Part Number: J721EXSOMXEVM
This thread has been locked.
If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.
Part Number: J721EXSOMXEVM
Hi,
May i know which flow are you using ? TIDL-RT or OSRT ?
Is this observation is specific to 8 bit or 16 bit quantization ?
What was calibration related settings you have ?
How many images you have used ?
what was the calibration iteration ?
Which quantization flow you are using ?
Could you please help me with few details.
Log of model compilation
Import config file
What is confidence wrt to 16 bit ? for the given model ?
Hi, Pratic,
Log of model compilation:
TFLite Model (Flatbuf) File : ../../test/testvecs/models/quant_mobilenet_v2_high_reg_batched1.tflite
TIDL Network File : ../../test/testvecs/out/quant_model/180_tidl_net.bin
TIDL IO Info File : ../../test/testvecs/out/quant_model/180_tidl_io_
70
Running tidl_optimizeNet
printing Current net
0|TIDL_DataLayer | |serving_default_input_1:0_original | 0| 0|
1|TIDL_BatchNormLayer |serving_default_input_1:0_original |serving_default_input_1:0 | 0| 1|
2|TIDL_ConvolutionLayer |serving_default_input_1:0 |model/Conv1_relu/Relu6;model/bn_Conv1/FusedBatchNormV3;model/block_5_project/Conv2D;model/Conv1/Conv2D| 1| 2|
3|TIDL_ConvolutionLayer |model/Conv1_relu/Relu6;model/bn_Conv1/FusedBatchNormV3;model/block_5_project/Conv2D;model/Conv1/Conv2D|model/expanded_conv_depthwise_relu/Relu6;model/expanded_conv_depthwise_BN/FusedBatchNormV3;model/expanded_conv_depthwise/depthwise;model/block_5_project/Conv2D| 2| 3|
4|TIDL_ConvolutionLayer |model/expanded_conv_depthwise_relu/Relu6;model/expanded_conv_depthwise_BN/FusedBatchNormV3;model/expanded_conv_depthwise/depthwise;model/block_5_project/Conv2D|model/expanded_conv_project_BN/FusedBatchNormV3;model/expanded_conv_project/Conv2D| 3| 4|
5|TIDL_ConvolutionLayer |model/expanded_conv_project_BN/FusedBatchNormV3;model/expanded_conv_project/Conv2D|model/block_1_expand_relu/Relu6;model/block_1_expand_BN/FusedBatchNormV3;model/block_12_project/Conv2D;model/block_1_expand/Conv2D| 4| 5|
6|TIDL_ConvolutionLayer |model/block_1_expand_relu/Relu6;model/block_1_expand_BN/FusedBatchNormV3;model/block_12_project/Conv2D;model/block_1_expand/Conv2D|model/block_1_depthwise_relu/Relu6;model/block_1_depthwise_BN/FusedBatchNormV3;model/block_12_project/Conv2D;model/block_1_depthwise/depthwise| 5| 6|
7|TIDL_ConvolutionLayer |model/block_1_depthwise_relu/Relu6;model/block_1_depthwise_BN/FusedBatchNormV3;model/block_12_project/Conv2D;model/block_1_depthwise/depthwise|model/block_1_project_BN/FusedBatchNormV3;model/block_2_project/Conv2D;model/block_1_project/Conv2D| 6| 7|
8|TIDL_ConvolutionLayer |model/block_1_project_BN/FusedBatchNormV3;model/block_2_project/Conv2D;model/block_1_project/Conv2D|model/block_2_expand_relu/Relu6;model/block_2_expand_BN/FusedBatchNormV3;model/block_3_depthwise/depthwise;model/block_2_expand/Conv2D| 7| 8|
9|TIDL_ConvolutionLayer |model/block_2_expand_relu/Relu6;model/block_2_expand_BN/FusedBatchNormV3;model/block_3_depthwise/depthwise;model/block_2_expand/Conv2D|model/block_2_depthwise_relu/Relu6;model/block_2_depthwise_BN/FusedBatchNormV3;model/block_3_depthwise/depthwise;model/block_2_depthwise/depthwise| 8| 9|
10|TIDL_ConvolutionLayer |model/block_2_depthwise_relu/Relu6;model/block_2_depthwise_BN/FusedBatchNormV3;model/block_3_depthwise/depthwise;model/block_2_depthwise/depthwise|model/block_2_project_BN/FusedBatchNormV3;model/block_2_project/Conv2D| 9| 10|
11|TIDL_EltWiseLayer |model/block_1_project_BN/FusedBatchNormV3;model/block_2_project/Conv2D;model/block_1_project/Conv2D|model/block_2_add/add | 7| 11|
12|TIDL_ConvolutionLayer |model/block_2_add/add |model/block_3_expand_relu/Relu6;model/block_3_expand_BN/FusedBatchNormV3;model/block_3_depthwise/depthwise;model/block_3_expand/Conv2D| 11| 12|
13|TIDL_ConvolutionLayer |model/block_3_expand_relu/Relu6;model/block_3_expand_BN/FusedBatchNormV3;model/block_3_depthwise/depthwise;model/block_3_expand/Conv2D|model/block_3_depthwise_relu/Relu6;model/block_3_depthwise_BN/FusedBatchNormV3;model/block_3_depthwise/depthwise| 12| 13|
14|TIDL_ConvolutionLayer |model/block_3_depthwise_relu/Relu6;model/block_3_depthwise_BN/FusedBatchNormV3;model/block_3_depthwise/depthwise|model/block_3_project_BN/FusedBatchNormV3;model/block_5_project/Conv2D;model/block_3_project/Conv2D| 13| 14|
15|TIDL_ConvolutionLayer |model/block_3_project_BN/FusedBatchNormV3;model/block_5_project/Conv2D;model/block_3_project/Conv2D|model/block_4_expand_relu/Relu6;model/block_4_expand_BN/FusedBatchNormV3;model/block_6_depthwise/depthwise;model/block_4_expand/Conv2D| 14| 15|
16|TIDL_ConvolutionLayer |model/block_4_expand_relu/Relu6;model/block_4_expand_BN/FusedBatchNormV3;model/block_6_depthwise/depthwise;model/block_4_expand/Conv2D|model/block_4_depthwise_relu/Relu6;model/block_4_depthwise_BN/FusedBatchNormV3;model/block_6_depthwise/depthwise;model/block_4_depthwise/depthwise| 15| 16|
17|TIDL_ConvolutionLayer |model/block_4_depthwise_relu/Relu6;model/block_4_depthwise_BN/FusedBatchNormV3;model/block_6_depthwise/depthwise;model/block_4_depthwise/depthwise|model/block_4_project_BN/FusedBatchNormV3;model/block_5_project/Conv2D;model/block_4_project/Conv2D| 16| 17|
18|TIDL_EltWiseLayer |model/block_3_project_BN/FusedBatchNormV3;model/block_5_project/Conv2D;model/block_3_project/Conv2D|model/block_4_add/add | 14| 18|
19|TIDL_ConvolutionLayer |model/block_4_add/add |model/block_5_expand_relu/Relu6;model/block_5_expand_BN/FusedBatchNormV3;model/block_6_depthwise/depthwise;model/block_5_expand/Conv2D| 18| 19|
20|TIDL_ConvolutionLayer |model/block_5_expand_relu/Relu6;model/block_5_expand_BN/FusedBatchNormV3;model/block_6_depthwise/depthwise;model/block_5_expand/Conv2D|model/block_5_depthwise_relu/Relu6;model/block_5_depthwise_BN/FusedBatchNormV3;model/block_6_depthwise/depthwise;model/block_5_depthwise/depthwise| 19| 20|
21|TIDL_ConvolutionLayer |model/block_5_depthwise_relu/Relu6;model/block_5_depthwise_BN/FusedBatchNormV3;model/block_6_depthwise/depthwise;model/block_5_depthwise/depthwise|model/block_5_project_BN/FusedBatchNormV3;model/block_5_project/Conv2D| 20| 21|
22|TIDL_EltWiseLayer |model/block_4_add/add |model/block_5_add/add | 18| 22|
23|TIDL_ConvolutionLayer |model/block_5_add/add |model/block_6_expand_relu/Relu6;model/block_6_expand_BN/FusedBatchNormV3;model/block_6_depthwise/depthwise;model/block_6_expand/Conv2D| 22| 23|
24|TIDL_ConvolutionLayer |model/block_6_expand_relu/Relu6;model/block_6_expand_BN/FusedBatchNormV3;model/block_6_depthwise/depthwise;model/block_6_expand/Conv2D|model/block_6_depthwise_relu/Relu6;model/block_6_depthwise_BN/FusedBatchNormV3;model/block_6_depthwise/depthwise| 23| 24|
25|TIDL_ConvolutionLayer |model/block_6_depthwise_relu/Relu6;model/block_6_depthwise_BN/FusedBatchNormV3;model/block_6_depthwise/depthwise|model/block_6_project_BN/FusedBatchNormV3;model/block_9_project/Conv2D;model/block_6_project/Conv2D| 24| 25|
26|TIDL_ConvolutionLayer |model/block_6_project_BN/FusedBatchNormV3;model/block_9_project/Conv2D;model/block_6_project/Conv2D|model/block_7_expand_relu/Relu6;model/block_7_expand_BN/FusedBatchNormV3;model/block_10_depthwise/depthwise;model/block_7_expand/Conv2D| 25| 26|
27|TIDL_ConvolutionLayer |model/block_7_expand_relu/Relu6;model/block_7_expand_BN/FusedBatchNormV3;model/block_10_depthwise/depthwise;model/block_7_expand/Conv2D|model/block_7_depthwise_relu/Relu6;model/block_7_depthwise_BN/FusedBatchNormV3;model/block_10_depthwise/depthwise;model/block_7_depthwise/depthwise| 26| 27|
28|TIDL_ConvolutionLayer |model/block_7_depthwise_relu/Relu6;model/block_7_depthwise_BN/FusedBatchNormV3;model/block_10_depthwise/depthwise;model/block_7_depthwise/depthwise|model/block_7_project_BN/FusedBatchNormV3;model/block_9_project/Conv2D;model/block_7_project/Conv2D| 27| 28|
29|TIDL_EltWiseLayer |model/block_6_project_BN/FusedBatchNormV3;model/block_9_project/Conv2D;model/block_6_project/Conv2D|model/block_7_add/add | 25| 29|
