Skip to main content

TF-Slim Inception models (v1, v2, v3, v4, and Inception ResNet v2)

Project description

Models

slim-inception-resnet-v2

Inception ResNet v2 classifier for TF-Slim

Operations

evaluate

Evaluate a trained Inception ResNet v2 model

Flags
dataset
Dataset to train with (cifar10, mnist, flowers, custom) (required)
max-batches
Maximum number of batches to evaluate (default is all)

export

Generate a Inception ResNet v2 graph def

Flags
dataset
Dataset to train with (cifar10, mnist, flowers, custom) (required)

finetune

Fine tune a Inception ResNet v2 model

Flags
batch-size
Number of samples in each batch (default is 32)
checkpoint
Run ID or path to checkpoint to resume training from.
dataset
Dataset to train with (cifar10, mnist, flowers, custom) (required)
learning-rate
Initial learning rate (default is 0.01)
learning-rate-decay-type
How the learning rate is decayed (default is ‘exponential’)
log-every-n-steps
Steps between status updates (default is 100)
max-steps
Maximum number of training steps (default is 1000)
optimizer
Training optimizer (adadelta, adagrad, adam, ftrl, momentum, sgd, rmsprop) (default is ‘rmsprop’)
save-model-secs
Seconds between model saves (default is 60)
save-summaries-secs
Seconds between summary saves (default is 60)
weight-decay
Weight decay on the model weights (default is 4e-05)

freeze

Generate a Inception ResNet v2 graph def with checkpoint weights

predict

Use TensorFlow label_image and Inception ResNet v2 to classify an image

Flags
dataset
Dataset name to use for labels and image transformation (required)
image
Path to the input image (required)
input-mean
Image mean to apply to the image (default is 0.0)
input-std
Image std deviation to apply to the image (default is 1.0)

train

Train a Inception ResNet v2 model

Flags
batch-size
Number of samples in each batch (default is 32)
checkpoint
Run ID or path to checkpoint to resume training from.
dataset
Dataset to train with (cifar10, mnist, flowers, custom) (required)
learning-rate
Initial learning rate (default is 0.01)
learning-rate-decay-type
How the learning rate is decayed (default is ‘exponential’)
log-every-n-steps
Steps between status updates (default is 100)
max-steps
Maximum number of training steps (default is 1000)
optimizer
Training optimizer (adadelta, adagrad, adam, ftrl, momentum, sgd, rmsprop) (default is ‘rmsprop’)
save-model-secs
Seconds between model saves (default is 60)
save-summaries-secs
Seconds between summary saves (default is 60)
weight-decay
Weight decay on the model weights (default is 4e-05)

slim-inception-v1

Inception v1 classifier for TF-Slim

Operations

evaluate

Evaluate a trained Inception v1 model

Flags
dataset
Dataset to train with (cifar10, mnist, flowers, custom) (required)
max-batches
Maximum number of batches to evaluate (default is all)

export

Generate a Inception v1 graph def

Flags
dataset
Dataset to train with (cifar10, mnist, flowers, custom) (required)

finetune

Fine tune a Inception v1 model

Flags
batch-size
Number of samples in each batch (default is 32)
checkpoint
Run ID or path to checkpoint to resume training from.
dataset
Dataset to train with (cifar10, mnist, flowers, custom) (required)
learning-rate
Initial learning rate (default is 0.01)
learning-rate-decay-type
How the learning rate is decayed (default is ‘exponential’)
log-every-n-steps
Steps between status updates (default is 100)
max-steps
Maximum number of training steps (default is 1000)
optimizer
Training optimizer (adadelta, adagrad, adam, ftrl, momentum, sgd, rmsprop) (default is ‘rmsprop’)
save-model-secs
Seconds between model saves (default is 60)
save-summaries-secs
Seconds between summary saves (default is 60)
weight-decay
Weight decay on the model weights (default is 4e-05)

freeze

Generate a Inception v1 graph def with checkpoint weights

predict

Use TensorFlow label_image and Inception v1 to classify an image

Flags
dataset
Dataset name to use for labels and image transformation (required)
image
Path to the input image (required)
input-mean
Image mean to apply to the image (default is 0.0)
input-std
Image std deviation to apply to the image (default is 1.0)

