Joe Freeman: Software Craft, Org Stuff, Tech Stuff
Joe Freeman: Software Craft, Org Stuff, Tech Stuff
  • 370
  • 236 321
Finding NVIDIA-card compatible NIM models and running their docker containers locally
NVIDIA NIMs are ready to run pre-packaged containerized models. The NIMs and their included models are available in a variety of profiles supporting different compute hardware configurations. You can run the NIMs in an interrogatory mode that will tell you which models are compatible with your GPU hardware. You can then run the NIM with the associated profile.
joe.blog.freemansoft.com/2024/08/manually-validating-compatibility-and.html
มุมมอง: 86

วีดีโอ

It's complicated - My sprawling Windows Software Engineering environment for ML and other stuffIt's complicated - My sprawling Windows Software Engineering environment for ML and other stuff
It's complicated - My sprawling Windows Software Engineering environment for ML and other stuff
มุมมอง 31หลายเดือนก่อน
Modern development is surprisingly complicated. My personal software development environment is suffering a severe case of urban sprawl. While working on some Container IAC scripting for ML and the cloud, I had to pop into a Linux environment. This environment map made me realize that I have a crazy set of different specialized sandboxes. Is yours like this? Windows, Linux, Containers, WSL, Ser...
See tabAutoComplete AI assist and Chat AI assist relying on local LLMs in 4 minutesSee tabAutoComplete AI assist and Chat AI assist relying on local LLMs in 4 minutes
See tabAutoComplete AI assist and Chat AI assist relying on local LLMs in 4 minutes
มุมมอง 46หลายเดือนก่อน
A demonstration of using tabAutoComplete and chat in Visual Studio code with local LLMs using Ollama and continue.dev Windows with NVidia 8GB: joe.blog.freemansoft.com/2024/08/get-ai-code-assist-vscode-with-local.html Mac: joe.blog.freemansoft.com/2024/08/run-ai-code-assist-locally-in-vscode.html Windows with NVidia 24GB: joe.blog.freemansoft.com/2024/08/rocking-older-titan-rtx-24gb-as-my.html
Using local Large Language Models for AI code assist in Visual Studio CodeUsing local Large Language Models for AI code assist in Visual Studio Code
Using local Large Language Models for AI code assist in Visual Studio Code
มุมมอง 87หลายเดือนก่อน
We can run LLM models locally, assuming we have enough VRAM or a big enough MAC, and use those models for Coding assist. No messages have to leave our network and we don't have to worry about unexpected costs due to orphan processes. This can also work for local servers Windows with NVidia 8GB: joe.blog.freemansoft.com/2024/08/get-ai-code-assist-vscode-with-local.html Mac: joe.blog.freemansoft....
Nvidia Inference Microservices - AI Workbench NIM-Anywhere Project ComponentsNvidia Inference Microservices - AI Workbench NIM-Anywhere Project Components
Nvidia Inference Microservices - AI Workbench NIM-Anywhere Project Components
มุมมอง 2072 หลายเดือนก่อน
The video walks you through the containers and networking that make up the NIM-Anywhere project. The project is evolving so the topology and components will change.. NVidia Inference Microservices is a model for deploying inference engines as microservice endpoints. They demonstrate using this in the AI Workbench NIM-Anywhere project available on GitHub. The project contains a set of applicatio...
Run the NVidia hybrid rag workbench example in a container using AI WorkbenchRun the NVidia hybrid rag workbench example in a container using AI Workbench
Run the NVidia hybrid rag workbench example in a container using AI Workbench
มุมมอง 2123 หลายเดือนก่อน
Download, choose an LLM, configure, run, and search using a hybrid RAG running in a container with its own localized vector database. This GitHub repository github.com/NVIDIA/workbench-example-hybrid-rag can be run as a containerized server NVidia AI Workbench on NVidia GPUs. This video records the first time I tried downloading, running, and using this RAG project. Some mistakes were deleted. ...
NVidia AI Workstation fixing the container config when the container build failsNVidia AI Workstation fixing the container config when the container build fails
NVidia AI Workstation fixing the container config when the container build fails
มุมมอง 1073 หลายเดือนก่อน
I broke my AI Workstation container image configuration so that I couldn't get an image built. This meant I couldn't edit my repository to edit back. Turns I could use the AI Workstation Environments tab to clean up the configuration to build a new image. NVidia AI Workstation provides managed containerized environments for CUDA and other Jupyter Notebook and Tensor Lab workloads. AI Workstatio...
