app.py

import gradio as gr
import datasets
import transformers
import torch

from transformers import AutoFeatureExtractor, AutoModelForImageClassification

dataset = datasets.load_dataset('beans')

extractor = AutoFeatureExtractor.from_pretrained("saved_model_files")
model = AutoModelForImageClassification.from_pretrained("saved_model_files")

labels = dataset['train'].features['labels'].names

def classify(im):
  features = feature_extractor(im, return_tensors='pt')
  with torch.no_grad():
    logits = model(**features).logits
  probability = torch.nn.functional.softmax(logits, dim=-1)
  probs = probability[0].detach().numpy()
  confidences = {label: float(probs[i]) for i, label in enumerate(labels)} 
  return confidences


gr.Interface(fn = classify,
             inputs = "image",
             outputs = "label",
             examples='/examples',
             title='Leaf classification on beans dataset',
             description='Fine-tuning a ViT for bean plant health classification'
                         ).launch(debug=True)