TensorFlow 分類モデルを使用したトレーニング後の量子化

この Jupyter ノートブックは、ローカルへのインストール後にのみ起動できます。

GitHub

この例では、301-tensorflow-training-openvino ノートブックで作成された OpenVINO モデルを量子化して推論速度を向上させる方法を示します。量子化は、NNCF を使用したトレーニング後の量子化によって実行されます。カスタム・データローダーとメトリックが定義され、元の IR モデルと量子化されたモデルの精度とパフォーマンスが計算されます。

目次

準備

ノートブックでは、トレーニング・ノートブックが実行され、中間表現 (IR) モデルが作成されている必要があります。IR モデルが存在しない場合、次のセルを実行するとトレーニング・ノートブックが実行されます。これにはしばらく時間がかかります。

%pip install -q tensorflow Pillow matplotlib numpy tqdm nncf
DEPRECATION: pytorch-lightning 1.6.5 has a non-standard dependency specifier torch>=1.8.*. pip 24.1 will enforce this behaviour change. A possible replacement is to upgrade to a newer version of pytorch-lightning or contact the author to suggest that they release a version with a conforming dependency specifiers. Discussion can be found at https://github.com/pypa/pip/issues/12063
Note: you may need to restart the kernel to use updated packages.
from pathlib import Path

import tensorflow as tf

model_xml = Path("model/flower/flower_ir.xml")
dataset_url = (
    "https://storage.googleapis.com/download.tensorflow.org/example_images/flower_photos.tgz"
)
data_dir = Path(tf.keras.utils.get_file("flower_photos", origin=dataset_url, untar=True))

if not model_xml.exists():
    print("Executing training notebook. This will take a while...")
    %run 301-tensorflow-training-openvino.ipynb
2024-02-10 01:09:00.730910: I tensorflow/core/util/port.cc:110] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable TF_ENABLE_ONEDNN_OPTS=0.
2024-02-10 01:09:00.766002: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
2024-02-10 01:09:01.406366: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
Executing training notebook. This will take a while...
DEPRECATION: pytorch-lightning 1.6.5 has a non-standard dependency specifier torch>=1.8.*. pip 24.1 will enforce this behaviour change. A possible replacement is to upgrade to a newer version of pytorch-lightning or contact the author to suggest that they release a version with a conforming dependency specifiers. Discussion can be found at https://github.com/pypa/pip/issues/12063
Note: you may need to restart the kernel to use updated packages.
3670
Found 3670 files belonging to 5 classes.
Using 2936 files for training.
2024-02-10 01:09:08.525687: E tensorflow/compiler/xla/stream_executor/cuda/cuda_driver.cc:266] failed call to cuInit: CUDA_ERROR_COMPAT_NOT_SUPPORTED_ON_DEVICE: forward compatibility was attempted on non supported HW
2024-02-10 01:09:08.525725: I tensorflow/compiler/xla/stream_executor/cuda/cuda_diagnostics.cc:168] retrieving CUDA diagnostic information for host: iotg-dev-workstation-07
2024-02-10 01:09:08.525729: I tensorflow/compiler/xla/stream_executor/cuda/cuda_diagnostics.cc:175] hostname: iotg-dev-workstation-07
2024-02-10 01:09:08.525856: I tensorflow/compiler/xla/stream_executor/cuda/cuda_diagnostics.cc:199] libcuda reported version is: 470.223.2
2024-02-10 01:09:08.525872: I tensorflow/compiler/xla/stream_executor/cuda/cuda_diagnostics.cc:203] kernel reported version is: 470.182.3
2024-02-10 01:09:08.525876: E tensorflow/compiler/xla/stream_executor/cuda/cuda_diagnostics.cc:312] kernel version 470.182.3 does not match DSO version 470.223.2 -- cannot find working devices in this configuration
Found 3670 files belonging to 5 classes.
Using 734 files for validation.
['daisy', 'dandelion', 'roses', 'sunflowers', 'tulips']
2024-02-10 01:09:08.855253: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'Placeholder/_0' with dtype string and shape [2936]
     [[{{node Placeholder/_0}}]]
2024-02-10 01:09:08.855534: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'Placeholder/_4' with dtype int32 and shape [2936]
     [[{{node Placeholder/_4}}]]
../_images/301-tensorflow-training-openvino-nncf-with-output_3_11.png
2024-02-10 01:09:09.711519: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'Placeholder/_4' with dtype int32 and shape [2936]
    [[{{node Placeholder/_4}}]]
2024-02-10 01:09:09.711766: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'Placeholder/_0' with dtype string and shape [2936]
    [[{{node Placeholder/_0}}]]
(32, 180, 180, 3)
(32,)
2024-02-10 01:09:10.063734: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'Placeholder/_4' with dtype int32 and shape [2936]
    [[{{node Placeholder/_4}}]]
2024-02-10 01:09:10.064340: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'Placeholder/_4' with dtype int32 and shape [2936]
    [[{{node Placeholder/_4}}]]
0.0 0.9970461
2024-02-10 01:09:10.875056: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'Placeholder/_0' with dtype string and shape [2936]
    [[{{node Placeholder/_0}}]]
2024-02-10 01:09:10.875365: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'Placeholder/_0' with dtype string and shape [2936]
    [[{{node Placeholder/_0}}]]
../_images/301-tensorflow-training-openvino-nncf-with-output_3_17.png
Model: "sequential_2"
_________________________________________________________________
Layer (type)                Output Shape              Param #
=================================================================
sequential_1 (Sequential)   (None, 180, 180, 3)       0
rescaling_2 (Rescaling)     (None, 180, 180, 3)       0
conv2d_3 (Conv2D)           (None, 180, 180, 16)      448
max_pooling2d_3 (MaxPooling  (None, 90, 90, 16)       0
2D)
conv2d_4 (Conv2D)           (None, 90, 90, 32)        4640
max_pooling2d_4 (MaxPooling  (None, 45, 45, 32)       0
2D)
conv2d_5 (Conv2D)           (None, 45, 45, 64)        18496
max_pooling2d_5 (MaxPooling  (None, 22, 22, 64)       0
2D)
dropout (Dropout)           (None, 22, 22, 64)        0
flatten_1 (Flatten)         (None, 30976)             0
dense_2 (Dense)             (None, 128)               3965056
outputs (Dense)             (None, 5)                 645
=================================================================
Total params: 3,989,285
Trainable params: 3,989,285
Non-trainable params: 0
_________________________________________________________________
Epoch 1/15
2024-02-10 01:09:11.882327: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'Placeholder/_0' with dtype string and shape [2936]
    [[{{node Placeholder/_0}}]]
2024-02-10 01:09:11.882802: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'Placeholder/_4' with dtype int32 and shape [2936]
    [[{{node Placeholder/_4}}]]
1/92 [..............................] - ETA: 1:32 - loss: 1.6315 - accuracy: 0.1562
   
2/92 [..............................] - ETA: 6s - loss: 1.7632 - accuracy: 0.2812
   
3/92 [..............................] - ETA: 5s - loss: 1.7516 - accuracy: 0.2708
   
4/92 [>.............................] - ETA: 5s - loss: 1.7249 - accuracy: 0.2578
   
5/92 [>.............................] - ETA: 5s - loss: 1.6968 - accuracy: 0.2750
   
6/92 [>.............................] - ETA: 5s - loss: 1.6729 - accuracy: 0.2917
   
7/92 [=>............................] - ETA: 5s - loss: 1.6459 - accuracy: 0.3080
   
8/92 [=>............................] - ETA: 5s - loss: 1.6410 - accuracy: 0.3008
   
9/92 [=>............................] - ETA: 4s - loss: 1.6246 - accuracy: 0.3125


10/92 [==>………………………] - ETA: 4s - loss: 1.6151 - accuracy: 0.3000



11/92 [==>………………………] - ETA: 4s - loss: 1.6065 - accuracy: 0.3011



12/92 [==>………………………] - ETA: 4s - loss: 1.5947 - accuracy: 0.3047



13/92 [===>……………………..] - ETA: 4s - loss: 1.5839 - accuracy: 0.3077



14/92 [===>……………………..] - ETA: 4s - loss: 1.5719 - accuracy: 0.3125



15/92 [===>……………………..] - ETA: 4s - loss: 1.5604 - accuracy: 0.3187



16/92 [====>…………………….] - ETA: 4s - loss: 1.5477 - accuracy: 0.3203



17/92 [====>…………………….] - ETA: 4s - loss: 1.5317 - accuracy: 0.3272



18/92 [====>…………………….] - ETA: 4s - loss: 1.5153 - accuracy: 0.3368



19/92 [=====>……………………] - ETA: 4s - loss: 1.5118 - accuracy: 0.3355



20/92 [=====>……………………] - ETA: 4s - loss: 1.4901 - accuracy: 0.3484



21/92 [=====>……………………] - ETA: 4s - loss: 1.4818 - accuracy: 0.3569



22/92 [======>…………………..] - ETA: 4s - loss: 1.4839 - accuracy: 0.3563



23/92 [======>…………………..] - ETA: 4s - loss: 1.4731 - accuracy: 0.3599



24/92 [======>…………………..] - ETA: 3s - loss: 1.4556 - accuracy: 0.3724



25/92 [=======>………………….] - ETA: 3s - loss: 1.4413 - accuracy: 0.3788



26/92 [=======>………………….] - ETA: 3s - loss: 1.4353 - accuracy: 0.3774



27/92 [=======>………………….] - ETA: 3s - loss: 1.4367 - accuracy: 0.3762



28/92 [========>…………………] - ETA: 3s - loss: 1.4293 - accuracy: 0.3750



29/92 [========>…………………] - ETA: 3s - loss: 1.4196 - accuracy: 0.3793



30/92 [========>…………………] - ETA: 3s - loss: 1.4177 - accuracy: 0.3813



31/92 [=========>………………..] - ETA: 3s - loss: 1.4057 - accuracy: 0.3872



32/92 [=========>………………..] - ETA: 3s - loss: 1.4028 - accuracy: 0.3868



33/92 [=========>………………..] - ETA: 3s - loss: 1.3896 - accuracy: 0.3950



34/92 [==========>……………….] - ETA: 3s - loss: 1.3879 - accuracy: 0.3963



35/92 [==========>……………….] - ETA: 3s - loss: 1.3886 - accuracy: 0.3966



36/92 [==========>……………….] - ETA: 3s - loss: 1.3839 - accuracy: 0.3969



37/92 [===========>………………] - ETA: 3s - loss: 1.3853 - accuracy: 0.4022



38/92 [===========>………………] - ETA: 3s - loss: 1.3812 - accuracy: 0.4023



39/92 [===========>………………] - ETA: 3s - loss: 1.3746 - accuracy: 0.4065



40/92 [============>……………..] - ETA: 3s - loss: 1.3733 - accuracy: 0.4049



41/92 [============>……………..] - ETA: 2s - loss: 1.3684 - accuracy: 0.4064



42/92 [============>……………..] - ETA: 2s - loss: 1.3665 - accuracy: 0.4064



43/92 [=============>…………….] - ETA: 2s - loss: 1.3624 - accuracy: 0.4108



44/92 [=============>…………….] - ETA: 2s - loss: 1.3590 - accuracy: 0.4121



45/92 [=============>…………….] - ETA: 2s - loss: 1.3533 - accuracy: 0.4148



46/92 [==============>……………] - ETA: 2s - loss: 1.3472 - accuracy: 0.4167



47/92 [==============>……………] - ETA: 2s - loss: 1.3448 - accuracy: 0.4164



48/92 [==============>……………] - ETA: 2s - loss: 1.3409 - accuracy: 0.4162



49/92 [==============>……………] - ETA: 2s - loss: 1.3383 - accuracy: 0.4186



50/92 [===============>…………..] - ETA: 2s - loss: 1.3381 - accuracy: 0.4190



51/92 [===============>…………..] - ETA: 2s - loss: 1.3341 - accuracy: 0.4212



52/92 [===============>…………..] - ETA: 2s - loss: 1.3292 - accuracy: 0.4245



53/92 [================>………….] - ETA: 2s - loss: 1.3286 - accuracy: 0.4277



54/92 [================>………….] - ETA: 2s - loss: 1.3246 - accuracy: 0.4302



55/92 [================>………….] - ETA: 2s - loss: 1.3228 - accuracy: 0.4309



56/92 [=================>…………] - ETA: 2s - loss: 1.3231 - accuracy: 0.4355



57/92 [=================>…………] - ETA: 2s - loss: 1.3221 - accuracy: 0.4350



58/92 [=================>…………] - ETA: 1s - loss: 1.3200 - accuracy: 0.4378



59/92 [==================>………..] - ETA: 1s - loss: 1.3177 - accuracy: 0.4394



60/92 [==================>………..] - ETA: 1s - loss: 1.3148 - accuracy: 0.4409



61/92 [==================>………..] - ETA: 1s - loss: 1.3140 - accuracy: 0.4408



62/92 [===================>……….] - ETA: 1s - loss: 1.3080 - accuracy: 0.4443



63/92 [===================>……….] - ETA: 1s - loss: 1.3096 - accuracy: 0.4447



64/92 [===================>……….] - ETA: 1s - loss: 1.3068 - accuracy: 0.4451



65/92 [====================>………] - ETA: 1s - loss: 1.3014 - accuracy: 0.4469



66/92 [====================>………] - ETA: 1s - loss: 1.3013 - accuracy: 0.4468



67/92 [====================>………] - ETA: 1s - loss: 1.2977 - accuracy: 0.4480



68/92 [=====================>……..] - ETA: 1s - loss: 1.2948 - accuracy: 0.4493



69/92 [=====================>……..] - ETA: 1s - loss: 1.2914 - accuracy: 0.4500



70/92 [=====================>……..] - ETA: 1s - loss: 1.2929 - accuracy: 0.4476



71/92 [======================>…….] - ETA: 1s - loss: 1.2929 - accuracy: 0.4479



72/92 [======================>…….] - ETA: 1s - loss: 1.2902 - accuracy: 0.4495



73/92 [======================>…….] - ETA: 1s - loss: 1.2864 - accuracy: 0.4506



74/92 [=======================>……] - ETA: 1s - loss: 1.2854 - accuracy: 0.4504



75/92 [=======================>……] - ETA: 0s - loss: 1.2853 - accuracy: 0.4494



76/92 [=======================>……] - ETA: 0s - loss: 1.2809 - accuracy: 0.4513



77/92 [========================>…..] - ETA: 0s - loss: 1.2779 - accuracy: 0.4532



78/92 [========================>…..] - ETA: 0s - loss: 1.2774 - accuracy: 0.4538



79/92 [========================>…..] - ETA: 0s - loss: 1.2739 - accuracy: 0.4540



80/92 [=========================>….] - ETA: 0s - loss: 1.2722 - accuracy: 0.4542



81/92 [=========================>….] - ETA: 0s - loss: 1.2669 - accuracy: 0.4582



82/92 [=========================>….] - ETA: 0s - loss: 1.2654 - accuracy: 0.4599



83/92 [==========================>…] - ETA: 0s - loss: 1.2625 - accuracy: 0.4622



84/92 [==========================>…] - ETA: 0s - loss: 1.2580 - accuracy: 0.4642



85/92 [==========================>…] - ETA: 0s - loss: 1.2573 - accuracy: 0.4653



86/92 [===========================>..] - ETA: 0s - loss: 1.2566 - accuracy: 0.4665



87/92 [===========================>..] - ETA: 0s - loss: 1.2564 - accuracy: 0.4665



88/92 [===========================>..] - ETA: 0s - loss: 1.2530 - accuracy: 0.4679



89/92 [============================>.] - ETA: 0s - loss: 1.2492 - accuracy: 0.4690



90/92 [============================>.] - ETA: 0s - loss: 1.2445 - accuracy: 0.4714



91/92 [============================>.] - ETA: 0s - loss: 1.2402 - accuracy: 0.4735



92/92 [==============================] - ETA: 0s - loss: 1.2400 - accuracy: 0.4741

2024-02-10 01:09:18.229567: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'Placeholder/_0' with dtype string and shape [734]
    [[{{node Placeholder/_0}}]]
2024-02-10 01:09:18.229847: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'Placeholder/_4' with dtype int32 and shape [734]
    [[{{node Placeholder/_4}}]]


92/92 [==============================] - 7s 66ms/step - loss: 1.2400 - accuracy: 0.4741 - val_loss: 1.3762 - val_accuracy: 0.5014

Epoch 2/15
1/92 [..............................] - ETA: 7s - loss: 1.2018 - accuracy: 0.5625
   
2/92 [..............................] - ETA: 5s - loss: 1.0597 - accuracy: 0.5469
   
3/92 [..............................] - ETA: 5s - loss: 1.0781 - accuracy: 0.5521
   
4/92 [>.............................] - ETA: 5s - loss: 0.9952 - accuracy: 0.5938
   
5/92 [>.............................] - ETA: 5s - loss: 0.9872 - accuracy: 0.6187
   
6/92 [>.............................] - ETA: 5s - loss: 0.9615 - accuracy: 0.6094
   
7/92 [=>............................] - ETA: 4s - loss: 0.9715 - accuracy: 0.6205
   
8/92 [=>............................] - ETA: 4s - loss: 0.9608 - accuracy: 0.6211
   
9/92 [=>............................] - ETA: 4s - loss: 0.9538 - accuracy: 0.6250


10/92 [==>………………………] - ETA: 4s - loss: 0.9487 - accuracy: 0.6250



11/92 [==>………………………] - ETA: 4s - loss: 0.9616 - accuracy: 0.6307



12/92 [==>………………………] - ETA: 4s - loss: 0.9475 - accuracy: 0.6302



13/92 [===>……………………..] - ETA: 4s - loss: 0.9476 - accuracy: 0.6346



14/92 [===>……………………..] - ETA: 4s - loss: 0.9426 - accuracy: 0.6295



15/92 [===>……………………..] - ETA: 4s - loss: 0.9547 - accuracy: 0.6250



16/92 [====>…………………….] - ETA: 4s - loss: 0.9611 - accuracy: 0.6211



17/92 [====>…………………….] - ETA: 4s - loss: 0.9566 - accuracy: 0.6176



18/92 [====>…………………….] - ETA: 4s - loss: 0.9654 - accuracy: 0.6111



19/92 [=====>……………………] - ETA: 4s - loss: 0.9686 - accuracy: 0.6118



20/92 [=====>……………………] - ETA: 4s - loss: 0.9718 - accuracy: 0.6078



21/92 [=====>……………………] - ETA: 4s - loss: 0.9820 - accuracy: 0.6012



22/92 [======>…………………..] - ETA: 4s - loss: 0.9866 - accuracy: 0.5994



23/92 [======>…………………..] - ETA: 4s - loss: 0.9839 - accuracy: 0.6005



24/92 [======>…………………..] - ETA: 3s - loss: 0.9903 - accuracy: 0.5964



25/92 [=======>………………….] - ETA: 3s - loss: 0.9938 - accuracy: 0.5950



26/92 [=======>………………….] - ETA: 3s - loss: 1.0051 - accuracy: 0.5901



27/92 [=======>………………….] - ETA: 3s - loss: 0.9990 - accuracy: 0.5938



28/92 [========>…………………] - ETA: 3s - loss: 0.9965 - accuracy: 0.5949



29/92 [========>…………………] - ETA: 3s - loss: 0.9986 - accuracy: 0.5948



30/92 [========>…………………] - ETA: 3s - loss: 0.9992 - accuracy: 0.5917



31/92 [=========>………………..] - ETA: 3s - loss: 0.9976 - accuracy: 0.5927



33/92 [=========>………………..] - ETA: 3s - loss: 0.9922 - accuracy: 0.5964



34/92 [==========>……………….] - ETA: 3s - loss: 0.9964 - accuracy: 0.5954



35/92 [==========>……………….] - ETA: 3s - loss: 1.0031 - accuracy: 0.5935



36/92 [==========>……………….] - ETA: 3s - loss: 1.0041 - accuracy: 0.5909



37/92 [===========>………………] - ETA: 3s - loss: 1.0075 - accuracy: 0.5884



38/92 [===========>………………] - ETA: 3s - loss: 1.0049 - accuracy: 0.5886



39/92 [===========>………………] - ETA: 3s - loss: 1.0093 - accuracy: 0.5879



40/92 [============>……………..] - ETA: 3s - loss: 1.0040 - accuracy: 0.5912



41/92 [============>……………..] - ETA: 2s - loss: 1.0038 - accuracy: 0.5920



42/92 [============>……………..] - ETA: 2s - loss: 1.0080 - accuracy: 0.5898



43/92 [=============>…………….] - ETA: 2s - loss: 1.0082 - accuracy: 0.5914



44/92 [=============>…………….] - ETA: 2s - loss: 1.0021 - accuracy: 0.5936



45/92 [=============>…………….] - ETA: 2s - loss: 1.0045 - accuracy: 0.5936



46/92 [==============>……………] - ETA: 2s - loss: 1.0064 - accuracy: 0.5922



47/92 [==============>……………] - ETA: 2s - loss: 1.0063 - accuracy: 0.5936



48/92 [==============>……………] - ETA: 2s - loss: 1.0006 - accuracy: 0.5975



49/92 [==============>……………] - ETA: 2s - loss: 0.9997 - accuracy: 0.5981



50/92 [===============>…………..] - ETA: 2s - loss: 0.9995 - accuracy: 0.5980



51/92 [===============>…………..] - ETA: 2s - loss: 0.9969 - accuracy: 0.5998



52/92 [===============>…………..] - ETA: 2s - loss: 0.9952 - accuracy: 0.5996



53/92 [================>………….] - ETA: 2s - loss: 0.9974 - accuracy: 0.5983



54/92 [================>………….] - ETA: 2s - loss: 1.0079 - accuracy: 0.5913



55/92 [================>………….] - ETA: 2s - loss: 1.0106 - accuracy: 0.5908



56/92 [=================>…………] - ETA: 2s - loss: 1.0121 - accuracy: 0.5880



57/92 [=================>…………] - ETA: 2s - loss: 1.0146 - accuracy: 0.5887



58/92 [=================>…………] - ETA: 1s - loss: 1.0174 - accuracy: 0.5882



59/92 [==================>………..] - ETA: 1s - loss: 1.0150 - accuracy: 0.5883



60/92 [==================>………..] - ETA: 1s - loss: 1.0120 - accuracy: 0.5884



61/92 [==================>………..] - ETA: 1s - loss: 1.0103 - accuracy: 0.5890



62/92 [===================>……….] - ETA: 1s - loss: 1.0095 - accuracy: 0.5901



63/92 [===================>……….] - ETA: 1s - loss: 1.0114 - accuracy: 0.5886



64/92 [===================>……….] - ETA: 1s - loss: 1.0114 - accuracy: 0.5897



65/92 [====================>………] - ETA: 1s - loss: 1.0108 - accuracy: 0.5907



66/92 [====================>………] - ETA: 1s - loss: 1.0115 - accuracy: 0.5903



67/92 [====================>………] - ETA: 1s - loss: 1.0117 - accuracy: 0.5894



68/92 [=====================>……..] - ETA: 1s - loss: 1.0125 - accuracy: 0.5881



69/92 [=====================>……..] - ETA: 1s - loss: 1.0151 - accuracy: 0.5850



70/92 [=====================>……..] - ETA: 1s - loss: 1.0146 - accuracy: 0.5865



71/92 [======================>…….] - ETA: 1s - loss: 1.0158 - accuracy: 0.5866



72/92 [======================>…….] - ETA: 1s - loss: 1.0140 - accuracy: 0.5871



73/92 [======================>…….] - ETA: 1s - loss: 1.0130 - accuracy: 0.5872



74/92 [=======================>……] - ETA: 1s - loss: 1.0127 - accuracy: 0.5873



75/92 [=======================>……] - ETA: 0s - loss: 1.0109 - accuracy: 0.5886



76/92 [=======================>……] - ETA: 0s - loss: 1.0074 - accuracy: 0.5895



77/92 [========================>…..] - ETA: 0s - loss: 1.0038 - accuracy: 0.5916



78/92 [========================>…..] - ETA: 0s - loss: 1.0003 - accuracy: 0.5936



79/92 [========================>…..] - ETA: 0s - loss: 0.9982 - accuracy: 0.5944



80/92 [=========================>….] - ETA: 0s - loss: 0.9978 - accuracy: 0.5929



81/92 [=========================>….] - ETA: 0s - loss: 1.0002 - accuracy: 0.5940



82/92 [=========================>….] - ETA: 0s - loss: 0.9975 - accuracy: 0.5944



83/92 [==========================>…] - ETA: 0s - loss: 0.9982 - accuracy: 0.5948



84/92 [==========================>…] - ETA: 0s - loss: 0.9979 - accuracy: 0.5963



85/92 [==========================>…] - ETA: 0s - loss: 0.9960 - accuracy: 0.5970



86/92 [===========================>..] - ETA: 0s - loss: 0.9941 - accuracy: 0.5977



