# from transformers import pipeline
# import gradio as gr

# pipe = pipeline("translation", model="HuggingFaceH4/zephyr-7b-gemma-v0.1")

# demo = gr.Interface.from_pipeline(pipe)
# demo.launch()

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

model = AutoModelForCausalLM.from_pretrained(
    "tiiuae/falcon-7b-instruct",
    torch_dtype=torch.bfloat16,
    trust_remote_code=True,
    device_map="auto",
    low_cpu_mem_usage=True,
)
tokenizer = AutoTokenizer.from_pretrained("tiiuae/falcon-7b-instruct")


def generate_textt(input_text):
    input_ids = tokenizer.encode(input_text, return_tensors="pt")
    attention_mask = torch.ones(input_ids.shape)

    output = model.generate(
        input_ids,
        attention_mask=attention_mask,
        max_length=200,
        do_sample=True,
        top_k=10,
        num_return_sequences=1,
        eos_token_id=tokenizer.eos_token_id,
    )

    output_text = tokenizer.decode(output[0], skip_special_tokens=True)
    # print(output_text)

    # Remove Prompt Echo from Generated Text
    cleaned_output_text = output_text.replace(input_text, "")
    return cleaned_output_text


with gr.Blocks() as text_generation_interface:
    with gr.Row():
        input_text = gr.Textbox(label="Input Text")
        output_text = gr.Textbox(label="Generated Text2")

    generate_button = gr.Button("Generate")
    generate_button.click(fn=generate_textt, inputs=input_text, outputs=output_text)

text_generation_interface.launch()

# text_generation_interface = gr.Interface(
#     fn=generate_text,
#     inputs=[
#         gr.inputs.Textbox(label="Input Text"),
#     ],
#     outputs=gr.inputs.Textbox(label="Generated Text"),
#     title="Falcon-7B Instruct",
# ).launch()