Nathaniel Felleke / Trash Image Detection 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

Raspberry Pi 4

100 ms

8388608 kB

33554432 kB

Filters

Status

DSP type

Model type

View

Data set

Variant

Sort

General

F1-score

Precision

Recall

83%
rgb-mobilenetv2-212
PERFORMANCE
LATENCY
14 ms of 100 ms
RAM
738 kB of 8388608 kB
ROM
586 kB of 33554432 kB
DSP NN Unused
IMAGE INPUT

160 |
160

IMAGE

RGB

ACCURACY (KERAS-TRANSFER-IMAGE)
TRANSFER LEARNING (IMAGES)

0.0005 | 20 | 83%

MobileNetV2 160x160 0.35
16 | 0.5

6/3/2022, 5:16:11 AM

82%
rgb-mobilenetv2-cc6
PERFORMANCE
LATENCY
12 ms of 100 ms
RAM
738 kB of 8388608 kB
ROM
649 kB of 33554432 kB
DSP NN Unused
IMAGE INPUT

160 |
160

IMAGE

RGB

ACCURACY (KERAS-TRANSFER-IMAGE)
TRANSFER LEARNING (IMAGES)

0.0005 | 20 | 82%

MobileNetV2 0.35
64 | 0.1

6/3/2022, 5:16:39 AM

82%
rgb-mobilenetv2-949
PERFORMANCE
LATENCY
11 ms of 100 ms
RAM
738 kB of 8388608 kB
ROM
586 kB of 33554432 kB
DSP NN Unused
IMAGE INPUT

160 |
160

IMAGE

RGB

ACCURACY (KERAS-TRANSFER-IMAGE)
TRANSFER LEARNING (IMAGES)

0.0005 | 20 | 82%

MobileNetV2 0.35
16 | 0.1 |

6/3/2022, 5:18:12 AM

82%
rgb-mobilenetv1-3f0
PERFORMANCE
LATENCY
5 ms of 100 ms
RAM
264 kB of 8388608 kB
ROM
324 kB of 33554432 kB
DSP NN Unused
IMAGE INPUT

160 |
160

IMAGE

RGB

ACCURACY (KERAS-TRANSFER-IMAGE)
TRANSFER LEARNING (IMAGES)

0.0005 | 20 | 82%

MobileNetV1 0.25
64 | 0.1 |

6/3/2022, 5:24:54 AM

81%
rgb-mobilenetv1-399
PERFORMANCE
LATENCY
6 ms of 100 ms
RAM
264 kB of 8388608 kB
ROM
324 kB of 33554432 kB
DSP NN Unused
IMAGE INPUT

160 |
160

IMAGE

RGB

ACCURACY (KERAS-TRANSFER-IMAGE)
TRANSFER LEARNING (IMAGES)

0.0005 | 20 | 81%

MobileNetV1 0.25
64 | 0.5

6/3/2022, 5:27:27 AM

79%
rgb-mobilenetv2-fd8
PERFORMANCE
LATENCY
12 ms of 100 ms
RAM
670 kB of 8388608 kB
ROM
1633 kB of 33554432 kB
DSP NN Unused
IMAGE INPUT

96 |
96

IMAGE

RGB

ACCURACY (KERAS-TRANSFER-IMAGE)
TRANSFER LEARNING (IMAGES)

0.0005 | 20 | 79%

MobileNetV2 160x160 0.75
16 | 0.5

6/3/2022, 5:23:23 AM

70%
grayscale-mobilenetv1-962
PERFORMANCE
LATENCY
6 ms of 100 ms
RAM
264 kB of 8388608 kB
ROM
312 kB of 33554432 kB
DSP NN Unused
IMAGE INPUT

160 |
160

IMAGE

Grayscale

ACCURACY (KERAS-TRANSFER-IMAGE)
TRANSFER LEARNING (IMAGES)

0.0005 | 20 | 70%

MobileNetV1 0.25
16 | 0.5

6/3/2022, 5:20:42 AM

69%
grayscale-mobilenetv1-b34
PERFORMANCE
LATENCY
6 ms of 100 ms
RAM
264 kB of 8388608 kB
ROM
312 kB of 33554432 kB
DSP NN Unused
IMAGE INPUT

160 |
160

IMAGE

Grayscale

ACCURACY (KERAS-TRANSFER-IMAGE)
TRANSFER LEARNING (IMAGES)

0.0005 | 20 | 69%

MobileNetV1 0.25
16 | 0.1 |

6/3/2022, 5:22:27 AM

67%
rgb-conv2d-085
PERFORMANCE
LATENCY
2 ms of 100 ms
RAM
51 kB of 8388608 kB
ROM
42 kB of 33554432 kB
DSP NN Unused
IMAGE INPUT

64 |
64

IMAGE

RGB

ACCURACY (KERAS)
CLASSIFICATION (KERAS)

0.0005 | 10 | 67%

Type Filters Kernel Rate
conv2d 8 3 -
conv2d 16 3 -
dropout - - 0.25

6/3/2022, 5:16:52 AM

62%
grayscale-conv2d-8b1
PERFORMANCE
LATENCY
44 ms of 100 ms
RAM
30 kB of 8388608 kB
ROM
59 kB of 33554432 kB
DSP NN Unused
IMAGE INPUT

32 |
32

IMAGE

Grayscale

ACCURACY (KERAS)
CLASSIFICATION (KERAS)

0.0005 | 10 | 62%

Type Filters Kernel Rate
conv2d 16 3 -
conv2d 32 3 -
conv2d 64 3 -
dropout - - 0.25

6/3/2022, 5:23:06 AM