File size: 48,560 Bytes
4a33762 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 |
Ticket Name: TDA2PXEVM: Is Conv with no relue supported In TDA2?
Query Text:
Part Number: TDA2PXEVM Other Parts Discussed in Thread: TDA2 Hi I get some accuracy problems with caffemodel importing. Here is my steps: a) Use tidl_model_import.out.exe convert regnet.prototxt/caffemodel to tidl_param/bin b) Compare each layer's output : onnx vs trace_dump_idx_wxh.y c) I find the first conv layer(with bn & relu) matched well(error < 1%) d) But conv_layer with no relu(which are inputs of eltwise_layer) cannot match original model layer's outputs I put a snapshot below : conv layers in green rect match, but conv layer without relu in red circle not match. regnet_import.zip So, is relu strictedly demanded to place after conv layer? Or this is just bugs in import tool? I upload my model and tools for analysis.
Responses:
Hi, Can you share the import output log to check the issue ? Thanks, Praveen
Hi Praveen Thanks for your replay! I uploaded LOG.txt. Geroge. 6675.LOG.txt .\tidl_model_import.out.exe .\ONNX_Reg200M_CIFAR\tidl_import.txt
Caffe Network File : ONNX_Reg200M_CIFAR\trained\regnetx200mf_cifar_Relu96.prototxt
Caffe Model File : ONNX_Reg200M_CIFAR\trained\regnetx200mf_cifar_Relu96.caffemodel
TIDL Network File : ONNX_Reg200M_CIFAR\model\tidl_net_reg200cifar_relu96.bin
TIDL Model File : ONNX_Reg200M_CIFAR\model\tidl_param_reg200cifar_relu96.bin
Name of the Network : REG200MCIFAR-ONNX
Num Inputs : 1
Num of Layer Detected : 71
0, TIDL_DataLayer , data 0, -1 , 1 , x , x , x , x , x , x , x , x , 0 , 0 , 0 , 0 , 0 , 1 , 3 , 32 , 32 , 0 ,
1, TIDL_ConvolutionLayer , Conv_0 1, 1 , 1 , 0 , x , x , x , x , x , x , x , 1 , 1 , 3 , 32 , 32 , 1 , 32 , 32 , 32 , 884736 ,
2, TIDL_ConvolutionLayer , Conv_2 1, 1 , 1 , 1 , x , x , x , x , x , x , x , 2 , 1 , 32 , 32 , 32 , 1 , 24 , 32 , 32 , 786432 ,
3, TIDL_ConvolutionLayer , Conv_4 1, 1 , 1 , 2 , x , x , x , x , x , x , x , 3 , 1 , 24 , 32 , 32 , 1 , 24 , 16 , 16 , 442368 ,
4, TIDL_ConvolutionLayer , Conv_6 1, 1 , 1 , 3 , x , x , x , x , x , x , x , 4 , 1 , 24 , 16 , 16 , 1 , 24 , 16 , 16 , 147456 ,
5, TIDL_ConvolutionLayer , Conv_7 1, 1 , 1 , 1 , x , x , x , x , x , x , x , 5 , 1 , 32 , 32 , 32 , 1 , 24 , 16 , 16 , 196608 ,
6, TIDL_EltWiseLayer , Add_8 1, 2 , 1 , 4 , 5 , x , x , x , x , x , x , 6 , 1 , 24 , 16 , 16 , 1 , 24 , 16 , 16 , 6144 ,
7, TIDL_BatchNormLayer , Relu_9 1, 1 , 1 , 6 , x , x , x , x , x , x , x , 7 , 1 , 24 , 16 , 16 , 1 , 24 , 16 , 16 , 6144 ,
8, TIDL_ConvolutionLayer , Conv_10 1, 1 , 1 , 7 , x , x , x , x , x , x , x , 8 , 1 , 24 , 16 , 16 , 1 , 56 , 16 , 16 , 344064 ,
9, TIDL_ConvolutionLayer , Conv_12 1, 1 , 1 , 8 , x , x , x , x , x , x , x , 9 , 1 , 56 , 16 , 16 , 1 , 56 , 16 , 16 , 1032192 ,
10, TIDL_ConvolutionLayer , Conv_14 1, 1 , 1 , 9 , x , x , x , x , x , x , x , 10 , 1 , 56 , 16 , 16 , 1 , 56 , 16 , 16 , 802816 ,
11, TIDL_ConvolutionLayer , Conv_15 1, 1 , 1 , 7 , x , x , x , x , x , x , x , 11 , 1 , 24 , 16 , 16 , 1 , 56 , 16 , 16 , 344064 ,
12, TIDL_EltWiseLayer , Add_16 1, 2 , 1 , 10 , 11 , x , x , x , x , x , x , 12 , 1 , 56 , 16 , 16 , 1 , 56 , 16 , 16 , 14336 ,
13, TIDL_BatchNormLayer , Relu_17 1, 1 , 1 , 12 , x , x , x , x , x , x , x , 13 , 1 , 56 , 16 , 16 , 1 , 56 , 16 , 16 , 14336 ,
14, TIDL_ConvolutionLayer , Conv_18 1, 1 , 1 , 13 , x , x , x , x , x , x , x , 14 , 1 , 56 , 16 , 16 , 1 , 152 , 16 , 16 , 2179072 ,
15, TIDL_ConvolutionLayer , Conv_20 1, 1 , 1 , 14 , x , x , x , x , x , x , x , 15 , 1 , 152 , 16 , 16 , 1 , 152 , 8 , 8 , 700416 ,
16, TIDL_ConvolutionLayer , Conv_22 1, 1 , 1 , 15 , x , x , x , x , x , x , x , 16 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 , 1478656 ,
17, TIDL_ConvolutionLayer , Conv_23 1, 1 , 1 , 13 , x , x , x , x , x , x , x , 17 , 1 , 56 , 16 , 16 , 1 , 152 , 8 , 8 , 544768 ,
18, TIDL_EltWiseLayer , Add_24 1, 2 , 1 , 16 , 17 , x , x , x , x , x , x , 18 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 , 9728 ,
19, TIDL_BatchNormLayer , Relu_25 1, 1 , 1 , 18 , x , x , x , x , x , x , x , 19 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 , 9728 ,
20, TIDL_ConvolutionLayer , Conv_26 1, 1 , 1 , 19 , x , x , x , x , x , x , x , 20 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 , 1478656 ,
21, TIDL_ConvolutionLayer , Conv_28 1, 1 , 1 , 20 , x , x , x , x , x , x , x , 21 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 , 700416 ,
22, TIDL_ConvolutionLayer , Conv_30 1, 1 , 1 , 21 , x , x , x , x , x , x , x , 22 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 , 1478656 ,
23, TIDL_EltWiseLayer , Add_31 1, 2 , 1 , 22 , 19 , x , x , x , x , x , x , 23 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 , 9728 ,
24, TIDL_BatchNormLayer , Relu_32 1, 1 , 1 , 23 , x , x , x , x , x , x , x , 24 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 , 9728 ,
25, TIDL_ConvolutionLayer , Conv_33 1, 1 , 1 , 24 , x , x , x , x , x , x , x , 25 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 , 1478656 ,
26, TIDL_ConvolutionLayer , Conv_35 1, 1 , 1 , 25 , x , x , x , x , x , x , x , 26 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 , 700416 ,
27, TIDL_ConvolutionLayer , Conv_37 1, 1 , 1 , 26 , x , x , x , x , x , x , x , 27 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 , 1478656 ,
28, TIDL_EltWiseLayer , Add_38 1, 2 , 1 , 27 , 24 , x , x , x , x , x , x , 28 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 , 9728 ,
29, TIDL_BatchNormLayer , Relu_39 1, 1 , 1 , 28 , x , x , x , x , x , x , x , 29 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 , 9728 ,
30, TIDL_ConvolutionLayer , Conv_40 1, 1 , 1 , 29 , x , x , x , x , x , x , x , 30 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 , 1478656 ,
31, TIDL_ConvolutionLayer , Conv_42 1, 1 , 1 , 30 , x , x , x , x , x , x , x , 31 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 , 700416 ,
32, TIDL_ConvolutionLayer , Conv_44 1, 1 , 1 , 31 , x , x , x , x , x , x , x , 32 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 , 1478656 ,
33, TIDL_EltWiseLayer , Add_45 1, 2 , 1 , 32 , 29 , x , x , x , x , x , x , 33 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 , 9728 ,
34, TIDL_BatchNormLayer , Relu_46 1, 1 , 1 , 33 , x , x , x , x , x , x , x , 34 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 , 9728 ,
35, TIDL_ConvolutionLayer , Conv_47 1, 1 , 1 , 34 , x , x , x , x , x , x , x , 35 , 1 , 152 , 8 , 8 , 1 , 368 , 8 , 8 , 3579904 ,
36, TIDL_ConvolutionLayer , Conv_49 1, 1 , 1 , 35 , x , x , x , x , x , x , x , 36 , 1 , 368 , 8 , 8 , 1 , 368 , 4 , 4 , 423936 ,
37, TIDL_ConvolutionLayer , Conv_51 1, 1 , 1 , 36 , x , x , x , x , x , x , x , 37 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 2166784 ,
38, TIDL_ConvolutionLayer , Conv_52 1, 1 , 1 , 34 , x , x , x , x , x , x , x , 38 , 1 , 152 , 8 , 8 , 1 , 368 , 4 , 4 , 894976 ,
39, TIDL_EltWiseLayer , Add_53 1, 2 , 1 , 37 , 38 , x , x , x , x , x , x , 39 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 5888 ,
40, TIDL_BatchNormLayer , Relu_54 1, 1 , 1 , 39 , x , x , x , x , x , x , x , 40 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 5888 ,
41, TIDL_ConvolutionLayer , Conv_55 1, 1 , 1 , 40 , x , x , x , x , x , x , x , 41 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 2166784 ,
42, TIDL_ConvolutionLayer , Conv_57 1, 1 , 1 , 41 , x , x , x , x , x , x , x , 42 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 423936 ,
43, TIDL_ConvolutionLayer , Conv_59 1, 1 , 1 , 42 , x , x , x , x , x , x , x , 43 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 2166784 ,
44, TIDL_EltWiseLayer , Add_60 1, 2 , 1 , 43 , 40 , x , x , x , x , x , x , 44 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 5888 ,
