Arnulfo Silva / TI-ML-Audio-Classification Public
Clone this project

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

Keyword spotting

TI LAUNCHXL-CC1352P (Cortex-M4F 48MHz)

500 ms

80 kB

352 kB

Filters

Status

DSP type

Network type

View

Data set

Precision

Sort

General

F1-score

Precision

Recall

93%
mfcc-conv1d-fb1
Show logs
PERFORMANCE
LATENCY
483 ms of 500 ms
RAM
30 kB of 80 kB
ROM
84 kB of 352 kB
DSP NN Unused
INPUT

2000 ms | 1000 ms

MFCC

0.05 | 0.025 | 32

ACCURACY
NEURAL NETWORK (KERAS)

0.005 | 100

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

4/4/2022, 6:27:16 PM

86%
mfe-conv2d-c3f
Show logs
PERFORMANCE
LATENCY
920 ms of 500 ms
Exceeds target by 420 ms
RAM
25 kB of 80 kB
ROM
61 kB of 352 kB
DSP NN Unused
INPUT

1000 ms | 500 ms

MFE

0.05 | 0.05 | 40

ACCURACY
NEURAL NETWORK (KERAS)

0.005 | 100

Type Filters Kernel Rate
Data augmentation
conv2d 8 3 -
conv2d 16 3 -
conv2d 32 3 -
conv2d 64 3 -
dropout - - 0.5

4/4/2022, 6:22:28 PM

78%
mfe-conv1d-d5e
Show logs
PERFORMANCE
LATENCY
116 ms of 500 ms
RAM
29 kB of 80 kB
ROM
78 kB of 352 kB
DSP NN Unused
INPUT

1000 ms | 1000 ms

MFE

0.032 | 0.016 | 40

ACCURACY
NEURAL NETWORK (KERAS)

0.005 | 100

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

4/4/2022, 6:34:42 PM

74%
mfe-conv2d-5b1
Show logs
PERFORMANCE
LATENCY
595 ms of 500 ms
Exceeds target by 95 ms
RAM
25 kB of 80 kB
ROM
61 kB of 352 kB
DSP NN Unused
INPUT

1000 ms | 1000 ms

MFE

0.05 | 0.05 | 40

ACCURACY
NEURAL NETWORK (KERAS)

0.005 | 100

Type Filters Kernel Rate
conv2d 8 3 -
conv2d 16 3 -
conv2d 32 3 -
conv2d 64 3 -
dropout - - 0.25

4/4/2022, 6:35:20 PM

72%
mfcc-conv1d-d2b
Show logs
PERFORMANCE
LATENCY
50 ms of 500 ms
RAM
24 kB of 80 kB
ROM
46 kB of 352 kB
DSP NN Unused
INPUT

1000 ms | 500 ms

MFCC

0.02 | 0.02 | 32

ACCURACY
NEURAL NETWORK (KERAS)

0.005 | 100

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

4/4/2022, 6:23:19 PM

68%
mfcc-conv1d-19b
Show logs
PERFORMANCE
LATENCY
59 ms of 500 ms
RAM
23 kB of 80 kB
ROM
45 kB of 352 kB
DSP NN Unused
INPUT

1000 ms | 500 ms

MFCC

0.02 | 0.02 | 32

ACCURACY
NEURAL NETWORK (KERAS)

0.005 | 100

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

4/4/2022, 6:24:32 PM

68%
mfcc-conv1d-d11
Show logs
PERFORMANCE
LATENCY
22 ms of 500 ms
RAM
26 kB of 80 kB
ROM
38 kB of 352 kB
DSP NN Unused
INPUT

1000 ms | 500 ms

MFCC

0.032 | 0.016 | 40

ACCURACY
NEURAL NETWORK (KERAS)

0.005 | 100

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

4/4/2022, 6:33:44 PM

67%
mfcc-conv1d-095
Show logs
PERFORMANCE
LATENCY
21 ms of 500 ms
RAM
21 kB of 80 kB
ROM
45 kB of 352 kB
DSP NN Unused
INPUT

