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: TDA2EXEVM
Other Parts Discussed in Thread: TDA2
Tool/software: TI C/C++ Compiler
hi Champs:
Customer tooks this caffe-Jacinto to train the "mobileNet" on the PC/NB, it is trained and working on the PC/NB
However, deploying it onto TDAEX17, it won't works.
Customer will post the error log on this thread later, please help.
BR Rio
Hi Rio,
Sure, we look into it as soon as customer post the error log.
Thanks,
Praveen
Hi Praveen,
I've trained 'mobiledetnet-1.0' caffe model with using train_mobilenet_object_detection.sh
model works well when simulating on PC
after using import tool and running on tda2, precision & frame rate become really low
import config was set by following TIDL user guide
how to modify import tool or config to overcome this issue
Hi,
Kindly refer to this below thread and update prototxt file and then import your model again and check,
https://e2e.ti.com/support/processors/f/791/t/689617
Also, follow FAQ in the user guide on "setting “layersGroupId” and “conv2dKernelType” parameters in import config file" and set them properly.
Thanks,
Praveen
Hi Praveen,
I followed the thread as you mentioned, update below two parameters
and follow the User Guide setting “layersGroupId” & “conv2dKernelType” but tda2 running situation seems quite the same
I wonder why both mAP & frame rate of mobileNet are lower than jacintoNet, and what is the advantage of mobileNet compare to jacintoNet
Hi,
I hope you had imported the model again after the above two parameters. Please confirm.
Also, can you please share the import config file used for importing?
Thanks,
Praveen
yes I had reimported the model after update parameters
frame rate indeed increased from 4 to 6, but still not reach half of jacintoNets frame rate (both model detect with same input image resolution)
import config attached below
# Default - 0 randParams = 0 # 0: Caffe, 1: TensorFlow, Default - 0 modelType = 0 # 0: Fixed quantization By tarininng Framework, 1: Dyanamic quantization by TIDL, Default - 1 quantizationStyle = 1 # quantRoundAdd/100 will be added while rounding to integer, Default - 50 quantRoundAdd = 50 numParamBits = 8 # 0 : 8bit Unsigned, 1 : 8bit Signed Default - 1 inElementType = 0 inputNetFile = "_trained_model\deploy.prototxt" inputParamsFile = "_trained_model\test_mobiledetnet-1.0_iter_3000.caffemodel" outputNetFile = "_output_model\tidl_net_jdetNet_ssd.bin" outputParamsFile = "_output_model\tidl_param_jdetNet_ssd.bin" rawSampleInData = 0 preProcType = 4 sampleInData = "_test_image\test.jpg" tidlStatsTool = "..\quantStatsTool\eve_test_dl_algo.out.exe" layersGroupId = 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 0 #layersGroupId = 0 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 0 conv2dKernelType = 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
Okay, thanks for confirmation and sharing the import config file, can you please also share the import tool log ?
Thanks,
Praveen
Okay, output log file attached below
Caffe Network File : _trained_model\deploy.prototxt
Caffe Model File : _trained_model\test_mobiledetnet-1.0_iter_3000.caffemodel
TIDL Network File : _output_model\tidl_net_jdetNet_ssd.bin
TIDL Model File : _output_model\tidl_param_jdetNet_ssd.bin
Name of the Network : mobiledetnet-1.0_deploy
Num Inputs : 1
Could not find detection_out Params
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 , 320 , 768 , 0 ,
1, TIDL_BatchNormLayer , data/bias 1, 1 , 1 , 0 , x , x , x , x , x , x , x , 1 , 1 , 3 , 320 , 768 , 1 , 3 , 320 , 768 , 737280 ,
2, TIDL_ConvolutionLayer , conv1 1, 1 , 1 , 1 , x , x , x , x , x , x , x , 2 , 1 , 3 , 320 , 768 , 1 , 32 , 160 , 384 , 53084160 ,
3, TIDL_ConvolutionLayer , conv2_1/dw 1, 1 , 1 , 2 , x , x , x , x , x , x , x , 3 , 1 , 32 , 160 , 384 , 1 , 32 , 160 , 384 , 17694720 ,
4, TIDL_ConvolutionLayer , conv2_1/sep 1, 1 , 1 , 3 , x , x , x , x , x , x , x , 4 , 1 , 32 , 160 , 384 , 1 , 64 , 160 , 384 , 125829120 ,
5, TIDL_ConvolutionLayer , conv2_2/dw 1, 1 , 1 , 4 , x , x , x , x , x , x , x , 5 , 1 , 64 , 160 , 384 , 1 , 64 , 80 , 192 , 8847360 ,
6, TIDL_ConvolutionLayer , conv2_2/sep 1, 1 , 1 , 5 , x , x , x , x , x , x , x , 6 , 1 , 64 , 80 , 192 , 1 , 128 , 80 , 192 , 125829120 ,
7, TIDL_ConvolutionLayer , conv3_1/dw 1, 1 , 1 , 6 , x , x , x , x , x , x , x , 7 , 1 , 128 , 80 , 192 , 1 , 128 , 80 , 192 , 17694720 ,
8, TIDL_ConvolutionLayer , conv3_1/sep 1, 1 , 1 , 7 , x , x , x , x , x , x , x , 8 , 1 , 128 , 80 , 192 , 1 , 128 , 80 , 192 , 251658240 ,
9, TIDL_ConvolutionLayer , conv3_2/dw 1, 1 , 1 , 8 , x , x , x , x , x , x , x , 9 , 1 , 128 , 80 , 192 , 1 , 128 , 40 , 96 , 4423680 ,
10, TIDL_ConvolutionLayer , conv3_2/sep 1, 1 , 1 , 9 , x , x , x , x , x , x , x , 10 , 1 , 128 , 40 , 96 , 1 , 256 , 40 , 96 , 125829120 ,
11, TIDL_ConvolutionLayer , conv4_1/dw 1, 1 , 1 , 10 , x , x , x , x , x , x , x , 11 , 1 , 256 , 40 , 96 , 1 , 256 , 40 , 96 , 8847360 ,
12, TIDL_ConvolutionLayer , conv4_1/sep 1, 1 , 1 , 11 , x , x , x , x , x , x , x , 12 , 1 , 256 , 40 , 96 , 1 , 256 , 40 , 96 , 251658240 ,
13, TIDL_ConvolutionLayer , conv4_2/dw 1, 1 , 1 , 12 , x , x , x , x , x , x , x , 13 , 1 , 256 , 40 , 96 , 1 , 256 , 20 , 48 , 2211840 ,
14, TIDL_ConvolutionLayer , conv4_2/sep 1, 1 , 1 , 13 , x , x , x , x , x , x , x , 14 , 1 , 256 , 20 , 48 , 1 , 512 , 20 , 48 , 125829120 ,
15, TIDL_ConvolutionLayer , conv5_1/dw 1, 1 , 1 , 14 , x , x , x , x , x , x , x , 15 , 1 , 512 , 20 , 48 , 1 , 512 , 20 , 48 , 4423680 ,
16, TIDL_ConvolutionLayer , conv5_1/sep 1, 1 , 1 , 15 , x , x , x , x , x , x , x , 16 , 1 , 512 , 20 , 48 , 1 , 512 , 20 , 48 , 251658240 ,
17, TIDL_ConvolutionLayer , conv5_2/dw 1, 1 , 1 , 16 , x , x , x , x , x , x , x , 17 , 1 , 512 , 20 , 48 , 1 , 512 , 20 , 48 , 4423680 ,
18, TIDL_ConvolutionLayer , conv5_2/sep 1, 1 , 1 , 17 , x , x , x , x , x , x , x , 18 , 1 , 512 , 20 , 48 , 1 , 512 , 20 , 48 , 251658240 ,
19, TIDL_ConvolutionLayer , conv5_3/dw 1, 1 , 1 , 18 , x , x , x , x , x , x , x , 19 , 1 , 512 , 20 , 48 , 1 , 512 , 20 , 48 , 4423680 ,
20, TIDL_ConvolutionLayer , conv5_3/sep 1, 1 , 1 , 19 , x , x , x , x , x , x , x , 20 , 1 , 512 , 20 , 48 , 1 , 512 , 20 , 48 , 251658240 ,
21, TIDL_ConvolutionLayer , conv5_4/dw 1, 1 , 1 , 20 , x , x , x , x , x , x , x , 21 , 1 , 512 , 20 , 48 , 1 , 512 , 20 , 48 , 4423680 ,
22, TIDL_ConvolutionLayer , conv5_4/sep 1, 1 , 1 , 21 , x , x , x , x , x , x , x , 22 , 1 , 512 , 20 , 48 , 1 , 512 , 20 , 48 , 251658240 ,
23, TIDL_ConvolutionLayer , conv5_5/dw 1, 1 , 1 , 22 , x , x , x , x , x , x , x , 23 , 1 , 512 , 20 , 48 , 1 , 512 , 20 , 48 , 4423680 ,
24, TIDL_ConvolutionLayer , conv5_5/sep 1, 1 , 1 , 23 , x , x , x , x , x , x , x , 24 , 1 , 512 , 20 , 48 , 1 , 512 , 20 , 48 , 251658240 ,
25, TIDL_ConvolutionLayer , conv5_6/dw 1, 1 , 1 , 24 , x , x , x , x , x , x , x , 25 , 1 , 512 , 20 , 48 , 1 , 512 , 10 , 24 , 1105920 ,
26, TIDL_ConvolutionLayer , conv5_6/sep 1, 1 , 1 , 25 , x , x , x , x , x , x , x , 26 , 1 , 512 , 10 , 24 , 1 , 1024 , 10 , 24 , 125829120 ,
27, TIDL_ConvolutionLayer , conv6/dw 1, 1 , 1 , 26 , x , x , x , x , x , x , x , 27 , 1 , 1024 , 10 , 24 , 1 , 1024 , 10 , 24 , 2211840 ,
28, TIDL_ConvolutionLayer , conv6/sep 1, 1 , 1 , 27 , x , x , x , x , x , x , x , 28 , 1 , 1024 , 10 , 24 , 1 , 1024 , 10 , 24 , 251658240 ,
29, TIDL_PoolingLayer , pool6 1, 1 , 1 , 28 , x , x , x , x , x , x , x , 29 , 1 , 1024 , 10 , 24 , 1 , 1024 , 5 , 12 , 245760 ,
30, TIDL_PoolingLayer , pool7 1, 1 , 1 , 29 , x , x , x , x , x , x , x , 30 , 1 , 1024 , 5 , 12 , 1 , 1024 , 3 , 6 , 73728 ,
31, TIDL_PoolingLayer , pool8 1, 1 , 1 , 30 , x , x , x , x , x , x , x , 31 , 1 , 1024 , 3 , 6 , 1 , 1024 , 2 , 3 , 24576 ,
32, TIDL_ConvolutionLayer , ctx_output1/dw 1, 1 , 1 , 24 , x , x , x , x , x , x , x , 32 , 1 , 512 , 20 , 48 , 1 , 512 , 20 , 48 , 4423680 ,
33, TIDL_ConvolutionLayer , ctx_output1/sep 1, 1 , 1 , 32 , x , x , x , x , x , x , x , 33 , 1 , 512 , 20 , 48 , 1 , 512 , 20 , 48 , 251658240 ,
34, TIDL_ConvolutionLayer , ctx_output2/dw 1, 1 , 1 , 28 , x , x , x , x , x , x , x , 34 , 1 , 1024 , 10 , 24 , 1 , 1024 , 10 , 24 , 2211840 ,
35, TIDL_ConvolutionLayer , ctx_output2/sep 1, 1 , 1 , 34 , x , x , x , x , x , x , x , 35 , 1 , 1024 , 10 , 24 , 1 , 512 , 10 , 24 , 125829120 ,
36, TIDL_ConvolutionLayer , ctx_output3/dw 1, 1 , 1 , 29 , x , x , x , x , x , x , x , 36 , 1 , 1024 , 5 , 12 , 1 , 1024 , 5 , 12 , 552960 ,
37, TIDL_ConvolutionLayer , ctx_output3/sep 1, 1 , 1 , 36 , x , x , x , x , x , x , x , 37 , 1 , 1024 , 5 , 12 , 1 , 512 , 5 , 12 , 31457280 ,
38, TIDL_ConvolutionLayer , ctx_output4/dw 1, 1 , 1 , 30 , x , x , x , x , x , x , x , 38 , 1 , 1024 , 3 , 6 , 1 , 1024 , 3 , 6 , 165888 ,
39, TIDL_ConvolutionLayer , ctx_output4/sep 1, 1 , 1 , 38 , x , x , x , x , x , x , x , 39 , 1 , 1024 , 3 , 6 , 1 , 512 , 3 , 6 , 9437184 ,
40, TIDL_ConvolutionLayer , ctx_output5/dw 1, 1 , 1 , 31 , x , x , x , x , x , x , x , 40 , 1 , 1024 , 2 , 3 , 1 , 1024 , 2 , 3 , 55296 ,
41, TIDL_ConvolutionLayer , ctx_output5/sep 1, 1 , 1 , 40 , x , x , x , x , x , x , x , 41 , 1 , 1024 , 2 , 3 , 1 , 512 , 2 , 3 , 3145728 ,
42, TIDL_ConvolutionLayer , ctx_output1/sep/relu_mbox_loc 1, 1 , 1 , 33 , x , x , x , x , x , x , x , 42 , 1 , 512 , 20 , 48 , 1 , 24 , 20 , 48 , 11796480 ,
43, TIDL_FlattenLayer , ctx_output1/sep/relu_mbox_loc_perm 1, 1 , 1 , 42 , x , x , x , x , x , x , x , 43 , 1 , 24 , 20 , 48 , 1 , 1 , 1 , 23040 , 1 ,
44, TIDL_ConvolutionLayer , ctx_output1/sep/relu_mbox_conf 1, 1 , 1 , 33 , x , x , x , x , x , x , x , 44 , 1 , 512 , 20 , 48 , 1 , 30 , 20 , 48 , 14745600 ,
45, TIDL_FlattenLayer , ctx_output1/sep/relu_mbox_conf_perm 1, 1 , 1 , 44 , x , x , x , x , x , x , x , 45 , 1 , 30 , 20 , 48 , 1 , 1 , 1 , 28800 , 1 ,
47, TIDL_ConvolutionLayer , ctx_output2/sep/relu_mbox_loc 1, 1 , 1 , 35 , x , x , x , x , x , x , x , 47 , 1 , 512 , 10 , 24 , 1 , 24 , 10 , 24 , 2949120 ,
48, TIDL_FlattenLayer , ctx_output2/sep/relu_mbox_loc_perm 1, 1 , 1 , 47 , x , x , x , x , x , x , x , 48 , 1 , 24 , 10 , 24 , 1 , 1 , 1 , 5760 , 1 ,
49, TIDL_ConvolutionLayer , ctx_output2/sep/relu_mbox_conf 1, 1 , 1 , 35 , x , x , x , x , x , x , x , 49 , 1 , 512 , 10 , 24 , 1 , 30 , 10 , 24 , 3686400 ,
50, TIDL_FlattenLayer , ctx_output2/sep/relu_mbox_conf_perm 1, 1 , 1 , 49 , x , x , x , x , x , x , x , 50 , 1 , 30 , 10 , 24 , 1 , 1 , 1 , 7200 , 1 ,
52, TIDL_ConvolutionLayer , ctx_output3/sep/relu_mbox_loc 1, 1 , 1 , 37 , x , x , x , x , x , x , x , 52 , 1 , 512 , 5 , 12 , 1 , 24 , 5 , 12 , 737280 ,
53, TIDL_FlattenLayer , ctx_output3/sep/relu_mbox_loc_perm 1, 1 , 1 , 52 , x , 1 file(s) copied.
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 , 320 , 768 ,
1, TIDL_BatchNormLayer , 1, 1 , 1 , 0 , x , x , x , x , x , x , x , 1 , 1 , 3 , 320 , 768 , 1 , 3 , 320 , 768 ,
2, TIDL_ConvolutionLayer , 1, 1 , 1 , 1 , x , x , x , x , x , x , x , 2 , 1 , 3 , 320 , 768 , 1 , 32 , 160 , 384 ,