45, TIDL_BatchNormLayer , Relu_61 1, 1 , 1 , 44 , x , x , x , x , x , x , x , 45 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 5888 ,
46, TIDL_ConvolutionLayer , Conv_62 1, 1 , 1 , 45 , x , x , x , x , x , x , x , 46 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 2166784 ,
47, TIDL_ConvolutionLayer , Conv_64 1, 1 , 1 , 46 , x , x , x , x , x , x , x , 47 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 423936 ,
48, TIDL_ConvolutionLayer , Conv_66 1, 1 , 1 , 47 , x , x , x , x , x , x , x , 48 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 2166784 ,
49, TIDL_EltWiseLayer , Add_67 1, 2 , 1 , 48 , 45 , x , x , x , x , x , x , 49 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 5888 ,
50, TIDL_BatchNormLayer , Relu_68 1, 1 , 1 , 49 , x , x , x , x , x , x , x , 50 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 5888 ,
51, TIDL_ConvolutionLayer , Conv_69 1, 1 , 1 , 50 , x , x , x , x , x , x , x , 51 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 2166784 ,
52, TIDL_ConvolutionLayer , Conv_71 1, 1 , 1 , 51 , x , x , x , x , x , x , x , 52 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 423936 ,
53, TIDL_ConvolutionLayer , Conv_73 1, 1 , 1 , 52 , x , x , x , x , x , x , x , 53 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 2166784 ,
54, TIDL_EltWiseLayer , Add_74 1, 2 , 1 , 53 , 50 , x , x , x , x , x , x , 54 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 5888 ,
55, TIDL_BatchNormLayer , Relu_75 1, 1 , 1 , 54 , x , x , x , x , x , x , x , 55 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 5888 ,
56, TIDL_ConvolutionLayer , Conv_76 1, 1 , 1 , 55 , x , x , x , x , x , x , x , 56 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 2166784 ,
57, TIDL_ConvolutionLayer , Conv_78 1, 1 , 1 , 56 , x , x , x , x , x , x , x , 57 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 423936 ,
58, TIDL_ConvolutionLayer , Conv_80 1, 1 , 1 , 57 , x , x , x , x , x , x , x , 58 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 2166784 ,
59, TIDL_EltWiseLayer , Add_81 1, 2 , 1 , 58 , 55 , x , x , x , x , x , x , 59 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 5888 ,
60, TIDL_BatchNormLayer , Relu_82 1, 1 , 1 , 59 , x , x , x , x , x , x , x , 60 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 5888 ,
61, TIDL_ConvolutionLayer , Conv_83 1, 1 , 1 , 60 , x , x , x , x , x , x , x , 61 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 2166784 ,
62, TIDL_ConvolutionLayer , Conv_85 1, 1 , 1 , 61 , x , x , x , x , x , x , x , 62 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 423936 ,
63, TIDL_ConvolutionLayer , Conv_87 1, 1 , 1 , 62 , x , x , x , x , x , x , x , 63 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 2166784 ,
64, TIDL_EltWiseLayer , Add_88 1, 2 , 1 , 63 , 60 , x , x , x , x , x , x , 64 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 5888 ,
65, TIDL_BatchNormLayer , Relu_89 1, 1 , 1 , 64 , x , x , x , x , x , x , x , 65 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 5888 ,
66, TIDL_ConvolutionLayer , Conv_90 1, 1 , 1 , 65 , x , x , x , x , x , x , x , 66 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 2166784 ,
67, TIDL_ConvolutionLayer , Conv_92 1, 1 , 1 , 66 , x , x , x , x , x , x , x , 67 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 423936 ,
68, TIDL_ConvolutionLayer , Conv_94 1, 1 , 1 , 67 , x , x , x , x , x , x , x , 68 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 2166784 ,
69, TIDL_EltWiseLayer , Add_95 1, 2 , 1 , 68 , 65 , x , x , x , x , x , x , 69 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 5888 ,
70, TIDL_BatchNormLayer , Relu_96 1, 1 , 1 , 69 , x , x , x , x , x , x , x , 70 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 , 5888 ,
Total Giga Macs : 0.0567
已复制 1 个文件。
Processing config file .\tempDir\qunat_stats_config.txt !
