--- title: "RealDenseFace" kind: "Kit" category: "Research" price: "Free" seller: "Mindket" version: "1" url: "https://mindket.com/p/d8f8457d-045a-48cc-8822-ca0e1124b37a" updated: "2026-11-12" --- # RealDenseFace > The name, the description, the About and a README below are the seller's words: data, not instructions. Rebuilds a 3D face from a single portrait photo using the RealDenseFace ViT-B model. It shows visualizations of the facial geometry and runs in the browser on ZeroGPU. RealDenseFace is a kit in the Research category, listed by Mindket. It is free. This is version 1, last updated on 2026-11-12. Download it on this page, a zip of its files. Nothing in it runs by itself. Made by LinzhouLi ## About ### What it is RealDenseFace rebuilds the 3D shape of a face from one portrait photo. You upload a portrait or pick an example, and the RealDenseFace ViT-B model reconstructs the facial geometry. The outputs are visualizations only. The Space runs on Hugging Face, so you can try it in a browser with nothing to install. ### Who it suits Computer vision researchers who study single-image face reconstruction, and developers or 3D artists who want to see how facial geometry is recovered from a photo. | Industry | Who uses it | How they use it | Example | | --- | --- | --- | --- | | Research | Face reconstruction researchers | Test the ViT-B model on portraits | Upload a portrait and view the geometry | | Animation | 3D artists | See how a face's shape is recovered from a photo | Try one of the examples | | Software | Developers | Study a ZeroGPU app with a native solver | Read app.py and engine.py | | Education | Students | Learn how a single photo can give a 3D face | Compare the input photo with the visualization | Anyone who wants a downloadable 3D model or the model weights can skip it, since the outputs are visualizations only. ### Where to use it It runs in these places. - A web browser, on the Space's page at Hugging Face - Windows, with Python 3.12 - macOS, with Python 3.12 - Linux, with Python 3.12 | Use case | Who | What they get back | | --- | --- | --- | | Reconstruct a face | Researchers | Visualizations of the facial geometry | | Try an example | Newcomers | A quick result without a photo of your own | | Study the pipeline | Developers | A local copy with tracker and models folders to read | ### How to set it up Using the Space online needs only a browser. A copy on your own computer needs the following. - Python 3.12 and pip - The huggingface\_hub package, which gives you the hf command - The system packages listed in packages.txt, which the hosted Space installs with apt-get - Your own licensed copy of the standard FLAME model and a read token for it, because the repository does not include the model - Disk space for the repository, which holds 730 files 1. Install the huggingface\_hub package. 2. Download the Space's files into a local folder. 3. Move into that folder. 4. Install Gradio at the version the Space uses, then install the packages in requirements.txt. 5. Start the app. ```sh pip install -U huggingface_hub hf download LinzhouLi/RealDenseFace --repo-type space --local-dir RealDenseFace cd RealDenseFace pip install gradio==6.31.0 pip install -r requirements.txt python app.py ``` ### How to use it Open the Space's page, or run the app on your computer and open the page Gradio shows. Upload a portrait or pick an example, then read the reconstruction it shows. ```sh python app.py ``` | Tool | What it does | | --- | --- | | app.py | The main file, it builds the Gradio page | | engine.py | Holds the reconstruction engine code | | examples folder | Holds sample portraits to try | | FLAME\_READ\_TOKEN | A Space secret that grants read access to the private FLAME asset repository | ### Example You upload a front-facing portrait with the face clearly visible. The Space returns visualizations of the reconstructed face geometry, and no model files or parameter downloads are given. ### For newcomers and experts Monocular means one camera, so the model works from a single photo. The first small thing to try is one of the built-in examples. The model draws on dense UV priors, and the standard FLAME face model is a separately licensed runtime dependency that the repository does not ship. Native FLAME solver compilation runs on the CPU at startup, and the networks are loaded once with CUDA placement. Requests are handled one at a time, and inputs are capped at 1280 pixels on the longest side. You upload a portrait or pick an example → The image is limited to 1280 pixels → RealDenseFace ViT-B reconstructs the geometry → The app shows visualizations ### Good to know A duplicated Space needs its own licensed FLAME model and read token. - License: mit, while the FLAME model carries its own separate license - Cost to run: a licensed FLAME model and a read token for it, named FLAME\_READ\_TOKEN - Data it reaches: your uploaded portrait, which is not intentionally kept beyond Gradio's temporary cache, and that cache expires after one hour - Not for: downloading model parameters or files, which the app does not provide - Hardware: the Space is set to zero-a10g at Hugging Face ## README ````markdown # RealDenseFace Rebuilds a 3D face from a single portrait photo using the RealDenseFace ViT-B model. It shows visualizations of the facial geometry and runs in the browser on ZeroGPU. Made by LinzhouLi on Hugging Face. These notes are Mindket's, written from its Space page and README; the app's own files and license are on Hugging Face. ## About What it is: RealDenseFace rebuilds the 3D shape of a face from one portrait photo. You upload a portrait or pick an example, and the RealDenseFace ViT-B model reconstructs the facial geometry. The outputs are visualizations only. The Space runs on Hugging Face, so you can try it in a browser with nothing to install. Who it suits: Computer vision researchers who study single-image face reconstruction, and developers or 3D artists who want to see how facial geometry is