Hugging Face Guide — Full Breakdown & Transcript

What Is Hugging Face and How To Use It

0h 08m video Published Mar 23, 2025 Transcribed Jun 14, 2026 Angel Poon Angel Poon
342K views Recent velocity 51.0 views/hour View full performance history →
Beginner 4 min read For: Beginners interested in AI, content creators, and developers looking to explore open-source AI tools.
AI Trust Score 85/100
✅ Highly Legit

"Title accurately reflects content: explains Hugging Face and shows how to use it to build an app."

AI Summary

Hugging Face is an open-source AI community and global hub for over 1.5 million AI models. This tutorial explores its main sections—Spaces, Models, and Datasets—and demonstrates how to build a custom AI image editing app using a Gemini-powered space, with code cloned and customized via Cursor.

[0:00]
Introduction to Hugging Face

Hugging Face is an open-source AI community with over 1.5 million models from companies like Meta, Google, and Alibaba. It's free to try, with paid GPU usage starting at $9/month.

[1:03]
Main Sections: Spaces, Models, Datasets

The three main sections are Spaces (ready-to-use apps), Models (open-source model hub), and Datasets (data for fine-tuning).

[1:16]
Spaces Overview

Spaces are interactive apps for tasks like image generation, video generation, text generation, speech synthesis, and 3D modeling. Users can filter by type and use them with a few clicks.

[1:49]
Example Space: LBM Relighting

A space that relights objects based on a selected background. Drop an image, choose a background, and it generates a realistic composite in seconds.

[2:27]
Example Space: Sesame CSM

A space that creates two speakers and makes them say custom text. Demonstrates voice generation with a sample conversation.

[3:05]
Model Hub

The Model Hub contains over 1.5 million open-source models sorted by categories like computer vision, NLP, and audio. Users can view code, deploy, and discuss models.

[3:42]
Datasets Section

Datasets provides data for model input and fine-tuning. Users can filter by language and task, preview data, run SQL queries, and download datasets.

[4:10]
Building an AI Image Editing App

Using a Gemini-based space, users can edit images (e.g., make a girl eat french fries). The space is free to try with a Gemini API key.

[4:39]
Duplicating a Space

Users can duplicate a space repo, add their own API key, and run it locally. The code is public and can be cloned via Cursor.

[5:09]
Cloning and Running Locally

Clone the repo using Cursor, install dependencies, set up a Python virtual environment, and run the app with a Gemini API key.

[6:43]
Customizing the App

Use Cursor to modify the app's appearance (e.g., Sailor Moon colors and cute fonts). Accept changes and rerun to see the customized version.

Hugging Face is a versatile platform for exploring and deploying AI models. With Spaces, Models, and Datasets, anyone can experiment with cutting-edge AI and build custom apps using open-source code.

Mentioned in this Video

Tutorial Checklist

1 4:39 Go to the Gemini image editing space on Hugging Face and click 'Duplicate Space'.
2 5:09 Clone the repo: copy the clone command from the space's 'Files' tab.
3 5:18 Open Cursor, clone the repo by pasting the command, and select a folder.
4 5:38 Install dependencies: run 'pip install -r requirements.txt' in the terminal.
5 6:18 Get a Gemini API key from the space's link and copy it.
6 6:29 Run the app: execute the run command (e.g., 'python app.py') and open the local URL.
7 6:51 Paste the API key into the app interface to enable image generation.
8 7:12 Customize the app: use Cursor agent with a prompt (e.g., 'Sailor Moon colors and cute fonts') and accept changes.

Study Flashcards (7)

What is Hugging Face?

easy Click to reveal answer

An open-source AI community and global hub for open-source large language models.

0:31

How many open-source AI models does Hugging Face host?

easy Click to reveal answer

Over 1.5 million.

0:15

What are the three main sections of Hugging Face?

easy Click to reveal answer

Models, Datasets, and Spaces.

1:03

What is the cost of Hugging Face's paid plan with GPU access?

medium Click to reveal answer

$9 per month.

0:57

What does the 'Spaces' section of Hugging Face offer?

medium Click to reveal answer

Ready-to-use interactive apps for tasks like image generation, video generation, text generation, speech synthesis, and 3D modeling.

