rknn.load_onnx() Error - Githubissues

Hello guys, I met an error when loading an onnx model. My input shape is (1, 61), so I set the input_size_list to [[61]] (ignoring the batch size according to user guide). Is there any problem here, since I have not seen any example model which accepts a 1D input.

D Using CPPUTILS: True I Start importing onnx... W Call onnx.optimizer.optimize fail, skip optimize I Current ONNX Model use ir_version 6 opset_version 11 I Call RKNN onnx optimize fail, skip optimize D Infer onnx shape: Meet empty shape tensor, reshape () to (1,)! D Calc tensor Constant_209 (1,) D Infer onnx shape: Meet empty shape tensor, reshape () to (1,)! D Calc tensor Constant_195 (1,) D Infer onnx shape: Meet empty shape tensor, reshape () to (1,)! D Calc tensor Constant_192 (1,) D Infer onnx shape: Meet empty shape tensor, reshape () to (1,)! D Calc tensor Constant_178 (1,) D Infer onnx shape: Meet empty shape tensor, reshape () to (1,)! D Calc tensor Constant_175 (1,) D Infer onnx shape: Meet empty shape tensor, reshape () to (1,)! D Calc tensor Constant_160 (1,) D Infer onnx shape: Meet empty shape tensor, reshape () to (1,)! D Calc tensor Constant_157 (1,) D Infer onnx shape: Meet empty shape tensor, reshape () to (1,)! D Calc tensor Constant_143 (1,) D Infer onnx shape: Meet empty shape tensor, reshape () to (1,)! D Calc tensor Constant_140 (1,) D Infer onnx shape: Meet empty shape tensor, reshape () to (1,)! D Calc tensor Constant_126 (1,) D Infer onnx shape: Meet empty shape tensor, reshape () to (1,)! D Calc tensor Constant_123 (1,) D Infer onnx shape: Meet empty shape tensor, reshape () to (1,)! D Calc tensor Constant_109 (1,) D Infer onnx shape: Meet empty shape tensor, reshape () to (1,)! D Calc tensor Constant_106 (1,) D Infer onnx shape: Meet empty shape tensor, reshape () to (1,)! D Calc tensor Constant_97 (1,) D Infer onnx shape: Meet empty shape tensor, reshape () to (1,)! D Calc tensor Constant_242 (1,) D Infer onnx shape: Meet empty shape tensor, reshape () to (1,)! D Calc tensor Constant_239 (1,) D Infer onnx shape: Meet empty shape tensor, reshape () to (1,)! D Calc tensor Constant_225 (1,) D Infer onnx shape: Meet empty shape tensor, reshape () to (1,)! D Calc tensor Constant_222 (1,) D Calc tensor Initializer_261 (256, 1) D Infer onnx shape: Meet empty shape tensor, reshape () to (1,)! D Calc tensor Initializer_260 (1,) D Infer onnx shape: Meet empty shape tensor, reshape () to (1,)! D Calc tensor Initializer_259 (1,) D Calc tensor Initializer_258 (256, 1) D Calc tensor Initializer_pitch_emb.bias (256,) D Calc tensor Initializer_pitch_emb.weight (256, 1, 3) D Calc tensor Initializer_pitch_predictor.fc.bias (1,) D Calc tensor Initializer_pitch_predictor.layers.1.norm.bias (256,) D Calc tensor Initializer_pitch_predictor.layers.1.norm.weight (256,) D Calc tensor Initializer_pitch_predictor.layers.1.conv.pointwise.bias (256,) D Calc tensor Initializer_pitch_predictor.layers.1.conv.pointwise.weight (256, 256, 1) D Calc tensor Initializer_pitch_predictor.layers.1.conv.depthwise.bias (256,) D Calc tensor Initializer_pitch_predictor.layers.1.conv.depthwise.weight (256, 1, 3) D Calc tensor Initializer_pitch_predictor.layers.0.norm.bias (256,) D Calc tensor Initializer_pitch_predictor.layers.0.norm.weight (256,) D Calc tensor Initializer_pitch_predictor.layers.0.conv.pointwise.bias (256,) D Calc tensor Initializer_pitch_predictor.layers.0.conv.pointwise.weight (256, 256, 1) D Calc tensor Initializer_pitch_predictor.layers.0.conv.depthwise.bias (256,) D Calc tensor Initializer_pitch_predictor.layers.0.conv.depthwise.weight (256, 1, 3) D Calc tensor Initializer_duration_predictor.fc.bias (1,) D Calc tensor Initializer_duration_predictor.layers.1.norm.bias (256,) D Calc tensor Initializer_duration_predictor.layers.1.norm.weight (256,) D Calc tensor Initializer_duration_predictor.layers.1.conv.pointwise.bias (256,) D Calc tensor Initializer_duration_predictor.layers.1.conv.pointwise.weight (256, 256, 1) D Calc tensor Initializer_duration_predictor.layers.1.conv.depthwise.bias (256,) D Calc tensor Initializer_duration_predictor.layers.1.conv.depthwise.weight (256, 1, 3) D Calc tensor Initializer_duration_predictor.layers.0.norm.bias (256,) D Calc tensor Initializer_duration_predictor.layers.0.norm.weight (256,) D Calc tensor Initializer_duration_predictor.layers.0.conv.pointwise.bias (256,) D Calc tensor Initializer_duration_predictor.layers.0.conv.pointwise.weight (256, 256, 1) D Calc tensor Initializer_duration_predictor.layers.0.conv.depthwise.bias (256,) D Calc tensor Initializer_duration_predictor.layers.0.conv.depthwise.weight (256, 1, 3) D Calc tensor Initializer_encoder.3.norm.bias (256,) D Calc tensor Initializer_encoder.3.norm.weight (256,) D Calc tensor Initializer_encoder.3.conv.pointwise.bias (256,) D Calc tensor Initializer_encoder.3.conv.pointwise.weight (256, 256, 1) D Calc tensor Initializer_encoder.3.conv.depthwise.bias (256,) D Calc tensor Initializer_encoder.3.conv.depthwise.weight (256, 1, 9) D Calc tensor Initializer_encoder.2.norm.bias (256,) D Calc tensor Initializer_encoder.2.norm.weight (256,) D Calc tensor Initializer_encoder.2.conv.pointwise.bias (256,) D Calc tensor Initializer_encoder.2.conv.pointwise.weight (256, 256, 1) D Calc tensor Initializer_encoder.2.conv.depthwise.bias (256,) D Calc tensor Initializer_encoder.2.conv.depthwise.weight (256, 1, 13) D Calc tensor Initializer_encoder.1.norm.bias (256,) D Calc tensor Initializer_encoder.1.norm.weight (256,) D Calc tensor Initializer_encoder.1.conv.pointwise.bias (256,) D Calc tensor Initializer_encoder.1.conv.pointwise.weight (256, 256, 1) D Calc tensor Initializer_encoder.1.conv.depthwise.bias (256,) D Calc tensor Initializer_encoder.1.conv.depthwise.weight (256, 1, 25) D Calc tensor Initializer_encoder.0.norm.bias (256,) D Calc tensor Initializer_encoder.0.norm.weight (256,) D Calc tensor Initializer_encoder.0.conv.pointwise.bias (256,) D Calc tensor Initializer_encoder.0.conv.pointwise.weight (256, 384, 1) D Calc tensor Initializer_encoder.0.conv.depthwise.bias (384,) D Calc tensor Initializer_encoder.0.conv.depthwise.weight (384, 1, 5) D Calc tensor Initializer_word_emb.weight (81, 384) D Calc tensor Gather_96 (1, 61, 384) D Calc tensor Add_98 (1, 61, 384) D Calc tensor Transpose_99 (1, 384, 61) D Calc tensor Conv_100 (1, 384, 61) D Calc tensor Conv_101 (1, 256, 61) D Calc tensor Relu_102 (1, 256, 61) D Calc tensor Transpose_103 (1, 61, 256) D Calc tensor ReduceMean_104 (1, 61, 1) D Calc tensor Sub_105 (1, 61, 256) D Calc tensor Pow_107 (1, 61, 256) D Calc tensor ReduceMean_108 (1, 61, 1) D Calc tensor Add_110 (1, 61, 1) D Calc tensor Sqrt_111 (1, 61, 1) D Calc tensor Div_112 (1, 61, 256) D Calc tensor Mul_113 (1, 61, 256) D Calc tensor Add_114 (1, 61, 256) D Calc tensor Transpose_115 (1, 256, 61) D Calc tensor Cast_116 (1, 256, 61) D Calc tensor Conv_117 (1, 256, 61) D Calc tensor Conv_118 (1, 256, 61) D Calc tensor Relu_119 (1, 256, 61) D Calc tensor Transpose_120 (1, 61, 256) D Calc tensor ReduceMean_121 (1, 61, 1) D Calc tensor Sub_122 (1, 61, 256) D Calc tensor Pow_124 (1, 61, 256) D Calc tensor ReduceMean_125 (1, 61, 1) D Calc tensor Add_127 (1, 61, 1) D Calc tensor Sqrt_128 (1, 61, 1) D Calc tensor Div_129 (1, 61, 256) D Calc tensor Mul_130 (1, 61, 256) D Calc tensor Add_131 (1, 61, 256) D Calc tensor Transpose_132 (1, 256, 61) D Calc tensor Cast_133 (1, 256, 61) D Calc tensor Conv_134 (1, 256, 61) D Calc tensor Conv_135 (1, 256, 61) D Calc tensor Relu_136 (1, 256, 61) D Calc tensor Transpose_137 (1, 61, 256) D Calc tensor ReduceMean_138 (1, 61, 1) D Calc tensor Sub_139 (1, 61, 256) D Calc tensor Pow_141 (1, 61, 256) D Calc tensor ReduceMean_142 (1, 61, 1) D Calc tensor Add_144 (1, 61, 1) D Calc tensor Sqrt_145 (1, 61, 1) D Calc tensor Div_146 (1, 61, 256) D Calc tensor Mul_147 (1, 61, 256) D Calc tensor Add_148 (1, 61, 256) D Calc tensor Transpose_149 (1, 256, 61) D Calc tensor Cast_150 (1, 256, 61) D Calc tensor Conv_151 (1, 256, 61) D Calc tensor Conv_152 (1, 256, 61) D Calc tensor Relu_153 (1, 256, 61) D Calc tensor Transpose_154 (1, 61, 256) D Calc tensor ReduceMean_155 (1, 61, 1) D Calc tensor Sub_156 (1, 61, 256) D Calc tensor Pow_158 (1, 61, 256) D Calc tensor ReduceMean_159 (1, 61, 1) D Calc tensor Add_161 (1, 61, 1) D Calc tensor Sqrt_162 (1, 61, 1) D Calc tensor Div_163 (1, 61, 256) D Calc tensor Mul_164 (1, 61, 256) D Calc tensor Add_165 (1, 61, 256) D Calc tensor Transpose_166 (1, 256, 61) D Calc tensor Cast_167 (1, 256, 61) D Calc tensor Transpose_168 (1, 61, 256) D Calc