The Cloud Girl

Month: June 2025

Uncategorized

Complete Beginners Guide to Hugging Face

by Priyanka Vergadia

hey everyone and welcome back to mychannel where we talk about cloud techand AI And today we're diving into aplatform that you must know about ifyou're doing anything to do with AI Itis Hugging Face Now Hugging Face hasbeen called the GitHub of machinelearning and for all the great reasonsIt is literally becoming the communitywhere AI models and creations are sharedacross everybody So by the end of thisvideo you will understand what it isexactly and why it matters to you evenif you are not somebody who codes everyday So stick around All right So this isDemothe hugging face homepage And the firstthing that you'll notice in here istheir tagline which is the AI communitybuilding the future That really sums upwhat they're really about right it's acollaborative platform where people aresharing AI tools models data sets andeven AI apps Now if you scroll downyou're able to see features models thatare trending and recently uploadedcontent Um and this gives you the tasteof what's popular in the AI communityAnd before we dive deeper I wouldrecommend that you sign up and create anaccount because you would need one Uhmost of the content you'll be able tojust see um without uh having an accountbut if you want to use the models andsave your favorites and things like thatyou will need an account Now um goinginto the models tab this is where allthe models are found You can see thatthey've got millions of models in hereand you can filter them by differenttasks in the categories like naturallanguage processing and classificationand audio and tabularAnd you could also filter them bylibraries and data sets and languagesandlicenses Now here let's check out uMicrosoft's popular um 54 reasoningmodel And I wanted to see how far I cango with this So each model has got itsown page with documentation and Iclicked on deploy and it literally justtook me right away into the machinelearning studio in Azure AI and I wasasked to create a workspace and as soonas I gave it all the details with thename and everything um it was able tocreate that workspace for me anddeployed that model the 54 model um fromhugging face into Azure AI machinelearning studioThat was absolutely amazing I just itjust took me a few clicks to do this Andyou can see it's creating that now Andonce it is created I can go to theworkspace And in this workspace if Iclick on endpoints if I click onendpoints I'm able to go into AzureOpenAI service And that's where my 54endpoint is And if I want to use thisendpoint I can click on continue Andthat takes me into Azure AI Foundrywhere I'll be able to um experiment withthis with this deployed model It tellsme my target URL the key that it createdfor me I'm able to see how to use thiswith my API key and um some samples ofhow to use this model Um I am also ableto go play with it in the playground andtest it out So I gave it a prompt Um andum I was really trying to go with likethe dog traveling uh to the mountainswhere he meets a robot and um that robotis helping a bird um survive in the coldum and they all become friends for lifeSo um I was just playing around withwith um a prompt but the idea here isthat you're able to go from looking at amodel in hugging phase to actually ableto deploy that model in Azure AI founduh foundry and uh Azure AI machinelearning studio Um and then I clicked ondeploy that um endpoint as a webapplicationAnd right now that is what you're seeingum with the Azure AI web applicationbeing created um as a part of thisdeployment It's able to deploy a webapplication right from that um thatmodel and endpoint that we just createdOnce the model this takes seconds maybea minute or so to to get created withthe deployment assets and stuff thingslike that And once the app is deployedI'm able to see that app in Azure AIfoundry in my web apps section There itis the 54 experiment I click on that appand there we have it An entire chatapplication built from hugging facechoosing a model 54 reasoning goes intoAzure AI machine learning studio foundryand builds it out for me as a webapplication Going back into our huggingWalkthroughface interface we're able to uh let'slook at the the the data sets tab Nowthis tab is where all your data sets areThere are thousands of these in thereand um you can preview the samples inthere You can also filter the data setsof through languages tasks libraries allof that Um and then if you click on oneyou're able to actually see the samplesof that of the data set um and startusing them Um and now the next thing isone of my favorites which is the spacessection of hugging face Now this iswhere things get really really excitingespecially for non-coders So spaces isthis interactive AI application thatanyone can use right within the huggingface browser experience And think ofthem as like readytouse AI tools Iclicked on one here which is calleddescribe anything um in uh by created byNvidia And when you go into that modelagain right in the browser I'm not doinganything else I can upload an image Anduh once I do that um I can type mydescription and I can get my descriptionfor for the regions of my images This ismy dog sitting on a chair in a park Andum I selected different parts of thatimage And um this model is able to thisdemo is able to tell me what are inthese different parts Um I selected thetree first and then I selected my doghimself and uh it was able to do areally good job at telling me what is inthis image Um and if I wanted I can takethis space and deploy it for myselfwhether locally or um or in