Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

improve handling for empty text on the tokenization step #502

Merged
merged 3 commits into from
Sep 19, 2023
Merged

improve handling for empty text on the tokenization step #502

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
d863fc3
improve handling for empty text on the tokenization step
winglian Aug 29, 2023
e136a7b
Update src/axolotl/prompt_tokenizers.py
winglian Sep 1, 2023
7571db3
update type annotation per PR feedback
winglian Sep 19, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
25 changes: 16 additions & 9 deletions src/axolotl/prompt_tokenizers.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@
import logging
from typing import Dict, List, Tuple, Union

from transformers import PreTrainedTokenizer
from transformers import BatchEncoding, PreTrainedTokenizer

from axolotl.prompters import IGNORE_TOKEN_ID

Expand Down Expand Up @@ -66,14 +66,21 @@ def _get_assistant_token(self):
pass
return False

def _tokenize(self, prompt: str, add_eos_token=True, strip_bos_token=False):
result = self.tokenizer(
prompt,
truncation=True,
max_length=self.sequence_len,
padding=False,
return_tensors=None,
)
def _tokenize(
self, prompt: str, add_eos_token: bool = True, strip_bos_token: bool = False
) -> BatchEncoding:
result: BatchEncoding
if not prompt.strip():
LOG.warning("Empty text requested for tokenization.")
result = BatchEncoding(data={"input_ids": [], "attention_mask": []})
else:
result = self.tokenizer(
prompt,
truncation=True,
max_length=self.sequence_len,
padding=False,
return_tensors=None,
)
if len(result["input_ids"]) == 0:
LOG.warning("Tokenizer result is empty. You may want to audit your dataset")
if (
Expand Down