0, TIDL_DataLayer , 0, -1 , 1 , x , x , x , x , x , x , x , x , 0 , 0 , 0 , 0 , 0 , 1 , 3 , 32 , 32 ,
1, TIDL_ConvolutionLayer , 1, 1 , 1 , 0 , x , x , x , x , x , x , x , 1 , 1 , 3 , 32 , 32 , 1 , 32 , 32 , 32 ,
2, TIDL_ConvolutionLayer , 1, 1 , 1 , 1 , x , x , x , x , x , x , x , 2 , 1 , 32 , 32 , 32 , 1 , 24 , 32 , 32 ,
3, TIDL_ConvolutionLayer , 1, 1 , 1 , 2 , x , x , x , x , x , x , x , 3 , 1 , 24 , 32 , 32 , 1 , 24 , 16 , 16 ,
4, TIDL_ConvolutionLayer , 1, 1 , 1 , 3 , x , x , x , x , x , x , x , 4 , 1 , 24 , 16 , 16 , 1 , 24 , 16 , 16 ,
5, TIDL_ConvolutionLayer , 1, 1 , 1 , 1 , x , x , x , x , x , x , x , 5 , 1 , 32 , 32 , 32 , 1 , 24 , 16 , 16 ,
6, TIDL_EltWiseLayer , 1, 2 , 1 , 4 , 5 , x , x , x , x , x , x , 6 , 1 , 24 , 16 , 16 , 1 , 24 , 16 , 16 ,
7, TIDL_BatchNormLayer , 1, 1 , 1 , 6 , x , x , x , x , x , x , x , 7 , 1 , 24 , 16 , 16 , 1 , 24 , 16 , 16 ,
8, TIDL_ConvolutionLayer , 1, 1 , 1 , 7 , x , x , x , x , x , x , x , 8 , 1 , 24 , 16 , 16 , 1 , 56 , 16 , 16 ,
9, TIDL_ConvolutionLayer , 1, 1 , 1 , 8 , x , x , x , x , x , x , x , 9 , 1 , 56 , 16 , 16 , 1 , 56 , 16 , 16 ,
10, TIDL_ConvolutionLayer , 1, 1 , 1 , 9 , x , x , x , x , x , x , x , 10 , 1 , 56 , 16 , 16 , 1 , 56 , 16 , 16 ,
11, TIDL_ConvolutionLayer , 1, 1 , 1 , 7 , x , x , x , x , x , x , x , 11 , 1 , 24 , 16 , 16 , 1 , 56 , 16 , 16 ,
12, TIDL_EltWiseLayer , 1, 2 , 1 , 10 , 11 , x , x , x , x , x , x , 12 , 1 , 56 , 16 , 16 , 1 , 56 , 16 , 16 ,
13, TIDL_BatchNormLayer , 1, 1 , 1 , 12 , x , x , x , x , x , x , x , 13 , 1 , 56 , 16 , 16 , 1 , 56 , 16 , 16 ,
14, TIDL_ConvolutionLayer , 1, 1 , 1 , 13 , x , x , x , x , x , x , x , 14 , 1 , 56 , 16 , 16 , 1 , 152 , 16 , 16 ,
15, TIDL_ConvolutionLayer , 1, 1 , 1 , 14 , x , x , x , x , x , x , x , 15 , 1 , 152 , 16 , 16 , 1 , 152 , 8 , 8 ,
16, TIDL_ConvolutionLayer , 1, 1 , 1 , 15 , x , x , x , x , x , x , x , 16 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 ,
17, TIDL_ConvolutionLayer , 1, 1 , 1 , 13 , x , x , x , x , x , x , x , 17 , 1 , 56 , 16 , 16 , 1 , 152 , 8 , 8 ,
18, TIDL_EltWiseLayer , 1, 2 , 1 , 16 , 17 , x , x , x , x , x , x , 18 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 ,
19, TIDL_BatchNormLayer , 1, 1 , 1 , 18 , x , x , x , x , x , x , x , 19 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 ,
20, TIDL_ConvolutionLayer , 1, 1 , 1 , 19 , x , x , x , x , x , x , x , 20 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 ,
21, TIDL_ConvolutionLayer , 1, 1 , 1 , 20 , x , x , x , x , x , x , x , 21 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 ,
22, TIDL_ConvolutionLayer , 1, 1 , 1 , 21 , x , x , x , x , x , x , x , 22 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 ,
23, TIDL_EltWiseLayer , 1, 2 , 1 , 22 , 19 , x , x , x , x , x , x , 23 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 ,
24, TIDL_BatchNormLayer , 1, 1 , 1 , 23 , x , x , x , x , x , x , x , 24 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 ,
25, TIDL_ConvolutionLayer , 1, 1 , 1 , 24 , x , x , x , x , x , x , x , 25 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 ,
26, TIDL_ConvolutionLayer , 1, 1 , 1 , 25 , x , x , x , x , x , x , x , 26 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 ,
27, TIDL_ConvolutionLayer , 1, 1 , 1 , 26 , x , x , x , x , x , x , x , 27 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 ,
28, TIDL_EltWiseLayer , 1, 2 , 1 , 27 , 24 , x , x , x , x , x , x , 28 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 ,
29, TIDL_BatchNormLayer , 1, 1 , 1 , 28 , x , x , x , x , x , x , x , 29 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 ,
30, TIDL_ConvolutionLayer , 1, 1 , 1 , 29 , x , x , x , x , x , x , x , 30 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 ,
