evaluate.py

from nltk.translate.bleu_score import sentence_bleu
import nltk
nltk.download('punkt')
nltk.download('punkt_tab')
from transformers import Blip2Processor, Blip2ForConditionalGeneration
from PIL import Image
import torch
import os

processor = Blip2Processor.from_pretrained("Salesforce/blip2-opt-2.7b")
model = Blip2ForConditionalGeneration.from_pretrained(
    "Salesforce/blip2-opt-2.7b", torch_dtype=torch.float16
)
device = "cuda" if torch.cuda.is_available() else "cpu"
model.to(device)

def generate_caption(image_path):
    image = Image.open(image_path).convert("RGB")
    inputs = processor(image, return_tensors="pt").to(device, torch.float16)
    generated_ids = model.generate(**inputs, max_new_tokens=50)
    caption = processor.batch_decode(generated_ids, skip_special_tokens=True)[0].strip()
    return caption

image_folder = "elc_testing_pic"
image_files = os.listdir(image_folder)

total_bleu = 0
count = 0

for image_file in image_files:
    image_path = os.path.join(image_folder, image_file)
    generated_caption = generate_caption(image_path)
    print(f"Image: {image_file}")
    print(f"Generated Caption: {generated_caption}")
    reference = input("Enter reference caption for this image: ")
    reference_tokens = [reference.lower().split()]
    generated_tokens = generated_caption.lower().split()
    bleu = sentence_bleu(reference_tokens, generated_tokens)
    print(f"BLEU Score: {bleu:.4f}")
    print("---")
    total_bleu += bleu
    count += 1

print(f"\nAverage BLEU Score: {total_bleu/count:.4f}")