Update scratchbook with upsampling example

5f92f14e · Jean-Luc Parouty · 0814d097 · 5f92f14e
Commit 5f92f14e authored 3 years ago by Jean-Luc Parouty
--- a/Misc/Scratchbook.ipynb
+++ b/Misc/Scratchbook.ipynb
@@ -233,10 +233,71 @@
    "    print(f'#{i} : {x} => {y}')"
   ]
  },
+  {
+   "cell_type": "markdown",
+   "id": "67e3c888-aaa4-4166-90a1-cdb63920fd7d",
+   "metadata": {},
+   "source": [
+    "## 3 - Upsampling"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 42,
+   "id": "20f12cf0-1fdb-4b53-92c6-d03b140e42d1",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "Initial :  (2, 2)\n",
+      "[[1 2]\n",
+      " [3 4]]\n",
+      "\n",
+      "Reshape as a batch of (2,2) vectors :  (1, 2, 2, 1)\n",
+      "[[[[1]\n",
+      "   [2]]\n",
+      "\n",
+      "  [[3]\n",
+      "   [4]]]]\n",
+      "\n",
+      "y shape :  (1, 4, 4, 1)\n",
+      "\n",
+      " After a (4,4) reshape :\n",
+      "[[1 1 2 2]\n",
+      " [1 1 2 2]\n",
+      " [3 3 4 4]\n",
+      " [3 3 4 4]]\n"
+     ]
+    }
+   ],
+   "source": [
+    "x = np.array([1,2,3,4])\n",
+    "x = x.reshape(2,2)\n",
+    "print('\\nInitial : ', x.shape)\n",
+    "print(x)\n",
+    "\n",
+    "x = x.reshape((1,2,2,1))\n",
+    "print('\\nReshape as a batch of (2,2) vectors : ', x.shape)\n",
+    "print(x)\n",
+    "\n",
+    "y = tf.keras.layers.UpSampling2D( size=(2, 2), interpolation=\"nearest\" )(x)\n",
+    "\n",
+    "y = np.array(y)\n",
+    "print('\\ny shape : ',y.shape)\n",
+    "\n",
+    "y = y.reshape(4,4)\n",
+    "print('\\n After a (4,4) reshape :')\n",
+    "print(y)\n",
+    "\n"
+   ]
+  },
  {
   "cell_type": "code",
   "execution_count": null,
-   "id": "4d94892b-d3a5-448d-aa2b-28c3a01a4b72",
+   "id": "09ac4e52-8953-41d9-b712-e6a83a9ae860",
   "metadata": {},
   "outputs": [],
   "source": []

 %% Cell type:markdown id:alpha-bahrain tags:

 <img width="800px" src="../fidle/img/00-Fidle-header-01.svg"></img>

 # <!-- TITLE --> [SCRATCH1] - Scratchbook
 <!-- DESC --> A scratchbook for small examples
 <!-- AUTHOR : Jean-Luc Parouty (CNRS/SIMaP) -->

 ## Objectives :
 - Take a quick look at thousands of little things

 ## Inside this scratchbook :

 [1 - LSTM Keras layer](#1---LSTM-Keras-layer)

 %% Cell type:markdown id:accessory-church tags:

 # One init to rule them all

 %% Cell type:code id:floppy-organic tags:

 ``` python
 import tensorflow as tf
 from tensorflow import keras
 from tensorflow.keras.callbacks import TensorBoard
 from tensorflow.keras.preprocessing.sequence import TimeseriesGenerator

 import numpy as np
 import math, random
 ```

 %% Cell type:markdown id:danish-rebound tags:

 ## 1 - LSTM Keras layer

 %% Cell type:code id:opposite-plasma tags:

 ``` python
 inputs  = tf.random.normal([32, 20, 8])

 lstm    = tf.keras.layers.LSTM(16)
 output  = lstm(inputs)

 print('Inputs shape is : ', inputs.shape)
 print('Output shape is : ', output.shape)
 ```

 %% Output

    Inputs shape is :  (32, 20, 8)
    Output shape is :  (32, 16)

 %% Cell type:code id:forbidden-murray tags:

 ``` python
 lstm = tf.keras.layers.LSTM(18, return_sequences=True, return_state=True)

 output, memory_state, carry_state = lstm(inputs)

 print('Output shape : ', output.shape)
 print('Memory state : ', memory_state.shape)
 print('Carry  state : ', carry_state.shape)
 ```

 %% Output

    Output shape :  (32, 20, 18)
    Memory state :  (32, 18)
    Carry  state :  (32, 18)

 %% Cell type:code id:verified-fruit tags:

 ``` python
 # --- See the last vector of the output
 output[-1,-1]
 ```

