whitney0507/unet-model

UNet Model for RemoveWeed Project Image Segmentation

A UNet Acrhitecture model for identifying Dock Weed in the paddy field.

Model Details

• Architecture: UNet
• Framework: PyTorch
• Input Shape: Variable, typically (C, H, W)
• Output: Segmentation map of same spatial dimensions as input

Intended Use

The model is suitable for:

• Dock Weed Detection and Classification

How to Use

pip install python, huggingface_hub and pytorch from huggingface_hub import hf_hub_download import torch

Download model weights

from huggingface_hub import hf_hub_download import torch model_path = hf_hub_download(repo_id="whitney0507/unet-model", filename="UNet_Model.pth")

Load model

model = torch.load(model_path) model.eval()

Input Image:

Output Image:

Training Details

• Dataset: (https://www.kaggle.com/datasets/whitneylimwanyee/dock-weed-instance-segementation-dataset)
• Epochs: 50
• Loss Function: Cross Entropy
• Optimizer: Adam
• Learning Rate: 1e-4