I got a desktop supercomputer? | NVIDIA DGX Spark overview

Video Description

Hey everyone!! Today, I can finally launch a video about my experience with the new *NVIDIA DGX Spark* , a bespoke device for specifically AI workloads, which is great because local AI is literally my day job! In this video, I go over my experience with the DGX Spark, the tools it comes with, and showcase some dead-simple capabilities running OpenAI's GPT-OSS 120B as well as fine-tuning a Gemma 4B model for a customer support use case with *Unsloth* . You can run and fine-tune much larger models on this device. If you have something you want me to try on this - let me know! ## Resources and Links NVIDIA DGX Spark Press Release: https://nvidianews.nvidia.com/ ### AnythingLLM https://anythingllm.com AnythingLLM Desktop: https://anythingllm.com/desktop AnythingLLM Mobile on Google Play: https://play.google.com/store/apps/details?id=com.anythingllm GitHub: https://github.com/Mintplex-Labs/anything-llm ### Unsloth (Fine-tuning) https://unsloth.ai FineTuning Notebook w/Granite4: https://colab.research.google.com/github/unslothai/notebooks/blob/main/nb/Granite4.0.ipynb ## CHAPTERS 0:00 Introduction 0:37 Unboxing the NVIDIA DGX Spark 1:45 DGX Overview, Look, Feel, and Specs 5:00 DGX Available Configurations 6:00 Reddit Criticisms - this is not a personal PC 7:20 The HomeLab setup use case 8:13 External form factor and craftsmanship 10:00 Using the built-in Operating system (DGX OS) 11:00 Running Ollama and GPT-OSS 120b 12:30 Document analysis use-case in AnythingLLM Desktop 16:00 Fine-Tuning demonstration on the DGX Spark! 25:54 Wrapping up my thoughts 26:32 What am I going to do with this now?