import gradio as gr import os import spaces from transformers import GemmaTokenizer, AutoModelForCausalLM from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer from threading import Thread # Set an environment variable HF_TOKEN = os.environ.get("HF_TOKEN", None) # Lê as variáveis de ambiente para autenticação e compartilhamento #auth_users = os.environ.get("GRADIO_AUTH_USERS") #auth_passwords = os.environ.get("GRADIO_AUTH_PASSWORDS") # Converte as strings de usuários e senhas em listas #auth_users = [user.strip() for user in auth_users.split(",")] #auth_passwords = [password.strip() for password in auth_passwords.split(",")] # Cria um dicionário de autenticação #auth_credentials = dict(zip(auth_users, auth_passwords)) DESCRIPTION = '''

Meta Llama3 8B

This Space demonstrates the instruction-tuned model Meta Llama3 8b Chat. Meta Llama3 is the new open LLM and comes in two sizes: 8b and 70b. Feel free to play with it, or duplicate to run privately!

🔎 For more details about the Llama3 release and how to use the model with transformers, take a look at our blog post.

🦕 Looking for an even more powerful model? Check out the Hugging Chat integration for Meta Llama 3 70b

''' LICENSE = """

--- CreativeWoks AI: Intelligence System for Advanced Dialogue and Organized Responses Assistance """ PLACEHOLDER = """

CreativeWorks Ai

Este espaço demonstra o modelo customizado para o português brasileiro Mistral-7B-v0.3. O Mistral-7B-v0.3 Large Language Model (LLM) é uma versão do Mistral-7B-v0.2 com vocabulário expandido. A CreativeWorks modificou e afinou o modelo para que seja mais rápido e alcance desempenho comparável aos principais modelos de código aberto existentes 10 vezes maiores, incluindo diversas melhorias e otimização para raciocínio lógico, com foco em RAG (Recuperação Aumentada por Geração).

🔎 Para mais detalhes sobre o modelo e como utilizá-lo com transformers, dê uma olhada em nosso model card..

🦕 Procurando um modelo ainda mais poderoso? Confira a integração do Hugging Chat para modelos maiores.

""" css = """ h1 { text-align: center; display: block; } #duplicate-button { margin: auto; color: white; background: #1565c0; border-radius: 100vh; } """ # Load the tokenizer and model tokenizer = AutoTokenizer.from_pretrained("CreativeWorksAi/CreativeWorks_Mistral_7b_Chat_V1") model = AutoModelForCausalLM.from_pretrained("CreativeWorksAi/CreativeWorks_Mistral_7b_Chat_V1", token=HF_TOKEN, device_map="auto") #model = AutoModelForCausalLM.from_pretrained("meta-llama/Meta-Llama-3-8B-Instruct", device_map="auto") # to("cuda:0") terminators = [ tokenizer.eos_token_id, tokenizer.convert_tokens_to_ids("") ] @spaces.GPU(duration=120) def CreativeWorks_Mistral_7b_Chat_V1(message: str, history: list, temperature: float, max_new_tokens: int ) -> str: """ Generate a streaming response using the Mistral model. Args: message (str): The input message. history (list): The conversation history used by ChatInterface. temperature (float): The temperature for generating the response. max_new_tokens (int): The maximum number of new tokens to generate. Returns: str: The generated response. """ conversation = [] for user, assistant in history: conversation.extend([{"from": "human", "value": user}, {"from": "assistant", "value": assistant}]) conversation.append({"from": "human", "value": message}) input_ids = tokenizer.apply_chat_template(conversation, return_tensors="pt").to(model.device) streamer = TextIteratorStreamer(tokenizer, timeout=10.0, skip_prompt=True, skip_special_tokens=True) generate_kwargs = dict( input_ids=input_ids, streamer=streamer, max_new_tokens=max_new_tokens, do_sample=True, temperature=temperature, eos_token_id=terminators, pad_token_id=tokenizer.eos_token_id ) # This will enforce greedy generation (do_sample=False) when the temperature is passed 0, avoiding the crash. if temperature == 0: generate_kwargs['do_sample'] = False t = Thread(target=model.generate, kwargs=generate_kwargs) t.start() outputs = [] for text in streamer: # Remove the unwanted prefix if present text = text.replace("<|im_start|>assistant", " ") outputs.append(text) yield "".join(outputs) # Gradio block chatbot=gr.Chatbot(height=450, placeholder=PLACEHOLDER, label='CreativeWorks Ai') with gr.Blocks(fill_height=True, css=css) as demo: #gr.Markdown(DESCRIPTION) #gr.DuplicateButton(value="Duplicate Space for private use", elem_id="duplicate-button") gr.ChatInterface( fn=CreativeWorks_Mistral_7b_Chat_V1, chatbot=chatbot, fill_height=True, additional_inputs_accordion=gr.Accordion(label="⚙️ Parameters", open=False, render=False), additional_inputs=[ gr.Slider(minimum=0, maximum=1, step=0.1, value=0.95, label="Temperature", render=False), gr.Slider(minimum=256, maximum=8192, step=1, value=512, label="Max new tokens", render=False ), ], examples=[ ['How to setup a human base on Mars? Give short answer.'], ['Explain theory of relativity to me like I’m 8 years old.'], ['What is 9,000 * 9,000?'], ['Write a pun-filled happy birthday message to my friend Alex.'], ['Justify why a penguin might make a good king of the jungle.'] ], cache_examples=False, ) gr.Markdown(LICENSE) if __name__ == "__main__": demo.launch(auth=("teste", "teste@teste"), share=True)