File size: 4,433 Bytes
7e06d6b
 
 
 
 
5478472
 
 
3ca8d26
7e06d6b
 
3ca8d26
7e06d6b
 
3ca8d26
 
3f4e721
3ca8d26
 
 
 
3f4e721
3ca8d26
7e06d6b
4da8dee
3ca8d26
 
4da8dee
3ca8d26
4da8dee
 
 
3ca8d26
8969c85
4da8dee
7e06d6b
4da8dee
3ca8d26
4603c42
3ca8d26
 
 
 
 
f4a2ddb
7e06d6b
ded78ff
5478472
 
 
 
 
fd66ee7
5478472
fd66ee7
 
 
5478472
7e06d6b
fd66ee7
 
 
c51f035
fd66ee7
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
5478472
 
 
 
fd66ee7
3ca8d26
5478472
fd66ee7
5478472
 
 
 
 
 
 
3ca8d26
 
 
7e06d6b
ded78ff
 
 
 
 
 
 
 
 
 
 
 
 
5478472
7e06d6b
5478472
 
 
 
 
3ca8d26
7b58632
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
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
import os
import gradio as gr
import torch
from PIL import Image
from pathlib import Path
import io
import sys
import traceback
from huggingface_hub import hf_hub_download

# =========================================
# 1. Define Hugging Face dataset + weights
# =========================================

HF_DATASET_REPO = "roll-ai/FloVD-weights"  # your dataset repo on HF

WEIGHT_FILES = {
    "ckpt/FVSM/FloVD_FVSM_Controlnet.pt": "FVSM/FloVD_FVSM_Controlnet.pt",
    "ckpt/OMSM/selected_blocks.safetensors": "OMSM/selected_blocks.safetensors",
    "ckpt/OMSM/pytorch_lora_weights.safetensors": "OMSM/pytorch_lora_weights.safetensors",
    "ckpt/others/depth_anything_v2_metric_hypersim_vitb.pth": "others/depth_anything_v2_metric_hypersim_vitb.pth"
}

def download_weights():
    print("πŸ”„ Downloading model weights via huggingface_hub...")
    for hf_path, local_rel_path in WEIGHT_FILES.items():
        local_path = Path("ckpt") / local_rel_path
        if not local_path.exists():
            print(f"πŸ“₯ Downloading {hf_path}")
            hf_hub_download(
                repo_id=HF_DATASET_REPO,
                repo_type="dataset",
                filename=hf_path,
                local_dir="./"
            )
        else:
            print(f"βœ… Already exists: {local_path}")

download_weights()

# =========================================
# 2. Import the FloVD generation pipeline
# =========================================

from inference.flovd_demo import generate_video

def run_inference(prompt, image, pose_type, speed, use_flow_integration, cam_pose_name):
    log_buffer = io.StringIO()
    sys_stdout = sys.stdout
    sys.stdout = log_buffer

    video_path = None
    try:
        print("πŸš€ Starting inference...")
        os.makedirs("input_images", exist_ok=True)
        image_path = "input_images/input_image.png"
        image.save(image_path)
        print(f"πŸ“Έ Saved input image to {image_path}")

        generate_video(
            prompt=prompt,
            image_path=image_path,
            fvsm_path="./ckpt/FVSM/FloVD_FVSM_Controlnet.pt",
            omsm_path="./ckpt/OMSM",
            output_path="./outputs",
            num_frames=49,
            fps=16,
            width=None,
            height=None,
            seed=42,
            guidance_scale=6.0,
            dtype=torch.float16,
            controlnet_guidance_end=0.4,
            use_dynamic_cfg=False,
            pose_type=pose_type,
            speed=float(speed),
            use_flow_integration=use_flow_integration,
            cam_pose_name=cam_pose_name,
            depth_ckpt_path="./ckpt/others/depth_anything_v2_metric_hypersim_vitb.pth"
        )

        video_name = f"{prompt[:30].strip().replace(' ', '_')}_{cam_pose_name or 'default'}.mp4"
        video_path = f"./outputs/generated_videos/{video_name}"
        print(f"βœ… Inference complete. Video saved to {video_path}")

    except Exception:
        print("πŸ”₯ Inference failed with exception:")
        traceback.print_exc()

    sys.stdout = sys_stdout
    logs = log_buffer.getvalue()
    log_buffer.close()

    return (video_path if video_path and os.path.exists(video_path) else None), logs

# =========================================
# 3. Gradio App Interface
# =========================================

with gr.Blocks() as demo:
    gr.Markdown("## πŸŽ₯ FloVD: Optical Flow + CogVideoX Video Generation")
    with gr.Row():
        with gr.Column():
            prompt = gr.Textbox(label="Prompt", value="A girl riding a bicycle through a park.")
            image = gr.Image(type="pil", label="Input Image")
            pose_type = gr.Radio(choices=["manual", "re10k"], value="manual", label="Camera Pose Type")
            cam_pose_name = gr.Textbox(label="Camera Trajectory Name", placeholder="e.g. zoom_in, tilt_up")
            speed = gr.Slider(minimum=0.1, maximum=2.0, step=0.1, value=0.5, label="Speed")
            use_flow_integration = gr.Checkbox(label="Use Flow Integration", value=False)
            submit = gr.Button("Generate Video")
        with gr.Column():
            output_video = gr.Video(label="Generated Video")
            output_logs = gr.Textbox(label="Logs", lines=20, interactive=False)

    submit.click(
        fn=run_inference,
        inputs=[prompt, image, pose_type, speed, use_flow_integration, cam_pose_name],
        outputs=[output_video, output_logs]
    )

demo.launch(show_error=True)