app.py

import torch 
import gradio as gr
from transformers import AutoTokenizer, ViTFeatureExtractor, VisionEncoderDecoderModel 

device = 'cpu'
encoder_checkpoint = "nlpconnect/vit-gpt2-image-captioning"
decoder_checkpoint = "nlpconnect/vit-gpt2-image-captioning"
model_checkpoint = "nlpconnect/vit-gpt2-image-captioning"
feature_extractor = ViTFeatureExtractor.from_pretrained(encoder_checkpoint)
tokenizer = AutoTokenizer.from_pretrained(decoder_checkpoint)
model = VisionEncoderDecoderModel.from_pretrained(model_checkpoint).to(device)

def predict(image, max_length=64, num_beams=4):
    image = image.convert('RGB')
    image = feature_extractor(image, return_tensors="pt").pixel_values.to(device)
    clean_text = lambda x: x.replace('', '').split('\n')[0]
    caption_ids = model.generate(image, max_length=max_length, num_beams=num_beams)[0]
    caption_text = clean_text(tokenizer.decode(caption_ids, skip_special_tokens=True))
    return caption_text

input_image = gr.inputs.Image(label="Upload your Image", type='pil', optional=True)
output_text = gr.outputs.Textbox(type="text", label="Captions")

examples = [f"example{i}.jpg" for i in range(1, 7)]

description = "Image captioning application made using transformers"
title = "Image Captioning 🖼️"
article = "Created By : Shreyas Dixit"

# Create the Gradio interface
interface = gr.Interface(
    fn=predict,
    inputs=input_image,
    outputs=output_text,
    examples=examples,
    title=title,
    description=description,
    article=article,
    theme="grass"
)

# Launch the interface
interface.launch(share=True)