Francesco De Giorgi

Honestly as a swe this is one of the most useful thing I did for software development. github.com/mfranzon/tdd

This is MicroCAD. The programming language for your 3D projects. You can design your own Lego block in a few lines of code, ready to be exported as an STL file for 3D printing.

Simulate a data center airflow. Gemini 3 + threejs. It works. The plan now is to work directly on the physics engine: Navier-Stokes on the web with WebGPU would be incredible. Imagine running full CFD simulations directly in the browser, this means countless applications.

Explore the data center digital twin with a VR headset. - Move the racks - See the airflow change - Make a digital twin cool like a VR videogame We are making CFD funny as a game.

A tool that pairs food with wines. That's what I build. Have a look winy.cloud

🤯Video chat with real Agents actions🤯 💬 Real-Time Language Assistance 🎯 Personalized Recommendations 📚 Interactive Learning ⚙️ Customizable Features ...and more! (I choose emoji by my own, appreciate the effort) Real-Time Language Assistance: One of the most significant benefits of incorporating LLM into your video chats is real-time language assistance. Whether you're struggling to find the right words, need help with translation, or want to clarify a complex concept, LLM can provide instant suggestions and guidance to ensure smooth communication. Personalized Recommendations: LLM can analyze the context of your conversation and offer personalized recommendations tailored to your needs. From suggesting relevant topics to providing additional information on a subject, LLM can enrich your discussions and make them more engaging. Interactive Learning: Whether you're conducting virtual training sessions, tutoring sessions, or language exchanges, LLM can enhance the learning experience by providing instant explanations, examples, and practice exercises. With LLM's support, learners can receive personalized guidance and feedback, accelerating their progress. Customizable Features: Depending on your preferences and needs, you can customize LLM's features to suit your video chat experience. Whether you want to focus on language assistance, brainstorming, or learning, you can adjust LLM's settings accordingly for optimal performance. Want to see it in action, stay tuned!😁

Hey @Docker lovers (or not)!🐳 If you're looking for a good place to find ready-to-go Dockerfiles, tips and tricks this repository is what you need! github.com/mfranzon/docker-i…

Today, I am at #mecspe in Bologna, Italy. Hope to meet other shine happy people to discuss about digital twin and generative AI🚀

How to make a RAG nowadays🧠 Technologies Powering a RAG: @langchain and @ollama: LangChain and Ollama serve as the backbone of a RAG, providing robust infrastructure for model development, training, and serving. LangChain offers a comprehensive suite of tools for natural language understanding and processing, while Ollama specializes in serving LLMs models, enabling a RAG to seamlessly integrate and process diverse data types. @FastAPI for REST API Service: FastAPI, known for its high performance and simplicity, serves as the interface for a RAG, facilitating seamless communication between the backend system and external applications. Its asynchronous capabilities ensure rapid response times, crucial for real-time interactions. @trychroma or @elastic : RAGs offers flexibility in data storage and retrieval by supporting both database like ChromaDB and Elasticsearch. ChromaDB provides a lightweight solution suitable for simpler tasks, while Elasticsearch caters to more complex operations involving vast amounts of data. @Docker : Docker will give you the possibility to containerize you infrastructure and make it scalable or deployed on cloud services. This is the next step to make in production a rag and not just use it in your local machine!

Hey devs! I'm thrilled to announce that I've officially been appointed as a @Docker Captain! 🐳⚓ It's an absolute honor to join the ranks of this prestigious community of innovators, educators, and passionate Docker enthusiasts. 🚀🤯 Shoutout to the super nice @eeveeta !

Exact-RAG for finance📊 FinTral integrates textual, numerical, tabular, and image data, and it is available on @huggingface FinTral was trained with domain-specific data, finetuned instructions and RLAIF training by exploiting a large collection of textual and visual datasets.

1-bit LLMs🤯 New models from @Microsoft that use ternary values (-1, 0, 1) instead of 16-bit. This makes them 2.7x faster, use 3.5x less GPU memory, and 71x less energy.🚀

Merged UI in exact-rag👌 I used @streamlit and it was super fast to make. Next step give the possibility to switch model in live from UI github.com/exactlab/exact-ra…

I got many different questions from who wants to approach the RAG's world. I am trying to summarize the answers with this post. 1⃣What is Retrieval Augmented Generation (RAG) for LLMs? ☑️Retrieval-augmented generation (RAG) for large language models (LLMs) aims to improve prediction quality by using an external datastore at inference time to build a richer prompt that includes some combination of context, history, and recent/relevant knowledge. RAG LLMs can outperform LLMs without retrieval by a large margin with much fewer parameters, and they can update their knowledge by replacing their retrieval corpora, and provide citations for users to easily verify and evaluate the predictions. 2⃣Why is there a need for RAG LLMs? ☑️Pre-trained LLMs (foundation models) do not learn over time, often hallucinate, and may leak private data from the training corpus. To overcome these limitations, there has been growing interest in retrieval-augmented generation which incorporate a vector database and/or feature store with their LLM to provide context to prompts, also known as RAG LLMs. 3⃣How does RAG compare with Fine-Tuning? ☑️Fine-tuning takes a pre-trained LLM and further trains the model on a smaller dataset, often with data not previously used to train the LLM, to improve the LLM’s performance for a particular task. In contrast, you should use RAG when you are able to augment your LLM prompt with data that was not known to your LLM at the time of training, such as real-time data, personal (user) data, or context information useful for the prompt.

Predictive maintenance in public. Speaking about one of the most concrete themes for industries in front of many entrepreneurs is always a moment to grow, learn and share. @dualistic_twin