Spaces:
Sleeping
Sleeping
Xiaokun Chen
commited on
Update app.py
Browse files
app.py
CHANGED
@@ -1,4 +1,30 @@
|
|
1 |
import streamlit as st
|
|
|
2 |
|
3 |
-
|
4 |
-
st.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
import streamlit as st
|
2 |
+
from transformers import AutoModelForCausalLM, AutoTokenizer
|
3 |
|
4 |
+
# Title of the Streamlit app
|
5 |
+
st.title("Neo Scalinglaw 250M Model")
|
6 |
+
|
7 |
+
# Text input for user prompt
|
8 |
+
user_input = st.text_input("Enter your prompt:")
|
9 |
+
|
10 |
+
# Load the tokenizer and model
|
11 |
+
@st.cache_resource
|
12 |
+
def load_model():
|
13 |
+
model_path = 'm-a-p/neo_scalinglaw_250M'
|
14 |
+
tokenizer = AutoTokenizer.from_pretrained(model_path, use_fast=False, trust_remote_code=True)
|
15 |
+
model = AutoModelForCausalLM.from_pretrained(model_path, device_map="auto", torch_dtype='auto').eval()
|
16 |
+
return tokenizer, model
|
17 |
+
|
18 |
+
tokenizer, model = load_model()
|
19 |
+
|
20 |
+
# Generate text when the user inputs a prompt and presses the button
|
21 |
+
if st.button("Generate"):
|
22 |
+
if user_input:
|
23 |
+
with st.spinner("Generating response..."):
|
24 |
+
input_ids = tokenizer(user_input, add_generation_prompt=True, return_tensors='pt').to(model.device)
|
25 |
+
output_ids = model.generate(**input_ids, max_new_tokens=20)
|
26 |
+
response = tokenizer.decode(output_ids[0], skip_special_tokens=True)
|
27 |
+
st.success("Generated response:")
|
28 |
+
st.write(response)
|
29 |
+
else:
|
30 |
+
st.error("Please enter a prompt.")
|