import tensorflow as tf
import tensorflow.keras as keras
from nodeconfeu_watch.reader import AccelerationReader
from nodeconfeu_watch.visual import plot_history, classification_report
from nodeconfeu_watch.convert import ExportModel
tf.random.set_seed(1)
dataset = AccelerationReader(
{"james": './data/james-v2'},
test_ratio=0.2,
validation_ratio=0.2,
classnames=['swiperight', 'swipeleft', 'upup', 'waggle', 'clap2'])
model = keras.Sequential()
model.add(
keras.Input(shape=(50, 1, 3),
name='acceleration',
dtype=dataset.train.x.dtype))
model.add(keras.layers.Conv2D(14, (5, 1), padding='valid', activation='relu'))
model.add(keras.layers.Dropout(0.2))
model.add(
keras.layers.Conv2D(len(dataset.classnames), (5, 1),
padding='same',
dilation_rate=2,
activation='relu'))
model.add(keras.layers.Dropout(0.1))
model.add(keras.layers.MaxPool2D((46, 1)))
model.add(keras.layers.Flatten())
model.add(keras.layers.Dense(len(dataset.classnames), use_bias=False))
import tensorflow as tf
import tensorflow.keras as keras
from nodeconfeu_watch.reader import AccelerationReader
dataset = AccelerationReader(
{
"james": ['./data/james-v2'],
"conor": ['./data/conor-v2']
},
test_ratio=0.2,
validation_ratio=0.2,
classnames=[
'swiperight', 'swipeleft', 'upup', 'waggle', 'clap2', 'random'
])
dataset.dataframe().to_csv('./exports/dataset.csv', index=False)
import sklearn.metrics
import tensorflow as tf
import tensorflow.keras as keras
from nodeconfeu_watch.reader import AccelerationReader
from nodeconfeu_watch.layer import MaskLastFeature, DirectionFeatures, MaskedConv
from nodeconfeu_watch.visual import plot_history, classification_report
tf.random.set_seed(0)
dataset = AccelerationReader(
{"gordon": './data/gordon-v1'},
test_ratio=0,
validation_ratio=0.25,
classnames=['nothing', 'clap2', 'upup', 'swiperight', 'swipeleft'],
input_shape='1d',
mask_dimention=True)
model = keras.Sequential()
model.add(keras.Input(shape=(50, 4), name='acceleration'))
model.add(MaskLastFeature())
model.add(DirectionFeatures())
model.add(MaskedConv(25, 1, padding='same'))
model.add(keras.layers.Dropout(0.1))
model.add(keras.layers.LayerNormalization(scale=False))
model.add(keras.layers.Activation('relu'))
model.add(MaskedConv(25, 3, padding='same'))
model.add(keras.layers.Dropout(0.1))
import tensorflow as tf
import tensorflow.keras as keras
from nodeconfeu_watch.reader import AccelerationReader
from nodeconfeu_watch.visual import plot_history, classification_report
from nodeconfeu_watch.convert import ExportModel
tf.random.set_seed(1)
dataset = AccelerationReader({
"james": ['./data/james-v2'],
"conor": ['./data/conor-v2'],
"extra": ['./data/extra-v2']
},
test_ratio=0.2, validation_ratio=0.2,
classnames=['swiperight', 'swipeleft', 'swipeup', 'upup', 'waggle', 'clap2', 'random'])
model = keras.Sequential()
model.add(keras.Input(shape=(50, 1, 3), name='acceleration', dtype=dataset.train.x.dtype))
model.add(keras.layers.Conv2D(14, (5, 1), padding='valid', activation='relu'))
model.add(keras.layers.Dropout(0.2))
model.add(keras.layers.Conv2D(len(dataset.classnames), (5, 1), padding='same', dilation_rate=2, activation='relu'))
model.add(keras.layers.Dropout(0.1))
model.add(keras.layers.MaxPool2D((46, 1)))
model.add(keras.layers.Flatten())
model.add(keras.layers.Dense(len(dataset.classnames), use_bias=False))
model.compile(optimizer=keras.optimizers.Adam(),
loss=keras.losses.SparseCategoricalCrossentropy(from_logits=True),
metrics=[keras.metrics.SparseCategoricalAccuracy()])