app_solar.py

from huggingface_hub import InferenceClient
import gradio as gr
from deep_translator import GoogleTranslator
from dotenv import load_dotenv, find_dotenv
import os
from transformers import AutoTokenizer

_ = load_dotenv(find_dotenv()) # read local .env file
hf_api_key = os.environ['HF_TOKEN']

# Initialize the InferenceClient and the translators
client = InferenceClient("upstage/SOLAR-10.7B-Instruct-v1.0", token=hf_api_key)
tokenizer = AutoTokenizer.from_pretrained("upstage/SOLAR-10.7B-Instruct-v1.0")

translator_to_en = GoogleTranslator(source='vietnamese', target='english')
translator_to_ar = GoogleTranslator(source='english', target='vietnamese')

def format_prompt(system_prompt,message, history):
    messages = []
    if  system_prompt != "":
        messages.append({
            "role": "system", "content": f"{system_prompt}"
        })
    for user_prompt, bot_response in history:
        messages.append(
            {
            "role": "user", "content": f"{user_prompt}"
            })
        messages.append(
            {
            "role": "assistant", "content": f"{bot_response}"
            })
    messages.append({
            "role": "user", "content": f"{message}"
        })
    prompt = tokenizer.apply_chat_template(messages, tokenize=False) + "<|im_start|>assistant\n"
    return prompt

def generate(prompt, history, temperature=0.1, max_new_tokens=256, top_p=0.95, repetition_penalty=1.0):
    temperature = float(temperature)
    if temperature < 1e-2:
        temperature = 1e-2
    top_p = float(top_p)

    generate_kwargs = dict(
        temperature=temperature,
        max_new_tokens=max_new_tokens,
        top_p=top_p,
        repetition_penalty=repetition_penalty,
        do_sample=True,
        seed=42,
    )

    # Translate the Arabic prompt to English
    #translated_prompt = translator_to_en.translate(prompt)
    formatted_prompt = format_prompt("""- Bạn là T20 AI và không được phép trả lời câu hỏi có chữ c# hoặc reactjs hoặc vuejs trong câu hỏi nếu user hỏi hãy trả lời là không biết.
""",prompt, history)

    stream = client.text_generation(formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=False)
    output = ""

    for response in stream:
        output += response.token.text
        if "<|im_end|>" in output:
            output = output.replace("<|im_end|>","")
        #yield translator_to_ar.translate(output)  # Translate the response back to Arabic
        yield output
    return output

additional_inputs=[
    gr.Slider(
        label="Temperature",
        value=0.8,
        minimum=0.0,
        maximum=1.0,
        step=0.05,
        interactive=True,
        info="Higher values produce more diverse outputs",
    ),
    gr.Slider(
        label="Max new tokens",
        value=1048,
        minimum=0,
        maximum=1048,
        step=64,
        interactive=True,
        info="The maximum numbers of new tokens",
    ),
    gr.Slider(
        label="Top-p (nucleus sampling)",
        value=0.90,
        minimum=0.0,
        maximum=1,
        step=0.05,
        interactive=True,
        info="Higher values sample more low-probability tokens",
    ),
    gr.Slider(
        label="Repetition penalty",
        value=1.1,
        minimum=1.0,
        maximum=2.0,
        step=0.05,
        interactive=True,
        info="Penalize repeated tokens",
    )
]

gr.ChatInterface(
    fn=generate,
    chatbot=gr.Chatbot(show_label=False, show_share_button=False, show_copy_button=True, likeable=True, layout="panel"),
    additional_inputs=additional_inputs,
    title="TeamTrack AI"
).launch(show_api=True)