Merge a543412 into 0f176f8

.github/workflows/install.yml

@@ -12,7 +12,7 @@ jobs:
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        python-version: [ 3.6, 3.7 ]
+        python-version: [3.7]
 
     steps:
       - name: Set up Python ${{ matrix.python-version }}

.github/workflows/tests.yml

@@ -12,8 +12,8 @@ jobs:
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        python-version: [3.6, 3.7]
-        tensorflow-version: [2.1, 2.2]
+        python-version: [3.7]
+        tensorflow-version: [2.9]
     steps:
       - uses: actions/checkout@v2
       - name: Set up Python ${{ matrix.python-version }}
@@ -61,8 +61,8 @@ jobs:
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        python-version: [3.6, 3.7]
-        tensorflow-version: [2.1, 2.2]
+        python-version: [3.7]
+        tensorflow-version: [2.9]
     steps:
       - uses: actions/checkout@v2
       - name: Set up Python ${{ matrix.python-version }}
@@ -110,7 +110,7 @@ jobs:
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        python-version: [3.6, 3.7]
+        python-version: [3.7]
     steps:
       - uses: actions/checkout@v2
       - name: Set up Python ${{ matrix.python-version }}

README.md

@@ -95,7 +95,7 @@ The [notebooks](notebooks) section provides some examples. For instance, just ru
 ```python
 import tensorflow as tf
 import tensorflow_datasets as tfds
-from tensorflow.keras.models import Sequential
+from keras.models import Sequential
 
 from keras_fsl.models.encoders import BasicCNN
 from keras_fsl.layers import GramMatrix

keras_fsl/callbacks/binary_statistics.py

@@ -2,7 +2,7 @@
 import numpy as np
 import tensorflow as tf
 from matplotlib.backends.backend_agg import FigureCanvasAgg
-from tensorflow.keras.callbacks import Callback
+from keras.callbacks import Callback
 
 
 class BinaryStatistics(Callback):

keras_fsl/layers/centroids_matrix.py

@@ -1,5 +1,5 @@
 import tensorflow as tf
-from tensorflow.keras import activations
+from keras import activations
 
 from keras_fsl.layers.support_layer import SupportLayer
 

keras_fsl/layers/classification.py

@@ -1,5 +1,5 @@
 import tensorflow as tf
-from tensorflow.keras.layers import Layer
+from keras.layers import Layer
 
 from keras_fsl.losses import class_consistency_loss
 

keras_fsl/layers/slicing.py

@@ -1,5 +1,5 @@
-from tensorflow.keras import Sequential
-from tensorflow.keras.layers import Lambda, Flatten
+from keras import Sequential
+from keras.layers import Lambda, Flatten
 
 
 def CenterSlicing2D():

keras_fsl/layers/support_layer.py

@@ -1,5 +1,5 @@
 import tensorflow as tf
-from tensorflow.keras.layers import Layer
+from keras.layers import Layer
 
 from keras_fsl.models import head_models
 

keras_fsl/losses/gram_matrix_losses.py

@@ -13,7 +13,7 @@
 import tensorflow as tf
 import tensorflow.keras.backend as K
 import tensorflow_probability as tfp
-from tensorflow.keras.losses import Loss
+from keras.losses import Loss
 
 
 class MeanScoreClassificationLoss(Loss):

keras_fsl/metrics/gram_matrix_metrics.py

@@ -3,7 +3,7 @@
 instance). y_true should be one-hot encoded
 """
 import tensorflow as tf
-from tensorflow.keras import backend as K
+from keras import backend as K
 
 
 def classification_accuracy(ascending=False):

keras_fsl/models/encoders/basic_cnn.py

@@ -1,5 +1,5 @@
-from tensorflow.keras.layers import Conv2D, Dense, Dropout, Flatten, MaxPooling2D
-from tensorflow.keras.models import Sequential
+from keras.layers import Conv2D, Dense, Dropout, Flatten, MaxPooling2D