Retaining Focus - How Deep Do We Need to OwnRetaining Focus - How Deep Do We Need to Own
Retaining Focus - How Deep Do We Need to Own
มุมมอง 443 หลายเดือนก่อน
Organizations must decide which processes are their secret sauce, their differentiators, things that make them stand out. They decide which criteria to use for that differentiation. Some are narrowly focused. They decided some processes or components are standard "off the shelf" and can be purchased or leased. Others decide everything is a differentiator. They need top-to-bottom ownership and c...
Camtasia 2022 experimentation with iPhone Continuity Camera on an M1 MacBookCamtasia 2022 experimentation with iPhone Continuity Camera on an M1 MacBook
Camtasia 2022 experimentation with iPhone Continuity Camera on an M1 MacBook
มุมมอง 533 หลายเดือนก่อน
I've not been satisfied with the Logitech C922 when hooked to my Macbook Pro. Camtasia has been offering to let me use my iOS devices but that seemed like such an Apple solution. Turns out an old iPhone can make a great Camtasia camera. My Samson Q2U is still the best microphone but the Camtasia noise reduction function renders that sort of mute Yeah, others probably already knew this! In this ...
Flutter messaging classes and their native platform peersFlutter messaging classes and their native platform peers
Flutter messaging classes and their native platform peers
มุมมอง 293 หลายเดือนก่อน
Communication between Flutter modules and Mobile Native code happens over platform channels. The docs are all over the place so I created this cheat sheet for the different channel types and their class documentation. joe.blog.freemansoft.com/2024/07/flutter-messaging-classes-and-their.html github.com/freemansoft/flutter-embedded
Bidirectional messaging between ios and FlutterBidirectional messaging between ios and Flutter
Bidirectional messaging between ios and Flutter
มุมมอง 223 หลายเดือนก่อน
The simplest of all demonstrations or at least a demonstration with the least amount of code. Sending JSON messages between Flutter and an iOS wrapper. This uses the Flutter template generated iOS application and edits one file on the iOS side. Repository github.com/freemansoft/flutter-embedded
Sending messages from Flutter to iOS using the BasicMessageChannelSending messages from Flutter to iOS using the BasicMessageChannel
Sending messages from Flutter to iOS using the BasicMessageChannel
มุมมอง 403 หลายเดือนก่อน
A short video demonstrating using the BasicMessageChannel to send messages from Flutter to an iOS application. The Flutter code is exactly the same code used for the Android bidirectional example. This video includes iOS receiving messages. It does not including sending the messages because I haven't created that code yet. Code: github.com/freemansoft/flutter-embedded Blog: joe.blog.freemansoft...
How big are the NVidia AI Workbench container imagesHow big are the NVidia AI Workbench container images
How big are the NVidia AI Workbench container images
มุมมอง 553 หลายเดือนก่อน
REPOSITORY SIZE project-rapids-cuda-12-0-1 20GB project-python-cuda-122 3.11GB project-pytorch-cuda-122 3.11GB project-python-basic 1.23GB
Using GIT from inside a Jupyter Notebook ML Container - What NVidia AI Workbench needs...Using GIT from inside a Jupyter Notebook ML Container - What NVidia AI Workbench needs...
Using GIT from inside a Jupyter Notebook ML Container - What NVidia AI Workbench needs...
มุมมอง 963 หลายเดือนก่อน
Jupyter Notebooks provide a browser-based experience that provides an integrated development environment in your local browser no matter where the work is being done. The AI Workbench lets you experiment, train, and run ML workloads in remote containers. This is an awesome way to support encapsulation and remote execution. Version control and GIT integration are the missing links when you are s...
Migrating Native Mobile to Flutter Entrypoint by EntrypointMigrating Native Mobile to Flutter Entrypoint by Entrypoint
Migrating Native Mobile to Flutter Entrypoint by Entrypoint
มุมมอง 273 หลายเดือนก่อน
Organizations with significant investments in native applications will probably migrate flutter feature-by-feature or navigation flow by navigation flow. They may be able to migrate from the back of the app to the front of the application with all of the Flutter code in a single package or bound to a single every-growing function set. The alternative approach is to migrate targets of opportunit...

