WEBVTT 00:00.120 --> 00:00.960 Welcome back. 00:01.400 --> 00:08.120 We reshaped the data, which is an essential because machine learning algorithms expect features to 00:08.120 --> 00:10.160 be in 2D format. 00:10.160 --> 00:15.240 So we transformed the 1D array into 2D array. 00:15.440 --> 00:19.720 Then we need to build the TensorFlow model. 00:19.760 --> 00:26.920 Here I want from you to focus with me because it's a very crucial and essential step. 00:26.960 --> 00:34.080 Let's start by model equals to dot dot sequential. 00:34.320 --> 00:38.840 And inside this function we need to pass an array. 00:39.040 --> 00:44.080 This array takes the tf.keras layers dot dense. 00:44.080 --> 00:49.520 And we passed one input shape and the name linear. 00:49.560 --> 00:52.320 Let me explain everything here. 00:52.360 --> 01:00.920 Here we are creating a simple model with exactly one on your own tf.keras dot sequential. 01:01.160 --> 01:10.460 This is a function that takes an array as a parameter creates a linear stack of layers that are executed 01:10.460 --> 01:15.900 in order from input to output, like stacking Lego blocks. 01:16.020 --> 01:23.540 And inside this array we have the layers dense, which is creating dense. 01:23.540 --> 01:26.220 Layer a fully connected layer. 01:26.220 --> 01:29.780 Every input connects to every output. 01:29.820 --> 01:35.980 Here we have the three parameters one input shape and the name. 01:36.220 --> 01:41.660 One number of neurons in the layer output dimension. 01:41.780 --> 01:47.620 For linear regression, one neuron equals to one output value. 01:47.660 --> 01:58.380 The predicted price, the input shape one and comma defines what kind of input the model expects. 01:58.620 --> 02:03.580 One expects one feature per sample, which is the house size. 02:03.580 --> 02:08.740 So here one is the one feature refers for one feature. 02:09.130 --> 02:17.490 This means each input is a single number like 1.5, 2.0, and so on. 02:17.530 --> 02:29.130 Okay, so we use this df TensorFlow Keras dot sequential function in order to create a linear stack 02:29.250 --> 02:35.570 of layers that are executed in order from input to output. 02:35.570 --> 02:44.730 And since our layers are very simple and our model is very simple, we have only one neuron in the layer, 02:44.850 --> 02:52.290 one neuron and the input shape we have only one input which is the size. 02:52.330 --> 02:54.890 This is the name name linear layer. 02:55.170 --> 02:58.290 It uh it's optional. 02:58.330 --> 02:59.810 You can specify it. 02:59.810 --> 03:09.290 But the main points here that you should pay attention to them are the number of neurons and the number.