Segformer B4 512x512 Ade 160k

🚀 分割模型庫（Segmentation-Models-PyTorch）

分割模型庫（Segmentation-Models-PyTorch）是一個基於 PyTorch 的圖像分割模型庫，提供了預訓練的圖像分割模型，可用於語義分割等任務。

🚀 快速開始

加載預訓練模型

你可以通過以下步驟加載並運行預訓練的 Segformer 模型。點擊下面的按鈕在 Colab 中打開示例：

1. 安裝依賴

pip install -U segmentation_models_pytorch albumentations

2. 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-b4-512x512-ade-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

💻 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-b4-512x512-ade-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

🔧 模型初始化參數

model_init_params = {
    "encoder_name": "mit_b4",
    "encoder_depth": 5,
    "encoder_weights": None,
    "decoder_segmentation_channels": 768,
    "in_channels": 3,
    "classes": 150,
    "activation": None,
    "aux_params": None
}

📚 詳細文檔

數據集

數據集名稱：ADE20K

更多信息

• 庫：https://github.com/qubvel/segmentation_models.pytorch
• 文檔：https://smp.readthedocs.io/en/latest/
• 許可證：https://github.com/NVlabs/SegFormer/blob/master/LICENSE

此模型已使用 PytorchModelHubMixin 推送到模型中心。

📄 許可證

本項目使用其他許可證，詳情請見：https://github.com/NVlabs/SegFormer/blob/master/LICENSE