Segmentation fault (core dumped) when converting to rknn

rknn-toolkit2 version: 2.0.0b0+9bab5682 --> Config model done --> Loading model I It is recommended onnx opset 19, but your onnx model opset is 11! I Loading : 100%|███████████████████████████████████████████████| 37/37 [00:00<00:00, 128680.97it/s] W load_onnx: The config.mean_values is None, zeros will be set for input 0! W load_onnx: The config.std_values is None, ones will be set for input 0! done --> Building model D base_optimize ... D base_optimize done. D D fold_constant ... D fold_constant done. D D correct_ops ... D correct_ops done. D D fuse_ops ... D fuse_ops results: D convert_squeeze_to_reshape: remove node = ['model_4/tf.unstack_3/unstack41'], add node = ['model_4/tf.unstack_3/unstack41_2reshape'] D convert_squeeze_to_reshape: remove node = ['model_4/tf.unstack_3/unstack42'], add node = ['model_4/tf.unstack_3/unstack42_2reshape'] D remove_invalid_slice: remove node = ['model_4/tf.operators.getitem_3/strided_slice1'] D swap_reshape_add: remove node = ['model_4/tf.operators.getitem_3/strided_slice', 'model_4/tf.operators.add_12/AddV2'], add node = ['model_4/tf.operators.add_12/AddV2', 'model_4/tf.operators.getitem_3/strided_slice'] D unsqueeze_to_4d_gather: remove node = [], add node = ['model_4/tf.compat.v1.gather_3/GatherV2;model_4/tf.compat.v1.gather_3/GatherV2/axis_0_unsqueeze0', 'model_4/tf.compat.v1.gather_3/GatherV2;model_4/tf.compat.v1.gather_3/GatherV2/axis_0_unsqueeze1'] D unsqueeze_to_4d_reducemean: remove node = [], add node = ['model_4/tf.math.reduce_mean_3/Mean_0_unsqueeze0', 'model_4/tf.math.reduce_mean_3/Mean_0_unsqueeze1'] D unsqueeze_to_4d_sub: remove node = [], add node = ['model_4/tf.math.subtract_9/Sub_0_unsqueeze0', 'model_4/tf.math.subtract_9/Sub_1_unsqueeze0', 'model_4/tf.math.subtract_9/Sub_0_unsqueeze1'] D unsqueeze_to_4d_split: remove node = [], add node = ['model_4/tf.unstack_3/unstack_0_unsqueeze0', 'model_4/tf.unstack_3/unstack_0_unsqueeze1', 'model_4/tf.unstack_3/unstack_1_unsqueeze1'] D convert_keepdims_to_1_reducemin: remove node = [], add node = ['model_4/tf.math.reduce_min_6/Min_reshape'] D convert_keepdims_to_1_reducemax: remove node = [], add node = ['model_4/tf.math.reduce_max_6/Max_reshape'] D convert_keepdims_to_1_reducemin: remove node = [], add node = ['model_4/tf.math.reduce_min_7/Min_reshape'] D convert_keepdims_to_1_reducemax: remove node = [], add node = ['model_4/tf.math.reduce_max_7/Max_reshape'] D unsqueeze_to_4d_div: remove node = [], add node = ['model_4/tf.math.truediv_3/truediv_0_unsqueeze0', 'model_4/tf.math.truediv_3/truediv_1_unsqueeze0', 'model_4/tf.math.truediv_3/truediv_0_unsqueeze1'] D convert_gemm_by_conv: remove node = ['model_4/dense_12/MatMul;model_4/dense_12/BiasAdd_Gemm32'], add node = ['model_4/dense_12/MatMul;model_4/dense_12/BiasAdd_Gemm__32_2conv_reshape1', 'model_4/dense_12/MatMul;model_4/dense_12/BiasAdd_Gemm32_2conv', 'model_4/dense_12/MatMul;model_4/dense_12/BiasAdd_Gemm32_2conv_reshape2'] D convert_gemm_by_conv: remove node = ['model_4/dense_12/MatMul;model_4/dense_12/BiasAdd;model_4/batch_normalization_10/batchnorm/mul_1;model_4/re_lu_10/Relu;model_4/batch_normalization_10/batchnorm/add_1_Gemm33'], add node = ['model_4/dense_12/MatMul;model_4/dense_12/BiasAdd;model_4/batch_normalization_10/batchnorm/mul_1;model_4/re_lu_10/Relu;model_4/batch_normalization_10/batchnorm/add_1_Gemm33_2conv_reshape1', 'model_4/dense_12/MatMul;model_4/dense_12/BiasAdd;model_4/batch_normalization_10/batchnorm/mul_1;model_4/re_lu_10/Relu;model_4/batch_normalization_10/batchnorm/add_1_Gemm__33_2conv', 'model_4/dense_12/MatMul;model_4/dense_12/BiasAdd;model_4/batch_normalization_10/batchnorm/mul_1;model_4/re_lu_10/Relu;model_4/batch_normalization_10/batchnorm/add_1_Gemm33_2conv_reshape2'] D unsqueeze_to_4d_relu: remove node = [], add node = ['Relu8_0_unsqueeze0', 'Relu8_0_unsqueeze1'] D convert_gemm_by_conv: remove node = ['model_4/dense_13/MatMul;model_4/dense_13/BiasAdd_Gemm34'], add node = ['model_4/dense_13/MatMul;model_4/dense_13/BiasAdd_Gemm__34_2conv_reshape1', 'model_4/dense_13/MatMul;model_4/dense_13/BiasAdd_Gemm34_2conv', 'model_4/dense_13/MatMul;model_4/dense_13/BiasAdd_Gemm34_2conv_reshape2'] D unsqueeze_to_4d_add: remove node = [], add node = ['model_4/tf.operators.add_13/AddV2_0_unsqueeze0', 'model_4/tf.operators.add_13/AddV2_1_unsqueeze0', 'model_4/tf.operators.add_13/AddV2_0_unsqueeze1'] D unsqueeze_to_4d_mul: remove node = [], add node = ['model_4/batch_normalization_11/batchnorm/mul_1_0_unsqueeze0', 'model_4/batch_normalization_11/batchnorm/mul_1_0_unsqueeze1'] D unsqueeze_to_4d_add: remove node = [], add node = ['model_4/re_lu_11/Relu;model_4/batch_normalization_11/batchnorm/add_1_0_unsqueeze0', 'model_4/re_lu_11/Relu;model_4/batch_normalization_11/batchnorm/add_1_0_unsqueeze1'] D unsqueeze_to_4d_relu: remove node = [], add node = ['Relu13_0_unsqueeze0', 'Relu13_0_unsqueeze1'] D convert_gemm_by_conv: remove node = ['model_4/dense_14/MatMul;model_4/dense_14/BiasAdd_Gemm35'], add node = ['model_4/dense_14/MatMul;model_4/dense_14/BiasAdd_Gemm35_2conv_reshape1', 'model_4/dense_14/MatMul;model_4/dense_14/BiasAdd_Gemm35_2conv', 'model_4/dense_14/MatMul;model_4/dense_14/BiasAdd_Gemm35_2conv_reshape2'] D unsqueeze_to_4d_add: remove node = [], add node = ['model_4/tf.operators.add_14/AddV2_0_unsqueeze0', 'model_4/tf.operators.add_14/AddV2_1_unsqueeze0', 'model_4/tf.operators.add_14/AddV2_0_unsqueeze1'] D unsqueeze_to_4d_mul: remove node = [], add node = ['model_4/batch_normalization_12/batchnorm/mul_1_0_unsqueeze0', 'model_4/batch_normalization_12/batchnorm/mul_1_0_unsqueeze1'] D unsqueeze_to_4d_add: remove node = [], add node = ['model_4/re_lu_12/Relu;model_4/batch_normalization_12/batchnorm/add_1_0_unsqueeze0', 'model_4/re_lu_12/Relu;model_4/batch_normalization_12/batchnorm/add_1_0_unsqueeze1'] D unsqueeze_to_4d_relu: remove node = [], add node = ['Relu16_0_unsqueeze0', 'Relu16_0_unsqueeze1'] D convert_gemm_by_conv: remove node = ['model_4/dense_15/MatMul;model_4/dense_15/BiasAdd_Gemm__36'], add node = ['model_4/dense_15/MatMul;model_4/dense_15/BiasAdd_Gemm36_2conv_reshape1', 'model_4/dense_15/MatMul;model_4/dense_15/BiasAdd_Gemm36_2conv', 'model_4/dense_15/MatMul;model_4/dense_15/BiasAdd_Gemm36_2conv_reshape2'] D unsqueeze_to_4d_add: remove node = [], add node = ['model_4/tf.operators.add_15/AddV2_0_unsqueeze0', 'model_4/tf.operators.add_15/AddV2_1_unsqueeze0', 'model_4/tf.operators.add_15/AddV2_0_unsqueeze1'] D unsqueeze_to_4d_mul: remove node = [], add