://hjfy.top/arxiv/2608.16157

https://github.com/FlashML-org/FreeToken (15847)

Differentiable Data Augmentation with Kornia

This paper reviews the Kornia differentiable data augmentation module for both 2D and 3D tensors, integrating with PyTorch's autograd and optim, and benchmarks it against other frameworks.

https://arxiv.org/abs/2011.09832

https://hjfy.top/arxiv/2011.09832

https://github.com/arraiyopensource/kornia (12116)

HairDeck: Animation-Ready Stylized Hair Cards from a Single Image

HairDeck generates stylized hair cards and spring-bone animations as unordered token sets using a diffusion transformer, improving accuracy and speed over prior methods.

https://arxiv.org/abs/2610.40599

https://hjfy.top/arxiv/2610.40599

https://github.com/hyz317/HairDeck (730)

Xiaomi-OCR-0 Technical Report

Xiaomi-OCR-0 is a compact vision-language model that unifies document parsing and OCR understanding through automated data synthesis, progressive training with text anchoring and reinforcement learning, and achieves state-of-the-art results on multiple benchmarks.

https://arxiv.org/abs/2609.36136

https://hjfy.top/arxiv/2609.36136

https://github.com/SeerRay-Lab/Xiaomi-OCR-0 (766)

Very Large-Scale Multi-Agent Simulation in AgentScope

Enhancements to the AgentScope platform improve scalability, efficiency, and ease of use for large-scale multi-agent simulations through distributed mechanisms, flexible environments, and user-friendly tools.

https://arxiv.org/abs/2407.17789

https://hjfy.top/arxiv/2407.17789

https://github.com/modelscope/agentscope (35223)

AgentScope 1.0: A Developer-Centric Framework for Building AgenticApplications

AgentScope enhances agentic applications by providing flexible tool-based interactions, unified interfaces, and advanced infrastructure based on the ReAct paradigm, supporting efficient and safe development and deployment.

https://arxiv.org/abs/2508.16279

https://hjfy.top/arxiv/2508.16279

https://github.com/agentscope-ai/agentscope (35223)

Efficient Memory Management for Large Language Model Serving withPagedAttention

PagedAttention algorithm and vLLM system enhance the throughput of large language models by efficiently managing memory and reducing waste in the key-value cache.

https://arxiv.org/abs/2309.06180

https://hjfy.top/arxiv/2309.06180

https://github.com/vllm-project/vllm (86094)

OpenDevin: An Open Platform for AI Software Developers as GeneralistAgents

OpenDevin is a platform for developing AI agents that interact with the world by writing code, using command lines, and browsing the web, with support for multiple agents and evaluation benchmarks.

https://arxiv.org/abs/2407.16741

https://hjfy.top/arxiv/2407.16741

https://github.com/opendevin/opendevin (92511)