vit-base-patch16-224-in21k-Brain_Tumors_Image_Classification

This model is a fine-tuned version of google/vit-base-patch16-224-in21k.

It achieves the following results on the evaluation set:

• Loss: 0.8584
• Accuracy: 0.8198
• Weighted f1: 0.7987
• Micro f1: 0.8198
• Macro f1: 0.8054
• Weighted recall: 0.8198
• Micro recall: 0.8198
• Macro recall: 0.8149
• Weighted precision: 0.8615
• Micro precision: 0.8198
• Macro precision: 0.8769

Model Description

Click here for the code that I used to create this model.

This project is part of a comparison of seventeen (17) transformers.

Click here to see the README markdown file for the full project.

Intended Uses & Limitations

This model is intended to demonstrate my ability to solve a complex problem using technology.

Training & Evaluation Data

Brain Tumor Image Classification Dataset

Sample Images

Class Distribution of Training Dataset

Class Distribution of Evaluation Dataset

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0002
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Weighted f1	Micro f1	Macro f1	Weighted recall	Micro recall	Macro recall	Weighted precision	Micro precision	Macro precision
1.3668	1.0	180	1.0736	0.6853	0.6524	0.6853	0.6428	0.6853	0.6853	0.6530	0.7637	0.6853	0.7866
1.3668	2.0	360	1.0249	0.7792	0.7335	0.7792	0.7411	0.7792	0.7792	0.7758	0.8391	0.7792	0.8528
0.1864	3.0	540	0.8584	0.8198	0.7987	0.8198	0.8054	0.8198	0.8198	0.8149	0.8615	0.8198	0.8769

Framework versions

• Transformers 4.28.1
• Pytorch 2.0.0
• Datasets 2.11.0
• Tokenizers 0.13.3