node = ['model_4/batch_normalization_13/batchnorm/mul_1_0_unsqueeze0', 'model_4/batch_normalization_13/batchnorm/mul_1_0_unsqueeze1'] D unsqueeze_to_4d_add: remove node = [], add node = ['model_4/re_lu_13/Relu;model_4/batch_normalization_13/batchnorm/add_1_0_unsqueeze0', 'model_4/re_lu_13/Relu;model_4/batch_normalization_13/batchnorm/add_1_0_unsqueeze1'] D unsqueeze_to_4d_relu: remove node = [], add node = ['Relu19_0_unsqueeze0', 'Relu19_0_unsqueeze1'] D convert_gemm_by_conv: remove node = ['model_4/dense_16/MatMul;model_4/dense_16/BiasAdd_Gemm37'], add node = ['model_4/dense_16/MatMul;model_4/dense_16/BiasAdd_Gemm__37_2conv_reshape1', 'model_4/dense_16/MatMul;model_4/dense_16/BiasAdd_Gemm37_2conv', 'model_4/dense_16/MatMul;model_4/dense_16/BiasAdd_Gemm37_2conv_reshape2'] D unsqueeze_to_4d_add: remove node = [], add node = ['model_4/tf.operators.add_16/AddV2_0_unsqueeze0', 'model_4/tf.operators.add_16/AddV2_1_unsqueeze0', 'model_4/tf.operators.add_16/AddV2_0_unsqueeze1'] D unsqueeze_to_4d_mul: remove node = [], add node = ['model_4/batch_normalization_14/batchnorm/mul_1_0_unsqueeze0', 'model_4/batch_normalization_14/batchnorm/mul_1_0_unsqueeze1'] D unsqueeze_to_4d_add: remove node = [], add node = ['model_4/re_lu_14/Relu;model_4/batch_normalization_14/batchnorm/add_1_0_unsqueeze0', 'model_4/re_lu_14/Relu;model_4/batch_normalization_14/batchnorm/add_1_0_unsqueeze1'] D unsqueeze_to_4d_relu: remove node = [], add node = ['Relu22_0_unsqueeze0', 'Relu22_0_unsqueeze1'] D convert_gemm_by_conv: remove node = ['model_4/dense_17/MatMul;model_4/dense_17/BiasAdd_Gemm38'], add node = ['model_4/dense_17/MatMul;model_4/dense_17/BiasAdd_Gemm38_2conv_reshape1', 'model_4/dense_17/MatMul;model_4/dense_17/BiasAdd_Gemm38_2conv', 'model_4/dense_17/MatMul;model_4/dense_17/BiasAdd_Gemm38_2conv_reshape2'] D unsqueeze_to_4d_add: remove node = [], add node = ['model_4/tf.operators.add_17/AddV2_0_unsqueeze0', 'model_4/tf.operators.add_17/AddV2_1_unsqueeze0', 'model_4/tf.operators.add_17/AddV2_0_unsqueeze1'] D unsqueeze_to_4d_mul: remove node = [], add node = ['model_4/batch_normalization_15/batchnorm/mul_1_0_unsqueeze0', 'model_4/batch_normalization_15/batchnorm/mul_1_0_unsqueeze1'] D unsqueeze_to_4d_add: remove node = [], add node = ['model_4/re_lu_15/Relu;model_4/batch_normalization_15/batchnorm/add_1_0_unsqueeze0', 'model_4/re_lu_15/Relu;model_4/batch_normalization_15/batchnorm/add_1_0_unsqueeze1'] D unsqueeze_to_4d_relu: remove node = [], add node = ['Relu25_0_unsqueeze0', 'Relu25_0_unsqueeze1'] D convert_gemm_by_conv: remove node = ['model_4/dense_18/MatMul;model_4/dense_18/BiasAdd_Gemm__39'], add node = ['model_4/dense_18/MatMul;model_4/dense_18/BiasAdd_Gemm39_2conv_reshape1', 'model_4/dense_18/MatMul;model_4/dense_18/BiasAdd_Gemm39_2conv', 'model_4/dense_18/MatMul;model_4/dense_18/BiasAdd_Gemm39_2conv_reshape2'] D unsqueeze_to_4d_add: remove node = [], add node = ['model_4/tf.operators.add_18/AddV2_0_unsqueeze0', 'model_4/tf.operators.add_18/AddV2_1_unsqueeze0', 'model_4/tf.operators.add_18/AddV2_0_unsqueeze1'] D unsqueeze_to_4d_mul: remove node = [], add node = ['model_4/batch_normalization_16/batchnorm/mul_1_0_unsqueeze0', 'model_4/batch_normalization_16/batchnorm/mul_1_0_unsqueeze1'] D unsqueeze_to_4d_add: remove