Merge pull request dreoporto#6 from dreoporto/caching-save-history

Caching save history

ptmlib/examples/computer_vision_caching.py

@@ -9,7 +9,7 @@
 from tensorflow.keras import layers
 
 import ptmlib.model_tools as modt
-
+import ptmlib.charts as pch
 
 class MyCallback(keras.callbacks.Callback):
 
@@ -103,6 +103,13 @@ def main():
     print(test_labels[0])
     print(max(classifications[0]))
 
+    # ensure history data is still available, even if cached
+    pch.show_history_chart(history, "accuracy")  # render again to be sure we have proper history data
+    print('type(model):', type(model))
+    print('type(history):', type(history))
+    print('history.history:', history.history)
+    print('history.params:', history.params)
+
 
 if __name__ == '__main__':
     main()

ptmlib/model_tools.py

@@ -1,4 +1,5 @@
 import os
+import pickle
 from typing import Any, List
 
 import matplotlib.image as mpimg
@@ -8,6 +9,8 @@
 import ptmlib.charts as pch
 from ptmlib.time import Stopwatch
 
+HISTORY_FILE_SUFFIX_EXTENSION = '_history.pkl'
+
 
 def get_file_path(model_file_name: str, model_file_format: str = ""):
     extension = _get_model_file_extension(model_file_format)
@@ -36,13 +39,12 @@ def load_or_fit_model(model: Any, model_file_name: str, x: Any, y: Any = None, v
                       model_file_format: str = "",
                       load_model_function=_default_load_model_function,
                       fit_model_function=_default_fit_model_function):
-    history = None
-
     file_extension = _get_model_file_extension(model_file_format)
 
     if os.path.exists(f'{model_file_name}{file_extension}'):
         print(f'Loading existing model file: {model_file_name}{file_extension}')
         model = load_model_function(model_file_name, model_file_format)
+        history = load_history_data(model_file_name)
         if images_enabled:
             _show_saved_images(metrics, model_file_name, fig_size)
     else:
@@ -52,12 +54,35 @@ def load_or_fit_model(model: Any, model_file_name: str, x: Any, y: Any = None, v
         stopwatch.stop()
         print(f'Saving new model file: {model_file_name}{file_extension}')
         model.save(f'{model_file_name}{file_extension}')
+        save_history_data(history, model_file_name)
         if images_enabled:
             _show_new_images(history, model_file_name, metrics)
 
     return model, history
 
 
+def save_history_data(history: Any, model_file_name: str):
+    with open(f'{model_file_name}{HISTORY_FILE_SUFFIX_EXTENSION}', 'wb') as history_file:
+        history_params_tuple = (history.history, history.params)
+        pickle.dump(history_params_tuple, history_file)
+
+
+def load_history_data(model_file_name: str):
+    if not os.path.exists(f'{model_file_name}{HISTORY_FILE_SUFFIX_EXTENSION}'):
+        return None
+
+    # create new history object for return value
+    history = keras.callbacks.History()
+
+    with open(f'{model_file_name}{HISTORY_FILE_SUFFIX_EXTENSION}', 'rb') as history_file:
+        # load from previously saved history_params_tuple
+        h, p = pickle.load(history_file)
+        history.history = h
+        history.params = p
+
+    return history
+
+
 def _show_new_images(history: Any, model_file_name: str, metrics: List[str]):
     if metrics is not None:
         for metric in metrics: