Tagged articles

RL training

13 articles · Page 1 of 1

Machine Learning Algorithms & Natural Language Processing

Aug 25, 2026 · Artificial Intelligence

Understanding Multi‑Round Rollouts, Context Reconstruction, and RL Training in Agentic RL

The article analyzes how Agentic RL decouples internal state, protocol requests, and token sequences, explains the inference pipeline, the challenges of preserving prefix relationships across multi‑round rollouts, and details a gateway‑based data collection and credit‑assignment pipeline for reinforcement‑learning training.

Agentic RLPiRL training

0 likes · 13 min read

Understanding Multi‑Round Rollouts, Context Reconstruction, and RL Training in Agentic RL

DataFunSummit

Jul 25, 2026 · Cloud Native

Evolution of Agent Infrastructure: Engineering Insights from Tencent Cloud Agent Runtime

The article analyzes how agents transition from demo to production, revealing that beyond model capabilities, stability, elasticity, security, and governance become critical, and explains the engineering challenges and solutions—including session management, state persistence, scheduling mismatches, sandbox isolation, and open‑source strategies—that underpin Tencent Cloud's Agent Runtime.

Agent RuntimeCloud NativeKubernetes

0 likes · 26 min read

Evolution of Agent Infrastructure: Engineering Insights from Tencent Cloud Agent Runtime

SuanNi

Jun 6, 2026 · Artificial Intelligence

Demystifying Harness, Scaffold, and Other Tricky AI Agent Terms

This article breaks down the core terminology of AI agents—Model, Scaffold, Harness, Context Engineering, Policy, Tool Use, Skills, Sub‑agents, and the training‑side concepts of RL Environment, Trainer, Rollout, and Reward—explaining their roles, differences, and how they combine to form functional agents.

AI agentHarnessRL training

0 likes · 12 min read

Demystifying Harness, Scaffold, and Other Tricky AI Agent Terms

SuanNi

May 16, 2026 · Artificial Intelligence

Can a 4B Small Model Replace Top‑Tier Closed‑Source LLMs? Microsoft’s Terminus‑4B Cuts Token Use by 30%

Microsoft’s research shows that a 4‑billion‑parameter small model, Terminus‑4B, can act as an execution sub‑agent for terminal tasks, trimming token consumption by about 30% while preserving performance on demanding SWE‑Bench benchmarks, demonstrating a practical alternative to costly large models.

AI programmingRL trainingSWE-bench

0 likes · 7 min read

Can a 4B Small Model Replace Top‑Tier Closed‑Source LLMs? Microsoft’s Terminus‑4B Cuts Token Use by 30%

DataFunSummit

Apr 15, 2026 · Artificial Intelligence

How Relax Powers Scalable Multi‑Modal RL Training with Full Asynchrony

Relax, an open‑source RL training engine built on Megatron‑LM and SGLang, tackles data heterogeneity, system fragility, and role coupling by using a service‑oriented fault‑tolerant architecture, asynchronous pipelines, and multimodal‑native support, achieving up to 76% end‑to‑end speedup over veRL.

AI InfrastructureDistributed SystemsRL training

0 likes · 11 min read

How Relax Powers Scalable Multi‑Modal RL Training with Full Asynchrony

CodeTrend

Apr 11, 2026 · Artificial Intelligence

Inside Hermes Agent: How Its Closed‑Loop Learning Architecture Transforms AI Assistants

Hermes Agent introduces a closed‑loop learning architecture that adds result evaluation, pattern extraction, and persistent user modeling to the traditional receive‑plan‑execute‑return cycle, offering searchable FTS5‑based memory, autonomous skill creation, multi‑platform messaging, provider‑agnostic model switching, and built‑in research tools for AI developers.

FTS5Hermes AgentLLM summarization

0 likes · 8 min read

Inside Hermes Agent: How Its Closed‑Loop Learning Architecture Transforms AI Assistants

PaperAgent

Mar 5, 2026 · Artificial Intelligence

Bridging Agent Runtime and RL: Inside the Claw‑R1 Training Framework

Claw‑R1, a new reinforcement‑learning framework from the USTC Cognitive Intelligence Lab, integrates the OpenClaw Agent Runtime with RL training to enable agents to learn directly in real environments, addressing the gap between simulated tasks and true tool‑calling, multi‑step reasoning, and stable long‑task execution.

