File tree Expand file tree Collapse file tree 1 file changed +2
-2
lines changed
onnxruntime/contrib_ops/webgpu/quantization Expand file tree Collapse file tree 1 file changed +2
-2
lines changed Original file line number Diff line number Diff line change @@ -137,9 +137,9 @@ Status GatherBlockQuantized::ComputeInternal(ComputeContext& context) const {
137137 bool is_signed = x_dtype == ONNX_TENSOR_ELEMENT_DATA_TYPE_INT8 || x_dtype == ONNX_TENSOR_ELEMENT_DATA_TYPE_INT4;
138138 bool is_int8 = x_dtype == ONNX_TENSOR_ELEMENT_DATA_TYPE_INT8 || x_dtype == ONNX_TENSOR_ELEMENT_DATA_TYPE_UINT8;
139139
140+ std::optional<Tensor> data_representation_4bit;
141+ std::optional<Tensor> zero_points_representation_4bit;
140142 if (bits_ == 4 && is_int8) {
141- std::optional<Tensor> data_representation_4bit;
142- std::optional<Tensor> zero_points_representation_4bit;
143143 TensorShape data_representation_4bit_shape{x->Shape ()};
144144 MLDataType new_dtype = (x_dtype == ONNX_TENSOR_ELEMENT_DATA_TYPE_UINT8) ? DataTypeImpl::GetType<UInt4x2>() : DataTypeImpl::GetType<Int4x2>();
145145 auto memory_info = OrtMemoryInfo{
You can’t perform that action at this time.
0 commit comments