Edge Impulse Experts / Solar Edge AI Forest Watcher for Logging and Poaching Public
Target: Cortex-M4F 80MHz 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

Run #1

Cortex-M4F 80MHz

100 ms

128 kB

1024 kB

Filters

Status

DSP type

Model type

View

Data set

Variant

Sort

General

F1-score

Precision

Recall

Run #2 (Sep 20 2025, 11:17:36, finished) Run #1 (Sep 20 2025, 11:00:13, failed)
68%
spectr-conv2d-01a
PERFORMANCE
LATENCY
337 ms of 100 ms
Exceeds target by 237 ms
RAM
89 kB of 128 kB
ROM
74 kB of 1024 kB
DSP NN Unused
TIME-SERIES INPUT

5000 ms |
5000 ms |
Enabled

SPECTROGRAM

0.075 | 0.075 | -72

ACCURACY (OVERALL)
CLASSIFICATION

0.005 | 100 | 68%

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

0 9/20/2025, 11:11:19 AM

52%
spectr-conv2d-26b
PERFORMANCE
LATENCY
873 ms of 100 ms
Exceeds target by 773 ms
RAM
129 kB of 128 kB
Exceeds target by 1 kB
ROM
84 kB of 1024 kB
DSP NN Unused
TIME-SERIES INPUT

5000 ms |
5000 ms |
Enabled

SPECTROGRAM

0.05 | 0.025 | -52

ACCURACY (OVERALL)
CLASSIFICATION

0.005 | 100 | 52%

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

0 9/20/2025, 11:11:38 AM

spectr-conv2d-063
PERFORMANCE
LATENCY
100 ms
RAM
128 kB
ROM
1024 kB
Unused
TIME-SERIES INPUT

5000 ms |
5000 ms |
Enabled

SPECTROGRAM

0.05 | 0.025 | -72

ACCURACY (KERAS)
CLASSIFICATION

0.005 | 100

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

0

spectr-conv2d-f9e
PERFORMANCE
LATENCY
100 ms
RAM
128 kB
ROM
1024 kB
Unused
TIME-SERIES INPUT

5000 ms |
5000 ms |
Enabled

SPECTROGRAM

0.05 | 0.025 | -32

ACCURACY (KERAS)
CLASSIFICATION

0.005 | 100

Type Filters Kernel Rate
Data augmentation
conv2d 16 3 -
conv2d 32 3 -
conv2d 64 3 -
conv2d 128 3 -
dropout - - 0.25

0

spectr-conv2d-d0b
PERFORMANCE
LATENCY
100 ms
RAM
128 kB
ROM
1024 kB
Unused
TIME-SERIES INPUT

5000 ms |
5000 ms |
Enabled

SPECTROGRAM

0.075 | 0.0375 | -52

ACCURACY (KERAS)
CLASSIFICATION

0.005 | 100

Type Filters Kernel Rate
conv2d 32 3 -
conv2d 64 3 -
conv2d 128 3 -
conv2d 256 3 -
dropout - - 0.25

0

spectr-conv2d-bbf
PERFORMANCE
LATENCY
100 ms
RAM
128 kB
ROM
1024 kB
Unused
TIME-SERIES INPUT

5000 ms |
5000 ms |
Enabled

SPECTROGRAM

0.075 | 0.0375 | -72

ACCURACY (KERAS)
CLASSIFICATION

0.005 | 100

Type Filters Kernel Rate
Data augmentation
conv2d 16 3 -
conv2d 32 3 -
dropout - - 0.25

0

spectr-conv2d-ff9
PERFORMANCE
LATENCY
100 ms
RAM
128 kB
ROM
1024 kB
Unused
TIME-SERIES INPUT

5000 ms |
5000 ms |
Enabled

SPECTROGRAM

0.05 | 0.025 | -52

ACCURACY (KERAS)
CLASSIFICATION

0.005 | 100

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

0

spectr-conv2d-d23
PERFORMANCE
LATENCY
100 ms
RAM
128 kB
ROM
1024 kB
Unused
TIME-SERIES INPUT

5000 ms |
5000 ms |
Enabled

SPECTROGRAM

0.05 | 0.025 | -52

ACCURACY (KERAS)
CLASSIFICATION

0.005 | 100

Type Filters Kernel Rate
conv2d 32 3 -
conv2d 64 3 -
conv2d 128 3 -
conv2d 256 3 -
dropout - - 0.5

0

spectr-conv2d-fc2
PERFORMANCE
LATENCY
100 ms
RAM
128 kB
ROM
1024 kB
Unused
TIME-SERIES INPUT

5000 ms |
5000 ms |
Enabled

SPECTROGRAM

0.05 | 0.025 | -32

ACCURACY (KERAS)
CLASSIFICATION

0.005 | 100

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

0

spectr-conv2d-b40
PERFORMANCE
LATENCY
100 ms
RAM
128 kB
ROM
1024 kB
Unused
TIME-SERIES INPUT

5000 ms |
5000 ms |
Enabled

SPECTROGRAM

0.025 | 0.0125 | -52

ACCURACY (KERAS)
CLASSIFICATION

0.005 | 100

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

0

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