sebastiansarasti/MRIFMFinetuned

Model Card: MRI Brain Tumor Classification (EfficientNet-B1)

Model Details

• Model Name: MRIModel
• Architecture: EfficientNet-B1-based model for MRI brain tumor classification
• Dataset: Brain Tumor MRI Dataset
• Batch Size: 32
• Loss Function: CrossEntropy Loss
• Optimizer: Adam (learning rate = 1e-3)
• Transfer Learning: No (trained from scratch)

Model Architecture

This model is based on EfficientNet-B1, a lightweight yet powerful convolutional neural network, and has been adapted for MRI-based brain tumor classification.

Modifications

• Input Channel Adaptation: The first convolutional layer is modified to accept single-channel (grayscale) MRI scans.
• Classifier Head: The fully connected (FC) layer is modified to output 756 features.

Implementation

Model Definition

import torch
import torch.nn as nn
from torchvision.models import efficientnet_b1

class MRIModel(nn.Module, PyTorchModelHubMixin):
    def __init__(self):
        super(MRIModel, self).__init__()
        self.base_model = efficientnet_b1(weights=False)  # No pretrained weights
        self.base_model.features[0] = nn.Sequential(
            nn.Conv2d(1, 32, kernel_size=(3, 3), stride=(2, 2), bias=False),
            nn.BatchNorm2d(
                32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True
            ),
            nn.ReLU6(inplace=True),
        )
        self.base_model.classifier = nn.Linear(1280, 756)

    def forward(self, x):
        x = self.base_model(x)
        return x

This model has been pushed to the Hub using the PytorchModelHubMixin integration: