- 100
- 177 179
Ominous Industries
Приєднався 16 жов 2023
Want to know more about the robot: www.ominousindustries.com
5 Reasons to Have a Local LLM Setup
00:00 - Intro
01:06 - Privacy
01:49 - Offline Accessibility
03:14 - No Subscriptions
04:19 - Customization and Control
05:27 - Educational Value
06:50 - Outro
In this video, we explore the top 5 reasons to set up your own local large language model (LLM) and how it can benefit you. From enhanced privacy to full customization, discover why a local LLM setup could be a game-changer for your workflow and education.
Privacy
Learn how having a local LLM ensures that your data stays private, with no need to send sensitive information to third-party services.
Offline Accessibility
Explore the convenience of using a local LLM without relying on an internet connection.
No Subscriptions
Say goodbye to monthly fees-once set up, a local LLM eliminates the need for expensive subscription services.
Customization and Control
See how a local LLM gives you the power to fine-tune your model and tailor it to your specific needs.
Educational Value
Discover how running a local LLM setup can be a valuable educational experience, helping you dive deep into AI and machine learning.
01:06 - Privacy
01:49 - Offline Accessibility
03:14 - No Subscriptions
04:19 - Customization and Control
05:27 - Educational Value
06:50 - Outro
In this video, we explore the top 5 reasons to set up your own local large language model (LLM) and how it can benefit you. From enhanced privacy to full customization, discover why a local LLM setup could be a game-changer for your workflow and education.
Privacy
Learn how having a local LLM ensures that your data stays private, with no need to send sensitive information to third-party services.
Offline Accessibility
Explore the convenience of using a local LLM without relying on an internet connection.
No Subscriptions
Say goodbye to monthly fees-once set up, a local LLM eliminates the need for expensive subscription services.
Customization and Control
See how a local LLM gives you the power to fine-tune your model and tailor it to your specific needs.
Educational Value
Discover how running a local LLM setup can be a valuable educational experience, helping you dive deep into AI and machine learning.
Переглядів: 75
Відео
Saving My Dream Laptop - Toshiba Qosmio X305 Teardown
Переглядів 4954 дні тому
I've been dreaming of owning a Toshiba Qosmio X305 ever since I first saw one years ago. Unfortunately, they were discontinued, and the only one I had access to was a Best Buy floor model, which I passed on. Ever since, I've wanted to get my hands on one. In this video, I finally receive this gaming behemoth, take a quick look at its aesthetics, and then tear it down to remove the broken LCD. S...
Raspberry Pi 5 vs Orange Pi 5B: Test of LLM Performance!
Переглядів 1,3 тис.5 днів тому
In this video, we put the Raspberry Pi 5 and Orange Pi 5B to the test by running Microsoft’s Phi-3-mini, a small yet powerful local Large Language Model (LLM) optimized for constrained devices. We dive into the performance differences between these two boards and explore the challenges of running LLMs on the less-supported Orange Pi 5B. For the Raspberry Pi 5, it's smooth sailing-using Ollama, ...
Python Testing Showdown: Gemini 1.5 Pro Vs GPT o1-Preview
Переглядів 2738 днів тому
In this video, we run a side-by-side test of two advanced AI models: GPT-4.0 Preview and Gemini 1.5 Pro 002. The task for both models is to generate a Python script that builds a 25-key piano keyboard using a graphical user interface (GUI), complete with mouse interaction and sound playback. The goal of this test is to determine which model can generate a fully functioning script that produces ...
ChatGPT Advanced Voice Test: Gimmick or Game-Changing Tool?
Переглядів 58013 днів тому
00:00 - Nonsense 00:20 - Intro 00:56 - Voice Options 01:44 - Laptop RAM Removal Walk-Through 07:07 - Speaking A Different Language In this video, I finally dive into ChatGPT’s advanced voice features, which I had been overlooking for a while. Despite using ChatGPT exclusively through its web interface since December 2022, I decided to try out the voice functionality by downloading the app on my...
Making a 3D Printed Delorean Case for The Raspberry Pi
Переглядів 29514 днів тому
In this video, we’re taking a nostalgic trip back to the future by creating a custom Delorean-inspired case for the Raspberry Pi. If you’ve ever thought your Raspberry Pi looks a bit like a time machine when placed just right, you’re not alone! Inspired by the iconic car from Back to the Future, we’ll build a 3D-printed case that makes your Pi look like it's ready to hit 88 miles per hour. This...