87/92 [===========================>..] - ETA: 0s - loss: 0.9931 - accuracy: 0.5973



88/92 [===========================>..] - ETA: 0s - loss: 0.9948 - accuracy: 0.5954



89/92 [============================>.] - ETA: 0s - loss: 0.9953 - accuracy: 0.5951



90/92 [============================>.] - ETA: 0s - loss: 0.9947 - accuracy: 0.5961



91/92 [============================>.] - ETA: 0s - loss: 0.9932 - accuracy: 0.5964



92/92 [==============================] - ETA: 0s - loss: 0.9956 - accuracy: 0.5974



92/92 [==============================] - 6s 64ms/step - loss: 0.9956 - accuracy: 0.5974 - val_loss: 0.9920 - val_accuracy: 0.6090

Epoch 3/15
1/92 [..............................] - ETA: 7s - loss: 1.2602 - accuracy: 0.4688
   
2/92 [..............................] - ETA: 5s - loss: 1.1814 - accuracy: 0.5781
   
3/92 [..............................] - ETA: 5s - loss: 1.1491 - accuracy: 0.5625
   
4/92 [>.............................] - ETA: 5s - loss: 1.0875 - accuracy: 0.5781
   
5/92 [>.............................] - ETA: 5s - loss: 1.0316 - accuracy: 0.5875
   
6/92 [>.............................] - ETA: 4s - loss: 1.0206 - accuracy: 0.5833
   
7/92 [=>............................] - ETA: 4s - loss: 0.9818 - accuracy: 0.5938
   
8/92 [=>............................] - ETA: 4s - loss: 1.0018 - accuracy: 0.5859
   
9/92 [=>............................] - ETA: 4s - loss: 0.9855 - accuracy: 0.5938


10/92 [==>………………………] - ETA: 4s - loss: 0.9760 - accuracy: 0.5969



11/92 [==>………………………] - ETA: 4s - loss: 0.9811 - accuracy: 0.5881



12/92 [==>………………………] - ETA: 4s - loss: 0.9836 - accuracy: 0.5859



13/92 [===>……………………..] - ETA: 4s - loss: 0.9757 - accuracy: 0.5889



14/92 [===>……………………..] - ETA: 4s - loss: 0.9660 - accuracy: 0.5893



15/92 [===>……………………..] - ETA: 4s - loss: 0.9619 - accuracy: 0.5938



16/92 [====>…………………….] - ETA: 4s - loss: 0.9688 - accuracy: 0.5898



17/92 [====>…………………….] - ETA: 4s - loss: 0.9691 - accuracy: 0.5919



18/92 [====>…………………….] - ETA: 4s - loss: 0.9729 - accuracy: 0.5938



20/92 [=====>……………………] - ETA: 4s - loss: 0.9704 - accuracy: 0.5934



21/92 [=====>……………………] - ETA: 4s - loss: 0.9652 - accuracy: 0.5934



22/92 [======>…………………..] - ETA: 4s - loss: 0.9528 - accuracy: 0.6006



23/92 [======>…………………..] - ETA: 4s - loss: 0.9511 - accuracy: 0.6044



24/92 [======>…………………..] - ETA: 3s - loss: 0.9597 - accuracy: 0.6026



25/92 [=======>………………….] - ETA: 3s - loss: 0.9707 - accuracy: 0.5972



26/92 [=======>………………….] - ETA: 3s - loss: 0.9649 - accuracy: 0.5971



27/92 [=======>………………….] - ETA: 3s - loss: 0.9528 - accuracy: 0.6016



28/92 [========>…………………] - ETA: 3s - loss: 0.9453 - accuracy: 0.6059



29/92 [========>…………………] - ETA: 3s - loss: 0.9458 - accuracy: 0.6065



30/92 [========>…………………] - ETA: 3s - loss: 0.9467 - accuracy: 0.6061



31/92 [=========>………………..] - ETA: 3s - loss: 0.9432 - accuracy: 0.6098



32/92 [=========>………………..] - ETA: 3s - loss: 0.9430 - accuracy: 0.6093



33/92 [=========>………………..] - ETA: 3s - loss: 0.9351 - accuracy: 0.6164



34/92 [==========>……………….] - ETA: 3s - loss: 0.9405 - accuracy: 0.6139



35/92 [==========>……………….] - ETA: 3s - loss: 0.9356 - accuracy: 0.6169



36/92 [==========>……………….] - ETA: 3s - loss: 0.9322 - accuracy: 0.6215



37/92 [===========>………………] - ETA: 3s - loss: 0.9363 - accuracy: 0.6199



38/92 [===========>………………] - ETA: 3s - loss: 0.9325 - accuracy: 0.6217



39/92 [===========>………………] - ETA: 3s - loss: 0.9322 - accuracy: 0.6234



40/92 [============>……………..] - ETA: 3s - loss: 0.9291 - accuracy: 0.6250



41/92 [============>……………..] - ETA: 2s - loss: 0.9252 - accuracy: 0.6242



42/92 [============>……………..] - ETA: 2s - loss: 0.9226 - accuracy: 0.6257



43/92 [=============>…………….] - ETA: 2s - loss: 0.9213 - accuracy: 0.6250



44/92 [=============>…………….] - ETA: 2s - loss: 0.9232 - accuracy: 0.6243



45/92 [=============>…………….] - ETA: 2s - loss: 0.9261 - accuracy: 0.6222



46/92 [==============>……………] - ETA: 2s - loss: 0.9277 - accuracy: 0.6209



47/92 [==============>……………] - ETA: 2s - loss: 0.9308 - accuracy: 0.6197



48/92 [==============>……………] - ETA: 2s - loss: 0.9392 - accuracy: 0.6145



49/92 [==============>……………] - ETA: 2s - loss: 0.9396 - accuracy: 0.6122



50/92 [===============>…………..] - ETA: 2s - loss: 0.9369 - accuracy: 0.6137



51/92 [===============>…………..] - ETA: 2s - loss: 0.9361 - accuracy: 0.6139



52/92 [===============>…………..] - ETA: 2s - loss: 0.9326 - accuracy: 0.6147



53/92 [================>………….] - ETA: 2s - loss: 0.9335 - accuracy: 0.6149



54/92 [================>………….] - ETA: 2s - loss: 0.9319 - accuracy: 0.6169



55/92 [================>………….] - ETA: 2s - loss: 0.9303 - accuracy: 0.6170



56/92 [=================>…………] - ETA: 2s - loss: 0.9289 - accuracy: 0.6183



57/92 [=================>…………] - ETA: 2s - loss: 0.9316 - accuracy: 0.6167



58/92 [=================>…………] - ETA: 1s - loss: 0.9314 - accuracy: 0.6163



59/92 [==================>………..] - ETA: 1s - loss: 0.9313 - accuracy: 0.6165



60/92 [==================>………..] - ETA: 1s - loss: 0.9282 - accuracy: 0.6182



61/92 [==================>………..] - ETA: 1s - loss: 0.9264 - accuracy: 0.6188



62/92 [===================>……….] - ETA: 1s - loss: 0.9249 - accuracy: 0.6194



63/92 [===================>……….] - ETA: 1s - loss: 0.9295 - accuracy: 0.6165



64/92 [===================>……….] - ETA: 1s - loss: 0.9288 - accuracy: 0.6157



65/92 [====================>………] - ETA: 1s - loss: 0.9234 - accuracy: 0.6178



66/92 [====================>………] - ETA: 1s - loss: 0.9228 - accuracy: 0.6179



67/92 [====================>………] - ETA: 1s - loss: 0.9232 - accuracy: 0.6170



68/92 [=====================>……..] - ETA: 1s - loss: 0.9240 - accuracy: 0.6172



69/92 [=====================>……..] - ETA: 1s - loss: 0.9200 - accuracy: 0.6205



70/92 [=====================>……..] - ETA: 1s - loss: 0.9244 - accuracy: 0.6196



71/92 [======================>…….] - ETA: 1s - loss: 0.9237 - accuracy: 0.6197



72/92 [======================>…….] - ETA: 1s - loss: 0.9253 - accuracy: 0.6193



73/92 [======================>…….] - ETA: 1s - loss: 0.9243 - accuracy: 0.6203



74/92 [=======================>……] - ETA: 1s - loss: 0.9259 - accuracy: 0.6195



75/92 [=======================>……] - ETA: 0s - loss: 0.9242 - accuracy: 0.6204



76/92 [=======================>……] - ETA: 0s - loss: 0.9216 - accuracy: 0.6213



77/92 [========================>…..] - ETA: 0s - loss: 0.9203 - accuracy: 0.6209



78/92 [========================>…..] - ETA: 0s - loss: 0.9200 - accuracy: 0.6214



79/92 [========================>…..] - ETA: 0s - loss: 0.9181 - accuracy: 0.6226



80/92 [=========================>….] - ETA: 0s - loss: 0.9185 - accuracy: 0.6226



81/92 [=========================>….] - ETA: 0s - loss: 0.9160 - accuracy: 0.6250



82/92 [=========================>….] - ETA: 0s - loss: 0.9186 - accuracy: 0.6250



83/92 [==========================>…] - ETA: 0s - loss: 0.9164 - accuracy: 0.6258



84/92 [==========================>…] - ETA: 0s - loss: 0.9167 - accuracy: 0.6269



85/92 [==========================>…] - ETA: 0s - loss: 0.9177 - accuracy: 0.6265



86/92 [===========================>..] - ETA: 0s - loss: 0.9183 - accuracy: 0.6276



87/92 [===========================>..] - ETA: 0s - loss: 0.9182 - accuracy: 0.6275



88/92 [===========================>..] - ETA: 0s - loss: 0.9156 - accuracy: 0.6278



89/92 [============================>.] - ETA: 0s - loss: 0.9135 - accuracy: 0.6292



90/92 [============================>.] - ETA: 0s - loss: 0.9121 - accuracy: 0.6302



91/92 [============================>.] - ETA: 0s - loss: 0.9129 - accuracy: 0.6298



92/92 [==============================] - ETA: 0s - loss: 0.9155 - accuracy: 0.6298



92/92 [==============================] - 6s 64ms/step - loss: 0.9155 - accuracy: 0.6298 - val_loss: 0.8959 - val_accuracy: 0.6621

Epoch 4/15
1/92 [..............................] - ETA: 7s - loss: 0.7704 - accuracy: 0.7812
   
2/92 [..............................] - ETA: 5s - loss: 0.8739 - accuracy: 0.6562
   
3/92 [..............................] - ETA: 5s - loss: 0.9644 - accuracy: 0.6146
   
4/92 [>.............................] - ETA: 5s - loss: 0.9070 - accuracy: 0.6484
   
5/92 [>.............................] - ETA: 4s - loss: 0.8696 - accuracy: 0.6625
   
6/92 [>.............................] - ETA: 4s - loss: 0.8536 - accuracy: 0.6562
   
7/92 [=>............................] - ETA: 4s - loss: 0.8587 - accuracy: 0.6473
   
8/92 [=>............................] - ETA: 4s - loss: 0.8727 - accuracy: 0.6523
   
9/92 [=>............................] - ETA: 4s - loss: 0.8413 - accuracy: 0.6701


10/92 [==>………………………] - ETA: 4s - loss: 0.8577 - accuracy: 0.6594



11/92 [==>………………………] - ETA: 4s - loss: 0.8386 - accuracy: 0.6733



12/92 [==>………………………] - ETA: 4s - loss: 0.8637 - accuracy: 0.6589



13/92 [===>……………………..] - ETA: 4s - loss: 0.8819 - accuracy: 0.6659



14/92 [===>……………………..] - ETA: 4s - loss: 0.8783 - accuracy: 0.6674



15/92 [===>……………………..] - ETA: 4s - loss: 0.8797 - accuracy: 0.6667



16/92 [====>…………………….] - ETA: 4s - loss: 0.8644 - accuracy: 0.6777



17/92 [====>…………………….] - ETA: 4s - loss: 0.8715 - accuracy: 0.6746



18/92 [====>…………………….] - ETA: 4s - loss: 0.8544 - accuracy: 0.6788



19/92 [=====>……………………] - ETA: 4s - loss: 0.8484 - accuracy: 0.6809



20/92 [=====>……………………] - ETA: 4s - loss: 0.8429 - accuracy: 0.6828



21/92 [=====>……………………] - ETA: 4s - loss: 0.8352 - accuracy: 0.6860



22/92 [======>…………………..] - ETA: 4s - loss: 0.8293 - accuracy: 0.6875



23/92 [======>…………………..] - ETA: 3s - loss: 0.8324 - accuracy: 0.6861



24/92 [======>…………………..] - ETA: 3s - loss: 0.8321 - accuracy: 0.6875



25/92 [=======>………………….] - ETA: 3s - loss: 0.8377 - accuracy: 0.6913



26/92 [=======>………………….] - ETA: 3s - loss: 0.8366 - accuracy: 0.6923



27/92 [=======>………………….] - ETA: 3s - loss: 0.8278 - accuracy: 0.6933



28/92 [========>…………………] - ETA: 3s - loss: 0.8303 - accuracy: 0.6942



29/92 [========>…………………] - ETA: 3s - loss: 0.8305 - accuracy: 0.6950



30/92 [========>…………………] - ETA: 3s - loss: 0.8342 - accuracy: 0.6958



31/92 [=========>………………..] - ETA: 3s - loss: 0.8350 - accuracy: 0.6956



32/92 [=========>………………..] - ETA: 3s - loss: 0.8386 - accuracy: 0.6914



33/92 [=========>………………..] - ETA: 3s - loss: 0.8354 - accuracy: 0.6922



34/92 [==========>……………….] - ETA: 3s - loss: 0.8424 - accuracy: 0.6921



35/92 [==========>……………….] - ETA: 3s - loss: 0.8367 - accuracy: 0.6920



36/92 [==========>……………….] - ETA: 3s - loss: 0.8349 - accuracy: 0.6936



37/92 [===========>………………] - ETA: 3s - loss: 0.8365 - accuracy: 0.6926



38/92 [===========>………………] - ETA: 3s - loss: 0.8451 - accuracy: 0.6891



39/92 [===========>………………] - ETA: 3s - loss: 0.8401 - accuracy: 0.6899



40/92 [============>……………..] - ETA: 3s - loss: 0.8397 - accuracy: 0.6891



41/92 [============>……………..] - ETA: 2s - loss: 0.8379 - accuracy: 0.6913



42/92 [============>……………..] - ETA: 2s - loss: 0.8459 - accuracy: 0.6868



43/92 [=============>…………….] - ETA: 2s - loss: 0.8410 - accuracy: 0.6860



44/92 [=============>…………….] - ETA: 2s - loss: 0.8352 - accuracy: 0.6889



46/92 [==============>……………] - ETA: 2s - loss: 0.8372 - accuracy: 0.6872



47/92 [==============>……………] - ETA: 2s - loss: 0.8341 - accuracy: 0.6885



48/92 [==============>……………] - ETA: 2s - loss: 0.8279 - accuracy: 0.6918



49/92 [==============>……………] - ETA: 2s - loss: 0.8294 - accuracy: 0.6917



50/92 [===============>…………..] - ETA: 2s - loss: 0.8300 - accuracy: 0.6928



51/92 [===============>…………..] - ETA: 2s - loss: 0.8275 - accuracy: 0.6940



52/92 [===============>…………..] - ETA: 2s - loss: 0.8267 - accuracy: 0.6944



53/92 [================>………….] - ETA: 2s - loss: 0.8255 - accuracy: 0.6961



54/92 [================>………….] - ETA: 2s - loss: 0.8220 - accuracy: 0.6977



55/92 [================>………….] - ETA: 2s - loss: 0.8198 - accuracy: 0.6975



56/92 [=================>…………] - ETA: 2s - loss: 0.8172 - accuracy: 0.6979



57/92 [=================>…………] - ETA: 2s - loss: 0.8155 - accuracy: 0.6982



58/92 [=================>…………] - ETA: 1s - loss: 0.8128 - accuracy: 0.6997



59/92 [==================>………..] - ETA: 1s - loss: 0.8129 - accuracy: 0.7000



60/92 [==================>………..] - ETA: 1s - loss: 0.8170 - accuracy: 0.6987



61/92 [==================>………..] - ETA: 1s - loss: 0.8182 - accuracy: 0.6980



62/92 [===================>……….] - ETA: 1s - loss: 0.8189 - accuracy: 0.6964



63/92 [===================>……….] - ETA: 1s - loss: 0.8193 - accuracy: 0.6952



64/92 [===================>……….] - ETA: 1s - loss: 0.8194 - accuracy: 0.6961



65/92 [====================>………] - ETA: 1s - loss: 0.8215 - accuracy: 0.6955



66/92 [====================>………] - ETA: 1s - loss: 0.8210 - accuracy: 0.6953



67/92 [====================>………] - ETA: 1s - loss: 0.8223 - accuracy: 0.6952



68/92 [=====================>……..] - ETA: 1s - loss: 0.8207 - accuracy: 0.6960



69/92 [=====================>……..] - ETA: 1s - loss: 0.8182 - accuracy: 0.6977



70/92 [=====================>……..] - ETA: 1s - loss: 0.8195 - accuracy: 0.6962



71/92 [======================>…….] - ETA: 1s - loss: 0.8194 - accuracy: 0.6961



72/92 [======================>…….] - ETA: 1s - loss: 0.8208 - accuracy: 0.6956



73/92 [======================>…….] - ETA: 1s - loss: 0.8190 - accuracy: 0.6963



74/92 [=======================>……] - ETA: 1s - loss: 0.8170 - accuracy: 0.6970



75/92 [=======================>……] - ETA: 0s - loss: 0.8158 - accuracy: 0.6977



76/92 [=======================>……] - ETA: 0s - loss: 0.8179 - accuracy: 0.6955



77/92 [========================>…..] - ETA: 0s - loss: 0.8176 - accuracy: 0.6958



78/92 [========================>…..] - ETA: 0s - loss: 0.8175 - accuracy: 0.6957



79/92 [========================>…..] - ETA: 0s - loss: 0.8188 - accuracy: 0.6952



80/92 [=========================>….] - ETA: 0s - loss: 0.8172 - accuracy: 0.6959



81/92 [=========================>….] - ETA: 0s - loss: 0.8217 - accuracy: 0.6939



82/92 [=========================>….] - ETA: 0s - loss: 0.8197 - accuracy: 0.6950



83/92 [==========================>…] - ETA: 0s - loss: 0.8180 - accuracy: 0.6952



84/92 [==========================>…] - ETA: 0s - loss: 0.8183 - accuracy: 0.6951



85/92 [==========================>…] - ETA: 0s - loss: 0.8190 - accuracy: 0.6947



86/92 [===========================>..] - ETA: 0s - loss: 0.8208 - accuracy: 0.6928



87/92 [===========================>..] - ETA: 0s - loss: 0.8207 - accuracy: 0.6931



88/92 [===========================>..] - ETA: 0s - loss: 0.8205 - accuracy: 0.6930



