Introduction to HuggingFace - The GitHub for ML

hey guys welcome to the first video in this hog andace playlist this one here will be an introduction to what we're going to cover and the hog andace platform so Hawken phase is pretty much just GitHub but for machine learning models and machine learning in general I'm going to cover all the different aspects of the hog andface platform all the different modules that they have all the different libraries and so on how you can use it what they are how to use it easily and use it in real world AI so hogen phas is doing a pretty good job in the open source world so I'm going to cover a bunch of different things both for different domains of AI and machine learning we have natural language processing computer vision also how you can start to combine all of that the different libraries that they have so they have a Transformer Library data set Library they have gradio demos and so on that you can build on top of it how you can do research look up new models they have the models regy where you basically just store all the models you can pull them down use them in a few lines of code with their pipelines so it's really easy to use Hawk and phas with a few lines of code you can actually like have machine learning computer vision natural language processing models up and running so just jump straight into the hawk and face platform and see the platform on a higher level and then we will dive into each individual part of it in separate videos so this is the main page for the hwen face websites up here at the top we have models so it's basically just a whole model registry tons of models out there like pretty much all models available out there you can find them in here we have data set so you can use data set directly in here you can pull them connect them together with the models do infer spin up models end points and use them in your own custom applications and projects we also have the spaces here to explore new models explore demos with gradio so you basically just have a very simple interface that you can share with your colleagues team members and so on for the models and projects that you're working on so immediately when I wake up I take out my phone go into the hogen face platform and then I go inside the daily papers to check out the previous day what are the new papers coming out maybe I missed a day I'll go back in time can see comparison what are the most uploaded ones and so on as well so I basically just used this tool to stay up to dat what is popular right now what are the newest research coming out and this is a really cool feature here we can see they also have upload you can submit papers in here as well but again you can see all the daily papers the most popular ones coming out every single day so I'm just scrolling through it reading the title once you have done this for weeks and even month you start to get a feeling of where is the research actually like going what is the direction for different domains could be lot language models computer vision and so on so here we have July 4th we can go back in time you can just keep on scrolling see all the research papers coming out they say that okay this is pretty interesting you can then tap into it read the abtract you can save them for later or go directly in and read the research papers I have tons of videos about how to read those as well in the most effective way and so on how you can use AI to make yourself more productive efficient and also just asking questions side by side so this is pretty cool definitely goad and check this out I haven't been able to find a tool that comes anywhere near the hog andace daily papers to stay up with the newest Technologies trends of research and the direction AI is going and if we just cover the docs quickly we have the transform module diffuses data set gradio Hub so those are the main modules and Frameworks that we're going to cover throughout this playlist all of this in here it will pretty much be impossible to cover in all details but they have very nice documentation so definitely go in and read about that as well here we're just going to cover the main parts how you can use that connect everything together so we have the model tab over here to the left we can see all the different domains so we have multimodalities computer vision natural language processing audio TBL reinforcement learning and so on and you can see each of the individual sub fields and sub areas within those domains so for our computer vision we have tons of different subdomains depth estimation image classification update detection image segmentation and so on pretty much every single domain and sub area is covered in here and if you just hit one of them let's go inside depth estimation it's just going to show you all of the different models available in here we have 102 models available for depth estimation so you can pull the models directly in here you just specify the path as if you're cloning a GitHub repository specify the path with the pipelines and it will just load the model automatically so you can use it in your own applications and projects but that's for another video we have the models we have a bunch of data sets available in here you can pull them directly as well with just specifying the path so we prettyy much just need to take each individual component test out different models glue them together and then build systems around it which is the direction AI is going and just machine learning in general where we actually want to do useful stuff out in the real world we also have this basis here I use that on a daily basis as well just to figure out what are the new models go through some quick demos see we can just drop in some images get the results on the right side they even have some examples but in here you can pretty much just see everything available you can sort them by spaces of the week trending most liked and so on you can even search for them let's just go inside this Gemma two so it's just a random one it's going to open up this gradio demo where you can do a bunch of different stuff have examples down here you can press on them Syle interface we submit it and then we get the response back we can both use this for chat Bots computer vision applications pretty much everything have the models we have the data set we have demos spin all of it up glue it together and this is the hawk andace platform so I'm really excited for it to just jump into it and show you guys all the different possibilities that you can do with Hawk andface after you have been through this whole playlist you'll get a way better understanding of what is the Hawkin face platform why is it so powerful and also how we can use it because we don't have to write a lot of code they have most of it available out there it is very well supported open source few lines of code and we building machine learning applications and projects if you want to get into an AI career you should definitely check out my AI career program the program is basically all my experiences from how I went from an average student to where I am today the program consists of three main categories we have my technical courses my personal branding course and then the AI career path we have a whole community in there with like-minded people supporting each other and every week we will have weekly live calls where I support and help all of you guys over time more courses resources code templates and so on will be added to the program you will give lifetime access so the sooner you join the more value you will get for your money let me help you take your AI career to the next levels

Channel: Nicolai Nielsen

Views: 444

Keywords: machine learning, hugging face, hugging face tutorial, deep learning, how to use hugging face, machine learning engineer, natural language processing, huggingface tensorflow, hugging face ai, hugging face tutorial for beginners, huggingface transformers, hugging face models, hugging face spaces, hugging face llm, hugging face tutorial 2024, how to use hugging face website, how to use hugging face spaces, huggingface pytorch, machine learning projects

Id: pgl85TzbALg

Length: 6min 55sec (415 seconds)

Published: Tue Jul 16 2024