1.18.0 User Documentation for Apache MADlib
Train Model Configurations
Contents

This module allows you to explore network architectures and hyperparameters by training many models a time across the database cluster.

The aim is to support efficient empirical comparison of multiple training configurations. This process is called model selection, and the implementation here is based on a parallel execution strategy called model hopper parallelism (MOP) [1,2].

Models are designed in Keras [3], which is a high-level neural network API written in Python. It can run on top of different backends and the one that is currently supported by MADlib is TensorFlow [4].

The main use case is classification using sequential models, which are made up of a linear stack of layers. This includes multilayer perceptrons (MLPs) and convolutional neural networks (CNNs). Regression is not currently supported.

Before using this model selection method, you will need to Prepocess Data which prepares data for use by models that support mini-batch as an optimization option. This is a one-time operation and you would only need to re-run the preprocessor if your input data has changed. The advantage of using mini-batching is that it can perform better than stochastic gradient descent because it uses more than one training example at a time, typically resulting faster and smoother convergence [5].

You can set up the models and hyperparameters to try with the Define Model Configurations function to define the unique combinations of model architectures, compile and fit parameters.

Note
1. If 'madlib_keras_fit_multiple_model()' is running on GPDB 5 and some versions of GPDB 6, the database will keep adding to the disk space (in proportion to model size) and will only release the disk space once the fit multiple query has completed execution. This is not the case for GPDB 6.5.0+ where disk space is released during the fit multiple query.
2. CUDA GPU memory cannot be released until the process holding it is terminated. When a MADlib deep learning function is called with GPUs, Greenplum internally creates a process (called a slice) which calls TensorFlow to do the computation. This process holds the GPU memory until one of the following two things happen: query finishes and user logs out of the Postgres client/session; or, query finishes and user waits for the timeout set by gp_vmem_idle_resource_timeout. The default value for this timeout is 18 sec [8]. So the recommendation is: log out/reconnect to the session after every GPU query; or wait for gp_vmem_idle_resource_timeout before you run another GPU query (you can also set it to a lower value).
3. This method does not currently support multi-input or multi-output neural networks.

Fit
The fit (training) function has the following format:
madlib_keras_fit_multiple_model(
source_table,
model_output_table,
model_selection_table,
num_iterations,
use_gpus,
validation_table,
metrics_compute_frequency,
warm_start,
name,
description,
use_caching
)


Arguments

source_table

TEXT. Name of the table containing the training data. This is the name of the output table from the data preprocessor. Independent and dependent variables are specified in the preprocessor step which is why you do not need to explictly state them here as part of the fit function.

model_output_table
TEXT. Name of the output table containing the multiple models created.
Note
pg_temp is not allowed as an output table schema for fit multiple. Details of output tables are shown below.
model_selection_table

TEXT. Name of the table containing model selection parameters to be tried. Here we mean both hyper-parameter tuning and model architecture search.

num_iterations

INTEGER. Number of iterations to train.

Note
This parameter is different than the number of passes over the dataset, which is commonly referred to as the number of epochs. Since MADlib operates in a distributed system, the number of epochs is actually equal to this parameter 'num_iterations' X 'epochs' as specified in the Keras fit parameter.
use_gpus (optional)

BOOLEAN, default: FALSE (i.e., CPU). Determines whether GPUs are to be used for training the neural network. Set to TRUE to use GPUs.

Note
This parameter must not conflict with how the distribution rules are set in the preprocessor function. For example, if you set a distribution rule to use certain segments on hosts that do not have GPUs attached, you will get an error if you set ‘use_gpus’ to TRUE. Also, we have seen some memory related issues when segments share GPU resources. For example, if you have 1 GPU per segment host and your cluster has 4 segments per segment host, it means that all 4 segments will share the same GPU on each host. The current recommended configuration is 1 GPU per segment.
validation_table (optional)

TEXT, default: none. Name of the table containing the validation dataset. Note that the validation dataset must be preprocessed in the same way as the training dataset, so this is the name of the output table from running the data preprocessor on the validation dataset. Using a validation dataset can mean a longer training time, depending on its size. This can be controlled using the 'metrics_compute_frequency' parameter described below.

metrics_compute_frequency (optional)

INTEGER, default: once at the end of training after 'num_iterations'. Frequency to compute per-iteration metrics for the training dataset and validation dataset (if specified). There can be considerable cost to computing metrics every iteration, especially if the training dataset is large. This parameter is a way of controlling the frequency of those computations. For example, if you specify 5, then metrics will be computed every 5 iterations as well as at the end of training after 'num_iterations'. If you use the default, metrics will be computed only once after 'num_iterations' have completed.

warm_start (optional)

BOOLEAN, default: FALSE. Initalize weights with the coefficients from the last call to the fit function. If set to TRUE, weights will be initialized from the model table generated by the previous training run.

Note
The warm start feature works based on the name of the model output table from a previous training run. When using warm start, do not drop the model output table or the model output summary table before calling the fit function, since these are needed to obtain the weights from the previous run. If you are not using warm start, the model output table and the model output table summary must be dropped in the usual way before calling the training function.
name (optional)

TEXT, default: NULL. Free text string to identify a name, if desired.

description (optional)

TEXT, default: NULL. Free text string to provide a description, if desired.

use_caching (optional)

BOOLEAN, default: FALSE. Use caching of data in memory on the segment in order to speed up processing.

Note
When set to TRUE, byte arrays on each segment are maintained in cache (SD). This can speed up training significantly, however the memory usage per segment increases. In effect, it requires enough available memory on a segment so that all training data residing on that segment can be read into memory.

Output tables
The model output table produced by fit contains the following columns. There is one row per model as per the rows in the 'model_selection_table':

mst_key INTEGER. ID that defines a unique tuple for model architecture-compile parameters-fit parameters, as defined in the 'model_selection_table'. BYTEA8. Byte array containing the weights of the neural net. TEXT. A JSON representation of the model architecture used in training.

An info table named <model_output_table>_info is also created, which has the following columns. There is one row per model as per the rows in the 'model_selection_table':

mst_key INTEGER. ID that defines a unique tuple for model architecture-compile parameters-fit parameters, as defined in the 'model_selection_table'. INTEGER. ID that defines model in the 'model_arch_table'. Compile parameters passed to Keras. Fit parameters passed to Keras. General identifier for type of model trained. Currently says 'madlib_keras'. Size of the model in KB. Models are stored in 'bytea' data format which is used for binary strings in PostgreSQL type databases. Array of elapsed time for metric computations as per the 'metrics_compute_frequency' parameter. Useful for drawing a curve showing loss, accuracy or other metrics as a function of time. For example, if 'metrics_compute_frequency=5' this would be an array of elapsed time for every 5th iteration, plus the last iteration. Note that this field reports the time for training + validation if there is a validation table provided. Metric specified in the 'compile_params'. Loss specified in the 'compile_params'. Final value of the training metric after all iterations have completed. The metric reported is the one specified in the 'metrics_type' parameter. Final value of the training loss after all iterations have completed. Array of training metrics as per the 'metrics_compute_frequency' parameter. For example, if 'metrics_compute_frequency=5' this would be an array of metrics for every 5th iteration, plus the last iteration. Array of training losses as per the 'metrics_compute_frequency' parameter. For example, if 'metrics_compute_frequency=5' this would be an array of losses for every 5th iteration, plus the last iteration. Final value of the validation metric after all iterations have completed. The metric reported is the one specified in the 'metrics_type' parameter. Final value of the validation loss after all iterations have completed. Array of validation metrics as per the 'metrics_compute_frequency' parameter. For example, if 'metrics_compute_frequency=5' this would be an array of metrics for every 5th iteration, plus the last iteration. Array of validation losses as per the 'metrics_compute_frequency' parameter. For example, if 'metrics_compute_frequency=5' this would be an array of losses for every 5th iteration, plus the last iteration.

A summary table named <model>_summary is also created, which has the following columns:

source_table Source table used for training. Name of the table containing the validation dataset (if specified). Name of the output table containing the model for each model selection tuple. Name of the output table containing the model performance and other info for each model selection tuple. Dependent variable column from the original source table in the data preprocessing step. Independent variables column from the original source table in the data preprocessing step. Name of the table containing the model architecture and (optionally) the initial model weights. Number of iterations of training completed. Frequency that per-iteration metrics are computed for the training dataset and validation datasets. Indicates whether warm start used or not. Name of the training run (free text). Description of the training run (free text). Timestamp for start of training. Timestamp for end of training. Version of MADlib used. Count of distinct classes values used. Array of actual class values used for a particular dependent variable. A column will be generated for each dependent variable. Data type of the dependent variable. Normalizing constant used from the data preprocessing step. Array indicating the iterations for which metrics are calculated, as derived from the parameters 'num_iterations' and 'metrics_compute_frequency'. For example, if 'num_iterations=5' and 'metrics_compute_frequency=2', then 'metrics_iters' value would be {2,4,5} indicating that metrics were computed at iterations 2, 4 and 5 (at the end). If 'num_iterations=5' and 'metrics_compute_frequency=1', then 'metrics_iters' value would be {1,2,3,4,5} indicating that metrics were computed at every iteration.

Evaluate
The evaluation function has the following format:
madlib_keras_evaluate(
model_table,
test_table,
output_table,
use_gpus,
mst_key
)


Arguments

model_table

TEXT. Name of the table containing the model to use for validation.

test_table

TEXT. Name of the table containing the evaluation dataset. Note that test/validation data must be preprocessed in the same way as the training dataset, so this is the name of the output table from the data preprocessor. Independent and dependent variables are specified in the preprocessor step which is why you do not need to explictly state them here as part of the fit function.

output_table
TEXT. Name of table that validation output will be written to. Table contains:
loss Loss value on evaluation dataset, where 'loss_type' below identifies the type of loss. Metric value on evaluation dataset, where 'metrics_type' below identifies the type of metric. Type of metric used that was used in the training step. (It means you cannot have a different metric in evaluate compared to training.) Type of loss function that was used in the training step. (It means you cannot have a different loss in evaluate compared to training.)
use_gpus (optional)

BOOLEAN, default: FALSE (i.e., CPU). Determines whether GPUs are to be used for training the neural network. Set to TRUE to use GPUs.

