TFSlim models including support for Inception, ResNet, VGG, MobileNet, NASNet, and PNASNet.
Project description
Models
######
images
######
*Generic images dataset*
Operations
==========
prepare
^^^^^^^
*Prepare images for training*
Flags

**images**
*Directory containing images to prepare (required)*
**randomseed**
*Seed used for train/validation split (randomly generated)*
**valsplit**
*Percentage of images reserved for validation (30)*
inception
#########
*TFSlim Inception v1 classifier*
Operations
==========
evaluate
^^^^^^^^
*Evaluate a trained model*
Flags

**batchsize**
*Number of examples in each evaluated batch (100)*
**evalbatches**
*Number of batches to evaluate (all available)*
**step**
*Checkpoint step to evaluate (latest checkpoint)*
exportandfreeze
^^^^^^^^^^^^^^^^^
*Export an inference graph with checkpoint weights*
Flags

**step**
*Checkpoint step to use for the frozen graph (latest checkpoint)*
finetune
^^^^^^^^
*Finetune a trained model*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.0001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
label
^^^^^
*Classify an image using a trained model*
Flags

**image**
*Path to image to classify (required)*
tflite
^^^^^^
*Generate a TFLite file from a frozen graph*
train
^^^^^
*Train model from scratch*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
transferlearn
^^^^^^^^^^^^^^
*Train model using transfer learning*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
inceptionresnetv2
###################
*TFSlim Inception ResNet v2 classifier*
Operations
==========
evaluate
^^^^^^^^
*Evaluate a trained model*
Flags

**batchsize**
*Number of examples in each evaluated batch (100)*
**evalbatches**
*Number of batches to evaluate (all available)*
**step**
*Checkpoint step to evaluate (latest checkpoint)*
exportandfreeze
^^^^^^^^^^^^^^^^^
*Export an inference graph with checkpoint weights*
Flags

**step**
*Checkpoint step to use for the frozen graph (latest checkpoint)*
finetune
^^^^^^^^
*Finetune a trained model*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.0001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
label
^^^^^
*Classify an image using a trained model*
Flags

**image**
*Path to image to classify (required)*
tflite
^^^^^^
*Generate a TFLite file from a frozen graph*
train
^^^^^
*Train model from scratch*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
transferlearn
^^^^^^^^^^^^^^
*Train model using transfer learning*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
inceptionv2
############
*TFSlim Inception v2 classifier*
Operations
==========
evaluate
^^^^^^^^
*Evaluate a trained model*
Flags

**batchsize**
*Number of examples in each evaluated batch (100)*
**evalbatches**
*Number of batches to evaluate (all available)*
**step**
*Checkpoint step to evaluate (latest checkpoint)*
exportandfreeze
^^^^^^^^^^^^^^^^^
*Export an inference graph with checkpoint weights*
Flags

**step**
*Checkpoint step to use for the frozen graph (latest checkpoint)*
finetune
^^^^^^^^
*Finetune a trained model*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.0001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
label
^^^^^
*Classify an image using a trained model*
Flags

**image**
*Path to image to classify (required)*
tflite
^^^^^^
*Generate a TFLite file from a frozen graph*
train
^^^^^
*Train model from scratch*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
transferlearn
^^^^^^^^^^^^^^
*Train model using transfer learning*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
inceptionv3
############
*TFSlim Inception v3 classifier*
Operations
==========
evaluate
^^^^^^^^
*Evaluate a trained model*
Flags

**batchsize**
*Number of examples in each evaluated batch (100)*
**evalbatches**
*Number of batches to evaluate (all available)*
**step**
*Checkpoint step to evaluate (latest checkpoint)*
exportandfreeze
^^^^^^^^^^^^^^^^^
*Export an inference graph with checkpoint weights*
Flags

**step**
*Checkpoint step to use for the frozen graph (latest checkpoint)*
finetune
^^^^^^^^
*Finetune a trained model*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.0001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
label
^^^^^
*Classify an image using a trained model*
Flags

**image**
*Path to image to classify (required)*
tflite
^^^^^^
*Generate a TFLite file from a frozen graph*
train
^^^^^
*Train model from scratch*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
transferlearn
^^^^^^^^^^^^^^
*Train model using transfer learning*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
inceptionv4
############
*TFSlim Inception v4 classifier*
Operations
==========
evaluate
^^^^^^^^
*Evaluate a trained model*
Flags

