Node names should not be changed

What commit version of aidge do you use

Latest development version of the module

Problem description

After visualizing my model, all the node names have changed. I expect the entire model to remain unchanged.

Using model.clone() as argument avoids this side effect, but is not user-friendly.

import aidge_model_explorer

print(sorted([node.name() for node in model.get_nodes()]))
> ['Add', 'Add_1', 'Constant_1_output_0', 'Constant_2_output_0', 'Constant_output_0', 'MHA_layer_Add', 'MHA_layer_Concat_output_0', 'MHA_layer_Constant_12_output_0', 'MHA_layer_Constant_15_output_0', 'MHA_layer_Constant_16_output_0', 'MHA_layer_Constant_17_output_0', 'MHA_layer_Constant_1_output_0', 'MHA_layer_Constant_output_0', 'MHA_layer_Gather', 'MHA_layer_Gather_1', 'MHA_layer_Gather_2', 'MHA_layer_Gemm', 'MHA_layer_MatMul', 'MHA_layer_MatMul_1', 'MHA_layer_MatMul_2', 'MHA_layer_Mul', 'MHA_layer_Reshape_2', 'MHA_layer_Reshape_3', 'MHA_layer_Reshape_4', 'MHA_layer_Reshape_5', 'MHA_layer_Reshape_6', 'MHA_layer_Reshape_7', 'MHA_layer_Softmax', 'MHA_layer_Transpose', 'MHA_layer_Transpose_1', 'MHA_layer_Transpose_2', 'MHA_layer_Transpose_3', 'MHA_layer_Transpose_4', 'MHA_layer_Transpose_5', 'MHA_layer_Transpose_6', 'MHA_layer_in_proj_bias', 'MHA_layer_out_proj_bias', 'MHA_layer_out_proj_weight', 'Reshape', 'Reshape_1', 'act_Sigmoid', 'batchnorm_BatchNormalization', 'batchnorm_bias', 'batchnorm_running_mean', 'batchnorm_running_var', 'batchnorm_weight', 'dense1_Gemm', 'dense1_bias', 'dense1_weight', 'dense2_Gemm', 'dense2_bias', 'dense2_weight', 'onnx__Add_137', 'onnx__MatMul_138']

aidge_model_explorer.visualize(model, "vBatch288_mod")

print(sorted([node.name() for node in model.get_nodes()]))
> ['Add_1_Add_2', 'Add_Add_1', 'Constant_1_output_0_Producer_8', 'Constant_2_output_0_Producer_22', 'Constant_output_0_Producer_4', 'MHA_layer_Add_Add_0', 'MHA_layer_Concat_output_0_Producer_1', 'MHA_layer_Constant_12_output_0_Producer_7', 'MHA_layer_Constant_15_output_0_Producer_9', 'MHA_layer_Constant_16_output_0_Producer_19', 'MHA_layer_Constant_17_output_0_Producer_14', 'MHA_layer_Constant_1_output_0_Producer_5', 'MHA_layer_Constant_output_0_Producer_3', 'MHA_layer_Gather_1', 'MHA_layer_Gather_2', 'MHA_layer_Gather_Gather_0', 'MHA_layer_Gemm_FC_1', 'MHA_layer_MatMul_1', 'MHA_layer_MatMul_2', 'MHA_layer_MatMul_MatMul_0', 'MHA_layer_Mul_Mul_0', 'MHA_layer_Reshape_2_Reshape_0', 'MHA_layer_Reshape_3_Reshape_1', 'MHA_layer_Reshape_4_Reshape_2', 'MHA_layer_Reshape_5_Reshape_3', 'MHA_layer_Reshape_6', 'MHA_layer_Reshape_7_Reshape_5', 'MHA_layer_Softmax_Softmax_0', 'MHA_layer_Transpose_1_Transpose_0', 'MHA_layer_Transpose_2', 'MHA_layer_Transpose_3_Transpose_4', 'MHA_layer_Transpose_4_Transpose_3', 'MHA_layer_Transpose_5_Transpose_6', 'MHA_layer_Transpose_6_Transpose_5', 'MHA_layer_Transpose_Transpose_1', 'MHA_layer_in_proj_bias_Producer_0', 'MHA_layer_out_proj_bias_Producer_18', 'MHA_layer_out_proj_weight_Producer_17', 'Reshape_1_Reshape_7', 'Reshape_Reshape_4', 'act_Sigmoid_Sigmoid_0', 'batchnorm_BatchNormalization_BatchNorm2D_0', 'batchnorm_bias_Producer_11', 'batchnorm_running_mean_Producer_12', 'batchnorm_running_var_Producer_13', 'batchnorm_weight_Producer_10', 'dense1_Gemm_FC_0', 'dense1_bias_Producer_16', 'dense1_weight_Producer_15', 'dense2_Gemm_FC_2', 'dense2_bias_Producer_21', 'dense2_weight_Producer_20', 'onnx__Add_137_Producer_6', 'onnx__MatMul_138_Producer_2']