app.py

import time
import gradio as gr
from ctransformers import AutoModelForCausalLM
from spellchecker import SpellChecker

llm = AutoModelForCausalLM.from_pretrained("TheBloke/WizardLM-7B-uncensored-GGUF", model_file="WizardLM-7B-uncensored.Q3_K_M.gguf", model_type="llama", stream=True)
history = [""]
spell = SpellChecker()

def correct_words(text):
    words = text.split()
    corrected_words = [spell.correction(word) for word in words]
    corrected_text = ' '.join(corrected_words)
    return corrected_text
import asyncio
import asyncio

# ... (previous code)

async def generate_async(tokens):
    for token in tokens:
        yield token

async def generate_response_async(message):
    global history
    for _ in range(2):
        tokens = [ord(char) for char in message]
        response = llm.generate_async(tokens=tokens, top_k=50, top_p=0.95, temperature=1.0, repetition_penalty=1.0, last_n_tokens=1)
        response_text = ' '.join(map(str, await asyncio.gather(*response)))
        time.sleep(2)
        corrected_response = correct_words(response_text)
        history.append(corrected_response)
        yield ' '.join(history)
    history = ["Chatbot:"]
    # Clear the history list after the last response
    history = ["Chatbot:"]
def chatbot(message, history):
    response_generator = generate_response(message)
    for response in response_generator:
        time.sleep(0.1)  # Optional delay for a natural chat feel
        yield response

iface = gr.ChatInterface(chatbot)
iface.launch()