node = [], add node = ['model_4/re_lu_16/Relu;model_4/batch_normalization_16/batchnorm/add_1_0_unsqueeze0', 'model_4/re_lu_16/Relu;model_4/batch_normalization_16/batchnorm/add_1_0_unsqueeze1'] D unsqueeze_to_4d_relu: remove node = [], add node = ['Relu28_0_unsqueeze0', 'Relu28_0_unsqueeze1'] D convert_gemm_by_conv: remove node = ['Identity_Gemm40'], add node = ['Identity_Gemm__40_2conv_reshape1', 'Identity_Gemm40_2conv', 'Identity_Gemm40_2conv_reshape2'] D unsqueeze_to_4d_relu: remove node = [], add node = ['Relu29_0_unsqueeze0', 'Relu29_0_unsqueeze1'] D remove_parallel_reshape: remove node = ['model_4/tf.math.subtract_9/Sub_0_unsqueeze0'] D convert_sub_to_conv_sub: remove node = [], add node = ['model_4/tf.math.subtract_9/Sub_conv'] D remove_parallel_reshape: remove node = ['model_4/tf.unstack_3/unstack_0_unsqueeze0'] D fuse_two_reshape: remove node = ['model_4/tf.unstack_3/unstack_0_unsqueeze1'] D unsqueeze_to_4d_reducemin: remove node = [], add node = ['model_4/tf.math.reduce_min_6/Min_0_unsqueeze0', 'model_4/tf.math.reduce_min_6/Min_0_unsqueeze1'] D unsqueeze_to_4d_reducemax: remove node = [], add node = ['model_4/tf.math.reduce_max_6/Max_0_unsqueeze0', 'model_4/tf.math.reduce_max_6/Max_0_unsqueeze1'] D fuse_two_reshape: remove node = ['model_4/tf.unstack_3/unstack_1_unsqueeze1'] D unsqueeze_to_4d_reducemin: remove node = [], add node = ['model_4/tf.math.reduce_min_7/Min_0_unsqueeze0', 'model_4/tf.math.reduce_min_7/Min_0_unsqueeze1'] D unsqueeze_to_4d_reducemax: remove node = [], add node = ['model_4/tf.math.reduce_max_7/Max_0_unsqueeze0', 'model_4/tf.math.reduce_max_7/Max_0_unsqueeze1'] D fuse_two_reshape: remove node = ['model_4/tf.operators.getitem_3/strided_slice'] D convert_div_to_expand_div: remove node = [], add node = ['model_4/tf.math.truediv_3/truediv_broadcast1'] D fuse_two_reshape: remove node = ['model_4/tf.math.truediv_3/truediv_0_unsqueeze1'] D remove_parallel_reshape: remove node = ['model_4/dense_12/MatMul;model_4/dense_12/BiasAdd_Gemm__32_2conv_reshape1'] D fuse_two_reshape: remove node = ['model_4/dense_12/MatMul;model_4/dense_12/BiasAdd_Gemm32_2conv_reshape2'] D remove_parallel_reshape: remove node = ['model_4/batch_normalization_11/batchnorm/mul_1_0_unsqueeze0'] D input_align_4D_mul: remove node = ['model_4/batch_normalization_11/batchnorm/mul_1'], add node = ['model_4/batch_normalization_11/batchnorm/mul_1'] D fuse_two_reshape: remove node = ['model_4/batch_normalization_11/batchnorm/mul_1_0_unsqueeze1'] D input_align_4D_add: remove node = ['model_4/re_lu_11/Relu;model_4/batch_normalization_11/batchnorm/add_1'], add node = ['model_4/re_lu_11/Relu;model_4/batch_normalization_11/batchnorm/add_1'] D fuse_two_reshape: remove node = ['model_4/re_lu_11/Relu;model_4/batch_normalization_11/batchnorm/add_1_0_unsqueeze1', 'Relu13_0_unsqueeze1', 'model_4/dense_14/MatMul;model_4/dense_14/BiasAdd_Gemm__35_2conv_reshape2'] D remove_parallel_reshape: remove node = ['model_4/batch_normalization_12/batchnorm/mul_1_0_unsqueeze0'] D input_align_4D_mul: remove node = ['model_4/batch_normalization_12/batchnorm/mul_1'], add node = ['model_4/batch_normalization_12/batchnorm/mul_1'] D fuse_two_reshape: remove node = ['model_4/batch_normalization_12/batchnorm/mul_1_0_unsqueeze1'] D input_align_4D_add: remove node = ['model_4/re_lu_12/Relu;model_4/batch_normalization_12/batchnorm/add_1'], add node = ['model_4/re_lu_12/Relu;model_4/batch_normalization_12/batchnorm/add_1'] D fuse_two_reshape: remove node = ['model_4/re_lu_12/Relu;model_4/batch_normalization_12/batchnorm/add_1_0_unsqueeze1', 'Relu16_0_unsqueeze1', 'model_4/dense_15/MatMul;model_4/dense_15/BiasAdd_Gemm36_2conv_reshape2'] D remove_parallel_reshape: remove node = ['model_4/batch_normalization_13/batchnorm/mul_1_0_unsqueeze0'] D input_align_4D_mul: remove node = ['model_4/batch_normalization_13/batchnorm/mul_1'], add node = ['model_4/batch_normalization_13/batchnorm/mul_1'] D fuse_two_reshape: remove node = ['model_4/batch_normalization_13/batchnorm/mul_1_0_unsqueeze1'] D input_align_4D_add: remove node = ['model_4/re_lu_13/Relu;model_4/batch_normalization_13/batchnorm/add_1'], add node = ['model_4/re_lu_13/Relu;model_4/batch_normalization_13/batchnorm/add_1'] D fuse_two_reshape: remove node = ['model_4/re_lu_13/Relu;model_4/batch_normalization_13/batchnorm/add_1_0_unsqueeze1', 'Relu__19_0_unsqueeze1', 'model_4/dense_16/MatMul;model_4/dense_16/BiasAdd_Gemm37_2conv_reshape2'] D remove_parallel_reshape: remove node = ['model_4/batch_normalization_14/batchnorm/mul_1_0_unsqueeze0'] D input_align_4D_mul: remove node = ['model_4/batch_normalization_14/batchnorm/mul_1'], add node = ['model_4/batch_normalization_14/batchnorm/mul_1'] D fuse_two_reshape: remove node = ['model_4/batch_normalization_14/batchnorm/mul_1_0_unsqueeze1'] D input_align_4D_add: remove node = ['model_4/re_lu_14/Relu;model_4/batch_normalization_14/batchnorm/add_1'], add node = ['model_4/re_lu_14/Relu;model_4/batch_normalization_14/batchnorm/add_1'] D fuse_two_reshape: remove node = ['model_4/re_lu_14/Relu;model_4/batch_normalization_14/batchnorm/add_1_0_unsqueeze1', 'Relu22_0_unsqueeze1', 'model_4/dense_17/MatMul;model_4/dense_17/BiasAdd_Gemm__38_2conv_reshape2'] D remove_parallel_reshape: remove node = ['model_4/batch_normalization_15/batchnorm/mul_1_0_unsqueeze0'] D input_align_4D_mul: remove node = ['model_4/batch_normalization_15/batchnorm/mul_1'], add node = ['model_4/batch_normalization_15/batchnorm/mul_1'] D fuse_two_reshape: remove node = ['model_4/batch_normalization_15/batchnorm/mul_1_0_unsqueeze1'] D input_align_4D_add: remove node = ['model_4/re_lu_15/Relu;model_4/batch_normalization_15/batchnorm/add_1'], add node = ['model_4/re_lu_15/Relu;model_4/batch_normalization_15/batchnorm/add_1'] D fuse_two_reshape: remove node = ['model_4/re_lu_15/Relu;model_4/batch_normalization_15/batchnorm/add_1_0_unsqueeze1', 'Relu25_0_unsqueeze1', 'model_4/dense_18/MatMul;model_4/dense_18/BiasAdd_Gemm39_2conv_reshape2', 'model_4/tf.operators.add_18/AddV2_0_unsqueeze1'] D input_align_4D_mul: remove node = ['model_4/batch_normalization_16/batchnorm/mul_1'], add node = ['model_4/batch_normalization_16/batchnorm/mul_1'] D fuse_two_reshape: remove node = ['model_4/batch_normalization_16/batchnorm/mul_1_0_unsqueeze1'] D input_align_4D_add: remove node = ['model_4/re_lu_16/Relu;model_4/batch_normalization_16/batchnorm/add_1'], add node = ['model_4/re_lu_16/Relu;model_4/batch_normalization_16/batchnorm/add_1'] D fuse_two_reshape: remove node = ['model_4/re_lu_16/Relu;model_4/batch_normalization_16/batchnorm/add_1_0_unsqueeze1', 'Relu__28_0_unsqueeze1', 'Identity_Gemm40_2conv_reshape2'] D fuse_two_reshape: remove node = ['model_4/tf.compat.v1.gather_3/GatherV2;model_4/tf.compat.v1.gather_3/GatherV2/axis_0_unsqueeze1'] D convert_reduce_mean_to_avgpool: remove node = ['model_4/tf.math.reduce_mean_3/Mean'], add node = ['model_4/tf.math.reduce_mean_3/Mean_2avgpool'] D fuse_two_reshape: remove node = ['model_4/tf.math.reduce_mean_3/Mean_0_unsqueeze1', 'model_4/tf.math.subtract_9/Sub_0_unsqueeze1'] D remove_parallel_reshape: remove node = ['model_4/tf.math.reduce_min_6/Min_0_unsqueeze0'] D fuse_two_reshape: remove node = ['model_4/tf.math.reduce_min_6/Min_0_unsqueeze1'] D remove_parallel_reshape: remove node = ['model_4/tf.math.reduce_min_7/Min_0_unsqueeze0'] D fuse_two_reshape: remove node = ['model_4/tf.math.reduce_min_7/Min_0_unsqueeze1', 'model_4/reshape_1/Reshape', 'model_4/dense_12/MatMul;model_4/dense_12/BiasAdd;model_4/batch_normalization_10/batchnorm/mul_1;model_4/re_lu_10/Relu;model_4/batch_normalization_10/batchnorm/add_1_Gemm33_2conv_reshape2', 'Relu8_0_unsqueeze1', 'model_4/dense_13/MatMul;model_4/dense_13/BiasAdd_Gemm34_2conv_reshape2'] D remove_invalid_reshape: remove node = ['model_4/tf.operators.add_13/AddV2_0_unsqueeze0'] D fuse_two_reshape: remove node = ['model_4/tf.operators.add_13/AddV2_0_unsqueeze1'] D remove_invalid_reshape: remove node = ['model_4/re_lu_11/Relu;model_4/batch_normalization_11/batchnorm/add_1_0_unsqueeze0', 'Relu__13_0_unsqueeze0', 'model_4/dense_14/MatMul;model_4/dense_14/BiasAdd_Gemm35_2conv_reshape1', 'model_4/tf.operators.add_14/AddV2_1_unsqueeze0'] D fuse_two_reshape: remove node = ['model_4/tf.operators.add_14/AddV2_0_unsqueeze1'] D remove_invalid_reshape: remove node = ['model_4/re_lu_12/Relu;model_4/batch_normalization_12/batchnorm/add_1_0_unsqueeze0', 'Relu16_0_unsqueeze0', 'model_4/dense_15/MatMul;model_4/dense_15/BiasAdd_Gemm__36_2conv_reshape1', 'model_4/tf.operators.add_15/AddV2_1_unsqueeze0'] D fuse_two_reshape: remove node = ['model_4/tf.operators.add_15/AddV2_0_unsqueeze1'] D remove_invalid_reshape: remove node = ['model_4/re_lu_13/Relu;model_4/batch_normalization_13/batchnorm/add_1_0_unsqueeze0', 'Relu19_0_unsqueeze0', 'model_4/dense_16/MatMul;model_4/dense_16/BiasAdd_Gemm37_2conv_reshape1', 'model_4/tf.operators.add_16/AddV2_1_unsqueeze0'] D fuse_two_reshape: remove node = ['model_4/tf.operators.add_16/AddV2_0_unsqueeze1'] D remove_invalid_reshape: remove node = ['model_4/re_lu_14/Relu;model_4/batch_normalization_14/batchnorm/add_1_0_unsqueeze0', 'Relu__22_0_unsqueeze0', 'model_4/dense_17/MatMul;model_4/dense_17/BiasAdd_Gemm38_2conv_reshape1', 'model_4/tf.operators.add_17/AddV2_1_unsqueeze0'] D fuse_two_reshape: remove node = ['model_4/tf.operators.add_17/AddV2_0_unsqueeze1'] D remove_invalid_reshape: remove node = ['model_4/re_lu_15/Relu;model_4/batch_normalization_15/batchnorm/add_1_0_unsqueeze0', 'Relu25_0_unsqueeze0', 'model_4/dense_18/MatMul;model_4/dense_18/BiasAdd_Gemm__39_2conv_reshape1', 'model_4/tf.operators.add_18/AddV2_1_unsqueeze0', 'model_4/batch_normalization_16/batchnorm/mul_1_0_unsqueeze0', 'model_4/re_lu_16/Relu;model_4/batch_normalization_16/batchnorm/add_1_0_unsqueeze0', 'Relu28_0_unsqueeze0', 