**batchsize**
*Number of examples in each evaluated batch (100)*
**evalbatches**
*Number of batches to evaluate (all available)*
**step**
*Checkpoint step to evaluate (latest checkpoint)*
exportandfreeze
^^^^^^^^^^^^^^^^^
*Export an inference graph with checkpoint weights*
Flags

**step**
*Checkpoint step to use for the frozen graph (latest checkpoint)*
finetune
^^^^^^^^
*Finetune a trained model*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.0001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
label
^^^^^
*Classify an image using a trained model*
Flags

**image**
*Path to image to classify (required)*
tflite
^^^^^^
*Generate a TFLite file from a frozen graph*
train
^^^^^
*Train model from scratch*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
transferlearn
^^^^^^^^^^^^^^
*Train model using transfer learning*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
mobilenet
#########
*TFSlim Mobilenet v1 classifier*
Operations
==========
evaluate
^^^^^^^^
*Evaluate a trained model*
Flags

**batchsize**
*Number of examples in each evaluated batch (100)*
**evalbatches**
*Number of batches to evaluate (all available)*
**step**
*Checkpoint step to evaluate (latest checkpoint)*
exportandfreeze
^^^^^^^^^^^^^^^^^
*Export an inference graph with checkpoint weights*
Flags

**step**
*Checkpoint step to use for the frozen graph (latest checkpoint)*
finetune
^^^^^^^^
*Finetune a trained model*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.0001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
label
^^^^^
*Classify an image using a trained model*
Flags

**image**
*Path to image to classify (required)*
tflite
^^^^^^
*Generate a TFLite file from a frozen graph*
train
^^^^^
*Train model from scratch*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
transferlearn
^^^^^^^^^^^^^^
*Train model using transfer learning*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
mobilenetv21.4
################
*TFSlim Mobilenet v2 classifier*
Operations
==========
evaluate
^^^^^^^^
*Evaluate a trained model*
Flags

**batchsize**
*Number of examples in each evaluated batch (100)*
**evalbatches**
*Number of batches to evaluate (all available)*
**step**
*Checkpoint step to evaluate (latest checkpoint)*
exportandfreeze
^^^^^^^^^^^^^^^^^
*Export an inference graph with checkpoint weights*
Flags

**step**
*Checkpoint step to use for the frozen graph (latest checkpoint)*
finetune
^^^^^^^^
*Finetune a trained model*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.0001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
label
^^^^^
*Classify an image using a trained model*
Flags

**image**
*Path to image to classify (required)*
tflite
^^^^^^
*Generate a TFLite file from a frozen graph*
train
^^^^^
*Train model from scratch*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
transferlearn
^^^^^^^^^^^^^^
*Train model using transfer learning*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
nasnetlarge
############
*TFSlim NASNet large classifier*
Operations
==========
evaluate
^^^^^^^^
*Evaluate a trained model*
Flags

**batchsize**
*Number of examples in each evaluated batch (100)*
**evalbatches**
*Number of batches to evaluate (all available)*
**step**
*Checkpoint step to evaluate (latest checkpoint)*
exportandfreeze
^^^^^^^^^^^^^^^^^
*Export an inference graph with checkpoint weights*
Flags

**step**
*Checkpoint step to use for the frozen graph (latest checkpoint)*
finetune
^^^^^^^^
*Finetune a trained model*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.0001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
label
^^^^^
*Classify an image using a trained model*
Flags

**image**
*Path to image to classify (required)*
tflite
^^^^^^
*Generate a TFLite file from a frozen graph*
train
^^^^^
*Train model from scratch*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
transferlearn
^^^^^^^^^^^^^^
*Train model using transfer learning*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
nasnetmobile
#############
*TFSlim NASNet mobile classifier*
Operations
==========
evaluate
^^^^^^^^
*Evaluate a trained model*
Flags

**batchsize**
*Number of examples in each evaluated batch (100)*
**evalbatches**
*Number of batches to evaluate (all available)*
**step**
*Checkpoint step to evaluate (latest checkpoint)*
exportandfreeze
^^^^^^^^^^^^^^^^^
*Export an inference graph with checkpoint weights*
Flags

