수행기록퀘스트1
가이드대로 따라한 화면을 캡쳐해 보았습니다.
.png)
.png)
.png)
<분석 내용 복사>
Analyzing model
C:/Users/USER/STM32Cube/Repository/Packs/STMicroelectronics/X-CUBE-AI/7.1.0/Utilities/windows/stm32ai analyze --name network -m D:/work/test/contest/4_cubeai/en.fp-ai-sensing1/STM32CubeFunctionPack_SENSING1_V4.0.3/Utilities/AI_Ressources/models/asc_93_Q.h5 --type keras --compression 1 --verbosity 1 --workspace C:\Users\USER\AppData\Local\Temp\mxAI_workspace23804064338050017264679071903012568 --output C:\Users\USER\.stm32cubemx\network_output
Neural Network Tools for STM32AI v1.6.0 (STM.ai v7.1.0-RC3)
Exec/report summary (analyze)
------------------------------------------------------------------------------------------------------------------------
model file : D:\work\test\contest\4_cubeai\en.fp-ai-sensing1\STM32CubeFunctionPack_SENSING1_V4.0.3\Utilities\AI_Ressources\models\asc_93_Q.h5
type : keras
c_name : network
compression : None
workspace dir : C:\Users\USER\AppData\Local\Temp\mxAI_workspace23804064338050017264679071903012568
output dir : C:\Users\USER\.stm32cubemx\network_output
model_name : asc_93_Q
model_hash : 70ee217758c82a261fc1ed12413ca77d
input 1/1 : 'quantize_conv2d_11_input'
960 items, 3.75 KiB, ai_float, float, (1,30,32,1), domain:user/
output 1/1 : 'softmax_10'
3 items, 12 B, ai_float, float, (1,1,1,3), domain:user/
params # : 7,703 items (30.09 KiB)
macc : 517,373
weights (ro) : 30,812 B (30.09 KiB) (1 segment)
activations (rw) : 17,280 B (16.88 KiB) (1 segment)
ram (total) : 21,132 B (20.64 KiB) = 17,280 + 3,840 + 12
Model name - asc_93_Q ['quantize_conv2d_11_input'] ['softmax_10']
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------
id layer (type) oshape param/size macc connected to | c_size c_macc c_type
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------
0 quantize_conv2d_11_input (Input) (None,30,32,1) |
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------
1 quantize_conv2d_11_conv2d (Conv2D) (None,28,30,16) 160/640 120,976 quantize_conv2d_11_input | +26,880(+22.2%) optimized_conv2d()[0]
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------
2 quantize_1 (Nonlinearity) (None,28,30,16) 13,440 quantize_conv2d_11_conv2d | -13,440(-100.0%)
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------
3 quantize_max_pooling2d_11 (Pool) (None,14,15,16) 13,440 quantize_1 | -13,440(-100.0%)
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------
5 quantize_conv2d_12_conv2d (Conv2D) (None,12,13,16) 2,320/9,280 359,440 quantize_max_pooling2d_11 | +4,800(+1.3%) optimized_conv2d()[1]
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------
6 quantize_4 (Nonlinearity) (None,12,13,16) 2,496 quantize_conv2d_12_conv2d | -2,496(-100.0%)
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------
7 quantize_max_pooling2d_12 (Pool) (None,6,6,16) 2,304 quantize_4 | -2,304(-100.0%)
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------
9 flatten_6 (Reshape) (None,1,1,576) quantize_max_pooling2d_12 |
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------
10 quantize_dense_11_dense (Dense) (None,1,1,9) 5,193/20,772 5,193 flatten_6 | dense()[2]
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------
11 quantize_8 (Nonlinearity) (None,1,1,9) 9 quantize_dense_11_dense | nl()[3]
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------
13 quantize_dense_12_dense (Dense) (None,1,1,3) 30/120 30 quantize_8 | dense()[4]
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------
14 softmax_10 (Nonlinearity) (None,1,1,3) 45 quantize_dense_12_dense | nl()/o[5]
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------
model/c-model: macc=517,373/517,373 weights=30,812/30,812 activations=--/17,280 io=--/3,852
Complexity report per layer - macc=517,373 weights=30,812 act=17,280 ram_io=3,852
------------------------------------------------------------------------------------------
id name c_macc c_rom c_id
------------------------------------------------------------------------------------------
1 quantize_conv2d_11_conv2d ||||||| 28.6% | 2.1% [0]
5 quantize_conv2d_12_conv2d |||||||||||||||| 70.4% ||||||| 30.1% [1]
10 quantize_dense_11_dense | 1.0% |||||||||||||||| 67.4% [2]
11 quantize_8 | 0.0% | 0.0% [3]
13 quantize_dense_12_dense | 0.0% | 0.4% [4]
14 softmax_10 | 0.0% | 0.0% [5]
Creating txt report file C:\Users\USER\.stm32cubemx\network_output\network_analyze_report.txt
elapsed time (analyze): 0.506s
Analyze complete on AI model
network graph 그림 캡쳐
위 그래프 전체 그림
.png)
이상입니다.
-Tiel-
- 첨부파일
- network_analyze_report.txt 다운로드
로그인 후
참가 상태를 확인할 수 있습니다.