Knowledge Base Resources
Use these resources “vetted” by the community. Additional Knowledge Base Resources are always welcome.
Fine-tuning LLMs with PEFT and LoRA
0
As LLMs get larger fine-tuning to the full extent can become difficult to train on consumer hardware. Storing and deploying these tuned models can also be quite expensive and difficult to store. With PEFT (parameter -efficent fine tuning), it approaches fine-tune on a smaller scale of model parameters while freezing most parameters of the pretrained LLMs. Basically it is providing full performance that which is similar if not better than full fine tuning while only having a small number of trainable parameters. This source explains that as well as going over LORA diagrams and a code walk through.
Recommended Libraries for Cyberinfrastructure Users Developing Jupyter Notebooks
0
This repository contains information about Jupyter Widgets and how they can be used to develop interactive workflows, data dashboards, and web applications that can be run on HPC systems and science gateways. Easy to build web applications are not only useful for scientists. They can also be used by software engineers and system admins who want to quickly create tools tools for file management and more!
Fundamentals of Cloud Computing
0
An introduction to Cloud Computing
ACCESS Getting Started Quick-Guide
0
A step-by-step guide to getting your first allocation for Access computing and storage resources.
DELTA Introductory Video
0
Introductory video about DELTA. Speaker Tim Boerner, Senior Assistant Director, NCSA
Better Scientific Software (BSSw)
0
The Better Scientific Software (BSSw) project provides a community to collaborate and learn about best practices in scientific software development. Software—the foundation of discovery in computational science & engineering—faces increasing complexity in computational models and computer architectures. BSSw provides a central hub for the community to address pressing challenges in software productivity, quality, and sustainability.
Data visualization with Matplotlib
0
Data visualization is a critical aspect of data analysis. It allows for a clear and concise representation of data, making it easier for users to understand and interpret complex datasets. One of the most popular libraries for data visualization in Python is Matplotlib. The included website aims to provide a brief overview of Matplotlib, its features, and examples/exercises to dive deeper into its functionalities.
Introduction to Probabilistic Graphical Models
0
This website summarizes the notes of Stanford's introductory course on probabilistic graphical models.
It starts from the very basics and concludes by explaining from first principles the variational auto-encoder, an important probabilistic model that is also one of the most influential recent results in deep learning.
ACES: Charliecloud Containers for Scientific Workflows (Tutorial)
0
This tutorial introduces the use of Containers using the Charliecloud software suite. This tutorial will provide participants with background and hands-on experience to use basic Charliecloud containers for HPC applications. We discuss what containers are, why they matter for HPC, and how they work. We'll give an overview of Charliecloud, the unprivileged container solution from Los Alamos National Laboratory's HPC Division. Students will learn how to build toy containers and containerize real HPC applications, and then run them on a cluster. Exercises are demonstrated using the ACES cluster, a composable accelerator testbed at Texas A&M University. Students with an allocation on the ACES cluster can follow along with the ACES-specific exercises.
Hour of Ci
0
Hour of Cyberinfrastructure (Hour of CI) is a nationwide campaign to introduce undergraduate and graduate students to cyberinfrastructure and geographic information science (GIS).
Raftlib: Open Source library for concurrent data processing pipelines
0
Raftlib is an open-source C++ Library that provides a framework for implementing parallel and concurrent data processing pipelines. It is designed to simplify the development of high-performance data processing applications by abstracting away the complexities of parallelism, concurrency, and data flow management.
It enables stream/data-flow parallel computation by linking parallel compute kernels together using simple right shift operators, similar to C++ streams for string manipulation. RaftLib eliminates the need for explicit usage of traditional threading libraries such as pthreads, std::thread, or OpenMP, which can lead to non-deterministic behavior when misused.
InsideHPC
0
InsideHPC is an informational site offers videos, research papers, articles, and other resources focused on machine learning and quantum computing among other topics within high performance computing.
AHPCC documentary
0
This link is a documentary website to use AHPCC.
Gaussian 16
0
Gaussian 16 is a computational chemistry package that is used in predicting molecular properties and understanding molecular behavior at a quantum mechanical level.
Thrust resources
0
Thrust is a CUDA library that optimizes parallelization on the GPU for you. The Thrust tutorial is great for beginners. The documentation is helpful for anyone using Thrust.
Displaying Scientific Data with Tableau
0
Tableau is a popular and capable software product for creating charts that present data and dashboards that allow you to explore data. It is typically used to present business or statistical data, but can also create compelling visualizations of scientific data. However, scientific data is often generated or stored in formats that are not immediately accessible by Tableau. This seminar will explore the data formats that work best with Tableau and the available mechanisms for generating scientific data in (or converting it to) those formats so that you can apply the full power of Tableau to create the best possible visualizations of your data.