Bambu Lab A1 Unboxing and Complete Setup Walkthrough
Переглядів 65322 дні тому
🚀 Unboxing the Bambu A1: Watch as we unveil the powerful and versatile Bambu A1, purchased on sale from MicroCenter. We’ll show you everything that comes in the box and share our first impressions of its design, build quality, and features. 🔧 Complete Setup Guide: Follow along with our step-by-step setup guide to get your Bambu A1 up and running quickly and efficiently. We’ll cover every detail...
FIRST LOOK: Testing OpenAI GPT-o1 'Strawberry' - Can It Build a Video Game from Scratch?!
Переглядів 3,1 тис.25 днів тому
In this video, I explore OpenAI's latest release, the 01 model, rumored to be built on the new Strawberry architecture, by testing its capabilities in creating a synthwave-themed game. I showcase the model's advanced reasoning features and compare its performance to GPT-4. The video includes a live demonstration of the AI generating game assets and mechanics. I wrap up with an analysis of the m...
Building a Retro-Futuristic AI Desktop Chat Buddy
Переглядів 1,3 тис.29 днів тому
Embark on an exciting journey into the world of AI-powered companionship as we build a unique, interactive desktop chat buddy. In this video, I'll guide you through the entire process of creating a voice-activated, 3D-printed AI companion that brings cutting-edge technology right to your desk. At the heart of our build is the versatile Raspberry Pi, complemented by an audio HAT for clear voice ...
FluxMusic Text To Music Generation - Local Test With NVIDIA 3090TI & Gradio
Переглядів 1,1 тис.Місяць тому
In this video, we take a deep dive into the "FluxMusic Text To Music Generation" model by running a local test on an NVIDIA 3090TI GPU using the Gradio interface. FluxMusic is an innovative research project exploring the extension of diffusion-based rectified flow Transformers for text-to-music generation. The model, developed with PyTorch, is designed to convert textual prompts into expressive...
Keychron K10 Pro Unboxing and Mini Review
Переглядів 402Місяць тому
In this video, I’m unboxing my second Keychron K10 Pro mechanical keyboard! After using my first K10 Pro for 6 months, I’m sharing my thoughts and experience with it as I unbox this new one. The reason for getting a second K10 Pro? Switching between it and a cheaper keyboard on my local LLM machine was getting awkward, leading to typing mistakes due to the different key profiles. Plus, it was o...
CogVideoX 2B & 5B Models Fully Tested! Local AI Video Gen Just Got WAY Better!
Переглядів 1,6 тис.Місяць тому
Explore the rapidly advancing world of open-source AI video generation with our in-depth test of CogVideoX, running locally on a powerful dual 3090 Ti setup with Ubuntu 22.04. Watch as we put both the 2B and 5B models to the test, showcasing the exciting capabilities and unique quirks of this emerging technology! 🚀🎥💻 What You’ll See: A comprehensive exploration of CogVideoX 2B & 5B models in ac...
Finishing My Dual 3090TI Custom Loop Water Cooling Build
Переглядів 249Місяць тому
In this thrilling conclusion to our dual 3090Ti AI PC water cooling upgrade, we complete the custom loop that will supercharge our 70B local LLM powerhouse. Building on the foundation laid in Part 1, this video takes you through the final, intricate steps of creating a high-performance cooling system. Watch as we tackle the challenges of hard tube bending, bringing our cooling dreams to life. W...
Prepping The Dual 3090Ti AI PC for Custom Loop Water Cooling
Переглядів 245Місяць тому
In this first part of the water cooling upgrade for my Dual 3090Ti AI PC, we take the next step in optimizing this powerhouse for running 70B localllm models and other GPU-intensive tasks. This build is all about maximizing performance and efficiency, and what better way to do that than by implementing a custom water cooling loop? In this video, we dive into the components that will be the back...
Building a Retro Clip-On Trackball For The Raspberry Pi Zero Laptop
Переглядів 5 тис.Місяць тому
In this follow-up video, we dive deeper into the world of retro-inspired computing by upgrading our unique Raspberry Pi Zero laptop and adding a custom-built clip-on trackball. If you enjoyed the original build, you're in for a treat as we enhance both the functionality and aesthetics of this compact, command-line interface device. Our journey begins with a revisit to Fusion 360, where we desig...
I agree and another awesome video!
Nice!
Your joy is pretty contageous and you've got a very soothing voice. This is a great build video.This is so cool! Its so cuuuteee
Great! Thanks for sharing, that’s exactly what I’m trying to do.
Sure thing! It is a fun setup indeed!
amazing project one thing though, tipying on a ortogonal keyboard has quite a steep learning curve
Thanks very much! Yes, it does indeed. I was lucky that I seemed to adapt to it fairly quickly!