89/92 [============================>.] - ETA: 0s - loss: 0.8188 - accuracy: 0.6937



90/92 [============================>.] - ETA: 0s - loss: 0.8180 - accuracy: 0.6939



91/92 [============================>.] - ETA: 0s - loss: 0.8167 - accuracy: 0.6946



92/92 [==============================] - ETA: 0s - loss: 0.8158 - accuracy: 0.6945



92/92 [==============================] - 6s 64ms/step - loss: 0.8158 - accuracy: 0.6945 - val_loss: 0.8530 - val_accuracy: 0.6757

Epoch 5/15
1/92 [..............................] - ETA: 7s - loss: 0.8907 - accuracy: 0.7188
   
2/92 [..............................] - ETA: 5s - loss: 0.8773 - accuracy: 0.6875
   
3/92 [..............................] - ETA: 5s - loss: 0.8330 - accuracy: 0.6771
   
4/92 [>.............................] - ETA: 5s - loss: 0.7960 - accuracy: 0.6953
   
5/92 [>.............................] - ETA: 5s - loss: 0.8390 - accuracy: 0.6812
   
6/92 [>.............................] - ETA: 5s - loss: 0.8144 - accuracy: 0.6771
   
7/92 [=>............................] - ETA: 5s - loss: 0.8024 - accuracy: 0.6920
   
8/92 [=>............................] - ETA: 4s - loss: 0.8119 - accuracy: 0.6914
   
9/92 [=>............................] - ETA: 4s - loss: 0.8164 - accuracy: 0.6875


10/92 [==>………………………] - ETA: 4s - loss: 0.7930 - accuracy: 0.7000



11/92 [==>………………………] - ETA: 4s - loss: 0.7694 - accuracy: 0.7102



12/92 [==>………………………] - ETA: 4s - loss: 0.7519 - accuracy: 0.7161



13/92 [===>……………………..] - ETA: 4s - loss: 0.7302 - accuracy: 0.7260



14/92 [===>……………………..] - ETA: 4s - loss: 0.7293 - accuracy: 0.7210



15/92 [===>……………………..] - ETA: 4s - loss: 0.7256 - accuracy: 0.7208



16/92 [====>…………………….] - ETA: 4s - loss: 0.7320 - accuracy: 0.7207



17/92 [====>…………………….] - ETA: 4s - loss: 0.7327 - accuracy: 0.7243



18/92 [====>…………………….] - ETA: 4s - loss: 0.7351 - accuracy: 0.7205



19/92 [=====>……………………] - ETA: 4s - loss: 0.7352 - accuracy: 0.7204



20/92 [=====>……………………] - ETA: 4s - loss: 0.7393 - accuracy: 0.7203



21/92 [=====>……………………] - ETA: 4s - loss: 0.7438 - accuracy: 0.7217



22/92 [======>…………………..] - ETA: 4s - loss: 0.7462 - accuracy: 0.7216



23/92 [======>…………………..] - ETA: 3s - loss: 0.7597 - accuracy: 0.7147



24/92 [======>…………………..] - ETA: 3s - loss: 0.7511 - accuracy: 0.7188



25/92 [=======>………………….] - ETA: 3s - loss: 0.7569 - accuracy: 0.7150



26/92 [=======>………………….] - ETA: 3s - loss: 0.7474 - accuracy: 0.7175



27/92 [=======>………………….] - ETA: 3s - loss: 0.7583 - accuracy: 0.7141



28/92 [========>…………………] - ETA: 3s - loss: 0.7548 - accuracy: 0.7154



29/92 [========>…………………] - ETA: 3s - loss: 0.7512 - accuracy: 0.7155



30/92 [========>…………………] - ETA: 3s - loss: 0.7478 - accuracy: 0.7156



31/92 [=========>………………..] - ETA: 3s - loss: 0.7464 - accuracy: 0.7157



32/92 [=========>………………..] - ETA: 3s - loss: 0.7567 - accuracy: 0.7129



33/92 [=========>………………..] - ETA: 3s - loss: 0.7518 - accuracy: 0.7159



34/92 [==========>……………….] - ETA: 3s - loss: 0.7569 - accuracy: 0.7160



35/92 [==========>……………….] - ETA: 3s - loss: 0.7563 - accuracy: 0.7152



36/92 [==========>……………….] - ETA: 3s - loss: 0.7563 - accuracy: 0.7153



37/92 [===========>………………] - ETA: 3s - loss: 0.7607 - accuracy: 0.7111



38/92 [===========>………………] - ETA: 3s - loss: 0.7650 - accuracy: 0.7072



39/92 [===========>………………] - ETA: 3s - loss: 0.7642 - accuracy: 0.7083



40/92 [============>……………..] - ETA: 3s - loss: 0.7700 - accuracy: 0.7078



41/92 [============>……………..] - ETA: 2s - loss: 0.7711 - accuracy: 0.7096



42/92 [============>……………..] - ETA: 2s - loss: 0.7661 - accuracy: 0.7128



43/92 [=============>…………….] - ETA: 2s - loss: 0.7633 - accuracy: 0.7129



44/92 [=============>…………….] - ETA: 2s - loss: 0.7632 - accuracy: 0.7124



45/92 [=============>…………….] - ETA: 2s - loss: 0.7625 - accuracy: 0.7125



46/92 [==============>……………] - ETA: 2s - loss: 0.7612 - accuracy: 0.7120



47/92 [==============>……………] - ETA: 2s - loss: 0.7590 - accuracy: 0.7108



48/92 [==============>……………] - ETA: 2s - loss: 0.7586 - accuracy: 0.7103



49/92 [==============>……………] - ETA: 2s - loss: 0.7561 - accuracy: 0.7111



50/92 [===============>…………..] - ETA: 2s - loss: 0.7645 - accuracy: 0.7050



52/92 [===============>…………..] - ETA: 2s - loss: 0.7645 - accuracy: 0.7047



53/92 [================>………….] - ETA: 2s - loss: 0.7692 - accuracy: 0.7038



54/92 [================>………….] - ETA: 2s - loss: 0.7705 - accuracy: 0.7035



55/92 [================>………….] - ETA: 2s - loss: 0.7802 - accuracy: 0.6986



56/92 [=================>…………] - ETA: 2s - loss: 0.7780 - accuracy: 0.6990



57/92 [=================>…………] - ETA: 2s - loss: 0.7770 - accuracy: 0.6982



58/92 [=================>…………] - ETA: 1s - loss: 0.7749 - accuracy: 0.6997



59/92 [==================>………..] - ETA: 1s - loss: 0.7784 - accuracy: 0.6984



60/92 [==================>………..] - ETA: 1s - loss: 0.7787 - accuracy: 0.6967



61/92 [==================>………..] - ETA: 1s - loss: 0.7795 - accuracy: 0.6955



62/92 [===================>……….] - ETA: 1s - loss: 0.7780 - accuracy: 0.6964



63/92 [===================>……….] - ETA: 1s - loss: 0.7771 - accuracy: 0.6977



64/92 [===================>……….] - ETA: 1s - loss: 0.7814 - accuracy: 0.6961



65/92 [====================>………] - ETA: 1s - loss: 0.7836 - accuracy: 0.6950



66/92 [====================>………] - ETA: 1s - loss: 0.7805 - accuracy: 0.6968



67/92 [====================>………] - ETA: 1s - loss: 0.7795 - accuracy: 0.6966



68/92 [=====================>……..] - ETA: 1s - loss: 0.7841 - accuracy: 0.6933



69/92 [=====================>……..] - ETA: 1s - loss: 0.7849 - accuracy: 0.6932



70/92 [=====================>……..] - ETA: 1s - loss: 0.7907 - accuracy: 0.6927



71/92 [======================>…….] - ETA: 1s - loss: 0.7915 - accuracy: 0.6917



72/92 [======================>…….] - ETA: 1s - loss: 0.7909 - accuracy: 0.6916



73/92 [======================>…….] - ETA: 1s - loss: 0.7939 - accuracy: 0.6903



74/92 [=======================>……] - ETA: 1s - loss: 0.7967 - accuracy: 0.6890



75/92 [=======================>……] - ETA: 0s - loss: 0.7993 - accuracy: 0.6873



76/92 [=======================>……] - ETA: 0s - loss: 0.8003 - accuracy: 0.6861



77/92 [========================>…..] - ETA: 0s - loss: 0.8013 - accuracy: 0.6853



78/92 [========================>…..] - ETA: 0s - loss: 0.8004 - accuracy: 0.6853



79/92 [========================>…..] - ETA: 0s - loss: 0.7983 - accuracy: 0.6861



80/92 [=========================>….] - ETA: 0s - loss: 0.8004 - accuracy: 0.6846



81/92 [=========================>….] - ETA: 0s - loss: 0.8001 - accuracy: 0.6854



82/92 [=========================>….] - ETA: 0s - loss: 0.7997 - accuracy: 0.6858



83/92 [==========================>…] - ETA: 0s - loss: 0.7988 - accuracy: 0.6869



84/92 [==========================>…] - ETA: 0s - loss: 0.7983 - accuracy: 0.6869



85/92 [==========================>…] - ETA: 0s - loss: 0.7975 - accuracy: 0.6881



86/92 [===========================>..] - ETA: 0s - loss: 0.7958 - accuracy: 0.6891



87/92 [===========================>..] - ETA: 0s - loss: 0.7945 - accuracy: 0.6902



88/92 [===========================>..] - ETA: 0s - loss: 0.7928 - accuracy: 0.6909



89/92 [============================>.] - ETA: 0s - loss: 0.7910 - accuracy: 0.6919



90/92 [============================>.] - ETA: 0s - loss: 0.7917 - accuracy: 0.6915



91/92 [============================>.] - ETA: 0s - loss: 0.7883 - accuracy: 0.6932



92/92 [==============================] - ETA: 0s - loss: 0.7896 - accuracy: 0.6931



92/92 [==============================] - 6s 63ms/step - loss: 0.7896 - accuracy: 0.6931 - val_loss: 0.8867 - val_accuracy: 0.6798

Epoch 6/15
1/92 [..............................] - ETA: 6s - loss: 0.5518 - accuracy: 0.7812
   
2/92 [..............................] - ETA: 5s - loss: 0.7630 - accuracy: 0.7500
   
3/92 [..............................] - ETA: 5s - loss: 0.7584 - accuracy: 0.7604
   
4/92 [>.............................] - ETA: 5s - loss: 0.7502 - accuracy: 0.7266
   
5/92 [>.............................] - ETA: 5s - loss: 0.7661 - accuracy: 0.7125
   
6/92 [>.............................] - ETA: 4s - loss: 0.7633 - accuracy: 0.7083
   
7/92 [=>............................] - ETA: 4s - loss: 0.7836 - accuracy: 0.7054
   
8/92 [=>............................] - ETA: 4s - loss: 0.7776 - accuracy: 0.7070
   
9/92 [=>............................] - ETA: 4s - loss: 0.7611 - accuracy: 0.7153


10/92 [==>………………………] - ETA: 4s - loss: 0.7587 - accuracy: 0.7125



11/92 [==>………………………] - ETA: 4s - loss: 0.7458 - accuracy: 0.7244



12/92 [==>………………………] - ETA: 4s - loss: 0.7555 - accuracy: 0.7266



13/92 [===>……………………..] - ETA: 4s - loss: 0.7522 - accuracy: 0.7308



14/92 [===>……………………..] - ETA: 4s - loss: 0.7398 - accuracy: 0.7277



15/92 [===>……………………..] - ETA: 4s - loss: 0.7376 - accuracy: 0.7312



16/92 [====>…………………….] - ETA: 4s - loss: 0.7344 - accuracy: 0.7285



17/92 [====>…………………….] - ETA: 4s - loss: 0.7325 - accuracy: 0.7298



18/92 [====>…………………….] - ETA: 4s - loss: 0.7301 - accuracy: 0.7292



19/92 [=====>……………………] - ETA: 4s - loss: 0.7525 - accuracy: 0.7237



20/92 [=====>……………………] - ETA: 4s - loss: 0.7644 - accuracy: 0.7188



21/92 [=====>……………………] - ETA: 4s - loss: 0.7723 - accuracy: 0.7158



22/92 [======>…………………..] - ETA: 4s - loss: 0.7636 - accuracy: 0.7202



23/92 [======>…………………..] - ETA: 3s - loss: 0.7545 - accuracy: 0.7228



24/92 [======>…………………..] - ETA: 3s - loss: 0.7522 - accuracy: 0.7227



25/92 [=======>………………….] - ETA: 3s - loss: 0.7536 - accuracy: 0.7200



26/92 [=======>………………….] - ETA: 3s - loss: 0.7700 - accuracy: 0.7103



27/92 [=======>………………….] - ETA: 3s - loss: 0.7697 - accuracy: 0.7106



28/92 [========>…………………] - ETA: 3s - loss: 0.7738 - accuracy: 0.7121



29/92 [========>…………………] - ETA: 3s - loss: 0.7679 - accuracy: 0.7155



30/92 [========>…………………] - ETA: 3s - loss: 0.7772 - accuracy: 0.7104



31/92 [=========>………………..] - ETA: 3s - loss: 0.7790 - accuracy: 0.7107



32/92 [=========>………………..] - ETA: 3s - loss: 0.7822 - accuracy: 0.7100



33/92 [=========>………………..] - ETA: 3s - loss: 0.7819 - accuracy: 0.7112



34/92 [==========>……………….] - ETA: 3s - loss: 0.7831 - accuracy: 0.7086



35/92 [==========>……………….] - ETA: 3s - loss: 0.7782 - accuracy: 0.7089



36/92 [==========>……………….] - ETA: 3s - loss: 0.7777 - accuracy: 0.7083



37/92 [===========>………………] - ETA: 3s - loss: 0.7794 - accuracy: 0.7078



38/92 [===========>………………] - ETA: 3s - loss: 0.7856 - accuracy: 0.7064



39/92 [===========>………………] - ETA: 3s - loss: 0.7847 - accuracy: 0.7067



40/92 [============>……………..] - ETA: 3s - loss: 0.7862 - accuracy: 0.7047



41/92 [============>……………..] - ETA: 2s - loss: 0.7830 - accuracy: 0.7058



42/92 [============>……………..] - ETA: 2s - loss: 0.7793 - accuracy: 0.7068



43/92 [=============>…………….] - ETA: 2s - loss: 0.7756 - accuracy: 0.7086



44/92 [=============>…………….] - ETA: 2s - loss: 0.7727 - accuracy: 0.7095



45/92 [=============>…………….] - ETA: 2s - loss: 0.7714 - accuracy: 0.7076



46/92 [==============>……………] - ETA: 2s - loss: 0.7723 - accuracy: 0.7079



47/92 [==============>……………] - ETA: 2s - loss: 0.7696 - accuracy: 0.7094



48/92 [==============>……………] - ETA: 2s - loss: 0.7686 - accuracy: 0.7109



49/92 [==============>……………] - ETA: 2s - loss: 0.7657 - accuracy: 0.7136



50/92 [===============>…………..] - ETA: 2s - loss: 0.7702 - accuracy: 0.7119



51/92 [===============>…………..] - ETA: 2s - loss: 0.7732 - accuracy: 0.7126



52/92 [===============>…………..] - ETA: 2s - loss: 0.7772 - accuracy: 0.7109



53/92 [================>………….] - ETA: 2s - loss: 0.7780 - accuracy: 0.7099



54/92 [================>………….] - ETA: 2s - loss: 0.7742 - accuracy: 0.7124



55/92 [================>………….] - ETA: 2s - loss: 0.7721 - accuracy: 0.7136



56/92 [=================>…………] - ETA: 2s - loss: 0.7717 - accuracy: 0.7132



57/92 [=================>…………] - ETA: 2s - loss: 0.7688 - accuracy: 0.7138



58/92 [=================>…………] - ETA: 1s - loss: 0.7733 - accuracy: 0.7101



59/92 [==================>………..] - ETA: 1s - loss: 0.7728 - accuracy: 0.7119



60/92 [==================>………..] - ETA: 1s - loss: 0.7746 - accuracy: 0.7109



61/92 [==================>………..] - ETA: 1s - loss: 0.7697 - accuracy: 0.7126



62/92 [===================>……….] - ETA: 1s - loss: 0.7746 - accuracy: 0.7117



63/92 [===================>……….] - ETA: 1s - loss: 0.7725 - accuracy: 0.7123



64/92 [===================>……….] - ETA: 1s - loss: 0.7711 - accuracy: 0.7134



65/92 [====================>………] - ETA: 1s - loss: 0.7713 - accuracy: 0.7130



66/92 [====================>………] - ETA: 1s - loss: 0.7690 - accuracy: 0.7135



67/92 [====================>………] - ETA: 1s - loss: 0.7682 - accuracy: 0.7141



68/92 [=====================>……..] - ETA: 1s - loss: 0.7670 - accuracy: 0.7146



69/92 [=====================>……..] - ETA: 1s - loss: 0.7655 - accuracy: 0.7156



70/92 [=====================>……..] - ETA: 1s - loss: 0.7639 - accuracy: 0.7161



71/92 [======================>…….] - ETA: 1s - loss: 0.7664 - accuracy: 0.7139



72/92 [======================>…….] - ETA: 1s - loss: 0.7650 - accuracy: 0.7153



73/92 [======================>…….] - ETA: 1s - loss: 0.7669 - accuracy: 0.7132



74/92 [=======================>……] - ETA: 1s - loss: 0.7677 - accuracy: 0.7124



76/92 [=======================>……] - ETA: 0s - loss: 0.7683 - accuracy: 0.7120



77/92 [========================>…..] - ETA: 0s - loss: 0.7676 - accuracy: 0.7113



78/92 [========================>…..] - ETA: 0s - loss: 0.7679 - accuracy: 0.7122



79/92 [========================>…..] - ETA: 0s - loss: 0.7655 - accuracy: 0.7127



80/92 [=========================>….] - ETA: 0s - loss: 0.7639 - accuracy: 0.7132



81/92 [=========================>….] - ETA: 0s - loss: 0.7643 - accuracy: 0.7136



82/92 [=========================>….] - ETA: 0s - loss: 0.7672 - accuracy: 0.7122



83/92 [==========================>…] - ETA: 0s - loss: 0.7681 - accuracy: 0.7115



84/92 [==========================>…] - ETA: 0s - loss: 0.7652 - accuracy: 0.7131



85/92 [==========================>…] - ETA: 0s - loss: 0.7664 - accuracy: 0.7124



86/92 [===========================>..] - ETA: 0s - loss: 0.7678 - accuracy: 0.7114



87/92 [===========================>..] - ETA: 0s - loss: 0.7660 - accuracy: 0.7118



88/92 [===========================>..] - ETA: 0s - loss: 0.7654 - accuracy: 0.7115



89/92 [============================>.] - ETA: 0s - loss: 0.7658 - accuracy: 0.7109



90/92 [============================>.] - ETA: 0s - loss: 0.7648 - accuracy: 0.7107



91/92 [============================>.] - ETA: 0s - loss: 0.7648 - accuracy: 0.7107



92/92 [==============================] - ETA: 0s - loss: 0.7647 - accuracy: 0.7115



92/92 [==============================] - 6s 64ms/step - loss: 0.7647 - accuracy: 0.7115 - val_loss: 0.7599 - val_accuracy: 0.7016

Epoch 7/15
1/92 [..............................] - ETA: 7s - loss: 0.4912 - accuracy: 0.8438
   
2/92 [..............................] - ETA: 5s - loss: 0.5197 - accuracy: 0.7812
   
3/92 [..............................] - ETA: 5s - loss: 0.6350 - accuracy: 0.7396
   
4/92 [>.............................] - ETA: 5s - loss: 0.6448 - accuracy: 0.7500
   
5/92 [>.............................] - ETA: 5s - loss: 0.6741 - accuracy: 0.7375
   
6/92 [>.............................] - ETA: 5s - loss: 0.7069 - accuracy: 0.7344
   
7/92 [=>............................] - ETA: 5s - loss: 0.7105 - accuracy: 0.7321
   
8/92 [=>............................] - ETA: 5s - loss: 0.7082 - accuracy: 0.7344
   
9/92 [=>............................] - ETA: 4s - loss: 0.7131 - accuracy: 0.7326


