Note

Go to the end to download the full example code.

Segmentation tutorial

This example demonstrates image segmentation with an Akida-compatible model as illustrated through person segmentation using the Portrait128 dataset.

Using pre-trained models for quick runtime, this example shows the evolution of model performance for a trained TF-Keras floating-point model, a TF-Keras quantized and Quantization Aware Trained (QAT) model, and an Akida-converted model. Notice that the performance of the original TF-Keras floating-point model is maintained throughout the model conversion flow.

1. Load the dataset

import os
import numpy as np
from akida_models import fetch_file

# Download validation set from Brainchip data server, it contains 10% of the original dataset
data_path = fetch_file(fname="val.tar.gz",
                       origin="https://data.brainchip.com/dataset-mirror/portrait128/val.tar.gz",
                       cache_subdir=os.path.join("datasets", "portrait128"),
                       extract=True)

data_dir = os.path.join(os.path.dirname(data_path), "val")
x_val = np.load(os.path.join(data_dir, "val_img.npy"))
y_val = np.load(os.path.join(data_dir, "val_msk.npy")).astype('uint8')
batch_size = 32
steps = x_val.shape[0] // 32

# Visualize some data
import matplotlib.pyplot as plt

rng = np.random.default_rng()
id = rng.integers(0, x_val.shape[0] - 2)

fig, axs = plt.subplots(3, 3, constrained_layout=True)
for col in range(3):
    axs[0, col].imshow(x_val[id + col] / 255.)
    axs[0, col].axis('off')
    axs[1, col].imshow(1 - y_val[id + col], cmap='Greys')
    axs[1, col].axis('off')
    axs[2, col].imshow(x_val[id + col] / 255. * y_val[id + col])
    axs[2, col].axis('off')

fig.suptitle('Image, mask and masked image', fontsize=10)
plt.show()
Image, mask and masked image
Downloading data from https://data.brainchip.com/dataset-mirror/portrait128/val.tar.gz.

