July 17,2023 release -- config updates

config_templates/gretel/synthetics/amplify.yml

@@ -16,3 +16,4 @@ models:
         params:
           num_records: null
           target_size_mb: null
+          auto_transform_datetimes: True

config_templates/gretel/synthetics/default.yml

@@ -1,17 +1,18 @@
 # deprecated: This configuration will be deprecated soon. Please use the "synthetics/tabular-lstm" configuration.
 
 schema_version: "1.0"
-name: "default-config"
+name: "tabular-lstm"
 models:
   - synthetics:
       data_source: __tmp__
       params:
-        epochs: 100
-        vocab_size: 20000
-        learning_rate: 0.01
-        validation_split: False
+        epochs: auto
+        vocab_size: auto
+        learning_rate: auto
+        batch_size: auto
+        rnn_units: auto
       generate:
         num_records: 5000
       privacy_filters:
         outliers: auto
-        similarity: auto
+        similarity: auto

config_templates/gretel/synthetics/high-dimensionality.yml

@@ -12,8 +12,9 @@ models:
             generator_lr: 0.0001
             discriminator_lr: .00033
             batch_size: auto
+            auto_transform_datetimes: False
         generate:
             num_records: 5000
         privacy_filters:
-            outliers: auto
-            similarity: auto
+            outliers: null # Set to "auto" for additional protections
+            similarity: null # Set to "auto" for additional protections

config_templates/gretel/synthetics/low-record-count.yml

@@ -4,16 +4,17 @@ schema_version: "1.0"
 name: "low-record-count"
 models:
   - actgan:
-        data_source: __tmp__
-        params:
-            epochs: auto
-            generator_dim: [1024, 1024]
-            discriminator_dim: [1024, 1024]
-            generator_lr: 0.0001
-            discriminator_lr: .00033
-            batch_size: auto
-        generate:
-            num_records: 5000
-        privacy_filters:
-            outliers: auto
-            similarity: auto
+      data_source: __tmp__
+      params:
+        epochs: auto
+        generator_dim: [1024, 1024]
+        discriminator_dim: [1024, 1024]
+        generator_lr: 0.0001
+        discriminator_lr: .00033
+        batch_size: auto
+        auto_transform_datetimes: False
+      generate:
+        num_records: 5000
+      privacy_filters:
+        outliers: null # Set to "auto" for additional protections
+        similarity: null # Set to "auto" for additional protections

config_templates/gretel/synthetics/tabular-actgan.yml

@@ -12,8 +12,9 @@ models:
             generator_lr: 0.0001
             discriminator_lr: .00033
             batch_size: auto
+            auto_transform_datetimes: False
         generate:
             num_records: 5000
         privacy_filters:
-            outliers: auto
-            similarity: auto
+            outliers: null # Set to "auto" for additional protections
+            similarity: null # Set to "auto" for additional protections

config_templates/gretel/synthetics/tabular-lstm-evaluate.yml

@@ -1,19 +1,10 @@
 # Blueprint configuration for evaluating synthetic data on classification and regression models
-
-### IMPORTANT: rename "target" to match the label/target header of your data ###
-
-# See https://docs.gretel.ai/reference/evaluate/evaluate-tasks/classification and
-# https://docs.gretel.ai/reference/evaluate/evaluate-tasks/regression  
-# for detailed info on all config options
-
+# You can substitute with any synthetic model. For example, to use Gretel ACTGAN or TabularDP models,
+# edit the default config and add the "evaluate" parameters below.
 
 schema_version: "1.0"
 name: "tabular-lstm-evaluate"
 models:
-
-  # This config uses the Gretel LSTM model to generate synthetic data. 
-  # You can substitute with any synthetic model. For example, to use Gretel ACTGAN model,
-  # copy the config from "tabular-actgan.yml"
   - synthetics:
       data_source: __tmp__
       params:
@@ -28,14 +19,11 @@ models:
         outliers: auto
         similarity: auto
       evaluate:
-
         ### Indicate classification or regression 
-        task: classification  
-
-        ### IMPORTANT: rename to match the label/target header of your data 
-        target: "y"           
+        task: classification # Indicate classification or regression
+        target: "y" # The column in your dataset containing the desired label
 
-        ### Optional metrics. See docs for all config options: https://docs.gretel.ai/reference/evaluate/evaluate-tasks/
+        ### Optional metrics
         # holdout: null       
         # metric: null       
         # models: null