zoo.sota
QuickNet¶
larq_zoo.sota.QuickNet(
*,
input_shape=None,
input_tensor=None,
weights="imagenet",
include_top=True,
num_classes=1000
)
Instantiates the QuickNet architecture.
Optionally loads weights pre-trained on ImageNet.
Model Summary
+quicknet stats----------------------------------------------------------------------------------------------------+
| Layer Input prec. Outputs # 1-bit # 32-bit Memory 1-bit MACs 32-bit MACs |
| (bit) x 1 x 1 (kB) |
+------------------------------------------------------------------------------------------------------------------+
| input_1 - ((None, 224, 224, 3),) 0 0 0 ? ? |
| conv2d - (-1, 112, 112, 8) 0 216 0.84 0 2709504 |
| depthwise_conv2d - (-1, 56, 56, 64) 0 576 2.25 0 1806336 |
| batch_normalization - (-1, 56, 56, 64) 0 128 0.50 0 0 |
| quant_conv2d 1 (-1, 56, 56, 64) 36864 0 4.50 115605504 0 |
| batch_normalization_1 - (-1, 56, 56, 64) 0 128 0.50 0 0 |
| add - (-1, 56, 56, 64) 0 0 0 ? ? |
| quant_conv2d_1 1 (-1, 56, 56, 64) 36864 0 4.50 115605504 0 |
| batch_normalization_2 - (-1, 56, 56, 64) 0 128 0.50 0 0 |
| add_1 - (-1, 56, 56, 64) 0 0 0 ? ? |
| max_pooling2d - (-1, 28, 28, 64) 0 0 0 0 0 |
| quant_conv2d_2 1 (-1, 28, 28, 64) 36864 0 4.50 28901376 0 |
| batch_normalization_3 - (-1, 28, 28, 64) 0 128 0.50 0 0 |
| batch_normalization_4 - (-1, 28, 28, 64) 0 128 0.50 0 0 |
| add_2 - (-1, 28, 28, 64) 0 0 0 ? ? |
| concatenate - (-1, 28, 28, 128) 0 0 0 ? ? |
| quant_conv2d_3 1 (-1, 28, 28, 128) 147456 0 18.00 115605504 0 |
| batch_normalization_5 - (-1, 28, 28, 128) 0 256 1.00 0 0 |
| add_3 - (-1, 28, 28, 128) 0 0 0 ? ? |
| quant_conv2d_4 1 (-1, 28, 28, 128) 147456 0 18.00 115605504 0 |
| batch_normalization_6 - (-1, 28, 28, 128) 0 256 1.00 0 0 |
| add_4 - (-1, 28, 28, 128) 0 0 0 ? ? |
| max_pooling2d_1 - (-1, 14, 14, 128) 0 0 0 0 0 |
| quant_conv2d_5 1 (-1, 14, 14, 128) 147456 0 18.00 28901376 0 |
| batch_normalization_7 - (-1, 14, 14, 128) 0 256 1.00 0 0 |
| batch_normalization_8 - (-1, 14, 14, 128) 0 256 1.00 0 0 |
| add_5 - (-1, 14, 14, 128) 0 0 0 ? ? |
| concatenate_1 - (-1, 14, 14, 256) 0 0 0 ? ? |
| quant_conv2d_6 1 (-1, 14, 14, 256) 589824 0 72.00 115605504 0 |
| batch_normalization_9 - (-1, 14, 14, 256) 0 512 2.00 0 0 |
| add_6 - (-1, 14, 14, 256) 0 0 0 ? ? |
| quant_conv2d_7 1 (-1, 14, 14, 256) 589824 0 72.00 115605504 0 |
| batch_normalization_10 - (-1, 14, 14, 256) 0 512 2.00 0 0 |
| add_7 - (-1, 14, 14, 256) 0 0 0 ? ? |
| quant_conv2d_8 1 (-1, 14, 14, 256) 589824 0 72.00 115605504 0 |
| batch_normalization_11 - (-1, 14, 14, 256) 0 512 2.00 0 0 |
| add_8 - (-1, 14, 14, 256) 0 0 0 ? ? |
| max_pooling2d_2 - (-1, 7, 7, 256) 0 0 0 0 0 |
| quant_conv2d_9 1 (-1, 7, 7, 256) 589824 0 72.00 28901376 0 |
| batch_normalization_12 - (-1, 7, 7, 256) 0 512 2.00 0 0 |
| batch_normalization_13 - (-1, 7, 7, 256) 0 512 2.00 0 0 |
| add_9 - (-1, 7, 7, 256) 0 0 0 ? ? |
| concatenate_2 - (-1, 7, 7, 512) 0 0 0 ? ? |
| quant_conv2d_10 1 (-1, 7, 7, 512) 2359296 0 288.00 115605504 0 |
| batch_normalization_14 - (-1, 7, 7, 512) 0 1024 4.00 0 0 |
| add_10 - (-1, 7, 7, 512) 0 0 0 ? ? |
| quant_conv2d_11 1 (-1, 7, 7, 512) 2359296 0 288.00 115605504 0 |
| batch_normalization_15 - (-1, 7, 7, 512) 0 1024 4.00 0 0 |
| add_11 - (-1, 7, 7, 512) 0 0 0 ? ? |
| quant_conv2d_12 1 (-1, 7, 7, 512) 2359296 0 288.00 115605504 0 |
| batch_normalization_16 - (-1, 7, 7, 512) 0 1024 4.00 0 0 |
| add_12 - (-1, 7, 7, 512) 0 0 0 ? ? |
