ZARZA We are Zarza, the prestigious firm behind major projects in information technology.

Critical Batch Size Revisited

Mechanical Dreams

03/06/2025 10 min Temporada 1 Episodio 58

Critical Batch Size Revisited

Listen "Critical Batch Size Revisited"

Descargar episodio

More episodes of the podcast Mechanical Dreams

Engram Paper 12/01/2026

From Entropy to Epiplexity- Rethinking Information for Computationally Bounded Intelligence 09/01/2026

Completed Hyperparameter Transfer across Modules, Width, Depth, Batch and Duration 08/01/2026

NorMuon- Making Muon more efficient and scalable 07/01/2026

Dion- Distributed Orthonormalized Updates 06/01/2026

How Learning Rate Decay Wastes Your Best Data in Curriculum-Based LLM Pretraining 05/01/2026

Latent State Models of Training Dynamics 28/10/2025

DeepSeek OCR 24/10/2025

The Coverage Principle - How Pre-training Enables Post-Training 23/10/2025

Continual Learning via Sparse Memory Finetuning 22/10/2025

Ver todos los episodios