1:16

How can you run a Hugging Face space locally?

hard Click to reveal answer

Duplicate the space repo, clone it, install dependencies, and run the app with an API key.

4:39

What is the purpose of the 'Datasets' section?

medium Click to reveal answer

To provide data for model input and fine-tuning, with options to filter, preview, run SQL queries, and download.

3:42

💡 Key Takeaways

📊

Hugging Face as Global AI Hub

Establishes Hugging Face as the primary platform for open-source AI model distribution.

0:31
💡

Spaces for Non-Technical Users

Makes advanced AI accessible to regular users with ready-to-use apps.

1:16
🔧

Building an AI App with Gemini

Demonstrates practical application of Hugging Face spaces for custom image editing.

4:10
🔧

Cloning and Customizing Code

Shows how to leverage open-source code and Cursor to build and personalize AI apps.

5:09

[00:00] hey, guys, today's, episode's, going, to, be

[00:02] super, useful, we're, going, to, dive, into, a

[00:05] powerful, AI, platform, that, anybody

[00:08] interested, in, AI, should, know, about, or

[00:10] must, know, about, and, that's, hugging, face

[00:12] hugging, face, has, more, than, 1.5, million

[00:15] open, source, AI, models, from, all, over, the

[00:18] world, I'll, even, show, you, how, you, can

[00:20] leverage, hugging, phase, to, build, your

[00:22] very, own, AI, image, editing, app, based, on

[00:25] the, latest, Google, Gemini

[00:28] update, so, what, is, hugging, face, hugging

[00:31] face, is, an, open, source, AI, community, that

[00:33] has, become, the, global, hub, for, open-

[00:35] Source, large, language, models, whether

[00:37] it's, meta, Google, deep, seek, Alibaba

[00:40] whenever, companies, want, to, release, a, new

[00:42] AI, model, hugging, face, is, where, they

[00:44] publish, them, first, hugging, face, is, free

[00:47] to, try, and, you, can, do, a, lot, of, things

[00:49] without, paying, anything, but, once, you

[00:51] start, using, their, gpus, and, stuff, you

[00:54] will, start, having, to, pay, for, $9, a, month

[00:57] which, is, the, plan, that, I'm, subscribed, to

[01:00] the, main, sections, of, hugging, face

[01:03] includes, models, data, sets, and, spaces

[01:06] let's, start, with, spaces, which, I, think, is

[01:08] where, majority, of, regular, people, people

[01:11] that, are, not, super, technical, who, are, not

[01:13] AI, researchers, will, visit, most, often

[01:16] when, we, open, up, the, spaces, page, what

[01:17] we're, going, to, see, is, spaces, of, the, week

[01:20] which, are, the, popular, spaces, that, you

[01:22] might, want, to, check, out, and, toy, around

[01:24] underneath, it, you'll, see, the, running

[01:26] apps, trending, first, and, there, are

[01:28] hundreds, of, pages, of, the, these, on, the

[01:30] top, you, can, also, filter, by, what, type, of

[01:33] spaces, it, is, it, could, be, image

[01:35] Generation, video, generation, text

[01:37] generation, or, language, generation, speech

[01:39] synthesis, 3D, modeling, there're, all, these

[01:41] goodies, in, spaces, and, they're, all, ready

[01:44] for, you, to, use, with, just, a, few, clicks

[01:46] let's, go, check, out, one, of, the, popular

[01:49] spaces, this, week, I, think, I'm, going, to

[01:51] try, this, lbm, relighting, this, is, a, model

[01:54] where, they, relight, the, object, based, on

[01:56] the, background, that, you, selected, I'm

[01:58] going, to, drop, an, image, of, myself, here

[02:01] and, then, selecting, a, background, and, then

[02:03] click, relight, and, in, 3, seconds, you'll

[02:06] see, this, image, has, generated, me, in, a

[02:09] different, background, and, it, will, make, it

[02:11] more, realistic, although, this, image, is, a

[02:13] little, bit, spooky, but, let's, try, again

[02:15] and, now, you, can, kind, of, see, before, and

[02:18] after, it, Blends, the, first, image, into, the

[02:21] second, image, it, doesn't, have, to, be, a

[02:22] person, it, could, be, a, product, or, whatnot

[02:25] another, one, that's, been, getting, a, lot, of

[02:27] hype, this, week, is, Sesame, CSM, basically

[02:30] can, create, two, speakers, and, make, them

[02:33] say, stuff, that, you, want, them, to, say, I'm

[02:35] just, going, to, tweak, this, and, now, let's

[02:38] listen, to, it, you, can, have, at, least, call

[02:40] James, I, waited, for, 2, hours, at, that

[02:41] restaurant, I, told, you, the, meeting, ran

[02:43] late, and, my, phone, died, was, I, supposed, to

[02:47] do, find, a, charger, someone's, phone, he's, a

[02:51] landline, literally, anything, other, than

[02:53] leaving, me, sitting, there, alone, I, know

[02:56] this, isn't, perfect, this, product, but, you

[02:58] could, at, least, test, it, with, a, few, clicks

[03:00] you, can, play, around, with, the, latest, AI

[03:03] tools, based, on, the, latest, research, then

[03:05] there's, the, model, Hub, here, you'll, find

[03:08] all, types, of, models, broken, down, by

[03:11] categories, computer, vision, natural

[03:13] language, processing, audio, and, there's, so

[03:16] many, of, them, 1.5, million, plus, models, and

[03:20] it's, sorted, by, trending, by, default

[03:22] remember, these, are, all, open, source, and

[03:24] free, that, means, you, can, even, look, into

[03:26] the, code, you, can, use, it, on, your, own, one

[03:29] once, you, click, into, a, model, you, can, jump

[03:32] into, the, spaces, that, use, this, particular

[03:35] model, you, can, deploy, it, you, can, use, this

[03:37] model, there's, a, community, section, where

[03:39] you, can, have, discussions, about, this, it's

[03:42] super, cool, then, we, can, go, into, data, sets

[03:45] so, what, is, data, sets, data, sets, is, the

[03:47] data, that, you, can, input, into, your, model

[03:49] you, can, filter, by, Language, by, the, task

[03:52] that, it, does, you, could, click, onto, one, of

[03:55] the, data, sets, and, look, inside, and, see

[03:57] what, kind, of, data, there, is, and, how, use

[04:00] it, you, can, even, run, SQL, queries, if, you

[04:02] want, through, the, data, set, and, obviously

[04:04] you, can, download, this, and, use, this, data

[04:06] set, these, data, sets, are, perfect, for, when

[04:08] you, want, to, fine-tune, your, own

[04:10] model, now, for, practice, why, don't, we, just

[04:13] take, a, look, at, what's, cool, for, example

[04:16] if, you, want, to, look, at, this, I, thought

[04:18] this, was, really, cool, I, uploaded, a

[04:20] picture, of, a, girl, and, I, want, to, make, the

[04:22] girl, eat, french, fries, instead, of, just

[04:24] holding, her, hand, out, and, we, can, see, that

[04:26] it, did, just, that, this, is, based, on, the

[04:28] latest, gem, Gemini, model, you, could, input

[04:32] your, own, Gemini, API, key, but, for, now

[04:34] they're, letting, you, try, it, for, free, this

[04:36] is, a, really, cool, app, cool, space, you, can

[04:39] duplicate, this, repo, add, your, own, secret

[04:41] and, click, duplicate, space, I've, already

[04:44] done, that, so, now, I, can, technically, input

[04:47] anything, make, her, cry, instead, of, smile

[04:49] and, it's, absolutely, adorable, say, you

[04:51] like, this, product, a, lot, and, you, want, to

[04:55] host, it, on, your, own, site, how, do, you, do

[04:57] something, like, that, the, code, is, public

[04:59] public, and, you, can, download, it, let's

[05:02] just, say, you, want, to, customize, this, all

[05:04] you, have, to, do, is, click, on, files, this, is

[05:06] the, code, you, will, want, to, download, to

[05:09] make, your, own, app, so, what, we're, going, to

[05:11] do, is, clone, repo, and, they, are, going, to

[05:14] give, you, the, instructions, over, here, just

[05:16] click, copy, and, then, we're, going, to, open

[05:18] up, cursor, hit, clone, repo, and, see, this

[05:21] pop, up, over, here, and

