BAE Juhyeong / Keyword spotting Public
Target: Arduino Nano 33 BLE Sense (Cortex-M4F 64MHz) Clone this project
Experiments EON Tuner
The EON Tuner helps you quickly run hyper-parameter sweeps that explore different pre-processing + model architectures optimized for your defined objectives. Clone this project to use the EON Tuner.

Target

No name set

Arduino Nano 33 BLE Sense (Cortex-M4F 64MHz)

100 ms

256 kB

1024 kB

Filters

Status

DSP type

Model type

View

Data set

Variant

Sort

General

F1-score

Precision

Recall

Run #1 (Sep 10 2024, 00:56:51, finished)
87%
mfcc-conv1d-b40
View
PERFORMANCE
LATENCY
5 ms of 100 ms
RAM
5 kB of 256 kB
ROM
40 kB of 1024 kB
DSP NN Unused
TIME-SERIES INPUT

1000 ms |
1000 ms |
Enabled

MFCC

0.05 | 0.05 | 32

ACCURACY (OVERALL)
CLASSIFICATION

0.005 | 100 | 87%

Type Filters Kernel Rate
Data augmentation
conv1d 32 3 -
conv1d 64 3 -
dropout - - 0.5

9/10/2024, 1:12:53 AM

87%
mfe-conv1d-132
PERFORMANCE
LATENCY
4 ms of 100 ms
RAM
6 kB of 256 kB
ROM
42 kB of 1024 kB
DSP NN Unused
TIME-SERIES INPUT

1000 ms |
1000 ms |
Enabled

MFE

0.05 | 0.05 | 32

ACCURACY (OVERALL)
CLASSIFICATION

0.005 | 100 | 87%

Type Filters Kernel Rate
conv1d 16 3 -
conv1d 32 3 -
conv1d 64 3 -
dropout - - 0.5

9/10/2024, 1:15:52 AM

85%
mfcc-conv1d-752
PERFORMANCE
LATENCY
7 ms of 100 ms
RAM
6 kB of 256 kB
ROM
42 kB of 1024 kB
DSP NN Unused
TIME-SERIES INPUT

1000 ms |
1000 ms |
Enabled

MFCC

0.05 | 0.025 | 40

ACCURACY (OVERALL)
CLASSIFICATION

0.005 | 100 | 85%

Type Filters Kernel Rate
conv1d 16 3 -
conv1d 32 3 -
conv1d 64 3 -
dropout - - 0.5

9/10/2024, 1:10:38 AM

84%
mfcc-conv1d-d80
PERFORMANCE
LATENCY
4 ms of 100 ms
RAM
4 kB of 256 kB
ROM
34 kB of 1024 kB
DSP NN Unused
TIME-SERIES INPUT

1000 ms |
1000 ms |
Enabled

MFCC

0.032 | 0.032 | 40

ACCURACY (OVERALL)
CLASSIFICATION

0.005 | 100 | 84%

Type Filters Kernel Rate
Data augmentation
conv1d 16 3 -
conv1d 32 3 -
dropout - - 0.5

9/10/2024, 1:07:18 AM

83%
mfcc-conv1d-4e5
PERFORMANCE
LATENCY
5 ms of 100 ms
RAM
4 kB of 256 kB
ROM
34 kB of 1024 kB
DSP NN Unused
TIME-SERIES INPUT

1000 ms |
1000 ms |
Enabled

MFCC

0.032 | 0.032 | 40

ACCURACY (OVERALL)
CLASSIFICATION

0.005 | 100 | 83%

Type Filters Kernel Rate
conv1d 16 3 -
conv1d 32 3 -
dropout - - 0.5

9/10/2024, 1:14:49 AM

83%
mfcc-conv1d-4a0
PERFORMANCE
LATENCY
6 ms of 100 ms
RAM
6 kB of 256 kB
ROM
42 kB of 1024 kB
DSP NN Unused
TIME-SERIES INPUT

1000 ms |
1000 ms |
Enabled

MFCC

0.05 | 0.025 | 40

ACCURACY (OVERALL)
CLASSIFICATION

0.005 | 100 | 83%

Type Filters Kernel Rate
Data augmentation
conv1d 8 3 -
conv1d 16 3 -
conv1d 32 3 -
conv1d 64 3 -
dropout - - 0.25

9/10/2024, 1:09:54 AM

82%
mfcc-conv1d-2e8
PERFORMANCE
LATENCY
25 ms of 100 ms
RAM
9 kB of 256 kB
ROM
69 kB of 1024 kB
DSP NN Unused
TIME-SERIES INPUT

1000 ms |
1000 ms |
Enabled

MFCC

0.05 | 0.025 | 40

ACCURACY (OVERALL)
CLASSIFICATION

0.005 | 100 | 82%

Type Filters Kernel Rate
Data augmentation
conv1d 32 3 -
conv1d 64 3 -
conv1d 128 3 -
dropout - - 0.5