        0/267313385 [..............................] - ETA: 0s
   229376/267313385 [..............................] - ETA: 1:00
   819200/267313385 [..............................] - ETA: 33s 
  1409024/267313385 [..............................] - ETA: 28s
  1835008/267313385 [..............................] - ETA: 29s
  2129920/267313385 [..............................] - ETA: 31s
  2719744/267313385 [..............................] - ETA: 29s
  2965504/267313385 [..............................] - ETA: 31s
  3424256/267313385 [..............................] - ETA: 31s
  4014080/267313385 [..............................] - ETA: 29s
  4603904/267313385 [..............................] - ETA: 29s
  5193728/267313385 [..............................] - ETA: 28s
  5783552/267313385 [..............................] - ETA: 27s
  6373376/267313385 [..............................] - ETA: 26s
  6963200/267313385 [..............................] - ETA: 26s
  7553024/267313385 [..............................] - ETA: 26s
  8142848/267313385 [..............................] - ETA: 25s
  8732672/267313385 [..............................] - ETA: 25s
  9306112/267313385 [>.............................] - ETA: 25s
  9895936/267313385 [>.............................] - ETA: 25s
 10485760/267313385 [>.............................] - ETA: 24s
 11075584/267313385 [>.............................] - ETA: 24s
 11665408/267313385 [>.............................] - ETA: 24s
 12189696/267313385 [>.............................] - ETA: 24s
 12812288/267313385 [>.............................] - ETA: 24s
 13402112/267313385 [>.............................] - ETA: 23s
 13991936/267313385 [>.............................] - ETA: 23s
 14581760/267313385 [>.............................] - ETA: 23s
 15155200/267313385 [>.............................] - ETA: 23s
 15745024/267313385 [>.............................] - ETA: 23s
 16334848/267313385 [>.............................] - ETA: 23s
 16924672/267313385 [>.............................] - ETA: 23s
 17530880/267313385 [>.............................] - ETA: 23s
 18120704/267313385 [=>............................] - ETA: 22s
 18710528/267313385 [=>............................] - ETA: 22s
 19283968/267313385 [=>............................] - ETA: 22s
 19742720/267313385 [=>............................] - ETA: 22s
 20480000/267313385 [=>............................] - ETA: 22s
 21086208/267313385 [=>............................] - ETA: 22s
 21659648/267313385 [=>............................] - ETA: 22s
 22249472/267313385 [=>............................] - ETA: 22s
 22839296/267313385 [=>............................] - ETA: 22s
 23429120/267313385 [=>............................] - ETA: 22s
 24018944/267313385 [=>............................] - ETA: 22s
 24608768/267313385 [=>............................] - ETA: 21s
 25198592/267313385 [=>............................] - ETA: 21s
 25788416/267313385 [=>............................] - ETA: 21s
 26378240/267313385 [=>............................] - ETA: 21s
 26968064/267313385 [==>...........................] - ETA: 21s
 27557888/267313385 [==>...........................] - ETA: 21s
 28147712/267313385 [==>...........................] - ETA: 21s
 28737536/267313385 [==>...........................] - ETA: 21s
 29294592/267313385 [==>...........................] - ETA: 21s
 29343744/267313385 [==>...........................] - ETA: 22s
 30294016/267313385 [==>...........................] - ETA: 22s
 32055296/267313385 [==>...........................] - ETA: 21s
 32636928/267313385 [==>...........................] - ETA: 21s
 33226752/267313385 [==>...........................] - ETA: 21s
 33701888/267313385 [==>...........................] - ETA: 21s
 34291712/267313385 [==>...........................] - ETA: 20s
 34881536/267313385 [==>...........................] - ETA: 20s
 35471360/267313385 [==>...........................] - ETA: 20s
 36061184/267313385 [===>..........................] - ETA: 20s
 36651008/267313385 [===>..........................] - ETA: 20s
 37240832/267313385 [===>..........................] - ETA: 20s
 37830656/267313385 [===>..........................] - ETA: 20s
 38420480/267313385 [===>..........................] - ETA: 20s
 39010304/267313385 [===>..........................] - ETA: 20s
 39600128/267313385 [===>..........................] - ETA: 20s
 40189952/267313385 [===>..........................] - ETA: 20s
 40779776/267313385 [===>..........................] - ETA: 20s
 41369600/267313385 [===>..........................] - ETA: 20s
 41959424/267313385 [===>..........................] - ETA: 20s
 42549248/267313385 [===>..........................] - ETA: 20s
 43139072/267313385 [===>..........................] - ETA: 19s
 43728896/267313385 [===>..........................] - ETA: 19s
 44318720/267313385 [===>..........................] - ETA: 19s
 44908544/267313385 [====>.........................] - ETA: 19s
 45498368/267313385 [====>.........................] - ETA: 19s
 46088192/267313385 [====>.........................] - ETA: 19s
 46678016/267313385 [====>.........................] - ETA: 19s
 47267840/267313385 [====>.........................] - ETA: 19s
 47857664/267313385 [====>.........................] - ETA: 19s
 48447488/267313385 [====>.........................] - ETA: 19s
 49037312/267313385 [====>.........................] - ETA: 19s
 49627136/267313385 [====>.........................] - ETA: 19s
 50216960/267313385 [====>.........................] - ETA: 19s
 50806784/267313385 [====>.........................] - ETA: 19s
 51396608/267313385 [====>.........................] - ETA: 19s
 51986432/267313385 [====>.........................] - ETA: 19s
 52576256/267313385 [====>.........................] - ETA: 18s
 53133312/267313385 [====>.........................] - ETA: 18s
 53739520/267313385 [=====>........................] - ETA: 18s
 54329344/267313385 [=====>........................] - ETA: 18s
 54919168/267313385 [=====>........................] - ETA: 18s
 55508992/267313385 [=====>........................] - ETA: 18s
 56098816/267313385 [=====>........................] - ETA: 18s
 56688640/267313385 [=====>........................] - ETA: 18s
 57278464/267313385 [=====>........................] - ETA: 18s
 57868288/267313385 [=====>........................] - ETA: 18s
 58458112/267313385 [=====>........................] - ETA: 18s
 59015168/267313385 [=====>........................] - ETA: 18s
 59670528/267313385 [=====>........................] - ETA: 18s
 60260352/267313385 [=====>........................] - ETA: 18s
 60850176/267313385 [=====>........................] - ETA: 18s
 61440000/267313385 [=====>........................] - ETA: 18s
 62029824/267313385 [=====>........................] - ETA: 18s
 62619648/267313385 [======>.......................] - ETA: 18s
 63209472/267313385 [======>.......................] - ETA: 17s
 63799296/267313385 [======>.......................] - ETA: 17s
 64389120/267313385 [======>.......................] - ETA: 17s
 64978944/267313385 [======>.......................] - ETA: 17s
 65568768/267313385 [======>.......................] - ETA: 17s
 66158592/267313385 [======>.......................] - ETA: 17s
 66568192/267313385 [======>.......................] - ETA: 17s
 67158016/267313385 [======>.......................] - ETA: 17s
 67747840/267313385 [======>.......................] - ETA: 17s
 68337664/267313385 [======>.......................] - ETA: 17s
 68927488/267313385 [======>.......................] - ETA: 17s
 69517312/267313385 [======>.......................] - ETA: 17s
 70107136/267313385 [======>.......................] - ETA: 17s
 70680576/267313385 [======>.......................] - ETA: 17s
 71270400/267313385 [======>.......................] - ETA: 17s
 71860224/267313385 [=======>......................] - ETA: 17s
 72433664/267313385 [=======>......................] - ETA: 17s
 73039872/267313385 [=======>......................] - ETA: 17s
 73629696/267313385 [=======>......................] - ETA: 17s
 74219520/267313385 [=======>......................] - ETA: 16s
 74809344/267313385 [=======>......................] - ETA: 16s
 75431936/267313385 [=======>......................] - ETA: 16s
 76021760/267313385 [=======>......................] - ETA: 16s
 76611584/267313385 [=======>......................] - ETA: 16s
 77201408/267313385 [=======>......................] - ETA: 16s
 77791232/267313385 [=======>......................] - ETA: 16s
 78381056/267313385 [=======>......................] - ETA: 16s
 78970880/267313385 [=======>......................] - ETA: 16s
 79560704/267313385 [=======>......................] - ETA: 16s
 80150528/267313385 [=======>......................] - ETA: 16s
 80740352/267313385 [========>.....................] - ETA: 16s
 81330176/267313385 [========>.....................] - ETA: 16s
 81920000/267313385 [========>.....................] - ETA: 16s
 82509824/267313385 [========>.....................] - ETA: 16s
 83099648/267313385 [========>.....................] - ETA: 16s
 83673088/267313385 [========>.....................] - ETA: 16s
 84262912/267313385 [========>.....................] - ETA: 16s
 84852736/267313385 [========>.....................] - ETA: 15s
 85442560/267313385 [========>.....................] - ETA: 15s
 86032384/267313385 [========>.....................] - ETA: 15s
 86622208/267313385 [========>.....................] - ETA: 15s
 87212032/267313385 [========>.....................] - ETA: 15s
 87801856/267313385 [========>.....................] - ETA: 15s
 88391680/267313385 [========>.....................] - ETA: 15s
