|
--- |
|
license: apache-2.0 |
|
tags: |
|
- generated_from_trainer |
|
datasets: |
|
- martinezomg/diabetic-retinopathy |
|
metrics: |
|
- accuracy |
|
pipeline_tag: image-classification |
|
base_model: google/vit-base-patch16-224-in21k |
|
model-index: |
|
- name: diabetic-retinopathy-224-procnorm-vit |
|
results: [] |
|
--- |
|
|
|
<!-- This model card has been generated automatically according to the information the Trainer had access to. You |
|
should probably proofread and complete it, then remove this comment. --> |
|
|
|
# diabetic-retinopathy-224-procnorm-vit |
|
|
|
This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on the [diabetic retinopathy](https://huggingface.co/datasets/martinezomg/diabetic-retinopathy) dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.7578 |
|
- Accuracy: 0.7431 |
|
|
|
## Model description |
|
|
|
More information needed |
|
|
|
## Intended uses & limitations |
|
|
|
More information needed |
|
|
|
## Training and evaluation data |
|
|
|
More information needed |
|
|
|
## Training procedure |
|
|
|
### Training hyperparameters |
|
|
|
The following hyperparameters were used during training: |
|
- learning_rate: 4e-05 |
|
- train_batch_size: 16 |
|
- eval_batch_size: 16 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 4 |
|
- total_train_batch_size: 64 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- lr_scheduler_warmup_ratio: 0.1 |
|
- num_epochs: 5 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Accuracy | |
|
|:-------------:|:-----:|:----:|:---------------:|:--------:| |
|
| 0.8619 | 1.0 | 50 | 0.8907 | 0.7143 | |
|
| 0.7831 | 2.0 | 100 | 0.7858 | 0.7393 | |
|
| 0.6906 | 3.0 | 150 | 0.7412 | 0.7531 | |
|
| 0.5934 | 4.0 | 200 | 0.7528 | 0.7393 | |
|
| 0.5276 | 5.0 | 250 | 0.7578 | 0.7431 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.28.1 |
|
- Pytorch 2.0.0 |
|
- Datasets 2.12.0 |
|
- Tokenizers 0.13.3 |
