Skip to content

Commit acd86ed

Browse files
committed
update
1 parent 2c608d1 commit acd86ed

File tree

3 files changed

+3
-3
lines changed

3 files changed

+3
-3
lines changed

tests/quantization/bnb/test_4bit.py

Lines changed: 2 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -45,6 +45,7 @@
4545
require_peft_backend,
4646
require_torch,
4747
require_torch_accelerator,
48+
require_torch_version_greater,
4849
require_transformers_version_greater,
4950
slow,
5051
torch_device,
@@ -860,7 +861,7 @@ def test_fp4_double_safe(self):
860861
self.test_serialization(quant_type="fp4", double_quant=True, safe_serialization=True)
861862

862863

863-
# @require_torch_version_greater("2.7.1")
864+
@require_torch_version_greater("2.7.1")
864865
class Bnb4BitCompileTests(QuantCompileTests):
865866
quantization_config = PipelineQuantizationConfig(
866867
quant_backend="bitsandbytes_8bit",

tests/quantization/test_torch_compile_utils.py

Lines changed: 0 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -74,7 +74,6 @@ def _test_torch_compile_with_group_offload_leaf(
7474
"onload_device": torch.device("cuda"),
7575
"offload_device": torch.device("cpu"),
7676
"offload_type": "leaf_level",
77-
"num_blocks_per_group": 1,
7877
"use_stream": use_stream,
7978
}
8079
pipe.transformer.enable_group_offload(**group_offload_kwargs)

tests/quantization/torchao/test_torchao.py

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -670,7 +670,7 @@ def test_torch_compile_with_group_offload_leaf(self):
670670
# weight.tensor_impl.{data,scale,zero_point}.device will be cpu
671671

672672
# For use_stream=True:
673-
# # NotImplementedError: AffineQuantizedTensor dispatch: attempting to run unimplemented operator/function: func=<OpOverload(op='aten.is_pinned', overload='default')>, types=(<class 'torchao.dtypes.affine_quantized_tensor.AffineQuantizedTensor'>,), arg_types=(<class 'torchao.dtypes.affine_quantized_tensor.AffineQuantizedTensor'>,), kwarg_types={}
673+
# NotImplementedError: AffineQuantizedTensor dispatch: attempting to run unimplemented operator/function: func=<OpOverload(op='aten.is_pinned', overload='default')>, types=(<class 'torchao.dtypes.affine_quantized_tensor.AffineQuantizedTensor'>,), arg_types=(<class 'torchao.dtypes.affine_quantized_tensor.AffineQuantizedTensor'>,), kwarg_types={}
674674
super()._test_torch_compile_with_group_offload_leaf(quantization_config=self.quantization_config)
675675

676676

0 commit comments

Comments
 (0)