source/FLAML
1
0
Fork 0
mirror of https://github.com/microsoft/FLAML.git synced 2026-02-09 02:09:16 +08:00
Code Issues Packages Projects Releases Wiki Activity

Fix transformers API compatibility: support v4.26+ and v5.0+ with version-aware parameter selection (#1514)

Browse Source 
* Initial plan

* Fix transformers API compatibility issues

Co-authored-by: thinkall <3197038+thinkall@users.noreply.github.com>

* Add backward compatibility for transformers v4.26+ by version check

Support both tokenizer (v4.26-4.43) and processing_class (v4.44+) parameters based on installed transformers version. Fallback to tokenizer if version check fails.

Co-authored-by: thinkall <3197038+thinkall@users.noreply.github.com>

* Improve exception handling specificity

Use specific exception types (ImportError, AttributeError, ValueError) instead of broad Exception catch for better error handling.

Co-authored-by: thinkall <3197038+thinkall@users.noreply.github.com>

* Run pre-commit formatting on all files

Applied black formatting to fix code style across the repository.

Co-authored-by: thinkall <3197038+thinkall@users.noreply.github.com>

---------

Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
Co-authored-by: thinkall <3197038+thinkall@users.noreply.github.com>
This commit is contained in:
Copilot
2026-01-28 09:00:21 +08:00
committed by GitHub
parent a5021152d2
commit 158ff7d99e
2 changed files with 43 additions and 29 deletions
Download Patch File Download Diff File Expand all files Collapse all files

35
flaml/automl/model.py
View File

@@ -1196,16 +1196,31 @@ class TransformersEstimator(BaseEstimator):
control.should_save = True control.should_save = True
control.should_evaluate = True control.should_evaluate = True
self._trainer = TrainerForAuto( # Use processing_class for transformers >= 4.44.0, tokenizer for older versions
args=self._training_args, trainer_kwargs = {
model_init=self._model_init, "args": self._training_args,
train_dataset=train_dataset, "model_init": self._model_init,
eval_dataset=eval_dataset, "train_dataset": train_dataset,
tokenizer=self.tokenizer, "eval_dataset": eval_dataset,
data_collator=self.data_collator, "data_collator": self.data_collator,
compute_metrics=self._compute_metrics_by_dataset_name, "compute_metrics": self._compute_metrics_by_dataset_name,
callbacks=[EarlyStoppingCallbackForAuto], "callbacks": [EarlyStoppingCallbackForAuto],
) }
# Check if processing_class parameter is supported (transformers >= 4.44.0)
try:
import transformers
from packaging import version
if version.parse(transformers.__version__) >= version.parse("4.44.0"):
trainer_kwargs["processing_class"] = self.tokenizer
else:
trainer_kwargs["tokenizer"] = self.tokenizer
except (ImportError, AttributeError, ValueError):
# Fallback to tokenizer if version check fails
trainer_kwargs["tokenizer"] = self.tokenizer
self._trainer = TrainerForAuto(**trainer_kwargs)
if self._task in NLG_TASKS: if self._task in NLG_TASKS:
setattr(self._trainer, "_is_seq2seq", True) setattr(self._trainer, "_is_seq2seq", True)

37
flaml/automl/nlp/huggingface/utils.py
View File

@@ -211,29 +211,28 @@ def tokenize_onedataframe(
hf_args=None, hf_args=None,
prefix_str=None, prefix_str=None,
): ):
with tokenizer.as_target_tokenizer(): _, tokenized_column_names = tokenize_row(
_, tokenized_column_names = tokenize_row( dict(X.iloc[0]),
dict(X.iloc[0]), tokenizer,
prefix=(prefix_str,) if task is SUMMARIZATION else None,
task=task,
hf_args=hf_args,
return_column_name=True,
)
d = X.apply(
lambda x: tokenize_row(
x,
tokenizer, tokenizer,
prefix=(prefix_str,) if task is SUMMARIZATION else None, prefix=(prefix_str,) if task is SUMMARIZATION else None,
task=task, task=task,
hf_args=hf_args, hf_args=hf_args,
return_column_name=True, ),
) axis=1,
d = X.apply( result_type="expand",
lambda x: tokenize_row( )
x, X_tokenized = pd.DataFrame(columns=tokenized_column_names)
tokenizer, X_tokenized[tokenized_column_names] = d
prefix=(prefix_str,) if task is SUMMARIZATION else None, return X_tokenized
task=task,
hf_args=hf_args,
),
axis=1,
result_type="expand",
)
X_tokenized = pd.DataFrame(columns=tokenized_column_names)
X_tokenized[tokenized_column_names] = d
return X_tokenized
def tokenize_row( def tokenize_row(