File size: 2,678 Bytes
d35cabc
 
71feab2
d35cabc
74cf048
 
d35cabc
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2c99d88
 
74cf048
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
d35cabc
 
 
 
74cf048
 
 
d91fe5f
d35cabc
 
d91fe5f
d35cabc
 
 
 
 
 
 
2c99d88
d35cabc
 
 
 
 
71feab2
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
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
import torch 
import re 
import gradio as gr
from transformers import AutoTokenizer, ViTFeatureExtractor, VisionEncoderDecoderModel 
from transformers import AutoProcessor, AutoTokenizer, BlipForConditionalGeneration
from huggingface_hub import hf_hub_download

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('<|endoftext|>','').split('\n')[0]
  caption_ids = model.generate(image, max_length = max_length)[0]
  caption_text = clean_text(tokenizer.decode(caption_ids))
  #caption_text2 = generate_captions(image)
  return caption_text

blip_processor_large = AutoProcessor.from_pretrained("Salesforce/blip-image-captioning-large")
blip_model_large = BlipForConditionalGeneration.from_pretrained("Salesforce/blip-image-captioning-large")

blip_model_large.to(device)


def generate_caption(processor, model, image, tokenizer=None, use_float_16=False):
    inputs = processor(images=image, return_tensors="pt").to(device)

    if use_float_16:
        inputs = inputs.to(torch.float16)
    
    generated_ids = model.generate(pixel_values=inputs.pixel_values, max_length=50)

    if tokenizer is not None:
        generated_caption = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
    else:
        generated_caption = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]
   
    return generated_caption



def generate_captions(image):


    caption_blip_large = generate_caption(blip_processor_large, blip_model_large, image)


    return caption_blip_large



input = gr.inputs.Image(label="Upload your Image", type = 'pil', optional=True)
#Two output boxes

output_1 = gr.outputs.Textbox(type="text",label="Caption - 1")
examples = [f"example{i}.png" for i in range(1,4)]

description= "Image caption Generator"
title = "Deep Learning and AI Intern Assignment for Listed Inc"

article = "Created By : Sravanth Kurmala"

interface = gr.Interface(
        fn=predict,
        inputs = input,
        theme="grass",
        outputs = output_1,
        examples = examples,
        title=title,
        description=description,
        article = article,
    )
interface.launch(debug=True)