This model takes text as input and returns the top five paraphrased versions of the input text. The T5 model is fine-tuned using persuasive ad transcripts.

Example usage:

import torch
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM

tokenizer = AutoTokenizer.from_pretrained("paragon-analytics/t5_para")  
model = AutoModelForSeq2SeqLM.from_pretrained("paragon-analytics/t5_para").to(device)

sentence = "This is something"

text =  "paraphrase: " + sentence + " </s>"

encoding = tokenizer.encode_plus(text,pad_to_max_length=True, return_tensors="pt")
input_ids, attention_masks = encoding["input_ids"].to("cuda"), encoding["attention_mask"].to("cuda")

outputs = model.generate(
    input_ids=input_ids, attention_mask=attention_masks,
    max_length=256,
    do_sample=True,
    top_k=120,
    top_p=0.95,
    early_stopping=True,
    num_return_sequences=5
)

for output in outputs:
    line = tokenizer.decode(output, skip_special_tokens=True,clean_up_tokenization_spaces=True)
    print(line)