search

jeffheaton / t81_558_deep_learning

T81-558: Keras - Applications of Deep Neural Networks @Washington University in St. Louis
https://sites.wustl.edu/jeffheaton/t81-558/
Other
5.71k stars 3.04k forks source link

t81_558_class_14_04_ids_kdd99: Epoch 00019: early stopping #86

Closed goish135 closed 3 years ago

goish135 commented 3 years ago

I don't know why early stopping . Is it Number of epochs with no improvement after which training will be stopped ?

And my result : 

Epoch 1/1000
2020-12-15 20:08:56.660904: I tensorflow/core/platform/profile_utils/cpu_utils.cc:94] CPU Frequency: 3392525000 Hz
2020-12-15 20:08:56.661195: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x77f4c30 executing computations on platform Host. Devices:
2020-12-15 20:08:56.661213: I tensorflow/compiler/xla/service/service.cc:175]   StreamExecutor device (0): <undefined>, <undefined>
2020-12-15 20:08:56.696438: W tensorflow/compiler/jit/mark_for_compilation_pass.cc:1412] (One-time warning): Not using XLA:CPU for cluster because envvar TF_XLA_FLAGS=--tf_xla_cpu_global_jit was not set.  If you want XLA:CPU, either set that envvar, or use experimental_jit_scope to enable XLA:CPU.  To confirm that XLA is active, pass --vmodule=xla_compilation_cache=1 (as a proper command-line flag, not via TF_XLA_FLAGS) or set the envvar XLA_FLAGS=--xla_hlo_profile.
370515/370515 - 13s - loss: 0.2022 - val_loss: 0.0460
Epoch 2/1000
370515/370515 - 12s - loss: 0.0500 - val_loss: 0.0332
Epoch 3/1000
370515/370515 - 12s - loss: 0.0509 - val_loss: 0.0302
Epoch 4/1000
370515/370515 - 13s - loss: 0.0283 - val_loss: 0.0269
Epoch 5/1000
370515/370515 - 13s - loss: 0.0436 - val_loss: 0.0256
Epoch 6/1000
370515/370515 - 13s - loss: 0.0247 - val_loss: 0.0254
Epoch 7/1000
370515/370515 - 13s - loss: 0.0335 - val_loss: 0.0250
Epoch 8/1000
370515/370515 - 13s - loss: 0.0224 - val_loss: 0.0248
Epoch 9/1000
370515/370515 - 13s - loss: 0.0274 - val_loss: 0.0231
Epoch 10/1000
370515/370515 - 13s - loss: 0.0214 - val_loss: 0.0226
Epoch 11/1000
370515/370515 - 13s - loss: 0.0230 - val_loss: 0.0226
Epoch 12/1000
370515/370515 - 13s - loss: 0.0200 - val_loss: 0.0221
Epoch 13/1000
370515/370515 - 13s - loss: 0.0198 - val_loss: 0.0236
Epoch 14/1000
370515/370515 - 13s - loss: 0.0195 - val_loss: 0.0208
Epoch 15/1000
370515/370515 - 13s - loss: 0.0237 - val_loss: 0.0257
Epoch 16/1000
370515/370515 - 13s - loss: 0.0184 - val_loss: 0.0218
Epoch 17/1000
370515/370515 - 13s - loss: 0.0184 - val_loss: 0.0217
Epoch 18/1000
370515/370515 - 13s - loss: 0.0187 - val_loss: 0.0215
Epoch 19/1000
Restoring model weights from the end of the best epoch.
370515/370515 - 12s - loss: 0.0175 - val_loss: 0.0240
Epoch 00019: early stopping
Validation score: 0.9969070328567033
goish135 commented 3 years ago

Hi : I know why XD https://www.tensorflow.org/api_docs/python/tf/keras/callbacks/EarlyStopping