Provides actions for training generative adversarial network models.
Trains a styleGAN model.
If a row includes a subparameter, you can specify the name, caslib, and so on in the subparameter. Otherwise, you can specify the name, caslib, and so on in the parameter.
|
Parameter |
Subparameter |
Description |
|---|---|---|
|
— |
specifies the checkpoint CAS table to be loaded to restore your model. If you omit this parameter, the model is trained from scratch. |
|
|
required parametertable |
— |
specifies the input CAS table of training images. |
|
Parameter |
Subparameter |
Description |
|---|---|---|
|
required parametercasOut |
— |
specifies the output CAS table in which to store the generated images from the trained model. |
|
— |
specifies the CAS table in which to save the current model as a checkpoint. |
|
|
names |
lists the names of results tables to save as CAS tables on the server. |
|
|
required parametersaveState |
— |
specifies the table in which to save the model state for model scoring. |
specifies the output CAS table in which to store the generated images from the trained model.
For more information about specifying the casOut parameter, see the common casouttable (Form 1) parameter (Appendix A: Common Parameters).
specifies the frequency at which to save a checkpoint during the training.
| Default | 5000 |
|---|---|
| Minimum value | 1 |
specifies the checkpoint CAS table to be loaded to restore your model. If you omit this parameter, the model is trained from scratch.
For more information about specifying the checkpointIn parameter, see the common castable (Form 1) parameter (Appendix A: Common Parameters).
specifies the CAS table in which to save the current model as a checkpoint.
For more information about specifying the checkpointOut parameter, see the common casouttable (Form 1) parameter (Appendix A: Common Parameters).
specifies a list of results tables to send to the client for display.
For more information about specifying the display parameter, see the common displayTables parameter (Appendix A: Common Parameters).
specifies GPU-related options.
The gpuOptions value can be one or more of the following:
when set to True, tells the NVIDIA CUDA Deep Neural Network library (CuDNN) that you want only the deterministic implementations. NOTE: Setting this parameter to True can significantly slow down the training process.
| Default | FALSE |
|---|
specifies the ID of one GPU device that is to be used.
| Alias | devices |
|---|---|
| Default | 0 |
| Minimum value | 0 |
| Requirement | The specified values must be unique. |
when set to True, specifies that a GPU device is to be used.
| Alias | useGPU |
|---|---|
| Default | TRUE |
specifies the image size (width of square images in pixels) in the training data.
| Range | 4–1024 |
|---|
when set to True, specifies that large networks are to be used.
| Default | TRUE |
|---|
specifies the number of images to generate.
| Alias | nSamples |
|---|---|
| Default | 1 |
| Minimum value | 1 |
specifies the number of latents to use in calculating the truncation mean.
| Alias | ntl |
|---|---|
| Default | 4096 |
| Minimum value (exclusive) | 0 |
specifies the optimization method to use. Currently only ADAM is supported.
| Alias | optimization |
|---|
The value that you specify for the method parameter determines the other parameters that apply.
lists the names of results tables to save as CAS tables on the server.
For more information about specifying the outputTables parameter, see the common outputTables parameter (Appendix A: Common Parameters).
specifies the frequency at which to print losses to the log and the iteration history table.
| Default | 1 |
|---|---|
| Minimum value | 0 |
specifies the frequency at which to regularize the discriminator.
| Default | 16 |
|---|---|
| Minimum value | 1 |
specifies the frequency at which to regularize the generator.
| Default | 4 |
|---|---|
| Minimum value | 1 |
specifies the weight for regularizing the path length loss.
| Default | 2 |
|---|---|
| Minimum value (exclusive) | 0 |
specifies the weight for regularizing the R1 loss.
| Default | 10 |
|---|---|
| Minimum value (exclusive) | 0 |
specifies the table in which to save the model state for model scoring.
| Long form | saveState={name="table-name"} |
|---|---|
| Shortcut form | saveState="table-name" |
The casouttable value can be one or more of the following:
specifies the name of the caslib for the output table.
specifies the descriptive label to associate with the table.
specifies the number of seconds to keep the table in memory after it is last accessed. The table is dropped if it is not accessed for the specified number of seconds.
| Default | 0 |
|---|---|
| Minimum value | 0 |
specifies the memory format for the output table.
| Default | INHERIT |
|---|
use the duplicate value reduction memory format. This memory format can reduce the memory consumption and file size when the input data contains duplicate values.
specifies the name for the output table.
when set to True, adds the output table with a global scope. This enables other sessions to access the table, subject to access controls. The target caslib must also have a global scope.
| Default | FALSE |
|---|
when set to True, overwrites an existing table that has the same name.
| Default | FALSE |
|---|
specifies the seed to use for the random number generator for scoring. A value of 0 means use the time from the computer's clock as the seed.
| Default | 0 |
|---|---|
| Minimum value | 0 |
specifies the seed to use for the random number generator for training. A value of 0 means use the time from the computer's clock as the seed.
| Default | 0 |
|---|---|
| Minimum value | 0 |
specifies the input CAS table of training images.
| Long form | table={name="table-name"} |
|---|---|
| Shortcut form | table="table-name" |
The castable value can be one or more of the following:
specifies the caslib for the input table that you want to use with the action. By default, the active caslib is used. Specify a value only if you need to access a table from a different caslib.
when set to True, creates the computed variables when the table is loaded instead of when the action begins.
| Alias | compOnDemand |
|---|---|
| Default | FALSE |
specifies the names of the computed variables to create. Specify an expression for each variable in the computedVarsProgram parameter. If you do not specify this parameter, then all variables from computedVarsProgram are automatically included.
| Alias | compVars |
|---|
The casinvardesc value can be one or more of the following:
specifies the format to apply to the variable.
specifies the length of the format field plus the length of the format precision.
specifies the descriptive label for the variable.
specifies the name for the variable.
specifies the length of the format precision.
specifies the length of the format field.
specifies an expression for each computed variable that you include in the computedVars parameter.
| Alias | compPgm |
|---|
specifies data source options.
| Aliases | options |
|---|---|
| dataSource |
specifies the settings for reading a table from a data source.
| Alias | import |
|---|
For more information about specifying the importOptions parameter, see the common importOptions parameter (Appendix A: Common Parameters).
specifies the name of the input table.
specifies the variables to use for ordering observations within partitions. This parameter applies to partitioned tables, or it can be combined with variables that are specified in the groupBy parameter when the value of the groupByMode parameter is set to REDISTRIBUTE.
The casinvardesc value can be one or more of the following:
specifies the format to apply to the variable.
specifies the length of the format field plus the length of the format precision.
specifies the descriptive label for the variable.
specifies the name for the variable.
specifies the length of the format precision.
specifies the length of the format field.
when set to True, does not create a transient table on the server. Setting this parameter to True can be efficient, but the data might not have stable ordering upon repeated runs.
| Default | FALSE |
|---|
specifies the variables to use in the action.
The casinvardesc value can be one or more of the following:
specifies the format to apply to the variable.
specifies the length of the format field plus the length of the format precision.
specifies the descriptive label for the variable.
specifies the name for the variable.
specifies the length of the format precision.
specifies the length of the format field.
specifies an expression for subsetting the input data.
specifies an input table that contains rows to use as a WHERE filter. If the vars parameter is not specified, then all the variable names that are common to the input table and the filtering table are used to find matching rows. If the where parameter for the input table and this parameter are specified, then this filtering table is applied first.
The groupbytable value can be one or more of the following:
specifies the caslib for the filter table. By default, the active caslib is used.
specifies data source options.
| Aliases | options |
|---|---|
| dataSource |
For more information about specifying the dataSourceOptions parameter, see the common dataSourceOptions parameter (Appendix A: Common Parameters).
specifies the settings for reading a table from a data source.
| Alias | import |
|---|
For more information about specifying the importOptions parameter, see the common importOptions parameter (Appendix A: Common Parameters).
specifies the name of the filter table.
specifies the variable names to use from the filter table.
The casinvardesc value can be one or more of the following:
specifies the format to apply to the variable.
specifies the length of the format field plus the length of the format precision.
specifies the descriptive label for the variable.
specifies the name for the variable.
specifies the length of the format precision.
specifies the length of the format field.
specifies an expression for subsetting the data from the filter table.
specifies the truncation ratio for image generation.
| Default | 1 |
|---|---|
| Range | (0–1] |
specifies the exponential decay rate for the first-moment estimates.
| Default | 0 |
|---|---|
| Range | [0–1) |
specifies the exponential decay rate for the second-moment estimates.
| Default | 0.99 |
|---|---|
| Range | [0–1) |
specifies the learning rate for the optimizer.
| Default | 0.002 |
|---|---|
| Range | (0–1] |
specifies the number of images in one minibatch.
| Default | 32 |
|---|---|
| Minimum value | 4 |
specifies the number of minibatches to be used for training the model.
| Default | 10 |
|---|---|
| Minimum value | 0 |
Trains a styleGAN model.
If a row includes a subparameter, you can specify the name, caslib, and so on in the subparameter. Otherwise, you can specify the name, caslib, and so on in the parameter.
|
Parameter |
Subparameter |
Description |
|---|---|---|
|
— |
specifies the checkpoint CAS table to be loaded to restore your model. If you omit this parameter, the model is trained from scratch. |
|
|
required parametertable |
— |
specifies the input CAS table of training images. |
|
Parameter |
Subparameter |
Description |
|---|---|---|
|
required parametercasOut |
— |
specifies the output CAS table in which to store the generated images from the trained model. |
|
— |
specifies the CAS table in which to save the current model as a checkpoint. |
|
|
names |
lists the names of results tables to save as CAS tables on the server. |
|
|
required parametersaveState |
— |
specifies the table in which to save the model state for model scoring. |
specifies the output CAS table in which to store the generated images from the trained model.
For more information about specifying the casOut parameter, see the common casouttable (Form 1) parameter (Appendix A: Common Parameters).
specifies the frequency at which to save a checkpoint during the training.
| Default | 5000 |
|---|---|
| Minimum value | 1 |
specifies the checkpoint CAS table to be loaded to restore your model. If you omit this parameter, the model is trained from scratch.
For more information about specifying the checkpointIn parameter, see the common castable (Form 1) parameter (Appendix A: Common Parameters).
specifies the CAS table in which to save the current model as a checkpoint.
For more information about specifying the checkpointOut parameter, see the common casouttable (Form 1) parameter (Appendix A: Common Parameters).
specifies a list of results tables to send to the client for display.
For more information about specifying the display parameter, see the common displayTables parameter (Appendix A: Common Parameters).
specifies GPU-related options.
The gpuOptions value can be one or more of the following:
when set to True, tells the NVIDIA CUDA Deep Neural Network library (CuDNN) that you want only the deterministic implementations. NOTE: Setting this parameter to True can significantly slow down the training process.
| Default | false |
|---|
specifies the ID of one GPU device that is to be used.
| Alias | devices |
|---|---|
| Default | 0 |
| Minimum value | 0 |
| Requirement | The specified values must be unique. |
when set to True, specifies that a GPU device is to be used.
| Alias | useGPU |
|---|---|
| Default | true |
specifies the image size (width of square images in pixels) in the training data.
| Range | 4–1024 |
|---|
when set to True, specifies that large networks are to be used.
| Default | true |
|---|
specifies the number of images to generate.
| Alias | nSamples |
|---|---|
| Default | 1 |
| Minimum value | 1 |
specifies the number of latents to use in calculating the truncation mean.
| Alias | ntl |
|---|---|
| Default | 4096 |
| Minimum value (exclusive) | 0 |
specifies the optimization method to use. Currently only ADAM is supported.
| Alias | optimization |
|---|
The value that you specify for the method parameter determines the other parameters that apply.
lists the names of results tables to save as CAS tables on the server.
For more information about specifying the outputTables parameter, see the common outputTables parameter (Appendix A: Common Parameters).
specifies the frequency at which to print losses to the log and the iteration history table.
| Default | 1 |
|---|---|
| Minimum value | 0 |
specifies the frequency at which to regularize the discriminator.
| Default | 16 |
|---|---|
| Minimum value | 1 |
specifies the frequency at which to regularize the generator.
| Default | 4 |
|---|---|
| Minimum value | 1 |
specifies the weight for regularizing the path length loss.
| Default | 2 |
|---|---|
| Minimum value (exclusive) | 0 |
specifies the weight for regularizing the R1 loss.
| Default | 10 |
|---|---|
| Minimum value (exclusive) | 0 |
specifies the table in which to save the model state for model scoring.
| Long form | saveState={name="table-name"} |
|---|---|
| Shortcut form | saveState="table-name" |
The casouttable value can be one or more of the following:
specifies the name of the caslib for the output table.
specifies the descriptive label to associate with the table.
specifies the number of seconds to keep the table in memory after it is last accessed. The table is dropped if it is not accessed for the specified number of seconds.
| Default | 0 |
|---|---|
| Minimum value | 0 |
specifies the memory format for the output table.
| Default | INHERIT |
|---|
use the duplicate value reduction memory format. This memory format can reduce the memory consumption and file size when the input data contains duplicate values.
specifies the name for the output table.
when set to True, adds the output table with a global scope. This enables other sessions to access the table, subject to access controls. The target caslib must also have a global scope.
| Default | false |
|---|
when set to True, overwrites an existing table that has the same name.
| Default | false |
|---|
specifies the seed to use for the random number generator for scoring. A value of 0 means use the time from the computer's clock as the seed.
| Default | 0 |
|---|---|
| Minimum value | 0 |
specifies the seed to use for the random number generator for training. A value of 0 means use the time from the computer's clock as the seed.
| Default | 0 |
|---|---|
| Minimum value | 0 |
specifies the input CAS table of training images.
| Long form | table={name="table-name"} |
|---|---|
| Shortcut form | table="table-name" |
The castable value can be one or more of the following:
specifies the caslib for the input table that you want to use with the action. By default, the active caslib is used. Specify a value only if you need to access a table from a different caslib.
when set to True, creates the computed variables when the table is loaded instead of when the action begins.
| Alias | compOnDemand |
|---|---|
| Default | false |
specifies the names of the computed variables to create. Specify an expression for each variable in the computedVarsProgram parameter. If you do not specify this parameter, then all variables from computedVarsProgram are automatically included.
| Alias | compVars |
|---|
The casinvardesc value can be one or more of the following:
specifies the format to apply to the variable.
specifies the length of the format field plus the length of the format precision.
specifies the descriptive label for the variable.
specifies the name for the variable.
specifies the length of the format precision.
specifies the length of the format field.
specifies an expression for each computed variable that you include in the computedVars parameter.
| Alias | compPgm |
|---|
specifies data source options.
| Aliases | options |
|---|---|
| dataSource |
specifies the settings for reading a table from a data source.
| Alias | import |
|---|
For more information about specifying the importOptions parameter, see the common importOptions parameter (Appendix A: Common Parameters).
specifies the name of the input table.
specifies the variables to use for ordering observations within partitions. This parameter applies to partitioned tables, or it can be combined with variables that are specified in the groupBy parameter when the value of the groupByMode parameter is set to REDISTRIBUTE.
The casinvardesc value can be one or more of the following:
specifies the format to apply to the variable.
specifies the length of the format field plus the length of the format precision.
specifies the descriptive label for the variable.
specifies the name for the variable.
specifies the length of the format precision.
specifies the length of the format field.
when set to True, does not create a transient table on the server. Setting this parameter to True can be efficient, but the data might not have stable ordering upon repeated runs.
| Default | false |
|---|
specifies the variables to use in the action.
The casinvardesc value can be one or more of the following:
specifies the format to apply to the variable.
specifies the length of the format field plus the length of the format precision.
specifies the descriptive label for the variable.
specifies the name for the variable.
specifies the length of the format precision.
specifies the length of the format field.
specifies an expression for subsetting the input data.
specifies an input table that contains rows to use as a WHERE filter. If the vars parameter is not specified, then all the variable names that are common to the input table and the filtering table are used to find matching rows. If the where parameter for the input table and this parameter are specified, then this filtering table is applied first.
The groupbytable value can be one or more of the following:
specifies the caslib for the filter table. By default, the active caslib is used.
specifies data source options.
| Aliases | options |
|---|---|
| dataSource |
For more information about specifying the dataSourceOptions parameter, see the common dataSourceOptions parameter (Appendix A: Common Parameters).
specifies the settings for reading a table from a data source.
| Alias | import |
|---|
For more information about specifying the importOptions parameter, see the common importOptions parameter (Appendix A: Common Parameters).
specifies the name of the filter table.
specifies the variable names to use from the filter table.
The casinvardesc value can be one or more of the following:
specifies the format to apply to the variable.
specifies the length of the format field plus the length of the format precision.
specifies the descriptive label for the variable.
specifies the name for the variable.
specifies the length of the format precision.
specifies the length of the format field.
specifies an expression for subsetting the data from the filter table.
specifies the truncation ratio for image generation.
| Default | 1 |
|---|---|
| Range | (0–1] |
specifies the exponential decay rate for the first-moment estimates.
| Default | 0 |
|---|---|
| Range | [0–1) |
specifies the exponential decay rate for the second-moment estimates.
| Default | 0.99 |
|---|---|
| Range | [0–1) |
specifies the learning rate for the optimizer.
| Default | 0.002 |
|---|---|
| Range | (0–1] |
specifies the number of images in one minibatch.
| Default | 32 |
|---|---|
| Minimum value | 4 |
specifies the number of minibatches to be used for training the model.
| Default | 10 |
|---|---|
| Minimum value | 0 |
Trains a styleGAN model.