30|TIDL_ConvolutionLayer |model/block_7_add/add |model/block_8_expand_relu/Relu6;model/block_8_expand_BN/FusedBatchNormV3;model/block_10_depthwise/depthwise;model/block_8_expand/Conv2D| 29| 30|
31|TIDL_ConvolutionLayer |model/block_8_expand_relu/Relu6;model/block_8_expand_BN/FusedBatchNormV3;model/block_10_depthwise/depthwise;model/block_8_expand/Conv2D|model/block_8_depthwise_relu/Relu6;model/block_8_depthwise_BN/FusedBatchNormV3;model/block_10_depthwise/depthwise;model/block_8_depthwise/depthwise| 30| 31|
32|TIDL_ConvolutionLayer |model/block_8_depthwise_relu/Relu6;model/block_8_depthwise_BN/FusedBatchNormV3;model/block_10_depthwise/depthwise;model/block_8_depthwise/depthwise|model/block_8_project_BN/FusedBatchNormV3;model/block_9_project/Conv2D;model/block_8_project/Conv2D| 31| 32|
33|TIDL_EltWiseLayer |model/block_7_add/add |model/block_8_add/add | 29| 33|
34|TIDL_ConvolutionLayer |model/block_8_add/add |model/block_9_expand_relu/Relu6;model/block_9_expand_BN/FusedBatchNormV3;model/block_10_depthwise/depthwise;model/block_9_expand/Conv2D| 33| 34|
35|TIDL_ConvolutionLayer |model/block_9_expand_relu/Relu6;model/block_9_expand_BN/FusedBatchNormV3;model/block_10_depthwise/depthwise;model/block_9_expand/Conv2D|model/block_9_depthwise_relu/Relu6;model/block_9_depthwise_BN/FusedBatchNormV3;model/block_10_depthwise/depthwise;model/block_9_depthwise/depthwise| 34| 35|
36|TIDL_ConvolutionLayer |model/block_9_depthwise_relu/Relu6;model/block_9_depthwise_BN/FusedBatchNormV3;model/block_10_depthwise/depthwise;model/block_9_depthwise/depthwise|model/block_9_project_BN/FusedBatchNormV3;model/block_9_project/Conv2D| 35| 36|
37|TIDL_EltWiseLayer |model/block_8_add/add |model/block_9_add/add | 33| 37|
38|TIDL_ConvolutionLayer |model/block_9_add/add |model/block_10_expand_relu/Relu6;model/block_10_expand_BN/FusedBatchNormV3;model/block_10_depthwise/depthwise;model/block_10_expand/Conv2D| 37| 38|
39|TIDL_ConvolutionLayer |model/block_10_expand_relu/Relu6;model/block_10_expand_BN/FusedBatchNormV3;model/block_10_depthwise/depthwise;model/block_10_expand/Conv2D|model/block_10_depthwise_relu/Relu6;model/block_10_depthwise_BN/FusedBatchNormV3;model/block_10_depthwise/depthwise| 38| 39|
40|TIDL_ConvolutionLayer |model/block_10_depthwise_relu/Relu6;model/block_10_depthwise_BN/FusedBatchNormV3;model/block_10_depthwise/depthwise|model/block_10_project_BN/FusedBatchNormV3;model/block_12_project/Conv2D;model/block_10_project/Conv2D| 39| 40|
41|TIDL_ConvolutionLayer |model/block_10_project_BN/FusedBatchNormV3;model/block_12_project/Conv2D;model/block_10_project/Conv2D|model/block_11_expand_relu/Relu6;model/block_11_expand_BN/FusedBatchNormV3;model/block_13_depthwise/depthwise;model/block_11_expand/Conv2D| 40| 41|
42|TIDL_ConvolutionLayer |model/block_11_expand_relu/Relu6;model/block_11_expand_BN/FusedBatchNormV3;model/block_13_depthwise/depthwise;model/block_11_expand/Conv2D|model/block_11_depthwise_relu/Relu6;model/block_11_depthwise_BN/FusedBatchNormV3;model/block_13_depthwise/depthwise;model/block_11_depthwise/depthwise| 41| 42|
43|TIDL_ConvolutionLayer |model/block_11_depthwise_relu/Relu6;model/block_11_depthwise_BN/FusedBatchNormV3;model/block_13_depthwise/depthwise;model/block_11_depthwise/depthwise|model/block_11_project_BN/FusedBatchNormV3;model/block_12_project/Conv2D;model/block_11_project/Conv2D| 42| 43|
44|TIDL_EltWiseLayer |model/block_10_project_BN/FusedBatchNormV3;model/block_12_project/Conv2D;model/block_10_project/Conv2D|model/block_11_add/add | 40| 44|
45|TIDL_ConvolutionLayer |model/block_11_add/add |model/block_12_expand_relu/Relu6;model/block_12_expand_BN/FusedBatchNormV3;model/block_13_depthwise/depthwise;model/block_12_expand/Conv2D| 44| 45|
46|TIDL_ConvolutionLayer |model/block_12_expand_relu/Relu6;model/block_12_expand_BN/FusedBatchNormV3;model/block_13_depthwise/depthwise;model/block_12_expand/Conv2D|model/block_12_depthwise_relu/Relu6;model/block_12_depthwise_BN/FusedBatchNormV3;model/block_13_depthwise/depthwise;model/block_12_depthwise/depthwise| 45| 46|
47|TIDL_ConvolutionLayer |model/block_12_depthwise_relu/Relu6;model/block_12_depthwise_BN/FusedBatchNormV3;model/block_13_depthwise/depthwise;model/block_12_depthwise/depthwise|model/block_12_project_BN/FusedBatchNormV3;model/block_12_project/Conv2D| 46| 47|
48|TIDL_EltWiseLayer |model/block_11_add/add |model/block_12_add/add | 44| 48|
49|TIDL_ConvolutionLayer |model/block_12_add/add |model/block_13_expand_relu/Relu6;model/block_13_expand_BN/FusedBatchNormV3;model/block_13_depthwise/depthwise;model/block_13_expand/Conv2D| 48| 49|
50|TIDL_ConvolutionLayer |model/block_13_expand_relu/Relu6;model/block_13_expand_BN/FusedBatchNormV3;model/block_13_depthwise/depthwise;model/block_13_expand/Conv2D|model/block_13_depthwise_relu/Relu6;model/block_13_depthwise_BN/FusedBatchNormV3;model/block_13_depthwise/depthwise| 49| 50|
51|TIDL_ConvolutionLayer |model/block_13_depthwise_relu/Relu6;model/block_13_depthwise_BN/FusedBatchNormV3;model/block_13_depthwise/depthwise|model/block_13_project_BN/FusedBatchNormV3;model/block_15_project/Conv2D;model/block_13_project/Conv2D| 50| 51|
52|TIDL_ConvolutionLayer |model/block_13_project_BN/FusedBatchNormV3;model/block_15_project/Conv2D;model/block_13_project/Conv2D|model/block_14_expand_relu/Relu6;model/block_14_expand_BN/FusedBatchNormV3;model/block_16_depthwise/depthwise;model/block_14_expand/Conv2D| 51| 52|
53|TIDL_ConvolutionLayer |model/block_14_expand_relu/Relu6;model/block_14_expand_BN/FusedBatchNormV3;model/block_16_depthwise/depthwise;model/block_14_expand/Conv2D|model/block_14_depthwise_relu/Relu6;model/block_14_depthwise_BN/FusedBatchNormV3;model/block_16_depthwise/depthwise;model/block_14_depthwise/depthwise| 52| 53|
54|TIDL_ConvolutionLayer |model/block_14_depthwise_relu/Relu6;model/block_14_depthwise_BN/FusedBatchNormV3;model/block_16_depthwise/depthwise;model/block_14_depthwise/depthwise|model/block_14_project_BN/FusedBatchNormV3;model/block_15_project/Conv2D;model/block_14_project/Conv2D| 53| 54|
55|TIDL_EltWiseLayer |model/block_13_project_BN/FusedBatchNormV3;model/block_15_project/Conv2D;model/block_13_project/Conv2D|model/block_14_add/add | 51| 55|
56|TIDL_ConvolutionLayer |model/block_14_add/add |model/block_15_expand_relu/Relu6;model/block_15_expand_BN/FusedBatchNormV3;model/block_16_depthwise/depthwise;model/block_15_expand/Conv2D| 55| 56|
57|TIDL_ConvolutionLayer |model/block_15_expand_relu/Relu6;model/block_15_expand_BN/FusedBatchNormV3;model/block_16_depthwise/depthwise;model/block_15_expand/Conv2D|model/block_15_depthwise_relu/Relu6;model/block_15_depthwise_BN/FusedBatchNormV3;model/block_16_depthwise/depthwise;model/block_15_depthwise/depthwise| 56| 57|
58|TIDL_ConvolutionLayer |model/block_15_depthwise_relu/Relu6;model/block_15_depthwise_BN/FusedBatchNormV3;model/block_16_depthwise/depthwise;model/block_15_depthwise/depthwise|model/block_15_project_BN/FusedBatchNormV3;model/block_15_project/Conv2D| 57| 58|
59|TIDL_EltWiseLayer |model/block_14_add/add |model/block_15_add/add | 55| 59|
60|TIDL_ConvolutionLayer |model/block_15_add/add |model/block_16_expand_relu/Relu6;model/block_16_expand_BN/FusedBatchNormV3;model/block_16_depthwise/depthwise;model/block_16_expand/Conv2D| 59| 60|
61|TIDL_ConvolutionLayer |model/block_16_expand_relu/Relu6;model/block_16_expand_BN/FusedBatchNormV3;model/block_16_depthwise/depthwise;model/block_16_expand/Conv2D|model/block_16_depthwise_relu/Relu6;model/block_16_depthwise_BN/FusedBatchNormV3;model/block_16_depthwise/depthwise| 60| 61|
62|TIDL_ConvolutionLayer |model/block_16_depthwise_relu/Relu6;model/block_16_depthwise_BN/FusedBatchNormV3;model/block_16_depthwise/depthwise|model/block_16_project_BN/FusedBatchNormV3;model/block_16_project/Conv2D| 61| 62|
63|TIDL_ConvolutionLayer |model/block_16_project_BN/FusedBatchNormV3;model/block_16_project/Conv2D|model/out_relu/Relu6;model/Conv_1_bn/FusedBatchNormV3;model/Conv_1/Conv2D| 62| 63|
64|TIDL_PoolingLayer |model/out_relu/Relu6;model/Conv_1_bn/FusedBatchNormV3;model/Conv_1/Conv2D|model/global_average_pooling2d/Mean | 63| 64|
