File size: 2,101 Bytes
2410db3
 
1cb6527
2410db3
1cb6527
b81c54e
2410db3
1cb6527
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2410db3
 
1cb6527
2410db3
1cb6527
 
 
 
2410db3
1cb6527
 
2410db3
 
 
1cb6527
2410db3
1cb6527
2410db3
 
 
 
 
 
 
 
 
1cb6527
2410db3
1cb6527
2410db3
1cb6527
2410db3
 
1cb6527
2410db3
 
 
 
1cb6527
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
import gradio as gr
from huggingface_hub import InferenceClient
from unsloth.chat_templates import get_chat_template

# Initialize the InferenceClient with the appropriate model
client = InferenceClient("wop/kosmox")

# Define the chat template and tokenizer configuration
tokenizer = get_chat_template(
    tokenizer=None,  # Assuming you need to pass an actual tokenizer here
    chat_template="phi-3",
    mapping={"role": "from", "content": "value", "user": "human", "assistant": "gpt"},
)

def format_messages(system_message, history, user_message):
    # Create a formatted string according to the specified chat template
    formatted_message = "<s>\n"
    if system_message:
        formatted_message += f"{system_message}\n"

    for user_msg, assistant_msg in history:
        if user_msg:
            formatted_message += f"{user_msg}\n"
        if assistant_msg:
            formatted_message += f"{assistant_msg}\n"
    
    formatted_message += f"{user_message}\n"
    return formatted_message

def respond(
    message: str,
    history: list[tuple[str, str]],
    system_message: str,
    max_tokens: int,
    temperature: float,
    top_p: float,
):
    # Format the messages
    formatted_message = format_messages(system_message, history, message)

    response = ""

    # Stream the response from the model
    for message in client.chat_completion(
        formatted_message,
        max_tokens=max_tokens,
        stream=True,
        temperature=temperature,
        top_p=top_p,
    ):
        token = message.choices[0].delta.content
        response += token
        yield response

# Define the Gradio interface
demo = gr.ChatInterface(
    fn=respond,
    additional_inputs=[
        gr.Textbox(value="You are AI.", label="System message"),
        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
        gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-p (nucleus sampling)"),
    ],
)

if __name__ == "__main__":
    demo.launch()