| activation - (-1, 7, 7, 512) 0 0 0 ? ? |
| average_pooling2d - (-1, 1, 1, 512) 0 0 0 0 0 |
| flatten - (-1, 512) 0 0 0 0 0 |
| dense - (-1, 1000) 0 513000 2003.91 0 512000 |
| activation_1 - (-1, 1000) 0 0 0 ? ? |
+------------------------------------------------------------------------------------------------------------------+
| Total 9990144 521088 3255.00 1242759168 5027840 |
+------------------------------------------------------------------------------------------------------------------+
+quicknet summary-----------------------------+
| Total params 10.5 M |
| Trainable params 10.5 M |
| Non-trainable params 7.3 k |
| Model size 3.18 MiB |
| Model size (8-bit FP weights) 1.69 MiB |
| Float-32 Equivalent 40.10 MiB |
| Compression Ratio of Memory 0.08 |
| Number of MACs 1.25 B |
| Ratio of MACs that are binarized 0.9960 |
+---------------------------------------------+
ImageNet Metrics
| Top-1 Accuracy | Top-5 Accuracy | Parameters | Memory |
|---|---|---|---|
| 58.6 % | 81.0 % | 10 518 528 | 3.21 MB |
Arguments
- input_shape
Optional[Sequence[Optional[int]]]: Optional shape tuple, to be specified if you would like to use a model with an input image resolution that is not (224, 224, 3). It should have exactly 3 inputs channels. - input_tensor
Optional[tf.Tensor]: optional Keras tensor (i.e. output oflayers.Input()) to use as image input for the model. - weights
Optional[str]: one ofNone(random initialization), "imagenet" (pre-training on ImageNet), or the path to the weights file to be loaded. - include_top
bool: whether to include the fully-connected layer at the top of the network. - num_classes
int: optional number of classes to classify images into, only to be specified ifinclude_topis True, and if noweightsargument is specified.
Returns
A Keras model instance.
Raises
- ValueError: in case of invalid argument for
weights, or invalid input shape.
QuickNetLarge¶
larq_zoo.sota.QuickNetLarge(
*,
input_shape=None,
input_tensor=None,
weights="imagenet",
include_top=True,
num_classes=1000
)
Instantiates the QuickNetLarge architecture.
Optionally loads weights pre-trained on ImageNet.
Model Summary
+quicknet_large stats-----------------------------------------------------------------------------------------------+
| Layer Input prec. Outputs # 1-bit # 32-bit Memory 1-bit MACs 32-bit MACs |
| (bit) x 1 x 1 (kB) |
+-------------------------------------------------------------------------------------------------------------------+
| input_1 - ((None, 224, 224, 3),) 0 0 0 ? ? |
| conv2d - (-1, 112, 112, 8) 0 216 0.84 0 2709504 |
| depthwise_conv2d - (-1, 56, 56, 64) 0 576 2.25 0 1806336 |
| batch_normalization - (-1, 56, 56, 64) 0 128 0.50 0 0 |
| quant_conv2d 1 (-1, 56, 56, 64) 36864 0 4.50 115605504 0 |
| batch_normalization_1 - (-1, 56, 56, 64) 0 128 0.50 0 0 |
| add - (-1, 56, 56, 64) 0 0 0 ? ? |
| quant_conv2d_1 1 (-1, 56, 56, 64) 36864 0 4.50 115605504 0 |
| batch_normalization_2 - (-1, 56, 56, 64) 0 128 0.50 0 0 |
| add_1 - (-1, 56, 56, 64) 0 0 0 ? ? |
| quant_conv2d_2 1 (-1, 56, 56, 64) 36864 0 4.50 115605504 0 |
| batch_normalization_3 - (-1, 56, 56, 64) 0 128 0.50 0 0 |
| add_2 - (-1, 56, 56, 64) 0 0 0 ? ? |
| quant_conv2d_3 1 (-1, 56, 56, 64) 36864 0 4.50 115605504 0 |
| batch_normalization_4 - (-1, 56, 56, 64) 0 128 0.50 0 0 |
| add_3 - (-1, 56, 56, 64) 0 0 0 ? ? |
| max_pooling2d - (-1, 28, 28, 64) 0 0 0 0 0 |
| quant_conv2d_4 1 (-1, 28, 28, 128) 73728 0 9.00 57802752 0 |