31, TIDL_ConvolutionLayer , 1, 1 , 1 , 30 , x , x , x , x , x , x , x , 31 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 ,
32, TIDL_ConvolutionLayer , 1, 1 , 1 , 31 , x , x , x , x , x , x , x , 32 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 ,
33, TIDL_EltWiseLayer , 1, 2 , 1 , 32 , 29 , x , x , x , x , x , x , 33 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 ,
34, TIDL_BatchNormLayer , 1, 1 , 1 , 33 , x , x , x , x , x , x , x , 34 , 1 , 152 , 8 , 8 , 1 , 152 , 8 , 8 ,
35, TIDL_ConvolutionLayer , 1, 1 , 1 , 34 , x , x , x , x , x , x , x , 35 , 1 , 152 , 8 , 8 , 1 , 368 , 8 , 8 ,
36, TIDL_ConvolutionLayer , 1, 1 , 1 , 35 , x , x , x , x , x , x , x , 36 , 1 , 368 , 8 , 8 , 1 , 368 , 4 , 4 ,
37, TIDL_ConvolutionLayer , 1, 1 , 1 , 36 , x , x , x , x , x , x , x , 37 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
38, TIDL_ConvolutionLayer , 1, 1 , 1 , 34 , x , x , x , x , x , x , x , 38 , 1 , 152 , 8 , 8 , 1 , 368 , 4 , 4 ,
39, TIDL_EltWiseLayer , 1, 2 , 1 , 37 , 38 , x , x , x , x , x , x , 39 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
40, TIDL_BatchNormLayer , 1, 1 , 1 , 39 , x , x , x , x , x , x , x , 40 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
41, TIDL_ConvolutionLayer , 1, 1 , 1 , 40 , x , x , x , x , x , x , x , 41 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
42, TIDL_ConvolutionLayer , 1, 1 , 1 , 41 , x , x , x , x , x , x , x , 42 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
43, TIDL_ConvolutionLayer , 1, 1 , 1 , 42 , x , x , x , x , x , x , x , 43 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
44, TIDL_EltWiseLayer , 1, 2 , 1 , 43 , 40 , x , x , x , x , x , x , 44 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
45, TIDL_BatchNormLayer , 1, 1 , 1 , 44 , x , x , x , x , x , x , x , 45 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
46, TIDL_ConvolutionLayer , 1, 1 , 1 , 45 , x , x , x , x , x , x , x , 46 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
47, TIDL_ConvolutionLayer , 1, 1 , 1 , 46 , x , x , x , x , x , x , x , 47 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
48, TIDL_ConvolutionLayer , 1, 1 , 1 , 47 , x , x , x , x , x , x , x , 48 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
49, TIDL_EltWiseLayer , 1, 2 , 1 , 48 , 45 , x , x , x , x , x , x , 49 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
50, TIDL_BatchNormLayer , 1, 1 , 1 , 49 , x , x , x , x , x , x , x , 50 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
51, TIDL_ConvolutionLayer , 1, 1 , 1 , 50 , x , x , x , x , x , x , x , 51 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
52, TIDL_ConvolutionLayer , 1, 1 , 1 , 51 , x , x , x , x , x , x , x , 52 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
53, TIDL_ConvolutionLayer , 1, 1 , 1 , 52 , x , x , x , x , x , x , x , 53 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
54, TIDL_EltWiseLayer , 1, 2 , 1 , 53 , 50 , x , x , x , x , x , x , 54 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
55, TIDL_BatchNormLayer , 1, 1 , 1 , 54 , x , x , x , x , x , x , x , 55 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
56, TIDL_ConvolutionLayer , 1, 1 , 1 , 55 , x , x , x , x , x , x , x , 56 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
57, TIDL_ConvolutionLayer , 1, 1 , 1 , 56 , x , x , x , x , x , x , x , 57 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
58, TIDL_ConvolutionLayer , 1, 1 , 1 , 57 , x , x , x , x , x , x , x , 58 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
59, TIDL_EltWiseLayer , 1, 2 , 1 , 58 , 55 , x , x , x , x , x , x , 59 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