3, TIDL_ConvolutionLayer , 1, 1 , 1 , 2 , x , x , x , x , x , x , x , 3 , 1 , 32 , 160 , 384 , 1 , 32 , 160 , 384 ,
4, TIDL_ConvolutionLayer , 1, 1 , 1 , 3 , x , x , x , x , x , x , x , 4 , 1 , 32 , 160 , 384 , 1 , 64 , 160 , 384 ,
5, TIDL_ConvolutionLayer , 1, 1 , 1 , 4 , x , x , x , x , x , x , x , 5 , 1 , 64 , 160 , 384 , 1 , 64 , 80 , 192 ,
6, TIDL_ConvolutionLayer , 1, 1 , 1 , 5 , x , x , x , x , x , x , x , 6 , 1 , 64 , 80 , 192 , 1 , 128 , 80 , 192 ,
7, TIDL_ConvolutionLayer , 1, 1 , 1 , 6 , x , x , x , x , x , x , x , 7 , 1 , 128 , 80 , 192 , 1 , 128 , 80 , 192 ,
8, TIDL_ConvolutionLayer , 1, 1 , 1 , 7 , x , x , x , x , x , x , x , 8 , 1 , 128 , 80 , 192 , 1 , 128 , 80 , 192 ,
9, TIDL_ConvolutionLayer , 1, 1 , 1 , 8 , x , x , x , x , x , x , x , 9 , 1 , 128 , 80 , 192 , 1 , 128 , 40 , 96 ,
10, TIDL_ConvolutionLayer , 1, 1 , 1 , 9 , x , x , x , x , x , x , x , 10 , 1 , 128 , 40 , 96 , 1 , 256 , 40 , 96 ,
11, TIDL_ConvolutionLayer , 1, 1 , 1 , 10 , x , x , x , x , x , x , x , 11 , 1 , 256 , 40 , 96 , 1 , 256 , 40 , 96 ,
12, TIDL_ConvolutionLayer , 1, 1 , 1 , 11 , x , x , x , x , x , x , x , 12 , 1 , 256 , 40 , 96 , 1 , 256 , 40 , 96 ,
13, TIDL_ConvolutionLayer , 1, 1 , 1 , 12 , x , x , x , x , x , x , x , 13 , 1 , 256 , 40 , 96 , 1 , 256 , 20 , 48 ,
14, TIDL_ConvolutionLayer , 1, 1 , 1 , 13 , x , x , x , x , x , x , x , 14 , 1 , 256 , 20 , 48 , 1 , 512 , 20 , 48 ,
15, TIDL_ConvolutionLayer , 1, 1 , 1 , 14 , x , x , x , x , x , x , x , 15 , 1 , 512 , 20 , 48 , 1 , 512 , 20 , 48 ,
16, TIDL_ConvolutionLayer , 1, 1 , 1 , 15 , x , x , x , x , x , x , x , 16 , 1 , 512 , 20 , 48 , 1 , 512 , 20 , 48 ,
17, TIDL_ConvolutionLayer , 1, 1 , 1 , 16 , x , x , x , x , x , x , x , 17 , 1 , 512 , 20 , 48 , 1 , 512 , 20 , 48 ,
18, TIDL_ConvolutionLayer , 1, 1 , 1 , 17 , x , x , x , x , x , x , x , 18 , 1 , 512 , 20 , 48 , 1 , 512 , 20 , 48 ,
19, TIDL_ConvolutionLayer , 1, 1 , 1 , 18 , x , x , x , x , x , x , x , 19 , 1 , 512 , 20 , 48 , 1 , 512 , 20 , 48 ,
20, TIDL_ConvolutionLayer , 1, 1 , 1 , 19 , x , x , x , x , x , x , x , 20 , 1 , 512 , 20 , 48 , 1 , 512 , 20 , 48 ,
21, TIDL_ConvolutionLayer , 1, 1 , 1 , 20 , x , x , x , x , x , x , x , 21 , 1 , 512 , 20 , 48 , 1 , 512 , 20 , 48 ,
22, TIDL_ConvolutionLayer , 1, 1 , 1 , 21 , x , x , x , x , x , x , x , 22 , 1 , 512 , 20 , 48 , 1 , 512 , 20 , 48 ,
23, TIDL_ConvolutionLayer , 1, 1 , 1 , 22 , x , x , x , x , x , x , x , 23 , 1 , 512 , 20 , 48 , 1 , 512 , 20 , 48 ,
24, TIDL_ConvolutionLayer , 1, 1 , 1 , 23 , x , x , x , x , x , x , x , 24 , 1 , 512 , 20 , 48 , 1 , 512 , 20 , 48 ,
25, TIDL_ConvolutionLayer , 1, 1 , 1 , 24 , x , x , x , x , x , x , x , 25 , 1 , 512 , 20 , 48 , 1 , 512 , 10 , 24 ,
26, TIDL_ConvolutionLayer , 1, 1 , 1 , 25 , x , x , x , x , x , x , x , 26 , 1 , 512 , 10 , 24 , 1 , 1024 , 10 , 24 ,
27, TIDL_ConvolutionLayer , 1, 1 , 1 , 26 , x , x , x , x , x , x , x , 27 , 1 , 1024 , 10 , 24 , 1 , 1024 , 10 , 24 ,
28, TIDL_ConvolutionLayer , 1, 1 , 1 , 27 , x , x , x , x , x , x , x , 28 , 1 , 1024 , 10 , 24 , 1 , 1024 , 10 , 24 ,
29, TIDL_PoolingLayer , 1, 1 , 1 , 28 , x , x , x , x , x , x , x , 29 , 1 , 1024 , 10 , 24 , 1 , 1024 , 5 , 12 ,
30, TIDL_PoolingLayer , 1, 1 , 1 , 29 , x , x , x , x , x , x , x , 30 , 1 , 1024 , 5 , 12 , 1 , 1024 , 3 , 6 ,
31, TIDL_PoolingLayer , 1, 1 , 1 , 30 , x , x , x , x , x , x , x , 31 , 1 , 1024 , 3 , 6 , 1 , 1024 , 2 , 3 ,
32, TIDL_ConvolutionLayer , 1, 1 , 1 , 24 , x , x , x , x , x , x , x , 32 , 1 , 512 , 20 , 48 , 1 , 512 , 20 , 48 ,
33, TIDL_ConvolutionLayer , 1, 1 , 1 , 32 , x , x , x , x , x , x , x , 33 , 1 , 512 , 20 , 48 , 1 , 512 , 20 , 48 ,
