app.py

import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
import random
import gradio as gr

# Replace with your actual model path
transformers_model_path = "AoEiuV020/MiniMind"

# Load the tokenizer and model
tokenizer = AutoTokenizer.from_pretrained(transformers_model_path)
model = AutoModelForCausalLM.from_pretrained(transformers_model_path, trust_remote_code=True).eval()

def setup_seed(seed):
    torch.manual_seed(seed)
    torch.cuda.manual_seed_all(seed)
    random.seed(seed)

def predict(prompt):
    messages = []
    max_seq_len = 128
    history_cnt = 0
    model_mode = 2
    setup_seed(random.randint(0, 2048))
    messages = messages[-history_cnt:] if history_cnt else []
    messages.append({"role": "user", "content": prompt})
    new_prompt = tokenizer.apply_chat_template(
        messages,
        tokenize=False,
        add_generation_prompt=True
    )[-max_seq_len - 1:] if model_mode != 0 else (tokenizer.bos_token + prompt)

    with torch.no_grad():
        x = torch.tensor(tokenizer(new_prompt)['input_ids'], device='cpu').unsqueeze(0)
        outputs = model.generate(
            x,
            eos_token_id=tokenizer.eos_token_id,
            max_new_tokens=max_seq_len,
            temperature=0.7,
            top_p=0.95,
            pad_token_id=tokenizer.pad_token_id
        )
        return tokenizer.decode(outputs.squeeze()[x.shape[1]:].tolist(), skip_special_tokens=True)

iface = gr.Interface(
    fn=predict,
    inputs=gr.Textbox(lines=2, placeholder="Enter your text here..."),
    outputs="text",
    title="MiniMind Chatbot",
    description="Enter text and see the model's response."
)

iface.launch()