ctx=ctx)
File "O:\PycharmProjects\catdogtf2.2\venv\lib\site-packages\tensorflow\python\eager\execute.py", line 60, in quick_execute
inputs, attrs, num_outputs)
tensorflow.python.framework.errors_impl.InternalError: Blas GEMM launch failed : a.shape=(32, 784), b.shape=(784, 512), m=32, n=512, k=784
[[node sequential/dense/MatMul (defined at O:/PycharmProjects/catdogtf2.2/006.py:51) ]] [Op:__inference_train_function_589]
Function call stack:
train_function
O:\PycharmProjects\catdogtf2.2\venv\Scripts\python.exe O:\PyCharm\plugins\python\helpers\pydev\pydevconsole.py --mode=client --port=54455
import sys; print('Python %s on %s' % (sys.version, sys.platform))
sys.path.extend(['O:\\PycharmProjects\\catdogtf2.2', 'O:/PycharmProjects/catdogtf2.2'])
PyDev console: starting.
Python 3.7.7 (default, May 6 2020, 11:45:54) [MSC v.1916 64 bit (AMD64)] on win32
>>> runfile('O:/PycharmProjects/catdogtf2.2/006.py', wdir='O:/PycharmProjects/catdogtf2.2')
2020-08-11 05:29:58.544667: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cudart64_101.dll
2.2.0
2020-08-11 05:30:01.282588: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library nvcuda.dll
2020-08-11 05:30:01.324897: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1561] Found device 0 with properties:
pciBusID: 0000:09:00.0 name: GeForce RTX 2080 SUPER computeCapability: 7.5
coreClock: 1.845GHz coreCount: 48 deviceMemorySize: 8.00GiB deviceMemoryBandwidth: 462.00GiB/s
2020-08-11 05:30:01.325397: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cudart64_101.dll
2020-08-11 05:30:01.333925: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cublas64_10.dll
2020-08-11 05:30:01.339713: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cufft64_10.dll
2020-08-11 05:30:01.342729: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library curand64_10.dll
2020-08-11 05:30:01.350152: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cusolver64_10.dll
2020-08-11 05:30:01.354987: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cusparse64_10.dll
2020-08-11 05:30:01.382853: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cudnn64_7.dll
2020-08-11 05:30:01.383178: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1703] Adding visible gpu devices: 0
2020-08-11 05:30:01.383630: I tensorflow/core/platform/cpu_feature_guard.cc:143] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2
2020-08-11 05:30:01.397994: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x1eabe93ddf0 initialized for platform Host (this does not guarantee that XLA will be used). Devices:
2020-08-11 05:30:01.398331: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): Host, Default Version
2020-08-11 05:30:01.398700: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1561] Found device 0 with properties:
pciBusID: 0000:09:00.0 name: GeForce RTX 2080 SUPER computeCapability: 7.5
coreClock: 1.845GHz coreCount: 48 deviceMemorySize: 8.00GiB deviceMemoryBandwidth: 462.00GiB/s
2020-08-11 05:30:01.399211: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cudart64_101.dll
2020-08-11 05:30:01.399492: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cublas64_10.dll
2020-08-11 05:30:01.399779: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cufft64_10.dll
2020-08-11 05:30:01.400141: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library curand64_10.dll
2020-08-11 05:30:01.400451: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cusolver64_10.dll
2020-08-11 05:30:01.400725: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cusparse64_10.dll
2020-08-11 05:30:01.401037: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cudnn64_7.dll
2020-08-11 05:30:01.401408: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1703] Adding visible gpu devices: 0
2020-08-11 05:30:02.056830: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1102] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-08-11 05:30:02.057100: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1108] 0
2020-08-11 05:30:02.057241: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1121] 0: N
