Training modern LLMs is extremely resource intensive, and customizing them
for various deployment scenarios characterized by limited compute and memory
resources through repeated training is impractical. In this paper, we introduce
Flextron, a network architecture and post-training model optimization framework
supporting flexible model deployment. The Flextron architecture utilizes a
nested elastic structure to rapidly adapt to specific user-defined latency and
accuracy targets during inference with no additional fine-tuning required. It
is also input-adaptive, and can automatically route tokens through its
sub-networks for improved performance and efficiency. We present a
sample-efficient training method and associated routing algorithms for
systematically transforming an existing trained LLM into a Flextron model. We
evaluate Flextron on the GPT-3 and LLama-2 family of LLMs, and demonstrate
superior performance over multiple end-to-end trained variants and other
state-of-the-art elastic networks, all with a single pretraining run that
consumes a mere 7.63% tokens compared to original pretraining.

Flextron 是一种网络架构和后训练模型优化框架，用于灵活的模型部署，并通过样本高效的训练方法和路由算法将现有的 LLM 转换为 Flextron 模型，实现卓越的性能，远优于其他弹性网络和多个端到端训练变体，并且仅需要进行一轮预训练消耗原始预训练的 7.63% 令牌。

Flextron: 多合一弹性大语言模型

Flextron: Many-in-One Flexible Large Language Model

The application of routing algorithms to real-world situations is a widely
studied research topic. Despite this, routing algorithms and applications are
usually developed for a general purpose, meaning that certain groups, such as
ageing people, are often marginalized due to the broad approach of the designed
algorithms. This situation may pose a problem in cities which are suffering a
slow but progressive ageing of their populations. With this motivation in mind,
this paper focuses on describing our implemented Age-Friendly Route Planner,
whose goal is to improve the experience in the city for senior citizens. In
order to measure the age-friendliness of a route, several variables have been
deemed, such as the number of amenities along the route, the amount of
comfortable elements found, or the avoidance of sloppy sections. In this paper,
we describe one of the main features of the Age-Friendly Route Planner: the
preference-based routes, and we also demonstrate how it can contribute to the
creation of adapted friendly routes.

这篇论文描述了我们所实现的老年人友好线路规划器的一个主要特性：基于偏好的路线，并展示了它如何促进适配友好的线路的创建。

面向老年公民的舒适路线规划器

Age-Friendly Route Planner: Calculating Comfortable Routes for Senior  Citizens

Global routing has been a historically challenging problem in electronic
circuit design, where the challenge is to connect a large and arbitrary number
of circuit components with wires without violating the design rules for the
printed circuit boards or integrated circuits. Similar routing problems also
exist in the design of complex hydraulic systems, pipe systems and logistic
networks. Existing solutions typically consist of greedy algorithms and
hard-coded heuristics. As such, existing approaches suffer from a lack of model
flexibility and non-optimum solutions. As an alternative approach, this work
presents a deep reinforcement learning method for solving the global routing
problem in a simulated environment. At the heart of the proposed method is deep
reinforcement learning that enables an agent to produce an optimal policy for
routing based on the variety of problems it is presented with leveraging the
conjoint optimization mechanism of deep reinforcement learning. Conjoint
optimization mechanism is explained and demonstrated in details; the best
network structure and the parameters of the learned model are explored. Based
on the fine-tuned model, routing solutions and rewards are presented and
analyzed. The results indicate that the approach can outperform the benchmark
method of a sequential A* method, suggesting a promising potential for deep
reinforcement learning for global routing and other routing or path planning
problems in general. Another major contribution of this work is the development
of a global routing problem sets generator with the ability to generate
parameterized global routing problem sets with different size and constraints,
enabling evaluation of different routing algorithms and the generation of
training datasets for future data-driven routing approaches.

本文提出了一种基于深度强化学习的全局路由方法，该方法可以在模拟环境中生成优化的路由策略，与传统的基于贪心算法的方法相比，该方法具有更高的灵活性和更优的效果，为未来数据驱动的路由方法的培训数据集和生成不同规模和约束条件的参数化全局路由问题集合提供了可能性。