vGiacomov/image-classifier-beans

ResNet18 Fine-tuned on Beans Dataset

This model was trained in Google Colab using a T4 GPU and tracked with MLflow.

Model Details

Dataset: Beans

Classes:

• Healthy
• Bean Rust
• Angular Leaf Spot

Validation Accuracy: 0.9398

Training Configuration

Overfitting Prevention Techniques:

• Data augmentation (rotation, flip, crop, color jitter)
• Dropout (30%)
• L2 regularization (weight decay: 1e-4)
• Learning rate scheduling (ReduceLROnPlateau)
• Best model selection based on validation accuracy

Hyperparameters:

• Learning Rate: 5e-05
• Epochs: 10
• Batch Size: 32
• Weight Decay: 0.0001
• Dropout: 0.3
• Optimizer: Adam

Artifacts

• resnet18_beans.pth - PyTorch model weights
• per_class_metrics.csv - Detailed per-class metrics
• confusion_matrix.png - Confusion matrix visualization

Usage

Download and load the model:

from huggingface_hub import hf_hub_download
import torch
from torchvision import models
from torch import nn

model_path = hf_hub_download(
    repo_id="vGiacomov/image-classifier-beans",
    filename="resnet18_beans.pth"
)

model = models.resnet18()
model.fc = nn.Sequential(
    nn.Dropout(0.3),
    nn.Linear(model.fc.in_features, 3)
)
model.load_state_dict(torch.load(model_path, map_location="cpu"))
model.eval()

Apache 2.0