9/10/2024, 1:18:24 AM

82%
mfe-conv1d-b94
PERFORMANCE
LATENCY
73 ms of 100 ms
RAM
15 kB of 256 kB
ROM
42 kB of 1024 kB
DSP NN Unused
TIME-SERIES INPUT

1000 ms |
1000 ms |
Enabled

MFE

0.05 | 0.05 | 32

ACCURACY (OVERALL)
CLASSIFICATION

0.005 | 100 | 82%

Type Filters Kernel Rate
conv1d 8 3 -
conv1d 16 3 -
conv1d 32 3 -
conv1d 64 3 -
dropout - - 0.5

9/10/2024, 1:14:39 AM

82%
mfcc-conv1d-9a6
PERFORMANCE
LATENCY
321 ms of 100 ms
Exceeds target by 221 ms
RAM
22 kB of 256 kB
ROM
69 kB of 1024 kB
DSP NN Unused
TIME-SERIES INPUT

1000 ms |
1000 ms |
Enabled

MFCC

0.05 | 0.025 | 40

ACCURACY (OVERALL)
CLASSIFICATION

0.005 | 100 | 82%

Type Filters Kernel Rate
Data augmentation
conv1d 16 3 -
conv1d 32 3 -
conv1d 64 3 -
conv1d 128 3 -
dropout - - 0.25

9/10/2024, 1:03:55 AM

81%
mfcc-conv2d-d10
PERFORMANCE
LATENCY
18 ms of 100 ms
RAM
6 kB of 256 kB
ROM
37 kB of 1024 kB
DSP NN Unused
TIME-SERIES INPUT

1000 ms |
1000 ms |
Enabled

MFCC

0.05 | 0.05 | 32

ACCURACY (OVERALL)
CLASSIFICATION

0.005 | 100 | 81%

Type Filters Kernel Rate
conv2d 8 3 -
conv2d 16 3 -
conv2d 32 3 -
dropout - - 0.25

9/10/2024, 1:10:57 AM

80%
mfcc-conv1d-a85
PERFORMANCE
LATENCY
145 ms of 100 ms
Exceeds target by 45 ms
RAM
22 kB of 256 kB
ROM
170 kB of 1024 kB
DSP NN Unused
TIME-SERIES INPUT

1000 ms |
1000 ms |
Enabled

MFCC

0.05 | 0.05 | 32

ACCURACY (OVERALL)
CLASSIFICATION

0.005 | 100 | 80%

Type Filters Kernel Rate
conv1d 32 3 -
conv1d 64 3 -
conv1d 128 3 -
conv1d 256 3 -
dropout - - 0.5

9/10/2024, 1:05:54 AM

78%
mfe-conv1d-a7e
PERFORMANCE
LATENCY
147 ms of 100 ms
Exceeds target by 47 ms
RAM
17 kB of 256 kB
ROM
43 kB of 1024 kB
DSP NN Unused
TIME-SERIES INPUT

1000 ms |
1000 ms |
Enabled

MFE

0.05 | 0.025 | 32

ACCURACY (OVERALL)
CLASSIFICATION

0.005 | 100 | 78%

Type Filters Kernel Rate
conv1d 32 3 -
conv1d 64 3 -
dropout - - 0.25

9/10/2024, 1:18:06 AM

77%
mfcc-conv2d-cf2
PERFORMANCE
LATENCY
21 ms of 100 ms
RAM
6 kB of 256 kB
ROM
33 kB of 1024 kB
DSP NN Unused
TIME-SERIES INPUT

1000 ms |
1000 ms |
Enabled

MFCC

0.032 | 0.032 | 40

ACCURACY (OVERALL)
CLASSIFICATION

0.005 | 100 | 77%

Type Filters Kernel Rate
conv2d 8 3 -
conv2d 16 3 -
dropout - - 0.5

9/10/2024, 1:15:02 AM

75%
mfcc-conv2d-a79
PERFORMANCE
LATENCY
271 ms of 100 ms
Exceeds target by 171 ms
RAM
18 kB of 256 kB
ROM
33 kB of 1024 kB
DSP NN Unused
TIME-SERIES INPUT

1000 ms |
1000 ms |
Enabled

MFCC

0.032 | 0.032 | 40

ACCURACY (OVERALL)
CLASSIFICATION

0.005 | 100 | 75%

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

9/10/2024, 1:08:03 AM

73%
mfcc-conv2d-786
PERFORMANCE
LATENCY
173 ms of 100 ms
Exceeds target by 73 ms
RAM
15 kB of 256 kB
ROM
32 kB of 1024 kB
DSP NN Unused
TIME-SERIES INPUT

1000 ms |
1000 ms |
Enabled

MFCC

0.05 | 0.05 | 40

ACCURACY (OVERALL)
CLASSIFICATION

0.005 | 100 | 73%

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

9/10/2024, 1:08:12 AM

Help & documentation Get help on the forum Send feedback Contact sales
Join us
Terms of Service Privacy policy