'Identity_Gemm40_2conv_reshape1', 'Relu29_0_unsqueeze0'] D remove_invalid_reshape: remove node = ['model_4/tf.math.reduce_mean_3/Mean_0_unsqueeze0', 'model_4/tf.math.subtract_9/Sub_1_unsqueeze0', 'model_4/tf.math.truediv_3/truediv_0_unsqueeze0'] D fuse_two_reshape: remove node = ['model_4/tf.unstack_3/unstack41_2reshape', 'model_4/tf.math.reduce_max_6/Max_0_unsqueeze1', 'model_4/tf.unstack_3/unstack42_2reshape', 'model_4/tf.math.reduce_max_7/Max_0_unsqueeze1'] D remove_invalid_reshape: remove node = ['Relu8_0_unsqueeze0', 'model_4/dense_13/MatMul;model_4/dense_13/BiasAdd_Gemm__34_2conv_reshape1', 'model_4/tf.operators.add_13/AddV2_1_unsqueeze0', 'model_4/tf.operators.add_14/AddV2_0_unsqueeze0', 'model_4/tf.operators.add_15/AddV2_0_unsqueeze0', 'model_4/tf.operators.add_16/AddV2_0_unsqueeze0', 'model_4/tf.operators.add_17/AddV2_0_unsqueeze0', 'model_4/tf.operators__.add_18/AddV2_0_unsqueeze0'] D remove_invalid_reshape: remove node = ['model_4/tf.math.reduce_max_6/Max_0_unsqueeze0', 'model_4/tf.math.reduce_max_7/Max_0_unsqueeze0'] D replace_mul_add_by_bn: remove node = ['model_4/batch_normalization_11/batchnorm/mul_1', 'model_4/re_lu_11/Relu;model_4/batch_normalization_11/batchnorm/add_1'], add node = ['model_4/re_lu_11/Relu;model_4/batch_normalization_11/batchnorm/add_1_bn'] D replace_mul_add_by_bn: remove node = ['model_4/batch_normalization_12/batchnorm/mul_1', 'model_4/re_lu_12/Relu;model_4/batch_normalization_12/batchnorm/add_1'], add node = ['model_4/re_lu_12/Relu;model_4/batch_normalization_12/batchnorm/add_1_bn'] D replace_mul_add_by_bn: remove node = ['model_4/batch_normalization_13/batchnorm/mul_1', 'model_4/re_lu_13/Relu;model_4/batch_normalization_13/batchnorm/add_1'], add node = ['model_4/re_lu_13/Relu;model_4/batch_normalization_13/batchnorm/add_1_bn'] D replace_mul_add_by_bn: remove node = ['model_4/batch_normalization_14/batchnorm/mul_1', 'model_4/re_lu_14/Relu;model_4/batch_normalization_14/batchnorm/add_1'], add node = ['model_4/re_lu_14/Relu;model_4/batch_normalization_14/batchnorm/add_1_bn'] D replace_mul_add_by_bn: remove node = ['model_4/batch_normalization_15/batchnorm/mul_1', 'model_4/re_lu_15/Relu;model_4/batch_normalization_15/batchnorm/add_1'], add node = ['model_4/re_lu_15/Relu;model_4/batch_normalization_15/batchnorm/add_1_bn'] D replace_mul_add_by_bn: remove node = ['model_4/batch_normalization_16/batchnorm/mul_1', 'model_4/re_lu_16/Relu;model_4/batch_normalization_16/batchnorm/add_1'], add node = ['model_4/re_lu_16/Relu;model_4/batch_normalization_16/batchnorm/add_1_bn'] D fold_constant ... D fold_constant done. D fuse_ops done. D D limit model tensor name to 128! D D sparse_weight ... D sparse_weight done. D I rknn building ... I RKNN: [16:54:56.433] compress = 0, conv_eltwise_activation_fuse = 1, global_fuse = 1, multi-core-model-mode = 7, output_optimize = 1, layout_match = 1, enable_argb_group = 0 I RKNN: librknnc version: 2.0.0b0 (35a6907d79@2024-03-24T02:34:11) D RKNN: [16:54:56.436] RKNN is invoked D RKNN: [16:54:56.448] >>>>>> start: rknn::RKNNExtractCustomOpAttrs D RKNN: [16:54:56.448] <<<<<<<< end: rknn::RKNNExtractCustomOpAttrs D RKNN: [16:54:56.448] >>>>>> start: rknn::RKNNSetOpTargetPass D RKNN: [16:54:56.448] <<<<<<<< end: rknn::RKNNSetOpTargetPass D RKNN: [16:54:56.448] >>>>>> start: rknn::RKNNBindNorm D RKNN: [16:54:56.448] <<<<<<<< end: rknn::RKNNBindNorm D RKNN: [16:54:56.448] >>>>>> start: rknn::RKNNAddFirstConv D RKNN: [16:54:56.448] <<<<<<<< end: rknn::RKNNAddFirstConv D RKNN: [16:54:56.448] >>>>>> start: rknn::RKNNEliminateQATDataConvert D RKNN: [16:54:56.448] <<<<<<<< end: rknn::RKNNEliminateQATDataConvert D RKNN: [16:54:56.448] >>>>>> start: rknn::RKNNTileGroupConv D RKNN: [16:54:56.448] <<<<<<<< end: rknn::RKNNTileGroupConv D RKNN: [16:54:56.448] >>>>>> start: rknn::RKNNAddConvBias D RKNN: [16:54:56.448] <<<<<<<< end: rknn::RKNNAddConvBias D RKNN: [16:54:56.448] >>>>>> start: rknn::RKNNTileChannel D RKNN: [16:54:56.448] <<<<<<<< end: rknn::RKNNTileChannel D RKNN: [16:54:56.448] >>>>>> start: rknn::RKNNPerChannelPrep D RKNN: [16:54:56.448] <<<<<<<< end: rknn::RKNNPerChannelPrep D RKNN: [16:54:56.448] >>>>>> start: rknn::RKNNBnQuant D RKNN: [16:54:56.448] <<<<<<<< end: rknn::RKNNBnQuant D RKNN: [16:54:56.448] >>>>>> start: rknn::RKNNFuseOptimizerPass D RKNN: [16:54:56.449] <<<<<<<< end: rknn::RKNNFuseOptimizerPass D RKNN: [16:54:56.449] >>>>>> start: rknn::RKNNTurnAutoPad D RKNN: [16:54:56.449] <<<<<<<< end: rknn::RKNNTurnAutoPad D RKNN: [16:54:56.449] >>>>>> start: rknn::RKNNInitRNNConst D RKNN: [16:54:56.449] <<<<<<<< end: rknn::RKNNInitRNNConst D RKNN: [16:54:56.449] >>>>>> start: rknn::RKNNInitCastConst D RKNN: [16:54:56.449] <<<<<<<< end: rknn::RKNNInitCastConst D RKNN: [16:54:56.449] >>>>>> start: rknn::RKNNMultiSurfacePass D RKNN: [16:54:56.449] <<<<<<<< end: rknn::RKNNMultiSurfacePass D RKNN: [16:54:56.449] >>>>>> start: rknn::RKNNReplaceConstantTensorPass D RKNN: [16:54:56.449] <<<<<<<< end: rknn::RKNNReplaceConstantTensorPass D RKNN: [16:54:56.449] >>>>>> start: rknn::RKNNTilingPass D RKNN: [16:54:56.449] <<<<<<<< end: rknn::RKNNTilingPass D RKNN: [16:54:56.449] >>>>>> start: rknn::RKNNSubgraphManager D RKNN: [16:54:56.449] <<<<<<<< end: rknn::RKNNSubgraphManager D RKNN: [16:54:56.449] >>>>>> start: OpEmit D RKNN: [16:54:56.449] <<<<<<<< end: OpEmit D RKNN: [16:54:56.449] >>>>>> start: rknn::RKNNLayoutMatchPass I RKNN: [16:54:56.450] AppointLayout: t->setNativeLayout(64), tname:[model_4/tf.math.reduce_mean_3/Mean] I RKNN: [16:54:56.450] AppointLayout: t->setNativeLayout(64), tname:[model_4/tf.math.subtract_9/Sub_conv_out] I RKNN: [16:54:56.450] AppointLayout: t->setNativeLayout(64), tname:[model_4/tf.math.subtract_9/Sub] I RKNN: [16:54:56.450] AppointLayout: t->setNativeLayout(64), tname:[Relu__8:0_shape4] I RKNN: [16:54:56.450] AppointLayout: t->setNativeLayout(64), tname:[Identity_shape4] I RKNN: [16:54:56.450] AppointLayout: t->setNativeLayout(0), tname:[Identity] D RKNN: [16:54:56.450] <<<<<<<< end: rknn::RKNNLayoutMatchPass D RKNN: [16:54:56.450] >>>>>> start: rknn::RKNNAddSecondaryNode D RKNN: [16:54:56.450] <<<<<<<< end: rknn::RKNNAddSecondaryNode D RKNN: [16:54:56.450] >>>>>> start: OpEmit W RKNN: [16:54:56.450] Meet unsupported sub operator with per-layer parameter is a feature for rknpu! W RKNN: [16:54:56.450] Meet unsupported sub operator with per-layer parameter is a feature for rknpu! Segmentation fault (core dumped)

logs listed here

airockchip / rknn-toolkit2

Segmentation fault (core dumped) when converting to rknn #81