app.py

from huggingface_hub import InferenceClient
import gradio as gr

client = InferenceClient(
    "mistralai/Mixtral-8x7B-Instruct-v0.1"
)

# Formats the prompt to hold all of the past messages
def format_prompt(message, history):
    prompt = "<s>"

    # String to add before every prompt
    prompt_prefix = "Correct any grammatical errors in the following sentence and provide the corrected version:\n\nSentence: "
    prompt_template = "[INST] " + prompt_prefix + ' "{}" [/INST]  Corrected Sentence:'
    
    
    myList = [["It is my friends house in England.", "It is my friend's house in England."], ["Every girl must bring their books to school.", "Every girl must bring her books to school."]] + history


    # Iterates through every past user input and response to be added to the prompt
    for user_prompt, bot_response in myList:
        prompt += prompt_template.format(user_prompt)
        prompt += f" {bot_response}</s> "

    prompt += prompt_template.format(message)
    print("PROMPT: \n\t{}\n".format(prompt))
    return prompt




def generate(prompt, history, system_prompt, temperature=0.9, max_new_tokens=256, top_p=0.95, repetition_penalty=1.0):
    print("\n\nSystem Prompt: '{}'".format(system_prompt))
    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,)
    
    formatted_prompt = format_prompt(f"{system_prompt}, {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
        yield output
    return output



additional_inputs=[
    gr.Textbox( label="System Prompt", value="Correct the following sentence to make it grammatically accurate while maintaining the original meaning. Output only the corrected sentence." , max_lines=1, interactive=True, ),
    gr.Slider( label="Temperature", value=0.9, 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=256, 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=1.2, minimum=1.0, maximum=2.0, step=0.05, interactive=True, info="Penalize repeated tokens", )
]

examples=[['Give me the grammatically correct version of the sentence: "We shood buy an car"', None, None, None, None, None, ],
          ["Give me an example exam question testing students on square roots on basic integers", None, None, None, None, None,],
          ["Would this block of HTML code run?\n```\n\n```", None, None, None, None, None,], 
          ["I have been to New York last summer.", None, None, None, None, None,],
          ["We shood buy an car.", None, None, None, None, None,],
          ["People is coming to my party.", None, None, None, None, None,],
          ["She is more taller.", None, None, None, None, None,], 
          ["Their were lot of sheeps.", None, None, None, None, None,], 
          ["I want to speak English good.", None, None, None, None, None,], 
          ["I must to buy a new cartoon book.", None, None, None, None, None,], 
]

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="Mixtral 46.7B",
    examples=examples,
    concurrency_limit=20,
).launch(show_api=False)