File size: 1,639 Bytes
70af6b0
 
947d2f8
70af6b0
c36694a
70af6b0
 
 
947d2f8
 
 
70af6b0
 
 
6bb6d88
 
947d2f8
c36694a
6bb6d88
 
 
70af6b0
d688b4b
 
947d2f8
70af6b0
6bb6d88
70af6b0
6bb6d88
70af6b0
6bb6d88
70af6b0
6bb6d88
70af6b0
6bb6d88
 
 
 
 
 
 
 
 
 
 
c36694a
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
import torch 
import gradio as gr
from transformers import AutoTokenizer, ViTImageProcessor, VisionEncoderDecoderModel 

device = 'cpu'
encoder_checkpoint = "nlpconnect/vit-gpt2-image-captioning"
decoder_checkpoint = "nlpconnect/vit-gpt2-image-captioning"
model_checkpoint = "nlpconnect/vit-gpt2-image-captioning"

# Replace ViTFeatureExtractor with ViTImageProcessor
feature_extractor = ViTImageProcessor.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(images=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

# Remove 'optional=True' from gr.Image
input_image = gr.Image(label="Upload your Image", type='pil')
output_text = gr.Textbox(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)