pslime/depth-anything-v2-large-api

1---2library_name: transformers3library: transformers4license: cc-by-nc-4.05tags:6- depth7- relative depth8pipeline_tag: depth-estimation9widget:10- inference: false11---12 13# Depth Anything V2 Base – Transformers Version14 15Depth Anything V2 is trained from 595K synthetic labeled images and 62M+ real unlabeled images, providing the most capable monocular depth estimation (MDE) model with the following features:16- more fine-grained details than Depth Anything V117- more robust than Depth Anything V1 and SD-based models (e.g., Marigold, Geowizard)18- more efficient (10x faster) and more lightweight than SD-based models19- impressive fine-tuned performance with our pre-trained models20 21This model checkpoint is compatible with the transformers library.22 23Depth Anything V2 was introduced in [the paper of the same name](https://arxiv.org/abs/2406.09414) by Lihe Yang et al. It uses the same architecture as the original Depth Anything release, but uses synthetic data and a larger capacity teacher model to achieve much finer and robust depth predictions. The original Depth Anything model was introduced in the paper [Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data](https://arxiv.org/abs/2401.10891) by Lihe Yang et al., and was first released in [this repository](https://github.com/LiheYoung/Depth-Anything).24 25[Online demo](https://huggingface.co/spaces/depth-anything/Depth-Anything-V2).26 27## Model description28 29Depth Anything V2 leverages the [DPT](https://huggingface.co/docs/transformers/model_doc/dpt) architecture with a [DINOv2](https://huggingface.co/docs/transformers/model_doc/dinov2) backbone.30 31The model is trained on ~600K synthetic labeled images and ~62 million real unlabeled images, obtaining state-of-the-art results for both relative and absolute depth estimation.32 33<img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/transformers/model_doc/depth_anything_overview.jpg"34alt="drawing" width="600"/>35 36<small> Depth Anything overview. Taken from the <a href="https://arxiv.org/abs/2401.10891">original paper</a>.</small>37 38## Intended uses & limitations39 40You can use the raw model for tasks like zero-shot depth estimation. See the [model hub](https://huggingface.co/models?search=depth-anything) to look for41other versions on a task that interests you.42 43### How to use44 45Here is how to use this model to perform zero-shot depth estimation:46 47```python48from transformers import pipeline49from PIL import Image50import requests51 52# load pipe53pipe = pipeline(task="depth-estimation", model="depth-anything/Depth-Anything-V2-Large-hf")54 55# load image56url = 'http://images.cocodataset.org/val2017/000000039769.jpg'57image = Image.open(requests.get(url, stream=True).raw)58 59# inference60depth = pipe(image)["depth"]61```62 63Alternatively, you can use the model and processor classes:64 65```python66from transformers import AutoImageProcessor, AutoModelForDepthEstimation67import torch68import numpy as np69from PIL import Image70import requests71 72url = "http://images.cocodataset.org/val2017/000000039769.jpg"73image = Image.open(requests.get(url, stream=True).raw)74 75image_processor = AutoImageProcessor.from_pretrained("depth-anything/Depth-Anything-V2-Large-hf")76model = AutoModelForDepthEstimation.from_pretrained("depth-anything/Depth-Anything-V2-Large-hf")77 78# prepare image for the model79inputs = image_processor(images=image, return_tensors="pt")80 81with torch.no_grad():82    outputs = model(**inputs)83    predicted_depth = outputs.predicted_depth84 85# interpolate to original size86prediction = torch.nn.functional.interpolate(87    predicted_depth.unsqueeze(1),88    size=image.size[::-1],89    mode="bicubic",90    align_corners=False,91)92```93 94For more code examples, please refer to the [documentation](https://huggingface.co/transformers/main/model_doc/depth_anything.html#).95 96 97### Citation98 99```bibtex100@misc{yang2024depth,101      title={Depth Anything V2}, 102      author={Lihe Yang and Bingyi Kang and Zilong Huang and Zhen Zhao and Xiaogang Xu and Jiashi Feng and Hengshuang Zhao},103      year={2024},104      eprint={2406.09414},105      archivePrefix={arXiv},106      primaryClass={id='cs.CV' full_name='Computer Vision and Pattern Recognition' is_active=True alt_name=None in_archive='cs' is_general=False description='Covers image processing, computer vision, pattern recognition, and scene understanding. Roughly includes material in ACM Subject Classes I.2.10, I.4, and I.5.'}107}108```109