Reproducibility code for Fluid Intelligence: scaling laws and cost analysis for CFD foundation models.
Generate Suno AI music from OpenCLI, download MP3 and lyrics, and manage song history and status from the terminal
Forecast-based detection of progressive neural network training degradation
A synthetic tabular and relational data generation framework
Provide on-device Apple Foundation Models inference in TypeScript with streaming, structured output, and chat-style APIs for secure, local AI processing.
🔥PhysInOne in Python (CVPR 2026)
The Silence of Intelligence — A comprehensive analysis of Anthropic CEO Dario Amodei's philosophy on Scaling Laws, AI safety, and the future of humanity. / Anthropic CEO ダリオ・アモディの思想を体系化したOSS書籍。スケーリング則の本質とAIの未来を解き明かす。
LLM Scaling Laws (2017-2026): review paper + machine-checked Lean 4 formalization (53 theorems, no sorry) | 大模型 Scaling Laws 综述与数学的机器验证
[EMNLP 2026] MDM-Prime-v2: Binary Encoding and Index Shuffling Enable Scaling of Diffusion Language Models
A controlled, cost-normalized benchmark of data-efficiency interventions, and the predictive law that organizes it: 2,800 cells, 14 datasets, 7 backbone families, 500 to 1.28M images. Companion code and artifacts for the Information Fusion submission.
Code and pretrained checkpoints for "Bigger Text Encoders Can Hurt CLIP Zero-Shot Performance": how the vision and text encoder capacity split affects CLIP zero-shot performance.
ByteBoost 2026
Independent study notes and an agent-readable index based on Geoffrey West’s Scale: scaling frameworks, chapter notes, a glossary and a cheatsheet.
Train small language models from scratch on a single RTX 5070 Ti. Scaling laws, distillation, tokenizers, evaluation.
Minimal reproduction of OneRec
Aletheion State Models: an attention-free causal state-model research family derived from DRM, evaluated through ablations and scaling laws.
Pretraining a family of LLMs on a laptop CPU: from-scratch BPE, RoPE/RMSNorm/SwiGLU, and a real fixed-compute scaling study where the 33M model loses to the 1.3M one
Package for estimating how forecasting performance scales with model size and training compute. It can be used to apply scaling-law methods to economics, finance, panel, and time-series data.
Awesome-RL-Reasoning
To associate your repository with the scaling-laws topic, visit your repo's landing page and select "manage topics."