+from keras.models import Sequential
 
 
 def BasicCNN(input_shape, classes=None):

keras_fsl/models/encoders/darknet.py

@@ -1,8 +1,8 @@
 from functools import wraps
 
-from tensorflow.keras import Sequential, Input, Model
-from tensorflow.keras.layers import Conv2D, LeakyReLU, ZeroPadding2D, Add, MaxPooling2D, BatchNormalization
-from tensorflow.keras.regularizers import l2
+from keras import Sequential, Input, Model
+from keras.layers import Conv2D, LeakyReLU, ZeroPadding2D, Add, MaxPooling2D, BatchNormalization
+from keras.regularizers import l2
 
 
 @wraps(Conv2D)

keras_fsl/models/encoders/koch_net.py

@@ -1,7 +1,7 @@
-from tensorflow.keras import Sequential, Input
-from tensorflow.keras.layers import Conv2D, MaxPooling2D, Dense, Flatten
-from tensorflow.keras.initializers import RandomNormal
-from tensorflow.keras.regularizers import l2
+from keras import Sequential, Input
+from keras.layers import Conv2D, MaxPooling2D, Dense, Flatten
+from keras.initializers import RandomNormal
+from keras.regularizers import l2
 
 
 def conv_2d(*args, **kwargs):

keras_fsl/models/encoders/single_conv_2d.py

@@ -1,5 +1,5 @@
-from tensorflow.keras import Sequential
-from tensorflow.keras.layers import Conv2D, GlobalAveragePooling2D
+from keras import Sequential
+from keras.layers import Conv2D, GlobalAveragePooling2D
 
 
 def SingleConv2D(input_shape):

keras_fsl/models/encoders/vinyals_net.py

@@ -1,5 +1,5 @@
-from tensorflow.keras import Sequential
-from tensorflow.keras.layers import Conv2D, Input, BatchNormalization, Activation, MaxPooling2D, Flatten
+from keras import Sequential
+from keras.layers import Conv2D, Input, BatchNormalization, Activation, MaxPooling2D, Flatten
 
 
 def conv_block(*args, **kwargs):

keras_fsl/models/head_models/dense_sigmoid.py

@@ -1,6 +1,6 @@
 import tensorflow as tf
-from tensorflow.keras.layers import Dense, Input, Lambda
-from tensorflow.keras.models import Model
+from keras.layers import Dense, Input, Lambda
+from keras.models import Model
 
 
 def DenseSigmoid(input_shape, use_bias=True):

keras_fsl/models/head_models/learnt_norms.py

@@ -1,17 +1,17 @@
-import numpy as np
 import tensorflow as tf
-from tensorflow.keras import activations
-from tensorflow.keras.layers import (
+import numpy as np
+from keras import activations
+from keras.layers import (
     Concatenate,
     Conv2D,
     Dense,
     Flatten,
     Input,
     Reshape,
 )
-from tensorflow.keras.mixed_precision.experimental import global_policy
-from tensorflow.keras.models import Model
-from tensorflow.python.keras.layers import Activation
+from keras.mixed_precision import global_policy
+from keras.models import Model
+from keras.layers import Activation
 
 
 def LearntNorms(input_shape, use_bias=True, activation="sigmoid"):

keras_fsl/models/head_models/mixed_norms.py

@@ -1,6 +1,6 @@
 import tensorflow as tf
-from tensorflow.keras import activations
-from tensorflow.keras.layers import (
+from keras import activations
+from keras.layers import (
     Concatenate,
     Conv2D,
     Dense,
@@ -10,7 +10,7 @@
     Lambda,
     Reshape,
 )
-from tensorflow.keras.models import Model
+from keras.models import Model
 
 
 def MixedNorms(input_shape, norms=None, use_bias=True, activation="sigmoid"):

keras_fsl/models/head_models/tests/learnt_norms_test.py

@@ -1,7 +1,8 @@
 import numpy as np
 import tensorflow as tf
 from absl.testing import parameterized
-from tensorflow.keras.optimizers import RMSprop
+from keras import mixed_precision