Note
This parameter must not conflict with how the distribution rules are set in the preprocessor function. For example, if you set a distribution rule to use certain segments on hosts that do not have GPUs attached, you will get an error if you set ‘use_gpus’ to TRUE. Also, we have seen some memory related issues when segments share GPU resources. For example, if you have 1 GPU per segment host and your cluster has 4 segments per segment host, it means that all 4 segments will share the same GPU on each host. The current recommended configuration is 1 GPU per segment.
mst_key (optional)
INTEGER, default: NULL. ID that defines a unique tuple for model architecture-compile parameters-fit parameters as defined in the model selection table.

Predict
The prediction function has the following format:
madlib_keras_predict(
model_table,
test_table,
id_col,
independent_varname,
output_table,
pred_type,
use_gpus,
mst_key
)


Arguments

model_table

TEXT. Name of the table containing the model to use for prediction.

test_table

TEXT. Name of the table containing the dataset to predict on. Note that test data is not preprocessed (unlike fit and evaluate) so put one test image per row for prediction. Also see the comment below for the 'independent_varname' parameter regarding normalization.

id_col

TEXT. Name of the id column in the test data table.

independent_varname

TEXT. Column with independent variables in the test table. If a 'normalizing_const' is specified when preprocessing the training dataset, this same normalization will be applied to the independent variables used in predict.

output_table
TEXT. Name of the table that prediction output will be written to. Table contains:
id Gives the 'id' for each prediction, corresponding to each row from the test_table. The estimated variable. The estimated class for classification. Probability of a given class. The rank of a given class based on the ordering of probabilities.
pred_type (optional)

TEXT or INTEGER or DOUBLE PRECISION default: 'prob'. The type and range of output desired. This parameter allows the following options.

• 'response': the actual prediction
• 'prob': the probability value for each class
• 0<value<1: the lower limit for the probability (double precision)
• 1<=value: the lower limit for the rank of the prediction (integer)

use_gpus (optional)

BOOLEAN, default: FALSE (i.e., CPU). Determines whether GPUs are to be used for prediction/inference. Set to TRUE to use GPUs.

Note
The prediction function uses the whole cluster. If you are using GPUs, it requires that GPUs are attached to all hosts, and that there are the same number of GPUs on each host (homogeneous cluster). This is different from the fit() and evaluate() functions that support GPUs on only some of the hosts (heterogeneous cluster). Therefore, if you have GPUs only on some of the hosts, or an uneven numbers of GPUs per host, then set this parameter to FALSE to use CPUs.
mst_key (optional)
INTEGER, default: NULL. ID that defines a unique tuple for model architecture-compile parameters-fit parameters as defined in the model selection table.

Examples
Note
Deep learning works best on very large datasets, but that is not convenient for a quick introduction to the syntax. So in this example we use an MLP on the well known iris data set from https://archive.ics.uci.edu/ml/datasets/iris. For more realistic examples with images please refer to the deep learning notebooks at https://github.com/apache/madlib-site/tree/asf-site/community-artifacts.

Classification

1. Create an input data set.
DROP TABLE IF EXISTS iris_data;
CREATE TABLE iris_data(
id serial,
attributes numeric[],
class_text varchar
);
INSERT INTO iris_data(id, attributes, class_text) VALUES
(1,ARRAY[5.1,3.5,1.4,0.2],'Iris-setosa'),
(2,ARRAY[4.9,3.0,1.4,0.2],'Iris-setosa'),
(3,ARRAY[4.7,3.2,1.3,0.2],'Iris-setosa'),
(4,ARRAY[4.6,3.1,1.5,0.2],'Iris-setosa'),
(5,ARRAY[5.0,3.6,1.4,0.2],'Iris-setosa'),
(6,ARRAY[5.4,3.9,1.7,0.4],'Iris-setosa'),
(7,ARRAY[4.6,3.4,1.4,0.3],'Iris-setosa'),
(8,ARRAY[5.0,3.4,1.5,0.2],'Iris-setosa'),
(9,ARRAY[4.4,2.9,1.4,0.2],'Iris-setosa'),
(10,ARRAY[4.9,3.1,1.5,0.1],'Iris-setosa'),
(11,ARRAY[5.4,3.7,1.5,0.2],'Iris-setosa'),
(12,ARRAY[4.8,3.4,1.6,0.2],'Iris-setosa'),
(13,ARRAY[4.8,3.0,1.4,0.1],'Iris-setosa'),
(14,ARRAY[4.3,3.0,1.1,0.1],'Iris-setosa'),
(15,ARRAY[5.8,4.0,1.2,0.2],'Iris-setosa'),
(16,ARRAY[5.7,4.4,1.5,0.4],'Iris-setosa'),
(17,ARRAY[5.4,3.9,1.3,0.4],'Iris-setosa'),
(18,ARRAY[5.1,3.5,1.4,0.3],'Iris-setosa'),
(19,ARRAY[5.7,3.8,1.7,0.3],'Iris-setosa'),
(20,ARRAY[5.1,3.8,1.5,0.3],'Iris-setosa'),
(21,ARRAY[5.4,3.4,1.7,0.2],'Iris-setosa'),
(22,ARRAY[5.1,3.7,1.5,0.4],'Iris-setosa'),
(23,ARRAY[4.6,3.6,1.0,0.2],'Iris-setosa'),
(24,ARRAY[5.1,3.3,1.7,0.5],'Iris-setosa'),
(25,ARRAY[4.8,3.4,1.9,0.2],'Iris-setosa'),
(26,ARRAY[5.0,3.0,1.6,0.2],'Iris-setosa'),
(27,ARRAY[5.0,3.4,1.6,0.4],'Iris-setosa'),
(28,ARRAY[5.2,3.5,1.5,0.2],'Iris-setosa'),
(29,ARRAY[5.2,3.4,1.4,0.2],'Iris-setosa'),
(30,ARRAY[4.7,3.2,1.6,0.2],'Iris-setosa'),
(31,ARRAY[4.8,3.1,1.6,0.2],'Iris-setosa'),
(32,ARRAY[5.4,3.4,1.5,0.4],'Iris-setosa'),
(33,ARRAY[5.2,4.1,1.5,0.1],'Iris-setosa'),
(34,ARRAY[5.5,4.2,1.4,0.2],'Iris-setosa'),
(35,ARRAY[4.9,3.1,1.5,0.1],'Iris-setosa'),
(36,ARRAY[5.0,3.2,1.2,0.2],'Iris-setosa'),
(37,ARRAY[5.5,3.5,1.3,0.2],'Iris-setosa'),
(38,ARRAY[4.9,3.1,1.5,0.1],'Iris-setosa'),
(39,ARRAY[4.4,3.0,1.3,0.2],'Iris-setosa'),
(40,ARRAY[5.1,3.4,1.5,0.2],'Iris-setosa'),
(41,ARRAY[5.0,3.5,1.3,0.3],'Iris-setosa'),
(42,ARRAY[4.5,2.3,1.3,0.3],'Iris-setosa'),
(43,ARRAY[4.4,3.2,1.3,0.2],'Iris-setosa'),
(44,ARRAY[5.0,3.5,1.6,0.6],'Iris-setosa'),
(45,ARRAY[5.1,3.8,1.9,0.4],'Iris-setosa'),
(46,ARRAY[4.8,3.0,1.4,0.3],'Iris-setosa'),
(47,ARRAY[5.1,3.8,1.6,0.2],'Iris-setosa'),
(48,ARRAY[4.6,3.2,1.4,0.2],'Iris-setosa'),
(49,ARRAY[5.3,3.7,1.5,0.2],'Iris-setosa'),
(50,ARRAY[5.0,3.3,1.4,0.2],'Iris-setosa'),
(51,ARRAY[7.0,3.2,4.7,1.4],'Iris-versicolor'),
(52,ARRAY[6.4,3.2,4.5,1.5],'Iris-versicolor'),
(53,ARRAY[6.9,3.1,4.9,1.5],'Iris-versicolor'),
(54,ARRAY[5.5,2.3,4.0,1.3],'Iris-versicolor'),
(55,ARRAY[6.5,2.8,4.6,1.5],'Iris-versicolor'),
(56,ARRAY[5.7,2.8,4.5,1.3],'Iris-versicolor'),
(57,ARRAY[6.3,3.3,4.7,1.6],'Iris-versicolor'),
(58,ARRAY[4.9,2.4,3.3,1.0],'Iris-versicolor'),
(59,ARRAY[6.6,2.9,4.6,1.3],'Iris-versicolor'),
(60,ARRAY[5.2,2.7,3.9,1.4],'Iris-versicolor'),
(61,ARRAY[5.0,2.0,3.5,1.0],'Iris-versicolor'),
(62,ARRAY[5.9,3.0,4.2,1.5],'Iris-versicolor'),
(63,ARRAY[6.0,2.2,4.0,1.0],'Iris-versicolor'),
(64,ARRAY[6.1,2.9,4.7,1.4],'Iris-versicolor'),
(65,ARRAY[5.6,2.9,3.6,1.3],'Iris-versicolor'),
(66,ARRAY[6.7,3.1,4.4,1.4],'Iris-versicolor'),
(67,ARRAY[5.6,3.0,4.5,1.5],'Iris-versicolor'),
(68,ARRAY[5.8,2.7,4.1,1.0],'Iris-versicolor'),
(69,ARRAY[6.2,2.2,4.5,1.5],'Iris-versicolor'),
(70,ARRAY[5.6,2.5,3.9,1.1],'Iris-versicolor'),
(71,ARRAY[5.9,3.2,4.8,1.8],'Iris-versicolor'),
(72,ARRAY[6.1,2.8,4.0,1.3],'Iris-versicolor'),
(73,ARRAY[6.3,2.5,4.9,1.5],'Iris-versicolor'),
(74,ARRAY[6.1,2.8,4.7,1.2],'Iris-versicolor'),
(75,ARRAY[6.4,2.9,4.3,1.3],'Iris-versicolor'),
(76,ARRAY[6.6,3.0,4.4,1.4],'Iris-versicolor'),
(77,ARRAY[6.8,2.8,4.8,1.4],'Iris-versicolor'),
(78,ARRAY[6.7,3.0,5.0,1.7],'Iris-versicolor'),
(79,ARRAY[6.0,2.9,4.5,1.5],'Iris-versicolor'),
(80,ARRAY[5.7,2.6,3.5,1.0],'Iris-versicolor'),
(81,ARRAY[5.5,2.4,3.8,1.1],'Iris-versicolor'),
(82,ARRAY[5.5,2.4,3.7,1.0],'Iris-versicolor'),
(83,ARRAY[5.8,2.7,3.9,1.2],'Iris-versicolor'),
(84,ARRAY[6.0,2.7,5.1,1.6],'Iris-versicolor'),
(85,ARRAY[5.4,3.0,4.5,1.5],'Iris-versicolor'),
(86,ARRAY[6.0,3.4,4.5,1.6],'Iris-versicolor'),
(87,ARRAY[6.7,3.1,4.7,1.5],'Iris-versicolor'),
(88,ARRAY[6.3,2.3,4.4,1.3],'Iris-versicolor'),
(89,ARRAY[5.6,3.0,4.1,1.3],'Iris-versicolor'),
(90,ARRAY[5.5,2.5,4.0,1.3],'Iris-versicolor'),
(91,ARRAY[5.5,2.6,4.4,1.2],'Iris-versicolor'),
(92,ARRAY[6.1,3.0,4.6,1.4],'Iris-versicolor'),
(93,ARRAY[5.8,2.6,4.0,1.2],'Iris-versicolor'),
(94,ARRAY[5.0,2.3,3.3,1.0],'Iris-versicolor'),
(95,ARRAY[5.6,2.7,4.2,1.3],'Iris-versicolor'),
(96,ARRAY[5.7,3.0,4.2,1.2],'Iris-versicolor'),
(97,ARRAY[5.7,2.9,4.2,1.3],'Iris-versicolor'),
(98,ARRAY[6.2,2.9,4.3,1.3],'Iris-versicolor'),
(99,ARRAY[5.1,2.5,3.0,1.1],'Iris-versicolor'),
(100,ARRAY[5.7,2.8,4.1,1.3],'Iris-versicolor'),
(101,ARRAY[6.3,3.3,6.0,2.5],'Iris-virginica'),
(102,ARRAY[5.8,2.7,5.1,1.9],'Iris-virginica'),
(103,ARRAY[7.1,3.0,5.9,2.1],'Iris-virginica'),
(104,ARRAY[6.3,2.9,5.6,1.8],'Iris-virginica'),
(105,ARRAY[6.5,3.0,5.8,2.2],'Iris-virginica'),
(106,ARRAY[7.6,3.0,6.6,2.1],'Iris-virginica'),
(107,ARRAY[4.9,2.5,4.5,1.7],'Iris-virginica'),
(108,ARRAY[7.3,2.9,6.3,1.8],'Iris-virginica'),
(109,ARRAY[6.7,2.5,5.8,1.8],'Iris-virginica'),
(110,ARRAY[7.2,3.6,6.1,2.5],'Iris-virginica'),
(111,ARRAY[6.5,3.2,5.1,2.0],'Iris-virginica'),
(112,ARRAY[6.4,2.7,5.3,1.9],'Iris-virginica'),
(113,ARRAY[6.8,3.0,5.5,2.1],'Iris-virginica'),
(114,ARRAY[5.7,2.5,5.0,2.0],'Iris-virginica'),
(115,ARRAY[5.8,2.8,5.1,2.4],'Iris-virginica'),
(116,ARRAY[6.4,3.2,5.3,2.3],'Iris-virginica'),
(117,ARRAY[6.5,3.0,5.5,1.8],'Iris-virginica'),
(118,ARRAY[7.7,3.8,6.7,2.2],'Iris-virginica'),
(119,ARRAY[7.7,2.6,6.9,2.3],'Iris-virginica'),
(120,ARRAY[6.0,2.2,5.0,1.5],'Iris-virginica'),
(121,ARRAY[6.9,3.2,5.7,2.3],'Iris-virginica'),
(122,ARRAY[5.6,2.8,4.9,2.0],'Iris-virginica'),
(123,ARRAY[7.7,2.8,6.7,2.0],'Iris-virginica'),
(124,ARRAY[6.3,2.7,4.9,1.8],'Iris-virginica'),
(125,ARRAY[6.7,3.3,5.7,2.1],'Iris-virginica'),
(126,ARRAY[7.2,3.2,6.0,1.8],'Iris-virginica'),
(127,ARRAY[6.2,2.8,4.8,1.8],'Iris-virginica'),
(128,ARRAY[6.1,3.0,4.9,1.8],'Iris-virginica'),
(129,ARRAY[6.4,2.8,5.6,2.1],'Iris-virginica'),
(130,ARRAY[7.2,3.0,5.8,1.6],'Iris-virginica'),
(131,ARRAY[7.4,2.8,6.1,1.9],'Iris-virginica'),
(132,ARRAY[7.9,3.8,6.4,2.0],'Iris-virginica'),
(133,ARRAY[6.4,2.8,5.6,2.2],'Iris-virginica'),
(134,ARRAY[6.3,2.8,5.1,1.5],'Iris-virginica'),
(135,ARRAY[6.1,2.6,5.6,1.4],'Iris-virginica'),
(136,ARRAY[7.7,3.0,6.1,2.3],'Iris-virginica'),
(137,ARRAY[6.3,3.4,5.6,2.4],'Iris-virginica'),
(138,ARRAY[6.4,3.1,5.5,1.8],'Iris-virginica'),
(139,ARRAY[6.0,3.0,4.8,1.8],'Iris-virginica'),
(140,ARRAY[6.9,3.1,5.4,2.1],'Iris-virginica'),
(141,ARRAY[6.7,3.1,5.6,2.4],'Iris-virginica'),
(142,ARRAY[6.9,3.1,5.1,2.3],'Iris-virginica'),
(143,ARRAY[5.8,2.7,5.1,1.9],'Iris-virginica'),
(144,ARRAY[6.8,3.2,5.9,2.3],'Iris-virginica'),
(145,ARRAY[6.7,3.3,5.7,2.5],'Iris-virginica'),
(146,ARRAY[6.7,3.0,5.2,2.3],'Iris-virginica'),
(147,ARRAY[6.3,2.5,5.0,1.9],'Iris-virginica'),
(148,ARRAY[6.5,3.0,5.2,2.0],'Iris-virginica'),
(149,ARRAY[6.2,3.4,5.4,2.3],'Iris-virginica'),
(150,ARRAY[5.9,3.0,5.1,1.8],'Iris-virginica');