 88981504/267313385 [========>.....................] - ETA: 15s
 89571328/267313385 [=========>....................] - ETA: 15s
 90161152/267313385 [=========>....................] - ETA: 15s
 90750976/267313385 [=========>....................] - ETA: 15s
 91340800/267313385 [=========>....................] - ETA: 15s
 91930624/267313385 [=========>....................] - ETA: 15s
 92438528/267313385 [=========>....................] - ETA: 15s
 92454912/267313385 [=========>....................] - ETA: 15s
 94568448/267313385 [=========>....................] - ETA: 15s
 95125504/267313385 [=========>....................] - ETA: 15s
 95682560/267313385 [=========>....................] - ETA: 15s
 96239616/267313385 [=========>....................] - ETA: 14s
 96698368/267313385 [=========>....................] - ETA: 14s
 97255424/267313385 [=========>....................] - ETA: 14s
 97796096/267313385 [=========>....................] - ETA: 14s
 98353152/267313385 [==========>...................] - ETA: 14s
 98910208/267313385 [==========>...................] - ETA: 14s
 99418112/267313385 [==========>...................] - ETA: 14s
 99991552/267313385 [==========>...................] - ETA: 14s
100532224/267313385 [==========>...................] - ETA: 14s
101089280/267313385 [==========>...................] - ETA: 14s
101662720/267313385 [==========>...................] - ETA: 14s
102227968/267313385 [==========>...................] - ETA: 14s
102793216/267313385 [==========>...................] - ETA: 14s
103350272/267313385 [==========>...................] - ETA: 14s
103874560/267313385 [==========>...................] - ETA: 14s
104497152/267313385 [==========>...................] - ETA: 14s
105054208/267313385 [==========>...................] - ETA: 14s
105611264/267313385 [==========>...................] - ETA: 14s
106168320/267313385 [==========>...................] - ETA: 14s
106725376/267313385 [==========>...................] - ETA: 14s
107282432/267313385 [===========>..................] - ETA: 14s
107839488/267313385 [===========>..................] - ETA: 14s
108396544/267313385 [===========>..................] - ETA: 14s
108953600/267313385 [===========>..................] - ETA: 13s
109494272/267313385 [===========>..................] - ETA: 13s
109953024/267313385 [===========>..................] - ETA: 13s
110575616/267313385 [===========>..................] - ETA: 13s
111165440/267313385 [===========>..................] - ETA: 13s
111755264/267313385 [===========>..................] - ETA: 13s
112345088/267313385 [===========>..................] - ETA: 13s
112951296/267313385 [===========>..................] - ETA: 13s
113541120/267313385 [===========>..................] - ETA: 13s
114147328/267313385 [===========>..................] - ETA: 13s
114737152/267313385 [===========>..................] - ETA: 13s
115326976/267313385 [===========>..................] - ETA: 13s
115916800/267313385 [============>.................] - ETA: 13s
116506624/267313385 [============>.................] - ETA: 13s
117096448/267313385 [============>.................] - ETA: 13s
117686272/267313385 [============>.................] - ETA: 13s
118276096/267313385 [============>.................] - ETA: 13s
118865920/267313385 [============>.................] - ETA: 13s
119455744/267313385 [============>.................] - ETA: 13s
120045568/267313385 [============>.................] - ETA: 12s
120553472/267313385 [============>.................] - ETA: 12s
121208832/267313385 [============>.................] - ETA: 12s
121798656/267313385 [============>.................] - ETA: 12s
122388480/267313385 [============>.................] - ETA: 12s
122978304/267313385 [============>.................] - ETA: 12s
123568128/267313385 [============>.................] - ETA: 12s
124141568/267313385 [============>.................] - ETA: 12s
124731392/267313385 [============>.................] - ETA: 12s
125321216/267313385 [=============>................] - ETA: 12s
125894656/267313385 [=============>................] - ETA: 12s
126451712/267313385 [=============>................] - ETA: 12s
126894080/267313385 [=============>................] - ETA: 12s
127614976/267313385 [=============>................] - ETA: 12s
128172032/267313385 [=============>................] - ETA: 12s
128729088/267313385 [=============>................] - ETA: 12s
129302528/267313385 [=============>................] - ETA: 12s
129908736/267313385 [=============>................] - ETA: 12s
130498560/267313385 [=============>................] - ETA: 12s
131088384/267313385 [=============>................] - ETA: 11s
131678208/267313385 [=============>................] - ETA: 11s
132235264/267313385 [=============>................] - ETA: 11s
132874240/267313385 [=============>................] - ETA: 11s
133464064/267313385 [=============>................] - ETA: 11s
134037504/267313385 [==============>...............] - ETA: 11s
134643712/267313385 [==============>...............] - ETA: 11s
135217152/267313385 [==============>...............] - ETA: 11s
135839744/267313385 [==============>...............] - ETA: 11s
136429568/267313385 [==============>...............] - ETA: 11s
137019392/267313385 [==============>...............] - ETA: 11s
137609216/267313385 [==============>...............] - ETA: 11s
138199040/267313385 [==============>...............] - ETA: 11s
138788864/267313385 [==============>...............] - ETA: 11s
139378688/267313385 [==============>...............] - ETA: 11s
139968512/267313385 [==============>...............] - ETA: 11s
140558336/267313385 [==============>...............] - ETA: 11s
141148160/267313385 [==============>...............] - ETA: 11s
141737984/267313385 [==============>...............] - ETA: 11s
142327808/267313385 [==============>...............] - ETA: 10s
142917632/267313385 [===============>..............] - ETA: 10s
143507456/267313385 [===============>..............] - ETA: 10s
144097280/267313385 [===============>..............] - ETA: 10s
144687104/267313385 [===============>..............] - ETA: 10s
145276928/267313385 [===============>..............] - ETA: 10s
145866752/267313385 [===============>..............] - ETA: 10s
146456576/267313385 [===============>..............] - ETA: 10s
147046400/267313385 [===============>..............] - ETA: 10s
147374080/267313385 [===============>..............] - ETA: 10s
147472384/267313385 [===============>..............] - ETA: 10s
148340736/267313385 [===============>..............] - ETA: 10s
149979136/267313385 [===============>..............] - ETA: 10s
150536192/267313385 [===============>..............] - ETA: 10s
151158784/267313385 [===============>..............] - ETA: 10s
151748608/267313385 [================>.............] - ETA: 10s
152338432/267313385 [================>.............] - ETA: 10s
152928256/267313385 [================>.............] - ETA: 10s
153518080/267313385 [================>.............] - ETA: 9s 
154091520/267313385 [================>.............] - ETA: 9s
154681344/267313385 [================>.............] - ETA: 9s
155271168/267313385 [================>.............] - ETA: 9s
155860992/267313385 [================>.............] - ETA: 9s
156450816/267313385 [================>.............] - ETA: 9s
157040640/267313385 [================>.............] - ETA: 9s
157630464/267313385 [================>.............] - ETA: 9s
158056448/267313385 [================>.............] - ETA: 9s
158646272/267313385 [================>.............] - ETA: 9s
159236096/267313385 [================>.............] - ETA: 9s
159825920/267313385 [================>.............] - ETA: 9s
160415744/267313385 [=================>............] - ETA: 9s
161005568/267313385 [=================>............] - ETA: 9s
161595392/267313385 [=================>............] - ETA: 9s
162185216/267313385 [=================>............] - ETA: 9s
162775040/267313385 [=================>............] - ETA: 9s
163364864/267313385 [=================>............] - ETA: 9s
163938304/267313385 [=================>............] - ETA: 9s
164528128/267313385 [=================>............] - ETA: 9s
165117952/267313385 [=================>............] - ETA: 8s
165707776/267313385 [=================>............] - ETA: 8s
166297600/267313385 [=================>............] - ETA: 8s
166887424/267313385 [=================>............] - ETA: 8s
167477248/267313385 [=================>............] - ETA: 8s
168067072/267313385 [=================>............] - ETA: 8s
168656896/267313385 [=================>............] - ETA: 8s
169246720/267313385 [=================>............] - ETA: 8s
169803776/267313385 [==================>...........] - ETA: 8s
170426368/267313385 [==================>...........] - ETA: 8s
171016192/267313385 [==================>...........] - ETA: 8s
171606016/267313385 [==================>...........] - ETA: 8s
172195840/267313385 [==================>...........] - ETA: 8s
172785664/267313385 [==================>...........] - ETA: 8s
173375488/267313385 [==================>...........] - ETA: 8s
173965312/267313385 [==================>...........] - ETA: 8s
174555136/267313385 [==================>...........] - ETA: 8s
175144960/267313385 [==================>...........] - ETA: 8s
175734784/267313385 [==================>...........] - ETA: 8s
176324608/267313385 [==================>...........] - ETA: 7s
176914432/267313385 [==================>...........] - ETA: 7s
177504256/267313385 [==================>...........] - ETA: 7s
177913856/267313385 [==================>...........] - ETA: 7s
178192384/267313385 [==================>...........] - ETA: 7s
