Skip to content

Commit

Permalink
FIX Avoid early import of torch extension by BOFT (#1879)
Browse files Browse the repository at this point in the history
  • Loading branch information
PhyscalX authored Jun 26, 2024
1 parent 27bc305 commit 7ffa43b
Showing 1 changed file with 3 additions and 1 deletion.
4 changes: 3 additions & 1 deletion src/peft/tuners/boft/layer.py
Original file line number Diff line number Diff line change
Expand Up @@ -27,7 +27,6 @@
import torch.nn as nn
import torch.nn.functional as F
from torch.autograd import Function
from torch.utils.cpp_extension import load

from peft.tuners.tuners_utils import BaseTunerLayer, check_adapters_to_merge

Expand Down Expand Up @@ -78,6 +77,9 @@ def get_fbd_cuda():
if _FBD_CUDA is not None:
return _FBD_CUDA

# This import initializes cuda context and should thus be local, see issue 1877
from torch.utils.cpp_extension import load

curr_dir = os.path.dirname(__file__)
# need ninja to build the extension
try:
Expand Down

0 comments on commit 7ffa43b

Please sign in to comment.