Published Dec 17, 2024
Domino: Communication-Free LLM Training Engine // Guanhua Wang // #278
Dive into the transformative world of large language model training with Guanhua Wang as he unveils Domino, an innovative engine that eliminates communication overhead, integrates with DeepSpeed, and leverages quantization for enhanced efficiency and scalability. Explore cutting-edge memory management, real-time learning techniques, and federated learning strategies that redefine multi-node training environments.
Topics covered
Popular Clips
Questions from this episode
- Asked by 4 people
- Asked by 1 person
- Asked by 1 person
- Asked by 1 person
- Asked by 1 person
- Asked by 1 person
Episode Highlights
Related Episodes
Reliable LLM Products, Fueled by Feedback // Chinar Movsisyan // #251
Answers 383 questions
The Art and Science of Training LLMs // Bandish Shah and Davis Blalock // #219
Answers 383 questions
LLMClient a New Way to Build Complex Workflows with LLMs // Vikram Rangnekar // #259
Answers 383 questions
Building LLM Products Panel // LLMs in Production Conference Part II
Answers 383 questions
Engineering MLOps // Emmanuel Raj // MLOps Meetup #69
Answers 383 questions
PyTorch's Combined Effort in Large Model Optimization // Michael Gschwind // #274
Answers 383 questions
Making Your Company LLM-native // Francisco Ingham // #266
Answers 383 questions[EXCLUSIVE EPISODE!] LLM Key Results
Answers 383 questions
Large Language Models in Production Round-table Conversation
Answers 383 questions
