Image-to-Video
File size: 803 Bytes
dc6a967
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
{
  "model_name": "Wan360",
  "base_model": "Wan-AI/Wan2.2-TI2V-5B",
  "task": "image-to-video",
  "description": "Camera-controlled panoramic video generation from a single panoramic image.",
  "implementation": {
    "repository": "https://github.com/AlayaLab/MUGEN",
    "entrypoint": "model/inference.py",
    "pipeline": "diffsynth.pipelines.wan_video.WanVideoPipeline"
  },
  "weights": {
    "dit": "Wan360.safetensors",
    "vae": "Wan2.2_VAE.pth",
    "text_encoder": "models_t5_umt5-xxl-enc-bf16.pth",
    "tokenizer": "google/umt5-xxl"
  },
  "inference_defaults": {
    "torch_dtype": "bfloat16",
    "is_360": true,
    "height": 960,
    "width": 1920,
    "num_frames": 161,
    "fps": 16,
    "num_inference_steps": 50,
    "cfg_scale": 5.0,
    "sigma_shift": 5.0,
    "seed": 42
  }
}