| conv2d_1 - (-1, 28, 28, 128) 0 8192 32.00 0 6422528 |
| batch_normalization_6 - (-1, 28, 28, 128) 0 256 1.00 0 0 |
| batch_normalization_5 - (-1, 28, 28, 128) 0 256 1.00 0 0 |
| add_4 - (-1, 28, 28, 128) 0 0 0 ? ? |
| quant_conv2d_5 1 (-1, 28, 28, 128) 147456 0 18.00 115605504 0 |
| batch_normalization_7 - (-1, 28, 28, 128) 0 256 1.00 0 0 |
| add_5 - (-1, 28, 28, 128) 0 0 0 ? ? |
| quant_conv2d_6 1 (-1, 28, 28, 128) 147456 0 18.00 115605504 0 |
| batch_normalization_8 - (-1, 28, 28, 128) 0 256 1.00 0 0 |
| add_6 - (-1, 28, 28, 128) 0 0 0 ? ? |
| quant_conv2d_7 1 (-1, 28, 28, 128) 147456 0 18.00 115605504 0 |
| batch_normalization_9 - (-1, 28, 28, 128) 0 256 1.00 0 0 |
| add_7 - (-1, 28, 28, 128) 0 0 0 ? ? |
| average_pooling2d - (-1, 1, 1, 128) 0 0 0 0 0 |
| flatten - (-1, 128) 0 0 0 0 0 |
| dense - (-1, 8) 0 1024 4.00 0 1024 |
| quant_conv2d_8 1 (-1, 14, 14, 256) 294912 0 36.00 57802752 0 |
| dense_1 - (-1, 256) 0 2048 8.00 0 2048 |
| max_pooling2d_1 - (-1, 14, 14, 128) 0 0 0 0 0 |
| batch_normalization_11 - (-1, 14, 14, 256) 0 512 2.00 0 0 |
| tf_op_layer_Reshape - ((None, 1, 1, 256),) 0 0 0 ? ? |
| conv2d_2 - (-1, 14, 14, 256) 0 32768 128.00 0 6422528 |
| tf_op_layer_Mul - ((None, 14, 14, 256),) 0 0 0 ? ? |
| batch_normalization_10 - (-1, 14, 14, 256) 0 512 2.00 0 0 |
| add_8 - (-1, 14, 14, 256) 0 0 0 ? ? |
| average_pooling2d_1 - (-1, 1, 1, 256) 0 0 0 0 0 |
| flatten_1 - (-1, 256) 0 0 0 0 0 |
| dense_2 - (-1, 16) 0 4096 16.00 0 4096 |
| quant_conv2d_9 1 (-1, 14, 14, 256) 589824 0 72.00 115605504 0 |
| dense_3 - (-1, 256) 0 4096 16.00 0 4096 |
| batch_normalization_12 - (-1, 14, 14, 256) 0 512 2.00 0 0 |
| tf_op_layer_Reshape_1 - ((None, 1, 1, 256),) 0 0 0 ? ? |
| tf_op_layer_Mul_1 - ((None, 14, 14, 256),) 0 0 0 ? ? |
| add_9 - (-1, 14, 14, 256) 0 0 0 ? ? |
| average_pooling2d_2 - (-1, 1, 1, 256) 0 0 0 0 0 |
| flatten_2 - (-1, 256) 0 0 0 0 0 |
| dense_4 - (-1, 16) 0 4096 16.00 0 4096 |
| quant_conv2d_10 1 (-1, 14, 14, 256) 589824 0 72.00 115605504 0 |
| dense_5 - (-1, 256) 0 4096 16.00 0 4096 |
| batch_normalization_13 - (-1, 14, 14, 256) 0 512 2.00 0 0 |
| tf_op_layer_Reshape_2 - ((None, 1, 1, 256),) 0 0 0 ? ? |
| tf_op_layer_Mul_2 - ((None, 14, 14, 256),) 0 0 0 ? ? |
| add_10 - (-1, 14, 14, 256) 0 0 0 ? ? |
| average_pooling2d_3 - (-1, 1, 1, 256) 0 0 0 0 0 |
| flatten_3 - (-1, 256) 0 0 0 0 0 |
| dense_6 - (-1, 16) 0 4096 16.00 0 4096 |
| quant_conv2d_11 1 (-1, 14, 14, 256) 589824 0 72.00 115605504 0 |
| dense_7 - (-1, 256) 0 4096 16.00 0 4096 |
| batch_normalization_14 - (-1, 14, 14, 256) 0 512 2.00 0 0 |
| tf_op_layer_Reshape_3 - ((None, 1, 1, 256),) 0 0 0 ? ? |
| tf_op_layer_Mul_3 - ((None, 14, 14, 256),) 0 0 0 ? ? |
| add_11 - (-1, 14, 14, 256) 0 0 0 ? ? |
| average_pooling2d_4 - (-1, 1, 1, 256) 0 0 0 0 0 |
| flatten_4 - (-1, 256) 0 0 0 0 0 |
| dense_8 - (-1, 16) 0 4096 16.00 0 4096 |
| quant_conv2d_12 1 (-1, 7, 7, 512) 1179648 0 144.00 57802752 0 |
| dense_9 - (-1, 512) 0 8192 32.00 0 8192 |
| max_pooling2d_2 - (-1, 7, 7, 256) 0 0 0 0 0 |
| batch_normalization_16 - (-1, 7, 7, 512) 0 1024 4.00 0 0 |
| tf_op_layer_Reshape_4 - ((None, 1, 1, 512),) 0 0 0 ? ? |
| conv2d_3 - (-1, 7, 7, 512) 0 131072 512.00 0 6422528 |
| tf_op_layer_Mul_4 - ((None, 7, 7, 512),) 0 0 0 ? ? |
| batch_normalization_15 - (-1, 7, 7, 512) 0 1024 4.00 0 0 |
| add_12 - (-1, 7, 7, 512) 0 0 0 ? ? |
| average_pooling2d_5 - (-1, 1, 1, 512) 0 0 0 0 0 |
| flatten_5 - (-1, 512) 0 0 0 0 0 |
