Tune-A-VideKO-anything

Github: Kyujinpy/Tune-A-VideKO

Model Description

• Base model: kyujinpy/KO-stable-diffusion-disney
• Training prompt: A bear is playing guitar 👁 sample-train
  

Samples

👁 sample-500
Test prompt: 토끼가 기타를 치고 있습니다, 모던한 디즈니 스타일

👁 sample-500
Test prompt: 잘생긴 왕자가 기타를 치고 있습니다, 모던한 디즈니 스타일

👁 sample-500
Test prompt: 사자가 기타를 치고 있습니다, 모던한 디즈니 스타일

Usage

Clone the github repo

git clone https://github.com/showlab/Tune-A-Video.git

Run inference code

from tuneavideo.pipelines.pipeline_tuneavideo import TuneAVideoPipeline
from tuneavideo.models.unet import UNet3DConditionModel
from tuneavideo.util import save_videos_grid
import torch

pretrained_model_path = "kyujinpy/KO-stable-diffusion-disney"
unet_model_path = "kyujinpy/Tune-A-VideKO-disney"
unet = UNet3DConditionModel.from_pretrained(unet_model_path, subfolder='unet', torch_dtype=torch.float16).to('cuda')
pipe = TuneAVideoPipeline.from_pretrained(pretrained_model_path, unet=unet, torch_dtype=torch.float16).to("cuda")
pipe.enable_xformers_memory_efficient_attention()

prompt = "사자가 기타를 치고 있습니다, 모던한 디즈니 스타일"
video = pipe(prompt, video_length=14, height=512, width=512, num_inference_steps=50, guidance_scale=7.5).videos

save_videos_grid(video, f"./{prompt}.gif")

Related Papers:

• Tune-A-Video: One-Shot Tuning of Image Diffusion Models for Text-to-Video Generation
• Stable Diffusion: High-Resolution Image Synthesis with Latent Diffusion Models