Create Digital Avatars and Face Swaps with EasyPhoto on JD Cloud

Following the recent OpenAI GPT‑4o launch, many developers are eager to explore AI‑driven creative apps. One popular direction is using the EasyPhoto plugin for Stable Diffusion to create digital avatars, perform face swaps, and generate short videos.

The EasyPhoto plugin is open‑source (https://github.com/aigc-apps/sd-webui-EasyPhoto) and can be added to a Stable Diffusion WebUI instance. To keep costs low, the author installed it on JD Cloud’s AI compute platform, which already includes a pre‑configured Stable Diffusion environment.

Installation steps:

Open the Extensions tab in the WebUI.

Click Install from URL , paste the GitHub URL above, and confirm.

After a few minutes the plugin appears in the UI.

Next, go to the Settings tab, set the ControlNet unit number to 4, click Apply settings , and then Reload UI to activate EasyPhoto.

To train a personal digital avatar, select the Train sub‑tab, upload 5‑20 half‑body photos of yourself, and click Start Training . The process takes about 25 minutes and saves the model under

/data/apps/stable-diffusion-webui/outputs/easyphoto-user-id-infos

.

After training, you can use the model for various creations:

Photo Inference – choose a template or upload a celebrity image, select your avatar ID, and click Start Generation to produce a swapped‑face picture (≈5 minutes).

Video Inference – pick a video scene, select your avatar ID, and generate a short video. The default output is a GIF, but you can switch to MP4 and adjust frame count for longer clips.

Below are example results (images shown are from the author’s experiments):

For users of JD Cloud, a two‑hour coupon is available via the platform’s assistant, allowing free experimentation with the above workflow.