10/92 [==>………………………] - ETA: 4s - loss: 0.7040 - accuracy: 0.7312



11/92 [==>………………………] - ETA: 4s - loss: 0.7117 - accuracy: 0.7244



12/92 [==>………………………] - ETA: 4s - loss: 0.7376 - accuracy: 0.7161



13/92 [===>……………………..] - ETA: 4s - loss: 0.7223 - accuracy: 0.7236



14/92 [===>……………………..] - ETA: 4s - loss: 0.7167 - accuracy: 0.7210



15/92 [===>……………………..] - ETA: 4s - loss: 0.7110 - accuracy: 0.7250



16/92 [====>…………………….] - ETA: 4s - loss: 0.6943 - accuracy: 0.7324



17/92 [====>…………………….] - ETA: 4s - loss: 0.6881 - accuracy: 0.7335



18/92 [====>…………………….] - ETA: 4s - loss: 0.6882 - accuracy: 0.7326



19/92 [=====>……………………] - ETA: 4s - loss: 0.6898 - accuracy: 0.7319



20/92 [=====>……………………] - ETA: 4s - loss: 0.6850 - accuracy: 0.7328



21/92 [=====>……………………] - ETA: 4s - loss: 0.6983 - accuracy: 0.7292



22/92 [======>…………………..] - ETA: 4s - loss: 0.6962 - accuracy: 0.7301



23/92 [======>…………………..] - ETA: 4s - loss: 0.6905 - accuracy: 0.7323



24/92 [======>…………………..] - ETA: 4s - loss: 0.6827 - accuracy: 0.7370



25/92 [=======>………………….] - ETA: 3s - loss: 0.6814 - accuracy: 0.7350



26/92 [=======>………………….] - ETA: 3s - loss: 0.6826 - accuracy: 0.7332



27/92 [=======>………………….] - ETA: 3s - loss: 0.6718 - accuracy: 0.7396



28/92 [========>…………………] - ETA: 3s - loss: 0.6691 - accuracy: 0.7388



29/92 [========>…………………] - ETA: 3s - loss: 0.6769 - accuracy: 0.7349



30/92 [========>…………………] - ETA: 3s - loss: 0.6747 - accuracy: 0.7365



31/92 [=========>………………..] - ETA: 3s - loss: 0.6848 - accuracy: 0.7339



32/92 [=========>………………..] - ETA: 3s - loss: 0.6793 - accuracy: 0.7383



33/92 [=========>………………..] - ETA: 3s - loss: 0.6826 - accuracy: 0.7377



34/92 [==========>……………….] - ETA: 3s - loss: 0.6777 - accuracy: 0.7408



35/92 [==========>……………….] - ETA: 3s - loss: 0.6865 - accuracy: 0.7357



36/92 [==========>……………….] - ETA: 3s - loss: 0.6899 - accuracy: 0.7352



37/92 [===========>………………] - ETA: 3s - loss: 0.6925 - accuracy: 0.7340



38/92 [===========>………………] - ETA: 3s - loss: 0.6955 - accuracy: 0.7327



39/92 [===========>………………] - ETA: 3s - loss: 0.6942 - accuracy: 0.7324



40/92 [============>……………..] - ETA: 3s - loss: 0.7007 - accuracy: 0.7297



41/92 [============>……………..] - ETA: 2s - loss: 0.7064 - accuracy: 0.7264



42/92 [============>……………..] - ETA: 2s - loss: 0.7118 - accuracy: 0.7262



43/92 [=============>…………….] - ETA: 2s - loss: 0.7098 - accuracy: 0.7267



44/92 [=============>…………….] - ETA: 2s - loss: 0.7088 - accuracy: 0.7294



45/92 [=============>…………….] - ETA: 2s - loss: 0.7112 - accuracy: 0.7292



46/92 [==============>……………] - ETA: 2s - loss: 0.7105 - accuracy: 0.7283



47/92 [==============>……………] - ETA: 2s - loss: 0.7076 - accuracy: 0.7294



48/92 [==============>……………] - ETA: 2s - loss: 0.7085 - accuracy: 0.7272



49/92 [==============>……………] - ETA: 2s - loss: 0.7099 - accuracy: 0.7277



50/92 [===============>…………..] - ETA: 2s - loss: 0.7086 - accuracy: 0.7287



51/92 [===============>…………..] - ETA: 2s - loss: 0.7087 - accuracy: 0.7298



52/92 [===============>…………..] - ETA: 2s - loss: 0.7098 - accuracy: 0.7302



53/92 [================>………….] - ETA: 2s - loss: 0.7106 - accuracy: 0.7300



54/92 [================>………….] - ETA: 2s - loss: 0.7093 - accuracy: 0.7315



55/92 [================>………….] - ETA: 2s - loss: 0.7085 - accuracy: 0.7318



56/92 [=================>…………] - ETA: 2s - loss: 0.7107 - accuracy: 0.7299



57/92 [=================>…………] - ETA: 2s - loss: 0.7082 - accuracy: 0.7308



58/92 [=================>…………] - ETA: 1s - loss: 0.7098 - accuracy: 0.7295



59/92 [==================>………..] - ETA: 1s - loss: 0.7162 - accuracy: 0.7251



60/92 [==================>………..] - ETA: 1s - loss: 0.7135 - accuracy: 0.7260



61/92 [==================>………..] - ETA: 1s - loss: 0.7132 - accuracy: 0.7259



62/92 [===================>……….] - ETA: 1s - loss: 0.7137 - accuracy: 0.7263



63/92 [===================>……….] - ETA: 1s - loss: 0.7139 - accuracy: 0.7257



64/92 [===================>……….] - ETA: 1s - loss: 0.7139 - accuracy: 0.7251



65/92 [====================>………] - ETA: 1s - loss: 0.7107 - accuracy: 0.7264



66/92 [====================>………] - ETA: 1s - loss: 0.7091 - accuracy: 0.7268



67/92 [====================>………] - ETA: 1s - loss: 0.7080 - accuracy: 0.7276



68/92 [=====================>……..] - ETA: 1s - loss: 0.7085 - accuracy: 0.7275



69/92 [=====================>……..] - ETA: 1s - loss: 0.7085 - accuracy: 0.7274



70/92 [=====================>……..] - ETA: 1s - loss: 0.7086 - accuracy: 0.7277



71/92 [======================>…….] - ETA: 1s - loss: 0.7068 - accuracy: 0.7293



72/92 [======================>…….] - ETA: 1s - loss: 0.7058 - accuracy: 0.7300



74/92 [=======================>……] - ETA: 1s - loss: 0.7033 - accuracy: 0.7292



75/92 [=======================>……] - ETA: 0s - loss: 0.7020 - accuracy: 0.7304



76/92 [=======================>……] - ETA: 0s - loss: 0.7010 - accuracy: 0.7306



77/92 [========================>…..] - ETA: 0s - loss: 0.6969 - accuracy: 0.7325



78/92 [========================>…..] - ETA: 0s - loss: 0.6963 - accuracy: 0.7327



79/92 [========================>…..] - ETA: 0s - loss: 0.6964 - accuracy: 0.7329



80/92 [=========================>….] - ETA: 0s - loss: 0.6943 - accuracy: 0.7343



81/92 [=========================>….] - ETA: 0s - loss: 0.6968 - accuracy: 0.7345



82/92 [=========================>….] - ETA: 0s - loss: 0.6975 - accuracy: 0.7343



83/92 [==========================>…] - ETA: 0s - loss: 0.7000 - accuracy: 0.7330



84/92 [==========================>…] - ETA: 0s - loss: 0.7012 - accuracy: 0.7328



85/92 [==========================>…] - ETA: 0s - loss: 0.6997 - accuracy: 0.7334



86/92 [===========================>..] - ETA: 0s - loss: 0.6986 - accuracy: 0.7336



87/92 [===========================>..] - ETA: 0s - loss: 0.6983 - accuracy: 0.7338



88/92 [===========================>..] - ETA: 0s - loss: 0.6983 - accuracy: 0.7340



89/92 [============================>.] - ETA: 0s - loss: 0.6967 - accuracy: 0.7338



90/92 [============================>.] - ETA: 0s - loss: 0.6940 - accuracy: 0.7357



91/92 [============================>.] - ETA: 0s - loss: 0.6932 - accuracy: 0.7362



92/92 [==============================] - ETA: 0s - loss: 0.6932 - accuracy: 0.7360



92/92 [==============================] - 6s 64ms/step - loss: 0.6932 - accuracy: 0.7360 - val_loss: 0.7731 - val_accuracy: 0.6853

Epoch 8/15
1/92 [..............................] - ETA: 7s - loss: 0.5886 - accuracy: 0.7500
   
2/92 [..............................] - ETA: 5s - loss: 0.6209 - accuracy: 0.7500
   
3/92 [..............................] - ETA: 5s - loss: 0.6834 - accuracy: 0.7396
   
4/92 [>.............................] - ETA: 5s - loss: 0.6812 - accuracy: 0.7266
   
5/92 [>.............................] - ETA: 5s - loss: 0.6540 - accuracy: 0.7312
   
6/92 [>.............................] - ETA: 5s - loss: 0.6633 - accuracy: 0.7240
   
7/92 [=>............................] - ETA: 4s - loss: 0.6415 - accuracy: 0.7411
   
8/92 [=>............................] - ETA: 4s - loss: 0.6256 - accuracy: 0.7539
   
9/92 [=>............................] - ETA: 4s - loss: 0.5989 - accuracy: 0.7604


10/92 [==>………………………] - ETA: 4s - loss: 0.6047 - accuracy: 0.7500



11/92 [==>………………………] - ETA: 4s - loss: 0.6068 - accuracy: 0.7500



12/92 [==>………………………] - ETA: 4s - loss: 0.6024 - accuracy: 0.7500



13/92 [===>……………………..] - ETA: 4s - loss: 0.6013 - accuracy: 0.7548



14/92 [===>……………………..] - ETA: 4s - loss: 0.5978 - accuracy: 0.7545



15/92 [===>……………………..] - ETA: 4s - loss: 0.6020 - accuracy: 0.7479



16/92 [====>…………………….] - ETA: 4s - loss: 0.5969 - accuracy: 0.7500



17/92 [====>…………………….] - ETA: 4s - loss: 0.6147 - accuracy: 0.7408



18/92 [====>…………………….] - ETA: 4s - loss: 0.6109 - accuracy: 0.7448



19/92 [=====>……………………] - ETA: 4s - loss: 0.6156 - accuracy: 0.7467



20/92 [=====>……………………] - ETA: 4s - loss: 0.6125 - accuracy: 0.7516



21/92 [=====>……………………] - ETA: 4s - loss: 0.6098 - accuracy: 0.7530



22/92 [======>…………………..] - ETA: 4s - loss: 0.6131 - accuracy: 0.7528



23/92 [======>…………………..] - ETA: 4s - loss: 0.6149 - accuracy: 0.7527



24/92 [======>…………………..] - ETA: 3s - loss: 0.6234 - accuracy: 0.7474



25/92 [=======>………………….] - ETA: 3s - loss: 0.6214 - accuracy: 0.7487



26/92 [=======>………………….] - ETA: 3s - loss: 0.6183 - accuracy: 0.7512



27/92 [=======>………………….] - ETA: 3s - loss: 0.6153 - accuracy: 0.7535



28/92 [========>…………………] - ETA: 3s - loss: 0.6128 - accuracy: 0.7556



29/92 [========>…………………] - ETA: 3s - loss: 0.6245 - accuracy: 0.7522



30/92 [========>…………………] - ETA: 3s - loss: 0.6246 - accuracy: 0.7510



31/92 [=========>………………..] - ETA: 3s - loss: 0.6228 - accuracy: 0.7530



32/92 [=========>………………..] - ETA: 3s - loss: 0.6247 - accuracy: 0.7539



33/92 [=========>………………..] - ETA: 3s - loss: 0.6282 - accuracy: 0.7509



34/92 [==========>……………….] - ETA: 3s - loss: 0.6380 - accuracy: 0.7491



35/92 [==========>……………….] - ETA: 3s - loss: 0.6360 - accuracy: 0.7500



36/92 [==========>……………….] - ETA: 3s - loss: 0.6347 - accuracy: 0.7526



37/92 [===========>………………] - ETA: 3s - loss: 0.6365 - accuracy: 0.7525



38/92 [===========>………………] - ETA: 3s - loss: 0.6336 - accuracy: 0.7549



39/92 [===========>………………] - ETA: 3s - loss: 0.6360 - accuracy: 0.7532



40/92 [============>……………..] - ETA: 3s - loss: 0.6311 - accuracy: 0.7547



41/92 [============>……………..] - ETA: 2s - loss: 0.6326 - accuracy: 0.7553



42/92 [============>……………..] - ETA: 2s - loss: 0.6372 - accuracy: 0.7545



43/92 [=============>…………….] - ETA: 2s - loss: 0.6439 - accuracy: 0.7536



44/92 [=============>…………….] - ETA: 2s - loss: 0.6455 - accuracy: 0.7521



46/92 [==============>……………] - ETA: 2s - loss: 0.6455 - accuracy: 0.7514



47/92 [==============>……………] - ETA: 2s - loss: 0.6464 - accuracy: 0.7520



48/92 [==============>……………] - ETA: 2s - loss: 0.6517 - accuracy: 0.7513



49/92 [==============>……………] - ETA: 2s - loss: 0.6550 - accuracy: 0.7513



50/92 [===============>…………..] - ETA: 2s - loss: 0.6535 - accuracy: 0.7519



51/92 [===============>…………..] - ETA: 2s - loss: 0.6543 - accuracy: 0.7525



52/92 [===============>…………..] - ETA: 2s - loss: 0.6592 - accuracy: 0.7482



53/92 [================>………….] - ETA: 2s - loss: 0.6656 - accuracy: 0.7447



54/92 [================>………….] - ETA: 2s - loss: 0.6636 - accuracy: 0.7459



55/92 [================>………….] - ETA: 2s - loss: 0.6646 - accuracy: 0.7454



56/92 [=================>…………] - ETA: 2s - loss: 0.6710 - accuracy: 0.7444



57/92 [=================>…………] - ETA: 2s - loss: 0.6695 - accuracy: 0.7450



58/92 [=================>…………] - ETA: 1s - loss: 0.6672 - accuracy: 0.7462



59/92 [==================>………..] - ETA: 1s - loss: 0.6702 - accuracy: 0.7452



60/92 [==================>………..] - ETA: 1s - loss: 0.6684 - accuracy: 0.7453



61/92 [==================>………..] - ETA: 1s - loss: 0.6693 - accuracy: 0.7454



62/92 [===================>……….] - ETA: 1s - loss: 0.6713 - accuracy: 0.7449



63/92 [===================>……….] - ETA: 1s - loss: 0.6722 - accuracy: 0.7440



64/92 [===================>……….] - ETA: 1s - loss: 0.6705 - accuracy: 0.7451



65/92 [====================>………] - ETA: 1s - loss: 0.6720 - accuracy: 0.7452



66/92 [====================>………] - ETA: 1s - loss: 0.6696 - accuracy: 0.7462



67/92 [====================>………] - ETA: 1s - loss: 0.6726 - accuracy: 0.7458



68/92 [=====================>……..] - ETA: 1s - loss: 0.6737 - accuracy: 0.7454



69/92 [=====================>……..] - ETA: 1s - loss: 0.6717 - accuracy: 0.7459



70/92 [=====================>……..] - ETA: 1s - loss: 0.6693 - accuracy: 0.7460



71/92 [======================>…….] - ETA: 1s - loss: 0.6691 - accuracy: 0.7460



72/92 [======================>…….] - ETA: 1s - loss: 0.6727 - accuracy: 0.7443



73/92 [======================>…….] - ETA: 1s - loss: 0.6732 - accuracy: 0.7440



74/92 [=======================>……] - ETA: 1s - loss: 0.6720 - accuracy: 0.7445



75/92 [=======================>……] - ETA: 0s - loss: 0.6715 - accuracy: 0.7458



76/92 [=======================>……] - ETA: 0s - loss: 0.6703 - accuracy: 0.7459



77/92 [========================>…..] - ETA: 0s - loss: 0.6712 - accuracy: 0.7455



78/92 [========================>…..] - ETA: 0s - loss: 0.6727 - accuracy: 0.7456



79/92 [========================>…..] - ETA: 0s - loss: 0.6713 - accuracy: 0.7460



80/92 [=========================>….] - ETA: 0s - loss: 0.6690 - accuracy: 0.7473



81/92 [=========================>….] - ETA: 0s - loss: 0.6698 - accuracy: 0.7457



82/92 [=========================>….] - ETA: 0s - loss: 0.6702 - accuracy: 0.7450



83/92 [==========================>…] - ETA: 0s - loss: 0.6702 - accuracy: 0.7455



84/92 [==========================>…] - ETA: 0s - loss: 0.6699 - accuracy: 0.7451



85/92 [==========================>…] - ETA: 0s - loss: 0.6690 - accuracy: 0.7463



86/92 [===========================>..] - ETA: 0s - loss: 0.6715 - accuracy: 0.7449



87/92 [===========================>..] - ETA: 0s - loss: 0.6732 - accuracy: 0.7439



88/92 [===========================>..] - ETA: 0s - loss: 0.6717 - accuracy: 0.7439



89/92 [============================>.] - ETA: 0s - loss: 0.6743 - accuracy: 0.7433



90/92 [============================>.] - ETA: 0s - loss: 0.6769 - accuracy: 0.7427



91/92 [============================>.] - ETA: 0s - loss: 0.6763 - accuracy: 0.7428



92/92 [==============================] - ETA: 0s - loss: 0.6821 - accuracy: 0.7398



92/92 [==============================] - 6s 64ms/step - loss: 0.6821 - accuracy: 0.7398 - val_loss: 0.7942 - val_accuracy: 0.6812

Epoch 9/15
1/92 [..............................] - ETA: 6s - loss: 0.5542 - accuracy: 0.7812
   
2/92 [..............................] - ETA: 5s - loss: 0.6572 - accuracy: 0.7500
   
3/92 [..............................] - ETA: 5s - loss: 0.6516 - accuracy: 0.7604
   
4/92 [>.............................] - ETA: 5s - loss: 0.6615 - accuracy: 0.7344
   
5/92 [>.............................] - ETA: 5s - loss: 0.6472 - accuracy: 0.7625
   
6/92 [>.............................] - ETA: 5s - loss: 0.6448 - accuracy: 0.7604
   
7/92 [=>............................] - ETA: 4s - loss: 0.6265 - accuracy: 0.7545
   
8/92 [=>............................] - ETA: 4s - loss: 0.6567 - accuracy: 0.7422
   
9/92 [=>............................] - ETA: 4s - loss: 0.6387 - accuracy: 0.7465


10/92 [==>………………………] - ETA: 4s - loss: 0.6596 - accuracy: 0.7406



11/92 [==>………………………] - ETA: 4s - loss: 0.6548 - accuracy: 0.7415



12/92 [==>………………………] - ETA: 4s - loss: 0.6587 - accuracy: 0.7370



13/92 [===>……………………..] - ETA: 4s - loss: 0.6604 - accuracy: 0.7356



14/92 [===>……………………..] - ETA: 4s - loss: 0.6458 - accuracy: 0.7433



15/92 [===>……………………..] - ETA: 4s - loss: 0.6432 - accuracy: 0.7479



16/92 [====>…………………….] - ETA: 4s - loss: 0.6368 - accuracy: 0.7500



17/92 [====>…………………….] - ETA: 4s - loss: 0.6318 - accuracy: 0.7574



18/92 [====>…………………….] - ETA: 4s - loss: 0.6263 - accuracy: 0.7604



19/92 [=====>……………………] - ETA: 4s - loss: 0.6300 - accuracy: 0.7599



20/92 [=====>……………………] - ETA: 4s - loss: 0.6349 - accuracy: 0.7578



21/92 [=====>……………………] - ETA: 4s - loss: 0.6289 - accuracy: 0.7589



22/92 [======>…………………..] - ETA: 4s - loss: 0.6234 - accuracy: 0.7614



23/92 [======>…………………..] - ETA: 4s - loss: 0.6200 - accuracy: 0.7609



24/92 [======>…………………..] - ETA: 3s - loss: 0.6159 - accuracy: 0.7630



25/92 [=======>………………….] - ETA: 3s - loss: 0.6139 - accuracy: 0.7625



26/92 [=======>………………….] - ETA: 3s - loss: 0.6118 - accuracy: 0.7620



27/92 [=======>………………….] - ETA: 3s - loss: 0.6134 - accuracy: 0.7581



28/92 [========>…………………] - ETA: 3s - loss: 0.6103 - accuracy: 0.7567



29/92 [========>…………………] - ETA: 3s - loss: 0.6040 - accuracy: 0.7586



30/92 [========>…………………] - ETA: 3s - loss: 0.6029 - accuracy: 0.7604



31/92 [=========>………………..] - ETA: 3s - loss: 0.6022 - accuracy: 0.7621



32/92 [=========>………………..] - ETA: 3s - loss: 0.6037 - accuracy: 0.7607



33/92 [=========>………………..] - ETA: 3s - loss: 0.6135 - accuracy: 0.7566



34/92 [==========>……………….] - ETA: 3s - loss: 0.6141 - accuracy: 0.7574



35/92 [==========>……………….] - ETA: 3s - loss: 0.6172 - accuracy: 0.7554



36/92 [==========>……………….] - ETA: 3s - loss: 0.6220 - accuracy: 0.7552



37/92 [===========>………………] - ETA: 3s - loss: 0.6279 - accuracy: 0.7525



38/92 [===========>………………] - ETA: 3s - loss: 0.6295 - accuracy: 0.7500



39/92 [===========>………………] - ETA: 3s - loss: 0.6328 - accuracy: 0.7492



40/92 [============>……………..] - ETA: 3s - loss: 0.6366 - accuracy: 0.7477



41/92 [============>……………..] - ETA: 2s - loss: 0.6435 - accuracy: 0.7470



42/92 [============>……………..] - ETA: 2s - loss: 0.6465 - accuracy: 0.7448



43/92 [=============>…………….] - ETA: 2s - loss: 0.6493 - accuracy: 0.7442



44/92 [=============>…………….] - ETA: 2s - loss: 0.6462 - accuracy: 0.7450



45/92 [=============>…………….] - ETA: 2s - loss: 0.6506 - accuracy: 0.7437



46/92 [==============>……………] - ETA: 2s - loss: 0.6490 - accuracy: 0.7432



47/92 [==============>……………] - ETA: 2s - loss: 0.6511 - accuracy: 0.7414



48/92 [==============>……………] - ETA: 2s - loss: 0.6528 - accuracy: 0.7415



49/92 [==============>……………] - ETA: 2s - loss: 0.6529 - accuracy: 0.7411



50/92 [===============>…………..] - ETA: 2s - loss: 0.6542 - accuracy: 0.7419



51/92 [===============>…………..] - ETA: 2s - loss: 0.6546 - accuracy: 0.7408



52/92 [===============>…………..] - ETA: 2s - loss: 0.6496 - accuracy: 0.7434



53/92 [================>………….] - ETA: 2s - loss: 0.6550 - accuracy: 0.7417



54/92 [================>………….] - ETA: 2s - loss: 0.6544 - accuracy: 0.7413



55/92 [================>………….] - ETA: 2s - loss: 0.6563 - accuracy: 0.7420



56/92 [=================>…………] - ETA: 2s - loss: 0.6577 - accuracy: 0.7416



57/92 [=================>…………] - ETA: 2s - loss: 0.6562 - accuracy: 0.7429



58/92 [=================>…………] - ETA: 1s - loss: 0.6543 - accuracy: 0.7441



59/92 [==================>………..] - ETA: 1s - loss: 0.6554 - accuracy: 0.7431



60/92 [==================>………..] - ETA: 1s - loss: 0.6605 - accuracy: 0.7401



61/92 [==================>………..] - ETA: 1s - loss: 0.6579 - accuracy: 0.7423



62/92 [===================>……….] - ETA: 1s - loss: 0.6568 - accuracy: 0.7414



64/92 [===================>……….] - ETA: 1s - loss: 0.6570 - accuracy: 0.7412



65/92 [====================>………] - ETA: 1s - loss: 0.6617 - accuracy: 0.7384



66/92 [====================>………] - ETA: 1s - loss: 0.6606 - accuracy: 0.7395



67/92 [====================>………] - ETA: 1s - loss: 0.6609 - accuracy: 0.7392



68/92 [=====================>……..] - ETA: 1s - loss: 0.6580 - accuracy: 0.7408



69/92 [=====================>……..] - ETA: 1s - loss: 0.6592 - accuracy: 0.7409



70/92 [=====================>……..] - ETA: 1s - loss: 0.6564 - accuracy: 0.7424



71/92 [======================>…….] - ETA: 1s - loss: 0.6546 - accuracy: 0.7434



72/92 [======================>…….] - ETA: 1s - loss: 0.6555 - accuracy: 0.7435



73/92 [======================>…….] - ETA: 1s - loss: 0.6536 - accuracy: 0.7444



74/92 [=======================>……] - ETA: 1s - loss: 0.6532 - accuracy: 0.7449



75/92 [=======================>……] - ETA: 0s - loss: 0.6585 - accuracy: 0.7425



76/92 [=======================>……] - ETA: 0s - loss: 0.6595 - accuracy: 0.7426



77/92 [========================>…..] - ETA: 0s - loss: 0.6558 - accuracy: 0.7447



78/92 [========================>…..] - ETA: 0s - loss: 0.6535 - accuracy: 0.7456



79/92 [========================>…..] - ETA: 0s - loss: 0.6509 - accuracy: 0.7472



80/92 [=========================>….] - ETA: 0s - loss: 0.6487 - accuracy: 0.7480



81/92 [=========================>….] - ETA: 0s - loss: 0.6491 - accuracy: 0.7473



82/92 [=========================>….] - ETA: 0s - loss: 0.6480 - accuracy: 0.7477



83/92 [==========================>…] - ETA: 0s - loss: 0.6493 - accuracy: 0.7481



84/92 [==========================>…] - ETA: 0s - loss: 0.6504 - accuracy: 0.7466



85/92 [==========================>…] - ETA: 0s - loss: 0.6491 - accuracy: 0.7474



86/92 [===========================>..] - ETA: 0s - loss: 0.6478 - accuracy: 0.7482