65|TIDL_InnerProductLayer |model/global_average_pooling2d/Mean |model/dense/MatMul;model/dense/BiasAdd;model/batch_normalization_154/batchnorm/mul_1;model/tf.nn.relu/Relu;model/batch_normalization_154/batchnorm/add_1| 64| 65|
66|TIDL_InnerProductLayer |model/dense/MatMul;model/dense/BiasAdd;model/batch_normalization_154/batchnorm/mul_1;model/tf.nn.relu/Relu;model/batch_normalization_154/batchnorm/add_1|model/dense_1/MatMul;model/dense_1/BiasAdd | 65| 66|
67|TIDL_SoftMaxLayer |model/dense_1/MatMul;model/dense_1/BiasAdd |StatefulPartitionedCall:0 | 66| 67|
68|TIDL_DataLayer |StatefulPartitionedCall:0 |StatefulPartitionedCall:0 | 67| 0|
Completed tidl_optimizeNet
Empty prototxt path, running calibration
Num of Layer Detected : 69
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Num|TIDL Layer Name |Out Data Name |Group |#Ins |#Outs |Inbuf Ids |Outbuf Id |In NCHW |Out NCHW |MACS |
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
0|TIDL_DataLayer |serving_default_input_1:0_original | 0| -1| 1| x x x x x x x x | 0 | 0 0 0 0 | 1 3 224 224 | 0 |
1|TIDL_BatchNormLayer |serving_default_input_1:0 | 0| 1| 1| 0 x x x x x x x | 1 | 1 3 224 224 | 1 3 224 224 | 150528 |
2|TIDL_ConvolutionLayer |V3;model/block_5_project/Conv2D;model/Conv1/Conv2D| 0| 1| 1| 1 x x x x x x x | 2 | 1 3 224 224 | 1 32 112 112 | 10838016 |
3|TIDL_ConvolutionLayer |v_depthwise/depthwise;model/block_5_project/Conv2D| 0| 1| 1| 2 x x x x x x x | 3 | 1 32 112 112 | 1 32 112 112 | 3612672 |
4|TIDL_ConvolutionLayer |usedBatchNormV3;model/expanded_conv_project/Conv2D| 0| 1| 1| 3 x x x x x x x | 4 | 1 32 112 112 | 1 16 112 112 | 6422528 |
5|TIDL_ConvolutionLayer |lock_12_project/Conv2D;model/block_1_expand/Conv2D| 0| 1| 1| 4 x x x x x x x | 5 | 1 16 112 112 | 1 96 112 112 | 19267584 |
6|TIDL_ConvolutionLayer |2_project/Conv2D;model/block_1_depthwise/depthwise| 0| 1| 1| 5 x x x x x x x | 6 | 1 96 112 112 | 1 96 56 56 | 2709504 |
7|TIDL_ConvolutionLayer |lock_2_project/Conv2D;model/block_1_project/Conv2D| 0| 1| 1| 6 x x x x x x x | 7 | 1 96 56 56 | 1 24 56 56 | 7225344 |
8|TIDL_ConvolutionLayer |_3_depthwise/depthwise;model/block_2_expand/Conv2D| 0| 1| 1| 7 x x x x x x x | 8 | 1 24 56 56 | 1 144 56 56 | 10838016 |
9|TIDL_ConvolutionLayer |thwise/depthwise;model/block_2_depthwise/depthwise| 0| 1| 1| 8 x x x x x x x | 9 | 1 144 56 56 | 1 144 56 56 | 4064256 |
10|TIDL_ConvolutionLayer |t_BN/FusedBatchNormV3;model/block_2_project/Conv2D| 0| 1| 1| 9 x x x x x x x | 10 | 1 144 56 56 | 1 24 56 56 | 10838016 |
11|TIDL_EltWiseLayer |model/block_2_add/add | 0| 2| 1| 7 10 x x x x x x | 11 | 1 24 56 56 | 1 24 56 56 | 75264 |
12|TIDL_ConvolutionLayer |_3_depthwise/depthwise;model/block_3_expand/Conv2D| 0| 1| 1| 11 x x x x x x x | 12 | 1 24 56 56 | 1 144 56 56 | 10838016 |
13|TIDL_ConvolutionLayer |FusedBatchNormV3;model/block_3_depthwise/depthwise| 0| 1| 1| 12 x x x x x x x | 13 | 1 144 56 56 | 1 144 28 28 | 1016064 |
14|TIDL_ConvolutionLayer |lock_5_project/Conv2D;model/block_3_project/Conv2D| 0| 1| 1| 13 x x x x x x x | 14 | 1 144 28 28 | 1 32 28 28 | 3612672 |
15|TIDL_ConvolutionLayer |_6_depthwise/depthwise;model/block_4_expand/Conv2D| 0| 1| 1| 14 x x x x x x x | 15 | 1 32 28 28 | 1 192 28 28 | 4816896 |
16|TIDL_ConvolutionLayer |thwise/depthwise;model/block_4_depthwise/depthwise| 0| 1| 1| 15 x x x x x x x | 16 | 1 192 28 28 | 1 192 28 28 | 1354752 |
17|TIDL_ConvolutionLayer |lock_5_project/Conv2D;model/block_4_project/Conv2D| 0| 1| 1| 16 x x x x x x x | 17 | 1 192 28 28 | 1 32 28 28 | 4816896 |
18|TIDL_EltWiseLayer |model/block_4_add/add | 0| 2| 1| 14 17 x x x x x x | 18 | 1 32 28 28 | 1 32 28 28 | 25088 |
19|TIDL_ConvolutionLayer |_6_depthwise/depthwise;model/block_5_expand/Conv2D| 0| 1| 1| 18 x x x x x x x | 19 | 1 32 28 28 | 1 192 28 28 | 4816896 |
20|TIDL_ConvolutionLayer |thwise/depthwise;model/block_5_depthwise/depthwise| 0| 1| 1| 19 x x x x x x x | 20 | 1 192 28 28 | 1 192 28 28 | 1354752 |
21|TIDL_ConvolutionLayer |t_BN/FusedBatchNormV3;model/block_5_project/Conv2D| 0| 1| 1| 20 x x x x x x x | 21 | 1 192 28 28 | 1 32 28 28 | 4816896 |
22|TIDL_EltWiseLayer |model/block_5_add/add | 0| 2| 1| 18 21 x x x x x x | 22 | 1 32 28 28 | 1 32 28 28 | 25088 |
23|TIDL_ConvolutionLayer |_6_depthwise/depthwise;model/block_6_expand/Conv2D| 0| 1| 1| 22 x x x x x x x | 23 | 1 32 28 28 | 1 192 28 28 | 4816896 |
24|TIDL_ConvolutionLayer |FusedBatchNormV3;model/block_6_depthwise/depthwise| 0| 1| 1| 23 x x x x x x x | 24 | 1 192 28 28 | 1 192 14 14 | 338688 |
25|TIDL_ConvolutionLayer |lock_9_project/Conv2D;model/block_6_project/Conv2D| 0| 1| 1| 24 x x x x x x x | 25 | 1 192 14 14 | 1 64 14 14 | 2408448 |
26|TIDL_ConvolutionLayer |10_depthwise/depthwise;model/block_7_expand/Conv2D| 0| 1| 1| 25 x x x x x x x | 26 | 1 64 14 14 | 1 384 14 14 | 4816896 |
27|TIDL_ConvolutionLayer |thwise/depthwise;model/block_7_depthwise/depthwise| 0| 1| 1| 26 x x x x x x x | 27 | 1 384 14 14 | 1 384 14 14 | 677376 |
28|TIDL_ConvolutionLayer |lock_9_project/Conv2D;model/block_7_project/Conv2D| 0| 1| 1| 27 x x x x x x x | 28 | 1 384 14 14 | 1 64 14 14 | 4816896 |
29|TIDL_EltWiseLayer |model/block_7_add/add | 0| 2| 1| 25 28 x x x x x x | 29 | 1 64 14 14 | 1 64 14 14 | 12544 |
30|TIDL_ConvolutionLayer |10_depthwise/depthwise;model/block_8_expand/Conv2D| 0| 1| 1| 29 x x x x x x x | 30 | 1 64 14 14 | 1 384 14 14 | 4816896 |
31|TIDL_ConvolutionLayer |thwise/depthwise;model/block_8_depthwise/depthwise| 0| 1| 1| 30 x x x x x x x | 31 | 1 384 14 14 | 1 384 14 14 | 677376 |
32|TIDL_ConvolutionLayer |lock_9_project/Conv2D;model/block_8_project/Conv2D| 0| 1| 1| 31 x x x x x x x | 32 | 1 384 14 14 | 1 64 14 14 | 4816896 |
33|TIDL_EltWiseLayer |model/block_8_add/add | 0| 2| 1| 29 32 x x x x x x | 33 | 1 64 14 14 | 1 64 14 14 | 12544 |
34|TIDL_ConvolutionLayer |10_depthwise/depthwise;model/block_9_expand/Conv2D| 0| 1| 1| 33 x x x x x x x | 34 | 1 64 14 14 | 1 384 14 14 | 4816896 |
35|TIDL_ConvolutionLayer |thwise/depthwise;model/block_9_depthwise/depthwise| 0| 1| 1| 34 x x x x x x x | 35 | 1 384 14 14 | 1 384 14 14 | 677376 |
36|TIDL_ConvolutionLayer |t_BN/FusedBatchNormV3;model/block_9_project/Conv2D| 0| 1| 1| 35 x x x x x x x | 36 | 1 384 14 14 | 1 64 14 14 | 4816896 |
37|TIDL_EltWiseLayer |model/block_9_add/add | 0| 2| 1| 33 36 x x x x x x | 37 | 1 64 14 14 | 1 64 14 14 | 12544 |
38|TIDL_ConvolutionLayer |0_depthwise/depthwise;model/block_10_expand/Conv2D| 0| 1| 1| 37 x x x x x x x | 38 | 1 64 14 14 | 1 384 14 14 | 4816896 |
39|TIDL_ConvolutionLayer |usedBatchNormV3;model/block_10_depthwise/depthwise| 0| 1| 1| 38 x x x x x x x | 39 | 1 384 14 14 | 1 384 14 14 | 677376 |
40|TIDL_ConvolutionLayer |ck_12_project/Conv2D;model/block_10_project/Conv2D| 0| 1| 1| 39 x x x x x x x | 40 | 1 384 14 14 | 1 96 14 14 | 7225344 |
41|TIDL_ConvolutionLayer |3_depthwise/depthwise;model/block_11_expand/Conv2D| 0| 1| 1| 40 x x x x x x x | 41 | 1 96 14 14 | 1 576 14 14 | 10838016 |
42|TIDL_ConvolutionLayer |hwise/depthwise;model/block_11_depthwise/depthwise| 0| 1| 1| 41 x x x x x x x | 42 | 1 576 14 14 | 1 576 14 14 | 1016064 |
43|TIDL_ConvolutionLayer |ck_12_project/Conv2D;model/block_11_project/Conv2D| 0| 1| 1| 42 x x x x x x x | 43 | 1 576 14 14 | 1 96 14 14 | 10838016 |
44|TIDL_EltWiseLayer |model/block_11_add/add | 0| 2| 1| 40 43 x x x x x x | 44 | 1 96 14 14 | 1 96 14 14 | 18816 |
45|TIDL_ConvolutionLayer |3_depthwise/depthwise;model/block_12_expand/Conv2D| 0| 1| 1| 44 x x x x x x x | 45 | 1 96 14 14 | 1 576 14 14 | 10838016 |
46|TIDL_ConvolutionLayer |hwise/depthwise;model/block_12_depthwise/depthwise| 0| 1| 1| 45 x x x x x x x | 46 | 1 576 14 14 | 1 576 14 14 | 1016064 |
47|TIDL_ConvolutionLayer |_BN/FusedBatchNormV3;model/block_12_project/Conv2D| 0| 1| 1| 46 x x x x x x x | 47 | 1 576 14 14 | 1 96 14 14 | 10838016 |
48|TIDL_EltWiseLayer |model/block_12_add/add | 0| 2| 1| 44 47 x x x x x x | 48 | 1 96 14 14 | 1 96 14 14 | 18816 |
49|TIDL_ConvolutionLayer |3_depthwise/depthwise;model/block_13_expand/Conv2D| 0| 1| 1| 48 x x x x x x x | 49 | 1 96 14 14 | 1 576 14 14 | 10838016 |
50|TIDL_ConvolutionLayer |usedBatchNormV3;model/block_13_depthwise/depthwise| 0| 1| 1| 49 x x x x x x x | 50 | 1 576 14 14 | 1 576 7 7 | 254016 |
51|TIDL_ConvolutionLayer |ck_15_project/Conv2D;model/block_13_project/Conv2D| 0| 1| 1| 50 x x x x x x x | 51 | 1 576 7 7 | 1 160 7 7 | 4515840 |
52|TIDL_ConvolutionLayer |6_depthwise/depthwise;model/block_14_expand/Conv2D| 0| 1| 1| 51 x x x x x x x | 52 | 1 160 7 7 | 1 960 7 7 | 7526400 |
53|TIDL_ConvolutionLayer |hwise/depthwise;model/block_14_depthwise/depthwise| 0| 1| 1| 52 x x x x x x x | 53 | 1 960 7 7 | 1 960 7 7 | 423360 |
54|TIDL_ConvolutionLayer |ck_15_project/Conv2D;model/block_14_project/Conv2D| 0| 1| 1| 53 x x x x x x x | 54 | 1 960 7 7 | 1 160 7 7 | 7526400 |
55|TIDL_EltWiseLayer |model/block_14_add/add | 0| 2| 1| 51 54 x x x x x x | 55 | 1 160 7 7 | 1 160 7 7 | 7840 |
56|TIDL_ConvolutionLayer |6_depthwise/depthwise;model/block_15_expand/Conv2D| 0| 1| 1| 55 x x x x x x x | 56 | 1 160 7 7 | 1 960 7 7 | 7526400 |
57|TIDL_ConvolutionLayer |hwise/depthwise;model/block_15_depthwise/depthwise| 0| 1| 1| 56 x x x x x x x | 57 | 1 960 7 7 | 1 960 7 7 | 423360 |
58|TIDL_ConvolutionLayer |_BN/FusedBatchNormV3;model/block_15_project/Conv2D| 0| 1| 1| 57 x x x x x x x | 58 | 1 960 7 7 | 1 160 7 7 | 7526400 |
59|TIDL_EltWiseLayer |model/block_15_add/add | 0| 2| 1| 55 58 x x x x x x | 59 | 1 160 7 7 | 1 160 7 7 | 7840 |
60|TIDL_ConvolutionLayer |6_depthwise/depthwise;model/block_16_expand/Conv2D| 0| 1| 1| 59 x x x x x x x | 60 | 1 160 7 7 | 1 960 7 7 | 7526400 |
61|TIDL_ConvolutionLayer |usedBatchNormV3;model/block_16_depthwise/depthwise| 0| 1| 1| 60 x x x x x x x | 61 | 1 960 7 7 | 1 960 7 7 | 423360 |
62|TIDL_ConvolutionLayer |_BN/FusedBatchNormV3;model/block_16_project/Conv2D| 0| 1| 1| 61 x x x x x x x | 62 | 1 960 7 7 | 1 320 7 7 | 15052800 |
63|TIDL_ConvolutionLayer |del/Conv_1_bn/FusedBatchNormV3;model/Conv_1/Conv2D| 0| 1| 1| 62 x x x x x x x | 63 | 1 320 7 7 | 1 1280 7 7 | 20070400 |
64|TIDL_PoolingLayer |model/global_average_pooling2d/Mean | 0| 1| 1| 63 x x x x x x x | 64 | 1 1280 7 7 | 1 1 1 1280 | 1280 |
65|TIDL_InnerProductLayer |Relu;model/batch_normalization_154/batchnorm/add_1| 0| 1| 1| 64 x x x x x x x | 65 | 1 1 1 1280 | 1 1 1 100 | 128100 |
66|TIDL_InnerProductLayer |model/dense_1/MatMul;model/dense_1/BiasAdd | 0| 1| 1| 65 x x x x x x x | 66 | 1 1 1 100 | 1 1 1 2 | 202 |
67|TIDL_SoftMaxLayer |StatefulPartitionedCall:0 | 0| 1| 1| 66 x x x x x x x | 67 | 1 1 1 2 | 1 1 1 2 | 2 |
68|TIDL_DataLayer |StatefulPartitionedCall:0 | 0| 1| -1| 67 x x x x x x x | 0 | 1 1 1 2 | 0 0 0 0 | 0 |
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Total Giga Macs : 0.3000
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Num of Layer Detected : 69
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Num|TIDL Layer Name |Out Data Name |Group |#Ins |#Outs |Inbuf Ids |Outbuf Id |In NCHW |Out NCHW |MACS |
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
0|TIDL_DataLayer |serving_default_input_1:0_original | 0| -1| 1| x x x x x x x x | 0 | 0 0 0 0 | 1 3 224 224 | 0 |
1|TIDL_BatchNormLayer |serving_default_input_1:0 | 0| 1| 1| 0 x x x x x x x | 1 | 1 3 224 224 | 1 3 224 224 | 150528 |
2|TIDL_ConvolutionLayer |V3;model/block_5_project/Conv2D;model/Conv1/Conv2D| 0| 1| 1| 1 x x x x x x x | 2 | 1 3 224 224 | 1 32 112 112 | 10838016 |
3|TIDL_ConvolutionLayer |v_depthwise/depthwise;model/block_5_project/Conv2D| 0| 1| 1| 2 x x x x x x x | 3 | 1 32 112 112 | 1 32 112 112 | 3612672 |
4|TIDL_ConvolutionLayer |usedBatchNormV3;model/expanded_conv_project/Conv2D| 0| 1| 1| 3 x x x x x x x | 4 | 1 32 112 112 | 1 16 112 112 | 6422528 |
5|TIDL_ConvolutionLayer |lock_12_project/Conv2D;model/block_1_expand/Conv2D| 0| 1| 1| 4 x x x x x x x | 5 | 1 16 112 112 | 1 96 112 112 | 19267584 |
6|TIDL_ConvolutionLayer |2_project/Conv2D;model/block_1_depthwise/depthwise| 0| 1| 1| 5 x x x x x x x | 6 | 1 96 112 112 | 1 96 56 56 | 2709504 |
7|TIDL_ConvolutionLayer |lock_2_project/Conv2D;model/block_1_project/Conv2D| 0| 1| 1| 6 x x x x x x x | 7 | 1 96 56 56 | 1 24 56 56 | 7225344 |
8|TIDL_ConvolutionLayer |_3_depthwise/depthwise;model/block_2_expand/Conv2D| 0| 1| 1| 7 x x x x x x x | 8 | 1 24 56 56 | 1 144 56 56 | 10838016 |
9|TIDL_ConvolutionLayer |thwise/depthwise;model/block_2_depthwise/depthwise| 0| 1| 1| 8 x x x x x x x | 9 | 1 144 56 56 | 1 144 56 56 | 4064256 |
10|TIDL_ConvolutionLayer |t_BN/FusedBatchNormV3;model/block_2_project/Conv2D| 0| 1| 1| 9 x x x x x x x | 10 | 1 144 56 56 | 1 24 56 56 | 10838016 |
11|TIDL_EltWiseLayer |model/block_2_add/add | 0| 2| 1| 7 10 x x x x x x | 11 | 1 24 56 56 | 1 24 56 56 | 75264 |
12|TIDL_ConvolutionLayer |_3_depthwise/depthwise;model/block_3_expand/Conv2D| 0| 1| 1| 11 x x x x x x x | 12 | 1 24 56 56 | 1 144 56 56 | 10838016 |
13|TIDL_ConvolutionLayer |FusedBatchNormV3;model/block_3_depthwise/depthwise| 0| 1| 1| 12 x x x x x x x | 13 | 1 144 56 56 | 1 144 28 28 | 1016064 |
14|TIDL_ConvolutionLayer |lock_5_project/Conv2D;model/block_3_project/Conv2D| 0| 1| 1| 13 x x x x x x x | 14 | 1 144 28 28 | 1 32 28 28 | 3612672 |
15|TIDL_ConvolutionLayer |_6_depthwise/depthwise;model/block_4_expand/Conv2D| 0| 1| 1| 14 x x x x x x x | 15 | 1 32 28 28 | 1 192 28 28 | 4816896 |
16|TIDL_ConvolutionLayer |thwise/depthwise;model/block_4_depthwise/depthwise| 0| 1| 1| 15 x x x x x x x | 16 | 1 192 28 28 | 1 192 28 28 | 1354752 |
17|TIDL_ConvolutionLayer |lock_5_project/Conv2D;model/block_4_project/Conv2D| 0| 1| 1| 16 x x x x x x x | 17 | 1 192 28 28 | 1 32 28 28 | 4816896 |
18|TIDL_EltWiseLayer |model/block_4_add/add | 0| 2| 1| 14 17 x x x x x x | 18 | 1 32 28 28 | 1 32 28 28 | 25088 |
19|TIDL_ConvolutionLayer |_6_depthwise/depthwise;model/block_5_expand/Conv2D| 0| 1| 1| 18 x x x x x x x | 19 | 1 32 28 28 | 1 192 28 28 | 4816896 |
20|TIDL_ConvolutionLayer |thwise/depthwise;model/block_5_depthwise/depthwise| 0| 1| 1| 19 x x x x x x x | 20 | 1 192 28 28 | 1 192 28 28 | 1354752 |
21|TIDL_ConvolutionLayer |t_BN/FusedBatchNormV3;model/block_5_project/Conv2D| 0| 1| 1| 20 x x x x x x x | 21 | 1 192 28 28 | 1 32 28 28 | 4816896 |
22|TIDL_EltWiseLayer |model/block_5_add/add | 0| 2| 1| 18 21 x x x x x x | 22 | 1 32 28 28 | 1 32 28 28 | 25088 |
23|TIDL_ConvolutionLayer |_6_depthwise/depthwise;model/block_6_expand/Conv2D| 0| 1| 1| 22 x x x x x x x | 23 | 1 32 28 28 | 1 192 28 28 | 4816896 |
24|TIDL_ConvolutionLayer |FusedBatchNormV3;model/block_6_depthwise/depthwise| 0| 1| 1| 23 x x x x x x x | 24 | 1 192 28 28 | 1 192 14 14 | 338688 |
25|TIDL_ConvolutionLayer |lock_9_project/Conv2D;model/block_6_project/Conv2D| 0| 1| 1| 24 x x x x x x x | 25 | 1 192 14 14 | 1 64 14 14 | 2408448 |
26|TIDL_ConvolutionLayer |10_depthwise/depthwise;model/block_7_expand/Conv2D| 0| 1| 1| 25 x x x x x x x | 26 | 1 64 14 14 | 1 384 14 14 | 4816896 |
27|TIDL_ConvolutionLayer |thwise/depthwise;model/block_7_depthwise/depthwise| 0| 1| 1| 26 x x x x x x x | 27 | 1 384 14 14 | 1 384 14 14 | 677376 |
28|TIDL_ConvolutionLayer |lock_9_project/Conv2D;model/block_7_project/Conv2D| 0| 1| 1| 27 x x x x x x x | 28 | 1 384 14 14 | 1 64 14 14 | 4816896 |
29|TIDL_EltWiseLayer |model/block_7_add/add | 0| 2| 1| 25 28 x x x x x x | 29 | 1 64 14 14 | 1 64 14 14 | 12544 |
30|TIDL_ConvolutionLayer |10_depthwise/depthwise;model/block_8_expand/Conv2D| 0| 1| 1| 29 x x x x x x x | 30 | 1 64 14 14 | 1 384 14 14 | 4816896 |
31|TIDL_ConvolutionLayer |thwise/depthwise;model/block_8_depthwise/depthwise| 0| 1| 1| 30 x x x x x x x | 31 | 1 384 14 14 | 1 384 14 14 | 677376 |
32|TIDL_ConvolutionLayer |lock_9_project/Conv2D;model/block_8_project/Conv2D| 0| 1| 1| 31 x x x x x x x | 32 | 1 384 14 14 | 1 64 14 14 | 4816896 |
33|TIDL_EltWiseLayer |model/block_8_add/add | 0| 2| 1| 29 32 x x x x x x | 33 | 1 64 14 14 | 1 64 14 14 | 12544 |
34|TIDL_ConvolutionLayer |10_depthwise/depthwise;model/block_9_expand/Conv2D| 0| 1| 1| 33 x x x x x x x | 34 | 1 64 14 14 | 1 384 14 14 | 4816896 |
