iiiorg

/

piiranha-v1-detect-personal-information

like 221

Follow

The Internet Integrity Initiative 36

Thank you for sharing this model which looks promising for my application !

Could you please confirm whether the max context length is 256 or 512 ?

• in the README.md I found "Piiranha is a fine-tuned version of microsoft/mdeberta-v3-base. The context length is 256 Deberta tokens. If your text is longer than that, just split it up."
• yet in config.json I found "max_position_embeddings": 512 as well as "position_buckets": 256
• and I got nothing when trying to print tokenizer.model_max_length in my Python terminal