34, TIDL_ConvolutionLayer , 1, 1 , 1 , 28 , x , x , x , x , x , x , x , 34 , 1 , 1024 , 10 , 24 , 1 , 1024 , 10 , 24 ,
35, TIDL_ConvolutionLayer , 1, 1 , 1 , 34 , x , x , x , x , x , x , x , 35 , 1 , 1024 , 10 , 24 , 1 , 512 , 10 , 24 ,
36, TIDL_ConvolutionLayer , 1, 1 , 1 , 29 , x , x , x , x , x , x , x , 36 , 1 , 1024 , 5 , 12 , 1 , 1024 , 5 , 12 ,
37, TIDL_ConvolutionLayer , 1, 1 , 1 , 36 , x , x , x , x , x , x , x , 37 , 1 , 1024 , 5 , 12 , 1 , 512 , 5 , 12 ,
38, TIDL_ConvolutionLayer , 1, 1 , 1 , 30 , x , x , x , x , x , x , x , 38 , 1 , 1024 , 3 , 6 , 1 , 1024 , 3 , 6 ,
39, TIDL_ConvolutionLayer , 1, 1 , 1 , 38 , x , x , x , x , x , x , x , 39 , 1 , 1024 , 3 , 6 , 1 , 512 , 3 , 6 ,
40, TIDL_ConvolutionLayer , 1, 1 , 1 , 31 , x , x , x , x , x , x , x , 40 , 1 , 1024 , 2 , 3 , 1 , 1024 , 2 , 3 ,
41, TIDL_ConvolutionLayer , 1, 1 , 1 , 40 , x , x , x , x , x , x , x , 41 , 1 , 1024 , 2 , 3 , 1 , 512 , 2 , 3 ,
42, TIDL_ConvolutionLayer , 1, 1 , 1 , 33 , x , x , x , x , x , x , x , 42 , 1 , 512 , 20 , 48 , 1 , 24 , 20 , 48 ,
43, TIDL_FlattenLayer , 1, 1 , 1 , 42 , x , x , x , x , x , x , x , 43 , 1 , 24 , 20 , 48 , 1 , 1 , 1 ,23040 ,
44, TIDL_ConvolutionLayer , 1, 1 , 1 , 33 , x , x , x , x , x , x , x , 44 , 1 , 512 , 20 , 48 , 1 , 30 , 20 , 48 ,
45, TIDL_FlattenLayer , 1, 1 , 1 , 44 , x , x , x , x , x , x , x , 45 , 1 , 30 , 20 , 48 , 1 , 1 , 1 ,28800 ,
46, TIDL_ConvolutionLayer , 1, 1 , 1 , 35 , x , x , x , x , x , x , x , 47 , 1 , 512 , 10 , 24 , 1 , 24 , 10 , 24 ,
47, TIDL_FlattenLayer , 1, 1 , 1 , 47 , x , x , x , x , x , x , x , 48 , 1 , 24 , 10 , 24 , 1 , 1 , 1 , 5760 ,
48, TIDL_ConvolutionLayer , 1, 1 , 1 , 35 , x , x , x , x , x , x , x , 49 , 1 , 512 , 10 , 24 , 1 , 30 , 10 , 24 ,
49, TIDL_FlattenLayer , 1, 1 , 1 , 49 , x , x , x , x , x , x , x , 50 , 1 , 30 , 10 , 24 , 1 , 1 , 1 , 7200 ,
50, TIDL_ConvolutionLayer , 1, 1 , 1 , 37 , x , x , x , x , x , x , x , 52 , 1 , 512 , 5 , 12 , 1 , 24 , 5 , 12 ,
51, TIDL_FlattenLayer , 1, 1 , 1 , 52 , x , x , x , x , x , x , x , 53 , 1 , 24 , 5 , 12 , 1 , 1 , 1 , 1440 ,
52, TIDL_ConvolutionLayer , 1, 1 , 1 , 37 , x , x , x , x , x , x , x , 54 , 1 , 512 , 5 , 12 , 1 , 30 , 5 , 12 ,
53, TIDL_FlattenLayer , 1, 1 , 1 , 54 , x , x , x , x , x , x , x , 55 , 1 , 30 , 5 , 12 , 1 , 1 , 1 , 1800 ,
54, TIDL_ConvolutionLayer , 1, 1 , 1 , 39 , x , x , x , x , x , x , x , 57 , 1 , 512 , 3 , 6 , 1 , 16 , 3 , 6 ,
55, TIDL_FlattenLayer , 1, 1 , 1 , 57 , x , x , x , x , x , x , x , 58 , 1 , 16 , 3 , 6 , 1 , 1 , 1 , 288 ,
56, TIDL_ConvolutionLayer , 1, 1 , 1 , 39 , x , x , x , x , x , x , x , 59 , 1 , 512 , 3 , 6 , 1 , 20 , 3 , 6 ,
57, TIDL_FlattenLayer , 1, 1 , 1 , 59 , x , x , x , x , x , x , x , 60 , 1 , 20 , 3 , 6 , 1 , 1 , 1 , 360 ,
58, TIDL_ConvolutionLayer , 1, 1 , 1 , 41 , x , x , x , x , x , x , x , 62 , 1 , 512 , 2 , 3 , 1 , 16 , 2 , 3 ,
59, TIDL_FlattenLayer , 1, 1 , 1 , 62 , x , x , x , x , x , x , x , 63 , 1 , 16 , 2 , 3 , 1 , 1 , 1 , 96 ,
60, TIDL_ConvolutionLayer , 1, 1 , 1 , 41 , x , x , x , x , x , x , x , 64 , 1 , 512 , 2 , 3 , 1 , 20 , 2 , 3 ,
61, TIDL_FlattenLayer , 1, 1 , 1 , 64 , x , x , x , x , x , x , x , 65 , 1 , 20 , 2 , 3 , 1 , 1 , 1 , 120 ,
62, TIDL_ConcatLayer , 1, 5 , 1 , 43 , 48 , 53 , 58 , 63 , x , x , x , 67 , 1 , 1 , 1 ,23040 , 1 , 1 , 1 ,30624 ,
63, TIDL_ConcatLayer , 1, 5 , 1 , 45 , 50 , 55 , 60 , 65 , x , x , x , 68 , 1 , 1 , 1 ,28800 , 1 , 1 , 1 ,38280 ,
64, TIDL_DetectionOutputLayer , 1, 2 , 1 , 67 , 68 , x , x , x , x , x , x , 70 , 1 , 1 , 1 ,30624 , 1 , 1 , 1 , 280 ,
65, TIDL_DataLayer , 0, 1 , -1 , 70 , x , x , x , x , x , x , x , 0 , 1 , 1 , 1 , 280 , 0 , 0 , 0 , 0 ,
