gpkg.slim.models

gpkg.slim.models provides models for each of the TF-Slim architectures support in the gpkg.slim base package. Install this package to use the models directly.

If you want to create your own models that extend TF-Slim architectures, refer to gpkg.slim for instructions.

Name gpkg.slim.models
Description TF-Slim models (Guild AI)
Version 0.5.1
Source https://github.com/guildai/packages/tree/master/gpkg/slim/models
Author Guild AI

images model

Generic images dataset.

Operations

prepare

Prepare images for training.
Flag Description Default
images Directory containing images to prepare. required
random‑seed Seed used for train/validation split.
val‑split Percentage of images reserved for validation. 30

Resources

inception model

TF-Slim Inception v1 classifier.

Operations

evaluate

Evaluate a trained model.
Flag Description Default
batch‑size Number of examples in each evaluated batch. 100
eval‑batches Number of batches to evaluate.
step Checkpoint step to evaluate.

export-and-freeze

Export an inference graph with checkpoint weights.
Flag Description Default
step Checkpoint step to use for the frozen graph.

finetune

Finetune a trained model.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.0001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

label

Classify an image using a trained model.
Flag Description Default
image Path to image to classify. required

tflite

Generate a TFLite file from a frozen graph.
Flag Description Default
output‑format TF Lite output format. tflite
quantized Whether or not output arrays are quantized. False
quantized‑inputs Whether or not input arrays are quantized. False

train

Train model from scratch.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

transfer-learn

Train model using transfer learning.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

Resources

examples

Dataset generated with images:prepare.

frozen-graph

Frozen inference graph.

label-image-script

labels

models-lib

trained-model

Trained model from train, transfer-learn, or finetune.

transfer-learn-checkpoint

inception-resnet-v2 model

TF-Slim Inception ResNet v2 classifier.

Operations

evaluate

Evaluate a trained model.
Flag Description Default
batch‑size Number of examples in each evaluated batch. 100
eval‑batches Number of batches to evaluate.
step Checkpoint step to evaluate.

export-and-freeze

Export an inference graph with checkpoint weights.
Flag Description Default
step Checkpoint step to use for the frozen graph.

finetune

Finetune a trained model.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.0001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

label

Classify an image using a trained model.
Flag Description Default
image Path to image to classify. required

tflite

Generate a TFLite file from a frozen graph.
Flag Description Default
output‑format TF Lite output format. tflite
quantized Whether or not output arrays are quantized. False
quantized‑inputs Whether or not input arrays are quantized. False

train

Train model from scratch.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

transfer-learn

Train model using transfer learning.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

Resources

examples

Dataset generated with images:prepare.

frozen-graph

Frozen inference graph.

label-image-script

labels

models-lib

trained-model

Trained model from train, transfer-learn, or finetune.

transfer-learn-checkpoint

inception-v2 model

TF-Slim Inception v2 classifier.

Operations

evaluate

Evaluate a trained model.
Flag Description Default
batch‑size Number of examples in each evaluated batch. 100
eval‑batches Number of batches to evaluate.
step Checkpoint step to evaluate.

export-and-freeze

Export an inference graph with checkpoint weights.
Flag Description Default
step Checkpoint step to use for the frozen graph.

finetune

Finetune a trained model.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.0001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

label

Classify an image using a trained model.
Flag Description Default
image Path to image to classify. required

tflite

Generate a TFLite file from a frozen graph.
Flag Description Default
output‑format TF Lite output format. tflite
quantized Whether or not output arrays are quantized. False
quantized‑inputs Whether or not input arrays are quantized. False

train

Train model from scratch.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

transfer-learn

Train model using transfer learning.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

Resources

examples

Dataset generated with images:prepare.

frozen-graph

Frozen inference graph.

label-image-script

labels

models-lib

trained-model

Trained model from train, transfer-learn, or finetune.

transfer-learn-checkpoint

inception-v3 model

TF-Slim Inception v3 classifier.

Operations

evaluate

Evaluate a trained model.
Flag Description Default
batch‑size Number of examples in each evaluated batch. 100
eval‑batches Number of batches to evaluate.
step Checkpoint step to evaluate.

export-and-freeze

Export an inference graph with checkpoint weights.
Flag Description Default
step Checkpoint step to use for the frozen graph.

finetune

Finetune a trained model.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.0001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

label

Classify an image using a trained model.
Flag Description Default
image Path to image to classify. required

tflite

Generate a TFLite file from a frozen graph.
Flag Description Default
output‑format TF Lite output format. tflite
quantized Whether or not output arrays are quantized. False
quantized‑inputs Whether or not input arrays are quantized. False

train

Train model from scratch.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

transfer-learn

Train model using transfer learning.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

Resources

examples

Dataset generated with images:prepare.

frozen-graph

Frozen inference graph.

label-image-script

labels

models-lib

trained-model

Trained model from train, transfer-learn, or finetune.

transfer-learn-checkpoint

inception-v4 model

TF-Slim Inception v4 classifier.