 %% Output

    <tf.Tensor: shape=(18,), dtype=float32, numpy=
    array([-0.20923303,  0.00193496,  0.05929745,  0.0429938 , -0.02835345,
            0.14096233,  0.07420755,  0.1777523 ,  0.1205566 , -0.03841979,
           -0.02402029,  0.16098973,  0.10468155, -0.06480312, -0.02497844,
            0.09700071, -0.24351674,  0.04884451], dtype=float32)>

 %% Cell type:code id:homeless-library tags:

 ``` python
 # ---- Memory state is the last output
 memory_state[-1]
 ```

 %% Output

    <tf.Tensor: shape=(18,), dtype=float32, numpy=
    array([-0.20923303,  0.00193496,  0.05929745,  0.0429938 , -0.02835345,
            0.14096233,  0.07420755,  0.1777523 ,  0.1205566 , -0.03841979,
           -0.02402029,  0.16098973,  0.10468155, -0.06480312, -0.02497844,
            0.09700071, -0.24351674,  0.04884451], dtype=float32)>

 %% Cell type:code id:preliminary-psychiatry tags:

 ``` python
 carry_state[-1]
 ```

 %% Output

    <tf.Tensor: shape=(18,), dtype=float32, numpy=
    array([-0.3245376 ,  0.00296011,  0.13041827,  0.10711877, -0.05223516,
            0.4009896 ,  0.21599025,  0.4260387 ,  0.30799934, -0.0799172 ,
           -0.06359857,  0.29457492,  0.18084048, -0.14462015, -0.04707906,
            0.15726675, -0.38622206,  0.09004797], dtype=float32)>

 %% Cell type:markdown id:41d326b2-376e-49d6-9429-07016d98dc09 tags:

 ## 2 - TimeseriesGenerator

 %% Cell type:code id:42276389-4ea6-42d1-93bc-6650062ef86a tags:

 ``` python
 from keras.preprocessing.sequence import TimeseriesGenerator

 # ---- Define a dataset

 series = np.array([1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15])

 # ---- Generator

 generator = TimeseriesGenerator(series, series, length=5, batch_size=1)

 # ---- Samples

 nb_batch = len(generator)

 print(f'Number of batch : {nb_batch}\n')
 for i in range(nb_batch):
    x, y = generator[i]
    print(f'#{i} : {x} => {y}')
 ```

 %% Output

    Number of batch : 10
    
    #0 : [[1 2 3 4 5]] => [6]
    #1 : [[2 3 4 5 6]] => [7]
    #2 : [[3 4 5 6 7]] => [8]
    #3 : [[4 5 6 7 8]] => [9]
    #4 : [[5 6 7 8 9]] => [10]
    #5 : [[ 6  7  8  9 10]] => [11]
    #6 : [[ 7  8  9 10 11]] => [12]
    #7 : [[ 8  9 10 11 12]] => [13]
    #8 : [[ 9 10 11 12 13]] => [14]
    #9 : [[10 11 12 13 14]] => [15]

-%% Cell type:code id:4d94892b-d3a5-448d-aa2b-28c3a01a4b72 tags:
+%% Cell type:markdown id:67e3c888-aaa4-4166-90a1-cdb63920fd7d tags:
+
+## 3 - Upsampling
+
+%% Cell type:code id:20f12cf0-1fdb-4b53-92c6-d03b140e42d1 tags:
+
+``` python
+x = np.array([1,2,3,4])
+x = x.reshape(2,2)
+print('\nInitial : ', x.shape)
+print(x)
+
+x = x.reshape((1,2,2,1))
+print('\nReshape as a batch of (2,2) vectors : ', x.shape)
+print(x)
+
+y = tf.keras.layers.UpSampling2D( size=(2, 2), interpolation="nearest" )(x)
+
+y = np.array(y)
+print('\ny shape : ',y.shape)
+
+y = y.reshape(4,4)
+print('\n After a (4,4) reshape :')
+print(y)
+
+```
+
+%% Output
+
+    
+    Initial :  (2, 2)
+    [[1 2]
+     [3 4]]
+    
+    Reshape as a batch of (2,2) vectors :  (1, 2, 2, 1)
+    [[[[1]
+       [2]]
+    
+      [[3]
+       [4]]]]
+    
+    y shape :  (1, 4, 4, 1)
+    
+     After a (4,4) reshape :
+    [[1 1 2 2]
+     [1 1 2 2]
+     [3 3 4 4]
+     [3 3 4 4]]
+
+%% Cell type:code id:09ac4e52-8953-41d9-b712-e6a83a9ae860 tags:

 ``` python
 ```