indonesian food_classification

This model is a fine-tuned version of torchvision.models.wide_resnet50_2 on the indonesian-food-dataset. It achieves the following results on the evaluation set:

Loss: 0.574
Accuracy: 0.8171

Model description

This model is based on the torchvision.models.wide_resnet50_2 architecture, which is a pre-trained model. The model has fine-tuned for a specific task using the Indonesian Food Dataset.

Intended uses & limitations

Intended Uses: This model is intended for image classification tasks, specifically for classifying Indonesian food items in the Indonesian Food Dataset. Limitations: Deep learning models require a large amount of data for training. They might not perform well if the dataset is small or not representative

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 0.01
train_batch_size: 50
eval_batch_size: 32
optimizer: SGD
lr_scheduler_type: MultiStepLr
num_epochs: 100

Training results

Best Accuracy Achieved: 0.8171