[05:24] paste, this, assuming, you've, already, set

[05:26] up, your, cursor, if, you, haven't, already

[05:28] you, could, watch, one, of, my, previous, vide

[05:30] on, cursor, to, set, it, up, select, a, folder

[05:33] that, you, want, to, open, it, now, that, you're

[05:35] in, here, you, can, follow, the, instructions

[05:38] and, make, sure, you, have, this, installed

[05:40] toggle, this, panel, which, you, can, also, use

[05:43] command, J, to, bring, up, this, is, just, a

[05:45] terminal, and, install, this, I, already, have

[05:48] this, I'm, going, to, use, cursor, agent, and

[05:51] ask, it, to, help, me, build, this, app, so, I

[05:54] can, run, it, locally, add, the, context

[05:56] attach, the, app, over, here, and, then, send

[05:59] it, even, if, you, don't, know, how, to, code

[06:01] cursor, is, going, to, go, step, by, step, and

[06:03] figure, out, what, they, need, to, do, so, first

[06:05] creating, a, python, virtual, environment

[06:08] Pyon, 3, I, already, have, it, install, the

[06:10] requirement, basically, you, just, let

[06:12] cursor, run, its, course, and, accept, the

[06:15] changes, you, need, to, get, a, Gemini, API, key

[06:18] to, use, with, this, app, and, to, get, this, key

[06:21] if, you, go, back, here, you, can, click, get, an

[06:24] API, key, you, can, create, an, API, key, over

[06:26] here, just, going, to, copy, it, and, follow

[06:29] the, instructions, it's, telling, you, to, run

[06:32] the, app, will, launch, in, the, local, server

[06:35] copy, and, paste, this, paste, it, as, cursor

[06:37] mentioned, this, app, will, be, running, on

[06:40] this, local, URL, click, on, it, and, follow, it

[06:43] and, now, you, can, see, that, this, is, on, my

[06:45] local, server, dropping, an, image, here, you

[06:47] can, click, generate, but, it's, not, going, to

[06:49] work, because, you, don't, have, an, API, key

[06:51] so, once, the, user, paste, and, API, key, you

[06:54] should, see, that, it, will, now, be

[06:56] processing, and, generating, based, on, the

[06:58] prompt, that, we, said, there, you, go, we, have

[07:00] a, girl, eating, a, banana, and, obviously, now

[07:03] this, part, is, required, about, inputting

[07:06] your, own, API, key, since, it's, your, own

[07:09] server, what, happens, if, you, want, cursor

[07:12] to, alter, this, and, customize, it, to, make

[07:15] it, your, own, I, just, told, cursor, I, like

[07:16] this, that, I, want, Sailor, Moon, colors, and

[07:18] cute, fonts, this, is, my, prompt, over, here

[07:21] we'll, see, what, cursor, comes, up, with, all

[07:23] you, have, to, do, is, command, enter, to

[07:25] accept, all, the, changes, now, run, the

[07:27] command, once, it's, done, click, to, follow

[07:30] the, link, you, can, see, that, the, background

[07:33] is, different, you, can, also, change, the

[07:36] title, of, this, or, whatnot, but, you, get, the

[07:38] idea, you, can, customize, your, own, site, and

[07:41] make, your, own, app, based, on, the, code, that

[07:43] people, shared, with, you, obviously, make

[07:44] sure, you, look, and, see, what, kind, of

[07:46] license, it, has, and, whether, you, need, to

[07:49] give, any, credits, but, that's, how, you, can

[07:52] download, the, code, and, tweak, it, in, cursor

[07:54] so, you, can, make, your, own, AI, app, now

[07:57] that's, it, for, today, this, is, a, quick

[07:58] tutorial, I, hope, hope, you, enjoyed, it

[08:00] leave, a, comment, below, letting, me, know, if

[08:02] you've, used, hugging, face, before, whether

[08:04] you, have, spaces, whether, I, could, check

[08:06] out, your, spaces, and, I, would, love, to, see

[08:08] what, cool, stuff, you, built, with, it, be

[08:10] sure, to, hit, this, like, subscribe, and

[08:13] notify, button, that, will, help, this, small

[08:14] Channel, out, a, lot, and, see, you, in, the

[08:16] next, video

⚡ Saved you 0h 08m reading this? Transcribe any YouTube video for free — no signup needed.