test4consistent.py

# Copyright (C) 2024 THL A29 Limited, a Tencent company.  All rights reserved.
#
# Licensed under the TENCENT HUNYUAN COMMUNITY LICENSE AGREEMENT (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     https://github.com/Tencent/Tencent-Hunyuan-Large/blob/main/License.docx
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# test tokenizer encode & decode consistency
from transformers import AutoTokenizer
tokenizer = AutoTokenizer.from_pretrained('/tokenizer_exp/other_tokenizer_vocab/hy', local_files_only=True, trust_remote_code=True)

test_data = [line.strip() for line in open('/tokenizer_exp/data/test.txt', 'r').readlines()]

num_origi_len = 0
num_token_len = 0

for d in test_data:
    a = tokenizer.encode(d)
    num_origi_len += len(d)
    num_token_len += len(a)
    b = tokenizer.decode(a)
    assert b == d, f"encode & decode not consistent: {d} vs {b}"
    
print(f" original length: {num_origi_len}")
print(f" token length: {num_token_len}")