File size: 2,036 Bytes
ebf036f 1a2089d ebf036f 1a2089d ebf036f 1a2089d ebf036f 1a2089d ebf036f 59e5fd8 ebf036f 1a2089d ebf036f 1a2089d 59e5fd8 ebf036f |
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 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 |
import gradio as gr
from transformers import AutoTokenizer, AutoModelForCausalLM
tokenizer = AutoTokenizer.from_pretrained("vinai/phobert-base")
# Define your models
models = {
"Lục Bát": AutoModelForCausalLM.from_pretrained(
"Libosa2707/vietnamese-poem-luc-bat-gpt2"
),
"Bảy Chữ": AutoModelForCausalLM.from_pretrained(
"Libosa2707/vietnamese-poem-bay-chu-gpt2"
),
"Tám Chữ": AutoModelForCausalLM.from_pretrained(
"Libosa2707/vietnamese-poem-tam-chu-gpt2"
),
"Năm Chữ": AutoModelForCausalLM.from_pretrained(
"Libosa2707/vietnamese-poem-nam-chu-gpt2"
),
}
def complete_poem(text, style):
# Preprocess the input text
text = text.strip()
text = text.lower()
# Choose the model based on the selected style
model = models[style]
# Tokenize the input line
input_ids = tokenizer.encode(text, return_tensors="pt")[:, :-1]
# Generate text
output = model.generate(input_ids, max_length=100, do_sample=True, temperature=0.7)
# Decode the output
generated_text = tokenizer.decode(
output[:, input_ids.shape[-1] :][0], skip_special_tokens=True
)
text = text + " " + generated_text
# Post-process the output
text = text.replace("<unk>", "\n")
pretty_text = ""
for idx, line in enumerate(text.split("\n")):
line = line.strip()
if not line:
continue
line = line[0].upper() + line[1:]
pretty_text += line + "\n"
return pretty_text
complete_poem_interface = gr.Interface(
title="Viết tiếp áng thơ hay...",
fn=complete_poem,
inputs=[
gr.components.Textbox(
lines=1,
placeholder="Tôi đâu có biết làm thơ",
label="Những áng thơ đầu tiên",
),
gr.components.Dropdown(
choices=["Lục Bát", "Bảy Chữ", "Tám Chữ", "Năm Chữ"],
label="Kiểu thơ",
value="Lục Bát",
),
],
outputs="text",
)
|