2020-08-11 05:30:02.057523: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1247] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 6198 MB memory) -> physical GPU (device: 0, name: GeForce RTX 2080 SUPER, pci bus id: 0000:09:00.0, compute capability: 7.5)
2020-08-11 05:30:02.061005: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x1ea874059e0 initialized for platform CUDA (this does not guarantee that XLA will be used). Devices:
2020-08-11 05:30:02.061296: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): GeForce RTX 2080 SUPER, Compute Capability 7.5
Model: "sequential"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
dense (Dense) (None, 512) 401920
_________________________________________________________________
dropout (Dropout) (None, 512) 0
_________________________________________________________________
dense_1 (Dense) (None, 10) 5130
=================================================================
Total params: 407,050
Trainable params: 407,050
Non-trainable params: 0
_________________________________________________________________
Epoch 1/10
2020-08-11 05:30:02.840239: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cublas64_10.dll
28/32 [=========================>....] - ETA: 0s - loss: 1.2189 - accuracy: 0.6618
Epoch 00001: saving model to training_1/cp.ckpt
32/32 [==============================] - 0s 13ms/step - loss: 1.1645 - accuracy: 0.6750 - val_loss: 0.7062 - val_accuracy: 0.7760
Epoch 2/10
24/32 [=====================>........] - ETA: 0s - loss: 0.4462 - accuracy: 0.8750
Epoch 00002: saving model to training_1/cp.ckpt
32/32 [==============================] - 0s 7ms/step - loss: 0.4426 - accuracy: 0.8740 - val_loss: 0.5803 - val_accuracy: 0.8140
Epoch 3/10
24/32 [=====================>........] - ETA: 0s - loss: 0.2971 - accuracy: 0.9180
Epoch 00003: saving model to training_1/cp.ckpt
32/32 [==============================] - 0s 7ms/step - loss: 0.2959 - accuracy: 0.9170 - val_loss: 0.4929 - val_accuracy: 0.8440
Epoch 4/10
22/32 [===================>..........] - ETA: 0s - loss: 0.2194 - accuracy: 0.9489
Epoch 00004: saving model to training_1/cp.ckpt
32/32 [==============================] - 0s 8ms/step - loss: 0.2136 - accuracy: 0.9510 - val_loss: 0.4599 - val_accuracy: 0.8460
Epoch 5/10
22/32 [===================>..........] - ETA: 0s - loss: 0.1591 - accuracy: 0.9631
Epoch 00005: saving model to training_1/cp.ckpt
32/32 [==============================] - 0s 7ms/step - loss: 0.1630 - accuracy: 0.9630 - val_loss: 0.4229 - val_accuracy: 0.8590
Epoch 6/10
25/32 [======================>.......] - ETA: 0s - loss: 0.1330 - accuracy: 0.9700
Epoch 00006: saving model to training_1/cp.ckpt
32/32 [==============================] - 0s 7ms/step - loss: 0.1269 - accuracy: 0.9710 - val_loss: 0.4153 - val_accuracy: 0.8710
Epoch 7/10
26/32 [=======================>......] - ETA: 0s - loss: 0.0856 - accuracy: 0.9844
Epoch 00007: saving model to training_1/cp.ckpt
32/32 [==============================] - 0s 7ms/step - loss: 0.0904 - accuracy: 0.9840 - val_loss: 0.4451 - val_accuracy: 0.8570
Epoch 8/10
23/32 [====================>.........] - ETA: 0s - loss: 0.0675 - accuracy: 0.9959
Epoch 00008: saving model to training_1/cp.ckpt
32/32 [==============================] - 0s 7ms/step - loss: 0.0716 - accuracy: 0.9950 - val_loss: 0.4232 - val_accuracy: 0.8630
Epoch 9/10
25/32 [======================>.......] - ETA: 0s - loss: 0.0572 - accuracy: 0.9975
Epoch 00009: saving model to training_1/cp.ckpt
32/32 [==============================] - 0s 7ms/step - loss: 0.0555 - accuracy: 0.9980 - val_loss: 0.4201 - val_accuracy: 0.8640
Epoch 10/10
24/32 [=====================>........] - ETA: 0s - loss: 0.0416 - accuracy: 1.0000
Epoch 00010: saving model to training_1/cp.ckpt
32/32 [==============================] - 0s 7ms/step - loss: 0.0432 - accuracy: 0.9980 - val_loss: 0.4192 - val_accuracy: 0.8590
32/32 - 0s - loss: 2.3390 - accuracy: 0.1360
훈련되지 않은 모델의 정확도: 13.60%
32/32 - 0s - loss: 0.4192 - accuracy: 0.8590
복원된 모델의 정확도: 85.90%
WARNING:tensorflow:`period` argument is deprecated. Please use `save_freq` to specify the frequency in number of batches seen.
