Searched defs:fused_w (Results 1 – 2 of 2) sorted by relevance
16 at::Tensor fused_w = p.linear_w * bn_scale.unsqueeze(-1); in computeUpdatedLinearWeightAndBias() local
100 …TensorType fused_w = create_tensor<TensorType>(w_shape, _data_type, 1, QuantizationInfo(), _data_… in compute_target() local