Layer ID ,inBlkWidth ,inBlkHeight ,inBlkPitch ,outBlkWidth ,outBlkHeight,outBlkPitch ,numInChs ,numOutChs ,numProcInChs,numLclInChs ,numLclOutChs,numProcItrs ,numAccItrs ,numHorBlock ,numVerBlock ,inBlkChPitch,outBlkChPitc,alignOrNot
2 72 68 72 32 32 32 3 32 3 1 8 1 3 12 5 4896 1024 1
3 40 34 40 32 32 32 1 1 1 1 1 1 1 12 5 1360 1024 1
4 32 32 32 32 32 32 32 64 32 7 8 1 5 12 5 1024 1024 1
5 72 44 72 32 20 32 1 1 1 1 1 1 1 6 4 3168 640 1
6 32 20 32 32 20 32 64 128 64 8 8 1 8 6 4 640 640 1
7 40 22 40 32 20 32 1 1 1 1 1 1 1 6 4 880 640 1
8 32 20 32 32 20 32 128 128 128 8 8 1 16 6 4 640 640 1
9 72 44 72 32 20 32 1 1 1 1 1 1 1 3 2 3168 640 1
10 32 20 32 32 20 32 128 256 128 8 8 1 16 3 2 640 640 1
11 40 22 40 32 20 32 1 1 1 1 1 1 1 3 2 880 640 1
12 32 20 32 32 20 32 256 256 256 8 8 1 32 3 2 640 640 1
13 104 44 104 48 20 48 1 1 1 1 1 1 1 1 1 4576 960 1
14 48 4 48 48 4 48 256 512 256 32 8 1 8 1 5 192 192 1
15 50 22 50 48 20 48 1 1 1 1 1 1 1 1 1 1100 960 1
16 48 4 48 48 4 48 512 512 512 32 16 1 16 1 5 192 192 1
17 50 22 50 48 20 48 1 1 1 1 1 1 1 1 1 1100 960 1
18 48 4 48 48 4 48 512 512 512 32 16 1 16 1 5 192 192 1
19 50 22 50 48 20 48 1 1 1 1 1 1 1 1 1 1100 960 1
20 48 4 48 48 4 48 512 512 512 32 16 1 16 1 5 192 192 1
21 50 22 50 48 20 48 1 1 1 1 1 1 1 1 1 1100 960 1
22 48 4 48 48 4 48 512 512 512 32 16 1 16 1 5 192 192 1
23 50 22 50 48 20 48 1 1 1 1 1 1 1 1 1 1100 960 1
24 48 4 48 48 4 48 512 512 512 32 16 1 16 1 5 192 192 1
25 72 24 72 32 10 32 1 1 1 1 1 1 1 1 1 1728 320 1
26 24 10 24 24 10 24 512 1024 512 32 32 1 16 1 1 240 240 1
27 26 12 26 24 10 24 1 1 1 1 1 1 1 1 1 312 240 1
28 24 10 24 24 10 24 1024 1024 1024 32 32 1 32 1 1 240 240 1
32 50 22 50 48 20 48 1 1 1 1 1 1 1 1 1 1100 960 1
33 48 4 48 48 4 48 512 512 512 32 16 1 16 1 5 192 192 1
34 26 12 26 24 10 24 1 1 1 1 1 1 1 1 1 312 240 1
35 24 10 24 24 10 24 1024 512 1024 32 32 1 32 1 1 240 240 1
36 14 7 14 12 5 12 1 1 1 1 1 1 1 1 1 98 60 1
37 12 5 12 12 5 12 1024 512 1024 32 32 1 32 1 1 60 60 1
38 8 5 8 6 3 6 1 1 1 1 1 1 1 1 1 40 18 1
39 6 3 6 6 3 6 1024 512 1024 32 32 1 32 1 1 18 18 1
40 5 4 5 3 2 3 1 1 1 1 1 1 1 1 1 20 6 1
41 3 2 3 3 2 3 1024 512 1024 32 32 1 32 1 1 6 6 1
42 48 4 48 48 4 48 512 32 512 32 16 1 16 1 5 192 192 1
44 48 4 48 48 4 48 512 32 512 32 16 1 16 1 5 192 192 1
46 24 10 24 24 10 24 512 24 512 32 24 1 16 1 1 240 240 1
48 24 10 24 24 10 24 512 30 512 32 30 1 16 1 1 240 240 1
50 12 5 12 12 5 12 512 24 512 32 24 1 16 1 1 60 60 1
52 12 5 12 12 5 12 512 30 512 32 30 1 16 1 1 60 60 1
54 6 3 6 6 3 6 512 16 512 32 16 1 16 1 1 18 18 1
56 6 3 6 6 3 6 512 20 512 32 20 1 16 1 1 18 18 1
58 3 2 3 3 2 3 512 16 512 32 16 1 16 1 1 6 6 1
60 3 2 3 3 2 3 512 20 512 32 20 1 16 1 1 6 6 1
Processing Frame Number : 0
Layer 1 : Out Q : 254 , TIDL_BatchNormLayer , PASSED #MMACs = 0.74, 0.74, Sparsity : 0.00
Layer 2 : Out Q : 17307 , TIDL_ConvolutionLayer, PASSED #MMACs = 53.08, 66.11, Sparsity : -24.54
Layer 3 : Out Q : 7617 , TIDL_ConvolutionLayer, PASSED #MMACs = 17.69, 17.69, Sparsity : 0.00
Layer 4 : Out Q : 13498 , TIDL_ConvolutionLayer, PASSED #MMACs = 125.83, 141.56, Sparsity : -12.50
Layer 5 : Out Q : 5489 , TIDL_ConvolutionLayer, PASSED #MMACs = 8.85, 8.85, Sparsity : 0.00
Layer 6 : Out Q : 16704 , TIDL_ConvolutionLayer, PASSED #MMACs = 125.83, 125.83, Sparsity : 0.00
Layer 7 : Out Q : 4949 , TIDL_ConvolutionLayer, PASSED #MMACs = 17.69, 17.69, Sparsity : 0.00
Layer 8 : Out Q : 20619 , TIDL_ConvolutionLayer, PASSED #MMACs = 251.66, 251.66, Sparsity : 0.00
Layer 9 : Out Q : 5545 , TIDL_ConvolutionLayer, PASSED #MMACs = 4.42, 4.42, Sparsity : 0.00
