Indot5 Base Paraphrase

🚀 Paraphrase Generation with IndoT5 Base

The IndoT5-base model trained on translated PAWS for paraphrase generation.

🚀 Quick Start

The IndoT5-base model is trained on translated PAWS, which can be used for paraphrase generation.

💻 Usage Examples

Basic Usage

from transformers import AutoTokenizer, AutoModelForSeq2SeqLM

tokenizer = AutoTokenizer.from_pretrained("Wikidepia/IndoT5-base-paraphrase")  
model = AutoModelForSeq2SeqLM.from_pretrained("Wikidepia/IndoT5-base-paraphrase")

sentence = "Anak anak melakukan piket kelas agar kebersihan kelas terjaga"
text =  "paraphrase: " + sentence + " </s>"

encoding = tokenizer(text, padding='longest', return_tensors="pt")
outputs = model.generate(
    input_ids=encoding["input_ids"], attention_mask=encoding["attention_mask"],
    max_length=512,
    do_sample=True,
    top_k=200,
    top_p=0.95,
    early_stopping=True,
    num_return_sequences=5
)

🔧 Technical Details

Limitations

⚠️ Important Note

Sometimes the generated paraphrase contains dates that don't exist in the original text.

📄 Acknowledgement

Thanks to Tensorflow Research Cloud for providing TPU v3-8s.