|
|
--- |
|
|
license: mit |
|
|
datasets: |
|
|
- ramnck/beer-classification |
|
|
base_model: |
|
|
- google/vit-base-patch16-224 |
|
|
pipeline_tag: image-classification |
|
|
library_name: transformers |
|
|
--- |
|
|
|
|
|
# Model Card for Model ID |
|
|
|
|
|
Student project |
|
|
|
|
|
Beer classification model |
|
|
Finetuned from vit-base |
|
|
65 classes |
|
|
Manually collected dataset |
|
|
|
|
|
## Model Details |
|
|
|
|
|
### Model Description |
|
|
|
|
|
|
|
|
|
|
|
- **Developed by:** Karetnikov Roman Khmilevsky Alexey Koreshkov Nicolay |
|
|
- **Model type:** Vit for image classification |
|
|
- **License:** MIT |
|
|
- **Finetuned from model:** google/vit-base-patch16-224 |
|
|
|
|
|
|
|
|
## Uses |
|
|
|
|
|
Use for classification of beer from enlisted types |
|
|
|
|
|
## Bias, Risks, and Limitations |
|
|
|
|
|
May not be 100% accurate |
|
|
|
|
|
### Recommendations |
|
|
|
|
|
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. |
|
|
|
|
|
## How to Get Started with the Model |
|
|
|
|
|
Use the code below to get started with the model. |
|
|
|
|
|
```python |
|
|
|
|
|
from transformers import ViTForImageClassification, ViTImageProcessor |
|
|
|
|
|
model = ViTForImageClassification.from_pretrained("ramnck/beer-classificator") |
|
|
processor = ViTImageProcessor.from_pretrained("ramnck/beer-classificator") |
|
|
|
|
|
``` |
|
|
|
|
|
## Training Details |
|
|
|
|
|
### Training Data |
|
|
|
|
|
65 classes, around 50 photos per class, all photos was taken by our hands |
|
|
|
|
|
### Training Procedure |
|
|
|
|
|
SFT |
|
|
|
|
|
#### Preprocessing [optional] |
|
|
|
|
|
Resize to 224x224 |
|
|
|
|
|
#### Training Hyperparameters |
|
|
|
|
|
look in .ipynb |
|
|
|
|
|
#### Speeds, Sizes, Times [optional] |
|
|
|
|
|
idk |
|
|
|
|
|
## Evaluation |
|
|
|
|
|
yes |
|
|
|
|
|
### Testing Data, Factors & Metrics |
|
|
|
|
|
#### Testing Data |
|
|
|
|
|
yes |
|
|
|
|
|
#### Factors |
|
|
|
|
|
it`s bad testing data (looks very alike as train data) |
|
|
|
|
|
#### Metrics |
|
|
|
|
|
Final Test Accuracy: 0.9982 |
|
|
|
|
|
### Results |
|
|
|
|
|
it works |
|
|
|
|
|
#### Summary |
|
|
|
|
|
have a nice day |
|
|
|
|
|
## Environmental Impact |
|
|
|
|
|
i ate 20 KFC wings during train |
|
|
i hope i spoiled air as much as it was possible |
|
|
|
|
|
|
|
|
## Model Card Contact |
|
|
|
|
|
write me on my HF/GH |
|
|
|
|
|
|