+from keras.optimizers import RMSprop
 from tensorflow.python.keras.keras_parameterized import TestCase, run_all_keras_modes, run_with_all_model_types
 
 from keras_fsl.models.head_models import LearntNorms
@@ -34,16 +35,16 @@ def test_should_fit(self, input_shape):
         ("float64", "float64", "float64"),
     )
     def test_last_activation_fp32_in_mixed_precision(self, mixed_precision_policy, expected_last_layer_dtype_policy):
-        policy = tf.keras.mixed_precision.experimental.Policy(mixed_precision_policy)
-        tf.keras.mixed_precision.experimental.set_policy(policy)
+        policy = mixed_precision.Policy(mixed_precision_policy)
+        mixed_precision.set_global_policy(policy)
         learnt_norms = LearntNorms(input_shape=(10,))
 
         # Check dtype policy of internal non-input layers
         for layer in learnt_norms.layers[2:-1]:
             assert layer._dtype_policy.name == mixed_precision_policy
 
         # Check dtype policy of last layer always at least FP32
-        assert learnt_norms.layers[-1]._dtype_policy.name == expected_last_layer_dtype_policy
+        assert learnt_norms.layers[-1].dtype_policy.name == expected_last_layer_dtype_policy
 
 
 if __name__ == "__main__":

keras_fsl/models/siamese_nets.py

@@ -1,5 +1,5 @@
-from tensorflow.keras import Model
-from tensorflow.keras.layers import Input
+from keras import Model
+from keras.layers import Input
 
 from keras_fsl.models import encoders, head_models
 

keras_fsl/sequences/abstract_sequence.py

@@ -3,8 +3,8 @@
 import imgaug.augmenters as iaa
 import numpy as np
 from abc import ABCMeta
-from tensorflow.keras.preprocessing.image import img_to_array, load_img
-from tensorflow.keras.utils import Sequence
+from keras.preprocessing.image import img_to_array, load_img
+from keras.utils import Sequence
 
 
 class AbstractSequence(Sequence, metaclass=ABCMeta):

keras_fsl/sequences/prediction/pairs/product_sequence.py

@@ -1,7 +1,7 @@
 import math
 
 import pandas as pd
-from tensorflow.keras.utils import Sequence
+from keras.utils import Sequence
 
 
 class ProductSequence(Sequence):

keras_fsl/sequences/training/pairs/random_product_sequence.py

@@ -1,7 +1,7 @@
 import math
 
 import pandas as pd
-from tensorflow.keras.utils import Sequence
+from keras.utils import Sequence
 
 
 class RandomProductSequence(Sequence):

notebooks/batch_gram_matrix_training.py

@@ -4,14 +4,14 @@
 import click
 import pandas as pd
 import tensorflow as tf
-from tensorflow.keras import applications as keras_applications
-from tensorflow.keras.callbacks import (
+from keras import applications as keras_applications
+from keras.callbacks import (
     ModelCheckpoint,
     ReduceLROnPlateau,
     TensorBoard,
 )
-from tensorflow.keras.models import Sequential
-from tensorflow.keras.optimizers import Adam
+from keras.models import Sequential
+from keras.optimizers import Adam
 
 from keras_fsl.dataframe.operators import ToKShotDataset
 from keras_fsl.layers import Classification, GramMatrix
@@ -22,8 +22,8 @@
 #%% Toggle some config if required
 # tf.config.experimental_run_functions_eagerly(True)
 # tf.config.optimizer.set_jit(True)
-# policy = tf.keras.mixed_precision.experimental.Policy("mixed_float16")
-# tf.keras.mixed_precision.experimental.set_policy(policy)
+# policy = tf.keras.mixed_precision.Policy("mixed_float16")
+# tf.keras.mixed_precision.set_policy(policy)
 
 
 #%% CLI args

notebooks/benchmark_caching_performance.ipynb

@@ -1897,8 +1897,8 @@
         "import tensorflow_datasets as tfds\n",
