app.py

import gradio as gr
import gradio_client
from gradio_client import Client 

title = """# Welcome To Tonic's Easy YI-6B-200K"""
description = """This is a [connector that looks like open ai](https://huggingface.co/spaces/Tonic1/TonicsYI-6B-200k) on the inputs , then ignores all that and connects to my [YI-6B-200K endpoint](https://huggingface.co/01-ai/Yi-6B-200K). This helps using [Yi for AGI here](https://huggingface.co/spaces/Tonic/AGYIntelligence)"""
client = Client("https://tonic1-tonicsyi-6b-200k.hf.space/--replicas/cvjqf/")
def chat_api_interface(messages, model, frequency_penalty=0, logit_bias=None, max_tokens=None, n=1, presence_penalty=0, response_format="text", seed=None, stop=None, stream=False, temperature=0.9, top_p=1, tools=None, tool_choice="none", user=None):
    result = client.predict(
        your_message = messages,
        system_prompt = "I am YiTonic, an AI language model created by Tonic-AI. I am a cautious assistant. I carefully follow instructions. I am helpful and harmless and I follow ethical guidelines and promote positive behavior.",
        max_new_tokens = 16000,
        temperature,
        top_p,
        top_k,
        disable_for_faster_inference,
        api_name="/predict"
    )
    return result

with gr.Blocks() as demo:
    with gr.Row():
        with gr.Column():
            messages = gr.Textbox(label="Messages (JSON format)", placeholder="Enter messages as a JSON array", value="[{\"role\": \"user\", \"content\": \"Hello, how can I help you today?\"}, {\"role\": \"assistant\", \"content\": \"I'm looking for information on renewable energy sources.\"}]")
            model = gr.Textbox(label="Model ID", placeholder="Enter the model ID", value="text-davinci-003")
            frequency_penalty = gr.Number(label="Frequency Penalty", value=0.5)
            logit_bias = gr.Textbox(label="Logit Bias (JSON format)", placeholder="Enter logit bias as JSON", value="{\"50256\": -100, \"50257\": 100}")
            max_tokens = gr.Number(label="Max Tokens", optional=True, value=150)
            n = gr.Number(label="N", value=3)
            presence_penalty = gr.Number(label="Presence Penalty", value=0.6)
            response_format = gr.Radio(label="Response Format", choices=["text", "json_object"], value="text")
            seed = gr.Number(label="Seed", optional=True, value=42)
            stop = gr.Textbox(label="Stop Sequence(s)", placeholder="Enter stop sequences", value="[\"\\n\", \" end\"]")
            stream = gr.Checkbox(label="Stream", value=False)
            temperature = gr.Slider(label="Temperature", minimum=0, maximum=2, value=0.7)
            top_p = gr.Slider(label="Top P", minimum=0, maximum=1, value=0.9)
            tools = gr.Textbox(label="Tools (JSON format)", optional=True, placeholder="Enter tools as JSON", value="{\"spellcheck\": true, \"grammar_check\": false}")
            tool_choice = gr.Textbox(label="Tool Choice", value="spellcheck", placeholder="Enter tool choice")
            user = gr.Textbox(label="User", optional=True, placeholder="Enter user identifier", value="user123")

        with gr.Column():
            submit_btn = gr.Button("Submit")
            output = gr.Textbox(label="API Response", readonly=True)

    submit_btn.click(
        chat_api_interface, 
        inputs=[messages, model, frequency_penalty, logit_bias, max_tokens, n, presence_penalty, response_format, seed, stop, stream, temperature, top_p, tools, tool_choice, user],
        outputs=output
    )

demo.launch()