| dense_10 - (-1, 32) 0 16384 64.00 0 16384 |
| quant_conv2d_13 1 (-1, 7, 7, 512) 2359296 0 288.00 115605504 0 |
| dense_11 - (-1, 512) 0 16384 64.00 0 16384 |
| batch_normalization_17 - (-1, 7, 7, 512) 0 1024 4.00 0 0 |
| tf_op_layer_Reshape_5 - ((None, 1, 1, 512),) 0 0 0 ? ? |
| tf_op_layer_Mul_5 - ((None, 7, 7, 512),) 0 0 0 ? ? |
| add_13 - (-1, 7, 7, 512) 0 0 0 ? ? |
| average_pooling2d_6 - (-1, 1, 1, 512) 0 0 0 0 0 |
| flatten_6 - (-1, 512) 0 0 0 0 0 |
| dense_12 - (-1, 32) 0 16384 64.00 0 16384 |
| quant_conv2d_14 1 (-1, 7, 7, 512) 2359296 0 288.00 115605504 0 |
| dense_13 - (-1, 512) 0 16384 64.00 0 16384 |
| batch_normalization_18 - (-1, 7, 7, 512) 0 1024 4.00 0 0 |
| tf_op_layer_Reshape_6 - ((None, 1, 1, 512),) 0 0 0 ? ? |
| tf_op_layer_Mul_6 - ((None, 7, 7, 512),) 0 0 0 ? ? |
| add_14 - (-1, 7, 7, 512) 0 0 0 ? ? |
| average_pooling2d_7 - (-1, 1, 1, 512) 0 0 0 0 0 |
| flatten_7 - (-1, 512) 0 0 0 0 0 |
| dense_14 - (-1, 32) 0 16384 64.00 0 16384 |
| quant_conv2d_15 1 (-1, 7, 7, 512) 2359296 0 288.00 115605504 0 |
| dense_15 - (-1, 512) 0 16384 64.00 0 16384 |
| batch_normalization_19 - (-1, 7, 7, 512) 0 1024 4.00 0 0 |
| tf_op_layer_Reshape_7 - ((None, 1, 1, 512),) 0 0 0 ? ? |
| tf_op_layer_Mul_7 - ((None, 7, 7, 512),) 0 0 0 ? ? |
| add_15 - (-1, 7, 7, 512) 0 0 0 ? ? |
| activation - (-1, 7, 7, 512) 0 0 0 ? ? |
| average_pooling2d_8 - (-1, 1, 1, 512) 0 0 0 0 0 |
| flatten_8 - (-1, 512) 0 0 0 0 0 |
| dense_16 - (-1, 1000) 0 513000 2003.91 0 512000 |
| activation_1 - (-1, 1000) 0 0 0 ? ? |
+-------------------------------------------------------------------------------------------------------------------+
| Total 10985472 833664 4597.50 1676279808 24433664 |
+-------------------------------------------------------------------------------------------------------------------+
+quicknet_large summary-----------------------+
| Total params 11.8 M |
| Trainable params 11.8 M |
| Non-trainable params 9.6 k |
| Model size 4.49 MiB |
| Model size (8-bit FP weights) 2.10 MiB |
| Float-32 Equivalent 45.09 MiB |
| Compression Ratio of Memory 0.10 |
| Number of MACs 1.7 B |
| Ratio of MACs that are binarized 0.9856 |
+---------------------------------------------+
ImageNet Metrics
| Top-1 Accuracy | Top-5 Accuracy | Parameters | Memory |
|---|---|---|---|
| 62.7 % | 84.0 % | 11 837 696 | 4.56 MB |
Arguments
- input_shape
Optional[Sequence[Optional[int]]]: Optional shape tuple, to be specified if you would like to use a model with an input image resolution that is not (224, 224, 3). It should have exactly 3 inputs channels. - input_tensor
Optional[tf.Tensor]: optional Keras tensor (i.e. output oflayers.Input()) to use as image input for the model. - weights
Optional[str]: one ofNone(random initialization), "imagenet" (pre-training on ImageNet), or the path to the weights file to be loaded. - include_top
bool: whether to include the fully-connected layer at the top of the network. - num_classes
int: optional number of classes to classify images into, only to be specified ifinclude_topis True, and if noweightsargument is specified.
Returns
A Keras model instance.
Raises
- ValueError: in case of invalid argument for
weights, or invalid input shape.
QuickNetXL¶
larq_zoo.sota.QuickNetXL(
*,
input_shape=None,
input_tensor=None,
weights="imagenet",
include_top=True,
num_classes=1000
)
Instantiates the QuickNetXL architecture.
Optionally loads weights pre-trained on ImageNet.
