Skip to content

Commit 07c294c

Browse files
hmellorvasqu
authored andcommitted
Fix peft lower bound (#46605)
Signed-off-by: Harry Mellor <[email protected]>
1 parent e0e7504 commit 07c294c

4 files changed

Lines changed: 5 additions & 5 deletions

File tree

‎docs/source/en/peft.md‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -15,7 +15,7 @@ rendered properly in your Markdown viewer.
1515

1616
Transformers integrates directly with the PEFT library through [`~integrations.PeftAdapterMixin`], added to all [`PreTrainedModel`] classes. You can load, add, train, switch, and delete adapters without wrapping your model in a separate [`~peft.PeftModel`]. All non-prompt-learning PEFT methods are supported (LoRA, IA3, AdaLoRA). Prompt-based methods like prompt tuning and prefix tuning require using the [PEFT library](https://huggingface.co/docs/peft/index) directly.
1717

18-
Install PEFT to get started. The integration requires `peft >= 0.18.0`.
18+
Install PEFT to get started. The integration requires `peft >= 0.19.0`.
1919

2020
```shell
2121
pip install -U peft

‎setup.py‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -104,7 +104,7 @@
104104
"pandas<2.3.0", # `datasets` requires `pandas` while `pandas==2.3.0` has issues with CircleCI on 2025/06/05
105105
"packaging>=20.0",
106106
"parameterized>=0.9", # older version of parameterized cause pytest collection to fail on .expand
107-
"peft>=0.18.0",
107+
"peft>=0.19.0",
108108
"phonemizer",
109109
"protobuf",
110110
"psutil",

‎src/transformers/dependency_versions_table.py‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -36,7 +36,7 @@
3636
"pandas": "pandas<2.3.0",
3737
"packaging": "packaging>=20.0",
3838
"parameterized": "parameterized>=0.9",
39-
"peft": "peft>=0.18.0",
39+
"peft": "peft>=0.19.0",
4040
"phonemizer": "phonemizer",
4141
"protobuf": "protobuf",
4242
"psutil": "psutil",

‎src/transformers/integrations/peft.py‎

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -58,7 +58,7 @@
5858
from accelerate.utils import get_balanced_memory, infer_auto_device_map
5959

6060
# Minimum PEFT version supported for the integration
61-
MIN_PEFT_VERSION = "0.18.2"
61+
MIN_PEFT_VERSION = "0.19.0"
6262

6363

6464
logger = logging.get_logger(__name__)
@@ -415,7 +415,7 @@ class PeftAdapterMixin:
415415
prompt tuning, prompt learning are out of scope as these adapters are not "injectable" into a torch module. For
416416
using these methods, please refer to the usage guide of PEFT library.
417417
418-
With this mixin, if the correct PEFT version is installed (>= 0.18.0), it is possible to:
418+
With this mixin, if the correct PEFT version is installed (>= 0.19.0), it is possible to:
419419
420420
- Load an adapter stored on a local path or in a remote Hub repository, and inject it in the model
421421
- Attach new adapters in the model and train them with Trainer or by your own.

0 commit comments

Comments
 (0)