EE-LLM: Large-Scale Training and Inference of Early-Exit Large Language Models with 3D Parallelism
akhaliq
Abstract
We present EE-LLM, a framework for large-scale training and inference of early-exit large language models (LLMs). While recent works have shown preliminary evidence for the efficacy of early exiting in accelerating LLM inference, EE-LLM makes a foundational step towards scaling up early-exit LLMs by supporting their training and inference with massive 3D parallelism. Built upon Megatron-LM, EE-LLM implements a variety of algorithmic innovations and performance optimizations tailored to early exiting, including a lightweight method that facilitates backpropagation for the early-exit training objective with pipeline parallelism, techniques of leveraging idle resources in the original pipeline schedule for computation related to early-exit layers, and two approaches of early-exit inference that are compatible with KV caching for autoregressive generation. Our analytical and empirical study shows that EE-LLM achieves great training efficiency with negligible computational overhead compared to standard LLM training, as well as outstanding inference speedup without compromising output quality. To facilitate further research and adoption, we release EE-LLM at https://github.com/pan-x-c/EE-LLM.
Community
Github Repo is empty lol
Github Repo is empty lol
Hi @puffy310 , the code has been uploaded to GitHub (which was delayed due to a technical issue).
Glad to hear! Excited to look at it.
Is there any plans to release pretrained models?
@puffy310 It is likely, though not guaranteed, that we will release our pre-trained models later on when they are ready. We will keep you posted on any update :)
Is there any plans to release pretrained models?
@puffy310 It is likely, though not guaranteed, that we will release our pre-trained models later on when they are ready. We will keep you posted on any update :)
Hi @puffy310 We have released the pre-trained models. You may find them at https://github.com/pan-x-c/EE-LLM?tab=readme-ov-file#checkpoints
Is there any plans to release pretrained models?
@puffy310 It is likely, though not guaranteed, that we will release our pre-trained models later on when they are ready. We will keep you posted on any update :)
Hi @puffy310 We have released the pre-trained models. You may find them at https://github.com/pan-x-c/EE-LLM?tab=readme-ov-file#checkpoints
Never thought i'd see the day. Thank you for releasing it and I deeply thank you for notifying me! Have a great day.
Models citing this paper 0
No model linking this paper
Datasets citing this paper 0
No dataset linking this paper
Spaces citing this paper 0
No Space linking this paper