         "from functools import partial\n",
         "from gpumonitor.callbacks.tf import TFGpuMonitorCallback\n",
-        "from tensorflow.keras.layers import Conv2D, Dropout, Flatten, GlobalMaxPooling2D, Input, MaxPooling2D\n",
-        "from tensorflow.keras.models import Sequential\n",
+        "from keras.layers import Conv2D, Dropout, Flatten, GlobalMaxPooling2D, Input, MaxPooling2D\n",
+        "from keras.models import Sequential\n",
         "\n",
         "from keras_fsl.utils.datasets import assign, cache, cache_with_tf_record, read_decode_and_crop_jpeg, transform"
       ],

notebooks/build_siamese_model_for_serving.py

@@ -1,6 +1,6 @@
 # flake8: noqa: E265
 import tensorflow as tf
-from tensorflow.keras.models import load_model
+from keras.models import load_model
 
 #%% Load siamese nets
 classifier = load_model("siamese_nets_classifier/1")

notebooks/centroids_similarity_training.py

@@ -3,15 +3,15 @@
 import click
 import pandas as pd
 import tensorflow as tf
-from tensorflow.keras import applications as keras_applications
-from tensorflow.keras.callbacks import (
+from keras import applications as keras_applications
+from keras.callbacks import (
     ModelCheckpoint,
     ReduceLROnPlateau,
     TensorBoard,
 )
-from tensorflow.keras.layers import Input
-from tensorflow.keras.models import Model
-from tensorflow.keras.optimizers import Adam
+from keras.layers import Input
+from keras.models import Model
+from keras.optimizers import Adam
 
 from keras_fsl.dataframe.operators import ToKShotDataset
 from keras_fsl.layers import CentroidsMatrix
@@ -22,8 +22,8 @@
 #%% Toggle some config if required
 # tf.config.experimental_run_functions_eagerly(True)
 # tf.config.optimizer.set_jit(True)
-# policy = tf.keras.mixed_precision.experimental.Policy("mixed_float16")
-# tf.keras.mixed_precision.experimental.set_policy(policy)
+# policy = tf.keras.mixed_precision.Policy("mixed_float16")
+# tf.keras.mixed_precision.set_policy(policy)
 
 
 #%% CLI args

notebooks/evaluate.py

@@ -3,8 +3,8 @@
 import numpy as np
 import pandas as pd
 import tensorflow as tf
-from tensorflow.keras.models import load_model
-from tensorflow.keras import metrics
+from keras.models import load_model
+from keras import metrics
 
 #%% Init
 k_shot = 4

notebooks/omniglot/basic_siamese_nets.py

@@ -4,9 +4,9 @@
 
 import imgaug.augmenters as iaa
 import pandas as pd
-from tensorflow.keras.callbacks import ModelCheckpoint, TensorBoard, ReduceLROnPlateau, EarlyStopping
-from tensorflow.keras.optimizer_v2.adam import Adam
-from tensorflow.keras.saving import load_model
+from keras.callbacks import ModelCheckpoint, TensorBoard, ReduceLROnPlateau, EarlyStopping
+from keras.optimizer_v2.adam import Adam
+from keras.saving import load_model
 
 from keras_fsl.datasets import omniglot
 from keras_fsl.models import SiameseNets

notebooks/siamese_nets_training.py

@@ -8,10 +8,10 @@
 import pandas as pd
 import tensorflow as tf
 import yaml
-from tensorflow.keras.models import load_model
-from tensorflow.keras import applications as keras_applications
-from tensorflow.keras.callbacks import ModelCheckpoint, ReduceLROnPlateau, TensorBoard
-from tensorflow.keras.optimizers import Adam
+from keras.models import load_model
+from keras import applications as keras_applications
+from keras.callbacks import ModelCheckpoint, ReduceLROnPlateau, TensorBoard
+from keras.optimizers import Adam
 
 from keras_fsl.models import SiameseNets
 from keras_fsl.sequences import training