35|TIDL_ConvolutionLayer |thwise/depthwise;model/block_9_depthwise/depthwise| 0| 1| 1| 34 x x x x x x x | 35 | 1 384 14 14 | 1 384 14 14 | 677376 |
36|TIDL_ConvolutionLayer |t_BN/FusedBatchNormV3;model/block_9_project/Conv2D| 0| 1| 1| 35 x x x x x x x | 36 | 1 384 14 14 | 1 64 14 14 | 4816896 |
37|TIDL_EltWiseLayer |model/block_9_add/add | 0| 2| 1| 33 36 x x x x x x | 37 | 1 64 14 14 | 1 64 14 14 | 12544 |
38|TIDL_ConvolutionLayer |0_depthwise/depthwise;model/block_10_expand/Conv2D| 0| 1| 1| 37 x x x x x x x | 38 | 1 64 14 14 | 1 384 14 14 | 4816896 |
39|TIDL_ConvolutionLayer |usedBatchNormV3;model/block_10_depthwise/depthwise| 0| 1| 1| 38 x x x x x x x | 39 | 1 384 14 14 | 1 384 14 14 | 677376 |
40|TIDL_ConvolutionLayer |ck_12_project/Conv2D;model/block_10_project/Conv2D| 0| 1| 1| 39 x x x x x x x | 40 | 1 384 14 14 | 1 96 14 14 | 7225344 |
41|TIDL_ConvolutionLayer |3_depthwise/depthwise;model/block_11_expand/Conv2D| 0| 1| 1| 40 x x x x x x x | 41 | 1 96 14 14 | 1 576 14 14 | 10838016 |
42|TIDL_ConvolutionLayer |hwise/depthwise;model/block_11_depthwise/depthwise| 0| 1| 1| 41 x x x x x x x | 42 | 1 576 14 14 | 1 576 14 14 | 1016064 |
43|TIDL_ConvolutionLayer |ck_12_project/Conv2D;model/block_11_project/Conv2D| 0| 1| 1| 42 x x x x x x x | 43 | 1 576 14 14 | 1 96 14 14 | 10838016 |
44|TIDL_EltWiseLayer |model/block_11_add/add | 0| 2| 1| 40 43 x x x x x x | 44 | 1 96 14 14 | 1 96 14 14 | 18816 |
45|TIDL_ConvolutionLayer |3_depthwise/depthwise;model/block_12_expand/Conv2D| 0| 1| 1| 44 x x x x x x x | 45 | 1 96 14 14 | 1 576 14 14 | 10838016 |
46|TIDL_ConvolutionLayer |hwise/depthwise;model/block_12_depthwise/depthwise| 0| 1| 1| 45 x x x x x x x | 46 | 1 576 14 14 | 1 576 14 14 | 1016064 |
47|TIDL_ConvolutionLayer |_BN/FusedBatchNormV3;model/block_12_project/Conv2D| 0| 1| 1| 46 x x x x x x x | 47 | 1 576 14 14 | 1 96 14 14 | 10838016 |
48|TIDL_EltWiseLayer |model/block_12_add/add | 0| 2| 1| 44 47 x x x x x x | 48 | 1 96 14 14 | 1 96 14 14 | 18816 |
49|TIDL_ConvolutionLayer |3_depthwise/depthwise;model/block_13_expand/Conv2D| 0| 1| 1| 48 x x x x x x x | 49 | 1 96 14 14 | 1 576 14 14 | 10838016 |
50|TIDL_ConvolutionLayer |usedBatchNormV3;model/block_13_depthwise/depthwise| 0| 1| 1| 49 x x x x x x x | 50 | 1 576 14 14 | 1 576 7 7 | 254016 |
51|TIDL_ConvolutionLayer |ck_15_project/Conv2D;model/block_13_project/Conv2D| 0| 1| 1| 50 x x x x x x x | 51 | 1 576 7 7 | 1 160 7 7 | 4515840 |
52|TIDL_ConvolutionLayer |6_depthwise/depthwise;model/block_14_expand/Conv2D| 0| 1| 1| 51 x x x x x x x | 52 | 1 160 7 7 | 1 960 7 7 | 7526400 |
53|TIDL_ConvolutionLayer |hwise/depthwise;model/block_14_depthwise/depthwise| 0| 1| 1| 52 x x x x x x x | 53 | 1 960 7 7 | 1 960 7 7 | 423360 |
54|TIDL_ConvolutionLayer |ck_15_project/Conv2D;model/block_14_project/Conv2D| 0| 1| 1| 53 x x x x x x x | 54 | 1 960 7 7 | 1 160 7 7 | 7526400 |
55|TIDL_EltWiseLayer |model/block_14_add/add | 0| 2| 1| 51 54 x x x x x x | 55 | 1 160 7 7 | 1 160 7 7 | 7840 |
56|TIDL_ConvolutionLayer |6_depthwise/depthwise;model/block_15_expand/Conv2D| 0| 1| 1| 55 x x x x x x x | 56 | 1 160 7 7 | 1 960 7 7 | 7526400 |
57|TIDL_ConvolutionLayer |hwise/depthwise;model/block_15_depthwise/depthwise| 0| 1| 1| 56 x x x x x x x | 57 | 1 960 7 7 | 1 960 7 7 | 423360 |
58|TIDL_ConvolutionLayer |_BN/FusedBatchNormV3;model/block_15_project/Conv2D| 0| 1| 1| 57 x x x x x x x | 58 | 1 960 7 7 | 1 160 7 7 | 7526400 |
59|TIDL_EltWiseLayer |model/block_15_add/add | 0| 2| 1| 55 58 x x x x x x | 59 | 1 160 7 7 | 1 160 7 7 | 7840 |
60|TIDL_ConvolutionLayer |6_depthwise/depthwise;model/block_16_expand/Conv2D| 0| 1| 1| 59 x x x x x x x | 60 | 1 160 7 7 | 1 960 7 7 | 7526400 |
61|TIDL_ConvolutionLayer |usedBatchNormV3;model/block_16_depthwise/depthwise| 0| 1| 1| 60 x x x x x x x | 61 | 1 960 7 7 | 1 960 7 7 | 423360 |
62|TIDL_ConvolutionLayer |_BN/FusedBatchNormV3;model/block_16_project/Conv2D| 0| 1| 1| 61 x x x x x x x | 62 | 1 960 7 7 | 1 320 7 7 | 15052800 |
63|TIDL_ConvolutionLayer |del/Conv_1_bn/FusedBatchNormV3;model/Conv_1/Conv2D| 0| 1| 1| 62 x x x x x x x | 63 | 1 320 7 7 | 1 1280 7 7 | 20070400 |
64|TIDL_PoolingLayer |model/global_average_pooling2d/Mean | 0| 1| 1| 63 x x x x x x x | 64 | 1 1280 7 7 | 1 1 1 1280 | 1280 |
65|TIDL_InnerProductLayer |Relu;model/batch_normalization_154/batchnorm/add_1| 0| 1| 1| 64 x x x x x x x | 65 | 1 1 1 1280 | 1 1 1 100 | 128100 |
66|TIDL_InnerProductLayer |model/dense_1/MatMul;model/dense_1/BiasAdd | 0| 1| 1| 65 x x x x x x x | 66 | 1 1 1 100 | 1 1 1 2 | 202 |
67|TIDL_SoftMaxLayer |StatefulPartitionedCall:0 | 0| 1| 1| 66 x x x x x x x | 67 | 1 1 1 2 | 1 1 1 2 | 2 |
68|TIDL_DataLayer |StatefulPartitionedCall:0 | 0| 1| -1| 67 x x x x x x x | 0 | 1 1 1 2 | 0 0 0 0 | 0 |
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Total Giga Macs : 0.3000
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
~~~~~Running TIDL in PC emulation mode to collect Activations range for each layer~~~~~
cd /home/users/mms00591/tisdk/ti-processor-sdk-rtos-j721e-evm-09_00_01_01/c7x-mma-tidl/ti_dl/test && ./PC_dsp_test_dl_algo.out s:/home/users/mms00591/tisdk/ti-processor-sdk-rtos-j721e-evm-09_00_01_01/c7x-mma-tidl/ti_dl/test/testvecs/out/quant_model/quant_mobilenet_v2.txt.qunat_stats_config.txt
Processing config file #0 : /home/users/mms00591/tisdk/ti-processor-sdk-rtos-j721e-evm-09_00_01_01/c7x-mma-tidl/ti_dl/test/testvecs/out/quant_model/quant_mobilenet_v2.txt.qunat_stats_config.txt
Input : dataId=0, name=serving_default_input_1:0_original, elementType 6, scale=1.000000, zero point=0, layout=1
Ouput : dataId=67, name=StatefulPartitionedCall:0, elementType 6, scale=1.000000, zero point=0, layout=1
worstCaseDelay for Pre-emption is 0.0000000
Freeing memory for user provided Net
Instance created for /home/users/mms00591/tisdk/ti-processor-sdk-rtos-j721e-evm-09_00_01_01/c7x-mma-tidl/ti_dl/test/testvecs/out/quant_model/quant_mobilenet_v2.txt.qunat_stats_config.txt
----------------------- TIDL Process with REF_ONLY FLOW ------------------------
# 0 . .. 0 1.00000 10.00000 254.00000 6
1 1.00000 -118.00000 126.00000 6
2 1.00000 0.00000 6.00000 6
3 1.00000 0.00000 6.00000 6
4 1.00000 -42.38116 47.43107 6
5 1.00000 0.00000 6.00000 6
6 1.00000 0.00000 6.00000 6
7 1.00000 -34.89307 27.23265 6
8 1.00000 0.00000 6.00000 6
9 1.00000 0.00000 6.00000 6
10 1.00000 -37.48398 27.05920 6
11 1.00000 -39.27068 34.50449 6
12 1.00000 0.00000 6.00000 6
13 1.00000 0.00000 6.00000 6
14 1.00000 -31.12851 27.29463 6
15 1.00000 0.00000 6.00000 6
16 1.00000 0.00000 6.00000 6
17 1.00000 -18.81322 28.36402 6
18 1.00000 -32.46035 33.45750 6
19 1.00000 0.00000 6.00000 6
20 1.00000 0.00000 6.00000 6
21 1.00000 -24.33548 14.44097 6
22 1.00000 -36.46158 36.49684 6
23 1.00000 0.00000 6.00000 6
24 1.00000 0.00000 6.00000 6
25 1.00000 -27.08207 25.49088 6
26 1.00000 0.00000 6.00000 6
27 1.00000 0.00000 6.00000 6
28 1.00000 -18.33643 18.51849 6
29 1.00000 -26.51540 26.30825 6
30 1.00000 0.00000 6.00000 6
31 1.00000 0.00000 6.00000 6
32 1.00000 -13.58280 16.95432 6
33 1.00000 -27.16744 24.12762 6
34 1.00000 0.00000 6.00000 6
35 1.00000 0.00000 6.00000 6
36 1.00000 -10.99764 11.45739 6
37 1.00000 -25.70851 25.61168 6
38 1.00000 0.00000 6.00000 6
39 1.00000 0.00000 6.00000 6
40 1.00000 -13.45807 11.65730 6
41 1.00000 0.00000 6.00000 6
42 1.00000 0.00000 6.00000 6
43 1.00000 -10.32781 9.48910 6
44 1.00000 -13.79831 13.18854 6
45 1.00000 0.00000 6.00000 6
46 1.00000 0.00000 6.00000 6
47 1.00000 -19.84853 14.03625 6
48 1.00000 -20.01550 16.97109 6
49 1.00000 0.00000 6.00000 6
50 1.00000 0.00000 5.30865 6
51 1.00000 -13.04135 9.50326 6
52 1.00000 0.00000 6.00000 6
53 1.00000 0.00000 6.00000 6
54 1.00000 -7.32819 7.49451 6
55 1.00000 -16.31997 13.05522 6
56 1.00000 0.00000 6.00000 6
57 1.00000 0.00000 3.43219 6
58 1.00000 -12.48979 8.95064 6
59 1.00000 -16.83730 13.32864 6
60 1.00000 0.00000 5.20456 6
61 1.00000 0.00000 6.00000 6
62 1.00000 -6.62132 6.39630 6
63 1.00000 0.00000 6.00000 6
64 1.00000 0.00000 3.45649 6
65 1.00000 0.00000 0.54558 6
66 1.00000 -0.37613 0.65395 6
67 1.00000 0.26307 0.73693 6
T 774.64 .... ..... ...
