How to Save and Restore Multiple Variables in Python

How to save all the variables in the current python session?

If you use shelve, you do not have to remember the order in which the objects are pickled, since shelve gives you a dictionary-like object:

To shelve your work:

import shelve

T='Hiya'
val=[1,2,3]

filename='/tmp/shelve.out'
my_shelf = shelve.open(filename,'n') # 'n' for new

for key in dir():
    try:
        my_shelf[key] = globals()[key]
    except TypeError:
        #
        # __builtins__, my_shelf, and imported modules can not be shelved.
        #
        print('ERROR shelving: {0}'.format(key))
my_shelf.close()

To restore:

my_shelf = shelve.open(filename)
for key in my_shelf:
    globals()[key]=my_shelf[key]
my_shelf.close()

print(T)
# Hiya
print(val)
# [1, 2, 3]

How to save and load selected and all variables in tensorflow 2.0 using tf.train.Checkpoint?

I'm not sure if this is what you mean, but you can create a tf.train.Checkpoint object specifically for the variables that you want to save and restore. See the following example:

import tensorflow as tf

class manyVariables:
    def __init__(self):
        self.initList = [None]*100
        for i in range(100):
            self.initList[i] = tf.Variable(tf.random.normal([5,5]))
        self.makeSomeMoreVariables()
        self.ckpt = self.makeCheckpoint()

    def makeSomeMoreVariables(self):
        self.moreList = [None]*10
        for i in range(10):
            self.moreList[i] = tf.Variable(tf.random.normal([3,3]))

    def makeCheckpoint(self):
        return tf.train.Checkpoint(
            init3=self.initList[3], init55=self.initList[55],
            init60=self.initList[60], more4=self.moreList[4])

    def saveVariables(self):
        self.ckpt.save('./ckpt')

    def restoreVariables(self):
        status = self.ckpt.restore(tf.train.latest_checkpoint('.'))
        status.assert_consumed()  # Optional check

# Create variables
v1 = manyVariables()
# Assigned fixed values
for i, v in enumerate(v1.initList):
    v.assign(i * tf.ones_like(v))
for i, v in enumerate(v1.moreList):
    v.assign(100 + i * tf.ones_like(v))
# Save them
v1.saveVariables()

# Create new variables
v2 = manyVariables()
# Check initial values
print(v2.initList[2].numpy())
# [[-1.9110833   0.05956204 -1.1753829  -0.3572553  -0.95049495]
#  [ 0.31409055  1.1262076   0.47890127 -0.1699607   0.4409122 ]
#  [-0.75385517 -0.13847834  0.97012395  0.42515194 -1.4371008 ]
#  [ 0.44205236  0.86158335  0.6919655  -2.5156968   0.16496429]
#  [-1.241602   -0.15177743  0.5603795  -0.3560254  -0.18536267]]
print(v2.initList[3].numpy())
# [[-3.3441594  -0.18425298 -0.4898144  -1.2330629   0.08798431]
#  [ 1.5002227   0.99475247  0.7817361   0.3849587  -0.59548247]
#  [-0.57121766 -1.277224    0.6957546  -0.67618763  0.0510064 ]
#  [ 0.85491985  0.13310803 -0.93152267  0.10205163  0.57520276]
#  [-1.0606447  -0.16966362 -1.0448577   0.56799036 -0.90726566]]

# Restore them
v2.restoreVariables()
# Check values after restoring
print(v2.initList[2].numpy())
# [[-1.9110833   0.05956204 -1.1753829  -0.3572553  -0.95049495]
#  [ 0.31409055  1.1262076   0.47890127 -0.1699607   0.4409122 ]
#  [-0.75385517 -0.13847834  0.97012395  0.42515194 -1.4371008 ]
#  [ 0.44205236  0.86158335  0.6919655  -2.5156968   0.16496429]
#  [-1.241602   -0.15177743  0.5603795  -0.3560254  -0.18536267]]
print(v2.initList[3].numpy())
# [[3. 3. 3. 3. 3.]
#  [3. 3. 3. 3. 3.]
#  [3. 3. 3. 3. 3.]
#  [3. 3. 3. 3. 3.]
#  [3. 3. 3. 3. 3.]]

If you want to save all the variables in the lists, you could replace makeCheckpoint with something like this:

def makeCheckpoint(self):
    return tf.train.Checkpoint(
        **{f'init{i}': v for i, v in enumerate(self.initList)},
        **{f'more{i}': v for i, v in enumerate(self.moreList)})

Note that you can have "nested" checkpoints so, more generally, you could have a function that makes a checkpoint for a list of variables, for example like this:

def listCheckpoint(varList):
    # Use 'item{}'.format(i) if using Python <3.6
    return tf.train.Checkpoint(**{f'item{i}': v for i, v in enumerate(varList)})

Then you could just have this:

def makeCheckpoint(self):
    return tf.train.Checkpoint(init=listCheckpoint(self.initList),
                               more=listCheckpoint(self.moreList))

Saving and loading multiple objects in pickle file?

Using a list, tuple, or dict is by far the most common way to do this:

import pickle
PIK = "pickle.dat"

data = ["A", "b", "C", "d"]
with open(PIK, "wb") as f:
    pickle.dump(data, f)
with open(PIK, "rb") as f:
    print pickle.load(f)

That prints:

['A', 'b', 'C', 'd']

However, a pickle file can contain any number of pickles. Here's code producing the same output. But note that it's harder to write and to understand:

with open(PIK, "wb") as f:
    pickle.dump(len(data), f)
    for value in data:
        pickle.dump(value, f)
data2 = []
with open(PIK, "rb") as f:
    for _ in range(pickle.load(f)):
        data2.append(pickle.load(f))
print data2

If you do this, you're responsible for knowing how many pickles are in the file you write out. The code above does that by pickling the number of list objects first.