cloud Um andbut before we do that let's look atanother example So I go back into myspaces I can really um you knowcategorize by image generation 3Dmodeling all the different options uptop Um and I went into stable diffusionwhich is another one of the very commonand very popular libraries in gener ofuh image generation models And um Itested this one out right here in spaceswith a prompt serene lake at sunset withmountains in the background and a goldenretriever watching the sunset I let itgenerate the image And there we have itUm I don't know if I like the first oneThe second one's okayUm but it it did what I wanted it to doUm and let's say I'm happy with withwhat it's I love the third and thefourth images Um they really do what Iasked it to do The the good part thebest part the part that I want to showyou is I can run this space Let's say Ilike it I can run it locally I can runit um I can clone the repo um and um andstart working with it right from herejust like how we deployed the um thefive for model in Azure AI and uh withthat um let's look at the docs the docssection is uh your knowledge center thisis where you are going to get deepdeeper technical information the docsare organized by different categorieslike the client libraries deploymentinterface core ML libraries like thetransformers which is one of the veryfamous libraries diffusers tokenizers umand a lot more like radio Um and thenthe next thing the last thing I want totalk to you about is the communitysection This is where people askquestions and share ideas and learn Theblog part of the of the community isamazing you'll see a lot of peoplecontributing to the blogs and you'll seeum what's happening right now um and andwhat's hot right now Then the learnsection is one of my favorites The LLMcourse and the agent course are some ofthe best courses out there on AI andmachine learning right now The LLMcourse goes from transformers all theway up to fine-tuning And then the agentcourse covers everything from intro toagents to to a lot more So that myfriends was hugging face and we'vetoured every major section of theplatform Whether you are just curiousabout AI want to use existing models orare developing something with AI or wantto contribute hugging face is definitelya platform to check out Now go exploreAnd if you liked this video and found ithelpful please hit that like andsubscribe button to get more tech and AIcontent And drop a comment if you havequestions and which AI platform I shouldcover next And thank you for watchingSee you next timehey everyone and welcome back to mychannel where we talk about cloud techand AI And today we're diving into aplatform that you must know about ifyou're doing anything to do with AI Itis Hugging Face Now Hugging Face hasbeen called the GitHub of machinelearning and for all the great reasonsIt is literally becoming the communitywhere AI models and creations are sharedacross everybody So by the end of thisvideo you will understand what it isexactly and why it matters to you evenif you are not somebody who codes everyday So stick around All right So this isDemothe hugging face homepage And the firstthing that you'll notice in here istheir tagline which is the AI communitybuilding the future That really sums upwhat they're really about right it's acollaborative platform where people aresharing AI tools models data sets andeven AI apps Now if you scroll downyou're able to see features models thatare trending and recently uploadedcontent Um and this gives you the tasteof what's popular in the AI communityAnd before we dive deeper I wouldrecommend that you sign up and create anaccount because you would need one Uhmost of the content you'll be able tojust see um without uh having an accountbut if you want to use the models andsave your favorites and things like thatyou will need an account Now um goinginto the models tab this is where allthe models are found You can see thatthey've got millions of models in hereand you can filter them by differenttasks in the categories like naturallanguage processing and classificationand audio and tabularAnd you could also filter them bylibraries and data sets and languagesandlicenses Now here let's check out uMicrosoft's popular um 54 reasoningmodel And I wanted to see how far I cango with this So each model has got itsown page with documentation and Iclicked on deploy and it literally justtook me right away into the machinelearning studio in Azure AI and I wasasked to create a workspace and as soonas I gave it all the details with thename and everything um it was able tocreate that workspace for me anddeployed that model the 54 model um fromhugging face into Azure AI machinelearning studioThat was absolutely amazing I just itjust took me a few clicks to do this Andyou can see it's creating that now Andonce it is created I can go to theworkspace And in this workspace if Iclick on endpoints if I click onendpoints I'm able to go into AzureOpenAI service And that's where my 54endpoint is And if I want to use thisendpoint I can click on continue Andthat takes me into Azure AI Foundrywhere I'll be able to um experiment withthis with this deployed model It tellsme my target URL the key that it createdfor me I'm able to see how to use thiswith my API key and um some samples ofhow to use this model Um I am also ableto go play with it in the playground andtest it out So I gave it a prompt Um andum I was really trying to go with likethe dog traveling uh to the mountainswhere he meets a robot and um that robotis helping a bird um survive in the coldum and they all become friends for lifeSo um I was just playing