Operations

evaluate

Evaluate a trained model.
Flag Description Default
batch‑size Number of examples in each evaluated batch. 100
eval‑batches Number of batches to evaluate.
step Checkpoint step to evaluate.

export-and-freeze

Export an inference graph with checkpoint weights.
Flag Description Default
step Checkpoint step to use for the frozen graph.

finetune

Finetune a trained model.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.0001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

label

Classify an image using a trained model.
Flag Description Default
image Path to image to classify. required

tflite

Generate a TFLite file from a frozen graph.
Flag Description Default
output‑format TF Lite output format. tflite
quantized Whether or not output arrays are quantized. False
quantized‑inputs Whether or not input arrays are quantized. False

train

Train model from scratch.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

transfer-learn

Train model using transfer learning.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

Resources

examples

Dataset generated with images:prepare.

frozen-graph

Frozen inference graph.

label-image-script

labels

models-lib

trained-model

Trained model from train, transfer-learn, or finetune.

transfer-learn-checkpoint

mobilenet model

TF-Slim Mobilenet v1 classifier.

Operations

evaluate

Evaluate a trained model.
Flag Description Default
batch‑size Number of examples in each evaluated batch. 100
eval‑batches Number of batches to evaluate.
step Checkpoint step to evaluate.

export-and-freeze

Export an inference graph with checkpoint weights.
Flag Description Default
step Checkpoint step to use for the frozen graph.

finetune

Finetune a trained model.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.0001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

label

Classify an image using a trained model.
Flag Description Default
image Path to image to classify. required

tflite

Generate a TFLite file from a frozen graph.
Flag Description Default
output‑format TF Lite output format. tflite
quantized Whether or not output arrays are quantized. False
quantized‑inputs Whether or not input arrays are quantized. False

train

Train model from scratch.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

transfer-learn

Train model using transfer learning.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

Resources

examples

Dataset generated with images:prepare.

frozen-graph

Frozen inference graph.

label-image-script

labels

models-lib

trained-model

Trained model from train, transfer-learn, or finetune.

transfer-learn-checkpoint

mobilenet-v2-1.4 model

TF-Slim Mobilenet v2 classifier.

Operations

evaluate

Evaluate a trained model.
Flag Description Default
batch‑size Number of examples in each evaluated batch. 100
eval‑batches Number of batches to evaluate.
step Checkpoint step to evaluate.

export-and-freeze

Export an inference graph with checkpoint weights.
Flag Description Default
step Checkpoint step to use for the frozen graph.

finetune

Finetune a trained model.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.0001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

label

Classify an image using a trained model.
Flag Description Default
image Path to image to classify. required

tflite

Generate a TFLite file from a frozen graph.
Flag Description Default
output‑format TF Lite output format. tflite
quantized Whether or not output arrays are quantized. False
quantized‑inputs Whether or not input arrays are quantized. False

train

Train model from scratch.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

transfer-learn

Train model using transfer learning.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

Resources

examples

Dataset generated with images:prepare.

frozen-graph

Frozen inference graph.

label-image-script

labels

models-lib

trained-model

Trained model from train, transfer-learn, or finetune.

transfer-learn-checkpoint

nasnet-large model

TF-Slim NASNet large classifier.

Operations

evaluate

Evaluate a trained model.
Flag Description Default
batch‑size Number of examples in each evaluated batch. 100
eval‑batches Number of batches to evaluate.
step Checkpoint step to evaluate.

export-and-freeze

Export an inference graph with checkpoint weights.
Flag Description Default
step Checkpoint step to use for the frozen graph.

finetune

Finetune a trained model.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.0001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

label

Classify an image using a trained model.
Flag Description Default
image Path to image to classify. required

tflite

Generate a TFLite file from a frozen graph.
Flag Description Default
output‑format TF Lite output format. tflite
quantized Whether or not output arrays are quantized. False
quantized‑inputs Whether or not input arrays are quantized. False

train

Train model from scratch.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

transfer-learn

Train model using transfer learning.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

Resources

examples

Dataset generated with images:prepare.

frozen-graph

Frozen inference graph.

label-image-script

labels

models-lib

trained-model

Trained model from train, transfer-learn, or finetune.

transfer-learn-checkpoint

nasnet-mobile model

TF-Slim NASNet mobile classifier.

