{ "model_id": "MagicAnimate", "description": "Temporally Consistent Human Image Animation using Diffusion Model", "license": "BSD-3-Clause", "version": "1.0", "model_type": "convnext", "configurations": [ { "model": "pretrained_models/stable-diffusion-v1-5", "description": "Stable Diffusion v1.5 model with various enhancements for temporal consistency.", "config_file": "pretrained_models/stable-diffusion-v1-5/config.json" }, { "model": "pretrained_models/sd-vae-ft-mse", "description": "VAE finetuned with MSE loss for better latent representations.", "config_file": "pretrained_models/sd-vae-ft-mse/config.json" }, { "model": "pretrained_models/MagicAnimate/appearance_encoder", "description": "Appearance encoder for extracting and encoding appearance features.", "config_file": "pretrained_models/MagicAnimate/appearance_encoder/config.json" }, { "model": "pretrained_models/MagicAnimate/densepose_controlnet", "description": "DensePose control net for controlling animations based on dense pose inputs.", "config_file": "pretrained_models/MagicAnimate/densepose_controlnet/config.json" }, { "model": "pretrained_models/MagicAnimate/temporal_attention", "description": "Temporal attention model for ensuring temporal consistency in animations.", "config_file": "pretrained_models/MagicAnimate/temporal_attention/config.json" } ], "dependencies": { "python_version": "3.10", "packages": [ "torch==2.0.1", "torchvision==0.15.2", "diffusers==0.21.4", "transformers==4.32.0", "fastapi==0.103.0", "gradio==3.41.2", "aiohttp==3.8.5", "uvicorn==0.23.2", "numpy==1.24.4", "opencv-python==4.8.0.76", "scipy==1.12.0", "sentence-transformers==2.2.2" ] } }