**step**
*Checkpoint step to use for the frozen graph (latest checkpoint)*
finetune
^^^^^^^^
*Finetune a trained model*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.0001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
label
^^^^^
*Classify an image using a trained model*
Flags

**image**
*Path to image to classify (required)*
tflite
^^^^^^
*Generate a TFLite file from a frozen graph*
train
^^^^^
*Train model from scratch*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
transferlearn
^^^^^^^^^^^^^^
*Train model using transfer learning*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
pnasnetlarge
#############
*TFSlim PNASNet classifier*
Operations
==========
evaluate
^^^^^^^^
*Evaluate a trained model*
Flags

**batchsize**
*Number of examples in each evaluated batch (100)*
**evalbatches**
*Number of batches to evaluate (all available)*
**step**
*Checkpoint step to evaluate (latest checkpoint)*
exportandfreeze
^^^^^^^^^^^^^^^^^
*Export an inference graph with checkpoint weights*
Flags

**step**
*Checkpoint step to use for the frozen graph (latest checkpoint)*
finetune
^^^^^^^^
*Finetune a trained model*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.0001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
label
^^^^^
*Classify an image using a trained model*
Flags

**image**
*Path to image to classify (required)*
tflite
^^^^^^
*Generate a TFLite file from a frozen graph*
train
^^^^^
*Train model from scratch*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
transferlearn
^^^^^^^^^^^^^^
*Train model using transfer learning*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
pnasnetmobile
##############
*TFSlim PNASNet mobile classifier*
Operations
==========
evaluate
^^^^^^^^
*Evaluate a trained model*
Flags

**batchsize**
*Number of examples in each evaluated batch (100)*
**evalbatches**
*Number of batches to evaluate (all available)*
**step**
*Checkpoint step to evaluate (latest checkpoint)*
exportandfreeze
^^^^^^^^^^^^^^^^^
*Export an inference graph with checkpoint weights*
Flags

**step**
*Checkpoint step to use for the frozen graph (latest checkpoint)*
finetune
^^^^^^^^
*Finetune a trained model*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.0001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
label
^^^^^
*Classify an image using a trained model*
Flags

**image**
*Path to image to classify (required)*
tflite
^^^^^^
*Generate a TFLite file from a frozen graph*
train
^^^^^
*Train model from scratch*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
transferlearn
^^^^^^^^^^^^^^
*Train model using transfer learning*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
resnet101
##########
*TFSlim ResNet v1 101 layer classifier*
Operations
==========
evaluate
^^^^^^^^
*Evaluate a trained model*
Flags

**batchsize**
*Number of examples in each evaluated batch (100)*
**evalbatches**
*Number of batches to evaluate (all available)*
**step**
*Checkpoint step to evaluate (latest checkpoint)*
exportandfreeze
^^^^^^^^^^^^^^^^^
*Export an inference graph with checkpoint weights*
Flags

**step**
*Checkpoint step to use for the frozen graph (latest checkpoint)*
finetune
^^^^^^^^
*Finetune a trained model*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.0001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
label
^^^^^
*Classify an image using a trained model*
Flags

**image**
*Path to image to classify (required)*
tflite
^^^^^^
*Generate a TFLite file from a frozen graph*
train
^^^^^
*Train model from scratch*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
transferlearn
^^^^^^^^^^^^^^
*Train model using transfer learning*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
resnet152
##########
*TFSlim ResNet v1 152 layer classifier*
Operations
==========
evaluate
^^^^^^^^
*Evaluate a trained model*
Flags

**batchsize**
*Number of examples in each evaluated batch (100)*
**evalbatches**
*Number of batches to evaluate (all available)*
**step**
*Checkpoint step to evaluate (latest checkpoint)*
exportandfreeze
^^^^^^^^^^^^^^^^^
*Export an inference graph with checkpoint weights*
Flags

**step**
*Checkpoint step to use for the frozen graph (latest checkpoint)*
finetune
^^^^^^^^
*Finetune a trained model*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.0001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
label
^^^^^
*Classify an image using a trained model*
Flags