Model Summary
+quicknet_xl stats--------------------------------------------------------------------------------------------------+
| Layer Input prec. Outputs # 1-bit # 32-bit Memory 1-bit MACs 32-bit MACs |
| (bit) x 1 x 1 (kB) |
+-------------------------------------------------------------------------------------------------------------------+
| input_1 - ((None, 224, 224, 3),) 0 0 0 ? ? |
| conv2d - (-1, 112, 112, 8) 0 216 0.84 0 2709504 |
| depthwise_conv2d - (-1, 56, 56, 64) 0 576 2.25 0 1806336 |
| batch_normalization - (-1, 56, 56, 64) 0 128 0.50 0 0 |
| quant_conv2d 1 (-1, 56, 56, 64) 36864 0 4.50 115605504 0 |
| batch_normalization_1 - (-1, 56, 56, 64) 0 128 0.50 0 0 |
| add - (-1, 56, 56, 64) 0 0 0 ? ? |
| quant_conv2d_1 1 (-1, 56, 56, 64) 36864 0 4.50 115605504 0 |
| batch_normalization_2 - (-1, 56, 56, 64) 0 128 0.50 0 0 |
| add_1 - (-1, 56, 56, 64) 0 0 0 ? ? |
| quant_conv2d_2 1 (-1, 56, 56, 64) 36864 0 4.50 115605504 0 |
| batch_normalization_3 - (-1, 56, 56, 64) 0 128 0.50 0 0 |
| add_2 - (-1, 56, 56, 64) 0 0 0 ? ? |
| quant_conv2d_3 1 (-1, 56, 56, 64) 36864 0 4.50 115605504 0 |
| batch_normalization_4 - (-1, 56, 56, 64) 0 128 0.50 0 0 |
| add_3 - (-1, 56, 56, 64) 0 0 0 ? ? |
| quant_conv2d_4 1 (-1, 56, 56, 64) 36864 0 4.50 115605504 0 |
| batch_normalization_5 - (-1, 56, 56, 64) 0 128 0.50 0 0 |
| add_4 - (-1, 56, 56, 64) 0 0 0 ? ? |
| quant_conv2d_5 1 (-1, 56, 56, 64) 36864 0 4.50 115605504 0 |
| batch_normalization_6 - (-1, 56, 56, 64) 0 128 0.50 0 0 |
| add_5 - (-1, 56, 56, 64) 0 0 0 ? ? |
| max_pooling2d - (-1, 28, 28, 64) 0 0 0 0 0 |
| quant_conv2d_6 1 (-1, 28, 28, 128) 73728 0 9.00 57802752 0 |
| conv2d_1 - (-1, 28, 28, 128) 0 8192 32.00 0 6422528 |
| batch_normalization_8 - (-1, 28, 28, 128) 0 256 1.00 0 0 |
| batch_normalization_7 - (-1, 28, 28, 128) 0 256 1.00 0 0 |
| add_6 - (-1, 28, 28, 128) 0 0 0 ? ? |
| quant_conv2d_7 1 (-1, 28, 28, 128) 147456 0 18.00 115605504 0 |
| batch_normalization_9 - (-1, 28, 28, 128) 0 256 1.00 0 0 |
| add_7 - (-1, 28, 28, 128) 0 0 0 ? ? |
| quant_conv2d_8 1 (-1, 28, 28, 128) 147456 0 18.00 115605504 0 |
| batch_normalization_10 - (-1, 28, 28, 128) 0 256 1.00 0 0 |
| add_8 - (-1, 28, 28, 128) 0 0 0 ? ? |
| quant_conv2d_9 1 (-1, 28, 28, 128) 147456 0 18.00 115605504 0 |
| batch_normalization_11 - (-1, 28, 28, 128) 0 256 1.00 0 0 |
| add_9 - (-1, 28, 28, 128) 0 0 0 ? ? |
| quant_conv2d_10 1 (-1, 28, 28, 128) 147456 0 18.00 115605504 0 |
| batch_normalization_12 - (-1, 28, 28, 128) 0 256 1.00 0 0 |
| add_10 - (-1, 28, 28, 128) 0 0 0 ? ? |
| quant_conv2d_11 1 (-1, 28, 28, 128) 147456 0 18.00 115605504 0 |
| batch_normalization_13 - (-1, 28, 28, 128) 0 256 1.00 0 0 |
| add_11 - (-1, 28, 28, 128) 0 0 0 ? ? |
| quant_conv2d_12 1 (-1, 28, 28, 128) 147456 0 18.00 115605504 0 |
| batch_normalization_14 - (-1, 28, 28, 128) 0 256 1.00 0 0 |
| add_12 - (-1, 28, 28, 128) 0 0 0 ? ? |
| quant_conv2d_13 1 (-1, 28, 28, 128) 147456 0 18.00 115605504 0 |
| batch_normalization_15 - (-1, 28, 28, 128) 0 256 1.00 0 0 |
| add_13 - (-1, 28, 28, 128) 0 0 0 ? ? |
| average_pooling2d - (-1, 1, 1, 128) 0 0 0 0 0 |
| flatten - (-1, 128) 0 0 0 0 0 |
| dense - (-1, 8) 0 1024 4.00 0 1024 |
| quant_conv2d_14 1 (-1, 14, 14, 256) 294912 0 36.00 57802752 0 |
| dense_1 - (-1, 256) 0 2048 8.00 0 2048 |
| max_pooling2d_1 - (-1, 14, 14, 128) 0 0 0 0 0 |
| batch_normalization_17 - (-1, 14, 14, 256) 0 512 2.00 0 0 |
