Llama 3: Scaling open LLMs to AGI
Interconnects
00:00
Analysis of Training Models and Meta's Efficiency
The chapter explores the training of 8B, 70B, and a forthcoming 400B model at Meta, emphasizing their significance and efficiency. It evaluates Meta's training infrastructure and compares the models based on data size, inference time, and training token utilization.
Transcript
Play full episode