178782208/267313385 [===================>..........] - ETA: 7s
179388416/267313385 [===================>..........] - ETA: 7s
179978240/267313385 [===================>..........] - ETA: 7s
180568064/267313385 [===================>..........] - ETA: 7s
181059584/267313385 [===================>..........] - ETA: 7s
181469184/267313385 [===================>..........] - ETA: 7s
182067200/267313385 [===================>..........] - ETA: 7s
182640640/267313385 [===================>..........] - ETA: 7s
183205888/267313385 [===================>..........] - ETA: 7s
183812096/267313385 [===================>..........] - ETA: 7s
184401920/267313385 [===================>..........] - ETA: 7s
184991744/267313385 [===================>..........] - ETA: 7s
185581568/267313385 [===================>..........] - ETA: 7s
186171392/267313385 [===================>..........] - ETA: 7s
186761216/267313385 [===================>..........] - ETA: 7s
187351040/267313385 [====================>.........] - ETA: 7s
187809792/267313385 [====================>.........] - ETA: 6s
188383232/267313385 [====================>.........] - ETA: 6s
188973056/267313385 [====================>.........] - ETA: 6s
189382656/267313385 [====================>.........] - ETA: 6s
189825024/267313385 [====================>.........] - ETA: 6s
190414848/267313385 [====================>.........] - ETA: 6s
191004672/267313385 [====================>.........] - ETA: 6s
191594496/267313385 [====================>.........] - ETA: 6s
192167936/267313385 [====================>.........] - ETA: 6s
192724992/267313385 [====================>.........] - ETA: 6s
193298432/267313385 [====================>.........] - ETA: 6s
193888256/267313385 [====================>.........] - ETA: 6s
194478080/267313385 [====================>.........] - ETA: 6s
195051520/267313385 [====================>.........] - ETA: 6s
195624960/267313385 [====================>.........] - ETA: 6s
196198400/267313385 [=====================>........] - ETA: 6s
196788224/267313385 [=====================>........] - ETA: 6s
197312512/267313385 [=====================>........] - ETA: 6s
198017024/267313385 [=====================>........] - ETA: 6s
198590464/267313385 [=====================>........] - ETA: 6s
199147520/267313385 [=====================>........] - ETA: 6s
199770112/267313385 [=====================>........] - ETA: 5s
200343552/267313385 [=====================>........] - ETA: 5s
200949760/267313385 [=====================>........] - ETA: 5s
201555968/267313385 [=====================>........] - ETA: 5s
202145792/267313385 [=====================>........] - ETA: 5s
202588160/267313385 [=====================>........] - ETA: 5s
203603968/267313385 [=====================>........] - ETA: 5s
204193792/267313385 [=====================>........] - ETA: 5s
204783616/267313385 [=====================>........] - ETA: 5s
205373440/267313385 [======================>.......] - ETA: 5s
205963264/267313385 [======================>.......] - ETA: 5s
206553088/267313385 [======================>.......] - ETA: 5s
207142912/267313385 [======================>.......] - ETA: 5s
207716352/267313385 [======================>.......] - ETA: 5s
208289792/267313385 [======================>.......] - ETA: 5s
208863232/267313385 [======================>.......] - ETA: 5s
209420288/267313385 [======================>.......] - ETA: 5s
209977344/267313385 [======================>.......] - ETA: 5s
210534400/267313385 [======================>.......] - ETA: 4s
211091456/267313385 [======================>.......] - ETA: 4s
211648512/267313385 [======================>.......] - ETA: 4s
212205568/267313385 [======================>.......] - ETA: 4s
212762624/267313385 [======================>.......] - ETA: 4s
213319680/267313385 [======================>.......] - ETA: 4s
213893120/267313385 [=======================>......] - ETA: 4s
214351872/267313385 [=======================>......] - ETA: 4s
214990848/267313385 [=======================>......] - ETA: 4s
215384064/267313385 [=======================>......] - ETA: 4s
216104960/267313385 [=======================>......] - ETA: 4s
216662016/267313385 [=======================>......] - ETA: 4s
217235456/267313385 [=======================>......] - ETA: 4s
217776128/267313385 [=======================>......] - ETA: 4s
218333184/267313385 [=======================>......] - ETA: 4s
218890240/267313385 [=======================>......] - ETA: 4s
219447296/267313385 [=======================>......] - ETA: 4s
220004352/267313385 [=======================>......] - ETA: 4s
220561408/267313385 [=======================>......] - ETA: 4s
221118464/267313385 [=======================>......] - ETA: 4s
221675520/267313385 [=======================>......] - ETA: 4s
222248960/267313385 [=======================>......] - ETA: 3s
222806016/267313385 [========================>.....] - ETA: 3s
223363072/267313385 [========================>.....] - ETA: 3s
223870976/267313385 [========================>.....] - ETA: 3s
224411648/267313385 [========================>.....] - ETA: 3s
224968704/267313385 [========================>.....] - ETA: 3s
225460224/267313385 [========================>.....] - ETA: 3s
225656832/267313385 [========================>.....] - ETA: 3s
227311616/267313385 [========================>.....] - ETA: 3s
227901440/267313385 [========================>.....] - ETA: 3s
228409344/267313385 [========================>.....] - ETA: 3s
228999168/267313385 [========================>.....] - ETA: 3s
229588992/267313385 [========================>.....] - ETA: 3s
230178816/267313385 [========================>.....] - ETA: 3s
230768640/267313385 [========================>.....] - ETA: 3s
231358464/267313385 [========================>.....] - ETA: 3s
231948288/267313385 [=========================>....] - ETA: 3s
232538112/267313385 [=========================>....] - ETA: 3s
233127936/267313385 [=========================>....] - ETA: 3s
233717760/267313385 [=========================>....] - ETA: 2s
234307584/267313385 [=========================>....] - ETA: 2s
234897408/267313385 [=========================>....] - ETA: 2s
235487232/267313385 [=========================>....] - ETA: 2s
236077056/267313385 [=========================>....] - ETA: 2s
236650496/267313385 [=========================>....] - ETA: 2s
237240320/267313385 [=========================>....] - ETA: 2s
237830144/267313385 [=========================>....] - ETA: 2s
238419968/267313385 [=========================>....] - ETA: 2s
239009792/267313385 [=========================>....] - ETA: 2s
239599616/267313385 [=========================>....] - ETA: 2s
240189440/267313385 [=========================>....] - ETA: 2s
240779264/267313385 [==========================>...] - ETA: 2s
241369088/267313385 [==========================>...] - ETA: 2s
241958912/267313385 [==========================>...] - ETA: 2s
242548736/267313385 [==========================>...] - ETA: 2s
243138560/267313385 [==========================>...] - ETA: 2s
243728384/267313385 [==========================>...] - ETA: 2s
244318208/267313385 [==========================>...] - ETA: 2s
244908032/267313385 [==========================>...] - ETA: 1s
245497856/267313385 [==========================>...] - ETA: 1s
246087680/267313385 [==========================>...] - ETA: 1s
246661120/267313385 [==========================>...] - ETA: 1s
247267328/267313385 [==========================>...] - ETA: 1s
247857152/267313385 [==========================>...] - ETA: 1s
248446976/267313385 [==========================>...] - ETA: 1s
248807424/267313385 [==========================>...] - ETA: 1s
249184256/267313385 [==========================>...] - ETA: 1s
249774080/267313385 [===========================>..] - ETA: 1s
250363904/267313385 [===========================>..] - ETA: 1s
250937344/267313385 [===========================>..] - ETA: 1s
251543552/267313385 [===========================>..] - ETA: 1s
252133376/267313385 [===========================>..] - ETA: 1s
252723200/267313385 [===========================>..] - ETA: 1s
253313024/267313385 [===========================>..] - ETA: 1s
253902848/267313385 [===========================>..] - ETA: 1s
254492672/267313385 [===========================>..] - ETA: 1s
255082496/267313385 [===========================>..] - ETA: 1s
255672320/267313385 [===========================>..] - ETA: 1s
256262144/267313385 [===========================>..] - ETA: 0s
256851968/267313385 [===========================>..] - ETA: 0s
257441792/267313385 [===========================>..] - ETA: 0s
258031616/267313385 [===========================>..] - ETA: 0s
258621440/267313385 [============================>.] - ETA: 0s
259211264/267313385 [============================>.] - ETA: 0s
259801088/267313385 [============================>.] - ETA: 0s
260390912/267313385 [============================>.] - ETA: 0s
260980736/267313385 [============================>.] - ETA: 0s
261570560/267313385 [============================>.] - ETA: 0s
262160384/267313385 [============================>.] - ETA: 0s
262750208/267313385 [============================>.] - ETA: 0s
263340032/267313385 [============================>.] - ETA: 0s
263929856/267313385 [============================>.] - ETA: 0s
264519680/267313385 [============================>.] - ETA: 0s
265109504/267313385 [============================>.] - ETA: 0s
265699328/267313385 [============================>.] - ETA: 0s
266289152/267313385 [============================>.] - ETA: 0s
266878976/267313385 [============================>.] - ETA: 0s
267313385/267313385 [==============================] - 24s 0us/step
Download complete.

