import os | |
from transformers import AutoTokenizer | |
os.environ["PROTOCOL_BUFFERS_PYTHON_IMPLEMENTATION"] = "python" | |
CURRENT_DIR = os.path.dirname(os.path.abspath(__file__)) | |
TOKENIZER_DIR = os.path.join(CURRENT_DIR, "tokenizer") | |
tokenizer = AutoTokenizer.from_pretrained(TOKENIZER_DIR, trust_remote_code=True) | |
# vocab_size = len(tokenizer.get_vocab()) | |
# vocab_size = tokenizer.vocab_size | |