Brainchip / Image Classification project using BrainChip MetaTF and Akidanet models Public
Target: BrainChip AKD1000 or AKD1500 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

test

BrainChip AKD1000 or AKD1500

100 ms

524288 kB

16384 kB

Filters

Status

DSP type

Model type

View

Data set

Variant

Sort

General

F1-score

Precision

Recall

95%
rgb-conv2d-e63
PERFORMANCE
LATENCY
2 ms of 100 ms
RAM
4 kB of 524288 kB
ROM
62 kB of 16384 kB
Allocated NPs
4
Activity sparsity
91.88%
MACs
3340809
DSP NN Unused
IMAGE INPUT

160 |
160

IMAGE

RGB

ACCURACY (KERAS-AKIDA)
CLASSIFICATION - BRAINCHIP AKIDA™

0.0005 | 30 | 95%

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

5/9/2024, 2:25:44 AM

93%
grayscale-conv2d-742
PERFORMANCE
LATENCY
2 ms of 100 ms
RAM
4 kB of 524288 kB
ROM
42 kB of 16384 kB
Allocated NPs
6
Activity sparsity
85.26%
MACs
4230409
DSP NN Unused
IMAGE INPUT

160 |
160

IMAGE

Grayscale

ACCURACY (KERAS-AKIDA)
CLASSIFICATION - BRAINCHIP AKIDA™

0.0005 | 30 | 93%

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

5/9/2024, 2:25:51 AM

87%
grayscale-other-373
PERFORMANCE
LATENCY
2 ms of 100 ms
RAM
4 kB of 524288 kB
ROM
715 kB of 16384 kB
Allocated NPs
25
Activity sparsity
74.34%
MACs
181742607
DSP NN Unused
IMAGE INPUT

160 |
160

IMAGE

Grayscale

ACCURACY (KERAS-AKIDA-TRANSFER-IMAGE)
TRANSFER LEARNING (IMAGES) - BRAINCHIP AKIDA™

0.0005 | 20 | 87%

Type Filters Kernel Rate
transfer_akidanet_imagenet_224_a50 64 - 0.1

5/9/2024, 2:37:16 AM

36%
rgb-other-013
PERFORMANCE
LATENCY
2 ms of 100 ms
RAM
4 kB of 524288 kB
ROM
697 kB of 16384 kB
Allocated NPs
25
Activity sparsity
69.90%
MACs
178582593
DSP NN Unused
IMAGE INPUT

160 |
160

IMAGE

RGB

ACCURACY (KERAS-AKIDA-TRANSFER-IMAGE)
TRANSFER LEARNING (IMAGES) - BRAINCHIP AKIDA™

0.0005 | 20 | 36%

Type Filters Kernel Rate
transfer_akidanet_imagenet_160_a50 16 - 0.1

5/9/2024, 2:34:26 AM

18%
grayscale-other-d17
PERFORMANCE
LATENCY
2 ms of 100 ms
RAM
4 kB of 524288 kB
ROM
715 kB of 16384 kB
Allocated NPs
25
Activity sparsity
74.39%
MACs
176764137
DSP NN Unused
IMAGE INPUT

160 |
160

IMAGE

Grayscale

ACCURACY (KERAS-AKIDA-TRANSFER-IMAGE)
TRANSFER LEARNING (IMAGES) - BRAINCHIP AKIDA™

0.0005 | 20 | 18%

Type Filters Kernel Rate
transfer_akidanet_imagenet_224_a50 64 - 0.5

5/9/2024, 2:34:16 AM

rgb-other-5c5
PERFORMANCE
LATENCY
100 ms
RAM
524288 kB
ROM
16384 kB
Unused
IMAGE INPUT

224 |
undefined

IMAGE

RGB

ACCURACY (KERAS-AKIDA-TRANSFER-IMAGE)
TRANSFER LEARNING (IMAGES) - BRAINCHIP AKIDA™

0.0005 | 20

Type Filters Kernel Rate
transfer_akidanet_imagenet_160_a50 64 - 0.5

rgb-other-71b
PERFORMANCE
LATENCY
100 ms
RAM
524288 kB
ROM
16384 kB
Unused
IMAGE INPUT

224 |
undefined

IMAGE

RGB

ACCURACY (KERAS-AKIDA-TRANSFER-IMAGE)
TRANSFER LEARNING (IMAGES) - BRAINCHIP AKIDA™

0.0005 | 20

Type Filters Kernel Rate
transfer_akidanet_imagenet_160_a50 64 - 0.1

grayscale-other-6da
PERFORMANCE
LATENCY
100 ms
RAM
524288 kB
ROM
16384 kB
Unused
IMAGE INPUT

224 |
undefined

IMAGE

Grayscale

ACCURACY (KERAS-AKIDA-TRANSFER-IMAGE)
TRANSFER LEARNING (IMAGES) - BRAINCHIP AKIDA™

0.0005 | 20

Type Filters Kernel Rate
transfer_akidanet_imagenet_160_a50 64 - 0.5

rgb-other-9be
PERFORMANCE
LATENCY
100 ms
RAM
524288 kB
ROM
16384 kB
Unused
IMAGE INPUT

224 |
undefined

IMAGE

RGB

ACCURACY (KERAS-AKIDA-TRANSFER-IMAGE)
TRANSFER LEARNING (IMAGES) - BRAINCHIP AKIDA™

0.0005 | 20

Type Filters Kernel Rate
transfer_akidanet_imagenet_160_a50 16 - 0.5

rgb-other-036
PERFORMANCE
LATENCY
100 ms
RAM
524288 kB
ROM
16384 kB
Unused
IMAGE INPUT

224 |
undefined

IMAGE

RGB

ACCURACY (KERAS-AKIDA-TRANSFER-IMAGE)
TRANSFER LEARNING (IMAGES) - BRAINCHIP AKIDA™

0.0005 | 20

Type Filters Kernel Rate
transfer_akidanet_imagenet_160_a50 16 - 0.1

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