Dpt Large

🚀 DPT-Large (MiDaS 3.0)

A Dense Prediction Transformer (DPT) model for monocular depth estimation trained on 1.4 million images.

🚀 Quick Start

The Dense Prediction Transformer (DPT) model is trained on 1.4 million images for monocular depth estimation. It uses the Vision Transformer (ViT) as the backbone and adds a neck + head for monocular depth estimation.

💻 Usage Examples

Basic Usage

The easiest way is to use the pipeline API:

from transformers import pipeline

pipe = pipeline(task="depth-estimation", model="Intel/dpt-large")
result = pipe(image)
result["depth"]

Advanced Usage

If you want to implement the entire logic yourself, here's how to do zero-shot depth estimation on an image:

from transformers import DPTImageProcessor, DPTForDepthEstimation
import torch
import numpy as np
from PIL import Image
import requests

url = "http://images.cocodataset.org/val2017/000000039769.jpg"
image = Image.open(requests.get(url, stream=True).raw)

processor = DPTImageProcessor.from_pretrained("Intel/dpt-large")
model = DPTForDepthEstimation.from_pretrained("Intel/dpt-large")

# prepare image for the model
inputs = processor(images=image, return_tensors="pt")

with torch.no_grad():
    outputs = model(**inputs)
    predicted_depth = outputs.predicted_depth

# interpolate to original size
prediction = torch.nn.functional.interpolate(
    predicted_depth.unsqueeze(1),
    size=image.size[::-1],
    mode="bicubic",
    align_corners=False,
)

# visualize the prediction
output = prediction.squeeze().cpu().numpy()
formatted = (output * 255 / np.max(output)).astype("uint8")
depth = Image.fromarray(formatted)

For more code examples, refer to the documentation.

📚 Documentation

Model Details

The model card is jointly written by the Hugging Face team and Intel.

Intended Use

Factors

Metrics

Training and Evaluation Data

Quantitative Analyses

Table 1. Comparison to the state of the art on monocular depth estimation. We evaluate zero-shot cross-dataset transfer according to the protocol defined in [30]. Relative performance is computed with respect to the original MiDaS model [30]. Lower is better for all metrics. (Ranftl et al., 2021)

Ethical Considerations

Caveats and Recommendations

Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. There are no additional caveats or recommendations for this model.

🔧 Technical Details

The DPT model was introduced in the paper Vision Transformers for Dense Prediction by Ranftl et al. (2021) and first released in this repository. It uses the Vision Transformer (ViT) as the backbone and adds a neck + head on top for monocular depth estimation.

📄 License

This model is licensed under the Apache 2.0 license.

BibTeX entry and citation info

@article{DBLP:journals/corr/abs-2103-13413,
  author    = {Ren{\'{e}} Ranftl and
               Alexey Bochkovskiy and
               Vladlen Koltun},
  title     = {Vision Transformers for Dense Prediction},
  journal   = {CoRR},
  volume    = {abs/2103.13413},
  year      = {2021},
  url       = {https://arxiv.org/abs/2103.13413},
  eprinttype = {arXiv},
  eprint    = {2103.13413},
  timestamp = {Wed, 07 Apr 2021 15:31:46 +0200},
  biburl    = {https://dblp.org/rec/journals/corr/abs-2103-13413.bib},
  bibsource = {dblp computer science bibliography, https://dblp.org}
}