60, TIDL_BatchNormLayer , 1, 1 , 1 , 59 , x , x , x , x , x , x , x , 60 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
61, TIDL_ConvolutionLayer , 1, 1 , 1 , 60 , x , x , x , x , x , x , x , 61 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
62, TIDL_ConvolutionLayer , 1, 1 , 1 , 61 , x , x , x , x , x , x , x , 62 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
63, TIDL_ConvolutionLayer , 1, 1 , 1 , 62 , x , x , x , x , x , x , x , 63 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
64, TIDL_EltWiseLayer , 1, 2 , 1 , 63 , 60 , x , x , x , x , x , x , 64 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
65, TIDL_BatchNormLayer , 1, 1 , 1 , 64 , x , x , x , x , x , x , x , 65 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
66, TIDL_ConvolutionLayer , 1, 1 , 1 , 65 , x , x , x , x , x , x , x , 66 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
67, TIDL_ConvolutionLayer , 1, 1 , 1 , 66 , x , x , x , x , x , x , x , 67 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
68, TIDL_ConvolutionLayer , 1, 1 , 1 , 67 , x , x , x , x , x , x , x , 68 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
69, TIDL_EltWiseLayer , 1, 2 , 1 , 68 , 65 , x , x , x , x , x , x , 69 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
70, TIDL_BatchNormLayer , 1, 1 , 1 , 69 , x , x , x , x , x , x , x , 70 , 1 , 368 , 4 , 4 , 1 , 368 , 4 , 4 ,
71, TIDL_DataLayer , 0, 1 , -1 , 70 , x , x , x , x , x , x , x , 0 , 1 , 368 , 4 , 4 , 0 , 0 , 0 , 0 ,
Layer ID ,inBlkWidth ,inBlkHeight ,inBlkPitch ,outBlkWidth ,outBlkHeight,outBlkPitch ,numInChs ,numOutChs ,numProcInChs,numLclInChs ,numLclOutChs,numProcItrs ,numAccItrs ,numHorBlock ,numVerBlock ,inBlkChPitch,outBlkChPitc,alignOrNot
1 40 34 40 32 32 32 3 32 3 1 8 1 3 1 1 1360 1024 1
2 32 32 32 32 32 32 32 24 32 7 8 1 5 1 1 1024 1024 1
3 40 36 40 16 16 16 8 8 8 4 8 1 2 1 1 1440 256 1
4 16 16 16 16 16 16 24 24 24 8 8 1 3 1 1 256 256 1
5 32 32 32 16 16 16 32 24 32 7 8 1 5 1 1 1024 256 1
8 16 16 16 16 16 16 24 56 24 8 8 1 3 1 1 256 256 1
9 24 18 24 16 16 16 8 8 8 4 8 1 2 1 1 432 256 1
10 16 16 16 16 16 16 56 56 56 8 8 1 7 1 1 256 256 1
11 16 16 16 16 16 16 24 56 24 8 8 1 3 1 1 256 256 1
14 16 16 16 16 16 16 56 152 56 8 8 1 7 1 1 256 256 1
15 40 20 40 16 8 16 8 8 8 4 8 1 2 1 1 800 128 1
16 16 8 16 16 8 16 152 152 152 8 8 1 19 1 1 128 128 1
17 32 16 32 16 8 16 56 152 56 8 8 1 7 1 1 512 128 1
20 16 8 16 16 8 16 152 152 152 8 8 1 19 1 1 128 128 1
21 24 10 24 16 8 16 8 8 8 4 8 1 2 1 1 240 128 1
22 16 8 16 16 8 16 152 152 152 8 8 1 19 1 1 128 128 1
25 16 8 16 16 8 16 152 152 152 8 8 1 19 1 1 128 128 1
26 24 10 24 16 8 16 8 8 8 4 8 1 2 1 1 240 128 1
27 16 8 16 16 8 16 152 152 152 8 8 1 19 1 1 128 128 1
30 16 8 16 16 8 16 152 152 152 8 8 1 19 1 1 128 128 1
31 24 10 24 16 8 16 8 8 8 4 8 1 2 1 1 240 128 1
32 16 8 16 16 8 16 152 152 152 8 8 1 19 1 1 128 128 1
35 16 8 16 16 8 16 152 368 152 8 8 1 19 1 1 128 128 1
36 40 12 40 16 4 16 8 8 8 4 8 1 2 1 1 480 64 1
37 16 4 16 16 4 16 368 368 368 8 8 1 46 1 1 64 64 1
38 32 8 32 16 4 16 152 368 152 8 8 1 19 1 1 256 64 1
41 16 4 16 16 4 16 368 368 368 8 8 1 46 1 1 64 64 1
42 24 6 24 16 4 16 8 8 8 4 8 1 2 1 1 144 64 1
43 16 4 16 16 4 16 368 368 368 8 8 1 46 1 1 64 64 1
46 16 4 16 16 4 16 368 368 368 8 8 1 46 1 1 64 64 1
47 24 6 24 16 4 16 8 8 8 4 8 1 2 1 1 144 64 1
48 16 4 16 16 4 16 368 368 368 8 8 1 46 1 1 64 64 1
51 16 4 16 16 4 16 368 368 368 8 8 1 46 1 1 64 64 1
52 24 6 24 16 4 16 8 8 8 4 8 1 2 1 1 144 64 1
53 16 4 16 16 4 16 368 368 368 8 8 1 46 1 1 64 64 1
56 16 4 16 16 4 16 368 368 368 8 8 1 46 1 1 64 64 1
57 24 6 24 16 4 16 8 8 8 4 8 1 2 1 1 144 64 1
