

























Wan 2.1 is a fast, open-source model for generating videos from text or images.
A particularly fun way to use Wan is for video style transfer. Whether you want to make dreamy Studio Ghibli-style loops, gritty cyberpunk trailers, or something entirely your own, Wan makes it easy to create stylized videos. You can start right away with a premade style, or train a custom one in minutes using your own images.
Replicate also supports fast inference with LoRAs on Wan 2.1, so you can apply styles to your videos with shorter wait times and no extra setup.
Here’s how you can get started.
You can run premade styles on top of these two text-to-video Wan 2.1 models:
To use a style, type your prompt, set fast_mode to “Fast”, and add a link to your weights in the lora_weights input. This link can be any arbitrary safetensors URL from the internet — from HuggingFace, CivitAI, or any other platform.
The lora_weights field even accepts finetuned Wan models you find on Replicate. Here’s an example with fofr/wan-14b-cyberpunk-realistic using Replicate’s Python client:
Here are a few premade styles you can try out:
On Replicate, you can also train Wan on photos that you curate to create your own stylized videos. Here’s how using ostris/wan-lora-trainer.
input_images field.trigger_word — this is what you’ll type in your prompt to activate the style.autocaption turned on. You can optionally add an autocaption_prefix or autocaption_suffix to guide the style more.In a few minutes, you’ll have your own version of Wan, trained on your style and ready to generate videos.
Whether you use a premade style or train your own, Wan 2.1 makes it simple to create stylized video that feels uniquely yours. Just pick a style, write a prompt, and hit run.
Share what you make on our Discord or tag us on X! We’re reposting our favorites.
此内容由惯性聚合(RSS阅读器)自动聚合整理,仅供阅读参考。 原文来自 — 版权归原作者所有。