ความคิดเห็น

  • @CarmenQuincy-g2h
    @CarmenQuincy-g2h 5 วันที่ผ่านมา

    Thomas David Thomas George Robinson Gary

  • @MO-ss7qt
    @MO-ss7qt 12 วันที่ผ่านมา

    This is an older video and I would imagine my question will fall on deaf ears. But just in case, I'll toss is out there anyway. Can you import a png file that is an export of a powerpoint block/line drawing containing images for the blocks and a variety of line styles representing everything from a phone line to hdmi? I want to move out of Microsoft Office but can't find a suitable alternative to handle this complex drawing without starting over.

  • @dcktater7847
    @dcktater7847 17 วันที่ผ่านมา

    Too bad you cant save the alignment pins as .stl...probs the only good thing about prusa is this cutting tool it's soo good and straighforward unbelieavable that there isnt a standalone stl editor that can do this and makes a usable stl file. I checked this meshmixer and blender and whatnot but that is an utter piece of crap relative to this cutter...

  • @anycaroliny7900
    @anycaroliny7900 25 วันที่ผ่านมา

    It works, thank you!

  • @xpectz
    @xpectz หลายเดือนก่อน

    This no longer works.

  • @Vapor_Space
    @Vapor_Space หลายเดือนก่อน

    Excellent tutorial Thanks.

  • @shivamjalotra7919
    @shivamjalotra7919 หลายเดือนก่อน

    this is easily the best video on this topic I have seen

  • @FreemanSoft
    @FreemanSoft หลายเดือนก่อน

    I made a mistake when I recorded the video and corrected it with some annotations. TP1, TP2, TP4, TP8 are measures of Tensor Parallelism and not tensor core generations. I have one video card so in general I look at the TP1 models.

  • @DavidEliasFloresEscalante
    @DavidEliasFloresEscalante หลายเดือนก่อน

    Greetings, Thanks for these videos. A question, is it possible to build a "custom" container, for example one using cuda 10, 11, or whatever? Thanks in advance

    • @FreemanSoft
      @FreemanSoft หลายเดือนก่อน

      You should be able to build an image using one of the NVidia base containers like they essentially do in their AI workstation. They have several Cuda versions and Rapids. Or, you could build one from scratch or find one a Container Repo.

  • @markfantone1253
    @markfantone1253 หลายเดือนก่อน

    Simply amazing! Thanks for the script. I was able to make this work with Windows 11.. But I have encountered some problems, I am using two 1TB HDD and one 120GB SSD. Whenever I try to install a more than 100GB game, the game detects that I don't have enough storage space. Is this expected since I'm using a 120GB SSD?

  • @13texasmom
    @13texasmom 2 หลายเดือนก่อน

    Hi Joe, Tony from your previous life!

  • @RyanKrausM
    @RyanKrausM 2 หลายเดือนก่อน

    Thanks for the shout out and your feedback, Joe!

  • @PuneetMakhija-b4q
    @PuneetMakhija-b4q 2 หลายเดือนก่อน

    can you please send me the project link

  • @PuneetMakhija-b4q
    @PuneetMakhija-b4q 2 หลายเดือนก่อน

    Hi Joe, can you please send me this project link

  • @marufhassan634
    @marufhassan634 2 หลายเดือนก่อน

    Your videos are very good, please keep making more.

  • @BrunoidGames
    @BrunoidGames 3 หลายเดือนก่อน

    I just fell in this hole.. thanks for your helping hand.

  • @mdsamidulislam1920
    @mdsamidulislam1920 3 หลายเดือนก่อน

    You know what? I am a regular viewer of your TH-cam Channel. I have seen all your videos and your content qualities are very good. I learned lots of things from you. Your channel has great quality contents but still it is not delivering its right audiences due to SEO problems. You should focus on your SEO so that You can get a big fanbase. Thanks.