Create a test/validation dataset from the training data:
DROP TABLE IF EXISTS iris_train, iris_test;
-- Set seed so results are reproducible
SELECT setseed(0);
'iris',          -- Output table root name
0.8,            -- Train proportion
NULL,           -- Test proportion (0.2)
NULL,           -- Strata definition
NULL,           -- Output all columns
NULL,           -- Sample without replacement
TRUE            -- Separate output tables
);
SELECT COUNT(*) FROM iris_train;

 count
------+
120

2. Call the preprocessor for deep learning. For the training dataset:
\x on
DROP TABLE IF EXISTS iris_train_packed, iris_train_packed_summary;
'iris_train_packed',  -- Output table
'class_text',         -- Dependent variable
'attributes'          -- Independent variable
);
SELECT * FROM iris_train_packed_summary;

-[ RECORD 1 ]-------+---------------------------------------------
source_table        | iris_train
output_table        | iris_train_packed
dependent_varname   | class_text
independent_varname | attributes
dependent_vartype   | character varying
class_values        | {Iris-setosa,Iris-versicolor,Iris-virginica}
buffer_size         | 60
normalizing_const   | 1.0
num_classes         | 3

For the validation dataset:
DROP TABLE IF EXISTS iris_test_packed, iris_test_packed_summary;
'iris_test_packed',   -- Output table
'class_text',         -- Dependent variable
'attributes',         -- Independent variable
'iris_train_packed'   -- From training preprocessor step
);
SELECT * FROM iris_test_packed_summary;

-[ RECORD 1 ]-------+---------------------------------------------
source_table        | iris_test
output_table        | iris_test_packed
dependent_varname   | class_text
independent_varname | attributes
dependent_vartype   | character varying
class_values        | {Iris-setosa,Iris-versicolor,Iris-virginica}
buffer_size         | 15
normalizing_const   | 1.0
num_classes         | 3

3. Define and load model architecture. Use Keras to define the model architecture with 1 hidden layer:
from tensorflow import keras
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense
model1 = Sequential()
model1.summary()
_________________________________________________________________
Layer (type)                 Output Shape              Param #
=================================================================
dense_1 (Dense)              (None, 10)                50
_________________________________________________________________
dense_2 (Dense)              (None, 10)                110
_________________________________________________________________
dense_3 (Dense)              (None, 3)                 33
=================================================================
Total params: 193
Trainable params: 193
Non-trainable params: 0


Export the model to JSON:
model1.to_json()

'{"class_name": "Sequential", "keras_version": "2.1.6", "config": [{"class_name": "Dense", "config": {"kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "uniform", "scale": 1.0, "seed": null, "mode": "fan_avg"}}, "name": "dense_1", "kernel_constraint": null, "bias_regularizer": null, "bias_constraint": null, "dtype": "float32", "activation": "relu", "trainable": true, "kernel_regularizer": null, "bias_initializer": {"class_name": "Zeros", "config": {}}, "units": 10, "batch_input_shape": [null, 4], "use_bias": true, "activity_regularizer": null}}, {"class_name": "Dense", "config": {"kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "uniform", "scale": 1.0, "seed": null, "mode": "fan_avg"}}, "name": "dense_2", "kernel_constraint": null, "bias_regularizer": null, "bias_constraint": null, "activation": "relu", "trainable": true, "kernel_regularizer": null, "bias_initializer": {"class_name": "Zeros", "config": {}}, "units": 10, "use_bias": true, "activity_regularizer": null}}, {"class_name": "Dense", "config": {"kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "uniform", "scale": 1.0, "seed": null, "mode": "fan_avg"}}, "name": "dense_3", "kernel_constraint": null, "bias_regularizer": null, "bias_constraint": null, "activation": "softmax", "trainable": true, "kernel_regularizer": null, "bias_initializer": {"class_name": "Zeros", "config": {}}, "units": 3, "use_bias": true, "activity_regularizer": null}}], "backend": "tensorflow"}'