train

Train a Inception v1 model

Flags
batch-size
Number of samples in each batch (default is 32)
checkpoint
Run ID or path to checkpoint to resume training from.
dataset
Dataset to train with (cifar10, mnist, flowers, custom) (required)
learning-rate
Initial learning rate (default is 0.01)
learning-rate-decay-type
How the learning rate is decayed (default is ‘exponential’)
log-every-n-steps
Steps between status updates (default is 100)
max-steps
Maximum number of training steps (default is 1000)
optimizer
Training optimizer (adadelta, adagrad, adam, ftrl, momentum, sgd, rmsprop) (default is ‘rmsprop’)
save-model-secs
Seconds between model saves (default is 60)
save-summaries-secs
Seconds between summary saves (default is 60)
weight-decay
Weight decay on the model weights (default is 4e-05)

slim-inception-v2

Inception v2 classifier for TF-Slim

Operations

evaluate

Evaluate a trained Inception v2 model

Flags
dataset
Dataset to train with (cifar10, mnist, flowers, custom) (required)
max-batches
Maximum number of batches to evaluate (default is all)

export

Generate a Inception v2 graph def

Flags
dataset
Dataset to train with (cifar10, mnist, flowers, custom) (required)

finetune

Fine tune a Inception v2 model

Flags
batch-size
Number of samples in each batch (default is 32)
checkpoint
Run ID or path to checkpoint to resume training from.
dataset
Dataset to train with (cifar10, mnist, flowers, custom) (required)
learning-rate
Initial learning rate (default is 0.01)
learning-rate-decay-type
How the learning rate is decayed (default is ‘exponential’)
log-every-n-steps
Steps between status updates (default is 100)
max-steps
Maximum number of training steps (default is 1000)
optimizer
Training optimizer (adadelta, adagrad, adam, ftrl, momentum, sgd, rmsprop) (default is ‘rmsprop’)
save-model-secs
Seconds between model saves (default is 60)
save-summaries-secs
Seconds between summary saves (default is 60)
weight-decay
Weight decay on the model weights (default is 4e-05)

freeze

Generate a Inception v2 graph def with checkpoint weights

predict

Use TensorFlow label_image and Inception v2 to classify an image

Flags
dataset
Dataset name to use for labels and image transformation (required)
image
Path to the input image (required)
input-mean
Image mean to apply to the image (default is 0.0)
input-std
Image std deviation to apply to the image (default is 1.0)

train

Train a Inception v2 model

Flags
batch-size
Number of samples in each batch (default is 32)
checkpoint
Run ID or path to checkpoint to resume training from.
dataset
Dataset to train with (cifar10, mnist, flowers, custom) (required)
learning-rate
Initial learning rate (default is 0.01)
learning-rate-decay-type
How the learning rate is decayed (default is ‘exponential’)
log-every-n-steps
Steps between status updates (default is 100)
max-steps
Maximum number of training steps (default is 1000)
optimizer
Training optimizer (adadelta, adagrad, adam, ftrl, momentum, sgd, rmsprop) (default is ‘rmsprop’)
save-model-secs
Seconds between model saves (default is 60)
save-summaries-secs
Seconds between summary saves (default is 60)
weight-decay
Weight decay on the model weights (default is 4e-05)

slim-inception-v3

Inception v3 classifier for TF-Slim

Operations

evaluate

Evaluate a trained Inception v3 model

Flags
dataset
Dataset to train with (cifar10, mnist, flowers, custom) (required)
max-batches
Maximum number of batches to evaluate (default is all)

export

Generate a Inception v3 graph def

Flags
dataset
Dataset to train with (cifar10, mnist, flowers, custom) (required)

finetune

Fine tune a Inception v3 model

Flags
batch-size
Number of samples in each batch (default is 32)
checkpoint
Run ID or path to checkpoint to resume training from.
dataset
Dataset to train with (cifar10, mnist, flowers, custom) (required)
learning-rate
Initial learning rate (default is 0.01)
learning-rate-decay-type
How the learning rate is decayed (default is ‘exponential’)
log-every-n-steps
Steps between status updates (default is 100)
max-steps
Maximum number of training steps (default is 1000)
optimizer
Training optimizer (adadelta, adagrad, adam, ftrl, momentum, sgd, rmsprop) (default is ‘rmsprop’)
save-model-secs
Seconds between model saves (default is 60)
save-summaries-secs
Seconds between summary saves (default is 60)
weight-decay
Weight decay on the model weights (default is 4e-05)