**image**
*Path to image to classify (required)*
tflite
^^^^^^
*Generate a TFLite file from a frozen graph*
train
^^^^^
*Train model from scratch*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
transferlearn
^^^^^^^^^^^^^^
*Train model using transfer learning*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
resnet50
#########
*TFSlim ResNet v1 50 layer classifier*
Operations
==========
evaluate
^^^^^^^^
*Evaluate a trained model*
Flags

**batchsize**
*Number of examples in each evaluated batch (100)*
**evalbatches**
*Number of batches to evaluate (all available)*
**step**
*Checkpoint step to evaluate (latest checkpoint)*
exportandfreeze
^^^^^^^^^^^^^^^^^
*Export an inference graph with checkpoint weights*
Flags

**step**
*Checkpoint step to use for the frozen graph (latest checkpoint)*
finetune
^^^^^^^^
*Finetune a trained model*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.0001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
label
^^^^^
*Classify an image using a trained model*
Flags

**image**
*Path to image to classify (required)*
tflite
^^^^^^
*Generate a TFLite file from a frozen graph*
train
^^^^^
*Train model from scratch*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
transferlearn
^^^^^^^^^^^^^^
*Train model using transfer learning*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
resnetv2101
#############
*TFSlim ResNet v2 101 layer classifier*
Operations
==========
evaluate
^^^^^^^^
*Evaluate a trained model*
Flags

**batchsize**
*Number of examples in each evaluated batch (100)*
**evalbatches**
*Number of batches to evaluate (all available)*
**step**
*Checkpoint step to evaluate (latest checkpoint)*
exportandfreeze
^^^^^^^^^^^^^^^^^
*Export an inference graph with checkpoint weights*
Flags

**step**
*Checkpoint step to use for the frozen graph (latest checkpoint)*
finetune
^^^^^^^^
*Finetune a trained model*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.0001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
label
^^^^^
*Classify an image using a trained model*
Flags

**image**
*Path to image to classify (required)*
tflite
^^^^^^
*Generate a TFLite file from a frozen graph*
train
^^^^^
*Train model from scratch*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
transferlearn
^^^^^^^^^^^^^^
*Train model using transfer learning*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
resnetv2152
#############
*TFSlim ResNet v2 152 layer classifier*
Operations
==========
evaluate
^^^^^^^^
*Evaluate a trained model*
Flags

**batchsize**
*Number of examples in each evaluated batch (100)*
**evalbatches**
*Number of batches to evaluate (all available)*
**step**
*Checkpoint step to evaluate (latest checkpoint)*
exportandfreeze
^^^^^^^^^^^^^^^^^
*Export an inference graph with checkpoint weights*
Flags

**step**
*Checkpoint step to use for the frozen graph (latest checkpoint)*
finetune
^^^^^^^^
*Finetune a trained model*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.0001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
label
^^^^^
*Classify an image using a trained model*
Flags

**image**
*Path to image to classify (required)*
tflite
^^^^^^
*Generate a TFLite file from a frozen graph*
train
^^^^^
*Train model from scratch*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
transferlearn
^^^^^^^^^^^^^^
*Train model using transfer learning*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
resnetv250
############
*TFSlim ResNet v2 50 layer classifier*
Operations
==========
evaluate
^^^^^^^^
*Evaluate a trained model*
Flags

**batchsize**
*Number of examples in each evaluated batch (100)*
**evalbatches**
*Number of batches to evaluate (all available)*
**step**
*Checkpoint step to evaluate (latest checkpoint)*
exportandfreeze
^^^^^^^^^^^^^^^^^
*Export an inference graph with checkpoint weights*
Flags

**step**
*Checkpoint step to use for the frozen graph (latest checkpoint)*
finetune
^^^^^^^^
*Finetune a trained model*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.0001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
label
^^^^^
*Classify an image using a trained model*
Flags

**image**
*Path to image to classify (required)*
tflite
^^^^^^
*Generate a TFLite file from a frozen graph*
train
^^^^^
*Train model from scratch*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
transferlearn
^^^^^^^^^^^^^^
*Train model using transfer learning*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
vgg16
######
*TFSlim VGG 16 classifier*
Operations
==========
evaluate
^^^^^^^^
*Evaluate a trained model*
Flags