If a row includes a subparameter, you can specify the name, caslib, and so on in the subparameter. Otherwise, you can specify the name, caslib, and so on in the parameter.
|
Parameter |
Subparameter |
Description |
|---|---|---|
|
— |
specifies the checkpoint CAS table to be loaded to restore your model. If you omit this parameter, the model is trained from scratch. |
|
|
required parametertable |
— |
specifies the input CAS table of training images. |
|
Parameter |
Subparameter |
Description |
|---|---|---|
|
required parametercasOut |
— |
specifies the output CAS table in which to store the generated images from the trained model. |
|
— |
specifies the CAS table in which to save the current model as a checkpoint. |
|
|
names |
lists the names of results tables to save as CAS tables on the server. |
|
|
required parametersaveState |
— |
specifies the table in which to save the model state for model scoring. |
specifies the output CAS table in which to store the generated images from the trained model.
For more information about specifying the casOut parameter, see the common casouttable (Form 1) parameter (Appendix A: Common Parameters).
specifies the frequency at which to save a checkpoint during the training.
| Default | 5000 |
|---|---|
| Minimum value | 1 |
specifies the checkpoint CAS table to be loaded to restore your model. If you omit this parameter, the model is trained from scratch.
For more information about specifying the checkpointIn parameter, see the common castable (Form 1) parameter (Appendix A: Common Parameters).
specifies the CAS table in which to save the current model as a checkpoint.
For more information about specifying the checkpointOut parameter, see the common casouttable (Form 1) parameter (Appendix A: Common Parameters).
specifies a list of results tables to send to the client for display.
For more information about specifying the display parameter, see the common displayTables parameter (Appendix A: Common Parameters).
specifies GPU-related options.
The gpuOptions value can be one or more of the following:
when set to True, tells the NVIDIA CUDA Deep Neural Network library (CuDNN) that you want only the deterministic implementations. NOTE: Setting this parameter to True can significantly slow down the training process.
| Default | False |
|---|
specifies the ID of one GPU device that is to be used.
| Alias | devices |
|---|---|
| Default | 0 |
| Minimum value | 0 |
| Requirement | The specified values must be unique. |
when set to True, specifies that a GPU device is to be used.
| Alias | useGPU |
|---|---|
| Default | True |
specifies the image size (width of square images in pixels) in the training data.
| Range | 4–1024 |
|---|
when set to True, specifies that large networks are to be used.
| Default | True |
|---|
specifies the number of images to generate.
| Alias | nSamples |
|---|---|
| Default | 1 |
| Minimum value | 1 |
specifies the number of latents to use in calculating the truncation mean.
| Alias | ntl |
|---|---|
| Default | 4096 |
| Minimum value (exclusive) | 0 |
specifies the optimization method to use. Currently only ADAM is supported.
| Alias | optimization |
|---|
The value that you specify for the method parameter determines the other parameters that apply.
lists the names of results tables to save as CAS tables on the server.
For more information about specifying the outputTables parameter, see the common outputTables parameter (Appendix A: Common Parameters).
specifies the frequency at which to print losses to the log and the iteration history table.
| Default | 1 |
|---|---|
| Minimum value | 0 |
specifies the frequency at which to regularize the discriminator.
| Default | 16 |
|---|---|
| Minimum value | 1 |
specifies the frequency at which to regularize the generator.
| Default | 4 |
|---|---|
| Minimum value | 1 |
specifies the weight for regularizing the path length loss.
| Default | 2 |
|---|---|
| Minimum value (exclusive) | 0 |
specifies the weight for regularizing the R1 loss.
| Default | 10 |
|---|---|
| Minimum value (exclusive) | 0 |
specifies the table in which to save the model state for model scoring.
| Long form | saveState={"name":"table-name"} |
|---|---|
| Shortcut form | saveState="table-name" |
The casouttable value can be one or more of the following:
specifies the name of the caslib for the output table.
specifies the descriptive label to associate with the table.
specifies the number of seconds to keep the table in memory after it is last accessed. The table is dropped if it is not accessed for the specified number of seconds.
| Default | 0 |
|---|---|
| Minimum value | 0 |
specifies the memory format for the output table.
| Default | INHERIT |
|---|
use the duplicate value reduction memory format. This memory format can reduce the memory consumption and file size when the input data contains duplicate values.
specifies the name for the output table.
when set to True, adds the output table with a global scope. This enables other sessions to access the table, subject to access controls. The target caslib must also have a global scope.
