Tagged articles

Gradient Clipping

3 articles · Page 1 of 1

Jun 18, 2026 · Artificial Intelligence

Can DP‑SGD’s Toughest Clip Threshold Auto‑Adjust? Inside the SlaClip Method

The article presents SlaClip, an adaptive gradient‑clipping technique for differential‑privacy SGD that leverages the slack between gradient norms and the clipping threshold as a privacy‑preserving indicator, eliminating extra privacy queries and dynamically adjusting the clipping bound, with experiments showing competitive accuracy across datasets and budgets.

Adaptive ClippingDP-SGDDifferential Privacy

0 likes · 8 min read

Can DP‑SGD’s Toughest Clip Threshold Auto‑Adjust? Inside the SlaClip Method

Kuaishou Large Model

Aug 19, 2025 · Artificial Intelligence

How Klear-Reasoner Achieves SOTA Math & Code Reasoning with GPPO

Klear-Reasoner, built on Qwen3‑8B‑Base, introduces the Gradient‑Preserving Clipping Policy Optimization (GPPO) algorithm to overcome traditional clip limitations, achieving state‑of‑the‑art performance on AIME2024/2025 and LiveCodeBench while providing detailed experimental analysis and data‑quality insights.

GPPOGradient ClippingLarge Language Models

0 likes · 11 min read

How Klear-Reasoner Achieves SOTA Math & Code Reasoning with GPPO

DataFunSummit

Feb 14, 2023 · Artificial Intelligence

Deep Learning Hyperparameter Tuning and Training Tips: Insights from Zhihu Experts

This article compiles practical deep learning training and hyperparameter tuning advice from Zhihu contributors, covering model debugging, learning‑rate strategies, optimizer choices, data preprocessing, regularization techniques, initialization methods, common pitfalls, recommended research papers, and ensemble approaches.

Gradient ClippingRegularizationdeep learning

0 likes · 13 min read

Deep Learning Hyperparameter Tuning and Training Tips: Insights from Zhihu Experts