Define model architecture with 2 hidden layers:
model2 = Sequential()
model2.summary()
Layer (type)                 Output Shape              Param #
=================================================================
dense_4 (Dense)              (None, 10)                50
_________________________________________________________________
dense_5 (Dense)              (None, 10)                110
_________________________________________________________________
dense_6 (Dense)              (None, 10)                110
_________________________________________________________________
dense_7 (Dense)              (None, 3)                 33
=================================================================
Total params: 303
Trainable params: 303
Non-trainable params: 0


Export the model to JSON:
model2.to_json()

'{"class_name": "Sequential", "keras_version": "2.1.6", "config": [{"class_name": "Dense", "config": {"kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "uniform", "scale": 1.0, "seed": null, "mode": "fan_avg"}}, "name": "dense_4", "kernel_constraint": null, "bias_regularizer": null, "bias_constraint": null, "dtype": "float32", "activation": "relu", "trainable": true, "kernel_regularizer": null, "bias_initializer": {"class_name": "Zeros", "config": {}}, "units": 10, "batch_input_shape": [null, 4], "use_bias": true, "activity_regularizer": null}}, {"class_name": "Dense", "config": {"kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "uniform", "scale": 1.0, "seed": null, "mode": "fan_avg"}}, "name": "dense_5", "kernel_constraint": null, "bias_regularizer": null, "bias_constraint": null, "activation": "relu", "trainable": true, "kernel_regularizer": null, "bias_initializer": {"class_name": "Zeros", "config": {}}, "units": 10, "use_bias": true, "activity_regularizer": null}}, {"class_name": "Dense", "config": {"kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "uniform", "scale": 1.0, "seed": null, "mode": "fan_avg"}}, "name": "dense_6", "kernel_constraint": null, "bias_regularizer": null, "bias_constraint": null, "activation": "relu", "trainable": true, "kernel_regularizer": null, "bias_initializer": {"class_name": "Zeros", "config": {}}, "units": 10, "use_bias": true, "activity_regularizer": null}}, {"class_name": "Dense", "config": {"kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "uniform", "scale": 1.0, "seed": null, "mode": "fan_avg"}}, "name": "dense_7", "kernel_constraint": null, "bias_regularizer": null, "bias_constraint": null, "activation": "softmax", "trainable": true, "kernel_regularizer": null, "bias_initializer": {"class_name": "Zeros", "config": {}}, "units": 3, "use_bias": true, "activity_regularizer": null}}], "backend": "tensorflow"}'

DROP TABLE IF EXISTS model_arch_library;
$${"class_name": "Sequential", "keras_version": "2.1.6", "config": [{"class_name": "Dense", "config": {"kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "uniform", "scale": 1.0, "seed": null, "mode": "fan_avg"}}, "name": "dense_1", "kernel_constraint": null, "bias_regularizer": null, "bias_constraint": null, "dtype": "float32", "activation": "relu", "trainable": true, "kernel_regularizer": null, "bias_initializer": {"class_name": "Zeros", "config": {}}, "units": 10, "batch_input_shape": [null, 4], "use_bias": true, "activity_regularizer": null}}, {"class_name": "Dense", "config": {"kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "uniform", "scale": 1.0, "seed": null, "mode": "fan_avg"}}, "name": "dense_2", "kernel_constraint": null, "bias_regularizer": null, "bias_constraint": null, "activation": "relu", "trainable": true, "kernel_regularizer": null, "bias_initializer": {"class_name": "Zeros", "config": {}}, "units": 10, "use_bias": true, "activity_regularizer": null}}, {"class_name": "Dense", "config": {"kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "uniform", "scale": 1.0, "seed": null, "mode": "fan_avg"}}, "name": "dense_3", "kernel_constraint": null, "bias_regularizer": null, "bias_constraint": null, "activation": "softmax", "trainable": true, "kernel_regularizer": null, "bias_initializer": {"class_name": "Zeros", "config": {}}, "units": 3, "use_bias": true, "activity_regularizer": null}}], "backend": "tensorflow"}$$
::json,         -- JSON blob
NULL,                  -- Weights
'Sophie',              -- Name
'MLP with 1 hidden layer'       -- Descr
);
$${"class_name": "Sequential", "keras_version": "2.1.6", "config": [{"class_name": "Dense", "config": {"kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "uniform", "scale": 1.0, "seed": null, "mode": "fan_avg"}}, "name": "dense_4", "kernel_constraint": null, "bias_regularizer": null, "bias_constraint": null, "dtype": "float32", "activation": "relu", "trainable": true, "kernel_regularizer": null, "bias_initializer": {"class_name": "Zeros", "config": {}}, "units": 10, "batch_input_shape": [null, 4], "use_bias": true, "activity_regularizer": null}}, {"class_name": "Dense", "config": {"kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "uniform", "scale": 1.0, "seed": null, "mode": "fan_avg"}}, "name": "dense_5", "kernel_constraint": null, "bias_regularizer": null, "bias_constraint": null, "activation": "relu", "trainable": true, "kernel_regularizer": null, "bias_initializer": {"class_name": "Zeros", "config": {}}, "units": 10, "use_bias": true, "activity_regularizer": null}}, {"class_name": "Dense", "config": {"kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "uniform", "scale": 1.0, "seed": null, "mode": "fan_avg"}}, "name": "dense_6", "kernel_constraint": null, "bias_regularizer": null, "bias_constraint": null, "activation": "relu", "trainable": true, "kernel_regularizer": null, "bias_initializer": {"class_name": "Zeros", "config": {}}, "units": 10, "use_bias": true, "activity_regularizer": null}}, {"class_name": "Dense", "config": {"kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "uniform", "scale": 1.0, "seed": null, "mode": "fan_avg"}}, "name": "dense_7", "kernel_constraint": null, "bias_regularizer": null, "bias_constraint": null, "activation": "softmax", "trainable": true, "kernel_regularizer": null, "bias_initializer": {"class_name": "Zeros", "config": {}}, "units": 3, "use_bias": true, "activity_regularizer": null}}], "backend": "tensorflow"}$$
::json,         -- JSON blob
NULL,                  -- Weights
'Maria',               -- Name
'MLP with 2 hidden layers'       -- Descr
);

4. Generate model configurations using grid search. The output table for grid search contains the unique combinations of model architectures, compile and fit parameters.
DROP TABLE IF EXISTS mst_table, mst_table_summary;
'model_arch_library', -- model architecture table
'mst_table',          -- model selection table output
ARRAY[1,2],          -- model ids from model architecture table
$${'loss': ['categorical_crossentropy'], 'optimizer_params_list': [ {'optimizer': ['Adam'], 'lr': [0.001, 0.01, 0.1]} ], 'metrics': ['accuracy']}$$,                  -- compile_param_grid
$${ 'batch_size': [4, 8], 'epochs': [1] }$$,                  -- fit_param_grid
'grid'               -- search_type
);
SELECT * FROM mst_table ORDER BY mst_key;

 mst_key | model_id |                                 compile_params                                  |      fit_params
---------+----------+---------------------------------------------------------------------------------+-----------------------
1 |        1 | optimizer='Adam(lr=0.001)',metrics=['accuracy'],loss='categorical_crossentropy' | epochs=1,batch_size=4
2 |        1 | optimizer='Adam(lr=0.001)',metrics=['accuracy'],loss='categorical_crossentropy' | epochs=1,batch_size=8
3 |        1 | optimizer='Adam(lr=0.01)',metrics=['accuracy'],loss='categorical_crossentropy'  | epochs=1,batch_size=4
4 |        1 | optimizer='Adam(lr=0.01)',metrics=['accuracy'],loss='categorical_crossentropy'  | epochs=1,batch_size=8
5 |        1 | optimizer='Adam(lr=0.1)',metrics=['accuracy'],loss='categorical_crossentropy'   | epochs=1,batch_size=4
6 |        1 | optimizer='Adam(lr=0.1)',metrics=['accuracy'],loss='categorical_crossentropy'   | epochs=1,batch_size=8
7 |        2 | optimizer='Adam(lr=0.001)',metrics=['accuracy'],loss='categorical_crossentropy' | epochs=1,batch_size=4
8 |        2 | optimizer='Adam(lr=0.001)',metrics=['accuracy'],loss='categorical_crossentropy' | epochs=1,batch_size=8
9 |        2 | optimizer='Adam(lr=0.01)',metrics=['accuracy'],loss='categorical_crossentropy'  | epochs=1,batch_size=4
10 |        2 | optimizer='Adam(lr=0.01)',metrics=['accuracy'],loss='categorical_crossentropy'  | epochs=1,batch_size=8
11 |        2 | optimizer='Adam(lr=0.1)',metrics=['accuracy'],loss='categorical_crossentropy'   | epochs=1,batch_size=4
12 |        2 | optimizer='Adam(lr=0.1)',metrics=['accuracy'],loss='categorical_crossentropy'   | epochs=1,batch_size=8
(12 rows)

This is the name of the model architecture table that corresponds to the model selection table:
SELECT * FROM mst_table_summary;

  model_arch_table  | object_table
--------------------+--------------
model_arch_library |

5. Train multiple models.
DROP TABLE IF EXISTS iris_multi_model, iris_multi_model_summary, iris_multi_model_info;
'iris_multi_model',     -- model_output_table
'mst_table',            -- model_selection_table
10,                     -- num_iterations
FALSE                   -- use gpus
);

