From 42df74738c3586f799d243b6c0a398a3bf0fcd88 Mon Sep 17 00:00:00 2001
From: Jan <jan.mandel@gmail.com>
Date: Sat, 2 Jul 2022 13:26:29 -0600
Subject: [PATCH] consolidate creating RNN

---
 fmda_kf_rnn.ipynb | 87 ++++++++++++++-----------------------------------------
 1 file changed, 21 insertions(+), 66 deletions(-)

diff --git a/fmda_kf_rnn.ipynb b/fmda_kf_rnn.ipynb
index 9077786..1db8668 100644
--- a/fmda_kf_rnn.ipynb
+++ b/fmda_kf_rnn.ipynb
@@ -1395,7 +1395,7 @@
         "    y = tf.keras.layers.Dense(hidden_units, activation=activation[1])(x)\n",
         "    outputs = tf.keras.layers.Dense(dense_units, activation=activation[1])(y)\n",
         "    model = tf.keras.Model(inputs=inputs, outputs=outputs)\n",
-        "    model.compile(loss='mean_squared_error', optimizer='sgd')\n",
+        "    model.compile(loss='mean_squared_error', optimizer='adam')\n",
         "    return model"
       ],
       "metadata": {
@@ -1412,84 +1412,39 @@
       },
       "outputs": [],
       "source": [
-        "fmda_model=create_RNN_2(hidden_units=7, dense_units=1, \n",
-        "                        batch_shape=(samples, timesteps, features),\n",
-        "                        input_shape=(timesteps, features),\n",
-        "                        stateful = True,\n",
-        "                        activation=['tanh', 'tanh'])\n",
-        "print(fmda_model.summary())"
+        "def create_fit_predict_RNN(hidden_units, dense_units, \n",
+        "                             samples, timesteps, features, activation):\n",
+        "  # statefull model version with with fixed number of batches\n",
+        "  model_fit=create_RNN_2(hidden_units=hidden_units, dense_units=dense_units, \n",
+        "                        batch_shape=(samples, timesteps, features),stateful = True,\n",
+        "                        activation=activation)\n",
+        "  print(model_fit.summary())\n",
+        "  # same model for prediction on the entire dataset\n",
+        "  model_predict=create_RNN_2(hidden_units=hidden_units, dense_units=dense_units,  \n",
+        "                        input_shape=(None,features),stateful = False,\n",
+        "                        activation=activation)\n",
+        "  print(model_predict.summary())\n",
+        "  return model_fit, model_predict\n",
+        "\n",
+        "fmda_model, fmda_model_eval = create_fit_predict_RNN(hidden_units=7, dense_units=1, \n",
+        "                        samples=samples, timesteps=timesteps, features=1, \n",
+        "                        activation=['tanh', 'tanh'])"
       ]
     },
     {
       "cell_type": "code",
       "source": [
-        "fmda_model_eval=create_RNN_2(hidden_units=7, dense_units=1, \n",
-        "                        batch_shape=(samples, timesteps, features),\n",
-        "                        input_shape=(None,features),\n",
-        "                        stateful = False,\n",
-        "                        activation=['tanh', 'tanh'])\n",
-        "print(fmda_model_eval.summary())"
-      ],
-      "metadata": {
-        "id": "Bq-e6mfZLfnE"
-      },
-      "execution_count": null,
-      "outputs": []
-    },
-    {
-      "cell_type": "code",
-      "source": [
-        "fmda_model.fit(x_train, y_train, epochs=40, verbose=2,batch_size=samples)"
-      ],
-      "metadata": {
-        "id": "VZoa3tlQWbBG"
-      },
-      "execution_count": null,
-      "outputs": []
-    },
-    {
-      "cell_type": "markdown",
-      "source": [
-        ""
-      ],
-      "metadata": {
-        "id": "Rl5lbJ7kM3tr"
-      }
-    },
-    {
-      "cell_type": "code",
-      "source": [
+        "fmda_model.fit(x_train, y_train, epochs=40, verbose=2,batch_size=samples)\n",
         "# Same model as stateless for prediction:\n",
         "w=fmda_model.get_weights()\n",
         "fmda_model_eval.set_weights(w)\n",
         "# prediction on the entire dataset from zero state\n",
         "mt = fmda_model_eval.predict(Et)\n",
-        "m = scalery.inverse_transform(mt)"
-      ],
-      "metadata": {
-        "id": "R2jkoZlAIaSb"
-      },
-      "execution_count": null,
-      "outputs": []
-    },
-    {
-      "cell_type": "code",
-      "source": [
+        "m = scalery.inverse_transform(mt)\n",
         "plot_m(m,title='RNN prediction')"
       ],
       "metadata": {
-        "id": "eTeOP9I7sBJ4"
-      },
-      "execution_count": null,
-      "outputs": []
-    },
-    {
-      "cell_type": "code",
-      "source": [
-        ""
-      ],
-      "metadata": {
-        "id": "jqlUYh8Zt8Cc"
+        "id": "R2jkoZlAIaSb"
       },
       "execution_count": null,
       "outputs": []
-- 
2.11.4.GIT