| Default | False |
|---|
when set to True, overwrites an existing table that has the same name.
| Default | False |
|---|
specifies the seed to use for the random number generator for scoring. A value of 0 means use the time from the computer's clock as the seed.
| Default | 0 |
|---|---|
| Minimum value | 0 |
specifies the seed to use for the random number generator for training. A value of 0 means use the time from the computer's clock as the seed.
| Default | 0 |
|---|---|
| Minimum value | 0 |
specifies the input CAS table of training images.
| Long form | table={"name":"table-name"} |
|---|---|
| Shortcut form | table="table-name" |
The castable value can be one or more of the following:
specifies the caslib for the input table that you want to use with the action. By default, the active caslib is used. Specify a value only if you need to access a table from a different caslib.
when set to True, creates the computed variables when the table is loaded instead of when the action begins.
| Alias | compOnDemand |
|---|---|
| Default | False |
specifies the names of the computed variables to create. Specify an expression for each variable in the computedVarsProgram parameter. If you do not specify this parameter, then all variables from computedVarsProgram are automatically included.
| Alias | compVars |
|---|
The casinvardesc value can be one or more of the following:
specifies the format to apply to the variable.
specifies the length of the format field plus the length of the format precision.
specifies the descriptive label for the variable.
specifies the name for the variable.
specifies the length of the format precision.
specifies the length of the format field.
specifies an expression for each computed variable that you include in the computedVars parameter.
| Alias | compPgm |
|---|
specifies data source options.
| Aliases | options |
|---|---|
| dataSource |
specifies the settings for reading a table from a data source.
| Alias | import_ |
|---|
For more information about specifying the importOptions parameter, see the common importOptions parameter (Appendix A: Common Parameters).
specifies the name of the input table.
specifies the variables to use for ordering observations within partitions. This parameter applies to partitioned tables, or it can be combined with variables that are specified in the groupBy parameter when the value of the groupByMode parameter is set to REDISTRIBUTE.
The casinvardesc value can be one or more of the following:
specifies the format to apply to the variable.
specifies the length of the format field plus the length of the format precision.
specifies the descriptive label for the variable.
specifies the name for the variable.
specifies the length of the format precision.
specifies the length of the format field.
when set to True, does not create a transient table on the server. Setting this parameter to True can be efficient, but the data might not have stable ordering upon repeated runs.
| Default | False |
|---|
specifies the variables to use in the action.
The casinvardesc value can be one or more of the following:
specifies the format to apply to the variable.
specifies the length of the format field plus the length of the format precision.
specifies the descriptive label for the variable.
specifies the name for the variable.
specifies the length of the format precision.
specifies the length of the format field.
specifies an expression for subsetting the input data.
specifies an input table that contains rows to use as a WHERE filter. If the vars parameter is not specified, then all the variable names that are common to the input table and the filtering table are used to find matching rows. If the where parameter for the input table and this parameter are specified, then this filtering table is applied first.
The groupbytable value can be one or more of the following:
specifies the caslib for the filter table. By default, the active caslib is used.
specifies data source options.
| Aliases | options |
|---|---|
| dataSource |
For more information about specifying the dataSourceOptions parameter, see the common dataSourceOptions parameter (Appendix A: Common Parameters).
specifies the settings for reading a table from a data source.
| Alias | import_ |
|---|
For more information about specifying the importOptions parameter, see the common importOptions parameter (Appendix A: Common Parameters).
specifies the name of the filter table.
specifies the variable names to use from the filter table.
The casinvardesc value can be one or more of the following:
specifies the format to apply to the variable.
specifies the length of the format field plus the length of the format precision.
specifies the descriptive label for the variable.
specifies the name for the variable.
specifies the length of the format precision.
specifies the length of the format field.
specifies an expression for subsetting the data from the filter table.
specifies the truncation ratio for image generation.
| Default | 1 |
|---|---|
| Range | (0–1] |
specifies the exponential decay rate for the first-moment estimates.
| Default | 0 |
|---|---|
| Range | [0–1) |
specifies the exponential decay rate for the second-moment estimates.
| Default | 0.99 |
|---|---|
| Range | [0–1) |
specifies the learning rate for the optimizer.
| Default | 0.002 |
|---|---|
| Range | (0–1] |
specifies the number of images in one minibatch.
| Default | 32 |
|---|---|
| Minimum value | 4 |
specifies the number of minibatches to be used for training the model.
| Default | 10 |
|---|---|
| Minimum value | 0 |
Trains a styleGAN model.