View the model summary:
SELECT * FROM iris_multi_model_summary;

source_table              | iris_train_packed
validation_table          |
model                     | iris_multi_model
model_info                | iris_multi_model_info
dependent_varname         | {class_text}
independent_varname       | {attributes}
model_arch_table          | model_arch_library
model_selection_table     | mst_table
object_table              |
num_iterations            | 10
metrics_compute_frequency | 10
warm_start                | f
name                      |
description               |
start_training_time       | 2021-02-05 00:40:42.695613
end_training_time         | 2021-02-05 00:42:20.796712
num_classes               | {1}
class_text_class_values   | {Iris-setosa,Iris-versicolor,Iris-virginica}
dependent_vartype         | {"character varying"}
normalizing_const         | 1
metrics_iters             | {10}

View results for each model:
SELECT * FROM iris_multi_model_info ORDER BY training_metrics_final DESC, training_loss_final;

  mst_key | model_id |                                 compile_params                                  |      fit_params       |  model_type  | model_size | metrics_elapsed_time | metrics_type |        loss_type         | training_metrics_final | training_loss_final |  training_metrics   |    training_loss     | validation_metrics_final | validation_loss_final | validation_metrics | validation_loss
---------+----------+---------------------------------------------------------------------------------+-----------------------+--------------+------------+----------------------+--------------+--------------------------+------------------------+---------------------+---------------------+----------------------+--------------------------+-----------------------+--------------------+-----------------
2 |        1 | loss='categorical_crossentropy',optimizer='Adam(lr=0.1)',metrics=['accuracy']   | batch_size=8,epochs=1 | madlib_keras | 0.75390625 | {96.2744579315186}   | {accuracy}   | categorical_crossentropy |      0.966666638851166 |  0.0771341994404793 | {0.966666638851166} | {0.0771341994404793} |                          |                       |                    |
3 |        1 | loss='categorical_crossentropy', optimizer='Adam(lr=0.01)',metrics=['accuracy'] | batch_size=4,epochs=1 | madlib_keras | 0.75390625 | {95.0798950195312}   | {accuracy}   | categorical_crossentropy |      0.958333313465118 |    0.14112713932991 | {0.958333313465118} | {0.14112713932991}   |                          |                       |                    |
10 |        2 | loss='categorical_crossentropy', optimizer='Adam(lr=0.01)',metrics=['accuracy'] | batch_size=8,epochs=1 | madlib_keras | 1.18359375 | {95.7743279933929}   | {accuracy}   | categorical_crossentropy |      0.949999988079071 |   0.126085489988327 | {0.949999988079071} | {0.126085489988327}  |                          |                       |                    |
5 |        1 | loss='categorical_crossentropy',optimizer='Adam(lr=0.001)',metrics=['accuracy'] | batch_size=4,epochs=1 | madlib_keras | 0.75390625 | {97.8191220760345}   | {accuracy}   | categorical_crossentropy |      0.866666674613953 |   0.459462374448776 | {0.866666674613953} | {0.459462374448776}  |                          |                       |                    |
4 |        1 | loss='categorical_crossentropy', optimizer='Adam(lr=0.01)',metrics=['accuracy'] | batch_size=8,epochs=1 | madlib_keras | 0.75390625 | {96.7445840835571}   | {accuracy}   | categorical_crossentropy |      0.858333349227905 |   0.279698997735977 | {0.858333349227905} | {0.279698997735977}  |                          |                       |                    |
9 |        2 | loss='categorical_crossentropy', optimizer='Adam(lr=0.01)',metrics=['accuracy'] | batch_size=4,epochs=1 | madlib_keras | 1.18359375 | {95.3640351295471}   | {accuracy}   | categorical_crossentropy |      0.824999988079071 |   0.325970768928528 | {0.824999988079071} | {0.325970768928528}  |                          |                       |                    |
7 |        2 | loss='categorical_crossentropy',optimizer='Adam(lr=0.1)',metrics=['accuracy']   | batch_size=4,epochs=1 | madlib_keras | 1.18359375 | {94.8350050449371}   | {accuracy}   | categorical_crossentropy |      0.800000011920929 |   0.458843886852264 | {0.800000011920929} | {0.458843886852264}  |                          |                       |                    |
12 |        2 | loss='categorical_crossentropy',optimizer='Adam(lr=0.001)',metrics=['accuracy'] | batch_size=8,epochs=1 | madlib_keras | 1.18359375 | {96.0411529541016}   | {accuracy}   | categorical_crossentropy |      0.783333361148834 |   0.766786217689514 | {0.783333361148834} | {0.766786217689514}  |                          |                       |                    |
11 |        2 | loss='categorical_crossentropy',optimizer='Adam(lr=0.001)',metrics=['accuracy'] | batch_size=4,epochs=1 | madlib_keras | 1.18359375 | {98.098680973053}    | {accuracy}   | categorical_crossentropy |      0.683333337306976 |   0.607033967971802 | {0.683333337306976} | {0.607033967971802}  |                          |                       |                    |
6 |        1 | loss='categorical_crossentropy',optimizer='Adam(lr=0.001)',metrics=['accuracy'] | batch_size=8,epochs=1 | madlib_keras | 0.75390625 | {96.5071561336517}   | {accuracy}   | categorical_crossentropy |      0.683333337306976 |   0.704851150512695 | {0.683333337306976} | {0.704851150512695}  |                          |                       |                    |
8 |        2 | loss='categorical_crossentropy',optimizer='Adam(lr=0.1)',metrics=['accuracy']   | batch_size=8,epochs=1 | madlib_keras | 1.18359375 | {97.4749901294708}   | {accuracy}   | categorical_crossentropy |      0.641666650772095 |   0.473412454128265 | {0.641666650772095} | {0.473412454128265}  |                          |                       |                    |
1 |        1 | loss='categorical_crossentropy',optimizer='Adam(lr=0.1)',metrics=['accuracy']   | batch_size=4,epochs=1 | madlib_keras | 0.75390625 | {96.9749960899353}   | {accuracy}   | categorical_crossentropy |      0.358333319425583 |    1.09744954109192 | {0.358333319425583} | {1.09744954109192}   |                          |                       |                    |
(12 rows)

6. Evaluate. Now run evaluate using models we built above:
DROP TABLE IF EXISTS iris_validate;
'iris_test_packed',  -- test table
'iris_validate',     -- output table
NULL,               -- use gpus
3                   -- mst_key to use
);
SELECT * FROM iris_validate;

        loss        | metric | metrics_type |        loss_type
--------------------+--------+--------------+--------------------------
0.0789926201105118 |      1 | {accuracy}   | categorical_crossentropy

7. Predict. Now predict using one of the models we built. We will use the validation data set for prediction as well, which is not usual but serves to show the syntax. The prediction is in the estimated_class_text column:
DROP TABLE IF EXISTS iris_predict;
'iris_test',        -- test_table
'id',               -- id column
'attributes',       -- independent var
'iris_predict',     -- output table
'response',        -- prediction type
FALSE,             -- use gpus
3                  -- mst_key to use
);
SELECT * FROM iris_predict ORDER BY id;

 id  | class_name |   class_value   |    prob
-----+------------+-----------------+------------
8 | class_text | Iris-setosa     |  0.9979984
15 | class_text | Iris-setosa     | 0.99929357
20 | class_text | Iris-setosa     |  0.9985701
21 | class_text | Iris-setosa     |  0.9984106
22 | class_text | Iris-setosa     |  0.9983991
27 | class_text | Iris-setosa     | 0.99763095
28 | class_text | Iris-setosa     |   0.998376
33 | class_text | Iris-setosa     | 0.99901116
37 | class_text | Iris-setosa     | 0.99871385
38 | class_text | Iris-setosa     | 0.99740535
46 | class_text | Iris-setosa     |  0.9968368
51 | class_text | Iris-versicolor | 0.93798196
57 | class_text | Iris-versicolor | 0.73391247
58 | class_text | Iris-versicolor |  0.9449931
59 | class_text | Iris-versicolor |  0.8938894
64 | class_text | Iris-versicolor |  0.6563309
65 | class_text | Iris-versicolor | 0.95828146
72 | class_text | Iris-versicolor |    0.94206
75 | class_text | Iris-versicolor | 0.93305194
87 | class_text | Iris-versicolor |  0.8458596
95 | class_text | Iris-versicolor | 0.76850986
97 | class_text | Iris-versicolor |  0.8467575
98 | class_text | Iris-versicolor |  0.9081358
120 | class_text | Iris-virginica  | 0.86913925
121 | class_text | Iris-virginica  | 0.96328884
122 | class_text | Iris-virginica  |  0.9474733
123 | class_text | Iris-virginica  | 0.97997576
140 | class_text | Iris-virginica  |  0.8721462
144 | class_text | Iris-virginica  |  0.9745266
147 | class_text | Iris-virginica  |  0.8978669
(30 rows)

Count missclassifications:
SELECT COUNT(*) FROM iris_predict JOIN iris_test USING (id)
WHERE iris_predict.class_value != iris_test.class_text;

 count
-------+
0

Accuracy:
SELECT round(count(*)*100/(150*0.2),2) as test_accuracy_percent from
(select iris_test.class_text as actual, iris_predict.class_value as estimated
from iris_predict inner join iris_test
on iris_test.id=iris_predict.id) q
WHERE q.actual=q.estimated;

 test_accuracy_percent
-----------------------+
100.00


Classification with Other Parameters

1. Validation dataset. Now use a validation dataset and compute metrics every 3rd iteration using the 'metrics_compute_frequency' parameter. This can help reduce run time if you do not need metrics computed at every iteration. Also turn on caching.
DROP TABLE IF EXISTS iris_multi_model, iris_multi_model_summary, iris_multi_model_info;
'iris_multi_model',     -- model_output_table
'mst_table',            -- model_selection_table
10,                     -- num_iterations
FALSE,                 -- use gpus
'iris_test_packed',     -- validation dataset
3,                     -- metrics compute frequency
FALSE,                 -- warm start
'Sophie L.',            -- name
'Model selection for iris dataset',  -- description
TRUE                   -- use caching
);

View the model summary:
SELECT * FROM iris_multi_model_summary;

