"Parameter Golf" competition attracts over 2,000 submissions on AI optimization
The Parameter Golf event successfully concluded with thousands of creative ideas on AI model optimization, including quantization, TTT LoRA, and SSMs.
Sources x.com
AutoTTS: Automating Inference Strategies, Cutting LLM Token Costs by 69.5%
The new AutoTTS framework enables large language models to automatically search for optimal inference strategies, cutting token consumption by up to 69.5% while enhancing problem-solving performance.
Sources venturebeat.com
llama.cpp b9235: Accelerating Inference with Speculative N-gram Tuning
The llama.cpp b9235 release introduces Speculative N-gram Tuning, significantly optimizing decode speeds when running large models like Qwen3.6 27B.
Sources x.com
New PIBO algorithm optimizes offshore wind farm layouts
The Permutation-Invariant Bayesian Optimization (PIBO) algorithm improves wind turbine placement and cuts computation time in half using Optimal Transport theory.
Sources arxiv.org
UniScale: Jointly Optimizing Model Routing and Test-Time Scaling
UniScale is an online framework that unifies model routing and test-time scaling into a single optimization space, achieving a better balance between quality and cost.
Sources arxiv.org
ECC — The Optimization Toolkit for Claude Code
ECC provides a collection of skills, commands, and hooks that help optimize token usage, enhance security, and boost productivity when using Claude Code.
Sources github.com
Understanding RAG: The LLM Optimization Solution from NVIDIA 🧠
RAG (Retrieval-Augmented Generation) improves the accuracy of large language models by enabling direct retrieval from trusted external data sources.
Sources blogs.nvidia.com
Transformer Reparameterizations Lab announces series of new reparameterization techniques 🛠️
Transformer Reparameterizations Lab has released new reparameterization techniques to optimize training and inference performance for the Transformer architecture.
Sources x.com
Optimizing CUDA Graph for Grouped GEMM with CLC Work Stealing
A new technique leveraging the CLC work-stealing mechanism enables CUDA Graph compatibility for grouped_gemm implementations, optimizing computational performance for complex AI models.
Sources x.com