Update app.py
Browse files
app.py
CHANGED
@@ -1,63 +1,67 @@
|
|
1 |
import gradio as gr
|
2 |
from huggingface_hub import InferenceClient
|
|
|
3 |
|
4 |
-
|
5 |
-
For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
|
6 |
-
"""
|
7 |
client = InferenceClient("wop/kosmox")
|
8 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
9 |
|
10 |
def respond(
|
11 |
-
message,
|
12 |
history: list[tuple[str, str]],
|
13 |
-
system_message,
|
14 |
-
max_tokens,
|
15 |
-
temperature,
|
16 |
-
top_p,
|
17 |
):
|
18 |
-
|
19 |
-
|
20 |
-
for val in history:
|
21 |
-
if val[0]:
|
22 |
-
messages.append({"role": "user", "content": val[0]})
|
23 |
-
if val[1]:
|
24 |
-
messages.append({"role": "assistant", "content": val[1]})
|
25 |
-
|
26 |
-
messages.append({"role": "user", "content": message})
|
27 |
|
28 |
response = ""
|
29 |
|
|
|
30 |
for message in client.chat_completion(
|
31 |
-
|
32 |
max_tokens=max_tokens,
|
33 |
stream=True,
|
34 |
temperature=temperature,
|
35 |
top_p=top_p,
|
36 |
):
|
37 |
token = message.choices[0].delta.content
|
38 |
-
|
39 |
response += token
|
40 |
yield response
|
41 |
|
42 |
-
|
43 |
-
For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
|
44 |
-
"""
|
45 |
demo = gr.ChatInterface(
|
46 |
-
respond,
|
47 |
additional_inputs=[
|
48 |
-
gr.Textbox(value="You are
|
49 |
gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
|
50 |
gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
|
51 |
-
gr.Slider(
|
52 |
-
minimum=0.1,
|
53 |
-
maximum=1.0,
|
54 |
-
value=0.95,
|
55 |
-
step=0.05,
|
56 |
-
label="Top-p (nucleus sampling)",
|
57 |
-
),
|
58 |
],
|
59 |
)
|
60 |
|
61 |
-
|
62 |
if __name__ == "__main__":
|
63 |
-
demo.launch()
|
|
|
1 |
import gradio as gr
|
2 |
from huggingface_hub import InferenceClient
|
3 |
+
from unsloth.chat_templates import get_chat_template
|
4 |
|
5 |
+
# Initialize the InferenceClient with the appropriate model
|
|
|
|
|
6 |
client = InferenceClient("wop/kosmox")
|
7 |
|
8 |
+
# Define the chat template and tokenizer configuration
|
9 |
+
tokenizer = get_chat_template(
|
10 |
+
tokenizer=None, # Assuming you need to pass an actual tokenizer here
|
11 |
+
chat_template="phi-3",
|
12 |
+
mapping={"role": "from", "content": "value", "user": "human", "assistant": "gpt"},
|
13 |
+
)
|
14 |
+
|
15 |
+
def format_messages(system_message, history, user_message):
|
16 |
+
# Create a formatted string according to the specified chat template
|
17 |
+
formatted_message = "<s>\n"
|
18 |
+
if system_message:
|
19 |
+
formatted_message += f"{system_message}\n"
|
20 |
+
|
21 |
+
for user_msg, assistant_msg in history:
|
22 |
+
if user_msg:
|
23 |
+
formatted_message += f"{user_msg}\n"
|
24 |
+
if assistant_msg:
|
25 |
+
formatted_message += f"{assistant_msg}\n"
|
26 |
+
|
27 |
+
formatted_message += f"{user_message}\n"
|
28 |
+
return formatted_message
|
29 |
|
30 |
def respond(
|
31 |
+
message: str,
|
32 |
history: list[tuple[str, str]],
|
33 |
+
system_message: str,
|
34 |
+
max_tokens: int,
|
35 |
+
temperature: float,
|
36 |
+
top_p: float,
|
37 |
):
|
38 |
+
# Format the messages
|
39 |
+
formatted_message = format_messages(system_message, history, message)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
40 |
|
41 |
response = ""
|
42 |
|
43 |
+
# Stream the response from the model
|
44 |
for message in client.chat_completion(
|
45 |
+
formatted_message,
|
46 |
max_tokens=max_tokens,
|
47 |
stream=True,
|
48 |
temperature=temperature,
|
49 |
top_p=top_p,
|
50 |
):
|
51 |
token = message.choices[0].delta.content
|
|
|
52 |
response += token
|
53 |
yield response
|
54 |
|
55 |
+
# Define the Gradio interface
|
|
|
|
|
56 |
demo = gr.ChatInterface(
|
57 |
+
fn=respond,
|
58 |
additional_inputs=[
|
59 |
+
gr.Textbox(value="You are AI.", label="System message"),
|
60 |
gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
|
61 |
gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
|
62 |
+
gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-p (nucleus sampling)"),
|
|
|
|
|
|
|
|
|
|
|
|
|
63 |
],
|
64 |
)
|
65 |
|
|
|
66 |
if __name__ == "__main__":
|
67 |
+
demo.launch()
|