Distill Any Depth Large hf
简介
核心亮点
- 通过知识蒸馏实现高性能与低延迟的平衡
- 精准的单目深度估计，支持 2D 转 3D 视觉分析
- 适配 MIT 协议，方便开发者快速集成至商业项目
- 兼容 Hugging Face 生态，部署与调用流程简单
使用方法
# 安装 Hugging Face transformers
pip install transformers torch# 使用 transformers 加载模型
from transformers import AutoModel, AutoTokenizer
model = AutoModel.from_pretrained("xingyang1/Distill-Any-Depth-Large-hf")
tokenizer = AutoTokenizer.from_pretrained("xingyang1/Distill-Any-Depth-Large-hf")Hugging Face 下载
我们推荐使用命令行或者 Hugging Face Hub SDK 来进行模型的下载。
操作指引：在下载前，请先通过如下命令安装 huggingface_hub：
pip install -U huggingface_hub命令行下载
下载完整模型库
huggingface-cli download xingyang1/Distill-Any-Depth-Large-hf下载单个文件到指定本地文件夹（以下载 config.json 到当前路径下 ./dir 目录为例）
huggingface-cli download xingyang1/Distill-Any-Depth-Large-hf config.json --local-dir ./dirSDK 下载
# 模型下载
from huggingface_hub import snapshot_download
model_dir = snapshot_download('xingyang1/Distill-Any-Depth-Large-hf')Git 下载
请确保 lfs 已经被正确安装
git lfs install
git clone https://huggingface.co/xingyang1/Distill-Any-Depth-Large-hf如果您希望跳过 lfs 大文件下载，可以使用如下命令
GIT_LFS_SKIP_SMUDGE=1 git clone https://huggingface.co/xingyang1/Distill-Any-Depth-Large-hf模型文件托管在 Hugging Face Hub，使用 HF CLI / SDK / Git 直接下载，不经过本站。
PyTorch / Transformers 使用
安装 Transformers
pip install -U transformers torch模型加载和推理
from transformers import AutoModelForCausalLM, AutoTokenizer
model = AutoModelForCausalLM.from_pretrained('xingyang1/Distill-Any-Depth-Large-hf')
tokenizer = AutoTokenizer.from_pretrained('xingyang1/Distill-Any-Depth-Large-hf')完整文档
---
library_name: transformers
license: mit
pipeline_tag: depth-estimation
arxiv: <2502.19204>
tags:
- distill-any-depth
- vision
---
Distill Any Depth Large - Transformers Version
Introduction
We present Distill-Any-Depth, a new SOTA monocular depth estimation model trained with our proposed knowledge distillation algorithms. It was introduced in the paper Distill Any Depth: Distillation Creates a Stronger Monocular Depth Estimator.This model checkpoint is compatible with the transformers library.
How to use
Here is how to use this model to perform zero-shot depth estimation:
from transformers import pipeline
from PIL import Image
import requests
load pipe
pipe = pipeline(task="depth-estimation", model="xingyang1/Distill-Any-Depth-Large-hf")
load image
url = 'http://images.cocodataset.org/val2017/000000039769.jpg'
image = Image.open(requests.get(url, stream=True).raw)
inference
depth = pipe(image)["depth"]Alternatively, you can use the model and processor classes:
from transformers import AutoImageProcessor, AutoModelForDepthEstimation
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)
image_processor = AutoImageProcessor.from_pretrained("xingyang1/Distill-Any-Depth-Large-hf")
model = AutoModelForDepthEstimation.from_pretrained("xingyang1/Distill-Any-Depth-Large-hf")
prepare image for the model
inputs = image_processor(images=image, return_tensors="pt")
with torch.no_grad():
outputs = model(**inputs)
interpolate to original size and visualize the prediction
post_processed_output = image_processor.post_process_depth_estimation(
outputs,
target_sizes=[(image.height, image.width)],
)
predicted_depth = post_processed_output[0]["predicted_depth"]
depth = (predicted_depth - predicted_depth.min()) / (predicted_depth.max() - predicted_depth.min())
depth = depth.detach().cpu().numpy() * 255
depth = Image.fromarray(depth.astype("uint8"))
)
If you find this project useful, please consider citing:
@article{he2025distill,
title = {Distill Any Depth: Distillation Creates a Stronger Monocular Depth Estimator},
author = {Xiankang He and Dongyan Guo and Hongji Li and Ruibo Li and Ying Cui and Chi Zhang},
year = {2025},
journal = {arXiv preprint arXiv: 2502.19204}
}