Edge Impulse Inc. / Tutorial: Adding sight to your sensors Public
The EON Tuner helps you find the most optimal architecture for your embedded machine-learning application. Clone this project to use the EON Tuner.

Target

No name set

Arduino Portenta H7 (Cortex-M7 480MHz)

100 ms

440 kB

2048 kB

Filters

Status

DSP type

Model type

View

Data set

Variant

Sort

General

F1-score

Precision

Recall

100%
rgb-mobilenetv2-4c4
PERFORMANCE
LATENCY
4772 ms of 100 ms
Exceeds target by 4672 ms
RAM
738 kB of 440 kB
Exceeds target by 298 kB
ROM
586 kB of 2048 kB
DSP NN Unused
IMAGE INPUT

160 |
160

IMAGE

RGB

ACCURACY (KERAS-TRANSFER-IMAGE)

12/17/2021, 8:44:38 PM

100%
grayscale-mobilenetv2-4f0
PERFORMANCE
LATENCY
1710 ms of 100 ms
Exceeds target by 1610 ms
RAM
351 kB of 440 kB
ROM
585 kB of 2048 kB
DSP NN Unused
IMAGE INPUT

96 |
96

IMAGE

Grayscale

ACCURACY (KERAS-TRANSFER-IMAGE)

12/17/2021, 8:37:58 PM

100%
rgb-mobilenetv2-154
PERFORMANCE
LATENCY
4772 ms of 100 ms
Exceeds target by 4672 ms
RAM
369 kB of 440 kB
ROM
2357 kB of 2048 kB
Exceeds target by 309 kB
DSP NN Unused
IMAGE INPUT

160 |
160

IMAGE

RGB

ACCURACY (KERAS-TRANSFER-IMAGE)

12/17/2021, 9:01:39 PM

98%
rgb-conv2d-6b2
PERFORMANCE
LATENCY
202 ms of 100 ms
Exceeds target by 102 ms
RAM
30 kB of 440 kB
ROM
122 kB of 2048 kB
DSP NN Unused
IMAGE INPUT

32 |
32

IMAGE

RGB

ACCURACY (KERAS)

12/17/2021, 8:34:46 PM

98%
rgb-mobilenetv2-297
PERFORMANCE
LATENCY
1718 ms of 100 ms
Exceeds target by 1618 ms
RAM
351 kB of 440 kB
ROM
586 kB of 2048 kB
DSP NN Unused
IMAGE INPUT

96 |
96

IMAGE

RGB

ACCURACY (KERAS-TRANSFER-IMAGE)

12/17/2021, 8:38:12 PM

98%
grayscale-conv2d-dd7
PERFORMANCE
LATENCY
440 ms of 100 ms
Exceeds target by 340 ms
RAM
99 kB of 440 kB
ROM
328 kB of 2048 kB
DSP NN Unused
IMAGE INPUT

96 |
96

IMAGE

Grayscale

ACCURACY (KERAS)

12/17/2021, 9:09:00 PM

96%
grayscale-mobilenetv1-81e
PERFORMANCE
LATENCY
638 ms of 100 ms
Exceeds target by 538 ms
RAM
135 kB of 440 kB
ROM
312 kB of 2048 kB
DSP NN Unused
IMAGE INPUT

96 |
96

IMAGE

Grayscale

ACCURACY (KERAS-TRANSFER-IMAGE)

12/17/2021, 8:41:12 PM

96%
rgb-conv2d-bdc
PERFORMANCE
LATENCY
3706 ms of 100 ms
Exceeds target by 3606 ms
RAM
369 kB of 440 kB
ROM
2357 kB of 2048 kB
Exceeds target by 309 kB
DSP NN Unused
IMAGE INPUT

96 |
96

IMAGE

RGB

ACCURACY (KERAS)

12/17/2021, 8:42:51 PM

96%
grayscale-mobilenetv1-4d9
PERFORMANCE
LATENCY
100 ms of 100 ms
RAM
64 kB of 440 kB
ROM
109 kB of 2048 kB
DSP NN Unused
IMAGE INPUT

96 |
96

IMAGE

Grayscale

ACCURACY (KERAS-TRANSFER-IMAGE)

12/17/2021, 8:45:47 PM

96%
rgb-mobilenetv2-ded
PERFORMANCE
LATENCY
500 ms of 100 ms
Exceeds target by 400 ms
RAM
369 kB of 440 kB
ROM
2357 kB of 2048 kB
Exceeds target by 309 kB
DSP NN Unused
IMAGE INPUT

96 |
96

IMAGE

RGB

ACCURACY (KERAS-TRANSFER-IMAGE)

12/17/2021, 8:51:20 PM

96%
grayscale-mobilenetv2-b13
PERFORMANCE
LATENCY
496 ms of 100 ms
Exceeds target by 396 ms
RAM
287 kB of 440 kB
ROM
173 kB of 2048 kB
DSP NN Unused
IMAGE INPUT

96 |
96

IMAGE

Grayscale

ACCURACY (KERAS-TRANSFER-IMAGE)

12/17/2021, 8:51:51 PM

96%
grayscale-mobilenetv1-238
PERFORMANCE
LATENCY
100 ms of 100 ms
RAM
64 kB of 440 kB
ROM
109 kB of 2048 kB
DSP NN Unused
IMAGE INPUT

96 |
96

IMAGE

Grayscale

ACCURACY (KERAS-TRANSFER-IMAGE)

12/17/2021, 8:51:03 PM

96%
rgb-mobilenetv1-2a3
PERFORMANCE
LATENCY
642 ms of 100 ms
Exceeds target by 542 ms
RAM
135 kB of 440 kB
ROM
312 kB of 2048 kB
DSP NN Unused
IMAGE INPUT

96 |
96

IMAGE

RGB

ACCURACY (KERAS-TRANSFER-IMAGE)

12/17/2021, 8:54:08 PM

96%
rgb-conv2d-76d
PERFORMANCE
LATENCY
726 ms of 100 ms
Exceeds target by 626 ms
RAM
52 kB of 440 kB
ROM
134 kB of 2048 kB
DSP NN Unused
IMAGE INPUT

32 |
32

IMAGE

RGB

ACCURACY (KERAS)

12/17/2021, 8:57:51 PM

96%
rgb-mobilenetv2-20d
PERFORMANCE
LATENCY
584 ms of 100 ms
Exceeds target by 484 ms
RAM
297 kB of 440 kB
ROM
223 kB of 2048 kB
DSP NN Unused
IMAGE INPUT

96 |
96

IMAGE

RGB

ACCURACY (KERAS-TRANSFER-IMAGE)

12/17/2021, 9:05:42 PM