サインイン
インサイト - Large Language Model Training
Smart-Infinity: Fast Large Language Model Training with Near-Storage Processing
Smart-Infinity addresses storage bandwidth bottleneck in large language model training using near-storage processing devices.
A Learning Rate Path Switching Training Paradigm for Efficient Version Updates of Large Language Models
This paper introduces a novel training paradigm for updating large language models (LLMs) that balances pre-training performance with reduced training cost by strategically switching learning rates.
COOKBOOK: A Framework for Enhancing Large Language Model Generative Abilities Using Programmatically Generated Data Templates
Programmatically generated training data based on simple patterns over random tokens can effectively improve the generative capabilities of large language models (LLMs) on various natural language tasks.