Layer 10 : Out Q : 27018 , TIDL_ConvolutionLayer, PASSED #MMACs = 125.83, 125.81, Sparsity : 0.01
Layer 11 : Out Q : 4561 , TIDL_ConvolutionLayer, PASSED #MMACs = 8.85, 8.85, Sparsity : 0.00
Layer 12 : Out Q : 29035 , TIDL_ConvolutionLayer, PASSED #MMACs = 251.66, 251.63, Sparsity : 0.01
Layer 13 : Out Q : 6761 , TIDL_ConvolutionLayer, PASSED #MMACs = 2.21, 2.21, Sparsity : 0.00
Layer 14 : Out Q : 45549 , TIDL_ConvolutionLayer, PASSED #MMACs = 125.83, 125.83, Sparsity : 0.00
Layer 15 : Out Q : 7168 , TIDL_ConvolutionLayer, PASSED #MMACs = 4.42, 4.42, Sparsity : 0.00
Layer 16 : Out Q : 42625 , TIDL_ConvolutionLayer, PASSED #MMACs = 251.66, 251.66, Sparsity : 0.00
Layer 17 : Out Q : 6019 , TIDL_ConvolutionLayer, PASSED #MMACs = 4.42, 4.42, Sparsity : 0.00
Layer 18 : Out Q : 67087 , TIDL_ConvolutionLayer, PASSED #MMACs = 251.66, 251.66, Sparsity : 0.00
Layer 19 : Out Q : 9130 , TIDL_ConvolutionLayer, PASSED #MMACs = 4.42, 4.42, Sparsity : 0.00
Layer 20 : Out Q : 58448 , TIDL_ConvolutionLayer, PASSED #MMACs = 251.66, 251.66, Sparsity : 0.00
Layer 21 : Out Q : 8141 , TIDL_ConvolutionLayer, PASSED #MMACs = 4.42, 4.42, Sparsity : 0.00
Layer 22 : Out Q : 55531 , TIDL_ConvolutionLayer, PASSED #MMACs = 251.66, 251.66, Sparsity : 0.00
Layer 23 : Out Q : 8091 , TIDL_ConvolutionLayer, PASSED #MMACs = 4.42, 4.42, Sparsity : 0.00
Layer 24 : Out Q : 56379 , TIDL_ConvolutionLayer, PASSED #MMACs = 251.66, 251.66, Sparsity : 0.00
Layer 25 : Out Q : 11226 , TIDL_ConvolutionLayer, PASSED #MMACs = 1.11, 1.11, Sparsity : 0.00
Layer 26 : Out Q : 78879 , TIDL_ConvolutionLayer, PASSED #MMACs = 125.83, 125.83, Sparsity : 0.00
Layer 27 : Out Q : 10630 , TIDL_ConvolutionLayer, PASSED #MMACs = 2.21, 2.21, Sparsity : 0.00
Layer 28 : Out Q : 52849 , TIDL_ConvolutionLayer, PASSED #MMACs = 251.66, 251.66, Sparsity : 0.00
Layer 29 :TIDL_PoolingLayer, PASSED #MMACs = 0.06, 0.06, Sparsity : 0.00
Layer 30 :TIDL_PoolingLayer, PASSED #MMACs = 0.02, 0.02, Sparsity : 0.00
Layer 31 :TIDL_PoolingLayer, PASSED #MMACs = 0.01, 0.01, Sparsity : 0.00
Layer 32 : Out Q : 10181 , TIDL_ConvolutionLayer, PASSED #MMACs = 4.42, 4.42, Sparsity : 0.00
Layer 33 : Out Q : 17244 , TIDL_ConvolutionLayer, PASSED #MMACs = 251.66, 251.66, Sparsity : 0.00
Layer 34 : Out Q : 11786 , TIDL_ConvolutionLayer, PASSED #MMACs = 2.21, 2.21, Sparsity : 0.00
Layer 35 : Out Q : 16098 , TIDL_ConvolutionLayer, PASSED #MMACs = 125.83, 125.83, Sparsity : 0.00
Layer 36 : Out Q : 28318 , TIDL_ConvolutionLayer, PASSED #MMACs = 0.55, 0.55, Sparsity : 0.00
Layer 37 : Out Q : 21740 , TIDL_ConvolutionLayer, PASSED #MMACs = 31.46, 31.46, Sparsity : 0.00
Layer 38 : Out Q : 25990 , TIDL_ConvolutionLayer, PASSED #MMACs = 0.17, 0.17, Sparsity : 0.00
Layer 39 : Out Q : 24629 , TIDL_ConvolutionLayer, PASSED #MMACs = 9.44, 9.44, Sparsity : 0.00
Layer 40 : Out Q : 26634 , TIDL_ConvolutionLayer, PASSED #MMACs = 0.06, 0.06, Sparsity : 0.00
Layer 41 : Out Q : 34531 , TIDL_ConvolutionLayer, PASSED #MMACs = 3.15, 3.15, Sparsity : 0.00
Layer 42 : Out Q : 2535 , TIDL_ConvolutionLayer, PASSED #MMACs = 15.73, 15.73, Sparsity : 0.00
Layer 43 :TIDL_FlattenLayer, PASSED #MMACs = 0.02, 0.02, Sparsity : 0.00
Layer 44 : Out Q : 4091 , TIDL_ConvolutionLayer, PASSED #MMACs = 15.73, 15.73, Sparsity : 0.00
Layer 45 :TIDL_FlattenLayer, PASSED #MMACs = 0.03, 0.03, Sparsity : 0.00
Layer 46 : Out Q : 9371 , TIDL_ConvolutionLayer, PASSED #MMACs = 2.95, 2.95, Sparsity : 0.00
Layer 47 :TIDL_FlattenLayer, PASSED #MMACs = 0.01, 0.01, Sparsity : 0.00
Layer 48 : Out Q : 5418 , TIDL_ConvolutionLayer, PASSED #MMACs = 3.69, 3.69, Sparsity : 0.00
Layer 49 :TIDL_FlattenLayer, PASSED #MMACs = 0.01, 0.01, Sparsity : 0.00