Official Documentation of VisIt
0
VisIt is a prominent open-source, interactive parallel visualization and graphical analysis tool predominantly used for viewing scientific data. Its GitHub repository offers a detailed insight into the software's source code, documentation, and contribution guidelines. In particular, it offers useful examples on how it
Open Storage Network
0
The Open Storage Network, a national resource available through the XSEDE resource allocation system, is high quality, sustainable, distributed storage cloud for the research community.
Advanced Mathematical Optimization Techniques
0
Mathematical optimization deals with the problem of finding numerically minimums or maximums of a functions. This tutorial provides the Python solutions for the optimization problems with examples.
File management of Visual Studio Code on clusters
0
Visual Studio Code, commonly known as VSCode, is a popular tool used by programmers worldwide. It serves as a text editor and an Integrated Development Environment (IDE) that supports a wide variety of programming languages. One of its key features is its extensive library of extensions. These extensions add on to the basic functionalities of VSCode, making coding more efficient and convenient.
However, there's a catch. When these extensions are installed and used frequently, they generate a multitude of files. These files are typically stored in a folder named .vscode-extension within your home directory. On a cluster computing facility such as the FASTER and Grace clusters at Texas A&M University, there's a limitation on how many files you can have in your home directory. For instance, the file number limit could be 10000, while the .vscode-extension directory can hold around 4000 temporary files even with just a few extensions. Thus, if the number of files in your home directory surpasses this limit due to VSCode extensions, you might face some issues. This restriction can discourage users from taking full advantage of the extensive features and extensions offered by the VSCode editor.
To overcome this, we can shift the .vscode-extension directory to the scratch space. The scratch space is another area in the cluster where you can store files and it usually has a much higher limit on the number of files compared to the home directory. We can perform this shift smoothly using a feature called symbolic links (or symlinks for short). Think of a symlink as a shortcut or a reference that points to another file or directory located somewhere else.
Here's a step-by-step guide on how to move the .vscode-extension directory to the scratch space and create a symbolic link to it in your home directory:
1. Copy the .vscode-extension directory to the scratch space: Using the cp command, you can copy the .vscode-extension directory (along with all its contents) to the scratch space. Here's how:
cp -r ~/.vscode-extension /scratch/user
Don't forget to replace /scratch/user with the actual path to your scratch directory.
2. Remove the original .vscode-extension directory: Once you've confirmed that the directory has been copied successfully to the scratch space, you can remove the original directory from your home space. You can do this using the rm command:
rm -r ~/.vscode-extension
It's important to make sure that the directory has been copied to the scratch space successfully before deleting the original.
3. Create a symbolic link in the home directory: Lastly, you'll create a symbolic link in your home directory that points to the .vscode-extension directory in the scratch space. You can do this as follows:
ln -s /scratch/user/.vscode-extension ~/.vscode-extension
By following this process, all the files generated by VSCode extensions will be stored in the scratch space. This prevents your home directory from exceeding its file limit. Now, when you access ~/.vscode-extension, the system will automatically redirect you to the directory in the scratch space, thanks to the symlink. This method ensures that you can use VSCode and its various extensions without worrying about hitting the file limit in your home directory.
ACCESS - Video for new ACCESS users
0
This is a short video on how to exchange ACCESS credits and connect to Jetstream 2 (please note this was created for Duke users but applies to all) .
GIS: What is a Geodetic Datums?
0
Often when working with GIS, or spatial data, one encounters the word "datum" and it may require that you choose a "datum" when doing GIS computation tasks. Below is a short video on what are datums from NOAA and UCAR.
Docker Container Library
0
The Docker container library, commonly known as Docker Hub, is a vast repository that hosts a multitude of pre-configured container images, streamlining the deployment process. It can drastically speed up a workflow, and gives you a consistent starting point each time. Check it out, they might have exactly what you are looking for!
Educause HEISC-800-171 Community Group
0
The purpose of this group is to provide a forum to discuss NIST 800-171 compliance. Participants are encouraged to collaborate and share effective practices and resources that help higher education institutions prepare for and comply with the NIST 800-171 standard as it relates to Federal Student Aid (FSA), CMMC, DFARS, NIH, and NSF activities.
Framework to help in scaling Machine Learning/Deep Learning/AI/NLP Models to Web Application level
0
This framework will help in scaling Machine Learning/Deep Learning/Artificial Intelligence/Natural Language Processing Models to Web Application level almost without any time.