AI InfrastructureAgent RuntimeClaw-R1

0 likes · 10 min read

Bridging Agent Runtime and RL: Inside the Claw‑R1 Training Framework

AI Engineering

Jan 7, 2026 · Artificial Intelligence

Unsloth-MLX: Fine‑Tune LLMs on Mac and Seamlessly Move Code to Cloud GPUs

Unsloth‑MLX leverages Apple’s MLX framework to let Mac users with Apple Silicon fine‑tune large language models locally with a single import change, offering zero‑cost migration to cloud GPUs, supporting SFT, DPO, ORPO, GRPO training, and export to HuggingFace or GGUF formats.

Apple SiliconGPU cloudLLM fine-tuning

0 likes · 4 min read

Unsloth-MLX: Fine‑Tune LLMs on Mac and Seamlessly Move Code to Cloud GPUs

Kuaishou Tech

Dec 19, 2025 · Artificial Intelligence

Why Sampling Noise, Not Train‑Inference Gap, Drives RL Instability in MOE Models

The article reveals that sampling noise, rather than train‑inference inconsistency, is the primary cause of reward collapse during RL training of MOE models, and demonstrates that suppressing this noise stabilizes training and speeds convergence.

AI codingMoE modelsModel Stability

0 likes · 6 min read

Why Sampling Noise, Not Train‑Inference Gap, Drives RL Instability in MOE Models

Baobao Algorithm Notes

Sep 23, 2025 · Artificial Intelligence

How LongCat-Flash-Thinking Sets New SOTA in Open‑Source AI Inference

LongCat-Flash-Thinking, the latest open‑source model from Meituan, introduces domain‑parallel RL training, a high‑throughput DORA infra, and a dual‑path inference framework that together achieve state‑of‑the‑art performance on logical, mathematical, coding, and agentic tasks while maintaining top‑tier speed.

LongCatRL trainingTool Use

0 likes · 10 min read

How LongCat-Flash-Thinking Sets New SOTA in Open‑Source AI Inference

AntTech

Aug 6, 2025 · Artificial Intelligence

Ring-lite-2507: Boosted Deep Reasoning and Balanced General Capabilities

The AntBailing team releases Ring-lite-2507, enhancing deep reasoning through a Two‑staged RL pipeline while simultaneously balancing overall model abilities, showcasing notable gains on benchmarks like ARC‑AGI‑v1 and offering the model as an open‑source resource across major platforms.

Large Language ModelRL trainingRing-lite

0 likes · 5 min read

Ring-lite-2507: Boosted Deep Reasoning and Balanced General Capabilities

Baobao Algorithm Notes

Jun 3, 2025 · Artificial Intelligence

How to Train a 671B‑Scale Model with RL: Insights from a verl Internship

This article shares a detailed, first‑hand analysis of the technical challenges, framework choices, memory management, weight conversion, precision alignment, and efficiency optimizations encountered while building reinforcement‑learning pipelines for a 671‑billion‑parameter model using the verl ecosystem.

GPU Memory ManagementLarge ModelsMegatron

0 likes · 16 min read

How to Train a 671B‑Scale Model with RL: Insights from a verl Internship

AI Algorithm Path

Apr 20, 2025 · Artificial Intelligence

Boosting Visual Reasoning in VLMs with Reinforcement Learning

The article analyzes how reinforcement learning, which transformed LLM reasoning in DeepSeek, can be applied to visual‑language models to overcome the limitations of traditional chain‑of‑thought prompting and supervised fine‑tuning, presenting concrete reward designs, training pipelines, and a critical assessment of their strengths and weaknesses.

LLMRL trainingchain-of-thought

0 likes · 10 min read

Boosting Visual Reasoning in VLMs with Reinforcement Learning