Operations

evaluate

Evaluate a trained model.
Flag Description Default
batch‑size Number of examples in each evaluated batch. 100
eval‑batches Number of batches to evaluate.
step Checkpoint step to evaluate.

export-and-freeze

Export an inference graph with checkpoint weights.
Flag Description Default
step Checkpoint step to use for the frozen graph.

finetune

Finetune a trained model.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.0001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

label

Classify an image using a trained model.
Flag Description Default
image Path to image to classify. required

tflite

Generate a TFLite file from a frozen graph.
Flag Description Default
output‑format TF Lite output format. tflite
quantized Whether or not output arrays are quantized. False
quantized‑inputs Whether or not input arrays are quantized. False

train

Train model from scratch.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

transfer-learn

Train model using transfer learning.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

Resources

examples

Dataset generated with images:prepare.

frozen-graph

Frozen inference graph.

label-image-script

labels

models-lib

trained-model

Trained model from train, transfer-learn, or finetune.

transfer-learn-checkpoint

pnasnet-large model

TF-Slim PNASNet classifier.

Operations

evaluate

Evaluate a trained model.
Flag Description Default
batch‑size Number of examples in each evaluated batch. 100
eval‑batches Number of batches to evaluate.
step Checkpoint step to evaluate.

export-and-freeze

Export an inference graph with checkpoint weights.
Flag Description Default
step Checkpoint step to use for the frozen graph.

finetune

Finetune a trained model.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.0001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

label

Classify an image using a trained model.
Flag Description Default
image Path to image to classify. required

tflite

Generate a TFLite file from a frozen graph.
Flag Description Default
output‑format TF Lite output format. tflite
quantized Whether or not output arrays are quantized. False
quantized‑inputs Whether or not input arrays are quantized. False

train

Train model from scratch.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

transfer-learn

Train model using transfer learning.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

Resources

examples

Dataset generated with images:prepare.

frozen-graph

Frozen inference graph.

label-image-script

labels

models-lib

trained-model

Trained model from train, transfer-learn, or finetune.

transfer-learn-checkpoint

pnasnet-mobile model

TF-Slim PNASNet mobile classifier.

Operations

evaluate

Evaluate a trained model.
Flag Description Default
batch‑size Number of examples in each evaluated batch. 100
eval‑batches Number of batches to evaluate.
step Checkpoint step to evaluate.

export-and-freeze

Export an inference graph with checkpoint weights.
Flag Description Default
step Checkpoint step to use for the frozen graph.

finetune

Finetune a trained model.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.0001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

label

Classify an image using a trained model.
Flag Description Default
image Path to image to classify. required

tflite

Generate a TFLite file from a frozen graph.
Flag Description Default
output‑format TF Lite output format. tflite
quantized Whether or not output arrays are quantized. False
quantized‑inputs Whether or not input arrays are quantized. False

train

Train model from scratch.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

transfer-learn

Train model using transfer learning.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

Resources

examples

Dataset generated with images:prepare.

frozen-graph

Frozen inference graph.

label-image-script

labels

models-lib

trained-model

Trained model from train, transfer-learn, or finetune.

transfer-learn-checkpoint

resnet-101 model

TF-Slim ResNet v1 101 layer classifier.

Operations

evaluate

Evaluate a trained model.
Flag Description Default
batch‑size Number of examples in each evaluated batch. 100
eval‑batches Number of batches to evaluate.
step Checkpoint step to evaluate.

export-and-freeze

Export an inference graph with checkpoint weights.
Flag Description Default
step Checkpoint step to use for the frozen graph.

finetune

Finetune a trained model.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.0001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

label

Classify an image using a trained model.
Flag Description Default
image Path to image to classify. required

tflite

Generate a TFLite file from a frozen graph.
Flag Description Default
output‑format TF Lite output format. tflite
quantized Whether or not output arrays are quantized. False
quantized‑inputs Whether or not input arrays are quantized. False

train

Train model from scratch.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

transfer-learn

Train model using transfer learning.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

Resources

examples

Dataset generated with images:prepare.

frozen-graph

Frozen inference graph.

label-image-script

labels

models-lib

trained-model

Trained model from train, transfer-learn, or finetune.

transfer-learn-checkpoint

resnet-152 model

TF-Slim ResNet v1 152 layer classifier.