2. Load a pre-trained native TF-Keras model

The model used in this example is AkidaUNet. It has an AkidaNet (0.5) backbone to extract features combined with a succession of separable transposed convolutional blocks to build an image segmentation map. A pre-trained floating-point TF-Keras model is downloaded to save training time.

Note

  • The “transposed” convolutional feature is new in Akida 2.0.

  • The “separable transposed” operation is realized through the combination of a QuantizeML custom DepthwiseConv2DTranspose layer with a standard pointwise convolution.

The performance of the model is evaluated using both pixel accuracy and Binary IoU. The pixel accuracy describes how well the model can predict the segmentation mask pixel by pixel and the Binary IoU takes into account how close the predicted mask is to the ground truth.

from akida_models.model_io import load_model

# Retrieve the model file from Brainchip data server
model_file = fetch_file(fname="akida_unet_portrait128.h5",
                        origin="https://data.brainchip.com/models/AkidaV2/akida_unet/akida_unet_portrait128.h5",
                        cache_subdir='models')

# Load the native TF-Keras pre-trained model
model_keras = load_model(model_file)
model_keras.summary()

Downloading data from https://data.brainchip.com/models/AkidaV2/akida_unet/akida_unet_portrait128.h5.

      0/4501976 [..............................] - ETA: 0s
 139264/4501976 [..............................] - ETA: 1s
 663552/4501976 [===>..........................] - ETA: 0s
1253376/4501976 [=======>......................] - ETA: 0s
1843200/4501976 [===========>..................] - ETA: 0s
2433024/4501976 [===============>..............] - ETA: 0s
3022848/4501976 [===================>..........] - ETA: 0s
3612672/4501976 [=======================>......] - ETA: 0s
4202496/4501976 [===========================>..] - ETA: 0s
4501976/4501976 [==============================] - 0s 0us/step
Download complete.
Model: "akida_unet"
_________________________________________________________________
 Layer (type)                Output Shape              Param #
=================================================================
 input (InputLayer)          [(None, 128, 128, 3)]     0

 rescaling (Rescaling)       (None, 128, 128, 3)       0

 conv_0 (Conv2D)             (None, 64, 64, 16)        432

 conv_0/BN (BatchNormalizat  (None, 64, 64, 16)        64
 ion)

 conv_0/relu (ReLU)          (None, 64, 64, 16)        0

 conv_1 (Conv2D)             (None, 64, 64, 32)        4608

 conv_1/BN (BatchNormalizat  (None, 64, 64, 32)        128
 ion)

 conv_1/relu (ReLU)          (None, 64, 64, 32)        0

 conv_2 (Conv2D)             (None, 32, 32, 64)        18432

 conv_2/BN (BatchNormalizat  (None, 32, 32, 64)        256
 ion)

 conv_2/relu (ReLU)          (None, 32, 32, 64)        0

 conv_3 (Conv2D)             (None, 32, 32, 64)        36864

 conv_3/BN (BatchNormalizat  (None, 32, 32, 64)        256
 ion)

 conv_3/relu (ReLU)          (None, 32, 32, 64)        0

 dw_separable_4 (DepthwiseC  (None, 16, 16, 64)        576
 onv2D)

 pw_separable_4 (Conv2D)     (None, 16, 16, 128)       8192

 pw_separable_4/BN (BatchNo  (None, 16, 16, 128)       512
 rmalization)

 pw_separable_4/relu (ReLU)  (None, 16, 16, 128)       0

 dw_separable_5 (DepthwiseC  (None, 16, 16, 128)       1152
 onv2D)

 pw_separable_5 (Conv2D)     (None, 16, 16, 128)       16384

 pw_separable_5/BN (BatchNo  (None, 16, 16, 128)       512
 rmalization)

 pw_separable_5/relu (ReLU)  (None, 16, 16, 128)       0

 dw_separable_6 (DepthwiseC  (None, 8, 8, 128)         1152
 onv2D)

 pw_separable_6 (Conv2D)     (None, 8, 8, 256)         32768

 pw_separable_6/BN (BatchNo  (None, 8, 8, 256)         1024
 rmalization)

 pw_separable_6/relu (ReLU)  (None, 8, 8, 256)         0

 dw_separable_7 (DepthwiseC  (None, 8, 8, 256)         2304
 onv2D)

 pw_separable_7 (Conv2D)     (None, 8, 8, 256)         65536

 pw_separable_7/BN (BatchNo  (None, 8, 8, 256)         1024
 rmalization)

 pw_separable_7/relu (ReLU)  (None, 8, 8, 256)         0

 dw_separable_8 (DepthwiseC  (None, 8, 8, 256)         2304
 onv2D)

 pw_separable_8 (Conv2D)     (None, 8, 8, 256)         65536

 pw_separable_8/BN (BatchNo  (None, 8, 8, 256)         1024
 rmalization)

 pw_separable_8/relu (ReLU)  (None, 8, 8, 256)         0

 dw_separable_9 (DepthwiseC  (None, 8, 8, 256)         2304
 onv2D)

 pw_separable_9 (Conv2D)     (None, 8, 8, 256)         65536

 pw_separable_9/BN (BatchNo  (None, 8, 8, 256)         1024
 rmalization)

 pw_separable_9/relu (ReLU)  (None, 8, 8, 256)         0

 dw_separable_10 (Depthwise  (None, 8, 8, 256)         2304
 Conv2D)

 pw_separable_10 (Conv2D)    (None, 8, 8, 256)         65536

 pw_separable_10/BN (BatchN  (None, 8, 8, 256)         1024
 ormalization)

 pw_separable_10/relu (ReLU  (None, 8, 8, 256)         0
 )

 dw_separable_11 (Depthwise  (None, 8, 8, 256)         2304
 Conv2D)

 pw_separable_11 (Conv2D)    (None, 8, 8, 256)         65536

 pw_separable_11/BN (BatchN  (None, 8, 8, 256)         1024
 ormalization)

 pw_separable_11/relu (ReLU  (None, 8, 8, 256)         0
 )

 dw_separable_12 (Depthwise  (None, 4, 4, 256)         2304
 Conv2D)