58 16 4 16 16 4 16 368 368 368 8 8 1 46 1 1 64 64 1
61 16 4 16 16 4 16 368 368 368 8 8 1 46 1 1 64 64 1
62 24 6 24 16 4 16 8 8 8 4 8 1 2 1 1 144 64 1
63 16 4 16 16 4 16 368 368 368 8 8 1 46 1 1 64 64 1
66 16 4 16 16 4 16 368 368 368 8 8 1 46 1 1 64 64 1
67 24 6 24 16 4 16 8 8 8 4 8 1 2 1 1 144 64 1
68 16 4 16 16 4 16 368 368 368 8 8 1 46 1 1 64 64 1
Processing Frame Number : 0
Layer 1 : Out Q : 239 , TIDL_ConvolutionLayer, PASSED #MMACs = 0.88, 1.17, Sparsity : -32.41
Layer 2 : Out Q : 324 , TIDL_ConvolutionLayer, PASSED #MMACs = 0.79, 0.88, Sparsity : -12.50
Layer 3 : Out Q : 212 , TIDL_ConvolutionLayer, PASSED #MMACs = 0.44, 0.44, Sparsity : 0.00
Layer 4 : Out Q : 95 , TIDL_ConvolutionLayer, PASSED #MMACs = 0.15, 0.15, Sparsity : 0.00
Layer 5 : Out Q : 142 , TIDL_ConvolutionLayer, PASSED #MMACs = 0.20, 0.22, Sparsity : -12.50
Layer 6 : Out Q : 87 , TIDL_EltWiseLayer, PASSED #MMACs = 0.01, 0.01, Sparsity : 0.00
Layer 7 : Out Q : 175 , TIDL_BatchNormLayer , PASSED #MMACs = 0.01, 0.01, Sparsity : 0.00
Layer 8 : Out Q : 452 , TIDL_ConvolutionLayer, PASSED #MMACs = 0.34, 0.34, Sparsity : 0.00
Layer 9 : Out Q : 544 , TIDL_ConvolutionLayer, PASSED #MMACs = 1.03, 1.03, Sparsity : 0.00
Layer 10 : Out Q : 239 , TIDL_ConvolutionLayer, PASSED #MMACs = 0.80, 0.80, Sparsity : 0.00
Layer 11 : Out Q : 148 , TIDL_ConvolutionLayer, PASSED #MMACs = 0.34, 0.34, Sparsity : 0.00
Layer 12 : Out Q : 139 , TIDL_EltWiseLayer, PASSED #MMACs = 0.03, 0.03, Sparsity : 0.00
Layer 13 : Out Q : 279 , TIDL_BatchNormLayer , PASSED #MMACs = 0.01, 0.01, Sparsity : 0.00
Layer 14 : Out Q : 512 , TIDL_ConvolutionLayer, PASSED #MMACs = 2.18, 2.18, Sparsity : 0.00
Layer 15 : Out Q : 570 , TIDL_ConvolutionLayer, PASSED #MMACs = 0.70, 0.70, Sparsity : 0.00
Layer 16 : Out Q : 236 , TIDL_ConvolutionLayer, PASSED #MMACs = 1.48, 1.48, Sparsity : 0.00
Layer 17 : Out Q : 235 , TIDL_ConvolutionLayer, PASSED #MMACs = 0.54, 0.54, Sparsity : 0.00
Layer 18 : Out Q : 124 , TIDL_EltWiseLayer, PASSED #MMACs = 0.02, 0.02, Sparsity : 0.00
Layer 19 : Out Q : 249 , TIDL_BatchNormLayer , PASSED #MMACs = 0.01, 0.01, Sparsity : 0.00
Layer 20 : Out Q : 621 , TIDL_ConvolutionLayer, PASSED #MMACs = 1.48, 1.48, Sparsity : 0.00
Layer 21 : Out Q : 670 , TIDL_ConvolutionLayer, PASSED #MMACs = 0.70, 0.70, Sparsity : 0.00
Layer 22 : Out Q : 267 , TIDL_ConvolutionLayer, PASSED #MMACs = 1.48, 1.48, Sparsity : 0.00
Layer 23 : Out Q : 131 , TIDL_EltWiseLayer, PASSED #MMACs = 0.02, 0.02, Sparsity : 0.00
Layer 24 : Out Q : 263 , TIDL_BatchNormLayer , PASSED #MMACs = 0.01, 0.01, Sparsity : 0.00
Layer 25 : Out Q : 680 , TIDL_ConvolutionLayer, PASSED #MMACs = 1.48, 1.48, Sparsity : 0.00
Layer 26 : Out Q : 271 , TIDL_ConvolutionLayer, PASSED #MMACs = 0.70, 0.70, Sparsity : 0.00
Layer 27 : Out Q : 224 , TIDL_ConvolutionLayer, PASSED #MMACs = 1.48, 1.48, Sparsity : 0.00
Layer 28 : Out Q : 158 , TIDL_EltWiseLayer, PASSED #MMACs = 0.02, 0.02, Sparsity : 0.00
Layer 29 : Out Q : 317 , TIDL_BatchNormLayer , PASSED #MMACs = 0.01, 0.01, Sparsity : 0.00
Layer 30 : Out Q : 884 , TIDL_ConvolutionLayer, PASSED #MMACs = 1.48, 1.48, Sparsity : 0.00
Layer 31 : Out Q : 665 , TIDL_ConvolutionLayer, PASSED #MMACs = 0.70, 0.70, Sparsity : 0.00
Layer 32 : Out Q : 556 , TIDL_ConvolutionLayer, PASSED #MMACs = 1.48, 1.48, Sparsity : 0.00
Layer 33 : Out Q : 150 , TIDL_EltWiseLayer, PASSED #MMACs = 0.02, 0.02, Sparsity : 0.00
Layer 34 : Out Q : 301 , TIDL_BatchNormLayer , PASSED #MMACs = 0.01, 0.01, Sparsity : 0.00
Layer 35 : Out Q : 870 , TIDL_ConvolutionLayer, PASSED #MMACs = 3.58, 3.58, Sparsity : 0.00
Layer 36 : Out Q : 409 , TIDL_ConvolutionLayer, PASSED #MMACs = 0.42, 0.42, Sparsity : 0.00
