Spaces:
No application file
No application file
Delete app.py
Browse files
app.py
DELETED
@@ -1,39 +0,0 @@
|
|
1 |
-
import torch
|
2 |
-
import re
|
3 |
-
import gradio as gr
|
4 |
-
from gradio.components import Image, Textbox
|
5 |
-
from transformers import AutoTokenizer, ViTImageProcessor, VisionEncoderDecoderModel
|
6 |
-
|
7 |
-
device=torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
8 |
-
local_path = "/app/model/nlpconnect_vit-gpt2-image-captioning/"
|
9 |
-
feature_extractor = ViTImageProcessor.from_pretrained(local_path)
|
10 |
-
tokenizer = AutoTokenizer.from_pretrained(local_path)
|
11 |
-
model = VisionEncoderDecoderModel.from_pretrained(local_path).to(device)
|
12 |
-
|
13 |
-
gen_kwargs = {"max_length": 128, "num_beams": 8}
|
14 |
-
def predict(image):
|
15 |
-
image = image.convert('RGB')
|
16 |
-
image = feature_extractor(image, return_tensors="pt").pixel_values.to(device)
|
17 |
-
clean_text = lambda x: x.replace('<|endoftext|>','').split('\n')[0]
|
18 |
-
caption_ids = model.generate(image, **gen_kwargs)[0]
|
19 |
-
caption_text = clean_text(tokenizer.decode(caption_ids))
|
20 |
-
return caption_text
|
21 |
-
|
22 |
-
|
23 |
-
_input = Image(label="Upload any Image", type = 'pil')
|
24 |
-
_output = Textbox(type="text",label="Captions")
|
25 |
-
examples = [f"example{i}.jpg" for i in range(1,7)]
|
26 |
-
|
27 |
-
title = "Image Captioning "
|
28 |
-
description = "Made by : 炼丹侠"
|
29 |
-
iface = gr.Interface(
|
30 |
-
|
31 |
-
fn=predict,
|
32 |
-
description=description,
|
33 |
-
inputs=_input,
|
34 |
-
outputs=_output,
|
35 |
-
title=title,
|
36 |
-
examples = examples
|
37 |
-
)
|
38 |
-
gr.close_all() #关闭所有正在运行的端口
|
39 |
-
iface.launch(server_name="0.0.0.0", server_port=7860, debug=True)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|