 pw_separable_12 (Conv2D)    (None, 4, 4, 512)         131072

 pw_separable_12/BN (BatchN  (None, 4, 4, 512)         2048
 ormalization)

 pw_separable_12/relu (ReLU  (None, 4, 4, 512)         0
 )

 dw_separable_13 (Depthwise  (None, 4, 4, 512)         4608
 Conv2D)

 pw_separable_13 (Conv2D)    (None, 4, 4, 512)         262144

 pw_separable_13/BN (BatchN  (None, 4, 4, 512)         2048
 ormalization)

 pw_separable_13/relu (ReLU  (None, 4, 4, 512)         0
 )

 dw_sepconv_t_0 (DepthwiseC  (None, 8, 8, 512)         5120
 onv2DTranspose)

 pw_sepconv_t_0 (Conv2D)     (None, 8, 8, 256)         131328

 pw_sepconv_t_0/BN (BatchNo  (None, 8, 8, 256)         1024
 rmalization)

 pw_sepconv_t_0/relu (ReLU)  (None, 8, 8, 256)         0

 dropout (Dropout)           (None, 8, 8, 256)         0

 dw_sepconv_t_1 (DepthwiseC  (None, 16, 16, 256)       2560
 onv2DTranspose)

 pw_sepconv_t_1 (Conv2D)     (None, 16, 16, 128)       32896

 pw_sepconv_t_1/BN (BatchNo  (None, 16, 16, 128)       512
 rmalization)

 pw_sepconv_t_1/relu (ReLU)  (None, 16, 16, 128)       0

 dropout_1 (Dropout)         (None, 16, 16, 128)       0

 dw_sepconv_t_2 (DepthwiseC  (None, 32, 32, 128)       1280
 onv2DTranspose)

 pw_sepconv_t_2 (Conv2D)     (None, 32, 32, 64)        8256

 pw_sepconv_t_2/BN (BatchNo  (None, 32, 32, 64)        256
 rmalization)

 pw_sepconv_t_2/relu (ReLU)  (None, 32, 32, 64)        0

 dropout_2 (Dropout)         (None, 32, 32, 64)        0

 dw_sepconv_t_3 (DepthwiseC  (None, 64, 64, 64)        640
 onv2DTranspose)

 pw_sepconv_t_3 (Conv2D)     (None, 64, 64, 32)        2080

 pw_sepconv_t_3/BN (BatchNo  (None, 64, 64, 32)        128
 rmalization)

 pw_sepconv_t_3/relu (ReLU)  (None, 64, 64, 32)        0

 dropout_3 (Dropout)         (None, 64, 64, 32)        0

 dw_sepconv_t_4 (DepthwiseC  (None, 128, 128, 32)      320
 onv2DTranspose)

 pw_sepconv_t_4 (Conv2D)     (None, 128, 128, 16)      528

 pw_sepconv_t_4/BN (BatchNo  (None, 128, 128, 16)      64
 rmalization)

 pw_sepconv_t_4/relu (ReLU)  (None, 128, 128, 16)      0

 dropout_4 (Dropout)         (None, 128, 128, 16)      0

 head (Conv2D)               (None, 128, 128, 1)       17

=================================================================
Total params: 1058865 (4.04 MB)
Trainable params: 1051889 (4.01 MB)
Non-trainable params: 6976 (27.25 KB)
_________________________________________________________________

from tf_keras.metrics import BinaryIoU

# Compile the native TF-Keras model (required to evaluate the metrics)
model_keras.compile(loss='binary_crossentropy', metrics=[BinaryIoU(), 'accuracy'])

# Check Keras model performance
_, biou, acc = model_keras.evaluate(x_val, y_val, steps=steps, verbose=0)

print(f"TF-Keras binary IoU / pixel accuracy: {biou:.4f} / {100*acc:.2f}%")

TF-Keras binary IoU / pixel accuracy: 0.9356 / 96.78%

3. Load a pre-trained quantized Keras model

The next step is to quantize and potentially perform Quantize Aware Training (QAT) on the TF-Keras model from the previous step. After the TF-Keras model is quantized to 8-bit for all weights and activations, QAT is used to maintain the performance of the quantized model. Again, a pre-trained model is downloaded to save runtime.

from akida_models import akida_unet_portrait128_pretrained

# Load the pre-trained quantized model
model_quantized_keras = akida_unet_portrait128_pretrained()
model_quantized_keras.summary()

Downloading data from https://data.brainchip.com/models/AkidaV2/akida_unet/akida_unet_portrait128_i8_w8_a8.h5.

      0/4527816 [..............................] - ETA: 0s
   8192/4527816 [..............................] - ETA: 1:04
  40960/4527816 [..............................] - ETA: 28s 
  98304/4527816 [..............................] - ETA: 16s
 188416/4527816 [>.............................] - ETA: 10s
 278528/4527816 [>.............................] - ETA: 8s 
 360448/4527816 [=>............................] - ETA: 7s
 458752/4527816 [==>...........................] - ETA: 6s
 557056/4527816 [==>...........................] - ETA: 5s
 671744/4527816 [===>..........................] - ETA: 5s
 786432/4527816 [====>.........................] - ETA: 4s
 901120/4527816 [====>.........................] - ETA: 4s
1032192/4527816 [=====>........................] - ETA: 4s
1146880/4527816 [======>.......................] - ETA: 3s
1294336/4527816 [=======>......................] - ETA: 3s
1425408/4527816 [========>.....................] - ETA: 3s
1589248/4527816 [=========>....................] - ETA: 3s
1736704/4527816 [==========>...................] - ETA: 2s
1900544/4527816 [===========>..................] - ETA: 2s
2080768/4527816 [============>.................] - ETA: 2s
2260992/4527816 [=============>................] - ETA: 2s
2457600/4527816 [===============>..............] - ETA: 1s
2654208/4527816 [================>.............] - ETA: 1s
2867200/4527816 [=================>............] - ETA: 1s
3096576/4527816 [===================>..........] - ETA: 1s
3325952/4527816 [=====================>........] - ETA: 0s
3571712/4527816 [======================>.......] - ETA: 0s
3817472/4527816 [========================>.....] - ETA: 0s
4079616/4527816 [==========================>...] - ETA: 0s
4096000/4527816 [==========================>...] - ETA: 0s
4358144/4527816 [===========================>..] - ETA: 0s
4374528/4527816 [===========================>..] - ETA: 0s
4527816/4527816 [==============================] - 3s 1us/step
Download complete.
Model: "akida_unet"
_________________________________________________________________
 Layer (type)                Output Shape              Param #
=================================================================
 input (InputLayer)          [(None, 128, 128, 3)]     0

 rescaling (QuantizedRescal  (None, 128, 128, 3)       0
 ing)

 conv_0 (QuantizedConv2D)    (None, 64, 64, 16)        448

 conv_0/relu (QuantizedReLU  (None, 64, 64, 16)        32
 )

 conv_1 (QuantizedConv2D)    (None, 64, 64, 32)        4640

 conv_1/relu (QuantizedReLU  (None, 64, 64, 32)        64
 )

 conv_2 (QuantizedConv2D)    (None, 32, 32, 64)        18496

 conv_2/relu (QuantizedReLU  (None, 32, 32, 64)        128
 )