| tf_op_layer_Reshape - ((None, 1, 1, 256),) 0 0 0 ? ? |
| conv2d_2 - (-1, 14, 14, 256) 0 32768 128.00 0 6422528 |
| tf_op_layer_Mul - ((None, 14, 14, 256),) 0 0 0 ? ? |
| batch_normalization_16 - (-1, 14, 14, 256) 0 512 2.00 0 0 |
| add_14 - (-1, 14, 14, 256) 0 0 0 ? ? |
| average_pooling2d_1 - (-1, 1, 1, 256) 0 0 0 0 0 |
| flatten_1 - (-1, 256) 0 0 0 0 0 |
| dense_2 - (-1, 16) 0 4096 16.00 0 4096 |
| quant_conv2d_15 1 (-1, 14, 14, 256) 589824 0 72.00 115605504 0 |
| dense_3 - (-1, 256) 0 4096 16.00 0 4096 |
| batch_normalization_18 - (-1, 14, 14, 256) 0 512 2.00 0 0 |
| tf_op_layer_Reshape_1 - ((None, 1, 1, 256),) 0 0 0 ? ? |
| tf_op_layer_Mul_1 - ((None, 14, 14, 256),) 0 0 0 ? ? |
| add_15 - (-1, 14, 14, 256) 0 0 0 ? ? |
| average_pooling2d_2 - (-1, 1, 1, 256) 0 0 0 0 0 |
| flatten_2 - (-1, 256) 0 0 0 0 0 |
| dense_4 - (-1, 16) 0 4096 16.00 0 4096 |
| quant_conv2d_16 1 (-1, 14, 14, 256) 589824 0 72.00 115605504 0 |
| dense_5 - (-1, 256) 0 4096 16.00 0 4096 |
| batch_normalization_19 - (-1, 14, 14, 256) 0 512 2.00 0 0 |
| tf_op_layer_Reshape_2 - ((None, 1, 1, 256),) 0 0 0 ? ? |
| tf_op_layer_Mul_2 - ((None, 14, 14, 256),) 0 0 0 ? ? |
| add_16 - (-1, 14, 14, 256) 0 0 0 ? ? |
| average_pooling2d_3 - (-1, 1, 1, 256) 0 0 0 0 0 |
| flatten_3 - (-1, 256) 0 0 0 0 0 |
| dense_6 - (-1, 16) 0 4096 16.00 0 4096 |
| quant_conv2d_17 1 (-1, 14, 14, 256) 589824 0 72.00 115605504 0 |
| dense_7 - (-1, 256) 0 4096 16.00 0 4096 |
| batch_normalization_20 - (-1, 14, 14, 256) 0 512 2.00 0 0 |
| tf_op_layer_Reshape_3 - ((None, 1, 1, 256),) 0 0 0 ? ? |
| tf_op_layer_Mul_3 - ((None, 14, 14, 256),) 0 0 0 ? ? |
| add_17 - (-1, 14, 14, 256) 0 0 0 ? ? |
| average_pooling2d_4 - (-1, 1, 1, 256) 0 0 0 0 0 |
| flatten_4 - (-1, 256) 0 0 0 0 0 |
| dense_8 - (-1, 16) 0 4096 16.00 0 4096 |
| quant_conv2d_18 1 (-1, 14, 14, 256) 589824 0 72.00 115605504 0 |
| dense_9 - (-1, 256) 0 4096 16.00 0 4096 |
| batch_normalization_21 - (-1, 14, 14, 256) 0 512 2.00 0 0 |
| tf_op_layer_Reshape_4 - ((None, 1, 1, 256),) 0 0 0 ? ? |
| tf_op_layer_Mul_4 - ((None, 14, 14, 256),) 0 0 0 ? ? |
| add_18 - (-1, 14, 14, 256) 0 0 0 ? ? |
| average_pooling2d_5 - (-1, 1, 1, 256) 0 0 0 0 0 |
| flatten_5 - (-1, 256) 0 0 0 0 0 |
| dense_10 - (-1, 16) 0 4096 16.00 0 4096 |
| quant_conv2d_19 1 (-1, 14, 14, 256) 589824 0 72.00 115605504 0 |
| dense_11 - (-1, 256) 0 4096 16.00 0 4096 |
| batch_normalization_22 - (-1, 14, 14, 256) 0 512 2.00 0 0 |
| tf_op_layer_Reshape_5 - ((None, 1, 1, 256),) 0 0 0 ? ? |
| tf_op_layer_Mul_5 - ((None, 14, 14, 256),) 0 0 0 ? ? |
| add_19 - (-1, 14, 14, 256) 0 0 0 ? ? |
| average_pooling2d_6 - (-1, 1, 1, 256) 0 0 0 0 0 |
| flatten_6 - (-1, 256) 0 0 0 0 0 |
| dense_12 - (-1, 16) 0 4096 16.00 0 4096 |
| quant_conv2d_20 1 (-1, 14, 14, 256) 589824 0 72.00 115605504 0 |
| dense_13 - (-1, 256) 0 4096 16.00 0 4096 |
| batch_normalization_23 - (-1, 14, 14, 256) 0 512 2.00 0 0 |
| tf_op_layer_Reshape_6 - ((None, 1, 1, 256),) 0 0 0 ? ? |
| tf_op_layer_Mul_6 - ((None, 14, 14, 256),) 0 0 0 ? ? |
| add_20 - (-1, 14, 14, 256) 0 0 0 ? ? |
| average_pooling2d_7 - (-1, 1, 1, 256) 0 0 0 0 0 |
| flatten_7 - (-1, 256) 0 0 0 0 0 |
| dense_14 - (-1, 16) 0 4096 16.00 0 4096 |
| quant_conv2d_21 1 (-1, 14, 14, 256) 589824 0 72.00 115605504 0 |
| dense_15 - (-1, 256) 0 4096 16.00 0 4096 |
| batch_normalization_24 - (-1, 14, 14, 256) 0 512 2.00 0 0 |
| tf_op_layer_Reshape_7 - ((None, 1, 1, 256),) 0 0 0 ? ? |