freeze

Generate a Inception v3 graph def with checkpoint weights

predict

Use TensorFlow label_image and Inception v3 to classify an image

Flags
dataset
Dataset name to use for labels and image transformation (required)
image
Path to the input image (required)
input-mean
Image mean to apply to the image (default is 0.0)
input-std
Image std deviation to apply to the image (default is 1.0)

train

Train a Inception v3 model

Flags
batch-size
Number of samples in each batch (default is 32)
checkpoint
Run ID or path to checkpoint to resume training from.
dataset
Dataset to train with (cifar10, mnist, flowers, custom) (required)
learning-rate
Initial learning rate (default is 0.01)
learning-rate-decay-type
How the learning rate is decayed (default is ‘exponential’)
log-every-n-steps
Steps between status updates (default is 100)
max-steps
Maximum number of training steps (default is 1000)
optimizer
Training optimizer (adadelta, adagrad, adam, ftrl, momentum, sgd, rmsprop) (default is ‘rmsprop’)
save-model-secs
Seconds between model saves (default is 60)
save-summaries-secs
Seconds between summary saves (default is 60)
weight-decay
Weight decay on the model weights (default is 4e-05)

slim-inception-v4

Inception v4 classifier for TF-Slim

Operations

evaluate

Evaluate a trained Inception v4 model

Flags
dataset
Dataset to train with (cifar10, mnist, flowers, custom) (required)
max-batches
Maximum number of batches to evaluate (default is all)

export

Generate a Inception v4 graph def

Flags
dataset
Dataset to train with (cifar10, mnist, flowers, custom) (required)

finetune

Fine tune a Inception v4 model

Flags
batch-size
Number of samples in each batch (default is 32)
checkpoint
Run ID or path to checkpoint to resume training from.
dataset
Dataset to train with (cifar10, mnist, flowers, custom) (required)
learning-rate
Initial learning rate (default is 0.01)
learning-rate-decay-type
How the learning rate is decayed (default is ‘exponential’)
log-every-n-steps
Steps between status updates (default is 100)
max-steps
Maximum number of training steps (default is 1000)
optimizer
Training optimizer (adadelta, adagrad, adam, ftrl, momentum, sgd, rmsprop) (default is ‘rmsprop’)
save-model-secs
Seconds between model saves (default is 60)
save-summaries-secs
Seconds between summary saves (default is 60)
weight-decay
Weight decay on the model weights (default is 4e-05)

freeze

Generate a Inception v4 graph def with checkpoint weights

predict

Use TensorFlow label_image and Inception v4 to classify an image

Flags
dataset
Dataset name to use for labels and image transformation (required)
image
Path to the input image (required)
input-mean
Image mean to apply to the image (default is 0.0)
input-std
Image std deviation to apply to the image (default is 1.0)

train

Train a Inception v4 model

Flags
batch-size
Number of samples in each batch (default is 32)
checkpoint
Run ID or path to checkpoint to resume training from.
dataset
Dataset to train with (cifar10, mnist, flowers, custom) (required)
learning-rate
Initial learning rate (default is 0.01)
learning-rate-decay-type
How the learning rate is decayed (default is ‘exponential’)
log-every-n-steps
Steps between status updates (default is 100)
max-steps
Maximum number of training steps (default is 1000)
optimizer
Training optimizer (adadelta, adagrad, adam, ftrl, momentum, sgd, rmsprop) (default is ‘rmsprop’)
save-model-secs
Seconds between model saves (default is 60)
save-summaries-secs
Seconds between summary saves (default is 60)
weight-decay
Weight decay on the model weights (default is 4e-05)

Project details


Release history Release notifications

This version
History Node

0.3.1

History Node

0.3.1.dev1

History Node

0.3.0

History Node

0.3.0.dev10

History Node

0.3.0.dev9

History Node

0.1.0.dev5

History Node

0.1.0.dev4

History Node

0.1.0.dev3

History Node

0.1.0.dev2

History Node

0.1.0.dev1

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Filename, size & hash SHA256 hash help File type Python version Upload date
gpkg.slim.inception-0.3.1-py2.py3-none-any.whl (7.1 kB) Copy SHA256 hash SHA256 Wheel py2.py3 Mar 28, 2018

Supported by

Elastic Elastic Search Pingdom Pingdom Monitoring Google Google BigQuery Sentry Sentry Error logging CloudAMQP CloudAMQP RabbitMQ AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN DigiCert DigiCert EV certificate StatusPage StatusPage Status page