-[ RECORD 1 ]-------------+---------------------------------------------
source_table              | iris_train_packed
validation_table          | iris_test_packed
model                     | iris_multi_model
model_info                | iris_multi_model_info
dependent_varname         | {class_text}
independent_varname       | {attributes}
model_arch_table          | model_arch_library
model_selection_table     | mst_table
object_table              |
num_iterations            | 10
metrics_compute_frequency | 3
warm_start                | f
name                      | Sophie L.
description               | Model selection for iris dataset
start_training_time       | 2021-02-05 01:03:11.337798
end_training_time         | 2021-02-05 01:05:14.988912
num_classes               | {1}
class_text_class_values   | {Iris-setosa,Iris-versicolor,Iris-virginica}
dependent_vartype         | {"character varying"}
normalizing_const         | 1
metrics_iters             | {3,6,9,10}

View results for each model:
SELECT * FROM iris_multi_model_info ORDER BY training_metrics_final DESC, training_loss_final;

 ---------+----------+---------------------------------------------------------------------------------+-----------------------+--------------+------------+-----------------------------------------------------------------------+--------------+--------------------------+------------------------+---------------------+---------------------------------------------------------------------------+---------------------------------------------------------------------------+--------------------------+-----------------------+---------------------------------------------------------------------------+------------------------------------------------------------------------------
10 |        2 | loss='categorical_crossentropy', optimizer='Adam(lr=0.01)',metrics=['accuracy'] | batch_size=8,epochs=1 | madlib_keras | 1.18359375 | {34.347088098526,69.4049510955811,105.285098075867,121.389348983765}  | {accuracy}   | categorical_crossentropy |      0.975000023841858 |   0.166684880852699 | {0.949999988079071,0.949999988079071,0.908333361148834,0.975000023841858} | {0.615781545639038,0.269571483135223,0.242876216769218,0.166684880852699} |                        1 |     0.107727721333504 | {1,1,0.966666638851166,1}                                                 | {0.620104968547821,0.227639734745026,0.155238434672356,0.107727721333504}
3 |        1 | loss='categorical_crossentropy', optimizer='Adam(lr=0.01)',metrics=['accuracy'] | batch_size=4,epochs=1 | madlib_keras | 0.75390625 | {33.7874810695648,68.735356092453,104.679323911667,120.741965055466}  | {accuracy}   | categorical_crossentropy |      0.958333313465118 |   0.151768147945404 | {0.958333313465118,0.816666662693024,0.941666662693024,0.958333313465118} | {0.232924744486809,0.356248170137405,0.160927340388298,0.151768147945404} |                        1 |    0.0933234840631485 | {1,0.866666674613953,0.966666638851166,1}                                 | {0.197644680738449,0.244517579674721,0.0969053283333778,0.0933234840631485}
4 |        1 | loss='categorical_crossentropy', optimizer='Adam(lr=0.01)',metrics=['accuracy'] | batch_size=8,epochs=1 | madlib_keras | 0.75390625 | {35.5827050209045,70.4194140434265,106.230206012726,122.374103069305} | {accuracy}   | categorical_crossentropy |      0.941666662693024 |   0.146068558096886 | {0.824999988079071,0.958333313465118,0.941666662693024,0.941666662693024} | {0.339193284511566,0.171411842107773,0.154963329434395,0.146068558096886} |                        1 |    0.0618178397417068 | {0.933333337306976,1,1,1}                                                 | {0.302341401576996,0.0991373136639595,0.0709080845117569,0.0618178397417068}
5 |        1 | loss='categorical_crossentropy',optimizer='Adam(lr=0.001)',metrics=['accuracy'] | batch_size=4,epochs=1 | madlib_keras | 0.75390625 | {36.3988909721375,71.1546559333801,107.219161987305,123.366652965546} | {accuracy}   | categorical_crossentropy |      0.908333361148834 |   0.656857788562775 | {0.358333319425583,0.683333337306976,0.766666650772095,0.908333361148834} | {0.996901094913483,0.852809607982635,0.694450318813324,0.656857788562775} |        0.899999976158142 |     0.630161166191101 | {0.233333334326744,0.600000023841858,0.633333325386047,0.899999976158142} | {1.05581676959991,0.876067101955414,0.700714349746704,0.630161166191101}
2 |        1 | loss='categorical_crossentropy',optimizer='Adam(lr=0.1)',metrics=['accuracy']   | batch_size=8,epochs=1 | madlib_keras | 0.75390625 | {35.0734770298004,69.949450969696,105.773796081543,121.901873111725}  | {accuracy}   | categorical_crossentropy |      0.866666674613953 |    0.26227542757988 | {0.558333337306976,0.683333337306976,0.708333313465118,0.866666674613953} | {0.755041897296906,0.551798105239868,0.504159986972809,0.26227542757988}  |        0.899999976158142 |     0.156955525279045 | {0.633333325386047,0.600000023841858,0.633333325386047,0.899999976158142} | {0.663675665855408,0.674827337265015,0.613502621650696,0.156955525279045}
6 |        1 | loss='categorical_crossentropy',optimizer='Adam(lr=0.001)',metrics=['accuracy'] | batch_size=8,epochs=1 | madlib_keras | 0.75390625 | {35.3355600833893,70.1827409267426,106.000793933868,122.135368108749} | {accuracy}   | categorical_crossentropy |      0.858333349227905 |   0.403681367635727 | {0.708333313465118,0.850000023841858,0.783333361148834,0.858333349227905} | {0.609176814556122,0.488206028938293,0.425172030925751,0.403681367635727} |        0.899999976158142 |     0.393621385097504 | {0.600000023841858,0.833333313465118,0.800000011920929,0.899999976158142} | {0.624664425849915,0.48302897810936,0.428876429796219,0.393621385097504}
9 |        2 | loss='categorical_crossentropy', optimizer='Adam(lr=0.01)',metrics=['accuracy'] | batch_size=4,epochs=1 | madlib_keras | 1.18359375 | {34.0762810707092,69.0332140922546,105.02664899826,121.034147024155}  | {accuracy}   | categorical_crossentropy |      0.816666662693024 |    0.40071240067482 | {0.716666638851166,0.975000023841858,0.941666662693024,0.816666662693024} | {0.530809044837952,0.112755678594112,0.178483173251152,0.40071240067482}  |        0.899999976158142 |     0.160617485642433 | {0.666666686534882,1,1,0.899999976158142}                                 | {0.510058879852295,0.0435655005276203,0.0271952152252197,0.160617485642433}
11 |        2 | loss='categorical_crossentropy',optimizer='Adam(lr=0.001)',metrics=['accuracy'] | batch_size=4,epochs=1 | madlib_keras | 1.18359375 | {36.664577960968,71.4262120723724,107.501835107803,123.646811962128}  | {accuracy}   | categorical_crossentropy |      0.816666662693024 |   0.683901190757751 | {0.358333319425583,0.574999988079071,0.758333325386047,0.816666662693024} | {0.969602465629578,0.828204095363617,0.7138671875,0.683901190757751}      |        0.833333313465118 |     0.720155417919159 | {0.233333334326744,0.466666668653488,0.666666686534882,0.833333313465118} | {1.04192113876343,0.878720223903656,0.755707621574402,0.720155417919159}
1 |        1 | loss='categorical_crossentropy',optimizer='Adam(lr=0.1)',metrics=['accuracy']   | batch_size=4,epochs=1 | madlib_keras | 0.75390625 | {35.8240220546722,70.6528959274292,106.680663108826,122.609847068787} | {accuracy}   | categorical_crossentropy |      0.683333337306976 |   0.479730814695358 | {0.675000011920929,0.683333337306976,0.683333337306976,0.683333337306976} | {0.507641136646271,0.467351347208023,0.507468938827515,0.479730814695358} |        0.600000023841858 |     0.503782331943512 | {0.600000023841858,0.600000023841858,0.600000023841858,0.600000023841858} | {0.504352450370789,0.448328793048859,0.561446607112885,0.503782331943512}
7 |        2 | loss='categorical_crossentropy',optimizer='Adam(lr=0.1)',metrics=['accuracy']   | batch_size=4,epochs=1 | madlib_keras | 1.18359375 | {33.5537171363831,68.2605030536652,104.443753004074,120.498863935471} | {accuracy}   | categorical_crossentropy |      0.683333337306976 |   0.940275907516479 | {0.733333349227905,0.641666650772095,0.933333337306976,0.683333337306976} | {0.419289141893387,0.746466338634491,0.25556743144989,0.940275907516479}  |        0.600000023841858 |       1.1002448797226 | {0.633333325386047,0.766666650772095,0.966666638851166,0.600000023841858} | {0.457020550966263,0.510054171085358,0.249405279755592,1.1002448797226}
12 |        2 | loss='categorical_crossentropy',optimizer='Adam(lr=0.001)',metrics=['accuracy'] | batch_size=8,epochs=1 | madlib_keras | 1.18359375 | {34.8103830814362,69.698233127594,105.547440052032,121.659696102142}  | {accuracy}   | categorical_crossentropy |      0.633333325386047 |   0.711242735385895 | {0.458333343267441,0.5,0.683333337306976,0.633333325386047}               | {1.71241450309753,0.98362809419632,0.776128530502319,0.711242735385895}   |        0.566666662693024 |     0.721668004989624 | {0.433333337306976,0.5,0.600000023841858,0.566666662693024}               | {1.30801546573639,0.950868666172028,0.841940879821777,0.721668004989624}
8 |        2 | loss='categorical_crossentropy',optimizer='Adam(lr=0.1)',metrics=['accuracy']   | batch_size=8,epochs=1 | madlib_keras | 1.18359375 | {36.0954039096832,70.921669960022,106.977710008621,123.112148046494}  | {accuracy}   | categorical_crossentropy |      0.358333319425583 |    1.10370969772339 | {0.324999988079071,0.358333319425583,0.358333319425583,0.358333319425583} | {1.10047388076782,1.09791541099548,1.09736382961273,1.10370969772339}     |        0.233333334326744 |      1.11518168449402 | {0.366666674613953,0.233333334326744,0.233333334326744,0.233333334326744} | {1.09480428695679,1.1150735616684,1.10827457904816,1.11518168449402}
(12 rows)