87/92 [===========================>..] - ETA: 0s - loss: 0.6478 - accuracy: 0.7486



88/92 [===========================>..] - ETA: 0s - loss: 0.6470 - accuracy: 0.7504



89/92 [============================>.] - ETA: 0s - loss: 0.6463 - accuracy: 0.7511



90/92 [============================>.] - ETA: 0s - loss: 0.6476 - accuracy: 0.7510



91/92 [============================>.] - ETA: 0s - loss: 0.6475 - accuracy: 0.7507



92/92 [==============================] - ETA: 0s - loss: 0.6469 - accuracy: 0.7510



92/92 [==============================] - 6s 63ms/step - loss: 0.6469 - accuracy: 0.7510 - val_loss: 0.7705 - val_accuracy: 0.6921

Epoch 10/15
1/92 [..............................] - ETA: 7s - loss: 0.5019 - accuracy: 0.8438
   
2/92 [..............................] - ETA: 5s - loss: 0.5805 - accuracy: 0.7969
   
3/92 [..............................] - ETA: 5s - loss: 0.6209 - accuracy: 0.7604
   
4/92 [>.............................] - ETA: 5s - loss: 0.6745 - accuracy: 0.7109
   
5/92 [>.............................] - ETA: 5s - loss: 0.6841 - accuracy: 0.7125
   
6/92 [>.............................] - ETA: 4s - loss: 0.6510 - accuracy: 0.7188
   
7/92 [=>............................] - ETA: 4s - loss: 0.6254 - accuracy: 0.7411
   
8/92 [=>............................] - ETA: 4s - loss: 0.6364 - accuracy: 0.7383
   
9/92 [=>............................] - ETA: 4s - loss: 0.6488 - accuracy: 0.7292


10/92 [==>………………………] - ETA: 4s - loss: 0.6263 - accuracy: 0.7406



11/92 [==>………………………] - ETA: 4s - loss: 0.6172 - accuracy: 0.7443



12/92 [==>………………………] - ETA: 4s - loss: 0.6176 - accuracy: 0.7422



13/92 [===>……………………..] - ETA: 4s - loss: 0.6043 - accuracy: 0.7452



14/92 [===>……………………..] - ETA: 4s - loss: 0.6265 - accuracy: 0.7433



15/92 [===>……………………..] - ETA: 4s - loss: 0.6181 - accuracy: 0.7479



16/92 [====>…………………….] - ETA: 4s - loss: 0.6257 - accuracy: 0.7520



17/92 [====>…………………….] - ETA: 4s - loss: 0.6240 - accuracy: 0.7574



18/92 [====>…………………….] - ETA: 4s - loss: 0.6256 - accuracy: 0.7535



19/92 [=====>……………………] - ETA: 4s - loss: 0.6189 - accuracy: 0.7566



20/92 [=====>……………………] - ETA: 4s - loss: 0.6213 - accuracy: 0.7578



21/92 [=====>……………………] - ETA: 4s - loss: 0.6196 - accuracy: 0.7589



22/92 [======>…………………..] - ETA: 4s - loss: 0.6144 - accuracy: 0.7642



23/92 [======>…………………..] - ETA: 3s - loss: 0.6133 - accuracy: 0.7649



24/92 [======>…………………..] - ETA: 3s - loss: 0.6115 - accuracy: 0.7669



25/92 [=======>………………….] - ETA: 3s - loss: 0.6141 - accuracy: 0.7638



26/92 [=======>………………….] - ETA: 3s - loss: 0.6078 - accuracy: 0.7656



27/92 [=======>………………….] - ETA: 3s - loss: 0.6107 - accuracy: 0.7639



28/92 [========>…………………] - ETA: 3s - loss: 0.6194 - accuracy: 0.7578



29/92 [========>…………………] - ETA: 3s - loss: 0.6195 - accuracy: 0.7575



30/92 [========>…………………] - ETA: 3s - loss: 0.6170 - accuracy: 0.7604



31/92 [=========>………………..] - ETA: 3s - loss: 0.6153 - accuracy: 0.7601



32/92 [=========>………………..] - ETA: 3s - loss: 0.6169 - accuracy: 0.7588



33/92 [=========>………………..] - ETA: 3s - loss: 0.6183 - accuracy: 0.7576



34/92 [==========>……………….] - ETA: 3s - loss: 0.6117 - accuracy: 0.7610



35/92 [==========>……………….] - ETA: 3s - loss: 0.6171 - accuracy: 0.7607



36/92 [==========>……………….] - ETA: 3s - loss: 0.6148 - accuracy: 0.7613



37/92 [===========>………………] - ETA: 3s - loss: 0.6160 - accuracy: 0.7601



38/92 [===========>………………] - ETA: 3s - loss: 0.6141 - accuracy: 0.7615



39/92 [===========>………………] - ETA: 3s - loss: 0.6149 - accuracy: 0.7612



40/92 [============>……………..] - ETA: 3s - loss: 0.6139 - accuracy: 0.7625



41/92 [============>……………..] - ETA: 2s - loss: 0.6141 - accuracy: 0.7630



42/92 [============>……………..] - ETA: 2s - loss: 0.6113 - accuracy: 0.7634



43/92 [=============>…………….] - ETA: 2s - loss: 0.6049 - accuracy: 0.7667



44/92 [=============>…………….] - ETA: 2s - loss: 0.6052 - accuracy: 0.7670



46/92 [==============>……………] - ETA: 2s - loss: 0.6061 - accuracy: 0.7678



47/92 [==============>……………] - ETA: 2s - loss: 0.6016 - accuracy: 0.7701



48/92 [==============>……………] - ETA: 2s - loss: 0.6008 - accuracy: 0.7703



49/92 [==============>……………] - ETA: 2s - loss: 0.6031 - accuracy: 0.7705



50/92 [===============>…………..] - ETA: 2s - loss: 0.6029 - accuracy: 0.7714



51/92 [===============>…………..] - ETA: 2s - loss: 0.6044 - accuracy: 0.7691



52/92 [===============>…………..] - ETA: 2s - loss: 0.6106 - accuracy: 0.7645



53/92 [================>………….] - ETA: 2s - loss: 0.6055 - accuracy: 0.7660



54/92 [================>………….] - ETA: 2s - loss: 0.6085 - accuracy: 0.7645



55/92 [================>………….] - ETA: 2s - loss: 0.6119 - accuracy: 0.7637



56/92 [=================>…………] - ETA: 2s - loss: 0.6117 - accuracy: 0.7640



57/92 [=================>…………] - ETA: 2s - loss: 0.6108 - accuracy: 0.7649



58/92 [=================>…………] - ETA: 1s - loss: 0.6119 - accuracy: 0.7641



59/92 [==================>………..] - ETA: 1s - loss: 0.6111 - accuracy: 0.7660



60/92 [==================>………..] - ETA: 1s - loss: 0.6074 - accuracy: 0.7667



61/92 [==================>………..] - ETA: 1s - loss: 0.6087 - accuracy: 0.7665



62/92 [===================>……….] - ETA: 1s - loss: 0.6106 - accuracy: 0.7672



63/92 [===================>……….] - ETA: 1s - loss: 0.6161 - accuracy: 0.7659



64/92 [===================>……….] - ETA: 1s - loss: 0.6146 - accuracy: 0.7672



65/92 [====================>………] - ETA: 1s - loss: 0.6118 - accuracy: 0.7688



66/92 [====================>………] - ETA: 1s - loss: 0.6131 - accuracy: 0.7681



67/92 [====================>………] - ETA: 1s - loss: 0.6130 - accuracy: 0.7678



68/92 [=====================>……..] - ETA: 1s - loss: 0.6126 - accuracy: 0.7680



69/92 [=====================>……..] - ETA: 1s - loss: 0.6167 - accuracy: 0.7655



70/92 [=====================>……..] - ETA: 1s - loss: 0.6165 - accuracy: 0.7652



71/92 [======================>…….] - ETA: 1s - loss: 0.6148 - accuracy: 0.7655



72/92 [======================>…….] - ETA: 1s - loss: 0.6191 - accuracy: 0.7635



73/92 [======================>…….] - ETA: 1s - loss: 0.6240 - accuracy: 0.7625



74/92 [=======================>……] - ETA: 1s - loss: 0.6252 - accuracy: 0.7614



75/92 [=======================>……] - ETA: 0s - loss: 0.6270 - accuracy: 0.7596



76/92 [=======================>……] - ETA: 0s - loss: 0.6247 - accuracy: 0.7603



77/92 [========================>…..] - ETA: 0s - loss: 0.6228 - accuracy: 0.7610



78/92 [========================>…..] - ETA: 0s - loss: 0.6224 - accuracy: 0.7609



79/92 [========================>…..] - ETA: 0s - loss: 0.6223 - accuracy: 0.7615



80/92 [=========================>….] - ETA: 0s - loss: 0.6221 - accuracy: 0.7614



81/92 [=========================>….] - ETA: 0s - loss: 0.6233 - accuracy: 0.7608



82/92 [=========================>….] - ETA: 0s - loss: 0.6214 - accuracy: 0.7615



83/92 [==========================>…] - ETA: 0s - loss: 0.6221 - accuracy: 0.7610



84/92 [==========================>…] - ETA: 0s - loss: 0.6238 - accuracy: 0.7616



85/92 [==========================>…] - ETA: 0s - loss: 0.6236 - accuracy: 0.7622



86/92 [===========================>..] - ETA: 0s - loss: 0.6234 - accuracy: 0.7624



87/92 [===========================>..] - ETA: 0s - loss: 0.6231 - accuracy: 0.7630



88/92 [===========================>..] - ETA: 0s - loss: 0.6217 - accuracy: 0.7632



89/92 [============================>.] - ETA: 0s - loss: 0.6214 - accuracy: 0.7641



90/92 [============================>.] - ETA: 0s - loss: 0.6248 - accuracy: 0.7639



91/92 [============================>.] - ETA: 0s - loss: 0.6238 - accuracy: 0.7645



92/92 [==============================] - ETA: 0s - loss: 0.6230 - accuracy: 0.7646



92/92 [==============================] - 6s 63ms/step - loss: 0.6230 - accuracy: 0.7646 - val_loss: 0.7725 - val_accuracy: 0.7153

Epoch 11/15
1/92 [..............................] - ETA: 7s - loss: 0.6668 - accuracy: 0.6875
   
2/92 [..............................] - ETA: 5s - loss: 0.5528 - accuracy: 0.7656
   
3/92 [..............................] - ETA: 5s - loss: 0.5535 - accuracy: 0.7917
   
4/92 [>.............................] - ETA: 5s - loss: 0.5296 - accuracy: 0.7969
   
5/92 [>.............................] - ETA: 5s - loss: 0.5133 - accuracy: 0.8062
   
6/92 [>.............................] - ETA: 4s - loss: 0.5098 - accuracy: 0.8073
   
7/92 [=>............................] - ETA: 4s - loss: 0.5332 - accuracy: 0.8036
   
8/92 [=>............................] - ETA: 4s - loss: 0.5426 - accuracy: 0.7969
   
9/92 [=>............................] - ETA: 4s - loss: 0.5770 - accuracy: 0.7778


10/92 [==>………………………] - ETA: 4s - loss: 0.6016 - accuracy: 0.7719



11/92 [==>………………………] - ETA: 4s - loss: 0.5975 - accuracy: 0.7699



12/92 [==>………………………] - ETA: 4s - loss: 0.5897 - accuracy: 0.7734



13/92 [===>……………………..] - ETA: 4s - loss: 0.5994 - accuracy: 0.7716



14/92 [===>……………………..] - ETA: 4s - loss: 0.5986 - accuracy: 0.7746



15/92 [===>……………………..] - ETA: 4s - loss: 0.5995 - accuracy: 0.7708



16/92 [====>…………………….] - ETA: 4s - loss: 0.6063 - accuracy: 0.7656



17/92 [====>…………………….] - ETA: 4s - loss: 0.6023 - accuracy: 0.7665



18/92 [====>…………………….] - ETA: 4s - loss: 0.6027 - accuracy: 0.7656



19/92 [=====>……………………] - ETA: 4s - loss: 0.5956 - accuracy: 0.7664



21/92 [=====>……………………] - ETA: 4s - loss: 0.5947 - accuracy: 0.7666



22/92 [======>…………………..] - ETA: 4s - loss: 0.5901 - accuracy: 0.7672



23/92 [======>…………………..] - ETA: 3s - loss: 0.5811 - accuracy: 0.7720



24/92 [======>…………………..] - ETA: 3s - loss: 0.5810 - accuracy: 0.7737



25/92 [=======>………………….] - ETA: 3s - loss: 0.5865 - accuracy: 0.7715



26/92 [=======>………………….] - ETA: 3s - loss: 0.5784 - accuracy: 0.7767



27/92 [=======>………………….] - ETA: 3s - loss: 0.5786 - accuracy: 0.7745



28/92 [========>…………………] - ETA: 3s - loss: 0.5726 - accuracy: 0.7770



29/92 [========>…………………] - ETA: 3s - loss: 0.5660 - accuracy: 0.7793



30/92 [========>…………………] - ETA: 3s - loss: 0.5637 - accuracy: 0.7805



31/92 [=========>………………..] - ETA: 3s - loss: 0.5661 - accuracy: 0.7785



32/92 [=========>………………..] - ETA: 3s - loss: 0.5745 - accuracy: 0.7766



33/92 [=========>………………..] - ETA: 3s - loss: 0.5816 - accuracy: 0.7748



34/92 [==========>……………….] - ETA: 3s - loss: 0.5781 - accuracy: 0.7750



35/92 [==========>……………….] - ETA: 3s - loss: 0.5772 - accuracy: 0.7734



36/92 [==========>……………….] - ETA: 3s - loss: 0.5729 - accuracy: 0.7753



37/92 [===========>………………] - ETA: 3s - loss: 0.5776 - accuracy: 0.7747



38/92 [===========>………………] - ETA: 3s - loss: 0.5781 - accuracy: 0.7781



39/92 [===========>………………] - ETA: 3s - loss: 0.5773 - accuracy: 0.7790



40/92 [============>……………..] - ETA: 3s - loss: 0.5783 - accuracy: 0.7799



41/92 [============>……………..] - ETA: 2s - loss: 0.5778 - accuracy: 0.7799



42/92 [============>……………..] - ETA: 2s - loss: 0.5755 - accuracy: 0.7792



43/92 [=============>…………….] - ETA: 2s - loss: 0.5769 - accuracy: 0.7792



44/92 [=============>…………….] - ETA: 2s - loss: 0.5776 - accuracy: 0.7786



45/92 [=============>…………….] - ETA: 2s - loss: 0.5752 - accuracy: 0.7793



46/92 [==============>……………] - ETA: 2s - loss: 0.5734 - accuracy: 0.7801



47/92 [==============>……………] - ETA: 2s - loss: 0.5765 - accuracy: 0.7787



48/92 [==============>……………] - ETA: 2s - loss: 0.5780 - accuracy: 0.7788



49/92 [==============>……………] - ETA: 2s - loss: 0.5799 - accuracy: 0.7776



50/92 [===============>…………..] - ETA: 2s - loss: 0.5836 - accuracy: 0.7770



51/92 [===============>…………..] - ETA: 2s - loss: 0.5823 - accuracy: 0.7777



52/92 [===============>…………..] - ETA: 2s - loss: 0.5808 - accuracy: 0.7790



53/92 [================>………….] - ETA: 2s - loss: 0.5838 - accuracy: 0.7778



54/92 [================>………….] - ETA: 2s - loss: 0.5808 - accuracy: 0.7791



55/92 [================>………….] - ETA: 2s - loss: 0.5803 - accuracy: 0.7780



56/92 [=================>…………] - ETA: 2s - loss: 0.5826 - accuracy: 0.7775



57/92 [=================>…………] - ETA: 2s - loss: 0.5896 - accuracy: 0.7742



58/92 [=================>…………] - ETA: 1s - loss: 0.5896 - accuracy: 0.7738



59/92 [==================>………..] - ETA: 1s - loss: 0.5893 - accuracy: 0.7739



60/92 [==================>………..] - ETA: 1s - loss: 0.5915 - accuracy: 0.7735



61/92 [==================>………..] - ETA: 1s - loss: 0.5931 - accuracy: 0.7726



62/92 [===================>……….] - ETA: 1s - loss: 0.5908 - accuracy: 0.7733



63/92 [===================>……….] - ETA: 1s - loss: 0.5886 - accuracy: 0.7744



64/92 [===================>……….] - ETA: 1s - loss: 0.5892 - accuracy: 0.7745



65/92 [====================>………] - ETA: 1s - loss: 0.5897 - accuracy: 0.7736



66/92 [====================>………] - ETA: 1s - loss: 0.5921 - accuracy: 0.7728



67/92 [====================>………] - ETA: 1s - loss: 0.5900 - accuracy: 0.7748



68/92 [=====================>……..] - ETA: 1s - loss: 0.5871 - accuracy: 0.7763



69/92 [=====================>……..] - ETA: 1s - loss: 0.5849 - accuracy: 0.7759



70/92 [=====================>……..] - ETA: 1s - loss: 0.5883 - accuracy: 0.7742



71/92 [======================>…….] - ETA: 1s - loss: 0.5860 - accuracy: 0.7752



72/92 [======================>…….] - ETA: 1s - loss: 0.5845 - accuracy: 0.7757



73/92 [======================>…….] - ETA: 1s - loss: 0.5850 - accuracy: 0.7762



74/92 [=======================>……] - ETA: 1s - loss: 0.5829 - accuracy: 0.7771



75/92 [=======================>……] - ETA: 0s - loss: 0.5818 - accuracy: 0.7780



76/92 [=======================>……] - ETA: 0s - loss: 0.5827 - accuracy: 0.7772



77/92 [========================>…..] - ETA: 0s - loss: 0.5818 - accuracy: 0.7761



78/92 [========================>…..] - ETA: 0s - loss: 0.5856 - accuracy: 0.7761



79/92 [========================>…..] - ETA: 0s - loss: 0.5870 - accuracy: 0.7758



80/92 [=========================>….] - ETA: 0s - loss: 0.5876 - accuracy: 0.7743



81/92 [=========================>….] - ETA: 0s - loss: 0.5868 - accuracy: 0.7748



82/92 [=========================>….] - ETA: 0s - loss: 0.5895 - accuracy: 0.7737



83/92 [==========================>…] - ETA: 0s - loss: 0.5908 - accuracy: 0.7727



84/92 [==========================>…] - ETA: 0s - loss: 0.5928 - accuracy: 0.7720



85/92 [==========================>…] - ETA: 0s - loss: 0.5920 - accuracy: 0.7718



86/92 [===========================>..] - ETA: 0s - loss: 0.5914 - accuracy: 0.7726



87/92 [===========================>..] - ETA: 0s - loss: 0.5921 - accuracy: 0.7723



88/92 [===========================>..] - ETA: 0s - loss: 0.5911 - accuracy: 0.7724



89/92 [============================>.] - ETA: 0s - loss: 0.5913 - accuracy: 0.7725



90/92 [============================>.] - ETA: 0s - loss: 0.5910 - accuracy: 0.7726



91/92 [============================>.] - ETA: 0s - loss: 0.5901 - accuracy: 0.7724



92/92 [==============================] - ETA: 0s - loss: 0.5883 - accuracy: 0.7725