| tf_op_layer_Mul_7 - ((None, 14, 14, 256),) 0 0 0 ? ? |
| add_21 - (-1, 14, 14, 256) 0 0 0 ? ? |
| average_pooling2d_8 - (-1, 1, 1, 256) 0 0 0 0 0 |
| flatten_8 - (-1, 256) 0 0 0 0 0 |
| dense_16 - (-1, 16) 0 4096 16.00 0 4096 |
| quant_conv2d_22 1 (-1, 14, 14, 256) 589824 0 72.00 115605504 0 |
| dense_17 - (-1, 256) 0 4096 16.00 0 4096 |
| batch_normalization_25 - (-1, 14, 14, 256) 0 512 2.00 0 0 |
| tf_op_layer_Reshape_8 - ((None, 1, 1, 256),) 0 0 0 ? ? |
| tf_op_layer_Mul_8 - ((None, 14, 14, 256),) 0 0 0 ? ? |
| add_22 - (-1, 14, 14, 256) 0 0 0 ? ? |
| average_pooling2d_9 - (-1, 1, 1, 256) 0 0 0 0 0 |
| flatten_9 - (-1, 256) 0 0 0 0 0 |
| dense_18 - (-1, 16) 0 4096 16.00 0 4096 |
| quant_conv2d_23 1 (-1, 14, 14, 256) 589824 0 72.00 115605504 0 |
| dense_19 - (-1, 256) 0 4096 16.00 0 4096 |
| batch_normalization_26 - (-1, 14, 14, 256) 0 512 2.00 0 0 |
| tf_op_layer_Reshape_9 - ((None, 1, 1, 256),) 0 0 0 ? ? |
| tf_op_layer_Mul_9 - ((None, 14, 14, 256),) 0 0 0 ? ? |
| add_23 - (-1, 14, 14, 256) 0 0 0 ? ? |
| average_pooling2d_10 - (-1, 1, 1, 256) 0 0 0 0 0 |
| flatten_10 - (-1, 256) 0 0 0 0 0 |
| dense_20 - (-1, 16) 0 4096 16.00 0 4096 |
| quant_conv2d_24 1 (-1, 14, 14, 256) 589824 0 72.00 115605504 0 |
| dense_21 - (-1, 256) 0 4096 16.00 0 4096 |
| batch_normalization_27 - (-1, 14, 14, 256) 0 512 2.00 0 0 |
| tf_op_layer_Reshape_10 - ((None, 1, 1, 256),) 0 0 0 ? ? |
| tf_op_layer_Mul_10 - ((None, 14, 14, 256),) 0 0 0 ? ? |
| add_24 - (-1, 14, 14, 256) 0 0 0 ? ? |
| average_pooling2d_11 - (-1, 1, 1, 256) 0 0 0 0 0 |
| flatten_11 - (-1, 256) 0 0 0 0 0 |
| dense_22 - (-1, 16) 0 4096 16.00 0 4096 |
| quant_conv2d_25 1 (-1, 14, 14, 256) 589824 0 72.00 115605504 0 |
| dense_23 - (-1, 256) 0 4096 16.00 0 4096 |
| batch_normalization_28 - (-1, 14, 14, 256) 0 512 2.00 0 0 |
| tf_op_layer_Reshape_11 - ((None, 1, 1, 256),) 0 0 0 ? ? |
| tf_op_layer_Mul_11 - ((None, 14, 14, 256),) 0 0 0 ? ? |
| add_25 - (-1, 14, 14, 256) 0 0 0 ? ? |
| average_pooling2d_12 - (-1, 1, 1, 256) 0 0 0 0 0 |
| flatten_12 - (-1, 256) 0 0 0 0 0 |
| dense_24 - (-1, 16) 0 4096 16.00 0 4096 |
| quant_conv2d_26 1 (-1, 7, 7, 512) 1179648 0 144.00 57802752 0 |
| dense_25 - (-1, 512) 0 8192 32.00 0 8192 |
| max_pooling2d_2 - (-1, 7, 7, 256) 0 0 0 0 0 |
| batch_normalization_30 - (-1, 7, 7, 512) 0 1024 4.00 0 0 |
| tf_op_layer_Reshape_12 - ((None, 1, 1, 512),) 0 0 0 ? ? |
| conv2d_3 - (-1, 7, 7, 512) 0 131072 512.00 0 6422528 |
| tf_op_layer_Mul_12 - ((None, 7, 7, 512),) 0 0 0 ? ? |
| batch_normalization_29 - (-1, 7, 7, 512) 0 1024 4.00 0 0 |
| add_26 - (-1, 7, 7, 512) 0 0 0 ? ? |
| average_pooling2d_13 - (-1, 1, 1, 512) 0 0 0 0 0 |
| flatten_13 - (-1, 512) 0 0 0 0 0 |
| dense_26 - (-1, 32) 0 16384 64.00 0 16384 |
| quant_conv2d_27 1 (-1, 7, 7, 512) 2359296 0 288.00 115605504 0 |
| dense_27 - (-1, 512) 0 16384 64.00 0 16384 |
| batch_normalization_31 - (-1, 7, 7, 512) 0 1024 4.00 0 0 |
| tf_op_layer_Reshape_13 - ((None, 1, 1, 512),) 0 0 0 ? ? |
| tf_op_layer_Mul_13 - ((None, 7, 7, 512),) 0 0 0 ? ? |
| add_27 - (-1, 7, 7, 512) 0 0 0 ? ? |
| average_pooling2d_14 - (-1, 1, 1, 512) 0 0 0 0 0 |
| flatten_14 - (-1, 512) 0 0 0 0 0 |
| dense_28 - (-1, 32) 0 16384 64.00 0 16384 |
| quant_conv2d_28 1 (-1, 7, 7, 512) 2359296 0 288.00 115605504 0 |
| dense_29 - (-1, 512) 0 16384 64.00 0 16384 |
| batch_normalization_32 - (-1, 7, 7, 512) 0 1024 4.00 0 0 |
| tf_op_layer_Reshape_14 - ((None, 1, 1, 512),) 0 0 0 ? ? |
