库名称:transformers
许可证:mit
任务标签:深度估计
arXiv:<2502.19204>
标签:
蒸馏任意深度大模型 - Transformers版本
简介
我们推出了Distill-Any-Depth,这是一种新的SOTA单目深度估计模型,采用我们提出的知识蒸馏算法训练而成。该模型在论文《蒸馏任意深度:蒸馏创造更强的单目深度估计器》中首次提出。
此模型检查点与transformers库兼容。
在线演示。
使用方法
以下是使用该模型进行零样本深度估计的方法:
from transformers import pipeline
from PIL import Image
import requests
pipe = pipeline(task="depth-estimation", model="xingyang1/Distill-Any-Depth-Large-hf")
url = 'http://images.cocodataset.org/val2017/000000039769.jpg'
image = Image.open(requests.get(url, stream=True).raw)
depth = pipe(image)["depth"]
或者,您可以使用模型和处理器类:
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")
inputs = image_processor(images=image, return_tensors="pt")
with torch.no_grad():
outputs = model(**inputs)
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"))
如果您觉得此项目有用,请考虑引用:
@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}
}
模型卡作者
Parteek Kamboj