Spaces:
Build error
Build error
Update app.py
Browse files
app.py
CHANGED
|
@@ -1,118 +1,44 @@
|
|
| 1 |
-
import
|
| 2 |
-
import
|
| 3 |
-
from PIL import Image
|
| 4 |
-
from PIL import ImageDraw
|
| 5 |
import gradio as gr
|
| 6 |
-
import
|
| 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 |
-
#['27-Books.jpg',['en']],
|
| 48 |
-
#['28-Books.jpg',['en']],
|
| 49 |
-
#['27-Games.jpg',['en']],
|
| 50 |
-
#['28-Games.jpg',['en']]]
|
| 51 |
-
|
| 52 |
-
#['29-Books-Science-Fiction.jpg',['en']],
|
| 53 |
-
#['30-Manga-Books.jpg',['en']],
|
| 54 |
-
#['31-Books.jpg',['en']],
|
| 55 |
-
#['32-Books.jpg',['en']],
|
| 56 |
-
#['33-Game-Night.jpg',['en']],
|
| 57 |
-
#['34-Games.jpg',['en']]
|
| 58 |
-
#['35-Favorite-Games.jpg',['en']],
|
| 59 |
-
#['36-Game-Night.jpg',['en']],
|
| 60 |
-
#['38-Strategy-Games.jpg',['en']],
|
| 61 |
-
#['39-Games.jpg',['en']],
|
| 62 |
-
#['40-Games.jpg',['en']],
|
| 63 |
-
#['41-Game-Rules.jpg',['en']],
|
| 64 |
-
#['42-Game-Rules.jpg',['en']],
|
| 65 |
-
#['43-Games-Pieces.jpg',['en']],
|
| 66 |
-
#['44-Game-Pieces.jpg',['en']],
|
| 67 |
-
#['45-Choose-Your-Own-Adventure.jpg',['en']],
|
| 68 |
-
#['46-Choose-Your-Own-Adventure.jpg',['en']]
|
| 69 |
-
#]
|
| 70 |
-
|
| 71 |
-
# Comment
|
| 72 |
-
#['english.png',['en']],
|
| 73 |
-
#['chinese.jpg',['ch_sim', 'en']],
|
| 74 |
-
#['japanese.jpg',['ja', 'en']],
|
| 75 |
-
#['Hindi.jpeg',['hi', 'en']]
|
| 76 |
-
|
| 77 |
-
def draw_boxes(image, bounds, color='yellow', width=2):
|
| 78 |
-
draw = ImageDraw.Draw(image)
|
| 79 |
-
for bound in bounds:
|
| 80 |
-
p0, p1, p2, p3 = bound[0]
|
| 81 |
-
draw.line([*p0, *p1, *p2, *p3, *p0], fill=color, width=width)
|
| 82 |
-
return image
|
| 83 |
-
|
| 84 |
-
def inference(img, lang):
|
| 85 |
-
reader = easyocr.Reader(lang)
|
| 86 |
-
bounds = reader.readtext(img.name)
|
| 87 |
-
im = PIL.Image.open(img.name)
|
| 88 |
-
draw_boxes(im, bounds)
|
| 89 |
-
im.save('result.jpg')
|
| 90 |
-
return ['result.jpg', pd.DataFrame(bounds).iloc[: , 1:]]
|
| 91 |
-
|
| 92 |
-
title = 'Image To Optical Character Recognition'
|
| 93 |
-
description = 'Multilingual OCR which works conveniently on all devices in multiple languages.'
|
| 94 |
-
article = "<p style='text-align: center'></p>"
|
| 95 |
-
css = ".output_image, .input_image {height: 40rem !important; width: 100% !important;}"
|
| 96 |
-
examples = [['20-Books.jpg',['en']],['21-Books.jpg',['en']],['22-Magazines.jpg',['en']],['23-Magazines.jpg',['en']]]
|
| 97 |
-
|
| 98 |
-
choices = [
|
| 99 |
-
"ch_sim",
|
| 100 |
-
"ch_tra",
|
| 101 |
-
"de",
|
| 102 |
-
"en",
|
| 103 |
-
"es",
|
| 104 |
-
"ja",
|
| 105 |
-
"hi",
|
| 106 |
-
"ru"
|
| 107 |
-
]
|
| 108 |
-
gr.Interface(
|
| 109 |
-
inference,
|
| 110 |
-
[gr.inputs.Image(type='file', label='Input'),gr.inputs.CheckboxGroup(choices, type="value", default=['en'], label='language')],
|
| 111 |
-
[gr.outputs.Image(type='file', label='Output'), gr.outputs.Dataframe(headers=['text', 'confidence'])],
|
| 112 |
-
title=title,
|
| 113 |
-
description=description,
|
| 114 |
-
article=article,
|
| 115 |
-
examples=examples,
|
| 116 |
-
css=css,
|
| 117 |
-
enable_queue=True
|
| 118 |
-
).launch(debug=True)
|
|
|
|
| 1 |
+
import torch
|
| 2 |
+
import re
|
|
|
|
|
|
|
| 3 |
import gradio as gr
|
| 4 |
+
from transformers import AutoTokenizer, ViTFeatureExtractor, VisionEncoderDecoderModel
|
| 5 |
+
|
| 6 |
+
device='cpu'
|
| 7 |
+
encoder_checkpoint = "nlpconnect/vit-gpt2-image-captioning"
|
| 8 |
+
decoder_checkpoint = "nlpconnect/vit-gpt2-image-captioning"
|
| 9 |
+
model_checkpoint = "nlpconnect/vit-gpt2-image-captioning"
|
| 10 |
+
feature_extractor = ViTFeatureExtractor.from_pretrained(encoder_checkpoint)
|
| 11 |
+
tokenizer = AutoTokenizer.from_pretrained(decoder_checkpoint)
|
| 12 |
+
model = VisionEncoderDecoderModel.from_pretrained(model_checkpoint).to(device)
|
| 13 |
+
|
| 14 |
+
|
| 15 |
+
def predict(image,max_length=64, num_beams=4):
|
| 16 |
+
image = image.convert('RGB')
|
| 17 |
+
image = feature_extractor(image, return_tensors="pt").pixel_values.to(device)
|
| 18 |
+
clean_text = lambda x: x.replace('<|endoftext|>','').split('\n')[0]
|
| 19 |
+
caption_ids = model.generate(image, max_length = max_length)[0]
|
| 20 |
+
caption_text = clean_text(tokenizer.decode(caption_ids))
|
| 21 |
+
return caption_text
|
| 22 |
+
|
| 23 |
+
|
| 24 |
+
|
| 25 |
+
input = gr.inputs.Image(label="Upload your Image", type = 'pil', optional=True)
|
| 26 |
+
output = gr.outputs.Textbox(type="auto",label="Captions")
|
| 27 |
+
examples = [f"example{i}.jpg" for i in range(1,7)]
|
| 28 |
+
|
| 29 |
+
description= "Image captioning application made using transformers"
|
| 30 |
+
title = "Image Captioning 🖼️"
|
| 31 |
+
|
| 32 |
+
article = "Created By : Shreyas Dixit "
|
| 33 |
+
|
| 34 |
+
interface = gr.Interface(
|
| 35 |
+
fn=predict,
|
| 36 |
+
inputs = input,
|
| 37 |
+
theme="grass",
|
| 38 |
+
outputs=output,
|
| 39 |
+
examples = examples,
|
| 40 |
+
title=title,
|
| 41 |
+
description=description,
|
| 42 |
+
article = article,
|
| 43 |
+
)
|
| 44 |
+
interface.launch(debug=True)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|