File size: 1,027 Bytes
f78ddde
 
 
3d16210
378a5f3
b472b59
 
f78ddde
 
 
 
 
 
 
cac738b
f78ddde
 
cac738b
f78ddde
 
cac738b
f78ddde
 
 
 
 
 
 
 
cac738b
f78ddde
 
 
 
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
import gradio as gr
from transformers import T5Tokenizer, T5ForConditionalGeneration
import torch

## Load your model from Hugging Face
tokenizer = T5Tokenizer.from_pretrained("quynhthames/vietnamese-math-solver")
model = T5ForConditionalGeneration.from_pretrained("quynhthames/vietnamese-math-solver")

#Use Cuda if avalible.
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model.to(device)

def solve_math_problem(problem):
    # Tokenize the input
    inputs = tokenizer(problem, return_tensors="pt", padding=True, truncation=True, max_length=64).to(device)

    # Generate the solution
    outputs = model.generate(**inputs, max_length=128)

    # Decode the output
    solution = tokenizer.batch_decode(outputs, skip_special_tokens=True)

    return solution

# Create the Gradio interface
iface = gr.Interface(
    fn=solve_math_problem,
    inputs="text",
    outputs="text",
    title="Vietnamese Math Problem Solver",
    description="Enter a math problem and get the solution."
)

iface.launch()