app.py

import gradio as gr
import torch
import numpy as np
from PIL import Image
from huggingface_hub import hf_hub_download
import matplotlib.pyplot as plt

from depth_anything_v2.dpt import DepthAnythingV2

# Load model as before
DEVICE = 'cuda' if torch.cuda.is_available() else 'cpu'
model_configs = {
    'vitl': {'encoder': 'vitl', 'features': 256, 'out_channels': [256, 512, 1024, 1024]},
}
encoder = 'vitl'
model = DepthAnythingV2(**model_configs[encoder])
model_path = hf_hub_download(
    repo_id="depth-anything/Depth-Anything-V2-Large",
    filename=f"depth_anything_v2_{encoder}.pth",
    repo_type="model"
)
state_dict = torch.load(model_path, map_location="cpu")
model.load_state_dict(state_dict)
model = model.to(DEVICE).eval()

# Use a matplotlib colormap
CMAP = plt.get_cmap('Spectral_r')

def infer(image: np.ndarray):
    # 1. Run the model (BGR to RGB if needed)
    with torch.no_grad():
        depth = model.infer_image(image[:, :, ::-1])
    # 2. Grayscale map (normalize to 0..255)
    depth_norm = (depth - depth.min()) / (depth.max() - depth.min()) * 255.0
    depth_norm = depth_norm.astype(np.uint8)
    gray = Image.fromarray(depth_norm)
    # 3. Color map
    colored = (CMAP(depth_norm)[:, :, :3] * 255).astype(np.uint8)
    color = Image.fromarray(colored)
    return gray, color

iface = gr.Interface(
    fn=infer,
    inputs=gr.Image(type="numpy", label="Input Image"),
    outputs=[
        gr.Image(label="Grayscale Depth"),
        gr.Image(label="Colored Depth"),
    ],
    title="Depth Anything V2 (Minimal, with Colored Output)",
    description="Upload an image, get depth as grayscale and colored."
)

iface.launch()