I almost exclusively use my pi via ssh. A 4bit llama3.2 3b runs 4-5tps and is more that plenty for batch inference jobs like summarizing docs. It also is decent at coding so I can use it to generate the code for it to batch process documents.
That is an awesome setup, and really a decent speed for that model on a pi. I generally ssh into them as well, as it makes it easier to copy/paste commands into the terminal lol
The audacity one must have to introduce so many people to something so anciently badass and not even being able to turn it on lmao. This thing looks sick, excited to see it boot!
LOL!! The screen and charger should be here by Wednesday, so hopefully Wednesday I can boot it up for the first time (and hopefully it actually works hahaha)
As a backup machine to move sd cards content to hdd or to create img files from sd cards , do you think it could be a viable option?
Either would be I suppose, however you would need an external sd card adapter as well since the existing sd slot on either would have to be occupied by the OS, unless you were to use an nvme with the pi, which I have never done so couldn't speak to whether or not the sd card slot would still work (though I don't see why not). I would likely go with the rpi for this since it has more community support for any troubleshooting needs!
Amazing laptop, never knew this thing even existed.
Thanks very much, there are so many cool niche older ones out there.
My favorite nestolgia pc is the hptc1100 my favorite
I had never seen that machine until now. That is awesome, way ahead of its time.
I found a HP HDX Dragon with the 20 inch screen in mint condition!
Holy crap, that is a monster laptop, I am extremely jealous at that find!
I have this laptop but the battery is completely dead. Where can I get a battery for it? It's a great multi-media laptop! I actually installed Ubuntu 24.04 LTS on it, and it runs fine! I still have the power adapter. So it runs, as long as I don't unplug it, since the battery no longer charges.
That is awesome, good call on the ubuntu install. I think for this one I may actually do a vista ultimate build lol... RE battery, couldn't say tbh, I planned to just use it only plugged in and didn't see any batteries on ebay at all.
I personally have a QOSMIO X500-122. I replaced the screen in 2017 but to this day, it still works. I changed the battery 3 times with the power supply brick that's pretty much it. I'm also under Linux.
That is an absolute beast of a machine, it is more common now to see 18" screens on high end gaming laptops but that was way ahead of its time in that regard.
I never heard of this laptop lol awesome videos as always!
Thanks very much! Yeah, I don't really see anything posted about these much so outside of remembering it from new for some reason they're pretty unknown. I'm glad I was able to get a screen and charger still.
balmbuab often sells it for $199 rn
Yeah, that's what I got mine for as well!
Something else you can try is image generation, but these little computers might struggle to do the job, as my laptop struggles.
Good idea for a test! Yeah, it is pretty computationally expensive to do image gen indeed.
Thanks for the tests, very helpful!
I am very glad to hear they were helpful, thanks!
I love your work , another upgrade , 🎉
Thanks very much!
Thanks for the setup tutorial. I am waiting for mine to arrive. I love the Suno song generator as I use it as well. I started with an Anet A8 3D printer years ago and have been seduced by Bamboo Labs.
Not a problem! Yes, suno is very cool, especially the newer versions of it. You will be blown away by the advancements in speed and quality that you see with the bambu. There is minimal to no actual printer tinkering required so it is essentially plug and play, a big departure from the older generations of printer!
have you coded advance voice into your chat robots? i guess it would need an api
I have not done that. They use MS Azure for speech and I have noticed some of the newer azure voices are a bit more "emotive", but not on the level of advanced voice mode. I am blown away by what I experienced with advanced voice so far.
I just got exact keyboard and have had it for 2 days now, and so far I enjoy it. I have yet to test it out for my main intended usage that is to write C++ code on Windows with this keyboard, and are currently modifying the keyboards QMK firmware in order to have my own custom macros to save on typing that are frequently used by me. Anyone other got experience with key modifiers because I got some strange issue with MO(0) + 'm' with somehow on release triggers Windows menu after the macro is triggered? I know I will eventually figure it out but the docs weren't clear about this key trapping. The VIA web app is quite decent for mainstream users but with only 16 macros that is way to little for me so the QMK firmware is the way to go for me.
I don't actually use any of the macro features or firmware stuff so unfortunately I wouldn't be of much assistance with this. Enjoy the keyboard though!
@@OminousIndustries Don't worry, it is a very niche problem that I stumbled on. It is replacing my CODE keyboard w/Green MX keys which I have used for 10 years. K10 Pro is probably my new favorite keyboard, I just need to get used to it. Cheers!
