Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism - Printable Version +- RISE Innovation Program (https://member.aiexosphere.com) +-- Forum: New Member Orientation (https://member.aiexosphere.com/forum-1.html) +--- Forum: Reading Recommendation (https://member.aiexosphere.com/forum-21.html) +--- Thread: Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism (/thread-6.html) |
Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism - aiexosphere - 09-08-2021 RE: Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism - csplusc - 10-26-2021 |