If a row includes a subparameter, you can specify the name, caslib, and so on in the subparameter. Otherwise, you can specify the name, caslib, and so on in the parameter.
|
Parameter |
Subparameter |
Description |
|---|---|---|
|
— |
specifies the checkpoint CAS table to be loaded to restore your model. If you omit this parameter, the model is trained from scratch. |
|
|
required parametertable |
— |
specifies the input CAS table of training images. |
|
Parameter |
Subparameter |
Description |
|---|---|---|
|
required parametercasOut |
— |
specifies the output CAS table in which to store the generated images from the trained model. |
|
— |
specifies the CAS table in which to save the current model as a checkpoint. |
|
|
names |
lists the names of results tables to save as CAS tables on the server. |
|
|
required parametersaveState |
— |
specifies the table in which to save the model state for model scoring. |
specifies the output CAS table in which to store the generated images from the trained model.
For more information about specifying the casOut parameter, see the common casouttable (Form 1) parameter (Appendix A: Common Parameters).
specifies the frequency at which to save a checkpoint during the training.
| Default | 5000 |
|---|---|
| Minimum value | 1 |
specifies the checkpoint CAS table to be loaded to restore your model. If you omit this parameter, the model is trained from scratch.
For more information about specifying the checkpointIn parameter, see the common castable (Form 1) parameter (Appendix A: Common Parameters).
specifies the CAS table in which to save the current model as a checkpoint.
For more information about specifying the checkpointOut parameter, see the common casouttable (Form 1) parameter (Appendix A: Common Parameters).
specifies a list of results tables to send to the client for display.
For more information about specifying the display parameter, see the common displayTables parameter (Appendix A: Common Parameters).
specifies GPU-related options.
The gpuOptions value can be one or more of the following:
when set to True, tells the NVIDIA CUDA Deep Neural Network library (CuDNN) that you want only the deterministic implementations. NOTE: Setting this parameter to True can significantly slow down the training process.
| Default | FALSE |
|---|
specifies the ID of one GPU device that is to be used.
| Alias | devices |
|---|---|
| Default | 0 |
| Minimum value | 0 |
| Requirement | The specified values must be unique. |
when set to True, specifies that a GPU device is to be used.
| Alias | useGPU |
|---|---|
| Default | TRUE |
specifies the image size (width of square images in pixels) in the training data.
| Range | 4–1024 |
|---|
when set to True, specifies that large networks are to be used.
| Default | TRUE |
|---|
specifies the number of images to generate.
| Alias | nSamples |
|---|---|
| Default | 1 |
| Minimum value | 1 |
specifies the number of latents to use in calculating the truncation mean.
| Alias | ntl |
|---|---|
| Default | 4096 |
| Minimum value (exclusive) | 0 |
specifies the optimization method to use. Currently only ADAM is supported.
| Alias | optimization |
|---|
The value that you specify for the method parameter determines the other parameters that apply.
lists the names of results tables to save as CAS tables on the server.
For more information about specifying the outputTables parameter, see the common outputTables parameter (Appendix A: Common Parameters).
specifies the frequency at which to print losses to the log and the iteration history table.
| Default | 1 |
|---|---|
| Minimum value | 0 |
specifies the frequency at which to regularize the discriminator.
| Default | 16 |
|---|---|
| Minimum value | 1 |
specifies the frequency at which to regularize the generator.
| Default | 4 |
|---|---|
| Minimum value | 1 |
specifies the weight for regularizing the path length loss.
| Default | 2 |
|---|---|
| Minimum value (exclusive) | 0 |
specifies the weight for regularizing the R1 loss.
| Default | 10 |
|---|---|
| Minimum value (exclusive) | 0 |
specifies the table in which to save the model state for model scoring.
| Long form | saveState=list(name="table-name") |
|---|---|
| Shortcut form | saveState="table-name" |
The casouttable value can be one or more of the following:
specifies the name of the caslib for the output table.
specifies the descriptive label to associate with the table.
specifies the number of seconds to keep the table in memory after it is last accessed. The table is dropped if it is not accessed for the specified number of seconds.
| Default | 0 |
|---|---|
| Minimum value | 0 |
specifies the memory format for the output table.
| Default | INHERIT |
|---|
use the duplicate value reduction memory format. This memory format can reduce the memory consumption and file size when the input data contains duplicate values.
specifies the name for the output table.
when set to True, adds the output table with a global scope. This enables other sessions to access the table, subject to access controls. The target caslib must also have a global scope.