 conv_3 (QuantizedConv2D)    (None, 32, 32, 64)        36928

 conv_3/relu (QuantizedReLU  (None, 32, 32, 64)        128
 )

 dw_separable_4 (QuantizedD  (None, 16, 16, 64)        704
 epthwiseConv2D)

 pw_separable_4 (QuantizedC  (None, 16, 16, 128)       8320
 onv2D)

 pw_separable_4/relu (Quant  (None, 16, 16, 128)       256
 izedReLU)

 dw_separable_5 (QuantizedD  (None, 16, 16, 128)       1408
 epthwiseConv2D)

 pw_separable_5 (QuantizedC  (None, 16, 16, 128)       16512
 onv2D)

 pw_separable_5/relu (Quant  (None, 16, 16, 128)       256
 izedReLU)

 dw_separable_6 (QuantizedD  (None, 8, 8, 128)         1408
 epthwiseConv2D)

 pw_separable_6 (QuantizedC  (None, 8, 8, 256)         33024
 onv2D)

 pw_separable_6/relu (Quant  (None, 8, 8, 256)         512
 izedReLU)

 dw_separable_7 (QuantizedD  (None, 8, 8, 256)         2816
 epthwiseConv2D)

 pw_separable_7 (QuantizedC  (None, 8, 8, 256)         65792
 onv2D)

 pw_separable_7/relu (Quant  (None, 8, 8, 256)         512
 izedReLU)

 dw_separable_8 (QuantizedD  (None, 8, 8, 256)         2816
 epthwiseConv2D)

 pw_separable_8 (QuantizedC  (None, 8, 8, 256)         65792
 onv2D)

 pw_separable_8/relu (Quant  (None, 8, 8, 256)         512
 izedReLU)

 dw_separable_9 (QuantizedD  (None, 8, 8, 256)         2816
 epthwiseConv2D)

 pw_separable_9 (QuantizedC  (None, 8, 8, 256)         65792
 onv2D)

 pw_separable_9/relu (Quant  (None, 8, 8, 256)         512
 izedReLU)

 dw_separable_10 (Quantized  (None, 8, 8, 256)         2816
 DepthwiseConv2D)

 pw_separable_10 (Quantized  (None, 8, 8, 256)         65792
 Conv2D)

 pw_separable_10/relu (Quan  (None, 8, 8, 256)         512
 tizedReLU)

 dw_separable_11 (Quantized  (None, 8, 8, 256)         2816
 DepthwiseConv2D)

 pw_separable_11 (Quantized  (None, 8, 8, 256)         65792
 Conv2D)

 pw_separable_11/relu (Quan  (None, 8, 8, 256)         512
 tizedReLU)

 dw_separable_12 (Quantized  (None, 4, 4, 256)         2816
 DepthwiseConv2D)

 pw_separable_12 (Quantized  (None, 4, 4, 512)         131584
 Conv2D)

 pw_separable_12/relu (Quan  (None, 4, 4, 512)         1024
 tizedReLU)

 dw_separable_13 (Quantized  (None, 4, 4, 512)         5632
 DepthwiseConv2D)

 pw_separable_13 (Quantized  (None, 4, 4, 512)         262656
 Conv2D)

 pw_separable_13/relu (Quan  (None, 4, 4, 512)         1024
 tizedReLU)

 dw_sepconv_t_0 (QuantizedD  (None, 8, 8, 512)         6144
 epthwiseConv2DTranspose)

 pw_sepconv_t_0 (QuantizedC  (None, 8, 8, 256)         131328
 onv2D)

 pw_sepconv_t_0/relu (Quant  (None, 8, 8, 256)         512
 izedReLU)

 dropout (QuantizedDropout)  (None, 8, 8, 256)         0

 dw_sepconv_t_1 (QuantizedD  (None, 16, 16, 256)       3072
 epthwiseConv2DTranspose)

 pw_sepconv_t_1 (QuantizedC  (None, 16, 16, 128)       32896
 onv2D)

 pw_sepconv_t_1/relu (Quant  (None, 16, 16, 128)       256
 izedReLU)

 dropout_1 (QuantizedDropou  (None, 16, 16, 128)       0
 t)

 dw_sepconv_t_2 (QuantizedD  (None, 32, 32, 128)       1536
 epthwiseConv2DTranspose)

 pw_sepconv_t_2 (QuantizedC  (None, 32, 32, 64)        8256
 onv2D)

 pw_sepconv_t_2/relu (Quant  (None, 32, 32, 64)        128
 izedReLU)

 dropout_2 (QuantizedDropou  (None, 32, 32, 64)        0
 t)

 dw_sepconv_t_3 (QuantizedD  (None, 64, 64, 64)        768
 epthwiseConv2DTranspose)

 pw_sepconv_t_3 (QuantizedC  (None, 64, 64, 32)        2080
 onv2D)

 pw_sepconv_t_3/relu (Quant  (None, 64, 64, 32)        64
 izedReLU)

 dropout_3 (QuantizedDropou  (None, 64, 64, 32)        0
 t)

 dw_sepconv_t_4 (QuantizedD  (None, 128, 128, 32)      384
 epthwiseConv2DTranspose)

 pw_sepconv_t_4 (QuantizedC  (None, 128, 128, 16)      528
 onv2D)

 pw_sepconv_t_4/relu (Quant  (None, 128, 128, 16)      32
 izedReLU)

 dropout_4 (QuantizedDropou  (None, 128, 128, 16)      0
 t)

 head (QuantizedConv2D)      (None, 128, 128, 1)       19

 dequantizer (Dequantizer)   (None, 128, 128, 1)       0

=================================================================
Total params: 1061603 (4.05 MB)
Trainable params: 1047905 (4.00 MB)
Non-trainable params: 13698 (53.51 KB)
_________________________________________________________________

# Compile the quantized TF-Keras model (required to evaluate the metrics)
model_quantized_keras.compile(loss='binary_crossentropy', metrics=[BinaryIoU(), 'accuracy'])

# Check Keras model performance
_, biou, acc = model_quantized_keras.evaluate(x_val, y_val, steps=steps, verbose=0)

print(f"TF-Keras quantized binary IoU / pixel accuracy: {biou:.4f} / {100*acc:.2f}%")

TF-Keras quantized binary IoU / pixel accuracy: 0.9344 / 96.77%

4. Conversion to Akida

Finally, the quantized TF-Keras model from the previous step is converted into an Akida model and its performance is evaluated. Note that the original performance of the TF-Keras floating-point model is maintained throughout the conversion process in this example.

from cnn2snn import convert

# Convert the model
model_akida = convert(model_quantized_keras)
model_akida.summary()

                  Model Summary
_________________________________________________
Input shape    Output shape   Sequences  Layers
=================================================
[128, 128, 3]  [128, 128, 1]  1          36
_________________________________________________

_____________________________________________________________________________
Layer (type)                               Output shape    Kernel shape