@@KimTiger777 It is a wonderful keyboard. The key cap profile took me a month+ to get used to but once I did, it became great to use
we might finally get decent technical support
Most definitely haha
Siiick
Thanks very much!
What motherboard are you using? I need to find one that will fit two big cards like yours.
It is an MSI Pro Z690. With that said, I think any ATX mobo with multiple PCIe slots would accommodate the two cards. Be mindful of temps with a setup like this however, as the cards were getting uncomfortably warm during training.
@@OminousIndustries Hi, what cpu are you using?
@@jesusleguiza77 It is a 12th gen i7
It's interesting.... #SomeNonsense 😂😂😂😂
LOL, I wasn't sold on leaving that in so I labeled it as such hahaha
Thanks
Of course!
I guess that's all the technical support call centers finished. Nice demo.
Yep!
Thanks very much! I had the same exact thought as I was being guided through the laptop fix. I would not want to be a lvl 1 tech right now. Once the option to show it video of what you are working on in combination with a voice mode like this it will be a big hit to those industries.
very cool. hope gemini will catch up soon
It would be good to have options. I have been testing the new Gemini with some programming am impressed at some of what I am seeing in terms of generating error-free code!
On Etsy with the rest of my 3d stuff: www.etsy.com/listing/1785360050/back-to-the-future-delorean-raspberry-pi
How do you hook it up can you use the upstream connector to hook up to the laptop😊
I am not entirely sure what you mean, but you should just be able to plug a regular hdmi or displayport cable into your laptop and connect to the monitor.
What about a color in Photoshop? Its ok?
I don't do any sort of graphic design work that requires complete accuracy, but I have not noticed anything odd in ps or graphically. I recently had some cards printed up that I made in ps on the monitor and they looked correct to what I designed them to be color wise.
4 seconds? as many a woman has said to unfortunate men, why so small?? Genuine question, I notice video ram seems to be a limitation with rendering length for a number of these things, which kind of puzzles me. I don't know code from adam so this is just me sptitballing from a know nothing perspective but if i were building software that needed a GODLY amount of video ram to get say 60 seconds of a,i, video at 360p, I would design it to shuffle into video ram the part you are building at the time and shuffle it out when done to conventional system ram while your gcard builds the next bit and then have some coding to strap it all together for final output. I mean.... 4 seconds... at 240p ? lol, jesus. Neet proof of concept though for sure but sadly it seems i will have to keep looking and hoping something better comes out in the future re hardware requirements. I have 3070 and am probably going to pick up a 3090 for christmas now they are nice and cheap, but 24 gig for 4 fricking seconds at 240p ?!?!? 😂😂
Oh and p.s. yes I know, moores law and all that, ten to 15 years this will be second nature / easy for all to do. But I want it now damnit! I'm getting old.
I understand your feelings here. Since I made this video a better model has been released called Cogvideo, which you can see here: ua-cam.com/video/MFLEQmvugKI/v-deo.html The advancements will happen as you say and in a decade it will likely be something that can be done on a mobile phone. A 3090 would be a great pickup if you're into this sort of stuff!
i see this on sale for 899 at microcenter....To think in 2019 1100$ got me a 1tb 9750h and a 1660ti....now for less you can get a 4060 which runs near the desktop range if not 1:1 as far as cuda cores and 8gb. While offering 8c cpus that game MUUUUCH better than the 3750h(trash) and the 4800h(meh) which ive both had/tested with 1660ti laptops. Even a 69xx apu with a 3070 max q zeph 14...also trash/cooling/maxq graphics....barely sustain 3ghz on cpu....while pushing 115w to gpu. Thing is when it comes to "budget range"/oriented packages im willing to overlook aspects of screen hz/speed.....even image uniformity edge to edge etc....keyboard/track pad quality...if im primarily going to be gaming from dedicated peripherals. Its a compromise that comes with the territory of trying to get a high chip count 1tb as a minimum...and a 4060 laptop at a GOOD PRICE. Any "offerings" at 900$ on sale...that happen to offer "MORE" or a better option as far as image quality, speakers, cooling transfer to keyboard...etc are welcomed...but IMO when BUDGET is purely a priority or budget:component ratio...IMO you cant really fault where theyve cut corners to offer these specs at this price point. Otherwise if you really cared about that you would be considering an OLED just the best value aspect from a pure content consumption/pve gaming standpoint IMO in 1 gpu/cpu generation OLED will be far more common in the 1500$+ segment ppl could definitely get a 60 series nv gpu on a 4k panel on sale by then. Sure 4k seems like a lot...but outside Of AAA titles or hardware hungry api/engines. Ive played Fallen order with my oled HP omen...2070s 100-115w....and with the right settings....it still POPS bc of oled/4k. No doubt a newer gen gpu with dlss+frame gen would be a worth while experience for PVE on an oled laptop when traveling etc....controller good viewing/seating angle. Even non DX12....witcher 3...XD with DLSS enabled and 4k....on mobile has to be a hoot. I wonder if a 4070 laptop could pull that off in dx12 with DLSS3 enabled. Even GTA V maxed with 1.5x on the res scaling slider so nice ppl really overlook the controller/ideal seating position on a laptop.