1000 ms | 500 ms

MFCC

0.05 | 0.025 | 40

ACCURACY
NEURAL NETWORK (KERAS)

0.005 | 100

Type Filters Kernel Rate
Data augmentation
conv1d 8 3 -
conv1d 16 3 -
dropout - - 0.5
dense 64 - -
dropout - - 0.5

4/4/2022, 6:35:48 PM

66%
mfcc-conv2d-041
Show logs
PERFORMANCE
LATENCY
233 ms of 500 ms
RAM
22 kB of 80 kB
ROM
82 kB of 352 kB
DSP NN Unused
INPUT

1000 ms | 500 ms

MFCC

0.05 | 0.05 | 32

ACCURACY
NEURAL NETWORK (KERAS)

0.005 | 100

Type Filters Kernel Rate
Data augmentation
conv2d 16 3 -
conv2d 32 3 -
conv2d 64 3 -
dropout - - 0.5
dense 64 - -
dropout - - 0.5

4/4/2022, 6:28:57 PM

65%
mfcc-conv1d-828
Show logs
PERFORMANCE
LATENCY
18 ms of 500 ms
RAM
24 kB of 80 kB
ROM
36 kB of 352 kB
DSP NN Unused
INPUT

2000 ms | 2000 ms

MFCC

0.032 | 0.032 | 32

ACCURACY
NEURAL NETWORK (KERAS)

0.005 | 100

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

4/4/2022, 6:25:49 PM

65%
mfcc-conv1d-c0a
Show logs
PERFORMANCE
LATENCY
552 ms of 500 ms
Exceeds target by 52 ms
RAM
26 kB of 80 kB
ROM
95 kB of 352 kB
DSP NN Unused
INPUT

1000 ms | 1000 ms

MFCC

0.02 | 0.02 | 40

ACCURACY
NEURAL NETWORK (KERAS)

0.005 | 100

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

4/4/2022, 6:21:14 PM

65%
mfcc-conv1d-c68
Show logs
PERFORMANCE
LATENCY
630 ms of 500 ms
Exceeds target by 130 ms
RAM
26 kB of 80 kB
ROM
45 kB of 352 kB
DSP NN Unused
INPUT

1000 ms | 1000 ms

MFCC

0.02 | 0.02 | 40

ACCURACY
NEURAL NETWORK (KERAS)

0.005 | 100

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

4/4/2022, 6:24:27 PM

59%
mfe-conv2d-fb9
Show logs
PERFORMANCE
LATENCY
360 ms of 500 ms
RAM
40 kB of 80 kB
ROM
44 kB of 352 kB
DSP NN Unused
INPUT

1000 ms | 1000 ms

MFE

0.02 | 0.02 | 40

ACCURACY
NEURAL NETWORK (KERAS)

0.005 | 100

Type Filters Kernel Rate
conv2d 8 3 -
conv2d 16 3 -
conv2d 32 3 -
dropout - - 0.5

4/4/2022, 6:38:04 PM

58%
mfcc-conv1d-cb4
Show logs
PERFORMANCE
LATENCY
62 ms of 500 ms
RAM
21 kB of 80 kB
ROM
38 kB of 352 kB
DSP NN Unused
INPUT

1000 ms | 1000 ms

MFCC

0.02 | 0.02 | 32

ACCURACY
NEURAL NETWORK (KERAS)

0.005 | 100

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

4/4/2022, 6:15:23 PM

56%
mfcc-conv2d-d10
Show logs
PERFORMANCE
LATENCY
2840 ms of 500 ms
Exceeds target by 2340 ms
RAM
30 kB of 80 kB
ROM
149 kB of 352 kB
DSP NN Unused
INPUT

1000 ms | 1000 ms

MFCC

0.032 | 0.032 | 32

ACCURACY
NEURAL NETWORK (KERAS)

0.005 | 100

Type Filters Kernel Rate
Data augmentation
conv2d 16 3 -
conv2d 32 3 -
conv2d 64 3 -
conv2d 128 3 -
dropout - - 0.5
dense 64 - -
dropout - - 0.5

4/4/2022, 6:16:33 PM

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