Spaces:
Runtime error
Runtime error
Upload TheBroCode.py
Browse files- pages/TheBroCode.py +64 -0
pages/TheBroCode.py
ADDED
@@ -0,0 +1,64 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import streamlit as st
|
2 |
+
import textwrap
|
3 |
+
import torch
|
4 |
+
from transformers import GPT2LMHeadModel, GPT2Tokenizer
|
5 |
+
|
6 |
+
DEVICE = torch.device("cpu")
|
7 |
+
# Load GPT-2 model and tokenizer
|
8 |
+
tokenizer = GPT2Tokenizer.from_pretrained('sberbank-ai/rugpt3small_based_on_gpt2')
|
9 |
+
model_finetuned = GPT2LMHeadModel.from_pretrained(
|
10 |
+
'sberbank-ai/rugpt3small_based_on_gpt2',
|
11 |
+
output_attentions = False,
|
12 |
+
output_hidden_states = False,
|
13 |
+
)
|
14 |
+
if torch.cuda.is_available():
|
15 |
+
model_finetuned.load_state_dict(torch.load('models/brat.pt'))
|
16 |
+
else:
|
17 |
+
model_finetuned.load_state_dict(torch.load('models/brat.pt', map_location=torch.device('cpu')))
|
18 |
+
model_finetuned.eval()
|
19 |
+
|
20 |
+
# Function to generate text
|
21 |
+
def generate_text(prompt, temperature, top_p, max_length, top_k):
|
22 |
+
input_ids = tokenizer.encode(prompt, return_tensors="pt")
|
23 |
+
|
24 |
+
with torch.no_grad():
|
25 |
+
out = model_finetuned.generate(
|
26 |
+
input_ids,
|
27 |
+
do_sample=True,
|
28 |
+
num_beams=5,
|
29 |
+
temperature=temperature,
|
30 |
+
top_p=top_p,
|
31 |
+
max_length=max_length,
|
32 |
+
top_k=top_k,
|
33 |
+
no_repeat_ngram_size=3,
|
34 |
+
num_return_sequences=1,
|
35 |
+
)
|
36 |
+
|
37 |
+
generated_text = list(map(tokenizer.decode, out))
|
38 |
+
return generated_text
|
39 |
+
|
40 |
+
# Streamlit app
|
41 |
+
def main():
|
42 |
+
st.title("Генерация текста 'Кодекс Братана'")
|
43 |
+
|
44 |
+
# User inputs
|
45 |
+
prompt = st.text_area("Введите начало текста")
|
46 |
+
temperature = st.slider("Temperature", min_value=0.2, max_value=2.5, value=1.8, step=0.1)
|
47 |
+
top_p = st.slider("Top-p", min_value=0.1, max_value=1.0, value=0.9, step=0.1)
|
48 |
+
max_length = st.slider("Max Length", min_value=10, max_value=300, value=100, step=10)
|
49 |
+
top_k = st.slider("Top-k", min_value=1, max_value=500, value=500, step=10)
|
50 |
+
num_return_sequences = st.slider("Number of Sequences", min_value=1, max_value=5, value=1, step=1)
|
51 |
+
|
52 |
+
if st.button("Generate Text"):
|
53 |
+
st.subheader("Generated Text:")
|
54 |
+
for i in range(num_return_sequences):
|
55 |
+
generated_text = generate_text(prompt, temperature, top_p, max_length, top_k)
|
56 |
+
st.write(f"Generated Text {i + 1}:")
|
57 |
+
wrapped_text = textwrap.fill(generated_text[0], width=80)
|
58 |
+
st.write(wrapped_text)
|
59 |
+
st.write("------------------")
|
60 |
+
|
61 |
+
st.sidebar.image('images/theBROcode.jpeg', use_column_width=True)
|
62 |
+
|
63 |
+
if __name__ == "__main__":
|
64 |
+
main()
|