Exploration Robotics

Solving Composite Problems in One Inference: The MeteoRA Architecture for Scalable Integration of Knowledge Modules in Large Language Models Based on MoE

by
Solving Composite Problems in One Inference: The MeteoRA Architecture for Scalable Integration of Knowledge Modules in Large Language Models Based on MoE

In the field of large language models, the pre-training + fine-tuning paradigm has become an important foundation for deploying various downstream applications. Within this framework, the use of low-rank adaptation (LoRA) methods for efficient fine-tuning of large model parameters (PEFT) has resulted in a large number of reusable LoRA adapters tailored for specific tasks. However, … Read more

Advantages of LoRaWAN Compared to LoRa Private Protocols

by
Advantages of LoRaWAN Compared to LoRa Private Protocols

This article is sourced from: Low Power Wide Area Network Review With the gradual promotion of LoRa technology, many applications have adopted LoRa technology chips, but for various reasons, they have not utilized the LoRaWAN network protocol. So, what advantages does the LoRaWAN network protocol have compared to LoRa private protocols? Many applications have adopted … Read more

Beware! LoRa Devices Suspected of Violations, Presenting Significant Risks

by
Beware! LoRa Devices Suspected of Violations, Presenting Significant Risks

The era of the Internet of Things (IoT) is accelerating, becoming a vast potential market. In the field of low-power wide-area networks, various connection technologies are fiercely competing. Currently, the LoRa connection method, based on American chip technology, faces significant limitations due to issues related to security and continuous supply. Many Chinese companies involved in … Read more

Guide to Calculating GPU Memory Requirements for LoRA and QLoRA Fine-Tuning: Understandable for Beginners

by
Guide to Calculating GPU Memory Requirements for LoRA and QLoRA Fine-Tuning: Understandable for Beginners

I have recently compiled a simple and easy-to-understand guide on the GPU memory requirements for fine-tuning with LoRA and QLoRA, which can help you estimate the memory needed when fine-tuning using LoRA and QLoRA. Below, we will explain step by step, requiring minimal background knowledge.1. What are LoRA and QLoRA? LoRA (Low-Rank Adaptation):This is a … Read more

Essential Tips for LoRA Fine-Tuning

by
Essential Tips for LoRA Fine-Tuning

As mentioned in previous articles, LoRA fine-tuning primarily targets the weight matrices of linear layers, such as the Q, K, and V projection matrices in the attention mechanism, as well as the weight matrices in the feedforward network (FFN). So, when fine-tuning a model with a Transformer architecture using LoRA, which weight matrices should we … Read more

What is the Low Power Mechanism of LoRa Devices?

by
What is the Low Power Mechanism of LoRa Devices?

Research on the Low Power Mechanism of LoRa Devices With the rapid development of Internet of Things (IoT) technology, Low Power Wide Area Network (LPWAN) technology has become one of the research hotspots in the field of wireless communication. LoRa technology, as one of the representatives of LPWAN technology, is widely used in smart cities, … Read more

Customer Service | Dedicated Debugging for LoRa Networking, Enabling Efficient Industrial IoT Implementation

by
Customer Service | Dedicated Debugging for LoRa Networking, Enabling Efficient Industrial IoT Implementation

The value of the industrial Internet of Things lies not only in thereliability of the technical solution, but also in theprofessionalism of the implementation services. When customers request equipment testing, Jixun IoT utilizes a comprehensive “Five Hearts Service” system that spans pre-sales and post-sales, completing atechnical breakthrough across thousands of miles, providing full support from … Read more

Solving Composite Problems in One Inference: The MeteoRA Architecture for Scalable Integration of Knowledge Modules in MoE-based Large Language Models

by
Solving Composite Problems in One Inference: The MeteoRA Architecture for Scalable Integration of Knowledge Modules in MoE-based Large Language Models

The AIxiv column is a section published by Machine Heart that features academic and technical content. Over the past few years, the AIxiv column has reported on more than 2000 pieces of content, covering top laboratories from major universities and companies worldwide, effectively promoting academic exchange and dissemination. If you have excellent work to share, … Read more

The Era of IoT Driven by NB-IoT/eMTC and LoRa: Everyone Can Become an ‘Operator’!

by
The Era of IoT Driven by NB-IoT/eMTC and LoRa: Everyone Can Become an 'Operator'!

Author: Zhao Xiaofei IoT Think Tank Original Reprint with source and origin indicated —— [Introduction] —— The greatest value of low-power wide-area networks lies in bringing new transformations to downstream demanders. Many industries and enterprises that have become ‘operators’ are based on the application of technologies such as NB-IoT/eMTC, LoRa, and the management of product … Read more

NB-IoT Accounts for 54%, LoRa Remains the Leader in Overseas Connections: Insights from the Latest LPWAN Research Report

by
NB-IoT Accounts for 54%, LoRa Remains the Leader in Overseas Connections: Insights from the Latest LPWAN Research Report

| Technology | Discussion | China’s leading media outlet in the telecommunications industry Author: Zhao Xiaofei Source: IoT Think Tank (iot101) Submission email for discussions: [email protected] The connection forms of the Internet of Things (IoT) are becoming increasingly diverse, with LPWAN being one of the most notable forms, catering to a large number of low-power … Read more