Spaces:
Sleeping
Sleeping
Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,29 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
|
2 |
+
from openai import OpenAI
|
3 |
+
import gradio as gr
|
4 |
+
|
5 |
+
api_key = "token-abc123" # Replace with your key
|
6 |
+
base_url = "http://localhost:8000/v1"
|
7 |
+
client = OpenAI(api_key=api_key, base_url=base_url)
|
8 |
+
|
9 |
+
def predict(message, history):
|
10 |
+
history_openai_format = []
|
11 |
+
for human, assistant in history:
|
12 |
+
history_openai_format.append({"role": "user", "content": human })
|
13 |
+
history_openai_format.append({"role": "assistant", "content":assistant})
|
14 |
+
history_openai_format.append({"role": "user", "content": message})
|
15 |
+
|
16 |
+
response = client.chat.completions.create(model='Llama-3-8B-UltraMedical',
|
17 |
+
messages= history_openai_format,
|
18 |
+
temperature=1.0,
|
19 |
+
stop=["<|eot_id|>"],
|
20 |
+
stream=True)
|
21 |
+
|
22 |
+
partial_message = ""
|
23 |
+
for chunk in response:
|
24 |
+
if chunk.choices[0].delta.content is not None:
|
25 |
+
partial_message = partial_message + chunk.choices[0].delta.content
|
26 |
+
yield partial_message
|
27 |
+
|
28 |
+
gr.ChatInterface(predict).launch()
|
29 |
+
|