====================== SW/conv_0-dequantizer (Software) =====================

conv_0 (InputConv2D)                       [64, 64, 16]    (3, 3, 3, 16)
_____________________________________________________________________________
conv_1 (Conv2D)                            [64, 64, 32]    (3, 3, 16, 32)
_____________________________________________________________________________
conv_2 (Conv2D)                            [32, 32, 64]    (3, 3, 32, 64)
_____________________________________________________________________________
conv_3 (Conv2D)                            [32, 32, 64]    (3, 3, 64, 64)
_____________________________________________________________________________
dw_separable_4 (DepthwiseConv2D)           [16, 16, 64]    (3, 3, 64, 1)
_____________________________________________________________________________
pw_separable_4 (Conv2D)                    [16, 16, 128]   (1, 1, 64, 128)
_____________________________________________________________________________
dw_separable_5 (DepthwiseConv2D)           [16, 16, 128]   (3, 3, 128, 1)
_____________________________________________________________________________
pw_separable_5 (Conv2D)                    [16, 16, 128]   (1, 1, 128, 128)
_____________________________________________________________________________
dw_separable_6 (DepthwiseConv2D)           [8, 8, 128]     (3, 3, 128, 1)
_____________________________________________________________________________
pw_separable_6 (Conv2D)                    [8, 8, 256]     (1, 1, 128, 256)
_____________________________________________________________________________
dw_separable_7 (DepthwiseConv2D)           [8, 8, 256]     (3, 3, 256, 1)
_____________________________________________________________________________
pw_separable_7 (Conv2D)                    [8, 8, 256]     (1, 1, 256, 256)
_____________________________________________________________________________
dw_separable_8 (DepthwiseConv2D)           [8, 8, 256]     (3, 3, 256, 1)
_____________________________________________________________________________
pw_separable_8 (Conv2D)                    [8, 8, 256]     (1, 1, 256, 256)
_____________________________________________________________________________
dw_separable_9 (DepthwiseConv2D)           [8, 8, 256]     (3, 3, 256, 1)
_____________________________________________________________________________
pw_separable_9 (Conv2D)                    [8, 8, 256]     (1, 1, 256, 256)
_____________________________________________________________________________
dw_separable_10 (DepthwiseConv2D)          [8, 8, 256]     (3, 3, 256, 1)
_____________________________________________________________________________
pw_separable_10 (Conv2D)                   [8, 8, 256]     (1, 1, 256, 256)
_____________________________________________________________________________
dw_separable_11 (DepthwiseConv2D)          [8, 8, 256]     (3, 3, 256, 1)
_____________________________________________________________________________
pw_separable_11 (Conv2D)                   [8, 8, 256]     (1, 1, 256, 256)
_____________________________________________________________________________
dw_separable_12 (DepthwiseConv2D)          [4, 4, 256]     (3, 3, 256, 1)
_____________________________________________________________________________
pw_separable_12 (Conv2D)                   [4, 4, 512]     (1, 1, 256, 512)
_____________________________________________________________________________
dw_separable_13 (DepthwiseConv2D)          [4, 4, 512]     (3, 3, 512, 1)
_____________________________________________________________________________
pw_separable_13 (Conv2D)                   [4, 4, 512]     (1, 1, 512, 512)
_____________________________________________________________________________
dw_sepconv_t_0 (DepthwiseConv2DTranspose)  [8, 8, 512]     (3, 3, 512, 1)
_____________________________________________________________________________
pw_sepconv_t_0 (Conv2D)                    [8, 8, 256]     (1, 1, 512, 256)
_____________________________________________________________________________
dw_sepconv_t_1 (DepthwiseConv2DTranspose)  [16, 16, 256]   (3, 3, 256, 1)
_____________________________________________________________________________
pw_sepconv_t_1 (Conv2D)                    [16, 16, 128]   (1, 1, 256, 128)
_____________________________________________________________________________
dw_sepconv_t_2 (DepthwiseConv2DTranspose)  [32, 32, 128]   (3, 3, 128, 1)
_____________________________________________________________________________
pw_sepconv_t_2 (Conv2D)                    [32, 32, 64]    (1, 1, 128, 64)
_____________________________________________________________________________
dw_sepconv_t_3 (DepthwiseConv2DTranspose)  [64, 64, 64]    (3, 3, 64, 1)
_____________________________________________________________________________
pw_sepconv_t_3 (Conv2D)                    [64, 64, 32]    (1, 1, 64, 32)
_____________________________________________________________________________
dw_sepconv_t_4 (DepthwiseConv2DTranspose)  [128, 128, 32]  (3, 3, 32, 1)
_____________________________________________________________________________
pw_sepconv_t_4 (Conv2D)                    [128, 128, 16]  (1, 1, 32, 16)
_____________________________________________________________________________
head (Conv2D)                              [128, 128, 1]   (1, 1, 16, 1)
_____________________________________________________________________________
dequantizer (Dequantizer)                  [128, 128, 1]   N/A
_____________________________________________________________________________

import tf_keras as keras

# Check Akida model performance
labels, pots = None, None

for s in range(steps):
    batch = x_val[s * batch_size: (s + 1) * batch_size, :]
    label_batch = y_val[s * batch_size: (s + 1) * batch_size, :]
    pots_batch = model_akida.predict(batch.astype('uint8'))

    if labels is None:
        labels = label_batch
        pots = pots_batch
    else:
        labels = np.concatenate((labels, label_batch))
        pots = np.concatenate((pots, pots_batch))
preds = keras.activations.sigmoid(pots)

m_binary_iou = keras.metrics.BinaryIoU(target_class_ids=[0, 1], threshold=0.5)
m_binary_iou.update_state(labels, preds)
binary_iou = m_binary_iou.result().numpy()

m_accuracy = keras.metrics.Accuracy()
m_accuracy.update_state(labels, preds > 0.5)
accuracy = m_accuracy.result().numpy()
print(f"Akida binary IoU / pixel accuracy: {binary_iou:.4f} / {100*accuracy:.2f}%")

# For non-regression purpose
assert binary_iou > 0.9

Akida binary IoU / pixel accuracy: 0.9249 / 96.71%

5. Segment a single image

For visualization of the person segmentation performed by the Akida model, display a single image along with the segmentation produced by the original floating-point model and the ground truth segmentation.

import matplotlib.pyplot as plt

# Estimate age on a random single image and display TF-Keras and Akida outputs
sample = np.expand_dims(x_val[id, :], 0)
keras_out = model_keras(sample)
akida_out = keras.activations.sigmoid(model_akida.forward(sample.astype('uint8')))

fig, axs = plt.subplots(1, 3, constrained_layout=True)
axs[0].imshow(keras_out[0] * sample[0] / 255.)
axs[0].set_title('Keras segmentation', fontsize=10)
axs[0].axis('off')

axs[1].imshow(akida_out[0] * sample[0] / 255.)
axs[1].set_title('Akida segmentation', fontsize=10)
axs[1].axis('off')

axs[2].imshow(y_val[id] * sample[0] / 255.)
axs[2].set_title('Expected segmentation', fontsize=10)
axs[2].axis('off')

plt.show()
Keras segmentation, Akida segmentation, Expected segmentation

Total running time of the script: (2 minutes 18.853 seconds)

Gallery generated by Sphinx-Gallery