metadata
library_name: UniDepth
tags:
- monocular-metric-depth-estimation
- pytorch_model_hub_mixin
- model_hub_mixin
repo_url: https://github.com/lpiccinelli-eth/UniDepth
This model has been pushed to the Hub using UniDepth:
Installation
First install the UniDepth package as follows:
!git clone -b add_hf https://github.com/NielsRogge/UniDepth.git
!cd UniDepth
!pip install -r requirements.txt
Usage
Next, one can load the model and perform inference as follows:
from unidepth.models import UniDepthV1HF
import numpy as np
from PIL import Image
model = UniDepthV1HF.from_pretrained("nielsr/unidepth-v1-convnext-large")
# Move to CUDA, if any
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model = model.to(device)
# Load the RGB image and the normalization will be taken care of by the model
rgb = torch.from_numpy(np.array(Image.open(image_path))).permute(2, 0, 1) # C, H, W
predictions = model.infer(rgb)
# Metric Depth Estimation
depth = predictions["depth"]
# Point Cloud in Camera Coordinate
xyz = predictions["points"]
# Intrinsics Prediction
intrinsics = predictions["intrinsics"]