app.py

import datasets
import torch
import numpy as np
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(input_image):
  features = extractor(input_image, return_tensors='pt')
  logits = model(features["pixel_values"])[-1]
  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

import gradio as gr

interface = gr.Interface(fn=classify, inputs='image', outputs='label', title='Leaf classification on beans dataset',
                         description='Sample fine-tuning a ViT for bean dataset classification')

interface.launch()