Updates the tokenizer configuration file
#3
by
lysandre
HF staff
- opened
The tokenizer configuration file is missing/incorrect and therefore leading to unforeseen errors after the migration of the canonical models.
Refer to the following issue for more information: transformers#29050
The current failing code is the following:
from transformers import AutoTokenizer
>>> previous_tokenizer = AutoTokenizer.from_pretrained("xlm-mlm-ende-1024")
>>> current_tokenizer = AutoTokenizer.from_pretrained("FacebookAI/xlm-mlm-ende-1024")
>>> print(previous_tokenizer.model_max_length, current_tokenizer.model_max_length)
512, 512
This is the result after the fix:
from transformers import AutoTokenizer
>>> previous_tokenizer = AutoTokenizer.from_pretrained("xlm-mlm-ende-1024")
>>> current_tokenizer = AutoTokenizer.from_pretrained("FacebookAI/xlm-mlm-ende-1024")
>>> print(previous_tokenizer.model_max_length, current_tokenizer.model_max_length)
512, 512