Spaces:
Runtime error
Runtime error
add code for image captioning
Browse files
README.md
CHANGED
@@ -1,5 +1,5 @@
|
|
1 |
---
|
2 |
-
title:
|
3 |
emoji: 💻
|
4 |
colorFrom: purple
|
5 |
colorTo: green
|
@@ -9,4 +9,6 @@ app_file: app.py
|
|
9 |
pinned: false
|
10 |
---
|
11 |
|
12 |
-
|
|
|
|
|
|
1 |
---
|
2 |
+
title: Image Captioning
|
3 |
emoji: 💻
|
4 |
colorFrom: purple
|
5 |
colorTo: green
|
|
|
9 |
pinned: false
|
10 |
---
|
11 |
|
12 |
+
Image Captioning
|
13 |
+
|
14 |
+
This space contains the code for image captioning. All you need to do is import an image and the system will generate the caption for you.
|
app.py
ADDED
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import pandas as pd
|
2 |
+
import gradio as gr
|
3 |
+
import torch
|
4 |
+
from torch.nn import functional as F
|
5 |
+
from transformers import AutoTokenizer, ViTFeatureExtractor, VisionEncoderDecoderModel
|
6 |
+
|
7 |
+
device="cpu"
|
8 |
+
feature_extractor = ViTFeatureExtractor.from_pretrained("nlpconnect/vit-gpt2-image-captioning")
|
9 |
+
cat_tokenizer = AutoTokenizer.from_pretrained("nlpconnect/vit-gpt2-image-captioning")
|
10 |
+
cap_model = VisionEncoderDecoderModel.from_pretrained("nlpconnect/vit-gpt2-image-captioning").to(device)
|
11 |
+
|
12 |
+
def predict(image, max_length=64, num_beams=4):
|
13 |
+
image = image.convert('RGB')
|
14 |
+
image = feature_extractor(image, return_tensors="pt").pixel_values.to(device)
|
15 |
+
clean_text = lambda x: x.replace('<|endoftext|>','').split('\n')[0]
|
16 |
+
caption_ids = cap_model.generate(image, max_length=max_length)[0]
|
17 |
+
caption_text = clean_text(cat_tokenizer.decode(caption_ids))
|
18 |
+
return caption_text
|
19 |
+
|
20 |
+
input = gr.components.Image(label="Upload Image", type = 'pil')
|
21 |
+
caption = gr.components.Textbox(type="text", label="Captions")
|
22 |
+
examples = [f"e{i}.jpg" for i in range(1,7)]
|
23 |
+
|
24 |
+
title = "Image Caption"
|
25 |
+
description = "Made by: Swapnil Tripathi"
|
26 |
+
|
27 |
+
interface = gr.Interface(
|
28 |
+
fn=predict,
|
29 |
+
description=description,
|
30 |
+
inputs=input,
|
31 |
+
theme=gr.themes.Default(
|
32 |
+
primary_hue=gr.themes.colors.orange,
|
33 |
+
secondary_hue=gr.themes.colors.slate
|
34 |
+
),
|
35 |
+
outputs=caption,
|
36 |
+
examples=examples,
|
37 |
+
title=title,
|
38 |
+
)
|
39 |
+
|
40 |
+
interface.launch(debug=True)
|
e1.jpg
ADDED
e2.jpg
ADDED
e3.jpg
ADDED
e4.jpg
ADDED
e5.jpg
ADDED
e6.jpg
ADDED
requirements.txt
ADDED
@@ -0,0 +1,2 @@
|
|
|
|
|
|
|
1 |
+
torch
|
2 |
+
transformers
|