Layer 50 : Out Q : 8851 , TIDL_ConvolutionLayer, PASSED #MMACs = 0.74, 0.74, Sparsity : 0.00
Layer 51 :TIDL_FlattenLayer, PASSED #MMACs = 0.00, 0.00, Sparsity : 0.00
Layer 52 : Out Q : 6013 , TIDL_ConvolutionLayer, PASSED #MMACs = 0.92, 0.92, Sparsity : 0.00
Layer 53 :TIDL_FlattenLayer, PASSED #MMACs = 0.00, 0.00, Sparsity : 0.00
Layer 54 : Out Q : 7562 , TIDL_ConvolutionLayer, PASSED #MMACs = 0.15, 0.15, Sparsity : 0.00
Layer 55 :TIDL_FlattenLayer, PASSED #MMACs = 0.00, 0.00, Sparsity : 0.00
Layer 56 : Out Q : 5387 , TIDL_ConvolutionLayer, PASSED #MMACs = 0.18, 0.18, Sparsity : 0.00
Layer 57 :TIDL_FlattenLayer, PASSED #MMACs = 0.00, 0.00, Sparsity : 0.00
Layer 58 : Out Q : 6626 , TIDL_ConvolutionLayer, PASSED #MMACs = 0.05, 0.05, Sparsity : 0.00
Layer 59 :TIDL_FlattenLayer, PASSED #MMACs = 0.00, 0.00, Sparsity : 0.00
Layer 60 : Out Q : 7298 , TIDL_ConvolutionLayer, PASSED #MMACs = 0.06, 0.06, Sparsity : 0.00
Layer 61 :TIDL_FlattenLayer, PASSED #MMACs = 0.00, 0.00, Sparsity : 0.00
Layer 62 : Out Q : 2545 , TIDL_ConcatLayer, PASSED #MMACs = 0.00, 0.00, Sparsity : -1.#J
Layer 63 : Out Q : 4107 , TIDL_ConcatLayer, PASSED #MMACs = 0.00, 0.00, Sparsity : -1.#J
Layer 64 : #MMACs = 0.00, 0.00, Sparsity : 0.00
End of config list found !
x , x , x , x , x , x , 53 , 1 , 24 , 5 , 12 , 1 , 1 , 1 , 1440 , 1 ,
54, TIDL_ConvolutionLayer , ctx_output3/sep/relu_mbox_conf 1, 1 , 1 , 37 , x , x , x , x , x , x , x , 54 , 1 , 512 , 5 , 12 , 1 , 30 , 5 , 12 , 921600 ,
55, TIDL_FlattenLayer , ctx_output3/sep/relu_mbox_conf_perm 1, 1 , 1 , 54 , x , x , x , x , x , x , x , 55 , 1 , 30 , 5 , 12 , 1 , 1 , 1 , 1800 , 1 ,
57, TIDL_ConvolutionLayer , ctx_output4/sep/relu_mbox_loc 1, 1 , 1 , 39 , x , x , x , x , x , x , x , 57 , 1 , 512 , 3 , 6 , 1 , 16 , 3 , 6 , 147456 ,
58, TIDL_FlattenLayer , ctx_output4/sep/relu_mbox_loc_perm 1, 1 , 1 , 57 , x , x , x , x , x , x , x , 58 , 1 , 16 , 3 , 6 , 1 , 1 , 1 , 288 , 1 ,
59, TIDL_ConvolutionLayer , ctx_output4/sep/relu_mbox_conf 1, 1 , 1 , 39 , x , x , x , x , x , x , x , 59 , 1 , 512 , 3 , 6 , 1 , 20 , 3 , 6 , 184320 ,
60, TIDL_FlattenLayer , ctx_output4/sep/relu_mbox_conf_perm 1, 1 , 1 , 59 , x , x , x , x , x , x , x , 60 , 1 , 20 , 3 , 6 , 1 , 1 , 1 , 360 , 1 ,
62, TIDL_ConvolutionLayer , ctx_output5/sep/relu_mbox_loc 1, 1 , 1 , 41 , x , x , x , x , x , x , x , 62 , 1 , 512 , 2 , 3 , 1 , 16 , 2 , 3 , 49152 ,
63, TIDL_FlattenLayer , ctx_output5/sep/relu_mbox_loc_perm 1, 1 , 1 , 62 , x , x , x , x , x , x , x , 63 , 1 , 16 , 2 , 3 , 1 , 1 , 1 , 96 , 1 ,
64, TIDL_ConvolutionLayer , ctx_output5/sep/relu_mbox_conf 1, 1 , 1 , 41 , x , x , x , x , x , x , x , 64 , 1 , 512 , 2 , 3 , 1 , 20 , 2 , 3 , 61440 ,
65, TIDL_FlattenLayer , ctx_output5/sep/relu_mbox_conf_perm 1, 1 , 1 , 64 , x , x , x , x , x , x , x , 65 , 1 , 20 , 2 , 3 , 1 , 1 , 1 , 120 , 1 ,
67, TIDL_ConcatLayer , mbox_loc 1, 5 , 1 , 43 , 48 , 53 , 58 , 63 , x , x , x , 67 , 1 , 1 , 1 , 23040 , 1 , 1 , 1 , 30624 , 1 ,
68, TIDL_ConcatLayer , mbox_conf 1, 5 , 1 , 45 , 50 , 55 , 60 , 65 , x , x , x , 68 , 1 , 1 , 1 , 28800 , 1 , 1 , 1 , 38280 , 1 ,
70, TIDL_DetectionOutputLayer , detection_out 2, 2 , 1 , 67 , 68 , x , x , x , x , x , x , 70 , 1 , 1 , 1 , 30624 , 1 , 1 , 1 , 280 , 1 ,
Total Giga Macs : 3.2459
Hi,
Sorry for late reply.
From import log I see that GMAC for this model is "Total Giga Macs : 3.2459", can you compare it with jacintoNets GMAC and I think jacintoNets could be less than the "mobiledetnet" GMAC, so this could be reason as more GMAC gives less fps.
Kindly compare GMAC requirement of both the models will give an idea on fps.
Thanks,
Praveen