|
--- |
|
license: apache-2.0 |
|
tags: |
|
- generated_from_trainer |
|
datasets: |
|
- food101 |
|
metrics: |
|
- accuracy |
|
model-index: |
|
- name: vit_101 |
|
results: |
|
- task: |
|
name: Image Classification |
|
type: image-classification |
|
dataset: |
|
name: food101 |
|
type: food101 |
|
config: default |
|
split: train[:5000] |
|
args: default |
|
metrics: |
|
- name: Accuracy |
|
type: accuracy |
|
value: 0.88 |
|
--- |
|
|
|
<!-- 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. --> |
|
|
|
# vit_101 |
|
|
|
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 food101 dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 1.6267 |
|
- Accuracy: 0.88 |
|
|
|
## 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: 5e-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: 3 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Accuracy | |
|
|:-------------:|:-----:|:----:|:---------------:|:--------:| |
|
| 2.7266 | 0.99 | 62 | 2.5317 | 0.814 | |
|
| 1.8315 | 2.0 | 125 | 1.7931 | 0.864 | |
|
| 1.5845 | 2.98 | 186 | 1.6267 | 0.88 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.27.2 |
|
- Pytorch 2.1.0.dev20230428 |
|
- Datasets 2.10.1 |
|
- Tokenizers 0.13.2 |
|
|
