Digital Twin #4: Wind Turbine Simulation, Industrial Metaverse, Parallel Mind & more!
A Newsletter on Engineering Simulation, Artificial Intelligence, and the Industrial Metaverse
Welcome to the Siml.ai Newsletter. It is a space of news, events, and curiosities from AI, scientific simulation, CFD, and the industrial metaverse.
What is Industrial Metaverse
The industrial metaverse emulates and replicates real-world systems, constructing a highly immersive, real-time, interactive digital environment. It leverages digital twins, artificial intelligence (AI), extended reality, blockchain, and cloud computing to bridge the divide between the tangible and virtual worlds. The field is on the cusp of rapid expansion and substantial investments in digital twin technology.
The industrial metaverse has the potential to transform work processes significantly, providing businesses with a platform to model, prototype, and test designs in a cost-effective and easily accessible digital space, addressing real-world challenges in a virtual landscape.
The industrial metaverse relies on digital twins—virtual replicas of processes in diverse fields. Examples like digital twins used by NASA for the Mars rover mission or digital twins used in Pfizer's cancer research showcase the evolution of digital twins into hyperrealistic physics-based models for accurate simulation. Case studies from Siml.ai highlight the potential of digital twins in numerous other applications.
The industrial metaverse aims to reshape industries by strongly emphasizing sustainability. Through digital twins, physical waste is reduced by enabling experimentation without consuming additional resources.
How do you envision the architecture of valuable and robust industrial metaverses?
Parallel Mind
DimensionLab’s CEO Michal Takáč (DimensionLab is the company behind Siml.ai) started a blog, where he is writing about expanded, and alternative ways of thinking about and reconfiguring technology.
While evolving mathematical methods can describe increasingly complex phenomena of our evolving world, the growing complexity of the problems outpaces current solving capabilities. Enter AI.
The blog covers nuanced insights into the recent advancements in AI, scientific machine learning (SciML), and decentralized scientific computing, foreseeing potential applications in training large-scale machine learning models in a cost-effective and accessible manner. Considering the vast ripples of aforementioned fields, the blog will surely continue to broaden horizons delving into groundbreaking new technologies and protocols.
Make sure to subscribe, read the articles, and stay tuned for more by clicking on the button below.👇
Exploring Wind Turbine Dynamics
Recent years have seen significant progress in renewable energy, with wind turbines becoming prominent symbols of eco-friendly power generation. They play a vital role in the global shift towards sustainable energy by converting wind's kinetic energy into clean electricity. However, understanding wind turbine dynamics is challenging. This blog post delves into wind turbine simulation and addresses its complexities.
What’s new in Siml.ai
We introduced some practical updates to the Siml.ai platform.
You can now view their compute credits usage for the last 24 hours or a month on the billing page.
Navigation through the platform got easier, with better naming and order of links in a way of the common workflow.
You can go through the tutorial of our Model Engineer and learn how to work with it to build fast and accurate simulators.
It is now possible to render only parts of the whole geometry model.
Looking forward to your feedback.
The fastest way to not only share your report but also get feedback and a solution is via our Discord community:
Alternatively, you can find us on ProductHunt, GitHub, and e-mail (hello@dimensionlab.org).
Want to try Siml.ai?
These are the plans we offer. They are modular, depending on the size of your organization and the challenges you face.
News from the field
🦅 Eagle 7B Unleashed: A Milestone in Open Source Language Models
In this prodigious development, the RWKV-v5 architecture welcomes a new era with the introduction of Eagle 7B, a formidable 7.52 billion parameter model — the world's most potent open-source, multi-lingual transformer, characterized by:
📐 Architecture: Eagle 7B, built on the RWKV-v5 architecture, boasts a 10-100 times lower inference cost compared to its counterparts.
🌱 Eco-Friendly Model: Eagle 7B is recognized as the world's greenest 7B model per token, an important contribution to sustainability in the technological landscape.
🦾 Extensive Training Data: Trained on 1.1 trillion tokens across 100 languages, Eagle 7B excels in multi-lingual benchmarks.
⚖️ Performance: In English evaluations, Eagle 7B rivals Falcon (1.5T) and LLaMA2 (2T), while also competing with Mistral (>2T?) and MPT-7B (1T).
💥 Attention-Free Transformer: Eagle 7B operates as an "Attention-Free Transformer," with innovative advancements in large language models.
⚪ Foundation Model: As a foundation model, Eagle 7B requires minimal fine-tuning, offering versatility for developers.
🔓 Open Source and Accessible: Released under the Apache 2.0 license, Eagle 7B is now part of the Linux Foundation for unrestricted personal and commercial use.
🧩 Easy Integration: Accessible via Huggingface, users can download Eagle 7B and deploy it anywhere, even locally. The model is accompanied by a reference pip inference package and various community inference options such as Desktop App and RWKV.cpp.
🔧 Fine-tuning: Developers can fine-tune Eagle 7B using the Infctx trainer for precise customization to specific use cases.
🤗 Community Engagement: A pending pull request aims to merge Eagle 7B into Huggingface transformers, enhancing accessibility for the developer community.
Eagle 7B's launch is a leap in open-source language models, presenting a robust and eco-conscious solution for diverse applications. Developers are urged to delve into Eagle 7B, playing a vital role in advancing the dynamic field of natural language processing.