Layer 37 : Out Q : 230 , TIDL_ConvolutionLayer, PASSED #MMACs = 2.17, 2.17, Sparsity : 0.00
Layer 38 : Out Q : 199 , TIDL_ConvolutionLayer, PASSED #MMACs = 0.89, 0.89, Sparsity : 0.00
Layer 39 : Out Q : 175 , TIDL_EltWiseLayer, PASSED #MMACs = 0.01, 0.01, Sparsity : 0.00
Layer 40 : Out Q : 351 , TIDL_BatchNormLayer , PASSED #MMACs = 0.01, 0.01, Sparsity : 0.00
Layer 41 : Out Q : 1196 , TIDL_ConvolutionLayer, PASSED #MMACs = 2.17, 2.17, Sparsity : 0.00
Layer 42 : Out Q : 483 , TIDL_ConvolutionLayer, PASSED #MMACs = 0.42, 0.42, Sparsity : 0.00
Layer 43 : Out Q : 220 , TIDL_ConvolutionLayer, PASSED #MMACs = 2.17, 2.17, Sparsity : 0.00
Layer 44 : Out Q : 158 , TIDL_EltWiseLayer, PASSED #MMACs = 0.01, 0.01, Sparsity : 0.00
Layer 45 : Out Q : 320 , TIDL_BatchNormLayer , PASSED #MMACs = 0.01, 0.01, Sparsity : 0.00
Layer 46 : Out Q : 1105 , TIDL_ConvolutionLayer, PASSED #MMACs = 2.17, 2.17, Sparsity : 0.00
Layer 47 : Out Q : 459 , TIDL_ConvolutionLayer, PASSED #MMACs = 0.42, 0.42, Sparsity : 0.00
Layer 48 : Out Q : 137 , TIDL_ConvolutionLayer, PASSED #MMACs = 2.17, 2.17, Sparsity : 0.00
Layer 49 : Out Q : 136 , TIDL_EltWiseLayer, PASSED #MMACs = 0.01, 0.01, Sparsity : 0.00
Layer 50 : Out Q : 287 , TIDL_BatchNormLayer , PASSED #MMACs = 0.01, 0.01, Sparsity : 0.00
Layer 51 : Out Q : 1024 , TIDL_ConvolutionLayer, PASSED #MMACs = 2.17, 2.17, Sparsity : 0.00
Layer 52 : Out Q : 631 , TIDL_ConvolutionLayer, PASSED #MMACs = 0.42, 0.42, Sparsity : 0.00
Layer 53 : Out Q : 178 , TIDL_ConvolutionLayer, PASSED #MMACs = 2.17, 2.17, Sparsity : 0.00
Layer 54 : Out Q : 109 , TIDL_EltWiseLayer, PASSED #MMACs = 0.01, 0.01, Sparsity : 0.00
Layer 55 : Out Q : 219 , TIDL_BatchNormLayer , PASSED #MMACs = 0.01, 0.01, Sparsity : 0.00
Layer 56 : Out Q : 1488 , TIDL_ConvolutionLayer, PASSED #MMACs = 2.17, 2.17, Sparsity : 0.00
Layer 57 : Out Q : 450 , TIDL_ConvolutionLayer, PASSED #MMACs = 0.42, 0.42, Sparsity : 0.00
Layer 58 : Out Q : 143 , TIDL_ConvolutionLayer, PASSED #MMACs = 2.17, 2.17, Sparsity : 0.00
Layer 59 : Out Q : 106 , TIDL_EltWiseLayer, PASSED #MMACs = 0.01, 0.01, Sparsity : 0.00
Layer 60 : Out Q : 213 , TIDL_BatchNormLayer , PASSED #MMACs = 0.01, 0.01, Sparsity : 0.00
Layer 61 : Out Q : 1198 , TIDL_ConvolutionLayer, PASSED #MMACs = 2.17, 2.17, Sparsity : 0.00
Layer 62 : Out Q : 556 , TIDL_ConvolutionLayer, PASSED #MMACs = 0.42, 0.42, Sparsity : 0.00
Layer 63 : Out Q : 150 , TIDL_ConvolutionLayer, PASSED #MMACs = 2.17, 2.17, Sparsity : 0.00
Layer 64 : Out Q : 102 , TIDL_EltWiseLayer, PASSED #MMACs = 0.01, 0.01, Sparsity : 0.00
Layer 65 : Out Q : 205 , TIDL_BatchNormLayer , PASSED #MMACs = 0.01, 0.01, Sparsity : 0.00
Layer 66 : Out Q : 1259 , TIDL_ConvolutionLayer, PASSED #MMACs = 2.17, 2.17, Sparsity : 0.00
Layer 67 : Out Q : 749 , TIDL_ConvolutionLayer, PASSED #MMACs = 0.42, 0.42, Sparsity : 0.00
Layer 68 : Out Q : 181 , TIDL_ConvolutionLayer, PASSED #MMACs = 2.17, 2.17, Sparsity : 0.00
Layer 69 : Out Q : 103 , TIDL_EltWiseLayer, PASSED #MMACs = 0.01, 0.01, Sparsity : 0.00
Layer 70 : Out Q : 207 , TIDL_BatchNormLayer , PASSED #MMACs = 0.01, 0.01, Sparsity : 0.00
Hi Praveen Is there any updates? lluo
Hi IIuo, I have looked at the log you shared, it looks okay to me, this configurations is supported on TDA2. Please follow steps mentioned in section 3.8 (Matching TIDL inference result) in the TIDL user guide. BTW, what is the TIDL release version that you are using? Thanks, Praveen
Hi Praveen I have been using REL.TIDL.01.01.03.00 . And the model & tools has been uploaded in regnet_import.zip. I will follow user guide to check my model on REL.TIDL.01.02.00.00 later.
Hi, Please check with REL.TIDL.01.02.00.00 and let us know. Thanks, Praveen
|