A : 1, 0.0000, 1.0000, 0 .... .....
# 1 . .. 0 1.00000 6.00000 238.00000 6
1 1.00000 -122.00000 110.00000 6
2 1.00000 0.00000 6.00000 6
3 1.00000 0.00000 6.00000 6
4 1.00000 -56.83550 47.43107 6
5 1.00000 0.00000 6.00000 6
6 1.00000 0.00000 6.00000 6
7 1.00000 -34.89307 33.39328 6
8 1.00000 0.00000 6.00000 6
9 1.00000 0.00000 6.00000 6
10 1.00000 -44.14635 30.28357 6
11 1.00000 -47.62183 42.27164 6
12 1.00000 0.00000 6.00000 6
13 1.00000 0.00000 6.00000 6
14 1.00000 -30.91902 29.01117 6
15 1.00000 0.00000 6.00000 6
16 1.00000 0.00000 6.00000 6
17 1.00000 -20.43439 24.35698 6
18 1.00000 -30.51571 36.35632 6
19 1.00000 0.00000 6.00000 6
20 1.00000 0.00000 6.00000 6
21 1.00000 -24.78173 18.32861 6
22 1.00000 -32.71653 39.62926 6
23 1.00000 0.00000 6.00000 6
24 1.00000 0.00000 6.00000 6
25 1.00000 -24.41302 21.62696 6
26 1.00000 0.00000 6.00000 6
27 1.00000 0.00000 6.00000 6
28 1.00000 -13.58287 19.01459 6
29 1.00000 -25.02938 21.29889 6
30 1.00000 0.00000 6.00000 6
31 1.00000 0.00000 6.00000 6
32 1.00000 -13.90108 16.95432 6
33 1.00000 -25.31912 24.92186 6
34 1.00000 0.00000 5.61877 6
35 1.00000 0.00000 6.00000 6
36 1.00000 -9.13276 9.71373 6
37 1.00000 -24.00728 23.94157 6
38 1.00000 0.00000 6.00000 6
39 1.00000 0.00000 6.00000 6
40 1.00000 -15.53822 13.09085 6
41 1.00000 0.00000 5.33918 6
42 1.00000 0.00000 6.00000 6
43 1.00000 -9.22113 6.48220 6
44 1.00000 -15.94719 12.69633 6
45 1.00000 0.00000 4.86096 6
46 1.00000 0.00000 6.00000 6
47 1.00000 -18.14042 12.37702 6
48 1.00000 -16.54464 15.00827 6
49 1.00000 0.00000 5.83915 6
50 1.00000 0.00000 5.95979 6
51 1.00000 -11.39833 11.78409 6
52 1.00000 0.00000 6.00000 6
53 1.00000 0.00000 5.90497 6
54 1.00000 -8.38301 6.46141 6
55 1.00000 -16.03459 12.23524 6
56 1.00000 0.00000 6.00000 6
57 1.00000 0.00000 3.81270 6
58 1.00000 -14.06860 10.74908 6
59 1.00000 -18.75942 15.33105 6
60 1.00000 0.00000 5.64081 6
61 1.00000 0.00000 6.00000 6
62 1.00000 -6.14268 6.22134 6
63 1.00000 0.00000 6.00000 6
64 1.00000 0.00000 3.45649 6
65 1.00000 0.00000 0.64553 6
66 1.00000 -0.75832 1.02371 6
67 1.00000 0.14405 0.85595 6
T 776.62 .... ..... ...
A : 0, 0.5000, 1.0000, 0 .... .....Warning :: Weight Scale Clipped - 1.94, 239.13, 123.40, 1, 1, 8, model/block_2_expand_relu/Relu6;model/block_2_expand_BN/FusedBatchNormV3;model/block_3_depthwise/depthwise;model/block_2_expand/Conv2D
Warning :: Weight Scale Clipped - 1.42, 35.43, 24.91, 0, 0, 10, model/block_2_project_BN/FusedBatchNormV3;model/block_2_project/Conv2D
Warning :: Weight Scale Clipped - 1.71, 280.96, 164.44, 1, 1, 12, model/block_3_expand_relu/Relu6;model/block_3_expand_BN/FusedBatchNormV3;model/block_3_depthwise/depthwise;model/block_3_expand/Conv2D
Warning :: Weight Scale Clipped - 1.67, 53.59, 32.00, 0, 1, 13, model/block_3_depthwise_relu/Relu6;model/block_3_depthwise_BN/FusedBatchNormV3;model/block_3_depthwise/depthwise
Warning :: Weight Scale Clipped - 1.16, 52.19, 44.87, 0, 0, 14, model/block_3_project_BN/FusedBatchNormV3;model/block_5_project/Conv2D;model/block_3_project/Conv2D
Warning :: Weight Scale Clipped - 1.77, 645.92, 365.12, 1, 1, 19, model/block_5_expand_relu/Relu6;model/block_5_expand_BN/FusedBatchNormV3;model/block_6_depthwise/depthwise;model/block_5_expand/Conv2D
Warning :: Weight Scale Clipped - 1.12, 54.09, 48.41, 0, 0, 21, model/block_5_project_BN/FusedBatchNormV3;model/block_5_project/Conv2D
Warning :: Weight Scale Clipped - 1.59, 59.78, 37.58, 0, 0, 25, model/block_6_project_BN/FusedBatchNormV3;model/block_9_project/Conv2D;model/block_6_project/Conv2D
Warning :: Weight Scale Clipped - 1.66, 13.25, 8.00, 0, 1, 27, model/block_7_depthwise_relu/Relu6;model/block_7_depthwise_BN/FusedBatchNormV3;model/block_10_depthwise/depthwise;model/block_7_depthwise/depthwise
Warning :: Weight Scale Clipped - 1.77, 771.38, 435.93, 1, 1, 34, model/block_9_expand_relu/Relu6;model/block_9_expand_BN/FusedBatchNormV3;model/block_10_depthwise/depthwise;model/block_9_expand/Conv2D
Warning :: Weight Scale Clipped - 1.79, 28.60, 16.00, 0, 1, 39, model/block_10_depthwise_relu/Relu6;model/block_10_depthwise_BN/FusedBatchNormV3;model/block_10_depthwise/depthwise
Warning :: Weight Scale Clipped - 1.83, 133.25, 72.75, 0, 0, 40, model/block_10_project_BN/FusedBatchNormV3;model/block_12_project/Conv2D;model/block_10_project/Conv2D
Warning :: Weight Scale Clipped - 1.58, 118.98, 75.27, 0, 0, 43, model/block_11_project_BN/FusedBatchNormV3;model/block_12_project/Conv2D;model/block_11_project/Conv2D
Warning :: Weight Scale Clipped - 1.61, 51.63, 32.00, 0, 1, 50, model/block_13_depthwise_relu/Relu6;model/block_13_depthwise_BN/FusedBatchNormV3;model/block_13_depthwise/depthwise
Warning :: Weight Scale Clipped - 1.82, 103.45, 56.80, 0, 0, 51, model/block_13_project_BN/FusedBatchNormV3;model/block_15_project/Conv2D;model/block_13_project/Conv2D
Warning :: Weight Scale Clipped - 1.60, 462.47, 288.44, 1, 1, 52, model/block_14_expand_relu/Relu6;model/block_14_expand_BN/FusedBatchNormV3;model/block_16_depthwise/depthwise;model/block_14_expand/Conv2D
Warning :: Tensorscale value is not met for the relu6/Clip Activation - 52
Warning :: Weight Scale Clipped - 1.57, 12.54, 8.00, 0, 1, 53, model/block_14_depthwise_relu/Relu6;model/block_14_depthwise_BN/FusedBatchNormV3;model/block_16_depthwise/depthwise;model/block_14_depthwise/depthwise
Warning :: Weight Scale Clipped - 1.10, 129.94, 118.57, 0, 0, 54, model/block_14_project_BN/FusedBatchNormV3;model/block_15_project/Conv2D;model/block_14_project/Conv2D
Warning :: Tensorscale value is not met for the relu6/Clip Activation - 56
Warning :: Weight Scale Clipped - 2.32, 184.84, 79.72, 0, 0, 62, model/block_16_project_BN/FusedBatchNormV3;model/block_16_project/Conv2D
Warning :: Weight Scale Clipped - 1.73, 177.31, 102.76, 1, 1, 63, model/out_relu/Relu6;model/Conv_1_bn/FusedBatchNormV3;model/Conv_1/Conv2D
Num of Layer Detected : 69
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Num|TIDL Layer Name |Out Data Name |Group |#Ins |#Outs |Inbuf Ids |Outbuf Id |In NCHW |Out NCHW |MACS |
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
0|TIDL_DataLayer |serving_default_input_1:0_original | 0| -1| 1| x x x x x x x x | 0 | 0 0 0 0 | 1 3 224 224 | 0 |
1|TIDL_BatchNormLayer |serving_default_input_1:0 | 0| 1| 1| 0 x x x x x x x | 1 | 1 3 224 224 | 1 3 224 224 | 150528 |
2|TIDL_ConvolutionLayer |V3;model/block_5_project/Conv2D;model/Conv1/Conv2D| 0| 1| 1| 1 x x x x x x x | 2 | 1 3 224 224 | 1 32 112 112 | 10838016 |
3|TIDL_ConvolutionLayer |v_depthwise/depthwise;model/block_5_project/Conv2D| 0| 1| 1| 2 x x x x x x x | 3 | 1 32 112 112 | 1 32 112 112 | 3612672 |
4|TIDL_ConvolutionLayer |usedBatchNormV3;model/expanded_conv_project/Conv2D| 0| 1| 1| 3 x x x x x x x | 4 | 1 32 112 112 | 1 16 112 112 | 6422528 |
5|TIDL_ConvolutionLayer |lock_12_project/Conv2D;model/block_1_expand/Conv2D| 0| 1| 1| 4 x x x x x x x | 5 | 1 16 112 112 | 1 96 112 112 | 19267584 |
6|TIDL_ConvolutionLayer |2_project/Conv2D;model/block_1_depthwise/depthwise| 0| 1| 1| 5 x x x x x x x | 6 | 1 96 112 112 | 1 96 56 56 | 2709504 |