92/92 [==============================] - 6s 64ms/step - loss: 0.5883 - accuracy: 0.7725 - val_loss: 0.7175 - val_accuracy: 0.7234

Epoch 12/15
1/92 [..............................] - ETA: 7s - loss: 0.5461 - accuracy: 0.7812
   
2/92 [..............................] - ETA: 5s - loss: 0.6078 - accuracy: 0.7500
   
3/92 [..............................] - ETA: 5s - loss: 0.5296 - accuracy: 0.7917
   
4/92 [>.............................] - ETA: 5s - loss: 0.5025 - accuracy: 0.8047
   
5/92 [>.............................] - ETA: 5s - loss: 0.5195 - accuracy: 0.7812
   
6/92 [>.............................] - ETA: 4s - loss: 0.4948 - accuracy: 0.7917
   
7/92 [=>............................] - ETA: 4s - loss: 0.4886 - accuracy: 0.7857
   
8/92 [=>............................] - ETA: 4s - loss: 0.5058 - accuracy: 0.7773
   
9/92 [=>............................] - ETA: 4s - loss: 0.4985 - accuracy: 0.7882


10/92 [==>………………………] - ETA: 4s - loss: 0.4993 - accuracy: 0.7969



11/92 [==>………………………] - ETA: 4s - loss: 0.4915 - accuracy: 0.8011



12/92 [==>………………………] - ETA: 4s - loss: 0.5063 - accuracy: 0.8047



13/92 [===>……………………..] - ETA: 4s - loss: 0.5257 - accuracy: 0.7981



14/92 [===>……………………..] - ETA: 4s - loss: 0.5310 - accuracy: 0.7969



15/92 [===>……………………..] - ETA: 4s - loss: 0.5420 - accuracy: 0.7875



16/92 [====>…………………….] - ETA: 4s - loss: 0.5323 - accuracy: 0.7910



17/92 [====>…………………….] - ETA: 4s - loss: 0.5475 - accuracy: 0.7831



18/92 [====>…………………….] - ETA: 4s - loss: 0.5417 - accuracy: 0.7830



19/92 [=====>……………………] - ETA: 4s - loss: 0.5430 - accuracy: 0.7829



20/92 [=====>……………………] - ETA: 4s - loss: 0.5497 - accuracy: 0.7766



21/92 [=====>……………………] - ETA: 4s - loss: 0.5450 - accuracy: 0.7783



22/92 [======>…………………..] - ETA: 4s - loss: 0.5501 - accuracy: 0.7784



23/92 [======>…………………..] - ETA: 3s - loss: 0.5590 - accuracy: 0.7799



24/92 [======>…………………..] - ETA: 3s - loss: 0.5482 - accuracy: 0.7865



25/92 [=======>………………….] - ETA: 3s - loss: 0.5465 - accuracy: 0.7875



26/92 [=======>………………….] - ETA: 3s - loss: 0.5421 - accuracy: 0.7885



27/92 [=======>………………….] - ETA: 3s - loss: 0.5454 - accuracy: 0.7882



28/92 [========>…………………] - ETA: 3s - loss: 0.5488 - accuracy: 0.7868



29/92 [========>…………………] - ETA: 3s - loss: 0.5551 - accuracy: 0.7812



30/92 [========>…………………] - ETA: 3s - loss: 0.5555 - accuracy: 0.7802



31/92 [=========>………………..] - ETA: 3s - loss: 0.5554 - accuracy: 0.7782



32/92 [=========>………………..] - ETA: 3s - loss: 0.5517 - accuracy: 0.7773



33/92 [=========>………………..] - ETA: 3s - loss: 0.5521 - accuracy: 0.7794



34/92 [==========>……………….] - ETA: 3s - loss: 0.5585 - accuracy: 0.7776



35/92 [==========>……………….] - ETA: 3s - loss: 0.5562 - accuracy: 0.7777



36/92 [==========>……………….] - ETA: 3s - loss: 0.5507 - accuracy: 0.7812



37/92 [===========>………………] - ETA: 3s - loss: 0.5586 - accuracy: 0.7779



38/92 [===========>………………] - ETA: 3s - loss: 0.5564 - accuracy: 0.7796



39/92 [===========>………………] - ETA: 3s - loss: 0.5534 - accuracy: 0.7812



40/92 [============>……………..] - ETA: 3s - loss: 0.5536 - accuracy: 0.7820



41/92 [============>……………..] - ETA: 2s - loss: 0.5572 - accuracy: 0.7805



42/92 [============>……………..] - ETA: 2s - loss: 0.5569 - accuracy: 0.7805



43/92 [=============>…………….] - ETA: 2s - loss: 0.5613 - accuracy: 0.7776



44/92 [=============>…………….] - ETA: 2s - loss: 0.5597 - accuracy: 0.7784



45/92 [=============>…………….] - ETA: 2s - loss: 0.5597 - accuracy: 0.7799



46/92 [==============>……………] - ETA: 2s - loss: 0.5566 - accuracy: 0.7812



47/92 [==============>……………] - ETA: 2s - loss: 0.5596 - accuracy: 0.7793



48/92 [==============>……………] - ETA: 2s - loss: 0.5574 - accuracy: 0.7806



49/92 [==============>……………] - ETA: 2s - loss: 0.5618 - accuracy: 0.7793



50/92 [===============>…………..] - ETA: 2s - loss: 0.5593 - accuracy: 0.7806



51/92 [===============>…………..] - ETA: 2s - loss: 0.5628 - accuracy: 0.7794



52/92 [===============>…………..] - ETA: 2s - loss: 0.5675 - accuracy: 0.7776



53/92 [================>………….] - ETA: 2s - loss: 0.5661 - accuracy: 0.7783



54/92 [================>………….] - ETA: 2s - loss: 0.5638 - accuracy: 0.7795



55/92 [================>………….] - ETA: 2s - loss: 0.5641 - accuracy: 0.7790



56/92 [=================>…………] - ETA: 2s - loss: 0.5638 - accuracy: 0.7785



57/92 [=================>…………] - ETA: 2s - loss: 0.5621 - accuracy: 0.7785



58/92 [=================>…………] - ETA: 1s - loss: 0.5647 - accuracy: 0.7775



59/92 [==================>………..] - ETA: 1s - loss: 0.5642 - accuracy: 0.7770



60/92 [==================>………..] - ETA: 1s - loss: 0.5652 - accuracy: 0.7760



61/92 [==================>………..] - ETA: 1s - loss: 0.5647 - accuracy: 0.7772



62/92 [===================>……….] - ETA: 1s - loss: 0.5618 - accuracy: 0.7792



63/92 [===================>……….] - ETA: 1s - loss: 0.5621 - accuracy: 0.7783



64/92 [===================>……….] - ETA: 1s - loss: 0.5602 - accuracy: 0.7788



65/92 [====================>………] - ETA: 1s - loss: 0.5594 - accuracy: 0.7793



66/92 [====================>………] - ETA: 1s - loss: 0.5603 - accuracy: 0.7784



67/92 [====================>………] - ETA: 1s - loss: 0.5598 - accuracy: 0.7789



68/92 [=====================>……..] - ETA: 1s - loss: 0.5633 - accuracy: 0.7776



69/92 [=====================>……..] - ETA: 1s - loss: 0.5660 - accuracy: 0.7767



70/92 [=====================>……..] - ETA: 1s - loss: 0.5666 - accuracy: 0.7763



71/92 [======================>…….] - ETA: 1s - loss: 0.5642 - accuracy: 0.7777



72/92 [======================>…….] - ETA: 1s - loss: 0.5633 - accuracy: 0.7778



73/92 [======================>…….] - ETA: 1s - loss: 0.5618 - accuracy: 0.7787



74/92 [=======================>……] - ETA: 1s - loss: 0.5634 - accuracy: 0.7779



75/92 [=======================>……] - ETA: 0s - loss: 0.5613 - accuracy: 0.7783



77/92 [========================>…..] - ETA: 0s - loss: 0.5630 - accuracy: 0.7769



78/92 [========================>…..] - ETA: 0s - loss: 0.5655 - accuracy: 0.7761



79/92 [========================>…..] - ETA: 0s - loss: 0.5639 - accuracy: 0.7770



80/92 [=========================>….] - ETA: 0s - loss: 0.5662 - accuracy: 0.7766



81/92 [=========================>….] - ETA: 0s - loss: 0.5626 - accuracy: 0.7783



82/92 [=========================>….] - ETA: 0s - loss: 0.5621 - accuracy: 0.7794



83/92 [==========================>…] - ETA: 0s - loss: 0.5608 - accuracy: 0.7806



84/92 [==========================>…] - ETA: 0s - loss: 0.5608 - accuracy: 0.7806



85/92 [==========================>…] - ETA: 0s - loss: 0.5613 - accuracy: 0.7802



86/92 [===========================>..] - ETA: 0s - loss: 0.5608 - accuracy: 0.7806



87/92 [===========================>..] - ETA: 0s - loss: 0.5642 - accuracy: 0.7803



88/92 [===========================>..] - ETA: 0s - loss: 0.5625 - accuracy: 0.7810



89/92 [============================>.] - ETA: 0s - loss: 0.5648 - accuracy: 0.7806



90/92 [============================>.] - ETA: 0s - loss: 0.5615 - accuracy: 0.7824



91/92 [============================>.] - ETA: 0s - loss: 0.5611 - accuracy: 0.7820



92/92 [==============================] - ETA: 0s - loss: 0.5609 - accuracy: 0.7827



92/92 [==============================] - 6s 64ms/step - loss: 0.5609 - accuracy: 0.7827 - val_loss: 0.6652 - val_accuracy: 0.7357

Epoch 13/15
1/92 [..............................] - ETA: 7s - loss: 0.5252 - accuracy: 0.8438
   
2/92 [..............................] - ETA: 5s - loss: 0.5595 - accuracy: 0.7969
   
3/92 [..............................] - ETA: 5s - loss: 0.5306 - accuracy: 0.8125
   
4/92 [>.............................] - ETA: 5s - loss: 0.5318 - accuracy: 0.8125
   
5/92 [>.............................] - ETA: 5s - loss: 0.4936 - accuracy: 0.8313
   
6/92 [>.............................] - ETA: 5s - loss: 0.4675 - accuracy: 0.8438
   
7/92 [=>............................] - ETA: 4s - loss: 0.4796 - accuracy: 0.8348
   
8/92 [=>............................] - ETA: 4s - loss: 0.5024 - accuracy: 0.8164
   
9/92 [=>............................] - ETA: 4s - loss: 0.4919 - accuracy: 0.8264


10/92 [==>………………………] - ETA: 4s - loss: 0.5071 - accuracy: 0.8219



11/92 [==>………………………] - ETA: 4s - loss: 0.5112 - accuracy: 0.8182



12/92 [==>………………………] - ETA: 4s - loss: 0.5037 - accuracy: 0.8203



13/92 [===>……………………..] - ETA: 4s - loss: 0.4893 - accuracy: 0.8245



14/92 [===>……………………..] - ETA: 4s - loss: 0.4904 - accuracy: 0.8281



15/92 [===>……………………..] - ETA: 4s - loss: 0.4893 - accuracy: 0.8271



16/92 [====>…………………….] - ETA: 4s - loss: 0.4908 - accuracy: 0.8242



17/92 [====>…………………….] - ETA: 4s - loss: 0.4963 - accuracy: 0.8180



18/92 [====>…………………….] - ETA: 4s - loss: 0.4981 - accuracy: 0.8177



19/92 [=====>……………………] - ETA: 4s - loss: 0.5066 - accuracy: 0.8141



20/92 [=====>……………………] - ETA: 4s - loss: 0.5055 - accuracy: 0.8125



21/92 [=====>……………………] - ETA: 4s - loss: 0.5156 - accuracy: 0.8065



22/92 [======>…………………..] - ETA: 4s - loss: 0.5282 - accuracy: 0.8054



23/92 [======>…………………..] - ETA: 3s - loss: 0.5264 - accuracy: 0.8084



24/92 [======>…………………..] - ETA: 3s - loss: 0.5195 - accuracy: 0.8099



25/92 [=======>………………….] - ETA: 3s - loss: 0.5105 - accuracy: 0.8138



26/92 [=======>………………….] - ETA: 3s - loss: 0.5050 - accuracy: 0.8149



27/92 [=======>………………….] - ETA: 3s - loss: 0.5063 - accuracy: 0.8171



28/92 [========>…………………] - ETA: 3s - loss: 0.5090 - accuracy: 0.8147



29/92 [========>…………………] - ETA: 3s - loss: 0.5016 - accuracy: 0.8179



30/92 [========>…………………] - ETA: 3s - loss: 0.4980 - accuracy: 0.8188



31/92 [=========>………………..] - ETA: 3s - loss: 0.5032 - accuracy: 0.8155



32/92 [=========>………………..] - ETA: 3s - loss: 0.4998 - accuracy: 0.8164



33/92 [=========>………………..] - ETA: 3s - loss: 0.4974 - accuracy: 0.8163



34/92 [==========>……………….] - ETA: 3s - loss: 0.5017 - accuracy: 0.8134



35/92 [==========>……………….] - ETA: 3s - loss: 0.5067 - accuracy: 0.8116



36/92 [==========>……………….] - ETA: 3s - loss: 0.5030 - accuracy: 0.8134



37/92 [===========>………………] - ETA: 3s - loss: 0.4970 - accuracy: 0.8159



38/92 [===========>………………] - ETA: 3s - loss: 0.4961 - accuracy: 0.8158



39/92 [===========>………………] - ETA: 3s - loss: 0.4938 - accuracy: 0.8165



40/92 [============>……………..] - ETA: 3s - loss: 0.4929 - accuracy: 0.8164



41/92 [============>……………..] - ETA: 2s - loss: 0.4964 - accuracy: 0.8171



42/92 [============>……………..] - ETA: 2s - loss: 0.4964 - accuracy: 0.8155



43/92 [=============>…………….] - ETA: 2s - loss: 0.4995 - accuracy: 0.8147



44/92 [=============>…………….] - ETA: 2s - loss: 0.5069 - accuracy: 0.8118



45/92 [=============>…………….] - ETA: 2s - loss: 0.5136 - accuracy: 0.8083



46/92 [==============>……………] - ETA: 2s - loss: 0.5124 - accuracy: 0.8084



47/92 [==============>……………] - ETA: 2s - loss: 0.5139 - accuracy: 0.8072



48/92 [==============>……………] - ETA: 2s - loss: 0.5154 - accuracy: 0.8073



49/92 [==============>……………] - ETA: 2s - loss: 0.5157 - accuracy: 0.8068



50/92 [===============>…………..] - ETA: 2s - loss: 0.5195 - accuracy: 0.8062



51/92 [===============>…………..] - ETA: 2s - loss: 0.5191 - accuracy: 0.8064



52/92 [===============>…………..] - ETA: 2s - loss: 0.5174 - accuracy: 0.8077



53/92 [================>………….] - ETA: 2s - loss: 0.5186 - accuracy: 0.8078



54/92 [================>………….] - ETA: 2s - loss: 0.5211 - accuracy: 0.8079



55/92 [================>………….] - ETA: 2s - loss: 0.5175 - accuracy: 0.8091



56/92 [=================>…………] - ETA: 2s - loss: 0.5185 - accuracy: 0.8092



57/92 [=================>…………] - ETA: 2s - loss: 0.5193 - accuracy: 0.8087



58/92 [=================>…………] - ETA: 1s - loss: 0.5222 - accuracy: 0.8071



59/92 [==================>………..] - ETA: 1s - loss: 0.5193 - accuracy: 0.8088



60/92 [==================>………..] - ETA: 1s - loss: 0.5179 - accuracy: 0.8094



61/92 [==================>………..] - ETA: 1s - loss: 0.5190 - accuracy: 0.8094



62/92 [===================>……….] - ETA: 1s - loss: 0.5203 - accuracy: 0.8095



63/92 [===================>……….] - ETA: 1s - loss: 0.5183 - accuracy: 0.8100



64/92 [===================>……….] - ETA: 1s - loss: 0.5221 - accuracy: 0.8091



65/92 [====================>………] - ETA: 1s - loss: 0.5257 - accuracy: 0.8077



66/92 [====================>………] - ETA: 1s - loss: 0.5256 - accuracy: 0.8078



67/92 [====================>………] - ETA: 1s - loss: 0.5246 - accuracy: 0.8083



68/92 [=====================>……..] - ETA: 1s - loss: 0.5257 - accuracy: 0.8065



69/92 [=====================>……..] - ETA: 1s - loss: 0.5277 - accuracy: 0.8057



70/92 [=====================>……..] - ETA: 1s - loss: 0.5288 - accuracy: 0.8049



71/92 [======================>…….] - ETA: 1s - loss: 0.5315 - accuracy: 0.8041



72/92 [======================>…….] - ETA: 1s - loss: 0.5324 - accuracy: 0.8030



73/92 [======================>…….] - ETA: 1s - loss: 0.5341 - accuracy: 0.8031



74/92 [=======================>……] - ETA: 1s - loss: 0.5340 - accuracy: 0.8032



75/92 [=======================>……] - ETA: 0s - loss: 0.5340 - accuracy: 0.8029



76/92 [=======================>……] - ETA: 0s - loss: 0.5359 - accuracy: 0.8014



77/92 [========================>…..] - ETA: 0s - loss: 0.5358 - accuracy: 0.8019



78/92 [========================>…..] - ETA: 0s - loss: 0.5377 - accuracy: 0.8021



79/92 [========================>…..] - ETA: 0s - loss: 0.5358 - accuracy: 0.8030



80/92 [=========================>….] - ETA: 0s - loss: 0.5352 - accuracy: 0.8035



81/92 [=========================>….] - ETA: 0s - loss: 0.5352 - accuracy: 0.8025



82/92 [=========================>….] - ETA: 0s - loss: 0.5368 - accuracy: 0.8018



83/92 [==========================>…] - ETA: 0s - loss: 0.5345 - accuracy: 0.8027



84/92 [==========================>…] - ETA: 0s - loss: 0.5322 - accuracy: 0.8036



86/92 [===========================>..] - ETA: 0s - loss: 0.5315 - accuracy: 0.8043



87/92 [===========================>..] - ETA: 0s - loss: 0.5294 - accuracy: 0.8044



88/92 [===========================>..] - ETA: 0s - loss: 0.5280 - accuracy: 0.8052



89/92 [============================>.] - ETA: 0s - loss: 0.5288 - accuracy: 0.8042



90/92 [============================>.] - ETA: 0s - loss: 0.5304 - accuracy: 0.8043



91/92 [============================>.] - ETA: 0s - loss: 0.5279 - accuracy: 0.8054



92/92 [==============================] - ETA: 0s - loss: 0.5255 - accuracy: 0.8072



92/92 [==============================] - 6s 64ms/step - loss: 0.5255 - accuracy: 0.8072 - val_loss: 0.7346 - val_accuracy: 0.7384

Epoch 14/15
1/92 [..............................] - ETA: 7s - loss: 0.3946 - accuracy: 0.8125
   
2/92 [..............................] - ETA: 5s - loss: 0.4186 - accuracy: 0.8438
   
3/92 [..............................] - ETA: 5s - loss: 0.5825 - accuracy: 0.7708
   
4/92 [>.............................] - ETA: 5s - loss: 0.5210 - accuracy: 0.8047
   
5/92 [>.............................] - ETA: 5s - loss: 0.5927 - accuracy: 0.7750
   
6/92 [>.............................] - ETA: 4s - loss: 0.5649 - accuracy: 0.7865
   
7/92 [=>............................] - ETA: 4s - loss: 0.5574 - accuracy: 0.7902
   
8/92 [=>............................] - ETA: 4s - loss: 0.5343 - accuracy: 0.7969
   
9/92 [=>............................] - ETA: 4s - loss: 0.5348 - accuracy: 0.7951