WARNING:tensorflow:Unresolved object in checkpoint: (root).optimizer.iter
WARNING:tensorflow:Unresolved object in checkpoint: (root).optimizer.beta_1
WARNING:tensorflow:Unresolved object in checkpoint: (root).optimizer.beta_2
WARNING:tensorflow:Unresolved object in checkpoint: (root).optimizer.decay
WARNING:tensorflow:Unresolved object in checkpoint: (root).optimizer.learning_rate
WARNING:tensorflow:A checkpoint was restored (e.g. tf.train.Checkpoint.restore or tf.keras.Model.load_weights) but not all checkpointed values were used. See above for specific issues. Use expect_partial() on the load status object, e.g. tf.train.Checkpoint.restore(...).expect_partial(), to silence these warnings, or use assert_consumed() to make the check explicit. See https://www.tensorflow.org/guide/checkpoint#loading_mechanics for details.
Epoch 00005: saving model to training_2/cp-0005.ckpt
Epoch 00010: saving model to training_2/cp-0010.ckpt
Epoch 00015: saving model to training_2/cp-0015.ckpt
Epoch 00020: saving model to training_2/cp-0020.ckpt
Epoch 00025: saving model to training_2/cp-0025.ckpt
Epoch 00030: saving model to training_2/cp-0030.ckpt
Epoch 00035: saving model to training_2/cp-0035.ckpt
Epoch 00040: saving model to training_2/cp-0040.ckpt
Epoch 00045: saving model to training_2/cp-0045.ckpt
Epoch 00050: saving model to training_2/cp-0050.ckpt
32/32 - 0s - loss: 0.4963 - accuracy: 0.8730
복원된 모델의 정확도: 87.30%
WARNING:tensorflow:Unresolved object in checkpoint: (root).optimizer.iter
WARNING:tensorflow:Unresolved object in checkpoint: (root).optimizer.beta_1
WARNING:tensorflow:Unresolved object in checkpoint: (root).optimizer.beta_2
WARNING:tensorflow:Unresolved object in checkpoint: (root).optimizer.decay
WARNING:tensorflow:Unresolved object in checkpoint: (root).optimizer.learning_rate
WARNING:tensorflow:A checkpoint was restored (e.g. tf.train.Checkpoint.restore or tf.keras.Model.load_weights) but not all checkpointed values were used. See above for specific issues. Use expect_partial() on the load status object, e.g. tf.train.Checkpoint.restore(...).expect_partial(), to silence these warnings, or use assert_consumed() to make the check explicit. See https://www.tensorflow.org/guide/checkpoint#loading_mechanics for details.
32/32 - 0s - loss: 0.4963 - accuracy: 0.8730
복원된 모델의 정확도: 87.30%
Epoch 1/5
32/32 [==============================] - 0s 2ms/step - loss: 1.1610 - accuracy: 0.6710
Epoch 2/5
32/32 [==============================] - 0s 3ms/step - loss: 0.4298 - accuracy: 0.8760
Epoch 3/5
32/32 [==============================] - 0s 2ms/step - loss: 0.2997 - accuracy: 0.9150
Epoch 4/5
32/32 [==============================] - 0s 2ms/step - loss: 0.2052 - accuracy: 0.9560
Epoch 5/5
32/32 [==============================] - 0s 2ms/step - loss: 0.1526 - accuracy: 0.9670
2020-08-11 05:30:14.908899: W tensorflow/python/util/util.cc:329] Sets are not currently considered sequences, but this may change in the future, so consider avoiding using them.
WARNING:tensorflow:From O:\PycharmProjects\catdogtf2.2\venv\lib\site-packages\tensorflow\python\ops\resource_variable_ops.py:1817: calling BaseResourceVariable.__init__ (from tensorflow.python.ops.resource_variable_ops) with constraint is deprecated and will be removed in a future version.
Instructions for updating:
If using Keras pass *_constraint arguments to layers.