7|TIDL_ConvolutionLayer |lock_2_project/Conv2D;model/block_1_project/Conv2D| 0| 1| 1| 6 x x x x x x x | 7 | 1 96 56 56 | 1 24 56 56 | 7225344 |
8|TIDL_ConvolutionLayer |_3_depthwise/depthwise;model/block_2_expand/Conv2D| 0| 1| 1| 7 x x x x x x x | 8 | 1 24 56 56 | 1 144 56 56 | 10838016 |
9|TIDL_ConvolutionLayer |thwise/depthwise;model/block_2_depthwise/depthwise| 0| 1| 1| 8 x x x x x x x | 9 | 1 144 56 56 | 1 144 56 56 | 4064256 |
10|TIDL_ConvolutionLayer |t_BN/FusedBatchNormV3;model/block_2_project/Conv2D| 0| 1| 1| 9 x x x x x x x | 10 | 1 144 56 56 | 1 24 56 56 | 10838016 |
11|TIDL_EltWiseLayer |model/block_2_add/add | 0| 2| 1| 7 10 x x x x x x | 11 | 1 24 56 56 | 1 24 56 56 | 75264 |
12|TIDL_ConvolutionLayer |_3_depthwise/depthwise;model/block_3_expand/Conv2D| 0| 1| 1| 11 x x x x x x x | 12 | 1 24 56 56 | 1 144 56 56 | 10838016 |
13|TIDL_ConvolutionLayer |FusedBatchNormV3;model/block_3_depthwise/depthwise| 0| 1| 1| 12 x x x x x x x | 13 | 1 144 56 56 | 1 144 28 28 | 1016064 |
14|TIDL_ConvolutionLayer |lock_5_project/Conv2D;model/block_3_project/Conv2D| 0| 1| 1| 13 x x x x x x x | 14 | 1 144 28 28 | 1 32 28 28 | 3612672 |
15|TIDL_ConvolutionLayer |_6_depthwise/depthwise;model/block_4_expand/Conv2D| 0| 1| 1| 14 x x x x x x x | 15 | 1 32 28 28 | 1 192 28 28 | 4816896 |
16|TIDL_ConvolutionLayer |thwise/depthwise;model/block_4_depthwise/depthwise| 0| 1| 1| 15 x x x x x x x | 16 | 1 192 28 28 | 1 192 28 28 | 1354752 |
17|TIDL_ConvolutionLayer |lock_5_project/Conv2D;model/block_4_project/Conv2D| 0| 1| 1| 16 x x x x x x x | 17 | 1 192 28 28 | 1 32 28 28 | 4816896 |
18|TIDL_EltWiseLayer |model/block_4_add/add | 0| 2| 1| 14 17 x x x x x x | 18 | 1 32 28 28 | 1 32 28 28 | 25088 |
19|TIDL_ConvolutionLayer |_6_depthwise/depthwise;model/block_5_expand/Conv2D| 0| 1| 1| 18 x x x x x x x | 19 | 1 32 28 28 | 1 192 28 28 | 4816896 |
20|TIDL_ConvolutionLayer |thwise/depthwise;model/block_5_depthwise/depthwise| 0| 1| 1| 19 x x x x x x x | 20 | 1 192 28 28 | 1 192 28 28 | 1354752 |
21|TIDL_ConvolutionLayer |t_BN/FusedBatchNormV3;model/block_5_project/Conv2D| 0| 1| 1| 20 x x x x x x x | 21 | 1 192 28 28 | 1 32 28 28 | 4816896 |
22|TIDL_EltWiseLayer |model/block_5_add/add | 0| 2| 1| 18 21 x x x x x x | 22 | 1 32 28 28 | 1 32 28 28 | 25088 |
23|TIDL_ConvolutionLayer |_6_depthwise/depthwise;model/block_6_expand/Conv2D| 0| 1| 1| 22 x x x x x x x | 23 | 1 32 28 28 | 1 192 28 28 | 4816896 |
24|TIDL_ConvolutionLayer |FusedBatchNormV3;model/block_6_depthwise/depthwise| 0| 1| 1| 23 x x x x x x x | 24 | 1 192 28 28 | 1 192 14 14 | 338688 |
25|TIDL_ConvolutionLayer |lock_9_project/Conv2D;model/block_6_project/Conv2D| 0| 1| 1| 24 x x x x x x x | 25 | 1 192 14 14 | 1 64 14 14 | 2408448 |
26|TIDL_ConvolutionLayer |10_depthwise/depthwise;model/block_7_expand/Conv2D| 0| 1| 1| 25 x x x x x x x | 26 | 1 64 14 14 | 1 384 14 14 | 4816896 |
27|TIDL_ConvolutionLayer |thwise/depthwise;model/block_7_depthwise/depthwise| 0| 1| 1| 26 x x x x x x x | 27 | 1 384 14 14 | 1 384 14 14 | 677376 |
28|TIDL_ConvolutionLayer |lock_9_project/Conv2D;model/block_7_project/Conv2D| 0| 1| 1| 27 x x x x x x x | 28 | 1 384 14 14 | 1 64 14 14 | 4816896 |
29|TIDL_EltWiseLayer |model/block_7_add/add | 0| 2| 1| 25 28 x x x x x x | 29 | 1 64 14 14 | 1 64 14 14 | 12544 |
30|TIDL_ConvolutionLayer |10_depthwise/depthwise;model/block_8_expand/Conv2D| 0| 1| 1| 29 x x x x x x x | 30 | 1 64 14 14 | 1 384 14 14 | 4816896 |
31|TIDL_ConvolutionLayer |thwise/depthwise;model/block_8_depthwise/depthwise| 0| 1| 1| 30 x x x x x x x | 31 | 1 384 14 14 | 1 384 14 14 | 677376 |
32|TIDL_ConvolutionLayer |lock_9_project/Conv2D;model/block_8_project/Conv2D| 0| 1| 1| 31 x x x x x x x | 32 | 1 384 14 14 | 1 64 14 14 | 4816896 |
33|TIDL_EltWiseLayer |model/block_8_add/add | 0| 2| 1| 29 32 x x x x x x | 33 | 1 64 14 14 | 1 64 14 14 | 12544 |
34|TIDL_ConvolutionLayer |10_depthwise/depthwise;model/block_9_expand/Conv2D| 0| 1| 1| 33 x x x x x x x | 34 | 1 64 14 14 | 1 384 14 14 | 4816896 |
35|TIDL_ConvolutionLayer |thwise/depthwise;model/block_9_depthwise/depthwise| 0| 1| 1| 34 x x x x x x x | 35 | 1 384 14 14 | 1 384 14 14 | 677376 |
36|TIDL_ConvolutionLayer |t_BN/FusedBatchNormV3;model/block_9_project/Conv2D| 0| 1| 1| 35 x x x x x x x | 36 | 1 384 14 14 | 1 64 14 14 | 4816896 |
37|TIDL_EltWiseLayer |model/block_9_add/add | 0| 2| 1| 33 36 x x x x x x | 37 | 1 64 14 14 | 1 64 14 14 | 12544 |
38|TIDL_ConvolutionLayer |0_depthwise/depthwise;model/block_10_expand/Conv2D| 0| 1| 1| 37 x x x x x x x | 38 | 1 64 14 14 | 1 384 14 14 | 4816896 |
39|TIDL_ConvolutionLayer |usedBatchNormV3;model/block_10_depthwise/depthwise| 0| 1| 1| 38 x x x x x x x | 39 | 1 384 14 14 | 1 384 14 14 | 677376 |
40|TIDL_ConvolutionLayer |ck_12_project/Conv2D;model/block_10_project/Conv2D| 0| 1| 1| 39 x x x x x x x | 40 | 1 384 14 14 | 1 96 14 14 | 7225344 |
41|TIDL_ConvolutionLayer |3_depthwise/depthwise;model/block_11_expand/Conv2D| 0| 1| 1| 40 x x x x x x x | 41 | 1 96 14 14 | 1 576 14 14 | 10838016 |
42|TIDL_ConvolutionLayer |hwise/depthwise;model/block_11_depthwise/depthwise| 0| 1| 1| 41 x x x x x x x | 42 | 1 576 14 14 | 1 576 14 14 | 1016064 |
43|TIDL_ConvolutionLayer |ck_12_project/Conv2D;model/block_11_project/Conv2D| 0| 1| 1| 42 x x x x x x x | 43 | 1 576 14 14 | 1 96 14 14 | 10838016 |
44|TIDL_EltWiseLayer |model/block_11_add/add | 0| 2| 1| 40 43 x x x x x x | 44 | 1 96 14 14 | 1 96 14 14 | 18816 |
45|TIDL_ConvolutionLayer |3_depthwise/depthwise;model/block_12_expand/Conv2D| 0| 1| 1| 44 x x x x x x x | 45 | 1 96 14 14 | 1 576 14 14 | 10838016 |
46|TIDL_ConvolutionLayer |hwise/depthwise;model/block_12_depthwise/depthwise| 0| 1| 1| 45 x x x x x x x | 46 | 1 576 14 14 | 1 576 14 14 | 1016064 |
47|TIDL_ConvolutionLayer |_BN/FusedBatchNormV3;model/block_12_project/Conv2D| 0| 1| 1| 46 x x x x x x x | 47 | 1 576 14 14 | 1 96 14 14 | 10838016 |
48|TIDL_EltWiseLayer |model/block_12_add/add | 0| 2| 1| 44 47 x x x x x x | 48 | 1 96 14 14 | 1 96 14 14 | 18816 |
49|TIDL_ConvolutionLayer |3_depthwise/depthwise;model/block_13_expand/Conv2D| 0| 1| 1| 48 x x x x x x x | 49 | 1 96 14 14 | 1 576 14 14 | 10838016 |
50|TIDL_ConvolutionLayer |usedBatchNormV3;model/block_13_depthwise/depthwise| 0| 1| 1| 49 x x x x x x x | 50 | 1 576 14 14 | 1 576 7 7 | 254016 |
51|TIDL_ConvolutionLayer |ck_15_project/Conv2D;model/block_13_project/Conv2D| 0| 1| 1| 50 x x x x x x x | 51 | 1 576 7 7 | 1 160 7 7 | 4515840 |
52|TIDL_ConvolutionLayer |6_depthwise/depthwise;model/block_14_expand/Conv2D| 0| 1| 1| 51 x x x x x x x | 52 | 1 160 7 7 | 1 960 7 7 | 7526400 |
53|TIDL_ConvolutionLayer |hwise/depthwise;model/block_14_depthwise/depthwise| 0| 1| 1| 52 x x x x x x x | 53 | 1 960 7 7 | 1 960 7 7 | 423360 |
54|TIDL_ConvolutionLayer |ck_15_project/Conv2D;model/block_14_project/Conv2D| 0| 1| 1| 53 x x x x x x x | 54 | 1 960 7 7 | 1 160 7 7 | 7526400 |
55|TIDL_EltWiseLayer |model/block_14_add/add | 0| 2| 1| 51 54 x x x x x x | 55 | 1 160 7 7 | 1 160 7 7 | 7840 |
56|TIDL_ConvolutionLayer |6_depthwise/depthwise;model/block_15_expand/Conv2D| 0| 1| 1| 55 x x x x x x x | 56 | 1 160 7 7 | 1 960 7 7 | 7526400 |
