import os import json import torch from PIL import Image from diffsynth.utils.data import save_video from diffsynth.pipelines.lingbot_video import LingBotVideoPipeline, ModelConfig from modelscope import dataset_snapshot_download pipe = LingBotVideoPipeline.from_pretrained( torch_dtype=torch.bfloat16, device="cuda", model_configs=[ ModelConfig(model_id="Robbyant/lingbot-video-dense-1.3b", origin_file_pattern="transformer/diffusion_pytorch_model.safetensors"), ModelConfig(model_id="Qwen/Qwen3-VL-4B-Instruct", origin_file_pattern="*.safetensors"), ModelConfig(model_id="Robbyant/lingbot-video-dense-1.3b", origin_file_pattern="vae/diffusion_pytorch_model.safetensors"), ], processor_config=ModelConfig(model_id="Qwen/Qwen3-VL-4B-Instruct", origin_file_pattern=""), ) # The condition first frame and its paired caption ship in the example dataset. dataset_snapshot_download( dataset_id="DiffSynth-Studio/diffsynth_example_dataset", local_dir="data/diffsynth_example_dataset", allow_file_pattern="lingbot_video/lingbot-video-dense-1.3b_ti2v/*", ) base = "data/diffsynth_example_dataset/lingbot_video/lingbot-video-dense-1.3b_ti2v" with open(os.path.join(base, "ti2v_example.json"), "r", encoding="utf-8") as f: caption = json.load(f) input_image = Image.open(os.path.join(base, "ti2v_first_frame.png")).convert("RGB") video = pipe( prompt=caption, negative_prompt=pipe.default_negative_prompt, input_image=input_image, height=480, width=832, num_frames=81, num_inference_steps=40, cfg_scale=3.0, seed=0, ) save_video(video, "video_lingbot-video-dense-1.3b_ti2v.mp4", fps=15, quality=10)