Model: "sequential_5"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
dense_10 (Dense) (None, 512) 401920
_________________________________________________________________
dropout_5 (Dropout) (None, 512) 0
_________________________________________________________________
dense_11 (Dense) (None, 10) 5130
=================================================================
Total params: 407,050
Trainable params: 407,050
Non-trainable params: 0
_________________________________________________________________
32/32 - 0s - loss: 0.4585 - accuracy: 0.8460
복원된 모델의 정확도: 84.60%
(1000, 10)
Epoch 1/5
32/32 [==============================] - 0s 2ms/step - loss: 1.1752 - accuracy: 0.6670
Epoch 2/5
32/32 [==============================] - 0s 2ms/step - loss: 0.4129 - accuracy: 0.8900
Epoch 3/5
32/32 [==============================] - 0s 2ms/step - loss: 0.2810 - accuracy: 0.9220
Epoch 4/5
32/32 [==============================] - 0s 2ms/step - loss: 0.2027 - accuracy: 0.9490
Epoch 5/5
32/32 [==============================] - 0s 2ms/step - loss: 0.1418 - accuracy: 0.9770
Model: "sequential_6"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
dense_12 (Dense) (None, 512) 401920
_________________________________________________________________
dropout_6 (Dropout) (None, 512) 0
_________________________________________________________________
dense_13 (Dense) (None, 10) 5130
=================================================================
Total params: 407,050
Trainable params: 407,050
Non-trainable params: 0
_________________________________________________________________
32/32 - 0s - loss: 0.4315 - accuracy: 0.8680
복원된 모델의 정확도: 86.80%
#pip install -q pyyaml h5py # HDF5 포맷으로 모델을 저장하기 위해서 필요합니다
import os
import tensorflow as tf
from tensorflow import keras
print(tf.version.VERSION)
(train_images, train_labels), (test_images, test_labels) = tf.keras.datasets.mnist.load_data()
train_labels = train_labels[:1000]
test_labels = test_labels[:1000]
train_images = train_images[:1000].reshape(-1, 28 * 28) / 255.0
test_images = test_images[:1000].reshape(-1, 28 * 28) / 255.0
# 간단한 Sequential 모델을 정의합니다
def create_model():
model = tf.keras.models.Sequential([
keras.layers.Dense(512, activation='relu', input_shape=(784,)),
keras.layers.Dropout(0.2),
keras.layers.Dense(10)
])
model.compile(optimizer='adam',
loss=tf.losses.SparseCategoricalCrossentropy(from_logits=True),
metrics=['accuracy'])
return model
# 모델 객체를 만듭니다
model = create_model()
# 모델 구조를 출력합니다
model.summary()
checkpoint_path = "training_1/cp.ckpt"
checkpoint_dir = os.path.dirname(checkpoint_path)
# 모델의 가중치를 저장하는 콜백 만들기
cp_callback = tf.keras.callbacks.ModelCheckpoint(filepath=checkpoint_path,
save_weights_only=True,
verbose=1)
