Edge Impulse Inc. / Tutorial - Syntiant TinyML Circular Motion Public

EON Tuner

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

Motion events

Cortex-M7 216MHz

100 ms

340 kB

1024 kB

Filters

Status

DSP type

Network type

View

Data set

Precision

Sort

General

F1-score

Precision

Recall

spectr-dense-f51
PERFORMANCE
LATENCY
100 ms
RAM
340 kB
ROM
1024 kB
Unused
INPUT

1000 ms | 1000 ms

SPECTRAL-ANALYSIS

1024

ACCURACY
NEURAL NETWORK (KERAS)

0.0005 | 30

Type Filters Kernel Rate
dense 80 - -
dense 40 - -
dense 20 - -

spectr-dense-f60
PERFORMANCE
LATENCY
100 ms
RAM
340 kB
ROM
1024 kB
Unused
INPUT

4000 ms | 4000 ms

SPECTRAL-ANALYSIS

1024

ACCURACY
NEURAL NETWORK (KERAS)

0.0005 | 30

Type Filters Kernel Rate
dense 80 - -
dropout - - 0.5
dense 40 - -
dropout - - 0.5

spectr-dense-4b4
PERFORMANCE
LATENCY
100 ms
RAM
340 kB
ROM
1024 kB
Unused
INPUT

1000 ms | 250 ms

SPECTRAL-ANALYSIS

1024

ACCURACY
NEURAL NETWORK (KERAS)

0.0005 | 30

Type Filters Kernel Rate
dense 80 - -
dropout - - 0.5
dense 40 - -
dropout - - 0.5

spectr-dense-05f
PERFORMANCE
LATENCY
100 ms
RAM
340 kB
ROM
1024 kB
Unused
INPUT

1000 ms | 500 ms

SPECTRAL-ANALYSIS

128

ACCURACY
NEURAL NETWORK (KERAS)

0.0005 | 30

Type Filters Kernel Rate
dense 20 - -
dense 10 - -

spectr-dense-75b
PERFORMANCE
LATENCY
100 ms
RAM
340 kB
ROM
1024 kB
Unused
INPUT

1000 ms | 500 ms

SPECTRAL-ANALYSIS

1024

ACCURACY
NEURAL NETWORK (KERAS)

0.0005 | 30

Type Filters Kernel Rate
dense 80 - -
dropout - - 0.25
dense 40 - -
dropout - - 0.25

spectr-dense-7e4
PERFORMANCE
LATENCY
100 ms
RAM
340 kB
ROM
1024 kB
Unused
INPUT

1000 ms | 250 ms

SPECTRAL-ANALYSIS

1024

ACCURACY
NEURAL NETWORK (KERAS)

0.0005 | 30

Type Filters Kernel Rate
dense 80 - -
dropout - - 0.25
dense 40 - -
dropout - - 0.25

spectr-dense-83b
PERFORMANCE
LATENCY
100 ms
RAM
340 kB
ROM
1024 kB
Unused
INPUT

1000 ms | 1000 ms

SPECTRAL-ANALYSIS

1024

ACCURACY
NEURAL NETWORK (KERAS)

0.0005 | 30

Type Filters Kernel Rate
dense 20 - -
dropout - - 0.5
dense 10 - -
dropout - - 0.5

spectr-dense-fa6
PERFORMANCE
LATENCY
100 ms
RAM
340 kB
ROM
1024 kB
Unused
INPUT

1000 ms | 1000 ms

SPECTRAL-ANALYSIS

128

ACCURACY
NEURAL NETWORK (KERAS)

0.0005 | 30

Type Filters Kernel Rate
dense 40 - -
dense 20 - -
dense 10 - -

spectr-dense-56d
PERFORMANCE
LATENCY
100 ms
RAM
340 kB
ROM
1024 kB
Unused
INPUT

4000 ms | 4000 ms

SPECTRAL-ANALYSIS

128

ACCURACY
NEURAL NETWORK (KERAS)

0.0005 | 30

Type Filters Kernel Rate
dense 40 - -
dense 20 - -
dense 10 - -

spectr-dense-438
PERFORMANCE
LATENCY
100 ms
RAM
340 kB
ROM
1024 kB
Unused
INPUT

4000 ms | 2000 ms

SPECTRAL-ANALYSIS

1024

ACCURACY
NEURAL NETWORK (KERAS)

0.0005 | 30

Type Filters Kernel Rate
dense 80 - -
dropout - - 0.5
dense 40 - -
dropout - - 0.5
dense 20 - -
dropout - - 0.5

spectr-dense-e87
PERFORMANCE
LATENCY
100 ms
RAM
340 kB
ROM
1024 kB
Unused
INPUT

4000 ms | 4000 ms

SPECTRAL-ANALYSIS

128

ACCURACY
NEURAL NETWORK (KERAS)

0.0005 | 30

Type Filters Kernel Rate
dense 80 - -
dropout - - 0.5
dense 40 - -
dropout - - 0.5

spectr-dense-d5e
PERFORMANCE
LATENCY
100 ms
RAM
340 kB
ROM
1024 kB
Unused
INPUT

1000 ms | 500 ms

SPECTRAL-ANALYSIS

1024

ACCURACY
NEURAL NETWORK (KERAS)

0.0005 | 30

Type Filters Kernel Rate
dense 80 - -
dropout - - 0.25
dense 40 - -
dropout - - 0.25
dense 20 - -
dropout - - 0.25

spectr-dense-0ed
PERFORMANCE
LATENCY
100 ms
RAM
340 kB
ROM
1024 kB
Unused
INPUT

1000 ms | 250 ms

SPECTRAL-ANALYSIS

128

ACCURACY
NEURAL NETWORK (KERAS)

0.0005 | 30

Type Filters Kernel Rate
dense 20 - -
dropout - - 0.5
dense 10 - -
dropout - - 0.5
dense 5 - -
dropout - - 0.5

spectr-dense-1de
PERFORMANCE
LATENCY
100 ms
RAM
340 kB
ROM
1024 kB
Unused
INPUT

1000 ms | 250 ms

SPECTRAL-ANALYSIS

1024

ACCURACY
NEURAL NETWORK (KERAS)

0.0005 | 30

Type Filters Kernel Rate
dense 80 - -
dropout - - 0.25
dense 40 - -
dropout - - 0.25
dense 20 - -
dropout - - 0.25

spectr-dense-3de
PERFORMANCE
LATENCY
100 ms
RAM
340 kB
ROM
1024 kB
Unused
INPUT

4000 ms | 1000 ms

SPECTRAL-ANALYSIS

1024

ACCURACY
NEURAL NETWORK (KERAS)

0.0005 | 30

Type Filters Kernel Rate
dense 40 - -
dense 20 - -