2504.07513

Total: 1

#1 GPT Carry-On: Training Foundation Model for Customization Could Be Simple, Scalable and Affordable [PDF4] [Copy] [Kimi3] [REL]

Author: Jianqiao Wangni

Modern large language foundation models (LLM) have now entered the daily lives of millions of users. We ask a natural question whether it is possible to customize LLM for every user or every task. From system and industrial economy consideration, general continue-training or fine-tuning still require substantial computation and memory of training GPU nodes, whereas most inference nodes under deployment, possibly with lower-end GPUs, are configured to make forward pass fastest possible. We propose a framework to take full advantages of existing LLMs and systems of online service. We train an additional branch of transformer blocks on the final-layer embedding of pretrained LLMs, which is the base, then a carry-on module merge the base models to compose a customized LLM. We can mix multiple layers, or multiple LLMs specialized in different domains such as chat, coding, math, to form a new mixture of LLM that best fit a new task. As the base model don't need to update parameters, we are able to outsource most computation of the training job on inference nodes, and only train a lightweight carry-on on training nodes, where we consume less than 1GB GPU memory to train a 100M carry-on layer on 30B LLM. We tested Qwen and DeepSeek opensourced models for continue-pretraining and got faster loss convergence. We use it to improve solving math questions with extremely small computation and model size, with 1000 data samples of chain-of-thoughts, and as small as 1 MB parameters of two layer layer carry-on, and the results are promising.

Subjects: Machine Learning , Artificial Intelligence , Distributed, Parallel, and Cluster Computing , Machine Learning

Publish: 2025-04-10 07:15:40 UTC