| tf_op_layer_Mul_14 - ((None, 7, 7, 512),) 0 0 0 ? ? |
| add_28 - (-1, 7, 7, 512) 0 0 0 ? ? |
| average_pooling2d_15 - (-1, 1, 1, 512) 0 0 0 0 0 |
| flatten_15 - (-1, 512) 0 0 0 0 0 |
| dense_30 - (-1, 32) 0 16384 64.00 0 16384 |
| quant_conv2d_29 1 (-1, 7, 7, 512) 2359296 0 288.00 115605504 0 |
| dense_31 - (-1, 512) 0 16384 64.00 0 16384 |
| batch_normalization_33 - (-1, 7, 7, 512) 0 1024 4.00 0 0 |
| tf_op_layer_Reshape_15 - ((None, 1, 1, 512),) 0 0 0 ? ? |
| tf_op_layer_Mul_15 - ((None, 7, 7, 512),) 0 0 0 ? ? |
| add_29 - (-1, 7, 7, 512) 0 0 0 ? ? |
| average_pooling2d_16 - (-1, 1, 1, 512) 0 0 0 0 0 |
| flatten_16 - (-1, 512) 0 0 0 0 0 |
| dense_32 - (-1, 32) 0 16384 64.00 0 16384 |
| quant_conv2d_30 1 (-1, 7, 7, 512) 2359296 0 288.00 115605504 0 |
| dense_33 - (-1, 512) 0 16384 64.00 0 16384 |
| batch_normalization_34 - (-1, 7, 7, 512) 0 1024 4.00 0 0 |
| tf_op_layer_Reshape_16 - ((None, 1, 1, 512),) 0 0 0 ? ? |
| tf_op_layer_Mul_16 - ((None, 7, 7, 512),) 0 0 0 ? ? |
| add_30 - (-1, 7, 7, 512) 0 0 0 ? ? |
| average_pooling2d_17 - (-1, 1, 1, 512) 0 0 0 0 0 |
| flatten_17 - (-1, 512) 0 0 0 0 0 |
| dense_34 - (-1, 32) 0 16384 64.00 0 16384 |
| quant_conv2d_31 1 (-1, 7, 7, 512) 2359296 0 288.00 115605504 0 |
| dense_35 - (-1, 512) 0 16384 64.00 0 16384 |
| batch_normalization_35 - (-1, 7, 7, 512) 0 1024 4.00 0 0 |
| tf_op_layer_Reshape_17 - ((None, 1, 1, 512),) 0 0 0 ? ? |
| tf_op_layer_Mul_17 - ((None, 7, 7, 512),) 0 0 0 ? ? |
| add_31 - (-1, 7, 7, 512) 0 0 0 ? ? |
| activation - (-1, 7, 7, 512) 0 0 0 ? ? |
| average_pooling2d_18 - (-1, 1, 1, 512) 0 0 0 0 0 |
| flatten_18 - (-1, 512) 0 0 0 0 0 |
| dense_36 - (-1, 1000) 0 513000 2003.91 0 512000 |
| activation_1 - (-1, 1000) 0 0 0 ? ? |
+-------------------------------------------------------------------------------------------------------------------+
| Total 21086208 972160 6371.50 3525967872 24564736 |
+-------------------------------------------------------------------------------------------------------------------+
+quicknet_xl summary--------------------------+
| Total params 22.1 M |
| Trainable params 22 M |
| Non-trainable params 17 k |
| Model size 6.22 MiB |
| Model size (8-bit FP weights) 3.44 MiB |
| Float-32 Equivalent 84.15 MiB |
| Compression Ratio of Memory 0.07 |
| Number of MACs 3.55 B |
| Ratio of MACs that are binarized 0.9931 |
+---------------------------------------------+
ImageNet Metrics
| Top-1 Accuracy | Top-5 Accuracy | Parameters | Memory |
|---|---|---|---|
| 67.0 % | 87.3 % | 22 058 368 | 6.22 MB |
Arguments
- input_shape
Optional[Sequence[Optional[int]]]: Optional shape tuple, to be specified if you would like to use a model with an input image resolution that is not (224, 224, 3). It should have exactly 3 inputs channels. - input_tensor
Optional[tf.Tensor]: optional Keras tensor (i.e. output oflayers.Input()) to use as image input for the model. - weights
Optional[str]: one ofNone(random initialization), "imagenet" (pre-training on ImageNet), or the path to the weights file to be loaded. - include_top
bool: whether to include the fully-connected layer at the top of the network. - num_classes
int: optional number of classes to classify images into, only to be specified ifinclude_topis True, and if noweightsargument is specified.
Returns
A Keras model instance.
Raises
- ValueError: in case of invalid argument for
weights, or invalid input shape.