Segformer B3 1024x1024 City 160k

🚀 Segformer模型卡片

Segformer是一個用於圖像分割的模型，藉助該模型可以實現對圖像中不同對象的精準分割，在計算機視覺領域有著廣泛的應用。

🚀 快速開始

加載預訓練模型

點擊下面的按鈕在Colab中打開示例：

安裝依賴

pip install -U segmentation_models_pytorch albumentations

運行推理

import torch
import requests
import numpy as np
import albumentations as A
import segmentation_models_pytorch as smp

from PIL import Image

device = "cuda" if torch.cuda.is_available() else "cpu"

# 加載預訓練模型和預處理函數
checkpoint = "smp-hub/segformer-b3-1024x1024-city-160k"
model = smp.from_pretrained(checkpoint).eval().to(device)
preprocessing = A.Compose.from_pretrained(checkpoint)

# 加載圖像
url = "https://huggingface.co/datasets/hf-internal-testing/fixtures_ade20k/resolve/main/ADE_val_00000001.jpg"
image = Image.open(requests.get(url, stream=True).raw)

# 預處理圖像
np_image = np.array(image)
normalized_image = preprocessing(image=np_image)["image"]
input_tensor = torch.as_tensor(normalized_image)
input_tensor = input_tensor.permute(2, 0, 1).unsqueeze(0)  # HWC -> BCHW
input_tensor = input_tensor.to(device)

# 進行推理
with torch.no_grad():
    output_mask = model(input_tensor)

# 後處理掩碼
mask = torch.nn.functional.interpolate(
    output_mask, size=(image.height, image.width), mode="bilinear", align_corners=False
)
mask = mask.argmax(1).cpu().numpy()  # 在預測的類別（通道維度）上取最大值

📦 安裝指南

安裝必要的依賴庫，命令如下：

pip install -U segmentation_models_pytorch albumentations

💻 使用示例

基礎用法

import torch
import requests
import numpy as np
import albumentations as A
import segmentation_models_pytorch as smp

from PIL import Image

device = "cuda" if torch.cuda.is_available() else "cpu"

# 加載預訓練模型和預處理函數
checkpoint = "smp-hub/segformer-b3-1024x1024-city-160k"
model = smp.from_pretrained(checkpoint).eval().to(device)
preprocessing = A.Compose.from_pretrained(checkpoint)

# 加載圖像
url = "https://huggingface.co/datasets/hf-internal-testing/fixtures_ade20k/resolve/main/ADE_val_00000001.jpg"
image = Image.open(requests.get(url, stream=True).raw)

# 預處理圖像
np_image = np.array(image)
normalized_image = preprocessing(image=np_image)["image"]
input_tensor = torch.as_tensor(normalized_image)
input_tensor = input_tensor.permute(2, 0, 1).unsqueeze(0)  # HWC -> BCHW
input_tensor = input_tensor.to(device)

# 進行推理
with torch.no_grad():
    output_mask = model(input_tensor)

# 後處理掩碼
mask = torch.nn.functional.interpolate(
    output_mask, size=(image.height, image.width), mode="bilinear", align_corners=False
)
mask = mask.argmax(1).cpu().numpy()  # 在預測的類別（通道維度）上取最大值

🔧 技術細節

模型初始化參數

model_init_params = {
    "encoder_name": "mit_b3",
    "encoder_depth": 5,
    "encoder_weights": None,
    "decoder_segmentation_channels": 768,
    "in_channels": 3,
    "classes": 19,
    "activation": None,
    "aux_params": None
}

數據集

數據集名稱：Cityscapes

📚 詳細文檔

• 庫：https://github.com/qubvel/segmentation_models.pytorch
• 文檔：https://smp.readthedocs.io/en/latest/

📄 許可證

許可證信息：https://github.com/NVlabs/SegFormer/blob/master/LICENSE

本模型已使用 PytorchModelHubMixin 推送到模型中心。