Operations

evaluate

Evaluate a trained model.
Flag Description Default
batch‑size Number of examples in each evaluated batch. 100
eval‑batches Number of batches to evaluate.
step Checkpoint step to evaluate.

export-and-freeze

Export an inference graph with checkpoint weights.
Flag Description Default
step Checkpoint step to use for the frozen graph.

finetune

Finetune a trained model.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.0001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

label

Classify an image using a trained model.
Flag Description Default
image Path to image to classify. required

tflite

Generate a TFLite file from a frozen graph.
Flag Description Default
output‑format TF Lite output format. tflite
quantized Whether or not output arrays are quantized. False
quantized‑inputs Whether or not input arrays are quantized. False

train

Train model from scratch.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

transfer-learn

Train model using transfer learning.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

Resources

examples

Dataset generated with images:prepare.

frozen-graph

Frozen inference graph.

label-image-script

labels

models-lib

trained-model

Trained model from train, transfer-learn, or finetune.

transfer-learn-checkpoint

resnet-50 model

TF-Slim ResNet v1 50 layer classifier.

Operations

evaluate

Evaluate a trained model.
Flag Description Default
batch‑size Number of examples in each evaluated batch. 100
eval‑batches Number of batches to evaluate.
step Checkpoint step to evaluate.

export-and-freeze

Export an inference graph with checkpoint weights.
Flag Description Default
step Checkpoint step to use for the frozen graph.

finetune

Finetune a trained model.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.0001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

label

Classify an image using a trained model.
Flag Description Default
image Path to image to classify. required

tflite

Generate a TFLite file from a frozen graph.
Flag Description Default
output‑format TF Lite output format. tflite
quantized Whether or not output arrays are quantized. False
quantized‑inputs Whether or not input arrays are quantized. False

train

Train model from scratch.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

transfer-learn

Train model using transfer learning.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

Resources

examples

Dataset generated with images:prepare.

frozen-graph

Frozen inference graph.

label-image-script

labels

models-lib

trained-model

Trained model from train, transfer-learn, or finetune.

transfer-learn-checkpoint

resnet-v2-101 model

TF-Slim ResNet v2 101 layer classifier.

Operations

evaluate

Evaluate a trained model.
Flag Description Default
batch‑size Number of examples in each evaluated batch. 100
eval‑batches Number of batches to evaluate.
step Checkpoint step to evaluate.

export-and-freeze

Export an inference graph with checkpoint weights.
Flag Description Default
step Checkpoint step to use for the frozen graph.

finetune

Finetune a trained model.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.0001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

label

Classify an image using a trained model.
Flag Description Default
image Path to image to classify. required

tflite

Generate a TFLite file from a frozen graph.
Flag Description Default
output‑format TF Lite output format. tflite
quantized Whether or not output arrays are quantized. False
quantized‑inputs Whether or not input arrays are quantized. False

train

Train model from scratch.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

transfer-learn

Train model using transfer learning.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

Resources

examples

Dataset generated with images:prepare.

frozen-graph

Frozen inference graph.

label-image-script

labels

models-lib

trained-model

Trained model from train, transfer-learn, or finetune.

transfer-learn-checkpoint

resnet-v2-152 model

TF-Slim ResNet v2 152 layer classifier.

Operations

evaluate

Evaluate a trained model.
Flag Description Default
batch‑size Number of examples in each evaluated batch. 100
eval‑batches Number of batches to evaluate.
step Checkpoint step to evaluate.

export-and-freeze

Export an inference graph with checkpoint weights.
Flag Description Default
step Checkpoint step to use for the frozen graph.

finetune

Finetune a trained model.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.0001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

label

Classify an image using a trained model.
Flag Description Default
image Path to image to classify. required

tflite

Generate a TFLite file from a frozen graph.
Flag Description Default
output‑format TF Lite output format. tflite
quantized Whether or not output arrays are quantized. False
quantized‑inputs Whether or not input arrays are quantized. False

train

Train model from scratch.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

transfer-learn

Train model using transfer learning.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

Resources

examples

Dataset generated with images:prepare.

frozen-graph

Frozen inference graph.

label-image-script

labels

models-lib

trained-model

Trained model from train, transfer-learn, or finetune.

transfer-learn-checkpoint

resnet-v2-50 model

TF-Slim ResNet v2 50 layer classifier.

