Mahmoud Amiri
commited on
Commit
·
dc732ec
1
Parent(s):
6896f48
Fix inputs, add OAuth token
Browse files
app.py
CHANGED
@@ -1,60 +1,83 @@
|
|
1 |
import gradio as gr
|
2 |
-
from huggingface_hub import InferenceClient
|
3 |
-
|
4 |
-
MODEL_ID = "Bocklitz-Lab/lit2vec-tldr-bart-model"
|
5 |
-
|
6 |
|
|
|
7 |
def respond(
|
8 |
message: str,
|
9 |
-
history:
|
10 |
system_message: str,
|
11 |
max_tokens: int,
|
12 |
temperature: float,
|
13 |
top_p: float,
|
14 |
-
hf_token: OAuthToken
|
15 |
):
|
16 |
-
"""
|
|
|
|
|
17 |
client = InferenceClient(
|
18 |
-
|
19 |
-
|
20 |
)
|
21 |
|
22 |
-
|
|
|
|
|
|
|
23 |
|
24 |
-
|
25 |
-
|
26 |
-
|
27 |
-
|
28 |
temperature=temperature,
|
29 |
top_p=top_p,
|
30 |
-
stream=True, # change to False if you prefer
|
31 |
):
|
32 |
-
|
33 |
-
|
34 |
-
|
35 |
-
|
36 |
-
with gr.Sidebar():
|
37 |
-
login_btn = gr.LoginButton()
|
38 |
|
39 |
-
|
40 |
-
|
41 |
-
|
42 |
-
|
43 |
-
|
44 |
-
|
|
|
|
|
|
|
45 |
),
|
46 |
-
|
47 |
-
|
48 |
-
|
49 |
-
|
50 |
-
|
51 |
-
|
52 |
-
|
53 |
-
|
54 |
-
|
55 |
-
|
56 |
-
|
57 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
58 |
|
|
|
59 |
if __name__ == "__main__":
|
60 |
-
demo.launch()
|
|
|
1 |
import gradio as gr
|
2 |
+
from huggingface_hub import InferenceClient
|
3 |
+
from typing import List, Dict
|
|
|
|
|
4 |
|
5 |
+
# Response function for the chatbot
|
6 |
def respond(
|
7 |
message: str,
|
8 |
+
history: List[Dict[str, str]],
|
9 |
system_message: str,
|
10 |
max_tokens: int,
|
11 |
temperature: float,
|
12 |
top_p: float,
|
13 |
+
hf_token: gr.OAuthToken,
|
14 |
):
|
15 |
+
"""
|
16 |
+
Sends a chat message to the Hugging Face Inference API using the provided token and parameters.
|
17 |
+
"""
|
18 |
client = InferenceClient(
|
19 |
+
token=hf_token.token,
|
20 |
+
model="Bocklitz-Lab/lit2vec-tldr-bart-model"
|
21 |
)
|
22 |
|
23 |
+
messages = [{"role": "system", "content": system_message}] + history
|
24 |
+
messages.append({"role": "user", "content": message})
|
25 |
+
|
26 |
+
response = ""
|
27 |
|
28 |
+
for message_chunk in client.chat_completion(
|
29 |
+
messages,
|
30 |
+
max_tokens=max_tokens,
|
31 |
+
stream=True,
|
32 |
temperature=temperature,
|
33 |
top_p=top_p,
|
|
|
34 |
):
|
35 |
+
if message_chunk.choices and message_chunk.choices[0].delta.content:
|
36 |
+
token = message_chunk.choices[0].delta.content
|
37 |
+
response += token
|
38 |
+
yield response
|
|
|
|
|
39 |
|
40 |
+
# Define the Gradio interface
|
41 |
+
chatbot = gr.ChatInterface(
|
42 |
+
fn=respond,
|
43 |
+
type="messages",
|
44 |
+
additional_inputs=[
|
45 |
+
gr.Textbox(
|
46 |
+
value="You are a friendly chatbot.",
|
47 |
+
label="System message",
|
48 |
+
lines=1
|
49 |
),
|
50 |
+
gr.Slider(
|
51 |
+
minimum=1,
|
52 |
+
maximum=2048,
|
53 |
+
value=512,
|
54 |
+
step=1,
|
55 |
+
label="Max new tokens"
|
56 |
+
),
|
57 |
+
gr.Slider(
|
58 |
+
minimum=0.1,
|
59 |
+
maximum=4.0,
|
60 |
+
value=0.7,
|
61 |
+
step=0.1,
|
62 |
+
label="Temperature"
|
63 |
+
),
|
64 |
+
gr.Slider(
|
65 |
+
minimum=0.1,
|
66 |
+
maximum=1.0,
|
67 |
+
value=0.95,
|
68 |
+
step=0.05,
|
69 |
+
label="Top-p (nucleus sampling)"
|
70 |
+
),
|
71 |
+
],
|
72 |
+
)
|
73 |
+
|
74 |
+
# Set up the full Gradio Blocks layout with login
|
75 |
+
with gr.Blocks() as demo:
|
76 |
+
with gr.Row():
|
77 |
+
with gr.Column(scale=1):
|
78 |
+
gr.LoginButton()
|
79 |
+
chatbot.render()
|
80 |
|
81 |
+
# Run the app
|
82 |
if __name__ == "__main__":
|
83 |
+
demo.launch()
|