RoboTF-AI

Just another engineer fooling around in a GPU powered Kubernetes based AI/LLM Lab. We play around with different GPUs, different Large Language Models, run test

[LocalAI @LocalAI_API]

LocalAI 3.0.0 just dropped!

- Backends as OCI plug-ins, installable in runtime.
- @vllm_project , @huggingface diffusers, kokoro, chatterbox backends
- Smaller and simpler images.
- Voice understanding.
- Latest llama.cpp updates @ggml_org

Enjoy!

[Ettore Di Giacinto @mudler_it]

super excited about this!

[RoboTF-AI @RoboTF_AI]

Building Ubuntu Server for AI LLMs Part 4: LLM Tools companion app for LocalAI. Model configuration editing, docker dashboard, and more!

https://t.co/GDmuSseNXi @YouTube

[RoboTF-AI @RoboTF_AI]

Updates, News, Local AI 2.27.0 Interface overhaul, Sycl Support, Mistral Small 24B during a quick coffee session

https://t.co/nYQay1gRTx via @YouTube

[RoboTF-AI @RoboTF_AI]

LocalAI LLM Testing: Deepseek R1 Distill Llama 70B Deepdive VRAM/RAM at different context levels! We are talking about Deepseek R1 Distill Llama 70B at the VRAM needed at different context levels. Providing more insight on hardware requirements #localai

https://t.co/Sgs3IY6thJ

[Dr. Daniel Bender @drdanielbender]

Mind-blown by this GitHub hack! 🤯

Simply change .com to .gg in any GitHub URL and instantly chat with the entire repo using AI. No more digging - just ask questions and get instant answers!

Big thank you to @nisten who develops it and @thursdai_pod where it was mentioned. 👏

[RoboTF-AI @RoboTF_AI]

Nvidia cutting support for Maxwell, Pascal, and Volta cards in a future release!
🪦
https://t.co/zsL5FO5SIY

[RoboTF-AI @RoboTF_AI]

Behind the Scenes: Unboxing MinisForum MS-01, Major Network upgrades, Deepseek R1 Workbench session This week in the RoboTF lab: We had a major switch failure which leads into snowball effect of full network upgrade to 2.5GB, and 10GB
https://t.co/GgJoegXYWv

[RoboTF-AI @RoboTF_AI]

Building Ubuntu Server for AI LLMs from scratch Part 3: ComfyUI, Open WebUI, Flowise, n8n, and more!

Come hang out with some coffee with other AI home labbers Sunday 11am EST/8am PST

#localai #flowise #n8n #comfyui #openwebui

https://t.co/pDbbzPGxkP via @YouTube

[NVIDIA Studio @NVIDIAStudio]

WIN THE NEXT GENERATION OF RTX ⚡

The #GeForceRTX50 sweepstakes is here & we're giving you multiple chances to WIN a GeForce RTX 5090!

Here's your first chance to enter:

⚫ Like this post
⚫ Comment #GeForceRTX50

[RoboTF-AI @RoboTF_AI]

@NVIDIAStudio #GeForceRTX50 https://t.co/lvLfL4eJ9C

[RoboTF-AI @RoboTF_AI]

Building Ubuntu Server for AI and LLMs from scratch Part 2: Nvidia Cuda Drivers, Toolkit, LocalAI! Getting the node bootstrapped with all the necessary software needed to run our first target of LocalAI (https://t.co/jkpKjOWEcy) with Docker.

https://t.co/lhOCTe2H3F

[RoboTF-AI @RoboTF_AI]

Building Ubuntu Server for AI and LLMs from scratch Part 1: Hardware build, Open Air Case, Multi-GPU
Using parts sitting around with some new ones to build a server and come hang out with a few AI/LLM folks

https://t.co/wdZLVvrZmb via @YouTube

[Josep Ferrer @iamjosepferrer]

The Transformer's encoder clearly explained 👇🏻

[RoboTF-AI @RoboTF_AI]

Cutting up a Dell R720xd to add 2x Nvidia Tesla M40 24GB GPUs - Llama 3.3 70B, Qwen QwQ 32B Preview, Mistral 7B, Gemma 2 testing. Along with real basic SD3 Medium and LTX Video generation testing.
Break out the old junk to do inference!

https://t.co/2mMa3HC2vS via @YouTube

[RoboTF-AI @RoboTF_AI]

3 Uncensored AI comedians do the 12 Days of Christmas and roast AI Engineers with homelabs! 🎄
Happy Holiday!

https://t.co/p5K5RH1QMM

[RoboTF-AI @RoboTF_AI]

LocalAI LLM Testing: Llama 3.3 70B Q8, Multi GPU 6x A4500, and PCIe Bandwidth during inference

#localai #gpuhardware

https://t.co/jErUWtaIzR via @YouTube

[RoboTF-AI @RoboTF_AI]

3 AI Comedians talk about Running a GPU Homelab for AI and Machine Learning using uncensored models! 🦾 🤣

https://t.co/lF7aEmHl0p via @YouTube

[RoboTF-AI @RoboTF_AI]

Project: Count your tokens for Huggingface models! 🪙 AutoTikTokenizer and Streamlit come together for a quick and handy application project.

#autotiktokenizer #streamlit #LLM #Huggingface

https://t.co/pJjnX17CCV via @YouTube

[RoboTF-AI @RoboTF_AI]

I was looking for a good way to estimate tokens on open-source LLM's. I found a nifty project called autotiktokenizer https://t.co/yvknNaWfQK which is fantastic! ⭐️
I wrapped that up in a quick Streamlit app project you can see here: https://t.co/NJ9SEpF61C