2. Predict probabilities for each class:
DROP TABLE IF EXISTS iris_predict;
'iris_test',        -- test_table
'id',               -- id column
'attributes',       -- independent var
'iris_predict',     -- output table
'prob',            -- prediction type
FALSE,             -- use gpus
3                  -- mst_key to use
);
SELECT * FROM iris_predict ORDER BY id, rank;

-----+------------+-----------------+---------------+------
8 | class_text | Iris-setosa     |    0.99961567 |    1
8 | class_text | Iris-versicolor | 0.00038426166 |    2
8 | class_text | Iris-virginica  | 5.4132368e-08 |    3
15 | class_text | Iris-setosa     |     0.9999677 |    1
15 | class_text | Iris-versicolor |  3.235117e-05 |    2
15 | class_text | Iris-virginica  | 7.5747697e-10 |    3
20 | class_text | Iris-setosa     |    0.99956757 |    1
20 | class_text | Iris-versicolor |  0.0004323488 |    2
20 | class_text | Iris-virginica  | 6.8698306e-08 |    3
21 | class_text | Iris-setosa     |    0.99978334 |    1
21 | class_text | Iris-versicolor | 0.00021666527 |    2
21 | class_text | Iris-virginica  |  1.741537e-08 |    3
22 | class_text | Iris-setosa     |    0.99955314 |    1
22 | class_text | Iris-versicolor | 0.00044669537 |    2
22 | class_text | Iris-virginica  |  8.287442e-08 |    3
27 | class_text | Iris-setosa     |      0.999393 |    1
27 | class_text | Iris-versicolor | 0.00060693553 |    2
27 | class_text | Iris-virginica  | 1.4125514e-07 |    3
28 | class_text | Iris-setosa     |     0.9997589 |    1
28 | class_text | Iris-versicolor | 0.00024109415 |    2
28 | class_text | Iris-virginica  | 2.3418018e-08 |    3
33 | class_text | Iris-setosa     |    0.99966764 |    1
33 | class_text | Iris-versicolor | 0.00033237415 |    2
33 | class_text | Iris-virginica  | 3.2712443e-08 |    3
37 | class_text | Iris-setosa     |     0.9999347 |    1
37 | class_text | Iris-versicolor | 6.5290136e-05 |    2
37 | class_text | Iris-virginica  | 2.6893372e-09 |    3
38 | class_text | Iris-setosa     |    0.99963343 |    1
38 | class_text | Iris-versicolor |  0.0003665546 |    2
38 | class_text | Iris-virginica  | 4.7346873e-08 |    3
46 | class_text | Iris-setosa     |     0.9995722 |    1
46 | class_text | Iris-versicolor | 0.00042761036 |    2
46 | class_text | Iris-virginica  | 8.4179504e-08 |    3
51 | class_text | Iris-versicolor |     0.9678325 |    1
51 | class_text | Iris-virginica  |   0.031179406 |    2
51 | class_text | Iris-setosa     |  0.0009880556 |    3
57 | class_text | Iris-versicolor |     0.7971095 |    1
57 | class_text | Iris-virginica  |    0.20281655 |    2
57 | class_text | Iris-setosa     |  7.402597e-05 |    3
58 | class_text | Iris-versicolor |    0.93808955 |    1
58 | class_text | Iris-virginica  |   0.045088027 |    2
58 | class_text | Iris-setosa     |   0.016822474 |    3
59 | class_text | Iris-versicolor |    0.95377713 |    1
59 | class_text | Iris-virginica  |   0.045621604 |    2
59 | class_text | Iris-setosa     |  0.0006012956 |    3
64 | class_text | Iris-versicolor |     0.8078371 |    1
64 | class_text | Iris-virginica  |    0.19210756 |    2
64 | class_text | Iris-setosa     | 5.5370016e-05 |    3
65 | class_text | Iris-versicolor |      0.946594 |    1
65 | class_text | Iris-virginica  |    0.04081257 |    2
65 | class_text | Iris-setosa     |   0.012593448 |    3
72 | class_text | Iris-versicolor |     0.9616088 |    1
72 | class_text | Iris-virginica  |      0.033955 |    2
72 | class_text | Iris-setosa     |   0.004436169 |    3
75 | class_text | Iris-versicolor |    0.96245867 |    1
75 | class_text | Iris-virginica  |    0.03556654 |    2
75 | class_text | Iris-setosa     |  0.0019747794 |    3
87 | class_text | Iris-versicolor |     0.9264334 |    1
87 | class_text | Iris-virginica  |    0.07328824 |    2
87 | class_text | Iris-setosa     | 0.00027841402 |    3
95 | class_text | Iris-versicolor |    0.85156035 |    1
95 | class_text | Iris-virginica  |    0.14813933 |    2
95 | class_text | Iris-setosa     | 0.00030031145 |    3
97 | class_text | Iris-versicolor |    0.87470025 |    1
97 | class_text | Iris-virginica  |     0.1248041 |    2
97 | class_text | Iris-setosa     |  0.0004957556 |    3
98 | class_text | Iris-versicolor |     0.9439469 |    1
98 | class_text | Iris-virginica  |    0.05491364 |    2
98 | class_text | Iris-setosa     |   0.001139452 |    3
120 | class_text | Iris-virginica  |    0.58107394 |    1
120 | class_text | Iris-versicolor |    0.41892523 |    2
120 | class_text | Iris-setosa     | 8.7891783e-07 |    3
121 | class_text | Iris-virginica  |    0.90112364 |    1
121 | class_text | Iris-versicolor |   0.098876335 |    2
121 | class_text | Iris-setosa     | 5.8106266e-09 |    3
122 | class_text | Iris-virginica  |     0.8664512 |    1
122 | class_text | Iris-versicolor |    0.13354866 |    2
122 | class_text | Iris-setosa     |  8.255242e-08 |    3
123 | class_text | Iris-virginica  |    0.90162355 |    1
123 | class_text | Iris-versicolor |    0.09837647 |    2
123 | class_text | Iris-setosa     |  2.745874e-10 |    3
140 | class_text | Iris-virginica  |     0.7306292 |    1
140 | class_text | Iris-versicolor |     0.2693706 |    2
140 | class_text | Iris-setosa     | 1.9480031e-07 |    3
144 | class_text | Iris-virginica  |    0.92295665 |    1
144 | class_text | Iris-versicolor |    0.07704339 |    2
144 | class_text | Iris-setosa     | 1.6369142e-09 |    3
147 | class_text | Iris-virginica  |    0.69545543 |    1
147 | class_text | Iris-versicolor |    0.30454406 |    2
147 | class_text | Iris-setosa     | 4.5714978e-07 |    3
(90 rows)

3. Warm start. Next, use the warm_start parameter to continue learning, using the coefficients from the run above. Note that we don't drop the model table or model summary table:
SELECT madlib.madlib_keras_fit_multiple_model('iris_train_packed',    -- source_table
'iris_multi_model',     -- model_output_table
'mst_table',            -- model_selection_table
3,                     -- num_iterations
FALSE,                 -- use gpus
'iris_test_packed',     -- validation dataset
1,                     -- metrics compute frequency
TRUE,                  -- warm start
'Sophie L.',            -- name
'Simple MLP for iris dataset',  -- description
TRUE                   -- use caching
);
SELECT * FROM iris_multi_model_summary;

-[ RECORD 1 ]-------------+---------------------------------------------
source_table              | iris_train_packed
validation_table          | iris_test_packed
model                     | iris_multi_model
model_info                | iris_multi_model_info
dependent_varname         | {class_text}
independent_varname       | {attributes}
model_arch_table          | model_arch_library
model_selection_table     | mst_table
object_table              |
num_iterations            | 3
metrics_compute_frequency | 1
warm_start                | t
name                      | Sophie L.
description               | Simple MLP for iris dataset
start_training_time       | 2021-02-05 01:17:19.432839
end_training_time         | 2021-02-05 01:18:09.062384
num_classes               | {1}
class_text_class_values   | {Iris-setosa,Iris-versicolor,Iris-virginica}
dependent_vartype         | {"character varying"}
normalizing_const         | 1
metrics_iters             | {1,2,3}

View results for each model:
SELECT * FROM iris_multi_model_info ORDER BY training_metrics_final DESC, training_loss_final;