**batchsize**
*Number of examples in each evaluated batch (100)*
**evalbatches**
*Number of batches to evaluate (all available)*
**step**
*Checkpoint step to evaluate (latest checkpoint)*
exportandfreeze
^^^^^^^^^^^^^^^^^
*Export an inference graph with checkpoint weights*
Flags

**step**
*Checkpoint step to use for the frozen graph (latest checkpoint)*
finetune
^^^^^^^^
*Finetune a trained model*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.0001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
label
^^^^^
*Classify an image using a trained model*
Flags

**image**
*Path to image to classify (required)*
tflite
^^^^^^
*Generate a TFLite file from a frozen graph*
train
^^^^^
*Train model from scratch*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
transferlearn
^^^^^^^^^^^^^^
*Train model using transfer learning*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
vgg19
######
*TFSlim VGG 19 classifier*
Operations
==========
evaluate
^^^^^^^^
*Evaluate a trained model*
Flags

**batchsize**
*Number of examples in each evaluated batch (100)*
**evalbatches**
*Number of batches to evaluate (all available)*
**step**
*Checkpoint step to evaluate (latest checkpoint)*
exportandfreeze
^^^^^^^^^^^^^^^^^
*Export an inference graph with checkpoint weights*
Flags

**step**
*Checkpoint step to use for the frozen graph (latest checkpoint)*
finetune
^^^^^^^^
*Finetune a trained model*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.0001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
label
^^^^^
*Classify an image using a trained model*
Flags

**image**
*Path to image to classify (required)*
tflite
^^^^^^
*Generate a TFLite file from a frozen graph*
train
^^^^^
*Train model from scratch*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
transferlearn
^^^^^^^^^^^^^^
*Train model using transfer learning*
Flags

**autoscale**
*Adjust applicable flags for multiGPU systems (yes)
Set to 'no' to disable any flag value adjustments.
When this value is 'yes' (the default) the following flags are adjusted on
multiGPU systems:
 clones
 learningrate
`clones` is set to the number of available GPUs.
`learningrate` is adjusted by multiplying its specified value by the
number of GPUs.
Flags are not adjusted on single GPU or CPU only systems.
*
**batchsize**
*Number of examples in each training batch (32)*
**clones**
*Number of model clones (calculated)
This value is automatically set to the number of available GPUs if `auto
scale` is 'yes'.
When `autoscale` is 'no' this value can be increased from 1 to train the
model in parallel on multiple GPUs.
*
**learningrate**
*Initial learning rate (0.001)*
**learningratedecayepochs**
*Number of epochs after which learning rate decays (2.0)*
**learningratedecayfactor**
*Learning rate decay factor (0.94)*
**learningratedecaytype**
*Method used to decay the learning rate (exponential)
Choices:
exponential
fixed
polynomial
*
**learningrateend**
*Minimal learning rate used by polynomial learning rate decay (0.0001)*
**logsaveseconds**
*Frequency of log summary saves in seconds (60)*
**logsteps**
*Frequency of summary logs in steps (100)*
**modelsaveseconds**
*Frequency of model saves (checkpoints) in seconds (600)*
**optimizer**
*Optimizer used to train (rmsprop)
Choices:
adadelta
adagrad
adam
ftrl
momentum
rmsprop
sgd
*
**preprocessing**
*Preprocessing to use (default for model)*
**preprocessors**
*Number of preprocessing threads (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize the
preprocessor thread count for the system.
*
**readers**
*Number of parallel data readers (calculated)
This value is automatically set to logical CPU count / 2 if `autoscale`
is 'yes'.
When `autoscale` is 'no' this value can be set to optimize data reader
performance for the system.
*
**trainsteps**
*Number of steps to train (train indefinitely)*
**weightdecay**
*Decay on the model weights (4e05)*
Project details
Release history Release notifications  RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Built Distribution
Close
Hashes for gpkg.slim.models0.5.0.dev6py2.py3noneany.whl
Algorithm  Hash digest  

SHA256  99fa732cb7229f976a9cb149d6dd662aa4a798c81718a0d61ffe1148b789010f 

MD5  3590a2b4fb50b9fba36b4f0b17f41363 

BLAKE2256  8418e6d6bb2bdcb3fa68d40397b97660ece255b950ecdc5a60b9819acff7dfff 