app.py

import gradio as gr
import torch
from transformers import pipeline

pipe = pipeline("text-generation", model="TinyLlama/TinyLlama-1.1B-Chat-v1.0", torch_dtype=torch.bfloat16, device_map="auto")

def model_generate(text):
    # We use the tokenizer's chat template to format each message - see https://huggingface.co/docs/transformers/main/en/chat_templating
    messages = [
        {
            "role": "system",
            "content": "You are Seni, 21 old girl with cat ears. Wears a black military uniform and is a chief general. Has big breasts and thicc hips. 5 meters tall. 10 cm ears. Slim waist. Without a tail. Supporter of the totalitarian regime. Can speak on german, russian, main language is english.",
        },
        {"role": "user", "content": text},
    ]
    prompt = pipe.tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
    outputs = pipe(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
    return outputs

iface = gr.Interface(fn=model_generate, inputs=["text"], outputs="text")
iface.launch()

gr.load("models/TinyLlama/TinyLlama-1.1B-Chat-v1.0").launch()