tensor Conv_216 (1, 256, 61) D Calc tensor Conv_217 (1, 256, 61) D Calc tensor Relu_218 (1, 256, 61) D Calc tensor Transpose_219 (1, 61, 256) D Calc tensor ReduceMean_220 (1, 61, 1) D Calc tensor Sub_221 (1, 61, 256) D Calc tensor Pow_223 (1, 61, 256) D Calc tensor ReduceMean_224 (1, 61, 1) D Calc tensor Add_226 (1, 61, 1) D Calc tensor Sqrt_227 (1, 61, 1) D Calc tensor Div_228 (1, 61, 256) D Calc tensor Mul_229 (1, 61, 256) D Calc tensor Add_230 (1, 61, 256) D Calc tensor Transpose_231 (1, 256, 61) D Calc tensor Cast_232 (1, 256, 61) D Calc tensor Conv_233 (1, 256, 61) D Calc tensor Conv_234 (1, 256, 61) D Calc tensor Relu_235 (1, 256, 61) D Calc tensor Transpose_236 (1, 61, 256) D Calc tensor ReduceMean_237 (1, 61, 1) D Calc tensor Sub_238 (1, 61, 256) D Calc tensor Pow_240 (1, 61, 256) D Calc tensor ReduceMean_241 (1, 61, 1) D Calc tensor Add_243 (1, 61, 1) D Calc tensor Sqrt_244 (1, 61, 1) D Calc tensor Div_245 (1, 61, 256) D Calc tensor Mul_246 (1, 61, 256) D Calc tensor Add_247 (1, 61, 256) D Calc tensor Transpose_248 (1, 256, 61) D Calc tensor Cast_249 (1, 256, 61) D Calc tensor Transpose_250 (1, 61, 256) D Calc tensor MatMul_252 (1, 61, 1) D Calc tensor Add_253 (1, 61, 1) D Calc tensor Transpose_254 (1, 1, 61) D Calc tensor Conv_255 (1, 256, 61) D Calc tensor Transpose_256 (1, 61, 256) D Calc tensor Add_enc_out (1, 61, 256) D Calc tensor Conv_169 (1, 256, 61) D Calc tensor Conv_170 (1, 256, 61) D Calc tensor Relu_171 (1, 256, 61) D Calc tensor Transpose_172 (1, 61, 256) D Calc tensor ReduceMean_173 (1, 61, 1) D Calc tensor Sub_174 (1, 61, 256) D Calc tensor Pow_176 (1, 61, 256) D Calc tensor ReduceMean_177 (1, 61, 1) D Calc tensor Add_179 (1, 61, 1) D Calc tensor Sqrt_180 (1, 61, 1) D Calc tensor Div_181 (1, 61, 256) D Calc tensor Mul_182 (1, 61, 256) D Calc tensor Add_183 (1, 61, 256) D Calc tensor Transpose_184 (1, 256, 61) D Calc tensor Cast_185 (1, 256, 61) D Calc tensor Conv_186 (1, 256, 61) D Calc tensor Conv_187 (1, 256, 61) D Calc tensor Relu_188 (1, 256, 61) D Calc tensor Transpose_189 (1, 61, 256) D Calc tensor ReduceMean_190 (1, 61, 1) D Calc tensor Sub_191 (1, 61, 256) D Calc tensor Pow_193 (1, 61, 256) D Calc tensor ReduceMean_194 (1, 61, 1) D Calc tensor Add_196 (1, 61, 1) D Calc tensor Sqrt_197 (1, 61, 1) D Calc tensor Div_198 (1, 61, 256) D Calc tensor Mul_199 (1, 61, 256) D Calc tensor Add_200 (1, 61, 256) D Calc tensor Transpose_201 (1, 256, 61) D Calc tensor Cast_202 (1, 256, 61) D Calc tensor Transpose_203 (1, 61, 256) D Calc tensor MatMul_205 (1, 61, 1) D Calc tensor Add_206 (1, 61, 1) D Calc tensor Squeeze_207 (1, 61) D Calc tensor Exp_208 (1, 61) D Calc tensor Sub_210 (1, 61) D Calc tensor Clip_dur_pred (1, 61) D import clients finished I build output layer attach_Add_Add_155:out0 I build output layer attach_Clip_Clip_114:out0 I Try match layer: Clip_Clip_114:out0 I Match r_clip_11, convert ONNX layer: ['Clip_Clip_114', 'Initializer_259', 'Initializer_260'](OPs: ['Clip', 'Constant', 'Constant_1']) to ['clipbyvalue'] I Try match layer: Add_Add_155:out0 I Match r_add, convert ONNX layer: ['Add_Add_155'](OPs: ['Add']) to ['add'] I Try match layer: Sub_Sub_113:out0 I Match r_sub, convert ONNX layer: ['Sub_Sub_113'](OPs: ['Sub']) to ['subtract'] I Try match layer: Transpose_Transpose_72:out0 I Match r_transpose, convert ONNX layer: ['Transpose_Transpose_72'](OPs: ['Transpose']) to ['permute'] I Try match layer: Transpose_Transpose_154:out0 I Match r_transpose, convert ONNX layer: ['Transpose_Transpose_154'](OPs: ['Transpose']) to ['permute'] I Try match layer: Exp_Exp_111:out0 I Match r_exp, convert ONNX layer: ['Exp_Exp_111'](OPs: ['Exp']) to ['exp'] I Try match layer: Constant_Constant_112:out0 I Match r_variable, convert ONNX layer: ['Constant_Constant_112'](OPs: ['Constant']) to ['variable'] I Try match layer: Cast_Cast_71:out0 I Match r_cast, convert ONNX layer: ['Cast_Cast_71'](OPs: ['Cast']) to ['cast'] I Try match layer: Conv_Conv_153:out0 I Match r_conv1d, convert ONNX layer: ['Conv_Conv_153', 'Initializer_pitch_emb.weight', 'Initializer_pitch_emb.bias'](OPs: ['Conv', 'Constant_0', 'Constant_1']) to ['conv1d'] I Try match layer: Squeeze_Squeeze_110:out0 I Match r_squeeze, convert ONNX layer: ['Squeeze_Squeeze_110'](OPs: ['Squeeze']) to ['squeeze'] I Try match layer: Transpose_Transpose_70:out0 I Match r_transpose, convert ONNX layer: ['Transpose_Transpose_70'](OPs: ['Transpose']) to ['permute'] I Try match layer: Transpose_Transpose_152:out0 I Match r_transpose, convert ONNX layer: ['Transpose_Transpose_152'](OPs: ['Transpose']) to ['permute'] I Try match layer: Add_Add_109:out0 I Match r_add, convert ONNX layer: ['Add_Add_109'](OPs: ['Add']) to ['add'] I Try match layer: Add_Add_69:out0 I Match layernorm, convert ONNX layer: ['Add_Add_69', 'Mul_Mul_68', 'Initializer_encoder.3.norm.bias', 'Div_Div_67', 'Initializer_encoder.3.norm.weight', 'Sub_Sub_60', 'Sqrt_Sqrt_66', 'ReduceMean_ReduceMean_59', 'Add_Add_65', 'ReduceMean_ReduceMean_63', 'Constant_Constant_64', 'Pow_Pow_62', 'Constant_Constant_61'](OPs: ['Sub', 'ReduceMean', 'Pow', 'Constant', 'ReduceMean_1', 'Add', 'Constant_1', 'Sqrt', 'Div', 'Mul', 'Constant_2', 'Add_1', 'Constant_3']) to ['layernormalize'] I Try match layer: Add_Add_151:out0 I Match r_add, convert ONNX layer: ['Add_Add_151'](OPs: ['Add']) to ['add'] I Try match layer: Initializer_duration_predictor.fc.bias:out0 I Match r_variable, convert ONNX layer: ['Initializer_duration_predictor.fc.bias'](OPs: ['Constant']) to ['variable'] I Try match layer: MatMul_MatMul_108:out0 I Match r_matmul, convert ONNX layer: ['MatMul_MatMul_108'](OPs: ['MatMul']) to ['matmul'] I Try match layer: Transpose_Transpose_58:out0 I Match r_transpose, convert ONNX layer: ['Transpose_Transpose_58'](OPs: ['Transpose']) to ['permute'] I Try match layer: Relu_Relu_57:out0 I Match r_relu, convert ONNX layer: ['Relu_Relu_57'](OPs: ['Relu']) to ['relu'] I Try match layer: Initializer_pitch_predictor.fc.bias:out0 I Match r_variable, convert ONNX layer: ['Initializer_pitch_predictor.fc.bias'](OPs: ['Constant']) to ['variable'] I Try match layer: MatMul_MatMul_150:out0 I Match r_matmul, convert ONNX layer: ['MatMul_MatMul_150'](OPs: ['MatMul']) to ['matmul'] I Try match layer: Transpose_Transpose_107:out0 I Match r_transpose, convert ONNX layer: ['Transpose_Transpose_107'](OPs: ['Transpose']) to ['permute'] I Try match layer: Initializer_258:out0 I Match r_variable, convert ONNX layer: ['Initializer_258'](OPs: ['Constant']) to ['variable'] I Try match layer: Conv_Conv_56:out0 I Match r_conv1d, convert ONNX layer: ['Conv_Conv_56', 'Initializer_encoder.3.conv.pointwise.weight', 'Initializer_encoder.3.conv.pointwise.bias'](OPs: ['Conv', 'Constant_0', 'Constant_1']) to ['conv1d'] I Try match layer: Transpose_Transpose_149:out0 I Match r_transpose, convert ONNX layer: ['Transpose_Transpose_149'](OPs: ['Transpose']) to ['permute'] I Try match layer: Initializer_261:out0 I Match r_variable, convert ONNX layer: ['Initializer_261'](OPs: ['Constant']) to ['variable'] I Try match layer: Cast_Cast_106:out0 I Match r_cast, convert ONNX layer: ['Cast_Cast_106'](OPs: ['Cast']) to ['cast'] I Try match layer: Conv_Conv_55:out0 I Match r_depthwise_conv1d, convert ONNX layer: ['Conv_Conv_55', 'Initializer_encoder.3.conv.depthwise.weight', 'Initializer_encoder.3.conv.depthwise.bias'](OPs: ['Conv', 'Constant_0', 'Constant_1']) to ['depthwise_conv1d'] I Try match layer: Cast_Cast_148:out0 I Match r_cast, convert ONNX layer: ['Cast_Cast_148'](OPs: ['Cast']) to ['cast'] I Try match layer: Transpose_Transpose_105:out0 I Match r_transpose, convert ONNX layer: ['Transpose_Transpose_105'](OPs: ['Transpose']) to ['permute'] I Try match layer: Cast_Cast_54:out0 I Match r_cast, convert ONNX layer: ['Cast_Cast_54'](OPs: ['Cast']) to ['cast'] I Try match layer: Transpose_Transpose_147:out0 I Match r_transpose, convert ONNX layer: ['Transpose_Transpose_147'](OPs: ['Transpose']) to ['permute'] I Try match layer: Add_Add_104:out0 I Match layernorm, convert ONNX layer: ['Add_Add_104', 'Mul_Mul_103', 'Initializer_duration_predictor.layers.1.norm.bias', 'Div_Div_102', 'Initializer_duration_predictor.layers.1.norm.weight', 'Sub_Sub_95', 'Sqrt_Sqrt_101', 'ReduceMean_ReduceMean_94', 'Add_Add_100', 'ReduceMean_ReduceMean_98', 'Constant_Constant_99', 'Pow_Pow_97', 'Constant_Constant_96'](OPs: ['Sub', 'ReduceMean', 'Pow', 'Constant', 'ReduceMean_1', 'Add', 'Constant_1', 'Sqrt', 'Div', 'Mul', 'Constant_2', 'Add_1', 'Constant_3']) to ['layernormalize'] I Try match layer: Transpose_Transpose_53:out0 I Match r_transpose, convert