  mst_key | model_id |                                 compile_params                                  |      fit_params       |  model_type  | model_size |                 metrics_elapsed_time                 | metrics_type |        loss_type         | training_metrics_final | training_loss_final |                    training_metrics                     |                      training_loss                      | validation_metrics_final | validation_loss_final |                   validation_metrics                    |                      validation_loss
---------+----------+---------------------------------------------------------------------------------+-----------------------+--------------+------------+------------------------------------------------------+--------------+--------------------------+------------------------+---------------------+---------------------------------------------------------+---------------------------------------------------------+--------------------------+-----------------------+---------------------------------------------------------+-----------------------------------------------------------
10 |        2 | loss='categorical_crossentropy', optimizer='Adam(lr=0.01)',metrics=['accuracy'] | batch_size=8,epochs=1 | madlib_keras | 1.18359375 | {15.5192940235138,31.2597029209137,47.3274641036987} | {accuracy}   | categorical_crossentropy |      0.966666638851166 |   0.122704714536667 | {0.925000011920929,0.808333337306976,0.966666638851166} | {0.172832280397415,0.354962348937988,0.122704714536667} |                        1 |    0.0603121444582939 | {1,0.866666674613953,1}                                 | {0.100369863212109,0.210344776511192,0.0603121444582939}
4 |        1 | loss='categorical_crossentropy', optimizer='Adam(lr=0.01)',metrics=['accuracy'] | batch_size=8,epochs=1 | madlib_keras | 0.75390625 | {16.485249042511,32.207494020462,48.3005399703979}   | {accuracy}   | categorical_crossentropy |      0.958333313465118 |   0.104984156787395 | {0.949999988079071,0.866666674613953,0.958333313465118} | {0.126872479915619,0.29683381319046,0.104984156787395}  |                        1 |    0.0291607566177845 | {1,0.933333337306976,1}                                 | {0.0433581434190273,0.127146035432816,0.0291607566177845}
3 |        1 | loss='categorical_crossentropy', optimizer='Adam(lr=0.01)',metrics=['accuracy'] | batch_size=4,epochs=1 | madlib_keras | 0.75390625 | {14.8933939933777,30.7211890220642,46.7588970661163} | {accuracy}   | categorical_crossentropy |      0.958333313465118 |   0.107885278761387 | {0.941666662693024,0.966666638851166,0.958333313465118} | {0.155905768275261,0.129546254873276,0.107885278761387} |                        1 |    0.0387893654406071 | {1,1,1}                                                 | {0.087645135819912,0.0574964620172977,0.0387893654406071}
7 |        2 | loss='categorical_crossentropy',optimizer='Adam(lr=0.1)',metrics=['accuracy']   | batch_size=4,epochs=1 | madlib_keras | 1.18359375 | {14.657429933548,30.4864449501038,46.5082159042358}  | {accuracy}   | categorical_crossentropy |      0.958333313465118 |   0.215500101447105 | {0.891666650772095,0.649999976158142,0.958333313465118} | {0.2352494597435,0.490339070558548,0.215500101447105}   |                        1 |     0.194652214646339 | {0.899999976158142,0.766666650772095,1}                 | {0.235657200217247,0.335934072732925,0.194652214646339}
6 |        1 | loss='categorical_crossentropy',optimizer='Adam(lr=0.001)',metrics=['accuracy'] | batch_size=8,epochs=1 | madlib_keras | 0.75390625 | {16.2535049915314,31.9760749340057,48.0661840438843} | {accuracy}   | categorical_crossentropy |      0.941666662693024 |   0.354692339897156 | {0.975000023841858,0.908333361148834,0.941666662693024} | {0.38957467675209,0.367432981729507,0.354692339897156}  |                        1 |     0.321579396724701 | {1,0.966666638851166,1}                                 | {0.348732680082321,0.345716863870621,0.321579396724701}
2 |        1 | loss='categorical_crossentropy',optimizer='Adam(lr=0.1)',metrics=['accuracy']   | batch_size=8,epochs=1 | madlib_keras | 0.75390625 | {16.0124208927155,31.7479128837585,47.8348250389099} | {accuracy}   | categorical_crossentropy |      0.933333337306976 |   0.238343968987465 | {0.975000023841858,0.975000023841858,0.933333337306976} | {0.17331151664257,0.0775784701108932,0.238343968987465} |        0.966666638851166 |     0.182112962007523 | {0.966666638851166,1,0.966666638851166}                 | {0.114220358431339,0.0159573350101709,0.182112962007523}
9 |        2 | loss='categorical_crossentropy', optimizer='Adam(lr=0.01)',metrics=['accuracy'] | batch_size=4,epochs=1 | madlib_keras | 1.18359375 | {15.1569800376892,30.9868409633636,47.0489699840546} | {accuracy}   | categorical_crossentropy |      0.908333361148834 |   0.254226505756378 | {0.858333349227905,0.833333313465118,0.908333361148834} | {0.450037389993668,0.500195503234863,0.254226505756378} |        0.966666638851166 |    0.0769383609294891 | {0.899999976158142,0.899999976158142,0.966666638851166} | {0.163723081350327,0.211927503347397,0.0769383609294891}
11 |        2 | loss='categorical_crossentropy',optimizer='Adam(lr=0.001)',metrics=['accuracy'] | batch_size=4,epochs=1 | madlib_keras | 1.18359375 | {17.7095139026642,33.4267060756683,49.6263670921326} | {accuracy}   | categorical_crossentropy |      0.899999976158142 |   0.602623522281647 | {0.824999988079071,0.899999976158142,0.899999976158142} | {0.65915185213089,0.633350729942322,0.602623522281647}  |        0.899999976158142 |     0.626452326774597 | {0.833333313465118,0.899999976158142,0.899999976158142} | {0.695301294326782,0.661069989204407,0.626452326774597}
5 |        1 | loss='categorical_crossentropy',optimizer='Adam(lr=0.001)',metrics=['accuracy'] | batch_size=4,epochs=1 | madlib_keras | 0.75390625 | {17.4354989528656,33.1478018760681,49.3430600166321} | {accuracy}   | categorical_crossentropy |      0.741666674613953 |   0.532543838024139 | {0.916666686534882,0.875,0.741666674613953}             | {0.610130310058594,0.571656167507172,0.532543838024139} |        0.633333325386047 |     0.540634453296661 | {0.966666638851166,0.899999976158142,0.633333325386047} | {0.601030588150024,0.567581355571747,0.540634453296661}
12 |        2 | loss='categorical_crossentropy',optimizer='Adam(lr=0.001)',metrics=['accuracy'] | batch_size=8,epochs=1 | madlib_keras | 1.18359375 | {15.7747659683228,31.5215599536896,47.6013031005859} | {accuracy}   | categorical_crossentropy |      0.675000011920929 |   0.581513583660126 | {0.683333337306976,0.683333337306976,0.675000011920929} | {0.660648465156555,0.623969316482544,0.581513583660126} |        0.600000023841858 |     0.594986796379089 | {0.600000023841858,0.600000023841858,0.600000023841858} | {0.694275736808777,0.665920257568359,0.594986796379089}
1 |        1 | loss='categorical_crossentropy',optimizer='Adam(lr=0.1)',metrics=['accuracy']   | batch_size=4,epochs=1 | madlib_keras | 0.75390625 | {16.8983609676361,32.4361009597778,48.5357549190521} | {accuracy}   | categorical_crossentropy |      0.608333349227905 |   0.687089145183563 | {0.583333313465118,0.699999988079071,0.608333349227905} | {0.816642761230469,0.455933183431625,0.687089145183563} |        0.533333361148834 |     0.692391216754913 | {0.533333361148834,0.600000023841858,0.533333361148834} | {0.85290253162384,0.465981602668762,0.692391216754913}
8 |        2 | loss='categorical_crossentropy',optimizer='Adam(lr=0.1)',metrics=['accuracy']   | batch_size=8,epochs=1 | madlib_keras | 1.18359375 | {17.1899569034576,32.8988509178162,49.0880739688873} | {accuracy}   | categorical_crossentropy |      0.358333319425583 |    1.11154329776764 | {0.358333319425583,0.324999988079071,0.358333319425583} | {1.09858739376068,1.09980893135071,1.11154329776764}    |        0.233333334326744 |      1.16109442710876 | {0.233333334326744,0.366666674613953,0.233333334326744} | {1.12281405925751,1.10376691818237,1.16109442710876}
(12 rows)

Note that the loss and accuracy values pick up from where the previous run left off.

Notes
1. Refer to the deep learning section of the Apache MADlib wiki [6] for important information including supported libraries and versions.
2. Classification is currently supported, not regression.
3. Reminder about the distinction between warm start and transfer learning. Warm start uses model state (weights) from the model output table from a previous training run - set the 'warm_start' parameter to TRUE in the fit function. Transfer learning uses initial model state (weights) stored in the 'model_arch_table' - in this case set the 'warm_start' parameter to FALSE in the fit function.
4. Here are some more details on how warm start works. These details are mostly applicable when implementing autoML algorithms on top of MADlib's model selection. In short, the 'model_selection_table' dictates which models get trained and output to the 'model_output_table' and associated summary and info tables. When 'warm_start' is TRUE, models are built for each 'mst_key' in the 'model_selection_table'. If there are prior runs for an 'mst_key' then the weights from that run will be used. If there are no prior runs for an 'mst_key' then random initialization will be used. For example, let's say we start with 'mst_keys' of 1, 2, 3, and 4 in the 'model_selection_table'. We run fit once to get model and info tables for 1, 2, 3, and 4. Then we modify the 'model_selection_table' as part of an autoML scheme, in which we remove the 'mst_key' for 1 and add a new 'mst_key' for 5. Next we run fit with warm start. The result will be models created for 'mst_keys' of 2, 3, 4, and 5. Warm start will be used for 2, 3, and 4 (using prior run) and random initialization will be used for 5 (no prior run). The 'mst_key' of 1 will be dropped.
5. The 'num_iterations' parameter and the Keras fit parameter 'epochs' can substantially affect accuracy and run-time. In general, increasing the number of 'epochs' for a fixed 'num_iterations' will speed up training, but may result in lower accuracy. It's best to do some experimentation to find out what works for your models and dataset.

Technical Background

For an introduction to deep learning foundations, including MLP and CNN, refer to [7].

This module trains many models a time across the database cluster in order to explore network architectures and hyperparameters. It uses model hopper parallelism (MOP) and has high convergence efficiency since it does not do model averaging [2].

On the effect of database cluster size: as the database cluster size increases, it will be faster to train a set of models, as long as you have at least as many model selection tuples as segments. This is because model state is "hopped" from segment to segment and training takes place in parallel [1,2]. If you have fewer model selection tuples to train than segments, then some segments may not be busy 100% of the time so speedup will not necessarily increase on a larger cluster. Inference (prediction) is an embarrassingly parallel operation so inference runtimes will be proportionally faster as the number of segments increases.

Literature

[1] "Cerebro: Efficient and Reproducible Model Selection on Deep Learning Systems," Supun Nakandala, Yuhao Zhang, and Arun Kumar, ACM SIGMOD 2019 DEEM Workshop, https://adalabucsd.github.io/papers/2019_Cerebro_DEEM.pdf

[2] "Cerebro: A Data System for Optimized Deep Learning Model Selection," Supun Nakandala, Yuhao Zhang, and Arun Kumar, Proceedings of the VLDB Endowment (2020), Vol. 13, No. 11 https://adalabucsd.github.io/papers/2020_Cerebro_VLDB.pdf

[5] "Neural Networks for Machine Learning", Lectures 6a and 6b on mini-batch gradient descent, Geoffrey Hinton with Nitish Srivastava and Kevin Swersky, http://www.cs.toronto.edu/~tijmen/csc321/slides/lecture_slides_lec6.pdf