    • @FreemanSoft
      @FreemanSoft 3 หลายเดือนก่อน

      Thanks for the comment. I have no idea how to do SEO😥

  • @iukeay
    @iukeay 3 หลายเดือนก่อน

    Have you played with wsl2 Dev drives? I am working with 80-144 models I'm noticing that my read speeds of just reading the models into gpu memory is about 10x slower then it should be. Just curious. Was going to move all my docker volumes to

  • @DaBo1
    @DaBo1 3 หลายเดือนก่อน

    Hello! I like your video about eye contact. Please tell me which model graphic card you have. Thank you!

    • @FreemanSoft
      @FreemanSoft 3 หลายเดือนก่อน

      developer.nvidia.com/blog/improve-human-connection-in-video-conferences-with-nvidia-maxine-eye-contact/

  • @GeorgiSapunkov
    @GeorgiSapunkov 3 หลายเดือนก่อน

    I`m new to 3d printing and I didn`t know that you can put joints in cuts :) Now I can make biiger thing on my 220x220mm plate \o/

    • @FreemanSoft
      @FreemanSoft 3 หลายเดือนก่อน

      I ran into issues with the sword because I only had a single dovetail. In that case, I needed to increase the fill and perimeter wall count because even model swords are used to whack things.

  • @racielrodriguez1544
    @racielrodriguez1544 3 หลายเดือนก่อน

    Excellent video

  • @sprintcog
    @sprintcog 3 หลายเดือนก่อน

    Great video, thanks for such a clear explanation!

  • @sebastianacevedosolano6804
    @sebastianacevedosolano6804 3 หลายเดือนก่อน

    Nice video, the explanation was very easy to understand

  • @AbdulRevuru
    @AbdulRevuru 3 หลายเดือนก่อน

    Wow!!! Super explanation. Finally got Openshift CRC installed on my WSL Oracle Linux, and yay! can access it from my Windows Host machine.

  • @1DoseFinance
    @1DoseFinance 3 หลายเดือนก่อน

    such a cool feature for talking head clips , video explainers commentaries etc

  • @ravi-mo6js
    @ravi-mo6js 4 หลายเดือนก่อน

    best explaination on the internet! subscribed

  • @sordidloam
    @sordidloam 4 หลายเดือนก่อน

    I spun this up, but now Entra authentication isn't working to my VPN users for the file shares over SMB. Been working on this for about 9 hours.

    • @Snowboardaccus
      @Snowboardaccus 2 หลายเดือนก่อน

      I ended up having to use DNS Private Resolver service which has a single inbound IP. Deploy it to the same vnet as your VPN GW and it solved all my issues immediately with the filehsare.

  • @jw27
    @jw27 4 หลายเดือนก่อน

    Hi there, good to find somebody who faced the same problem with me. I got a esp32-c3 super mini dev board, I can use it to scan for WIFI, it works great. But it can't connect to any of home Tenda-2.4G-Wifi, Iphone11 hotspot, Huawei hotspot, and iphone can't even find the AP c3 created. As you mentioned, Wifi created by windows cant be utilized by c3, but I dont have a Windows laptop, there is no alternative for similar app on Macbook, I just cant try it. Have you found the reason why only Windows hot spot? What makes it so special.

  • @muhammadamirulsamsulbahar2544
    @muhammadamirulsamsulbahar2544 5 หลายเดือนก่อน

    Short and precise. Really helpful. Thanks!

  • @robertshoop8181
    @robertshoop8181 5 หลายเดือนก่อน

    Does anyone know if you can add storage spaces tiering to an existing storage spaces configuration created on Server 2019 Standard?

  • @akashtriz
    @akashtriz 6 หลายเดือนก่อน

    Wow what a well explained diagram that finally shows the method to the madness

  • @mahendranmani3951
    @mahendranmani3951 6 หลายเดือนก่อน

    Well explained🙂

  • @nallwhite4146
    @nallwhite4146 6 หลายเดือนก่อน

    Windows no longer has a seat at the executive table. There is a lot that needs to be removed, combined and improved.