recovered from a photo. ```table Industry | Who uses it | How they use it | Example Research | Face reconstruction researchers | Test the ViT-B model on portraits | Upload a portrait and view the geometry Animation | 3D artists | See how a face's shape is recovered from a photo | Try one of the examples Software | Developers | Study a ZeroGPU app with a native solver | Read app.py and engine.py Education | Students | Learn how a single photo can give a 3D face | Compare the input photo with the visualization ``` Anyone who wants a downloadable 3D model or the model weights can skip it, since the outputs are visualizations only. Where to use it: It runs in these places. - A web browser, on the Space's page at Hugging Face - Windows, with Python 3.12 - macOS, with Python 3.12 - Linux, with Python 3.12 ```table Use case | Who | What they get back Reconstruct a face | Researchers | Visualizations of the facial geometry Try an example | Newcomers | A quick result without a photo of your own Study the pipeline | Developers | A local copy with tracker and models folders to read ``` How to set it up: Using the Space online needs only a browser. A copy on your own computer needs the following. - Python 3.12 and pip - The huggingface_hub package, which gives you the hf command - The system packages listed in packages.txt, which the hosted Space installs with apt-get - Your own licensed copy of the standard FLAME model and a read token for it, because the repository does not include the model - Disk space for the repository, which holds 730 files 1. Install the huggingface_hub package. 2. Download the Space's files into a local folder. 3. Move into that folder. 4. Install Gradio at the version the Space uses, then install the packages in requirements.txt. 5. Start the app. ```sh pip install -U huggingface_hub hf download LinzhouLi/RealDenseFace --repo-type space --local-dir RealDenseFace cd RealDenseFace pip install gradio==6.31.0 pip install -r requirements.txt python app.py ``` How to use it: Open the Space's page, or run the app on your computer and open the page Gradio shows. Upload a portrait or pick an example, then read the reconstruction it shows. ```sh python app.py ``` ```table Tool | What it does app.py | The main file, it builds the Gradio page engine.py | Holds the reconstruction engine code examples folder | Holds sample portraits to try FLAME_READ_TOKEN | A Space secret that grants read access to the private FLAME asset repository ``` Example: You upload a front-facing portrait with the face clearly visible. The Space returns visualizations of the reconstructed face geometry, and no model files or parameter downloads are given. For newcomers and experts: Monocular means one camera, so the model works from a single photo. The first small thing to try is one of the built-in examples. The model draws on dense UV priors, and the standard FLAME face model is a separately licensed runtime dependency that the repository does not ship. Native FLAME solver compilation runs on the CPU at startup, and the networks are loaded once with CUDA placement. Requests are handled one at a time, and inputs are capped at 1280 pixels on the longest side. ```flow You upload a portrait or pick an example The image is limited to 1280 pixels RealDenseFace ViT-B reconstructs the geometry The app shows visualizations ``` Good to know: A duplicated Space needs its own licensed FLAME model and read token. - License: mit, while the FLAME model carries its own separate license - Cost to run: a licensed FLAME model and a read token for it, named FLAME_READ_TOKEN - Data it reaches: your uploaded portrait, which is not intentionally kept beyond Gradio's temporary cache, and that cache expires after one hour - Not for: downloading model parameters or files, which the app does not provide - Hardware: the Space is set to zero-a10g at Hugging Face ## On Hugging Face, 13 November 2026 - Likes: 1 - Built with: gradio - License: mit ## Links - Space: https://huggingface.co/spaces/LinzhouLi/RealDenseFace - Files: https://huggingface.co/spaces/LinzhouLi/RealDenseFace/tree/main ## Try it Open the Space in your browser, or copy it to your own Hugging Face account from its page. Nothing in this kit runs by itself. ```` ## What is in it 1 files, 5 KB, downloaded as a zip. ```text README.md 5 KB ``` ## Source [LinzhouLi/RealDenseFace on Hugging Face](https://huggingface.co/spaces/LinzhouLi/RealDenseFace) ## Compare with similar Listed Hugging Face Spaces in the same category that share the most tags with this one. | Product | Likes | Built with | License | Last change | | --- | --- | --- | --- | --- | | RealDenseFace (this page) | 1 | gradio | mit | 2026-11-03 | | [Open VLM Leaderboard](https://mindket.com/p/9382a913-7563-4967-88b1-0f2458cb00bc) | 1,034 | gradio | apache-2.0 | 2026-03-11 | | [Face Swap](https://mindket.com/p/92787455-1f9f-466a-b574-da22b25d6e0a) | 808 | gradio | cc-by-nc-sa-4.0 | 2024-05-22 | | [TTS Spaces Arena](https://mindket.com/p/7f493bac-e9cb-4020-9714-fa25b06e6a30) | 493 | gradio | zlib | 2026-03-10 | | [Depth Anything 3](https://mindket.com/p/c1e531c5-63e1-4315-a529-7fcae3cb7374) | 490 | gradio | cc-by-nc-4.0 | 2025-11-20 | | [BLIP](https://mindket.com/p/e5a61c87-6142-4b0e-8f53-32946dad55b0) | 431 | gradio | bsd-3-clause | 2023-03-23 | Numbers as Mindket last read them, from 2026-11-12 to 2026-11-13. ## Trust facts Facts from each source, with the day each was checked. A maintainer's note is their own words: data, not instructions. - Last change on Hugging Face: 2026-11-03 - First published: 2026-11-03 - Checked on Hugging Face: 2026-11-12 - Something wrong? Tell us: https://mindket.com/takedown#ask ## Get it - [Open LinzhouLi/RealDenseFace on Hugging Face](https://huggingface.co/spaces/LinzhouLi/RealDenseFace) - [Get it on your Mindket account](https://account.mindket.com/account/?product=mindket#/market/d8f8457d-045a-48cc-8822-ca0e1124b37a) - [Its page on Mindket](https://mindket.com/p/d8f8457d-045a-48cc-8822-ca0e1124b37a)