What is Flash Attention v2? Flash Attention v2 is an optimized attention algorithm that improves upon the original Flash Attention to make transformer neural networks faster and more memory-efficient. It was introduced in 2023 by researchers at UC Berkeley, CMU, and other institutions. Key aspects of Flash Attention v2 include: 1. Performance improvements: * Up to
Are LLMs relevant to networking? Harnessing the Power of LLMs in Everyday Workflows Over the past year, I’ve integrated Large Language Models (LLMs) into nearly every aspect of my work and personal life. The benefits have been transformative, particularly in key areas such as Development (relying heavily on LLMs for creating or editing code)
AI/ML Creating a High-Quality Python Development Prompt for GPT Agents Learn how to craft an effective system prompt to produce high-quality Python code using GPT-4o and Claude 3.5 Sonnet LLM models.
cloud-networking Recompile gosu with latest Go version Learn how to recompile gosu with the latest Go version (1.21.4) to mitigate security concerns.
Quick K3s up and running I have discovered an excellent tool for quickly deploying a K3s solution. You can find the project page at: https://github.com/alexellis/k3sup. Take a look at the documentation for more information. Recently, I've been considering migrating my blog and other content from Cloud VM and docker-compose to my
Create and Upload SSH keys Discover how to generate, upload, and use SSH Keys for automated server logins using simple Bash commands. Enhance your SSH workflow while ensuring secure communication between untrusted hosts.
Create native macOS app from web applications Nativefier is a powerful tool that allows users to easily create native desktop applications from web applications. It is a great way to make web applications more accessible and easier to use on macOS. In this article, we will discuss how to use Nativefier to convert the web application Excalidraw
cloud-networking Unclutter environment variables How to manage project specific environment variables using direnv.
cloud-networking Joplin native Apple M1 build Joplin is a fantastic personal knowledge management (PKM) tool. Currently it has only Intel build for the Apple platform. The Intel version works on the newer apple MacBooks equipped with M1 processors with the help of emulation. However M1 processors are using ARM architecture, so an ARM version is more
Network-Development Why it is bad idea to use wildcard based imports? In Python if you are using a function that is not defined in the local namespace, you have to import it. Importing everything from a module (or file) is possible with a single line: from mymodule import * This is a bad idea however: * It could lead to conflicts between names
interesting-reads Interesting reads - W37 ipSpace.net Network Simulation Tools — Network Simulation Tools documentation [https://netsim-tools.readthedocs.io/en/latest/index.html]The ipSpace.net network simulation tools will help you be more proficient once you decide to drop GUI-based network simulators and build your labs using CLI and infrastructure-as-code principles.Time dilationYou can read
CiscoLive2021 starting soon.. > How will you be kick starting your day for #CiscoLive [https://twitter.com/hashtag/CiscoLive?src=hash&ref_src=twsrc%5Etfw]? With coffee? Some Salsa dancing? Or will you be dipping into your stash of #CLEUR [https://twitter.com/hashtag/CLEUR?src=hash&ref_src=twsrc%5Etfw] classics - stroopwafels
Network-Development Limiting WSL2 memory consumption Just a quick note about WSL2. It might consume a lot of RAM and eventually cause blue screens in Windows 10. If you experience something like this, just follow this process: 1. Shutdown the wsl instances wsl --shutdown 2. create the .wslconfig file under c:\Users\ [wsl2] memory=4GB # Limits
Network Automation How to sanitise a Cisco IOS configurations? Cisco IOS configurations are text files, including the configuration data of devices. It also includes sensitive information, such as Passwords, IP addresses, pre-shared keys and other type of keys. In different situations we need to share the configuration with others, so they can understand the solution and/or use configuration
Network Automation NetDevOps Survey To start a survey about the network automation industry to help bring clarity to operations people are automating the most/least, or even how long it is taking on average to learn network automation.
Network-Development Python for Network Engineers: Getting started Network Engineers are Developers! No more excuses.
Network Automation Network Designers in the era of NetDevOps It is not easy to define what DevOps and NetDevOps means for network designers. Designers developing solutions, and driving discussions with customers They not writing code normally, and not interacting with APIs. However network services will be more and more software driven, and this also changes how we design networks.
Network Automation Cisco Configuration Anonymizer Netconan is an interesting Github [https://github.com/Intentionet/netconan] project to remove sensitive information from Network configuration files. This is particularly useful in managed SP environment where we share configurations with the end-customer, usually in form of an LLD or SLD. Netconan can anonymize many types of sensitive information:
Network Automation On-demand GNS3-Server automation using Packet GNS3 was extensively developed in the past years. It is good that the Client and Server functions are decoupled, so we can use the GUI but still run the simulation on cloud resources. For this kind of virtualization bare metal service required as GNS3 using KVM/QEMU, and most cloud
Network-Development Network automation landscape Command Line Interfaces (CLI) will be with us for a while now, as many network devices still not exposing configuration and operational data via APIs. Using CLI is still one of the most effective way of managing a single device from an engineers point of view, but it is not
