File size: 2,789 Bytes
83ddd7e | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 | # Copyright 2025 HuggingFace Inc. and the LlamaFactory team.
#
# This code is inspired by the HuggingFace's transformers library.
# https://github.com/huggingface/transformers/blob/v4.40.0/examples/pytorch/language-modeling/run_clm.py
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# 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.
from dataclasses import dataclass
from itertools import chain
from typing import Any
from .processor_utils import DatasetProcessor
@dataclass
class PretrainDatasetProcessor(DatasetProcessor):
def preprocess_dataset(self, examples: dict[str, list[Any]]) -> dict[str, list[Any]]:
# build grouped texts with format `X1 X2 X3 ...` if packing is enabled
eos_token = "<|end_of_text|>" if self.data_args.template == "llama3" else self.tokenizer.eos_token
text_examples = [messages[0]["content"] + eos_token for messages in examples["_prompt"]]
if not self.data_args.packing:
if getattr(self.tokenizer, "add_bos_token", False):
text_examples = [self.tokenizer.bos_token + example for example in text_examples]
result = self.tokenizer(
text_examples, add_special_tokens=False, truncation=True, max_length=self.data_args.cutoff_len
)
else:
tokenized_examples = self.tokenizer(text_examples, add_special_tokens=False)
concatenated_examples = {k: list(chain(*tokenized_examples[k])) for k in tokenized_examples.keys()}
total_length = len(concatenated_examples[list(concatenated_examples.keys())[0]])
block_size = self.data_args.cutoff_len
total_length = (total_length // block_size) * block_size
result = {
k: [t[i : i + block_size] for i in range(0, total_length, block_size)]
for k, t in concatenated_examples.items()
}
if getattr(self.tokenizer, "add_bos_token", False):
for i in range(len(result["input_ids"])):
result["input_ids"][i][0] = self.tokenizer.bos_token_id
return result
def print_data_example(self, example: dict[str, list[int]]) -> None:
print("input_ids:\n{}".format(example["input_ids"]))
print("inputs:\n{}".format(self.tokenizer.decode(example["input_ids"], skip_special_tokens=False)))
|