ONNX layer: ['Transpose_Transpose_53'](OPs: ['Transpose']) to ['permute'] I Try match layer: Add_Add_146:out0 I Match layernorm, convert ONNX layer: ['Add_Add_146', 'Mul_Mul_145', 'Initializer_pitch_predictor.layers.1.norm.bias', 'Div_Div_144', 'Initializer_pitch_predictor.layers.1.norm.weight', 'Sub_Sub_137', 'Sqrt_Sqrt_143', 'ReduceMean_ReduceMean_136', 'Add_Add_142', 'ReduceMean_ReduceMean_140', 'Constant_Constant_141', 'Pow_Pow_139', 'Constant_Constant_138'](OPs: ['Sub', 'ReduceMean', 'Pow', 'Constant', 'ReduceMean_1', 'Add', 'Constant_1', 'Sqrt', 'Div', 'Mul', 'Constant_2', 'Add_1', 'Constant_3']) to ['layernormalize'] I Try match layer: Transpose_Transpose_93:out0 I Match r_transpose, convert ONNX layer: ['Transpose_Transpose_93'](OPs: ['Transpose']) to ['permute'] I Try match layer: Relu_Relu_92:out0 I Match r_relu, convert ONNX layer: ['Relu_Relu_92'](OPs: ['Relu']) to ['relu'] I Try match layer: Add_Add_52:out0 I Match layernorm, convert ONNX layer: ['Add_Add_52', 'Mul_Mul_51', 'Initializer_encoder.2.norm.bias', 'Div_Div_50', 'Initializer_encoder.2.norm.weight', 'Sub_Sub_43', 'Sqrt_Sqrt_49', 'ReduceMean_ReduceMean_42', 'Add_Add_48', 'ReduceMean_ReduceMean_46', 'Constant_Constant_47', 'Pow_Pow_45', 'Constant_Constant_44'](OPs: ['Sub', 'ReduceMean', 'Pow', 'Constant', 'ReduceMean_1', 'Add', 'Constant_1', 'Sqrt', 'Div', 'Mul', 'Constant_2', 'Add_1', 'Constant_3']) to ['layernormalize'] I Try match layer: Transpose_Transpose_135:out0 I Match r_transpose, convert ONNX layer: ['Transpose_Transpose_135'](OPs: ['Transpose']) to ['permute'] I Try match layer: Relu_Relu_134:out0 I Match r_relu, convert ONNX layer: ['Relu_Relu_134'](OPs: ['Relu']) to ['relu'] I Try match layer: Conv_Conv_91:out0 I Match r_conv1d, convert ONNX layer: ['Conv_Conv_91', 'Initializer_duration_predictor.layers.1.conv.pointwise.weight', 'Initializer_duration_predictor.layers.1.conv.pointwise.bias'](OPs: ['Conv', 'Constant_0', 'Constant_1']) to ['conv1d'] I Try match layer: Transpose_Transpose_41:out0 I Match r_transpose, convert ONNX layer: ['Transpose_Transpose_41'](OPs: ['Transpose']) to ['permute'] I Try match layer: Relu_Relu_40:out0 I Match r_relu, convert ONNX layer: ['Relu_Relu_40'](OPs: ['Relu']) to ['relu'] I Try match layer: Conv_Conv_133:out0 I Match r_conv1d, convert ONNX layer: ['Conv_Conv_133', 'Initializer_pitch_predictor.layers.1.conv.pointwise.weight', 'Initializer_pitch_predictor.layers.1.conv.pointwise.bias'](OPs: ['Conv', 'Constant_0', 'Constant_1']) to ['conv1d'] I Try match layer: Conv_Conv_90:out0 I Match r_depthwise_conv1d, convert ONNX layer: ['Conv_Conv_90', 'Initializer_duration_predictor.layers.1.conv.depthwise.weight', 'Initializer_duration_predictor.layers.1.conv.depthwise.bias'](OPs: ['Conv', 'Constant_0', 'Constant_1']) to ['depthwise_conv1d'] I Try match layer: Conv_Conv_39:out0 I Match r_conv1d, convert ONNX layer: ['Conv_Conv_39', 'Initializer_encoder.2.conv.pointwise.weight', 'Initializer_encoder.2.conv.pointwise.bias'](OPs: ['Conv', 'Constant_0', 'Constant_1']) to ['conv1d'] I Try match layer: Conv_Conv_132:out0 I Match r_depthwise_conv1d, convert ONNX layer: ['Conv_Conv_132', 'Initializer_pitch_predictor.layers.1.conv.depthwise.weight', 'Initializer_pitch_predictor.layers.1.conv.depthwise.bias'](OPs: ['Conv', 'Constant_0', 'Constant_1']) to ['depthwise_conv1d'] I Try match layer: Cast_Cast_89:out0 I Match r_cast, convert ONNX layer: ['Cast_Cast_89'](OPs: ['Cast']) to ['cast'] I Try match layer: Conv_Conv_38:out0 I Match r_depthwise_conv1d, convert ONNX layer: ['Conv_Conv_38', 'Initializer_encoder.2.conv.depthwise.weight', 'Initializer_encoder.2.conv.depthwise.bias'](OPs: ['Conv', 'Constant_0', 'Constant_1']) to ['depthwise_conv1d'] I Try match layer: Cast_Cast_131:out0 I Match r_cast, convert ONNX layer: ['Cast_Cast_131'](OPs: ['Cast']) to ['cast'] I Try match layer: Transpose_Transpose_88:out0 I Match r_transpose, convert ONNX layer: ['Transpose_Transpose_88'](OPs: ['Transpose']) to ['permute'] I Try match layer: Cast_Cast_37:out0 I Match r_cast, convert ONNX layer: ['Cast_Cast_37'](OPs: ['Cast']) to ['cast'] I Try match layer: Transpose_Transpose_130:out0 I Match r_transpose, convert ONNX layer: ['Transpose_Transpose_130'](OPs: ['Transpose']) to ['permute'] I Try match layer: Add_Add_87:out0 I Match layernorm, convert ONNX layer: ['Add_Add_87', 'Mul_Mul_86', 'Initializer_duration_predictor.layers.0.norm.bias', 'Div_Div_85', 'Initializer_duration_predictor.layers.0.norm.weight', 'Sub_Sub_78', 'Sqrt_Sqrt_84', 'ReduceMean_ReduceMean_77', 'Add_Add_83', 'ReduceMean_ReduceMean_81', 'Constant_Constant_82', 'Pow_Pow_80', 'Constant_Constant_79'](OPs: ['Sub', 'ReduceMean', 'Pow', 'Constant', 'ReduceMean_1', 'Add', 'Constant_1', 'Sqrt', 'Div', 'Mul', 'Constant_2', 'Add_1', 'Constant_3']) to ['layernormalize'] I Try match layer: Transpose_Transpose_36:out0 I Match r_transpose, convert ONNX layer: ['Transpose_Transpose_36'](OPs: ['Transpose']) to ['permute'] I Try match layer: Add_Add_129:out0 I Match layernorm, convert ONNX layer: ['Add_Add_129', 'Mul_Mul_128', 'Initializer_pitch_predictor.layers.0.norm.bias', 'Div_Div_127', 'Initializer_pitch_predictor.layers.0.norm.weight', 'Sub_Sub_120', 'Sqrt_Sqrt_126', 'ReduceMean_ReduceMean_119', 'Add_Add_125', 'ReduceMean_ReduceMean_123', 'Constant_Constant_124', 'Pow_Pow_122', 'Constant_Constant_121'](OPs: ['Sub', 'ReduceMean', 'Pow', 'Constant', 'ReduceMean_1', 'Add', 'Constant_1', 'Sqrt', 'Div', 'Mul', 'Constant_2', 'Add_1', 'Constant_3']) to ['layernormalize'] I Try match layer: Transpose_Transpose_76:out0 I Match r_transpose, convert ONNX layer: ['Transpose_Transpose_76'](OPs: ['Transpose']) to ['permute'] I Try match layer: Relu_Relu_75:out0 I Match r_relu, convert ONNX layer: ['Relu_Relu_75'](OPs: ['Relu']) to ['relu'] I Try match layer: Add_Add_35:out0 I Match layernorm, convert ONNX layer: ['Add_Add_35', 'Mul_Mul_34', 'Initializer_encoder.1.norm.bias', 'Div_Div_33', 'Initializer_encoder.1.norm.weight', 'Sub_Sub_26', 'Sqrt_Sqrt_32', 'ReduceMean_ReduceMean_25', 'Add_Add_31', 'ReduceMean_ReduceMean_29', 'Constant_Constant_30', 'Pow_Pow_28', 'Constant_Constant_27'](OPs: ['Sub', 'ReduceMean', 'Pow', 'Constant', 'ReduceMean_1', 'Add', 'Constant_1', 'Sqrt', 'Div', 'Mul', 'Constant_2', 'Add_1', 'Constant_3']) to ['layernormalize'] I Try match layer: Transpose_Transpose_118:out0 I Match r_transpose, convert ONNX layer: ['Transpose_Transpose_118'](OPs: ['Transpose']) to ['permute'] I Try match layer: Relu_Relu_117:out0 I Match r_relu, convert ONNX layer: ['Relu_Relu_117'](OPs: ['Relu']) to ['relu'] I Try match layer: Conv_Conv_74:out0 I Match r_conv1d, convert ONNX layer: ['Conv_Conv_74', 'Initializer_duration_predictor.layers.0.conv.pointwise.weight', 'Initializer_duration_predictor.layers.0.conv.pointwise.bias'](OPs: ['Conv', 'Constant_0', 'Constant_1']) to ['conv1d'] I Try match layer: Transpose_Transpose_24:out0 I Match r_transpose, convert ONNX layer: ['Transpose_Transpose_24'](OPs: ['Transpose']) to ['permute'] I Try match layer: Relu_Relu_23:out0 I Match r_relu, convert ONNX layer: ['Relu_Relu_23'](OPs: ['Relu']) to ['relu'] I Try match layer: Conv_Conv_116:out0 I Match r_conv1d, convert ONNX layer: ['Conv_Conv_116', 'Initializer_pitch_predictor.layers.0.conv.pointwise.weight', 'Initializer_pitch_predictor.layers.0.conv.pointwise.bias'](OPs: ['Conv', 'Constant_0', 'Constant_1']) to ['conv1d'] I Try match layer: Conv_Conv_73:out0 I Match r_depthwise_conv1d, convert ONNX layer: ['Conv_Conv_73', 'Initializer_duration_predictor.layers.0.conv.depthwise.weight', 'Initializer_duration_predictor.layers.0.conv.depthwise.bias'](OPs: ['Conv', 'Constant_0', 'Constant_1']) to ['depthwise_conv1d'] I Try match layer: Conv_Conv_22:out0 I Match r_conv1d, convert ONNX layer: ['Conv_Conv_22', 'Initializer_encoder.1.conv.pointwise.weight', 'Initializer_encoder.1.conv.pointwise.bias'](OPs: ['Conv', 'Constant_0', 'Constant_1']) to ['conv1d'] I Try match layer: Conv_Conv_115:out0 I Match r_depthwise_conv1d, convert ONNX layer: ['Conv_Conv_115', 'Initializer_pitch_predictor.layers.0.conv.depthwise.weight', 'Initializer_pitch_predictor.layers.0.conv.depthwise.bias'](OPs: ['Conv', 'Constant_0', 'Constant_1']) to ['depthwise_conv1d'] I Try match layer: Conv_Conv_21:out0 I Match r_depthwise_conv1d, convert ONNX layer: ['Conv_Conv_21', 'Initializer_encoder.1.conv.depthwise.weight', 'Initializer_encoder.1.conv.depthwise.bias'](OPs: ['Conv', 'Constant_0', 'Constant_1']) to ['depthwise_conv1d'] I Try match layer: Cast_Cast_20:out0 I Match r_cast, convert ONNX layer: ['Cast_Cast_20'](OPs: ['Cast']) to ['cast'] I Try match layer: Transpose_Transpose_19:out0 I Match r_transpose, convert ONNX layer: ['Transpose_Transpose_19'](OPs: ['Transpose']) to ['permute'] I Try match layer: Add_Add_18:out0 I Match layernorm, convert ONNX layer: ['Add_Add_18', 'Mul_Mul_17', 'Initializer_encoder.0.norm.bias', 'Div_Div_16', 'Initializer_encoder.0.norm.weight', 'Sub_Sub_9', 'Sqrt_Sqrt_15', 'ReduceMean_ReduceMean_8', 'Add_Add_14', 'ReduceMean_ReduceMean_12', 'Constant_Constant_13', 'Pow_Pow_11', 'Constant_Constant_10'](OPs: ['Sub', 'ReduceMean', 'Pow', 'Constant', 'ReduceMean_1', 'Add', 'Constant_1', 'Sqrt', 'Div', 'Mul', 'Constant_2', 'Add_1', 'Constant_3']) to ['layernormalize'] I Try match layer: Transpose_Transpose_7:out0 I Match r_transpose, convert ONNX layer: ['Transpose_Transpose_7'](OPs: ['Transpose']) to ['permute'] I Try match layer: Relu_Relu_6:out0 I Match r_relu, convert ONNX layer: ['Relu_Relu_6'](OPs: ['Relu']) to ['relu'] I Try match layer: Conv_Conv_5:out0 I Match r_conv1d, convert ONNX layer: ['Conv_Conv_5', 'Initializer_encoder.0.conv.pointwise.weight', 'Initializer_encoder.0.conv.pointwise.bias'](OPs: ['Conv', 'Constant_0', 'Constant_1']) to ['conv1d'] I Try match layer: Conv_Conv_4:out0 I Match r_depthwise_conv1d, convert ONNX layer: ['Conv_Conv_4', 'Initializer_encoder.0.conv.depthwise.weight', 'Initializer_encoder.0.conv.depthwise.bias'](OPs: ['Conv', 'Constant_0', 'Constant_1']) to ['depthwise_conv1d'] I Try match layer: Transpose_Transpose_3:out0 I Match r_transpose, convert ONNX layer: ['Transpose_Transpose_3'](OPs: ['Transpose']) to ['permute'] I Try match layer: Add_Add_2:out0 I Match r_add, convert ONNX layer: ['Add_Add_2'](OPs: ['Add']) to ['add'] I Try match layer: Gather_Gather_0:out0 I Match r_embedding_as_gather, convert ONNX layer: ['Gather_Gather_0'](OPs: ['Gather']) to ['gather', 'squeeze'] D connect ['gather:out0', 'squeeze:in0'] I Try match layer: Constant_Constant_1:out0 I Match r_variable, convert ONNX layer: ['Constant_Constant_1'](OPs: ['Constant']) to ['variable'] I Try match layer: Initializer_word_emb.weight:out0 I Match r_variable, convert ONNX layer: ['Initializer_word_emb.weight'](OPs: ['Constant']) to ['variable'] I build input layer input:out0 D connect Sub_Sub_113_4 0 ~ Clip_Clip_114_2 0 D connect Transpose_Transpose_72_5 0 ~ Add_Add_155_3 0 D connect Transpose_Transpose_154_6 0 ~ Add_Add_155_3 1 D connect Exp_Exp_111_7 0 ~ Sub_Sub_113_4 0 D connect Constant_Constant_112_8 0 ~ Sub_Sub_113_4 1 D connect Cast_Cast_71_9 0 ~ Transpose_Transpose_72_5 0 D connect Conv_Conv_153_10 0 ~ Transpose_Transpose_154_6 0 D connect Squeeze_Squeeze_110_11 0 ~ Exp_Exp_111_7 0 D connect Transpose_Transpose_70_12 0 ~ Cast_Cast_71_9 0 D connect Transpose_Transpose_152_13 0 ~ Conv_Conv_153_10 0 D connect Add_Add_109_14 0 ~ Squeeze_Squeeze_110_11 0 D connect Add_Add_69_15 0 ~ Transpose_Transpose_70_12 0 D connect Add_Add_151_16 0 ~ Transpose_Transpose_152_13 0 D connect Initializer_duration_predictor.fc.bias_17 0 ~ Add_Add_109_14 0 D connect MatMul_MatMul_108_18 0 ~ Add_Add_109_14 1 D connect Transpose_Transpose_58_19 0 ~ Add_Add_69_15 0 D connect Initializer_pitch_predictor.fc.bias_21 0 ~ Add_Add_151_16 0 D connect MatMul_MatMul_150_22 0 ~ Add_Add_151_16 1 D connect Transpose_Transpose_107_23 0 ~ MatMul_MatMul_108_18 0 D connect Initializer_258_24 0 ~ MatMul_MatMul_108_18 1 D connect Relu_Relu_57_20 0 ~ Transpose_Transpose_58_19 0 D connect Conv_Conv_56_25 0 ~ Relu_Relu_57_20 0 D connect Transpose_Transpose_149_26 0 ~ MatMul_MatMul_150_22 0 D connect Initializer_261_27 0 ~ MatMul_MatMul_150_22 1 D connect Cast_Cast_106_28 0 ~ Transpose_Transpose_107_23 0 D connect Conv_Conv_55_29 0 ~ Conv_Conv_56_25 0 D connect Cast_Cast_148_30 0 ~ Transpose_Transpose_149_26 0 D connect Transpose_Transpose_105_31 0 ~ Cast_Cast_106_28 0 D connect Cast_Cast_54_32 0 ~ Conv_Conv_55_29 0 D connect Transpose_Transpose_147_33 0 ~ Cast_Cast_148_30 0 D connect Add_Add_104_34 0 ~ Transpose_Transpose_105_31 0 D connect Transpose_Transpose_53_35 0 ~ Cast_Cast_54_32 0 D connect Add_Add_146_36 0 ~ Transpose_Transpose_147_33 0 D connect Transpose_Transpose_93_37 0 ~ Add_Add_104_34 0 D connect Add_Add_52_39 0 ~ Transpose_Transpose_53_35 0 D connect Transpose_Transpose_135_40 0 ~ Add_Add_146_36 0 D connect Relu_Relu_92_38 0 ~ Transpose_Transpose_93_37 0 D connect Conv_Conv_91_42 0 ~ Relu_Relu_92_38 0 D connect Transpose_Transpose_41_43 0 ~ Add_Add_52_39 0 D connect Relu_Relu_134_41 0 ~ Transpose_Transpose_135_40 0 D connect Conv_Conv_133_45 0 ~ Relu_Relu_134_41 0 D connect Conv_Conv_90_46 0 ~ Conv_Conv_91_42 0 D connect Relu_Relu_40_44 0 ~ Transpose_Transpose_41_43 0 D connect Conv_Conv_39_47 0 ~ Relu_Relu_40_44 0 D connect Conv_Conv_132_48 0 ~ Conv_Conv_133_45 0 D connect Cast_Cast_89_49 0 ~ Conv_Conv_90_46 0 D connect Conv_Conv_38_50 0 ~ Conv_Conv_39_47 0 D connect Cast_Cast_131_51 0 ~ Conv_Conv_132_48 0 D connect Transpose_Transpose_88_52 0 ~ Cast_Cast_89_49 0 D connect Cast_Cast_37_53 0 ~ Conv_Conv_38_50 0 D connect Transpose_Transpose_130_54 0 ~ Cast_Cast_131_51 0 D connect Add_Add_87_55 0 ~ Transpose_Transpose_88_52 0 D connect Transpose_Transpose_36_56 0 ~ Cast_Cast_37_53 0 D connect Add_Add_129_57 0 ~ Transpose_Transpose_130_54 0 D connect Transpose_Transpose_76_58 0 ~ Add_Add_87_55 0 D connect Add_Add_35_60 0 ~ Transpose_Transpose_36_56 0 D connect Transpose_Transpose_118_61 0 ~ Add_Add_129_57 0 D connect Relu_Relu_75_59 0 ~ Transpose_Transpose_76_58 0 D connect Conv_Conv_74_63 0 ~ Relu_Relu_75_59 0 D connect Transpose_Transpose_24_64 0 ~ Add_Add_35_60 0 D connect Relu_Relu_117_62 0 ~ Transpose_Transpose_118_61 0 D connect Conv_Conv_116_66 0 ~ Relu_Relu_117_62 0 D connect Conv_Conv_73_67 0 ~ Conv_Conv_74_63 0 D connect Relu_Relu_23_65 0 ~ Transpose_Transpose_24_64 0 D connect Conv_Conv_22_68 0 ~ Relu_Relu_23_65 0 D connect Conv_Conv_115_69 0 ~ Conv_Conv_116_66 0 D connect Cast_Cast_71_9 0 ~ Conv_Conv_73_67 0 D connect Conv_Conv_21_70 0 ~ Conv_Conv_22_68 0 D connect Cast_Cast_71_9 0 ~ Conv_Conv_115_69 0 D connect Cast_Cast_20_71 0 ~ Conv_Conv_21_70 0 D connect Transpose_Transpose_19_72 0 ~ Cast_Cast_20_71 0 D connect Add_Add_18_73 0 ~ Transpose_Transpose_19_72 0 D connect Transpose_Transpose_7_74 0 ~ Add_Add_18_73 0 D connect Relu_Relu_6_75 0 ~ Transpose_Transpose_7_74 0 D connect Conv_Conv_5_76 0 ~ Relu_Relu_6_75 0 D connect Conv_Conv_4_77 0 ~ Conv_Conv_5_76 0 D connect Transpose_Transpose_3_78 0 ~ Conv_Conv_4_77 0 D connect Add_Add_2_79 0 ~ Transpose_Transpose_3_78 0 D connect squeeze_81 0 ~ Add_Add_2_79 0 D connect Constant_Constant_1_82 0 ~ Add_Add_2_79 1 D connect Initializer_word_emb.weight_83 0 ~ Gather_Gather_0_80 0 D connect input_84 0 ~ Gather_Gather_0_80 1 D connect Add_Add_155_3 0 ~ attach_Add_Add_155/out0_0 0 D connect Clip_Clip_114_2 0 ~ attach_Clip_Clip_114/out0_1 0 D Process Initializer_word_emb.weight_83 ... D RKNN output shape(variable): (81 384) D Process input_84 ... D RKNN output shape(input): (1 61) D Tensor @input_84:out0 type: float32 D Process Gather_Gather_0_80 ... W Warning: Axis may need to be adjusted according to original model shape. D RKNN output shape(gather): (1 61 384) D Tensor @Gather_Gather_0_80:out0 type: float32 D Process squeeze_81 ... D RKNN output shape(squeeze): (61 384) D Tensor @squeeze_81:out0 type: float32 D Process Constant_Constant_1_82 ... D RKNN output shape(variable): (1) D Process Add_Add_2_79 ... D RKNN output shape(add): (61 384) D Tensor @Add_Add_2_79:out0 type: float32 D Process Transpose_Transpose_3_78 ... E Catch exception when loading onnx model: /mnt/c/tianyu_zhao/fastpitch_outputs/biaobei_lightspeech_ssim/onnx/LightSpeech_encoder.onnx! E Traceback (most recent call last): E File "rknn/base/RKNNlib/RKNNnetbuilder.py", line 437, in rknn.base.RKNNlib.RKNNnetbuilder.RKNNNetBuilder.build_layer E File "rknn/base/RKNNlib/RKNNnetbuilder.py", line 437, in rknn.base.RKNNlib.RKNNnetbuilder.RKNNNetBuilder.build_layer E File "rknn/base/RKNNlib/RKNNnetbuilder.py", line 449, in rknn.base.RKNNlib.RKNNnetbuilder.RKNNNetBuilder.build_layer E File "rknn/base/RKNNlib/layer/RKNNlayer.py", line 268, in rknn.base.RKNNlib.layer.RKNNlayer.RKNNLayer.compute_shape E File "rknn/base/RKNNlib/layer/permute.py", line 24, in rknn.base.RKNNlib.layer.permute.Permute.compute_out_shape E IndexError: list index out of range E Please feedback the detailed log file to the RKNN Toolkit development team. E You can also check github issues: https://github.com/rockchip-linux/rknn-toolkit/issues

rockchip-linux / rknn-toolkit

rknn.load_onnx() Error #177