I agree it was a decent deal, I only use it while sitting and as an actual laptop meaning kb/mouse and only play BeamNG on it so can only speak to that. It has been good so far and I agree I was willing to overlook a higher res screen etc since the laptop is basically used for a single game hahaha.
Nice! Need more videos like this 😶😶
Thanks very much !!
You put Freewrite to shame!
I hadn't heard of that until your comment, thanks very much hahah! Those are cool, remind me of the old alphasmart systems they gave some elementary school children in the early 2000's to type on.
Stable diffusion 1.5 are done much better than these results..
Looks amazing and thinking of buying the build on etsy! What would you say the rough cost is of the entire build?
Thanks very much! Pi - 20, Display about 45-50, battery & Powerboost Charger - 35, MISC about 30, so say in the realm of about $130-150 depending on chosen parts!
Do you know how many m.2 SSD slots it has?
From what I see on the micro center page and some reviews with images, there are 2 ssd slots.
friggin HILARIOUS LOL
LOL
How are you liking it? Temps and fans good? I’m gonna grab one as well!
I am very happy with it so far. I haven't even monitored temps so can't speak to that, fans are pretty loud on full blast but that's to be expected for a laptop like this I suppose. It is a great buy at the price point I got it for!
is the stand height adjustable?
Unfortunately it is not. The height with the stand installed is very high, so I would heavily suggest a wall mount setup for this monitor.
@@OminousIndustries thanks for the reply!
@@Brian-qo8bq Not a problem!
hey there, I was wondering how did you find the screen? any light flooding or colour issues? thanks
When the screen is all black, you can see a small bit of backlight bleeding on the bottom bezel edge, but nothing horrible. I have not noticed anything bad otherwise.
where can i buy the robot???
ominousindustries.com/collections/robots/products/r1d-the-intelligent-companion-robot-for-everyone
You can't beat Bambu Lab. Love my A1 mini. Tip : Drop the initial layer & initial layer infill to 30. The small extra time taken, makes a big quality change to that first surface.
Agreed. I was initially discouraged by how the bambu owners seemed like a "cult", but after having a Creality K1 experience I decided to try a bambu and haven't looked back. The A1 Mini is IMO the best one when factoring in price and ease of setup, it is my favorite. Thanks for the tip, I will try that!
this thing looks fantastic, what a beauty
Thank you very much!
Do you sell a kit that comes with the webcam attachment
No, as this was just experimental and the robot doesn't ship with any webcam related code :)
Is this one available on your shop
This is only available as a DIY 3D printed project that you print yourself and get all the parts for: www.etsy.com/listing/1791303031/retro-futuristic-ai-desktop-chat-buddy
Love your stuff man. I literally bought this printer like 2 months ago but I haven't had the time to set it up and this is exactly what I needed. Also, I've been waiting to buy one of your desktop AI faces because I had the feeling that you were on the vertative releasing a newer version of it. Is there a newer version that's going to be coming out anytime soon? Where I should have just purchased the one that you have out now
Thanks so much for the kind words, I really appreciate them! It is an awesome printer and you will be very happy with it. As for the robot, I am not currently planning any upgrades to it. I was considering an external case but it is not cost effective and wouldn't really add anything to the robot. They will likely stay the same for the foreseeable future - sans updating the code to accommodate API changes if necessary.
This is very cool. I know almost nothing of programming and used the old GPT 4o to make a playable game. I just told it to make a silly game and it came up with "Alien Cowbell Hunt" on its own. I generated some simple graphics+title screen with an image generator and a soundtrack for the title screen with Udio and downloaded some free sound effects. Not that hard at all. The hardest part was to make Windows not see it as malware...
That is a fantastic example of the usefulness of this technology. It can accelerate both productivity and learning compared to being limited to "traditional" methods like one would have been a few years ago.
So Junior programmer job Still safe 😊
Yes, just made more productive with this haha.
Yeah, it can't made a good and original game by itself. Even if you go google you can find more complete good games