| Default | FALSE |
|---|
when set to True, overwrites an existing table that has the same name.
| Default | FALSE |
|---|
specifies the seed to use for the random number generator for scoring. A value of 0 means use the time from the computer's clock as the seed.
| Default | 0 |
|---|---|
| Minimum value | 0 |
specifies the seed to use for the random number generator for training. A value of 0 means use the time from the computer's clock as the seed.
| Default | 0 |
|---|---|
| Minimum value | 0 |
specifies the input CAS table of training images.
| Long form | table=list(name="table-name") |
|---|---|
| Shortcut form | table="table-name" |
The castable value can be one or more of the following:
specifies the caslib for the input table that you want to use with the action. By default, the active caslib is used. Specify a value only if you need to access a table from a different caslib.
when set to True, creates the computed variables when the table is loaded instead of when the action begins.
| Alias | compOnDemand |
|---|---|
| Default | FALSE |
specifies the names of the computed variables to create. Specify an expression for each variable in the computedVarsProgram parameter. If you do not specify this parameter, then all variables from computedVarsProgram are automatically included.
| Alias | compVars |
|---|
The casinvardesc value can be one or more of the following:
specifies the format to apply to the variable.
specifies the length of the format field plus the length of the format precision.
specifies the descriptive label for the variable.
specifies the name for the variable.
specifies the length of the format precision.
specifies the length of the format field.
specifies an expression for each computed variable that you include in the computedVars parameter.
| Alias | compPgm |
|---|
specifies data source options.
| Aliases | options |
|---|---|
| dataSource |
specifies the settings for reading a table from a data source.
| Alias | import |
|---|
For more information about specifying the importOptions parameter, see the common importOptions parameter (Appendix A: Common Parameters).
specifies the name of the input table.
specifies the variables to use for ordering observations within partitions. This parameter applies to partitioned tables, or it can be combined with variables that are specified in the groupBy parameter when the value of the groupByMode parameter is set to REDISTRIBUTE.
The casinvardesc value can be one or more of the following:
specifies the format to apply to the variable.
specifies the length of the format field plus the length of the format precision.
specifies the descriptive label for the variable.
specifies the name for the variable.
specifies the length of the format precision.
specifies the length of the format field.
when set to True, does not create a transient table on the server. Setting this parameter to True can be efficient, but the data might not have stable ordering upon repeated runs.
| Default | FALSE |
|---|
specifies the variables to use in the action.
The casinvardesc value can be one or more of the following:
specifies the format to apply to the variable.
specifies the length of the format field plus the length of the format precision.
specifies the descriptive label for the variable.
specifies the name for the variable.
specifies the length of the format precision.
specifies the length of the format field.
specifies an expression for subsetting the input data.
specifies an input table that contains rows to use as a WHERE filter. If the vars parameter is not specified, then all the variable names that are common to the input table and the filtering table are used to find matching rows. If the where parameter for the input table and this parameter are specified, then this filtering table is applied first.
The groupbytable value can be one or more of the following:
specifies the caslib for the filter table. By default, the active caslib is used.
specifies data source options.
| Aliases | options |
|---|---|
| dataSource |
For more information about specifying the dataSourceOptions parameter, see the common dataSourceOptions parameter (Appendix A: Common Parameters).
specifies the settings for reading a table from a data source.
| Alias | import |
|---|
For more information about specifying the importOptions parameter, see the common importOptions parameter (Appendix A: Common Parameters).
specifies the name of the filter table.
specifies the variable names to use from the filter table.
The casinvardesc value can be one or more of the following:
specifies the format to apply to the variable.
specifies the length of the format field plus the length of the format precision.
specifies the descriptive label for the variable.
specifies the name for the variable.
specifies the length of the format precision.
specifies the length of the format field.
specifies an expression for subsetting the data from the filter table.
specifies the truncation ratio for image generation.
| Default | 1 |
|---|---|
| Range | (0–1] |
specifies the exponential decay rate for the first-moment estimates.
| Default | 0 |
|---|---|
| Range | [0–1) |
specifies the exponential decay rate for the second-moment estimates.
| Default | 0.99 |
|---|---|
| Range | [0–1) |
specifies the learning rate for the optimizer.
| Default | 0.002 |
|---|---|
| Range | (0–1] |
specifies the number of images in one minibatch.
| Default | 32 |
|---|---|
| Minimum value | 4 |
specifies the number of minibatches to be used for training the model.
| Default | 10 |
|---|---|
| Minimum value | 0 |