57|TIDL_ConvolutionLayer |hwise/depthwise;model/block_15_depthwise/depthwise| 0| 1| 1| 56 x x x x x x x | 57 | 1 960 7 7 | 1 960 7 7 | 423360 |
58|TIDL_ConvolutionLayer |_BN/FusedBatchNormV3;model/block_15_project/Conv2D| 0| 1| 1| 57 x x x x x x x | 58 | 1 960 7 7 | 1 160 7 7 | 7526400 |
59|TIDL_EltWiseLayer |model/block_15_add/add | 0| 2| 1| 55 58 x x x x x x | 59 | 1 160 7 7 | 1 160 7 7 | 7840 |
60|TIDL_ConvolutionLayer |6_depthwise/depthwise;model/block_16_expand/Conv2D| 0| 1| 1| 59 x x x x x x x | 60 | 1 160 7 7 | 1 960 7 7 | 7526400 |
61|TIDL_ConvolutionLayer |usedBatchNormV3;model/block_16_depthwise/depthwise| 0| 1| 1| 60 x x x x x x x | 61 | 1 960 7 7 | 1 960 7 7 | 423360 |
62|TIDL_ConvolutionLayer |_BN/FusedBatchNormV3;model/block_16_project/Conv2D| 0| 1| 1| 61 x x x x x x x | 62 | 1 960 7 7 | 1 320 7 7 | 15052800 |
63|TIDL_ConvolutionLayer |del/Conv_1_bn/FusedBatchNormV3;model/Conv_1/Conv2D| 0| 1| 1| 62 x x x x x x x | 63 | 1 320 7 7 | 1 1280 7 7 | 20070400 |
64|TIDL_PoolingLayer |model/global_average_pooling2d/Mean | 0| 1| 1| 63 x x x x x x x | 64 | 1 1280 7 7 | 1 1 1 1280 | 1280 |
65|TIDL_InnerProductLayer |Relu;model/batch_normalization_154/batchnorm/add_1| 0| 1| 1| 64 x x x x x x x | 65 | 1 1 1 1280 | 1 1 1 100 | 128100 |
66|TIDL_InnerProductLayer |model/dense_1/MatMul;model/dense_1/BiasAdd | 0| 1| 1| 65 x x x x x x x | 66 | 1 1 1 100 | 1 1 1 2 | 202 |
67|TIDL_SoftMaxLayer |StatefulPartitionedCall:0 | 0| 1| 1| 66 x x x x x x x | 67 | 1 1 1 2 | 1 1 1 2 | 2 |
68|TIDL_DataLayer |StatefulPartitionedCall:0 | 0| 1| -1| 67 x x x x x x x | 0 | 1 1 1 2 | 0 0 0 0 | 0 |
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Total Giga Macs : 0.3000
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
cp /home/users/mms00591/tisdk/ti-processor-sdk-rtos-j721e-evm-09_00_01_01/c7x-mma-tidl/ti_dl/test/testvecs/config/import/device_config.cfg /home/users/mms00591/tisdk/ti-processor-sdk-rtos-j721e-evm-09_00_01_01/c7x-mma-tidl/ti_dl/test/testvecs/out/quant_model/quant_mobilenet_v2.txt.perf_sim_config.txt
cd /home/users/mms00591/tisdk/ti-processor-sdk-rtos-j721e-evm-09_00_01_01/c7x-mma-tidl/ti_dl/utils/perfsim && ./ti_cnnperfsim.out /home/users/mms00591/tisdk/ti-processor-sdk-rtos-j721e-evm-09_00_01_01/c7x-mma-tidl/ti_dl/test/testvecs/out/quant_model/quant_mobilenet_v2.txt.perf_sim_config.txt 1601 0 2
------------------ Network Compiler Traces -----------------------------
successful Memory allocation
Network buffer segments:
buf(0), offset( 0), size( 124416)
buf(1), offset( 3072000), size( 2352900)
buf(2), offset( 124416), size( 544960)
buf(3), offset( 669376), size( 1380152)
/home/users/mms00591/tisdk/ti-processor-sdk-rtos-j721e-evm-09_00_01_01/c7x-mma-tidl/ti_dl/utils/tidlModelGraphviz/out/tidl_graphVisualiser.out ../../test/testvecs/out/quant_model/180_tidl_net.bin
/home/users/mms00591/tisdk/ti-processor-sdk-rtos-j721e-evm-09_00_01_01/c7x-mma-tidl/ti_dl/utils/tidlModelDump/out/tidl_dump.out -model ../../test/testvecs/out/quant_model/180_tidl_net.bin -io ../../test/testvecs/out/quant_model/180_tidl_io_1.bin -perfsim -o ../../test/testvecs/out/quant_model/180_tidl_net.bin.txt
****************************************************
** TIDL Model Checker **
****************************************************
****************************************************
** ALL MODEL CHECK PASSED **
****************************************************
**********************************************************************
c7x-mma-tidl/ti_dl/test/testvecs/config/quant_mobilenet_v2.txt:
modelType = 3
inputNetFile = "../../test/testvecs/models/quant_mobilenet_v2_high_reg_batched1.tflite"
outputNetFile = "../../test/testvecs/out/quant_model/180_tidl_net.bin"
outputParamsFile = "../../test/testvecs/out/quant_model/180_tidl_io_"
inData = "../../test/testvecs/config/one.txt"
quantizationStyle = 2
calibrationOption = 7
biasCalibrationIterations = 0
resizeWidth = 224
resizeHeight = 224
numParamBits = 8
# numFeatureBits = 8
inElementType = 0
outElementType = 0
inWidth = 224
inHeight = 224
inNumChannels = 3
inDataNorm = 1
inMean = 128 128 128
inScale = 1 1 1
postProcType = 1
writeTraceLevel = 1
debugTraceLevel = 1
inLayout = 1
outLayout = 1
foldPreBnConv2D = 0
***********************************
Our model is already quantized (int8 format) using the tensorflow lite library. The only thing we want to do is to convert it to an equivalent TIDL model. That is why I think confidence wrt 16 bit is not sensible measure in this context. Please do not hesitate to correct me if I have misunderstood something!
Best regards!
Hi, Pratic,
modelType = 3
# inputNetFile = "../../test/testvecs/models/quant_mobilenet_v2_high_reg_batched1.tflite"
inputNetFile = "../../test/testvecs/models/quant_model.tflite"
outputNetFile = "../../test/testvecs/out/quant_model/180_tidl_net.bin"
outputParamsFile = "../../test/testvecs/out/quant_model/180_tidl_io_"
inData = "../../test/testvecs/config/one.txt"
quantizationStyle = 2
calibrationOption = 7
biasCalibrationIterations = 0
resizeWidth = 224
resizeHeight = 224
numParamBits = 8
# numFeatureBits = 8
inElementType = 0
outElementType = 0
inWidth = 224
inHeight = 224
inNumChannels = 3
inDataNorm = 1
inMean = 128 128 128
inScale = 1 1 1
postProcType = 1
writeTraceLevel = 1
debugTraceLevel = 1
inLayout = 1
outLayout = 1
foldPreBnConv2D = 0
BR
Could you please help me with few details.
Log of model compilation
Import config file
What is confidence wrt to 16 bit ? for the given model ?
I can see the model compilation logs, can you help me with other details as mentioned in previous response (please share the zip file here)
I will try to reproduce this issue at my end.
Hi Pratik,
I want to send you some zip files with quant model, bitmaps and config data, so you can reproduce this issue. But I can not upload tflite model here - it is 4MB.
How can I send it to you?
Best Regards!
Hi,
You can create a zip file and share it via e2e forum.
But before we delve into details, can you help me with which sdk version you are using ?
Also i would like ask for the accuracy related observation on latest 9.1 sdk as there are potential fixes added in it, that might resolve this issue, looking forward for this info from you, post this i will try to reproducing the issue at my end.
Hi Pratic,
Currently we use SDK 08.06, but I tested this model on SDK 09.00.01.
If you believe that SDK 9.1 will help, I'll go to test with this latest version.
Can you help me with a link for uploading a zip file (4MB)?
Best Regards!
If you believe that SDK 9.1 will help, I'll go to test with this latest version.
Looking forward for the result logs.
Can you help me with a link for uploading a zip file (4MB)?
You can refer this FAQ : https://e2e.ti.com/support/site-support-group/site-support/f/site-support-forum/761619/faq-how-do-i-attach-a-file-or-link-to-my-post#:~:text=Need%20to%20attach%20a%20file,one%20file%20at%20a%20time).
Hi Pratik,
Here are the model, configs and input pictures:
They are got from SDK 9.0.1 and here are results from tidl test during convert the model.
Best Regards!