From 786e4b5c41634cfd60e3b82bd5f637ff44efe2ea Mon Sep 17 00:00:00 2001 From: Kyle Sayers Date: Mon, 23 Feb 2026 12:22:10 -0500 Subject: [PATCH 1/2] remove name Signed-off-by: Kyle Sayers --- src/llmcompressor/modifiers/utils/helpers.py | 14 ++++++++------ 1 file changed, 8 insertions(+), 6 deletions(-) diff --git a/src/llmcompressor/modifiers/utils/helpers.py b/src/llmcompressor/modifiers/utils/helpers.py index c7e1fc4bc2..d67dd30cde 100644 --- a/src/llmcompressor/modifiers/utils/helpers.py +++ b/src/llmcompressor/modifiers/utils/helpers.py @@ -8,8 +8,8 @@ """ import torch +from compressed_tensors.offload import align_modules, update_offload_parameter from compressed_tensors.quantization import QuantizationStrategy, is_attention_module -from compressed_tensors.utils import align_modules, update_parameter_data from torch.nn import Linear, Module __all__ = ["update_fused_layer_weight_global_scales"] @@ -80,9 +80,9 @@ def _valid_tensor_group_quant(layer_list: list[Linear]): ) ).reshape([1]) - update_parameter_data(submodule.k_proj, global_scale, "weight_global_scale") - update_parameter_data(submodule.q_proj, global_scale, "weight_global_scale") - update_parameter_data(submodule.v_proj, global_scale, "weight_global_scale") + update_offload_parameter(submodule.k_proj, global_scale, "weight_global_scale") + update_offload_parameter(submodule.q_proj, global_scale, "weight_global_scale") + update_offload_parameter(submodule.v_proj, global_scale, "weight_global_scale") del global_scale @@ -100,7 +100,9 @@ def _valid_tensor_group_quant(layer_list: list[Linear]): ) ).reshape([1]) - update_parameter_data(submodule.gate_proj, global_scale, "weight_global_scale") - update_parameter_data(submodule.up_proj, global_scale, "weight_global_scale") + update_offload_parameter( + submodule.gate_proj, global_scale, "weight_global_scale" + ) + update_offload_parameter(submodule.up_proj, global_scale, "weight_global_scale") del global_scale From df2ef948bbf196790d617463b4d4c858edcbddbf Mon Sep 17 00:00:00 2001 From: Kyle Sayers Date: Mon, 23 Feb 2026 12:23:41 -0500 Subject: [PATCH 2/2] change arg order Signed-off-by: Kyle Sayers --- src/llmcompressor/modifiers/utils/helpers.py | 12 +++++++----- 1 file changed, 7 insertions(+), 5 deletions(-) diff --git a/src/llmcompressor/modifiers/utils/helpers.py b/src/llmcompressor/modifiers/utils/helpers.py index d67dd30cde..cbba632d94 100644 --- a/src/llmcompressor/modifiers/utils/helpers.py +++ b/src/llmcompressor/modifiers/utils/helpers.py @@ -80,9 +80,9 @@ def _valid_tensor_group_quant(layer_list: list[Linear]): ) ).reshape([1]) - update_offload_parameter(submodule.k_proj, global_scale, "weight_global_scale") - update_offload_parameter(submodule.q_proj, global_scale, "weight_global_scale") - update_offload_parameter(submodule.v_proj, global_scale, "weight_global_scale") + update_offload_parameter(submodule.k_proj, "weight_global_scale", global_scale) + update_offload_parameter(submodule.q_proj, "weight_global_scale", global_scale) + update_offload_parameter(submodule.v_proj, "weight_global_scale", global_scale) del global_scale @@ -101,8 +101,10 @@ def _valid_tensor_group_quant(layer_list: list[Linear]): ).reshape([1]) update_offload_parameter( - submodule.gate_proj, global_scale, "weight_global_scale" + submodule.gate_proj, + "weight_global_scale", + global_scale, ) - update_offload_parameter(submodule.up_proj, global_scale, "weight_global_scale") + update_offload_parameter(submodule.up_proj, "weight_global_scale", global_scale) del global_scale