File size: 1,598 Bytes
ab3c637 bc57463 49e23eb e15cc06 49e23eb ab3c637 dcf4887 673400c b6f4106 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 |
import gradio as gr
from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
def translate(text):
model_name = 'hackathon-pln-es/t5-small-finetuned-spanish-to-quechua'
model = AutoModelForSeq2SeqLM.from_pretrained(model_name)
tokenizer = AutoTokenizer.from_pretrained(model_name)
input = tokenizer(text, return_tensors="pt")
output = model.generate(input["input_ids"], max_length=40, num_beams=4, early_stopping=True)
return tokenizer.decode(output[0], skip_special_tokens=True)
title = "Spanish to Quechua translation 🦙"
inputs = gr.inputs.Textbox(lines=1, label="Text in Spanish")
outputs = [gr.outputs.Textbox(label="Translated text in Quechua")]
description = "Here use the [t5-small-finetuned-spanish-to-quechua-model](https://huggingface.co/hackathon-pln-es/t5-small-finetuned-spanish-to-quechua) that was trained with [spanish-to-quechua dataset](https://huggingface.co/datasets/hackathon-pln-es/spanish-to-quechua)."
article = '''
## Challenges
- Create a dataset, as there are different variants of Quechua.
- Training of the model to optimize results using the least amount of computational resources.
## Team members
- [Sara Benel](https://huggingface.co/sbenel)
- [Jose Vílchez](https://huggingface.co/JCarlos)
'''
examples=[
'Dios ama a los hombres',
'A pesar de todo, soy feliz',
'¿Qué harán allí?',
'Debes aprender a respetar',
]
iface = gr.Interface(fn=translate, inputs=inputs, outputs=outputs, theme="grass", css="styles.css", examples=examples, title=title, description=description, article=article)
iface.launch(enable_queue=True) |