10/92 [==>………………………] - ETA: 4s - loss: 0.5328 - accuracy: 0.7937



11/92 [==>………………………] - ETA: 4s - loss: 0.5417 - accuracy: 0.7898



12/92 [==>………………………] - ETA: 4s - loss: 0.5423 - accuracy: 0.7839



13/92 [===>……………………..] - ETA: 4s - loss: 0.5293 - accuracy: 0.7957



14/92 [===>……………………..] - ETA: 4s - loss: 0.5237 - accuracy: 0.8013



15/92 [===>……………………..] - ETA: 4s - loss: 0.5192 - accuracy: 0.7958



16/92 [====>…………………….] - ETA: 4s - loss: 0.5227 - accuracy: 0.7969



17/92 [====>…………………….] - ETA: 4s - loss: 0.5219 - accuracy: 0.7996



18/92 [====>…………………….] - ETA: 4s - loss: 0.5389 - accuracy: 0.7951



20/92 [=====>……………………] - ETA: 4s - loss: 0.5362 - accuracy: 0.7927



21/92 [=====>……………………] - ETA: 4s - loss: 0.5310 - accuracy: 0.7937



22/92 [======>…………………..] - ETA: 4s - loss: 0.5365 - accuracy: 0.7945



23/92 [======>…………………..] - ETA: 3s - loss: 0.5421 - accuracy: 0.7940



24/92 [======>…………………..] - ETA: 3s - loss: 0.5364 - accuracy: 0.7947



25/92 [=======>………………….] - ETA: 3s - loss: 0.5400 - accuracy: 0.7942



26/92 [=======>………………….] - ETA: 3s - loss: 0.5380 - accuracy: 0.7949



27/92 [=======>………………….] - ETA: 3s - loss: 0.5330 - accuracy: 0.7967



28/92 [========>…………………] - ETA: 3s - loss: 0.5423 - accuracy: 0.7928



29/92 [========>…………………] - ETA: 3s - loss: 0.5431 - accuracy: 0.7935



30/92 [========>…………………] - ETA: 3s - loss: 0.5458 - accuracy: 0.7920



31/92 [=========>………………..] - ETA: 3s - loss: 0.5496 - accuracy: 0.7907



32/92 [=========>………………..] - ETA: 3s - loss: 0.5505 - accuracy: 0.7884



33/92 [=========>………………..] - ETA: 3s - loss: 0.5535 - accuracy: 0.7872



34/92 [==========>……………….] - ETA: 3s - loss: 0.5620 - accuracy: 0.7870



35/92 [==========>……………….] - ETA: 3s - loss: 0.5640 - accuracy: 0.7878



36/92 [==========>……………….] - ETA: 3s - loss: 0.5689 - accuracy: 0.7858



37/92 [===========>………………] - ETA: 3s - loss: 0.5683 - accuracy: 0.7849



38/92 [===========>………………] - ETA: 3s - loss: 0.5653 - accuracy: 0.7848



39/92 [===========>………………] - ETA: 3s - loss: 0.5606 - accuracy: 0.7863



40/92 [============>……………..] - ETA: 2s - loss: 0.5612 - accuracy: 0.7862



41/92 [============>……………..] - ETA: 2s - loss: 0.5642 - accuracy: 0.7860



42/92 [============>……………..] - ETA: 2s - loss: 0.5674 - accuracy: 0.7844



43/92 [=============>…………….] - ETA: 2s - loss: 0.5612 - accuracy: 0.7880



44/92 [=============>…………….] - ETA: 2s - loss: 0.5614 - accuracy: 0.7879



45/92 [=============>…………….] - ETA: 2s - loss: 0.5615 - accuracy: 0.7863



46/92 [==============>……………] - ETA: 2s - loss: 0.5615 - accuracy: 0.7855



47/92 [==============>……………] - ETA: 2s - loss: 0.5600 - accuracy: 0.7861



48/92 [==============>……………] - ETA: 2s - loss: 0.5585 - accuracy: 0.7880



49/92 [==============>……………] - ETA: 2s - loss: 0.5641 - accuracy: 0.7846



50/92 [===============>…………..] - ETA: 2s - loss: 0.5635 - accuracy: 0.7864



51/92 [===============>…………..] - ETA: 2s - loss: 0.5630 - accuracy: 0.7857



52/92 [===============>…………..] - ETA: 2s - loss: 0.5604 - accuracy: 0.7874



53/92 [================>………….] - ETA: 2s - loss: 0.5621 - accuracy: 0.7855



54/92 [================>………….] - ETA: 2s - loss: 0.5613 - accuracy: 0.7855



55/92 [================>………….] - ETA: 2s - loss: 0.5630 - accuracy: 0.7842



56/92 [=================>…………] - ETA: 2s - loss: 0.5613 - accuracy: 0.7853



57/92 [=================>…………] - ETA: 2s - loss: 0.5607 - accuracy: 0.7858



58/92 [=================>…………] - ETA: 1s - loss: 0.5580 - accuracy: 0.7873



59/92 [==================>………..] - ETA: 1s - loss: 0.5560 - accuracy: 0.7872



60/92 [==================>………..] - ETA: 1s - loss: 0.5529 - accuracy: 0.7887



61/92 [==================>………..] - ETA: 1s - loss: 0.5507 - accuracy: 0.7896



62/92 [===================>……….] - ETA: 1s - loss: 0.5497 - accuracy: 0.7900



63/92 [===================>……….] - ETA: 1s - loss: 0.5490 - accuracy: 0.7903



64/92 [===================>……….] - ETA: 1s - loss: 0.5505 - accuracy: 0.7887



65/92 [====================>………] - ETA: 1s - loss: 0.5467 - accuracy: 0.7901



66/92 [====================>………] - ETA: 1s - loss: 0.5484 - accuracy: 0.7899



67/92 [====================>………] - ETA: 1s - loss: 0.5478 - accuracy: 0.7903



68/92 [=====================>……..] - ETA: 1s - loss: 0.5458 - accuracy: 0.7911



69/92 [=====================>……..] - ETA: 1s - loss: 0.5454 - accuracy: 0.7914



70/92 [=====================>……..] - ETA: 1s - loss: 0.5482 - accuracy: 0.7899



71/92 [======================>…….] - ETA: 1s - loss: 0.5472 - accuracy: 0.7902



72/92 [======================>…….] - ETA: 1s - loss: 0.5475 - accuracy: 0.7888



73/92 [======================>…….] - ETA: 1s - loss: 0.5450 - accuracy: 0.7904



74/92 [=======================>……] - ETA: 1s - loss: 0.5426 - accuracy: 0.7911



75/92 [=======================>……] - ETA: 0s - loss: 0.5456 - accuracy: 0.7901



76/92 [=======================>……] - ETA: 0s - loss: 0.5460 - accuracy: 0.7892



77/92 [========================>…..] - ETA: 0s - loss: 0.5432 - accuracy: 0.7911



78/92 [========================>…..] - ETA: 0s - loss: 0.5412 - accuracy: 0.7914



79/92 [========================>…..] - ETA: 0s - loss: 0.5420 - accuracy: 0.7909



80/92 [=========================>….] - ETA: 0s - loss: 0.5455 - accuracy: 0.7888



81/92 [=========================>….] - ETA: 0s - loss: 0.5470 - accuracy: 0.7891



82/92 [=========================>….] - ETA: 0s - loss: 0.5454 - accuracy: 0.7898



83/92 [==========================>…] - ETA: 0s - loss: 0.5431 - accuracy: 0.7908



84/92 [==========================>…] - ETA: 0s - loss: 0.5453 - accuracy: 0.7907



85/92 [==========================>…] - ETA: 0s - loss: 0.5463 - accuracy: 0.7906



86/92 [===========================>..] - ETA: 0s - loss: 0.5460 - accuracy: 0.7905



87/92 [===========================>..] - ETA: 0s - loss: 0.5441 - accuracy: 0.7903



88/92 [===========================>..] - ETA: 0s - loss: 0.5444 - accuracy: 0.7899



89/92 [============================>.] - ETA: 0s - loss: 0.5429 - accuracy: 0.7901



90/92 [============================>.] - ETA: 0s - loss: 0.5423 - accuracy: 0.7904



91/92 [============================>.] - ETA: 0s - loss: 0.5448 - accuracy: 0.7893



92/92 [==============================] - ETA: 0s - loss: 0.5438 - accuracy: 0.7895



92/92 [==============================] - 6s 64ms/step - loss: 0.5438 - accuracy: 0.7895 - val_loss: 0.7761 - val_accuracy: 0.7275

Epoch 15/15
1/92 [..............................] - ETA: 7s - loss: 0.5374 - accuracy: 0.8750
   
2/92 [..............................] - ETA: 5s - loss: 0.4638 - accuracy: 0.8906
   
3/92 [..............................] - ETA: 5s - loss: 0.4361 - accuracy: 0.8750
   
4/92 [>.............................] - ETA: 5s - loss: 0.4714 - accuracy: 0.8281
   
5/92 [>.............................] - ETA: 4s - loss: 0.4472 - accuracy: 0.8375
   
6/92 [>.............................] - ETA: 4s - loss: 0.4562 - accuracy: 0.8281
   
7/92 [=>............................] - ETA: 4s - loss: 0.4228 - accuracy: 0.8438
   
8/92 [=>............................] - ETA: 4s - loss: 0.4377 - accuracy: 0.8359
   
9/92 [=>............................] - ETA: 4s - loss: 0.4744 - accuracy: 0.8264


10/92 [==>………………………] - ETA: 4s - loss: 0.4706 - accuracy: 0.8313



11/92 [==>………………………] - ETA: 4s - loss: 0.4714 - accuracy: 0.8324



12/92 [==>………………………] - ETA: 4s - loss: 0.4935 - accuracy: 0.8255



13/92 [===>……………………..] - ETA: 4s - loss: 0.4925 - accuracy: 0.8245



14/92 [===>……………………..] - ETA: 4s - loss: 0.4784 - accuracy: 0.8281



15/92 [===>……………………..] - ETA: 4s - loss: 0.4763 - accuracy: 0.8271



16/92 [====>…………………….] - ETA: 4s - loss: 0.4744 - accuracy: 0.8301



17/92 [====>…………………….] - ETA: 4s - loss: 0.4797 - accuracy: 0.8254



18/92 [====>…………………….] - ETA: 4s - loss: 0.4824 - accuracy: 0.8264



19/92 [=====>……………………] - ETA: 4s - loss: 0.4777 - accuracy: 0.8273



20/92 [=====>……………………] - ETA: 4s - loss: 0.4756 - accuracy: 0.8297



21/92 [=====>……………………] - ETA: 4s - loss: 0.4717 - accuracy: 0.8304



22/92 [======>…………………..] - ETA: 4s - loss: 0.4740 - accuracy: 0.8281



23/92 [======>…………………..] - ETA: 3s - loss: 0.4742 - accuracy: 0.8261



24/92 [======>…………………..] - ETA: 3s - loss: 0.4757 - accuracy: 0.8229



25/92 [=======>………………….] - ETA: 3s - loss: 0.4786 - accuracy: 0.8213



26/92 [=======>………………….] - ETA: 3s - loss: 0.4786 - accuracy: 0.8209



27/92 [=======>………………….] - ETA: 3s - loss: 0.4836 - accuracy: 0.8183



28/92 [========>…………………] - ETA: 3s - loss: 0.4816 - accuracy: 0.8181



29/92 [========>…………………] - ETA: 3s - loss: 0.4780 - accuracy: 0.8190



30/92 [========>…………………] - ETA: 3s - loss: 0.4789 - accuracy: 0.8177



31/92 [=========>………………..] - ETA: 3s - loss: 0.4751 - accuracy: 0.8175



32/92 [=========>………………..] - ETA: 3s - loss: 0.4703 - accuracy: 0.8184



33/92 [=========>………………..] - ETA: 3s - loss: 0.4701 - accuracy: 0.8191



34/92 [==========>……………….] - ETA: 3s - loss: 0.4664 - accuracy: 0.8199



35/92 [==========>……………….] - ETA: 3s - loss: 0.4676 - accuracy: 0.8196



36/92 [==========>……………….] - ETA: 3s - loss: 0.4686 - accuracy: 0.8203



37/92 [===========>………………] - ETA: 3s - loss: 0.4684 - accuracy: 0.8209



38/92 [===========>………………] - ETA: 3s - loss: 0.4648 - accuracy: 0.8199



39/92 [===========>………………] - ETA: 3s - loss: 0.4673 - accuracy: 0.8181



40/92 [============>……………..] - ETA: 2s - loss: 0.4649 - accuracy: 0.8188



41/92 [============>……………..] - ETA: 2s - loss: 0.4656 - accuracy: 0.8186



42/92 [============>……………..] - ETA: 2s - loss: 0.4694 - accuracy: 0.8177



43/92 [=============>…………….] - ETA: 2s - loss: 0.4750 - accuracy: 0.8154



44/92 [=============>…………….] - ETA: 2s - loss: 0.4788 - accuracy: 0.8118



45/92 [=============>…………….] - ETA: 2s - loss: 0.4772 - accuracy: 0.8132



46/92 [==============>……………] - ETA: 2s - loss: 0.4774 - accuracy: 0.8132



47/92 [==============>……………] - ETA: 2s - loss: 0.4781 - accuracy: 0.8145



48/92 [==============>……………] - ETA: 2s - loss: 0.4757 - accuracy: 0.8158



49/92 [==============>……………] - ETA: 2s - loss: 0.4758 - accuracy: 0.8157



50/92 [===============>…………..] - ETA: 2s - loss: 0.4741 - accuracy: 0.8169



51/92 [===============>…………..] - ETA: 2s - loss: 0.4775 - accuracy: 0.8168



52/92 [===============>…………..] - ETA: 2s - loss: 0.4794 - accuracy: 0.8149



53/92 [================>………….] - ETA: 2s - loss: 0.4896 - accuracy: 0.8125



54/92 [================>………….] - ETA: 2s - loss: 0.4885 - accuracy: 0.8137



55/92 [================>………….] - ETA: 2s - loss: 0.4864 - accuracy: 0.8136



56/92 [=================>…………] - ETA: 2s - loss: 0.4861 - accuracy: 0.8136



57/92 [=================>…………] - ETA: 2s - loss: 0.4869 - accuracy: 0.8130



58/92 [=================>…………] - ETA: 1s - loss: 0.4843 - accuracy: 0.8141



59/92 [==================>………..] - ETA: 1s - loss: 0.4882 - accuracy: 0.8130



60/92 [==================>………..] - ETA: 1s - loss: 0.4926 - accuracy: 0.8109



61/92 [==================>………..] - ETA: 1s - loss: 0.4920 - accuracy: 0.8115



62/92 [===================>……….] - ETA: 1s - loss: 0.4945 - accuracy: 0.8095



63/92 [===================>……….] - ETA: 1s - loss: 0.4951 - accuracy: 0.8095



64/92 [===================>……….] - ETA: 1s - loss: 0.4949 - accuracy: 0.8101



65/92 [====================>………] - ETA: 1s - loss: 0.4933 - accuracy: 0.8096



66/92 [====================>………] - ETA: 1s - loss: 0.4981 - accuracy: 0.8073



67/92 [====================>………] - ETA: 1s - loss: 0.4964 - accuracy: 0.8083



68/92 [=====================>……..] - ETA: 1s - loss: 0.4934 - accuracy: 0.8093



69/92 [=====================>……..] - ETA: 1s - loss: 0.4972 - accuracy: 0.8093



70/92 [=====================>……..] - ETA: 1s - loss: 0.4996 - accuracy: 0.8080



71/92 [======================>…….] - ETA: 1s - loss: 0.5009 - accuracy: 0.8072



72/92 [======================>…….] - ETA: 1s - loss: 0.5067 - accuracy: 0.8064



73/92 [======================>…….] - ETA: 1s - loss: 0.5055 - accuracy: 0.8061



74/92 [=======================>……] - ETA: 1s - loss: 0.5039 - accuracy: 0.8066



75/92 [=======================>……] - ETA: 0s - loss: 0.5070 - accuracy: 0.8046



76/92 [=======================>……] - ETA: 0s - loss: 0.5080 - accuracy: 0.8047



77/92 [========================>…..] - ETA: 0s - loss: 0.5113 - accuracy: 0.8040



78/92 [========================>…..] - ETA: 0s - loss: 0.5107 - accuracy: 0.8045



79/92 [========================>…..] - ETA: 0s - loss: 0.5141 - accuracy: 0.8026



80/92 [=========================>….] - ETA: 0s - loss: 0.5121 - accuracy: 0.8035



81/92 [=========================>….] - ETA: 0s - loss: 0.5107 - accuracy: 0.8044



82/92 [=========================>….] - ETA: 0s - loss: 0.5102 - accuracy: 0.8041



83/92 [==========================>…] - ETA: 0s - loss: 0.5116 - accuracy: 0.8035



84/92 [==========================>…] - ETA: 0s - loss: 0.5117 - accuracy: 0.8032



85/92 [==========================>…] - ETA: 0s - loss: 0.5099 - accuracy: 0.8037



86/92 [===========================>..] - ETA: 0s - loss: 0.5129 - accuracy: 0.8016



87/92 [===========================>..] - ETA: 0s - loss: 0.5155 - accuracy: 0.8006



89/92 [============================>.] - ETA: 0s - loss: 0.5151 - accuracy: 0.8011



90/92 [============================>.] - ETA: 0s - loss: 0.5154 - accuracy: 0.8012



91/92 [============================>.] - ETA: 0s - loss: 0.5143 - accuracy: 0.8017



92/92 [==============================] - ETA: 0s - loss: 0.5165 - accuracy: 0.8004



92/92 [==============================] - 6s 64ms/step - loss: 0.5165 - accuracy: 0.8004 - val_loss: 0.7822 - val_accuracy: 0.7289

../_images/301-tensorflow-training-openvino-nncf-with-output_3_1452.png
1/1 [==============================] - ETA: 0s

1/1 [==============================] - 0s 74ms/step
This image most likely belongs to sunflowers with a 99.24 percent confidence.
2024-02-10 01:10:41.607321: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'random_flip_input' with dtype float and shape [?,180,180,3]
     [[{{node random_flip_input}}]]
2024-02-10 01:10:41.692936: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'inputs' with dtype float and shape [?,180,180,3]
     [[{{node inputs}}]]
2024-02-10 01:10:41.703478: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'random_flip_input' with dtype float and shape [?,180,180,3]
     [[{{node random_flip_input}}]]
2024-02-10 01:10:41.714441: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'inputs' with dtype float and shape [?,180,180,3]
     [[{{node inputs}}]]
2024-02-10 01:10:41.722136: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'inputs' with dtype float and shape [?,180,180,3]
     [[{{node inputs}}]]
2024-02-10 01:10:41.728943: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'inputs' with dtype float and shape [?,180,180,3]
     [[{{node inputs}}]]
2024-02-10 01:10:41.739850: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'inputs' with dtype float and shape [?,180,180,3]
     [[{{node inputs}}]]
2024-02-10 01:10:41.778944: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'sequential_1_input' with dtype float and shape [?,180,180,3]
     [[{{node sequential_1_input}}]]
2024-02-10 01:10:41.847949: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'inputs' with dtype float and shape [?,180,180,3]
     [[{{node inputs}}]]
2024-02-10 01:10:41.868730: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'sequential_1_input' with dtype float and shape [?,180,180,3]
     [[{{node sequential_1_input}}]]
2024-02-10 01:10:41.907533: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'inputs' with dtype float and shape [?,22,22,64]
     [[{{node inputs}}]]
2024-02-10 01:10:41.933427: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'inputs' with dtype float and shape [?,180,180,3]
     [[{{node inputs}}]]
2024-02-10 01:10:42.007195: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'inputs' with dtype float and shape [?,180,180,3]
     [[{{node inputs}}]]
2024-02-10 01:10:42.149634: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'inputs' with dtype float and shape [?,180,180,3]
     [[{{node inputs}}]]
2024-02-10 01:10:42.286901: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'inputs' with dtype float and shape [?,22,22,64]
     [[{{node inputs}}]]
2024-02-10 01:10:42.489392: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'inputs' with dtype float and shape [?,180,180,3]
     [[{{node inputs}}]]
2024-02-10 01:10:42.517820: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'inputs' with dtype float and shape [?,180,180,3]
     [[{{node inputs}}]]
2024-02-10 01:10:42.563861: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'inputs' with dtype float and shape [?,180,180,3]
     [[{{node inputs}}]]
WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op, _update_step_xla while saving (showing 4 of 4). These functions will not be directly callable after loading.
INFO:tensorflow:Assets written to: model/flower/saved_model/assets
INFO:tensorflow:Assets written to: model/flower/saved_model/assets
output/A_Close_Up_Photo_of_a_Dandelion.jpg:   0%|          | 0.00/21.7k [00:00<?, ?B/s]
(1, 180, 180, 3)
[1,180,180,3]
This image most likely belongs to dandelion with a 97.96 percent confidence.
../_images/301-tensorflow-training-openvino-nncf-with-output_3_1464.png

インポート

トレーニング後の量子化 API は nncf ライブラリーに実装されています。

import sys

import matplotlib.pyplot as plt
import numpy as np
import nncf
from openvino.runtime import Core
from openvino.runtime import serialize
from PIL import Image
from sklearn.metrics import accuracy_score

sys.path.append("../utils")
from notebook_utils import download_file
INFO:nncf:NNCF initialized successfully. Supported frameworks detected: torch, tensorflow, onnx, openvino

NNCF によるトレーニング後の量子化

NNCF は、精度の低下を最小限に抑えながら、OpenVINO でニューラル・ネットワーク推論を最適化する一連の高度なアルゴリズムを提供します。

事前トレーニングされた FP32 モデルとキャリブレーション・データセットから量子化モデルを作成します。最適化プロセスには次の手順が含まれます。

  1. 量子化用のデータセットを作成します。

  2. 最適化されたモデルを取得するには、nncf.quantize を実行します。

検証データセットはトレーニング・ノートブックですでに定義されています。

img_height = 180
img_width = 180
val_dataset = tf.keras.preprocessing.image_dataset_from_directory(
  data_dir,
  validation_split=0.2,
  subset="validation",
  seed=123,
  image_size=(img_height, img_width),
  batch_size=1
)

for a, b in val_dataset:
    print(type(a), type(b))
    break
Found 3670 files belonging to 5 classes.
Using 734 files for validation.
<class 'tensorflow.python.framework.ops.EagerTensor'> <class 'tensorflow.python.framework.ops.EagerTensor'>
2024-02-10 01:10:45.668839: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'Placeholder/_0' with dtype string and shape [734]
    [[{{node Placeholder/_0}}]]
2024-02-10 01:10:45.669302: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'Placeholder/_0' with dtype string and shape [734]
    [[{{node Placeholder/_0}}]]

また、検証データセットは量子化プロセスで再利用できます。ただし、calibration_dataset は画像のみを返すのに対し、ここではタプル (画像、ラベル) を返します。変換関数は、ユーザー検証データセットをキャリブレーション・データセットに変換するのに役立ちます。

def transform_fn(data_item):
    """
    The transformation function transforms a data item into model input data.
    This function should be passed when the data item cannot be used as model's input.
    """
    images, _ = data_item
    return images.numpy()


calibration_dataset = nncf.Dataset(val_dataset, transform_fn)

中間表現 (IR) モデルをダウンロードします。

core = Core()
ir_model = core.read_model(model_xml)