around withwith um a prompt but the idea here isthat you're able to go from looking at amodel in hugging phase to actually ableto deploy that model in Azure AI founduh foundry and uh Azure AI machinelearning studio Um and then I clicked ondeploy that um endpoint as a webapplicationAnd right now that is what you're seeingum with the Azure AI web applicationbeing created um as a part of thisdeployment It's able to deploy a webapplication right from that um thatmodel and endpoint that we just createdOnce the model this takes seconds maybea minute or so to to get created withthe deployment assets and stuff thingslike that And once the app is deployedI'm able to see that app in Azure AIfoundry in my web apps section There itis the 54 experiment I click on that appand there we have it An entire chatapplication built from hugging facechoosing a model 54 reasoning goes intoAzure AI machine learning studio foundryand builds it out for me as a webapplication Going back into our huggingWalkthroughface interface we're able to uh let'slook at the the the data sets tab Nowthis tab is where all your data sets areThere are thousands of these in thereand um you can preview the samples inthere You can also filter the data setsof through languages tasks libraries allof that Um and then if you click on oneyou're able to actually see the samplesof that of the data set um and startusing them Um and now the next thing isone of my favorites which is the spacessection of hugging face Now this iswhere things get really really excitingespecially for non-coders So spaces isthis interactive AI application thatanyone can use right within the huggingface browser experience And think ofthem as like readytouse AI tools Iclicked on one here which is calleddescribe anything um in uh by created byNvidia And when you go into that modelagain right in the browser I'm not doinganything else I can upload an image Anduh once I do that um I can type mydescription and I can get my descriptionfor for the regions of my images This ismy dog sitting on a chair in a park Andum I selected different parts of thatimage And um this model is able to thisdemo is able to tell me what are inthese different parts Um I selected thetree first and then I selected my doghimself and uh it was able to do areally good job at telling me what is inthis image Um and if I wanted I can takethis space and deploy it for myselfwhether locally or um or in cloud Um andbut before we do that let's look atanother example So I go back into myspaces I can really um you knowcategorize by image generation 3Dmodeling all the different options uptop Um and I went into stable diffusionwhich is another one of the very commonand very popular libraries in gener ofuh image generation models And um Itested this one out right here in spaceswith a prompt serene lake at sunset withmountains in the background and a goldenretriever watching the sunset I let itgenerate the image And there we have itUm I don't know if I like the first oneThe second one's okayUm but it it did what I wanted it to doUm and let's say I'm happy with withwhat it's I love the third and thefourth images Um they really do what Iasked it to do The the good part thebest part the part that I want to showyou is I can run this space Let's say Ilike it I can run it locally I can runit um I can clone the repo um and um andstart working with it right from herejust like how we deployed the um thefive for model in Azure AI and uh withthat um let's look at the docs the docssection is uh your knowledge center thisis where you are going to get deepdeeper technical information the docsare organized by different categorieslike the client libraries deploymentinterface core ML libraries like thetransformers which is one of the veryfamous libraries diffusers tokenizers umand a lot more like radio Um and thenthe next thing the last thing I want totalk to you about is the communitysection This is where people askquestions and share ideas and learn Theblog part of the of the community isamazing you'll see a lot of peoplecontributing to the blogs and you'll seeum what's happening right now um and andwhat's hot right now Then the learnsection is one of my favorites The LLMcourse and the agent course are some ofthe best courses out there on AI andmachine learning right now The LLMcourse goes from transformers all theway up to fine-tuning And then the agentcourse covers everything from intro toagents to to a lot more So that myfriends was hugging face and we'vetoured every major section of theplatform Whether you are just curiousabout AI want to use existing models orare developing something with AI or wantto contribute hugging face is definitelya platform to check out Now go exploreAnd if you liked this video and found ithelpful please hit that like andsubscribe button to get more tech and AIcontent And drop a comment if you havequestions and which AI platform I shouldcover next And thank you for watching

Read more »

Uncategorized

What is Synthetic Data and how to use it effectively in your AI Projects

by Priyanka Vergadia

Researchers predict we'll exhaust all fresh text data on the internet in less than 30 years. This looming "data cliff" is why synthetic data is becoming the secret sauce of AI development—our escape hatch from running out of training material.If you're working with AI systems or curious about how modern language models are trained, understanding synthetic data isn't just helpful—it's becoming essential. Let's dive into what it is, how it works, and why it might be the key to AI's future.

Read more »

Cart

Your cart is empty.

Start Shopping