GIPBERT / app.py
rwheel's picture
Create app.py
b550c2c
raw
history blame
1.52 kB
import gradio as gr
import torch
from peft import PeftModel, PeftConfig
from transformers import AutoModelForCausalLM, AutoTokenizer
peft_model_id = "rwheel/discriminacion_gitana_intervenciones"
config = PeftConfig.from_pretrained(peft_model_id)
model = AutoModelForCausalLM.from_pretrained(config.base_model_name_or_path, return_dict=True, load_in_8bit=True, device_map='auto')
tokenizer = AutoTokenizer.from_pretrained(peft_model_id)
# Load the Lora model
model = PeftModel.from_pretrained(model, peft_model_id)
def predecir_intervencion(text):
text = "<SH>" + text + " Intervenci贸n: "
batch = tokenizer(text, return_tensors='pt')
with torch.cuda.amp.autocast():
output_tokens = model.generate(**batch, max_new_tokens=256, eos_token_id=50258)
output = tokenizer.decode(output_tokens[0], skip_special_tokens=False)
aux = output.split("Intervenci贸n:")[1].strip()
intervencion = aux.split("Resultado:")[0].strip()
resultado = aux.split("Resultado:")[1].split("<EH>")[0].strip()
return intervencion, resultado
with gr.Blocks() as demo:
gr.Markdown("Predicci贸n de intervenciones para mitigar el da帽o racista en el pueblo gitano")
with gr.Row():
hechos = gr.Textbox(placeholder="Un alumno gitano de un Instituto...")
with gr.Row():
intervencion = gr.Textbox()
resultado = gr.Textbox()
btn = gr.Button("Go")
btn.click(fn=predecir_intervencion, inputs=hechos, outputs=[intervencion, resultado])
demo.launch(share=True)