Upload test-st.py
Browse files- test-st.py +60 -0
test-st.py
ADDED
@@ -0,0 +1,60 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import streamlit as st
|
2 |
+
import torch
|
3 |
+
from transformers import AutoModelForCausalLM, AutoTokenizer, GenerationConfig
|
4 |
+
|
5 |
+
|
6 |
+
model_id = "Narrativaai/BioGPT-Large-finetuned-chatdoctor"
|
7 |
+
tokenizer = AutoTokenizer.from_pretrained("microsoft/BioGPT-Large")
|
8 |
+
model = AutoModelForCausalLM.from_pretrained(model_id)
|
9 |
+
|
10 |
+
|
11 |
+
def answer_question(prompt, temperature=0.1, top_p=0.75, top_k=40, num_beams=2, **kwargs):
|
12 |
+
inputs = tokenizer(prompt, return_tensors="pt")
|
13 |
+
input_ids = inputs["input_ids"].to("cpu")
|
14 |
+
attention_mask = inputs["attention_mask"].to("cpu")
|
15 |
+
generation_config = GenerationConfig(
|
16 |
+
temperature=temperature, top_p=top_p, top_k=top_k, num_beams=num_beams, **kwargs
|
17 |
+
)
|
18 |
+
with torch.no_grad():
|
19 |
+
generation_output = model.generate(
|
20 |
+
input_ids=input_ids,
|
21 |
+
attention_mask=attention_mask,
|
22 |
+
generation_config=generation_config,
|
23 |
+
return_dict_in_generate=True,
|
24 |
+
output_scores=True,
|
25 |
+
max_new_tokens=512,
|
26 |
+
eos_token_id=tokenizer.eos_token_id,
|
27 |
+
)
|
28 |
+
s = generation_output.sequences[0]
|
29 |
+
output = tokenizer.decode(s, skip_special_tokens=True)
|
30 |
+
return output.split(" Response:")[1]
|
31 |
+
|
32 |
+
|
33 |
+
st.set_page_config(page_title="Medical Chat Bot", page_icon=":ambulance:", layout="wide")
|
34 |
+
st.title("Medical Chat Bot")
|
35 |
+
st.caption("Talk your way to better health")
|
36 |
+
|
37 |
+
with open("ui/sidebar.md", "r") as sidebar_file:
|
38 |
+
sidebar_content = sidebar_file.read()
|
39 |
+
|
40 |
+
with open("ui/styles.md", "r") as styles_file:
|
41 |
+
styles_content = styles_file.read()
|
42 |
+
|
43 |
+
# Display the DDL for the selected table
|
44 |
+
st.sidebar.markdown(sidebar_content)
|
45 |
+
|
46 |
+
st.write(styles_content, unsafe_allow_html=True)
|
47 |
+
|
48 |
+
|
49 |
+
st.write("Please enter your question below:")
|
50 |
+
|
51 |
+
# get user input
|
52 |
+
user_input = st.text_input("You: ")
|
53 |
+
|
54 |
+
if user_input:
|
55 |
+
# generate response
|
56 |
+
bot_response = answer_question(f"Input: {user_input}\nResponse:")
|
57 |
+
st.write("")
|
58 |
+
st.write("Bot:", bot_response)
|
59 |
+
|
60 |
+
|