# 새로운 콜백으로 모델 훈련하기
model.fit(train_images,
train_labels,
epochs=10,
validation_data=(test_images,test_labels),
callbacks=[cp_callback]) # 콜백을 훈련에 전달합니다
# 옵티마이저의 상태를 저장하는 것과 관련되어 경고가 발생할 수 있습니다.
# 이 경고는 (그리고 이 노트북의 다른 비슷한 경고는) 이전 사용 방식을 권장하지 않기 위함이며 무시해도 좋습니다.
#ls {checkpoint_dir}
# 기본 모델 객체를 만듭니다
model = create_model()
# 모델을 평가합니다
loss, acc = model.evaluate(test_images, test_labels, verbose=2)
print("훈련되지 않은 모델의 정확도: {:5.2f}%".format(100*acc))
# 가중치 로드
model.load_weights(checkpoint_path)
# 모델 재평가
loss,acc = model.evaluate(test_images, test_labels, verbose=2)
print("복원된 모델의 정확도: {:5.2f}%".format(100*acc))
# 파일 이름에 에포크 번호를 포함시킵니다(`str.format` 포맷)
checkpoint_path = "training_2/cp-{epoch:04d}.ckpt"
checkpoint_dir = os.path.dirname(checkpoint_path)
# 다섯 번째 에포크마다 가중치를 저장하기 위한 콜백을 만듭니다
cp_callback = tf.keras.callbacks.ModelCheckpoint(
filepath=checkpoint_path,
verbose=1,
save_weights_only=True,
period=5)
# 새로운 모델 객체를 만듭니다
model = create_model()
# `checkpoint_path` 포맷을 사용하는 가중치를 저장합니다
model.save_weights(checkpoint_path.format(epoch=0))
# 새로운 콜백을 사용하여 모델을 훈련합니다
model.fit(train_images,
train_labels,
epochs=50,
callbacks=[cp_callback],
validation_data=(test_images,test_labels),
verbose=0)
#ls {checkpoint_dir}
latest = tf.train.latest_checkpoint(checkpoint_dir)
latest
# 새로운 모델 객체를 만듭니다
model = create_model()
# 이전에 저장한 가중치를 로드합니다
model.load_weights(latest)
# 모델을 재평가합니다
loss, acc = model.evaluate(test_images, test_labels, verbose=2)
print("복원된 모델의 정확도: {:5.2f}%".format(100*acc))
# 가중치를 저장합니다
model.save_weights('./checkpoints/my_checkpoint')
# 새로운 모델 객체를 만듭니다
model = create_model()
# 가중치를 복원합니다
model.load_weights('./checkpoints/my_checkpoint')
# 모델을 평가합니다
loss,acc = model.evaluate(test_images, test_labels, verbose=2)
print("복원된 모델의 정확도: {:5.2f}%".format(100*acc))
# 새로운 모델 객체를 만들고 훈련합니다
model = create_model()
model.fit(train_images, train_labels, epochs=5)
# SavedModel로 전체 모델을 저장합니다
#!mkdir -p saved_model
model.save('saved_model/my_model')
# my_model 디렉토리
#!ls saved_model
# assests 폴더, saved_model.pb, variables 폴더
#!ls saved_model/my_model
new_model = tf.keras.models.load_model('saved_model/my_model')
# 모델 구조를 확인합니다
new_model.summary()
# 복원된 모델을 평가합니다
loss, acc = new_model.evaluate(test_images, test_labels, verbose=2)
print('복원된 모델의 정확도: {:5.2f}%'.format(100*acc))
print(new_model.predict(test_images).shape)
# 새로운 모델 객체를 만들고 훈련합니다
model = create_model()
model.fit(train_images, train_labels, epochs=5)
# 전체 모델을 HDF5 파일로 저장합니다
# '.h5' 확장자는 이 모델이 HDF5로 저장되었다는 것을 나타냅니다
model.save('my_model.h5')
# 가중치와 옵티마이저를 포함하여 정확히 동일한 모델을 다시 생성합니다
new_model = tf.keras.models.load_model('my_model.h5')
# 모델 구조를 출력합니다
new_model.summary()
loss, acc = new_model.evaluate(test_images, test_labels, verbose=2)
print('복원된 모델의 정확도: {:5.2f}%'.format(100*acc))
#
# Copyright (c) 2017 François Chollet
#
# Permission is hereby granted, free of charge, to any person obtaining a
# copy of this software and associated documentation files (the "Software"),
# to deal in the Software without restriction, including without limitation
# the rights to use, copy, modify, merge, publish, distribute, sublicense,
# and/or sell copies of the Software, and to permit persons to whom the
# Software is furnished to do so, subject to the following conditions:
#
# The above copyright notice and this permission notice shall be included in
# all copies or substantial portions of the Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
# THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
# FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
# DEALINGS IN THE SOFTWARE.
'진행 프로젝트 > [진행] Tensorflow2 "해볼까?"' 카테고리의 다른 글
| tutorials 08 (0) | 2020.08.11 |
|---|---|
| tutorials 07 (0) | 2020.08.11 |
| tutorials 05 (0) | 2020.08.11 |
| tutorials 04 (0) | 2020.08.11 |
| tutorials 03 (0) | 2020.08.11 |