基本的な量子化フローを使用します。精度制御を備えたモデルに 8 ビットの量子化を適用する最も高度な量子化フローを使用するには、精度制御による量子化を参照してください。

quantized_model = nncf.quantize(
    ir_model,
    calibration_dataset,
    subset_size=1000
)
Output()
Exception in thread Thread-88:
Traceback (most recent call last):
  File "/usr/lib/python3.8/threading.py", line 932, in _bootstrap_inner
    self.run()
  File
"/opt/home/k8sworker/ci-ai/cibuilds/ov-notebook/OVNotebookOps-609/.workspace/scm/ov-notebook/.venv/lib/python3.8/si
te-packages/rich/live.py", line 32, in run
    self.live.refresh()
  File
"/opt/home/k8sworker/ci-ai/cibuilds/ov-notebook/OVNotebookOps-609/.workspace/scm/ov-notebook/.venv/lib/python3.8/si
te-packages/rich/live.py", line 223, in refresh
    self._live_render.set_renderable(self.renderable)
  File
"/opt/home/k8sworker/ci-ai/cibuilds/ov-notebook/OVNotebookOps-609/.workspace/scm/ov-notebook/.venv/lib/python3.8/si
te-packages/rich/live.py", line 203, in renderable
    renderable = self.get_renderable()
  File
"/opt/home/k8sworker/ci-ai/cibuilds/ov-notebook/OVNotebookOps-609/.workspace/scm/ov-notebook/.venv/lib/python3.8/si
te-packages/rich/live.py", line 98, in get_renderable
    self._get_renderable()
  File
"/opt/home/k8sworker/ci-ai/cibuilds/ov-notebook/OVNotebookOps-609/.workspace/scm/ov-notebook/.venv/lib/python3.8/si
te-packages/rich/progress.py", line 1537, in get_renderable
    renderable = Group(*self.get_renderables())
  File
"/opt/home/k8sworker/ci-ai/cibuilds/ov-notebook/OVNotebookOps-609/.workspace/scm/ov-notebook/.venv/lib/python3.8/si
te-packages/rich/progress.py", line 1542, in get_renderables
    table = self.make_tasks_table(self.tasks)
  File
"/opt/home/k8sworker/ci-ai/cibuilds/ov-notebook/OVNotebookOps-609/.workspace/scm/ov-notebook/.venv/lib/python3.8/si
te-packages/rich/progress.py", line 1566, in make_tasks_table
    table.add_row(
  File
"/opt/home/k8sworker/ci-ai/cibuilds/ov-notebook/OVNotebookOps-609/.workspace/scm/ov-notebook/.venv/lib/python3.8/si
te-packages/rich/progress.py", line 1571, in <genexpr>
    else column(task)
  File
"/opt/home/k8sworker/ci-ai/cibuilds/ov-notebook/OVNotebookOps-609/.workspace/scm/ov-notebook/.venv/lib/python3.8/si
te-packages/rich/progress.py", line 528, in __call__
    renderable = self.render(task)
  File
"/opt/home/k8sworker/ci-ai/cibuilds/ov-notebook/OVNotebookOps-609/.workspace/scm/ov-notebook/.venv/lib/python3.8/si
te-packages/nncf/common/logging/track_progress.py", line 58, in render
    text = super().render(task)
  File
"/opt/home/k8sworker/ci-ai/cibuilds/ov-notebook/OVNotebookOps-609/.workspace/scm/ov-notebook/.venv/lib/python3.8/si
te-packages/rich/progress.py", line 787, in render
    task_time = task.time_remaining
  File
"/opt/home/k8sworker/ci-ai/cibuilds/ov-notebook/OVNotebookOps-609/.workspace/scm/ov-notebook/.venv/lib/python3.8/si
te-packages/rich/progress.py", line 1039, in time_remaining
    estimate = ceil(remaining / speed)
  File
"/opt/home/k8sworker/ci-ai/cibuilds/ov-notebook/OVNotebookOps-609/.workspace/scm/ov-notebook/.venv/lib/python3.8/si
te-packages/tensorflow/python/util/traceback_utils.py", line 153, in error_handler
    raise e.with_traceback(filtered_tb) from None
  File
"/opt/home/k8sworker/ci-ai/cibuilds/ov-notebook/OVNotebookOps-609/.workspace/scm/ov-notebook/.venv/lib/python3.8/si
te-packages/tensorflow/python/ops/math_ops.py", line 1569, in _truediv_python3
    raise TypeError(f"`x` and `y` must have the same dtype, "
TypeError: `x` and `y` must have the same dtype, got tf.int64 != tf.float32.
Output()

量子化されたモデルをベンチマークに保存します。

compressed_model_dir = Path("model/optimized")
compressed_model_dir.mkdir(parents=True, exist_ok=True)
compressed_model_xml = compressed_model_dir / "flower_ir.xml"
serialize(quantized_model, str(compressed_model_xml))

推論デバイスの選択

OpenVINO を使用して推論を実行するためにドロップダウン・リストからデバイスを選択します。

import ipywidgets as widgets

device = widgets.Dropdown(
        options=core.available_devices + ["AUTO"],
        value='AUTO',
        description='Device:',
        disabled=False,
)

device
Dropdown(description='Device:', index=1, options=('CPU', 'AUTO'), value='AUTO')

メトリックを比較

モデルのパフォーマンスを判断するメトリックを定義します。

このデモでは、精度メトリックを計算する検証関数を定義します。

def validate(model, validation_loader):
    """
    Evaluate model and compute accuracy metrics.

    :param model: Model to validate
    :param validation_loader: Validation dataset
    :returns: Accuracy scores
    """
    predictions = []
    references = []

    output = model.outputs[0]

    for images, target in validation_loader:
        pred = model(images.numpy())[output]

        predictions.append(np.argmax(pred, axis=1))
        references.append(target)

    predictions = np.concatenate(predictions, axis=0)
    references = np.concatenate(references, axis=0)

    scores = accuracy_score(references, predictions)

    return scores

元のモデルと量子化されたモデルの精度を計算します。

original_compiled_model = core.compile_model(model=ir_model, device_name=device.value)
quantized_compiled_model = core.compile_model(model=quantized_model, device_name=device.value)

original_accuracy = validate(original_compiled_model, val_dataset)
quantized_accuracy = validate(quantized_compiled_model, val_dataset)

print(f"Accuracy of the original model: {original_accuracy:.3f}")
print(f"Accuracy of the quantized model: {quantized_accuracy:.3f}")
Accuracy of the original model: 0.729
Accuracy of the quantized model: 0.729

モデルのファイルサイズを比較します。

original_model_size = model_xml.with_suffix(".bin").stat().st_size / 1024
quantized_model_size = compressed_model_xml.with_suffix(".bin").stat().st_size / 1024

print(f"Original model size: {original_model_size:.2f} KB")
print(f"Quantized model size: {quantized_model_size:.2f} KB")
Original model size: 7791.65 KB
Quantized model size: 3897.08 KB

したがって、元のモデルと量子化されたモデルでは、量子化されたモデルのサイズがはるかに小さいにもかかわらず、同様の精度を持っていることが分かります。

量子化モデルで推論を実行

トレーニング・ノートブックから前処理関数をコピーし、OpenVINO を使用して量子化モデルで推論を実行します。OpenVINO Python API を使用した推論の実行の詳細については、OpenVINO API チュートリアルを参照してください。

def pre_process_image(imagePath, img_height=180):
    # Model input format
    n, c, h, w = [1, 3, img_height, img_height]
    image = Image.open(imagePath)
    image = image.resize((h, w), resample=Image.BILINEAR)

    # Convert to array and change data layout from HWC to CHW
    image = np.array(image)

    input_image = image.reshape((n, h, w, c))

    return input_image
# Get the names of the input and output layer
# model_pot = ie.read_model(model="model/optimized/flower_ir.xml")
input_layer = quantized_compiled_model.input(0)
output_layer = quantized_compiled_model.output(0)

# Get the class names: a list of directory names in alphabetical order
class_names = sorted([item.name for item in Path(data_dir).iterdir() if item.is_dir()])

# Run inference on an input image...
inp_img_url = (
    "https://upload.wikimedia.org/wikipedia/commons/4/48/A_Close_Up_Photo_of_a_Dandelion.jpg"
)
directory = "output"
inp_file_name = "A_Close_Up_Photo_of_a_Dandelion.jpg"
file_path = Path(directory)/Path(inp_file_name)
# Download the image if it does not exist yet
if not Path(inp_file_name).exists():
    download_file(inp_img_url, inp_file_name, directory=directory)

# Pre-process the image and get it ready for inference.
input_image = pre_process_image(imagePath=file_path)
print(f'input image shape: {input_image.shape}')
print(f'input layer shape: {input_layer.shape}')

res = quantized_compiled_model([input_image])[output_layer]

score = tf.nn.softmax(res[0])

# Show the results
image = Image.open(file_path)
plt.imshow(image)
print(
    "This image most likely belongs to {} with a {:.2f} percent confidence.".format(
        class_names[np.argmax(score)], 100 * np.max(score)
    )
)
'output/A_Close_Up_Photo_of_a_Dandelion.jpg' already exists.
input image shape: (1, 180, 180, 3)
input layer shape: [1,180,180,3]
This image most likely belongs to dandelion with a 98.03 percent confidence.
../_images/301-tensorflow-training-openvino-nncf-with-output_27_1.png

推論速度の比較

OpenVINO ベンチマーク・アプリで推論速度を測定します。

ベンチマーク・アプリは、指定された OpenVINO IR モデルの生の推論パフォーマンスを測定するコマンドライン・ツールです。使用可能なパラメーターのリストを表示するには、benchmark_app --help を実行します。デフォルトでは、ベンチマーク・アプリは、CPU 上で非同期推論を使用して、-m パラメーターで指定されたモデルのパフォーマンスを 1 分間テストします。-d パラメーターを使用して、インテル® 統合グラフィックス (iGPU) などのデバイスでパフォーマンスをテストし、-t を使用して推論を実行する秒数を設定します。詳細についてはドキュメントを参照してください。

このチュートリアルでは、ノートブック・ユーティリティーのラッパー関数を使用します。選択したパラメーターを使用して benchmark_app コマンドを出力します。

次のセルでは、CPU 上で元のモデルと量子化されたモデルの推論速度を測定します。iGPU が利用可能な場合、CPU + GPU の推論速度も測定されます。秒数は 15 に設定されています。

注: 最も正確なパフォーマンス推定を行うには、他のアプリケーションを閉じた後、ターミナル/コマンドプロンプトで benchmark_app を実行することを推奨します。

# print the available devices on this system
print("Device information:")
print(core.get_property("CPU", "FULL_DEVICE_NAME"))
if "GPU" in core.available_devices:
    print(core.get_property("GPU", "FULL_DEVICE_NAME"))
Device information:
Intel(R) Core(TM) i9-10920X CPU @ 3.50GHz
# Original model - CPU
! benchmark_app -m $model_xml -d CPU -t 15 -api async
[Step 1/11] Parsing and validating input arguments
[ INFO ] Parsing input parameters
[Step 2/11] Loading OpenVINO Runtime
[ INFO ] OpenVINO:
[ INFO ] Build ................................. 2023.3.0-13775-ceeafaf64f3-releases/2023/3
[ INFO ]
[ INFO ] Device info:
[ INFO ] CPU
[ INFO ] Build ................................. 2023.3.0-13775-ceeafaf64f3-releases/2023/3
[ INFO ]
[ INFO ]
[Step 3/11] Setting device configuration
[ WARNING ] Performance hint was not explicitly specified in command line. Device(CPU) performance hint will be set to PerformanceMode.THROUGHPUT.
[Step 4/11] Reading model files
[ INFO ] Loading model files
[ INFO ] Read model took 12.98 ms
[ INFO ] Original model I/O parameters:
[ INFO ] Model inputs:
[ INFO ]     sequential_1_input (node: sequential_1_input) : f32 / [...] / [1,180,180,3]
[ INFO ] Model outputs:
[ INFO ]     outputs (node: sequential_2/outputs/BiasAdd) : f32 / [...] / [1,5]
[Step 5/11] Resizing model to match image sizes and given batch
[ INFO ] Model batch size: 1
[Step 6/11] Configuring input of the model
[ INFO ] Model inputs:
[ INFO ]     sequential_1_input (node: sequential_1_input) : u8 / [N,H,W,C] / [1,180,180,3]
[ INFO ] Model outputs:
[ INFO ]     outputs (node: sequential_2/outputs/BiasAdd) : f32 / [...] / [1,5]
[Step 7/11] Loading the model to the device
[ INFO ] Compile model took 71.95 ms
[Step 8/11] Querying optimal runtime parameters
[ INFO ] Model:
[ INFO ]   NETWORK_NAME: TensorFlow_Frontend_IR
[ INFO ]   OPTIMAL_NUMBER_OF_INFER_REQUESTS: 12
[ INFO ]   NUM_STREAMS: 12
[ INFO ]   AFFINITY: Affinity.CORE
[ INFO ]   INFERENCE_NUM_THREADS: 24
[ INFO ]   PERF_COUNT: NO
[ INFO ]   INFERENCE_PRECISION_HINT: <Type: 'float32'>
[ INFO ]   PERFORMANCE_HINT: THROUGHPUT
[ INFO ]   EXECUTION_MODE_HINT: ExecutionMode.PERFORMANCE
[ INFO ]   PERFORMANCE_HINT_NUM_REQUESTS: 0
[ INFO ]   ENABLE_CPU_PINNING: True
[ INFO ]   SCHEDULING_CORE_TYPE: SchedulingCoreType.ANY_CORE
[ INFO ]   ENABLE_HYPER_THREADING: True
[ INFO ]   EXECUTION_DEVICES: ['CPU']
[ INFO ]   CPU_DENORMALS_OPTIMIZATION: False
[ INFO ]   CPU_SPARSE_WEIGHTS_DECOMPRESSION_RATE: 1.0
[Step 9/11] Creating infer requests and preparing input tensors
[ WARNING ] No input files were given for input 'sequential_1_input'!. This input will be filled with random values!
[ INFO ] Fill input 'sequential_1_input' with random values
[Step 10/11] Measuring performance (Start inference asynchronously, 12 inference requests, limits: 15000 ms duration)
[ INFO ] Benchmarking in inference only mode (inputs filling are not included in measurement loop).
[ INFO ] First inference took 7.48 ms
[Step 11/11] Dumping statistics report
[ INFO ] Execution Devices:['CPU']
[ INFO ] Count:            57660 iterations
[ INFO ] Duration:         15004.59 ms
[ INFO ] Latency:
[ INFO ]    Median:        2.95 ms
[ INFO ]    Average:       2.95 ms
[ INFO ]    Min:           1.69 ms
[ INFO ]    Max:           12.80 ms
[ INFO ] Throughput:   3842.82 FPS
# Quantized model - CPU
! benchmark_app -m $compressed_model_xml -d CPU -t 15 -api async
[Step 1/11] Parsing and validating input arguments
[ INFO ] Parsing input parameters
[Step 2/11] Loading OpenVINO Runtime
[ INFO ] OpenVINO:
[ INFO ] Build ................................. 2023.3.0-13775-ceeafaf64f3-releases/2023/3
[ INFO ]
[ INFO ] Device info:
[ INFO ] CPU
[ INFO ] Build ................................. 2023.3.0-13775-ceeafaf64f3-releases/2023/3
[ INFO ]
[ INFO ]
[Step 3/11] Setting device configuration
[ WARNING ] Performance hint was not explicitly specified in command line. Device(CPU) performance hint will be set to PerformanceMode.THROUGHPUT.
[Step 4/11] Reading model files
[ INFO ] Loading model files
[ INFO ] Read model took 15.15 ms
[ INFO ] Original model I/O parameters:
[ INFO ] Model inputs:
[ INFO ]     sequential_1_input (node: sequential_1_input) : f32 / [...] / [1,180,180,3]
[ INFO ] Model outputs:
[ INFO ]     outputs (node: sequential_2/outputs/BiasAdd) : f32 / [...] / [1,5]
[Step 5/11] Resizing model to match image sizes and given batch
[ INFO ] Model batch size: 1
[Step 6/11] Configuring input of the model
[ INFO ] Model inputs:
[ INFO ]     sequential_1_input (node: sequential_1_input) : u8 / [N,H,W,C] / [1,180,180,3]
[ INFO ] Model outputs:
[ INFO ]     outputs (node: sequential_2/outputs/BiasAdd) : f32 / [...] / [1,5]
[Step 7/11] Loading the model to the device
[ INFO ] Compile model took 67.57 ms
[Step 8/11] Querying optimal runtime parameters
[ INFO ] Model:
[ INFO ]   NETWORK_NAME: TensorFlow_Frontend_IR
[ INFO ]   OPTIMAL_NUMBER_OF_INFER_REQUESTS: 12
[ INFO ]   NUM_STREAMS: 12
[ INFO ]   AFFINITY: Affinity.CORE
[ INFO ]   INFERENCE_NUM_THREADS: 24
[ INFO ]   PERF_COUNT: NO
[ INFO ]   INFERENCE_PRECISION_HINT: <Type: 'float32'>
[ INFO ]   PERFORMANCE_HINT: THROUGHPUT
[ INFO ]   EXECUTION_MODE_HINT: ExecutionMode.PERFORMANCE
[ INFO ]   PERFORMANCE_HINT_NUM_REQUESTS: 0
[ INFO ]   ENABLE_CPU_PINNING: True
[ INFO ]   SCHEDULING_CORE_TYPE: SchedulingCoreType.ANY_CORE
[ INFO ]   ENABLE_HYPER_THREADING: True
[ INFO ]   EXECUTION_DEVICES: ['CPU']
[ INFO ]   CPU_DENORMALS_OPTIMIZATION: False
[ INFO ]   CPU_SPARSE_WEIGHTS_DECOMPRESSION_RATE: 1.0
[Step 9/11] Creating infer requests and preparing input tensors
[ WARNING ] No input files were given for input 'sequential_1_input'!. This input will be filled with random values!
[ INFO ] Fill input 'sequential_1_input' with random values
[Step 10/11] Measuring performance (Start inference asynchronously, 12 inference requests, limits: 15000 ms duration)
[ INFO ] Benchmarking in inference only mode (inputs filling are not included in measurement loop).
[ INFO ] First inference took 1.99 ms
[Step 11/11] Dumping statistics report
[ INFO ] Execution Devices:['CPU']
[ INFO ] Count:            178152 iterations
[ INFO ] Duration:         15001.85 ms
[ INFO ] Latency:
[ INFO ]    Median:        0.94 ms
[ INFO ]    Average:       0.98 ms
[ INFO ]    Min:           0.55 ms
[ INFO ]    Max:           11.77 ms
[ INFO ] Throughput:   11875.34 FPS

MULTI のベンチマーク: CPU、GPU

最近のインテル CPU では、OpenVINO のマルチデバイス・プラグインを使用して、CPU と iGPU の両方で推論を実行することで、最高のパフォーマンスを実現できる場合があります。GPU でモデルをロードするには CPU よりも少し時間がかかるため、このベンチマークを初めて実行する場合、完了するまでに CPU ベンチマークよりも少し時間がかかります。ベンチマーク・アプリは、--cdir パラメーターを指定することでキャッシュをサポートします。下のセルでは、モデルは model_cache ディレクトリーにキャッシュされます。

# Original model - MULTI:CPU,GPU
if "GPU" in core.available_devices:
    ! benchmark_app -m $model_xml -d MULTI:CPU,GPU -t 15 -api async
else:
    print("A supported integrated GPU is not available on this system.")
A supported integrated GPU is not available on this system.
# Quantized model - MULTI:CPU,GPU
if "GPU" in core.available_devices:
    ! benchmark_app -m $compressed_model_xml -d MULTI:CPU,GPU -t 15 -api async
else:
    print("A supported integrated GPU is not available on this system.")
A supported integrated GPU is not available on this system.
# print the available devices on this system
print("Device information:")
print(core.get_property("CPU", "FULL_DEVICE_NAME"))
if "GPU" in core.available_devices:
    print(core.get_property("GPU", "FULL_DEVICE_NAME"))
Device information:
Intel(R) Core(TM) i9-10920X CPU @ 3.50GHz

元の IR モデル - CPU

benchmark_output = %sx benchmark_app -m $model_xml -t 15 -api async
# Remove logging info from benchmark_app output and show only the results
benchmark_result = benchmark_output[-8:]
print("\n".join(benchmark_result))
[ INFO ] Count:            57840 iterations
[ INFO ] Duration:         15004.24 ms
[ INFO ] Latency:
[ INFO ]    Median:        2.94 ms
[ INFO ]    Average:       2.94 ms
[ INFO ]    Min:           1.98 ms
[ INFO ]    Max:           12.12 ms
[ INFO ] Throughput:   3854.91 FPS

量子化された IR モデル - CPU

benchmark_output = %sx benchmark_app -m $compressed_model_xml -t 15 -api async
# Remove logging info from benchmark_app output and show only the results
benchmark_result = benchmark_output[-8:]
print("\n".join(benchmark_result))
[ INFO ] Count:            178836 iterations
[ INFO ] Duration:         15001.19 ms
[ INFO ] Latency:
[ INFO ]    Median:        0.94 ms
[ INFO ]    Average:       0.97 ms
[ INFO ]    Min:           0.58 ms
[ INFO ]    Max:           6.85 ms
[ INFO ] Throughput:   11921.45 FPS

元の IR モデル - MULTI: CPU、GPU

最近のインテル CPU では、OpenVINO のマルチデバイス・プラグインを使用して、CPU と iGPU の両方で推論を実行することで、最高のパフォーマンスを実現できる場合があります。GPU でモデルをロードするには CPU よりも少し時間がかかるため、このベンチマークは CPU ベンチマークよりも完了するまでに少し時間を要します。

if "GPU" in core.available_devices:
    benchmark_output = %sx benchmark_app -m $model_xml -d MULTI:CPU,GPU -t 15 -api async
    # Remove logging info from benchmark_app output and show only the results
    benchmark_result = benchmark_output[-8:]
    print("\n".join(benchmark_result))
else:
    print("GPU is not available on this system.")
 GPU is not available on this system.

量子化された IR モデル - MULTI: CPU、GPU

if "GPU" in core.available_devices:
    benchmark_output = %sx benchmark_app -m $compressed_model_xml -d MULTI:CPU,GPU -t 15 -api async
    # Remove logging info from benchmark_app output and show only the results
    benchmark_result = benchmark_output[-8:]
    print("\n".join(benchmark_result))
else:
    print("GPU is not available on this system.")
 GPU is not available on this system.