Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,65 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
import os
|
3 |
+
import subprocess
|
4 |
+
from modelscope import AutoModelForSequenceClassification,AutoTokenizer
|
5 |
+
model_names = ['agront-1b','plant-nucleotide-transformer','plant-dnagpt',
|
6 |
+
'plant-dnagemma','plant-dnabert','nucleotide-transformer-v2-100m','dnabert2']
|
7 |
+
|
8 |
+
def inference(seq,model,task):
|
9 |
+
if not seq:
|
10 |
+
gr.Warning("No sequence provided, use the default sequence.")
|
11 |
+
seq = placeholder
|
12 |
+
# Load model and tokenizer
|
13 |
+
model_name = f'zhangtaolab/{model}-{task}'
|
14 |
+
model = AutoModelForSequenceClassification.from_pretrained(model_name,ignore_mismatched_sizes=True)
|
15 |
+
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
16 |
+
|
17 |
+
# Inference
|
18 |
+
inputs = tokenizer(seq, return_tensors='pt', padding=True, truncation=True, max_length=1024)
|
19 |
+
outputs = model(**inputs)
|
20 |
+
result = outputs.logits.item()
|
21 |
+
return result
|
22 |
+
|
23 |
+
placeholder = 'TACTCTAATCGTATCAGCTGCACTTGCGTACAGGCTACCGGCGTCCTCAGCCACGTAAGAAAAGGCCCAATAAAGGCCCAACTACAACCAGCGGATATATATACTGGAGCCTGGCGAGATCACCCTAACCCCTCACACTCCCATCCAGCCGCCACCAGGTGCAGAGTGTT'
|
24 |
+
css = """
|
25 |
+
.gradio-container {
|
26 |
+
max-width: 900px;
|
27 |
+
margin: auto;
|
28 |
+
padding: 20px;
|
29 |
+
}
|
30 |
+
.btn-primary {
|
31 |
+
background-color: #8e44ad;
|
32 |
+
border-color: #8e44ad;
|
33 |
+
}
|
34 |
+
"""
|
35 |
+
# 创建 Gradio 接口
|
36 |
+
|
37 |
+
with gr.Blocks(css=css) as demo:
|
38 |
+
gr.HTML(
|
39 |
+
"""
|
40 |
+
<h1 style="text-align: center;">Promoter strength in protoplast predicted by plant LLMs</h1>
|
41 |
+
"""
|
42 |
+
)
|
43 |
+
with gr.Row():
|
44 |
+
with gr.Column(scale=1):
|
45 |
+
drop1 = gr.Dropdown(choices=['promoter_strength_protoplast'],
|
46 |
+
label="Selected Task",
|
47 |
+
interactive=False,
|
48 |
+
value="promoter_strength_protoplast")
|
49 |
+
with gr.Column(scale=1):
|
50 |
+
drop2 = gr.Dropdown(choices=model_names,
|
51 |
+
label="Select Model",
|
52 |
+
interactive=True,
|
53 |
+
value=model_names[0])
|
54 |
+
with gr.Row():
|
55 |
+
with gr.Column(scale=1):
|
56 |
+
input_box = gr.Textbox(placeholder=placeholder, label="Enter promoter Sequence", lines=4)
|
57 |
+
with gr.Column(scale=1):
|
58 |
+
output_box = gr.Textbox(label="Output", lines=4)
|
59 |
+
with gr.Row():
|
60 |
+
submit_button = gr.Button("Submit", variant="primary")
|
61 |
+
clear_button = gr.Button("Clear")
|
62 |
+
submit_button.click(inference, inputs=[input_box,drop2,drop1], outputs=output_box)
|
63 |
+
clear_button.click(lambda: ("", ""), inputs=None, outputs=[input_box, output_box])
|
64 |
+
# 启动 Gradio 接口
|
65 |
+
demo.launch()
|