  • @thomastempe3301
    @thomastempe3301 6 หลายเดือนก่อน

    So, how do you make that behavior go away?

  • @dlross173
    @dlross173 7 หลายเดือนก่อน

    Interesting about the eye contact. Can it be quickly turned on and off? Like if I want to do a powerpoint, and I can turn it off when I'm gesturing and have to turn my head, but if someone asks a question, I can look directly into the camera and answer it while I'm checking my notes off to the side? then turn it off when I get back to my presentation? If the answer is yes, I'm getting it! Hahaha... thanks for the video.

  • @hotwired2424
    @hotwired2424 7 หลายเดือนก่อน

    No need to build vm or container anymore. Checkout Azure DNS Private Resolver.

  • @IvanTsaryk-p3f
    @IvanTsaryk-p3f 7 หลายเดือนก่อน

    Thank you, Joe! This is so usefull 👍

  • @AL2013
    @AL2013 7 หลายเดือนก่อน

    Rtx 4060 Mobile 8gb nao deu certo. ...chatbot

    • @FreemanSoft
      @FreemanSoft 7 หลายเดือนก่อน

      Not sure I understand the question. Google translate didn't tell me enough. Shared memory and physical VRAM are not equivalent. Integrated video memory adjustments in the bios are only for an integrated GPU and not the NVIDIA chipset As far as I know the 8GB or 12GB VRAM requirement form ML is for actual attached video memory and you can't change it to allocate more memory to the NVIDIA card.

  • @althaf1
    @althaf1 8 หลายเดือนก่อน

    Great video, I appreciate if you can help me some solution to my Query

  • @TahaAlars
    @TahaAlars 8 หลายเดือนก่อน

    u r the best o all the time thank you

  • @user-fr5b3w734
    @user-fr5b3w734 8 หลายเดือนก่อน

    😂😂😂😂🤣🤣🤣 ikea Alarm clock DIY with lights

  • @manaswidaksha137
    @manaswidaksha137 9 หลายเดือนก่อน

    Thanks for the tutorial!

  • @RamdanCode
    @RamdanCode 9 หลายเดือนก่อน

    Thanks a ton for this guide! Clear and concise steps for Flutter SSL setup. Real-life saver! 🚀 Your straightforward instructions made it a breeze. Kudos for keeping it human and not overwhelming. Much appreciated!

  • @RamdanCode
    @RamdanCode 9 หลายเดือนก่อน

    Absolutely impressed by the seamless migration process! Your detailed walkthrough made it easy to grasp the complexities involved. Kudos for the clear explanations and practical tips. Thanks a ton for sharing this invaluable knowledge, it's a game-changer for anyone considering Flutter for web development. Much appreciated!

  • @stephenzzz
    @stephenzzz 9 หลายเดือนก่อน

    👊

  • @tiddybagel
    @tiddybagel 9 หลายเดือนก่อน

    I bought this unit and it only reaches 102 F at the max setting...not even hot enough for PLA

    • @FreemanSoft
      @FreemanSoft 9 หลายเดือนก่อน

      I melted the outer layer of my PLA when I accidentally let it run overnight on max

    • @tiddybagel
      @tiddybagel 9 หลายเดือนก่อน

      @FreemanSoft I had to cover it with foil to get it to a decent temp. After that it maintained 105. I'm just using the lid it comes with with a 3d printed ring

  • @WilsonMar1
    @WilsonMar1 10 หลายเดือนก่อน

    As if all this isn't confusing enough, Microsoft renamed AAD to Entra ID in 2023.

  • @marktompkins3180
    @marktompkins3180 10 หลายเดือนก่อน

    Thank you for the amazing discussion.

  • @mehdi-vl5nn
    @mehdi-vl5nn 10 หลายเดือนก่อน

    0:05 lol "do" in Farsi means two =)))))

  • @DenysonData
    @DenysonData 10 หลายเดือนก่อน

    Super helpful. Landed at your blog while preparing for an AWS exam and now find this video applicable to the situation I found myself inside of the new org.