Operations

evaluate

Evaluate a trained model.
Flag Description Default
batch‑size Number of examples in each evaluated batch. 100
eval‑batches Number of batches to evaluate.
step Checkpoint step to evaluate.

export-and-freeze

Export an inference graph with checkpoint weights.
Flag Description Default
step Checkpoint step to use for the frozen graph.

finetune

Finetune a trained model.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.0001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

label

Classify an image using a trained model.
Flag Description Default
image Path to image to classify. required

tflite

Generate a TFLite file from a frozen graph.
Flag Description Default
output‑format TF Lite output format. tflite
quantized Whether or not output arrays are quantized. False
quantized‑inputs Whether or not input arrays are quantized. False

train

Train model from scratch.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

transfer-learn

Train model using transfer learning.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

Resources

examples

Dataset generated with images:prepare.

frozen-graph

Frozen inference graph.

label-image-script

labels

models-lib

trained-model

Trained model from train, transfer-learn, or finetune.

transfer-learn-checkpoint

vgg-16 model

TF-Slim VGG 16 classifier.

Operations

evaluate

Evaluate a trained model.
Flag Description Default
batch‑size Number of examples in each evaluated batch. 100
eval‑batches Number of batches to evaluate.
step Checkpoint step to evaluate.

export-and-freeze

Export an inference graph with checkpoint weights.
Flag Description Default
step Checkpoint step to use for the frozen graph.

finetune

Finetune a trained model.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.0001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

label

Classify an image using a trained model.
Flag Description Default
image Path to image to classify. required

tflite

Generate a TFLite file from a frozen graph.
Flag Description Default
output‑format TF Lite output format. tflite
quantized Whether or not output arrays are quantized. False
quantized‑inputs Whether or not input arrays are quantized. False

train

Train model from scratch.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

transfer-learn

Train model using transfer learning.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

Resources

examples

Dataset generated with images:prepare.

frozen-graph

Frozen inference graph.

label-image-script

labels

models-lib

trained-model

Trained model from train, transfer-learn, or finetune.

transfer-learn-checkpoint

vgg-19 model

TF-Slim VGG 19 classifier.

Operations

evaluate

Evaluate a trained model.
Flag Description Default
batch‑size Number of examples in each evaluated batch. 100
eval‑batches Number of batches to evaluate.
step Checkpoint step to evaluate.

export-and-freeze

Export an inference graph with checkpoint weights.
Flag Description Default
step Checkpoint step to use for the frozen graph.

finetune

Finetune a trained model.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.0001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

label

Classify an image using a trained model.
Flag Description Default
image Path to image to classify. required

tflite

Generate a TFLite file from a frozen graph.
Flag Description Default
output‑format TF Lite output format. tflite
quantized Whether or not output arrays are quantized. False
quantized‑inputs Whether or not input arrays are quantized. False

train

Train model from scratch.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

transfer-learn

Train model using transfer learning.
Flag Description Default
auto‑scale Adjust applicable flags for multi-GPU systems Set to 'no' to disable any flag value adjustments. When this value is 'yes' (the default) the following flags are adjusted on multi-GPU systems: - clones - learning-rate `clones` is set to the number of available GPUs. `learning-rate` is adjusted by multiplying its specified value by the number of GPUs. Flags are not adjusted on single GPU or CPU only systems. . True
batch‑size Number of examples in each training batch. 32
clones Number of model clones This value is automatically set to the number of available GPUs if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be increased from 1 to train the model in parallel on multiple GPUs. .
learning‑rate Initial learning rate. 0.001
learning‑rate‑decay‑epochs Number of epochs after which learning rate decays. 2.0
learning‑rate‑decay‑factor Learning rate decay factor. 0.94
learning‑rate‑decay‑type Method used to decay the learning rate. exponential
learning‑rate‑end Minimal learning rate used by polynomial learning rate decay. 0.0001
log‑save‑seconds Frequency of log summary saves in seconds. 60
log‑steps Frequency of summary logs in steps. 100
model‑save‑seconds Frequency of model saves (checkpoints) in seconds. 600
optimizer Optimizer used to train. rmsprop
preprocessing Preprocessing to use.
preprocessors Number of preprocessing threads This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize the preprocessor thread count for the system. .
readers Number of parallel data readers This value is automatically set to logical CPU count / 2 if `auto-scale` is 'yes'. When `auto-scale` is 'no' this value can be set to optimize data reader performance for the system. .
train‑steps Number of steps to train.
weight‑decay Decay on the model weights. 4e‑05

Resources

examples

Dataset generated with images:prepare.

frozen-graph

Frozen inference graph.

label-image-script

labels

models-lib

trained-model

Trained model from train, transfer-learn, or finetune.

transfer-learn-checkpoint