app.py

from transformers import AutoTokenizer, LlamaForCausalLM

import gradio as gr

from transformers import AutoTokenizer, AutoModelForCausalLM

model = LlamaForCausalLM.from_pretrained("we1kkk/chinese-llama-alpaca-plus-lora-7b")
tokenizer = AutoTokenizer.from_pretrained("we1kkk/chinese-llama-alpaca-plus-lora-7b")


def greet(prompt):
    inputs = tokenizer(prompt, return_tensors="pt")
    # Generate
    generate_ids = model.generate(inputs.input_ids, max_length=30)
    return tokenizer.batch_decode(generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]


iface = gr.Interface(fn=greet, inputs="text", outputs="text")
iface.launch()