tianyuz's picture
fix tokenizer
9b9b353
|
raw
history blame
574 Bytes

japanese-gpt2-medium

rinna-icon

This repository provides a medium-sized Japanese GPT-2 model trained on Japanese CC-100. The model is provided by rinna.

Use the model

NOTE: Use T5Tokenizer to initiate the tokenizer with argument extra_ids=0.

from transformers import T5Tokenizer, AutoModelForCausalLM

tokenizer = T5Tokenizer.from_pretrained("rinna/japanese-gpt2-medium", extra_ids=0)

model = AutoModelForCausalLM.from_pretrained("rinna/japanese-gpt2-medium")