Plain-text, Python-syntax file to save and restore some class variables?

You can import the module into your class like:

Code:

class Teost:
    import a as _a_py_attrs

    def __init__(self):
        for name in dir(Teost._a_py_attrs):
            if not name.startswith('__'):
                setattr(self, name, getattr(Teost._a_py_attrs, name))

Test Code:

xx = Teost()
print(xx.__dict__)
print(xx.a)

Results:

{'a': 'hello', 'b': 134}
hello

As Class Attributes:

If it would be preferable to have these as class attributes, instead of instance attributes, you can do:

class Teost:
    """ My Test Class """

import a as _a_py_attrs
for name in dir(_a_py_attrs):
    if not name.startswith('__'):
        setattr(Teost, name, getattr(_a_py_attrs, name))

Test Code:

xx = Teost()
print(xx.__dict__)
print(xx.a)

Results:

{}
hello

How to save/restore a model after training?

Tensorflow 2 Docs

Saving Checkpoints

Adapted from the docs

# -------------------------
# -----  Toy Context  -----
# -------------------------
import tensorflow as tf

class Net(tf.keras.Model):
    """A simple linear model."""

    def __init__(self):
        super(Net, self).__init__()
        self.l1 = tf.keras.layers.Dense(5)

    def call(self, x):
        return self.l1(x)

def toy_dataset():
    inputs = tf.range(10.0)[:, None]
    labels = inputs * 5.0 + tf.range(5.0)[None, :]
    return (
        tf.data.Dataset.from_tensor_slices(dict(x=inputs, y=labels)).repeat().batch(2)
    )

def train_step(net, example, optimizer):
    """Trains `net` on `example` using `optimizer`."""
    with tf.GradientTape() as tape:
        output = net(example["x"])
        loss = tf.reduce_mean(tf.abs(output - example["y"]))
    variables = net.trainable_variables
    gradients = tape.gradient(loss, variables)
    optimizer.apply_gradients(zip(gradients, variables))
    return loss

# ----------------------------
# -----  Create Objects  -----
# ----------------------------

net = Net()
opt = tf.keras.optimizers.Adam(0.1)
dataset = toy_dataset()
iterator = iter(dataset)
ckpt = tf.train.Checkpoint(
    step=tf.Variable(1), optimizer=opt, net=net, iterator=iterator
)
manager = tf.train.CheckpointManager(ckpt, "./tf_ckpts", max_to_keep=3)

# ----------------------------
# -----  Train and Save  -----
# ----------------------------

ckpt.restore(manager.latest_checkpoint)
if manager.latest_checkpoint:
    print("Restored from {}".format(manager.latest_checkpoint))
else:
    print("Initializing from scratch.")

for _ in range(50):
    example = next(iterator)
    loss = train_step(net, example, opt)
    ckpt.step.assign_add(1)
    if int(ckpt.step) % 10 == 0:
        save_path = manager.save()
        print("Saved checkpoint for step {}: {}".format(int(ckpt.step), save_path))
        print("loss {:1.2f}".format(loss.numpy()))

# ---------------------
# -----  Restore  -----
# ---------------------

# In another script, re-initialize objects
opt = tf.keras.optimizers.Adam(0.1)
net = Net()
dataset = toy_dataset()
iterator = iter(dataset)
ckpt = tf.train.Checkpoint(
    step=tf.Variable(1), optimizer=opt, net=net, iterator=iterator
)
manager = tf.train.CheckpointManager(ckpt, "./tf_ckpts", max_to_keep=3)

# Re-use the manager code above ^

ckpt.restore(manager.latest_checkpoint)
if manager.latest_checkpoint:
    print("Restored from {}".format(manager.latest_checkpoint))
else:
    print("Initializing from scratch.")

for _ in range(50):
    example = next(iterator)
    # Continue training or evaluate etc.

Tensorflow < 2

From the docs:

Save

# Create some variables.
v1 = tf.get_variable("v1", shape=[3], initializer = tf.zeros_initializer)
v2 = tf.get_variable("v2", shape=[5], initializer = tf.zeros_initializer)

inc_v1 = v1.assign(v1+1)
dec_v2 = v2.assign(v2-1)

# Add an op to initialize the variables.
init_op = tf.global_variables_initializer()

# Add ops to save and restore all the variables.
saver = tf.train.Saver()

# Later, launch the model, initialize the variables, do some work, and save the
# variables to disk.
with tf.Session() as sess:
  sess.run(init_op)
  # Do some work with the model.
  inc_v1.op.run()
  dec_v2.op.run()
  # Save the variables to disk.
  save_path = saver.save(sess, "/tmp/model.ckpt")
  print("Model saved in path: %s" % save_path)

Restore

tf.reset_default_graph()

# Create some variables.
v1 = tf.get_variable("v1", shape=[3])
v2 = tf.get_variable("v2", shape=[5])

# Add ops to save and restore all the variables.
saver = tf.train.Saver()

# Later, launch the model, use the saver to restore variables from disk, and
# do some work with the model.
with tf.Session() as sess:
  # Restore variables from disk.
  saver.restore(sess, "/tmp/model.ckpt")
  print("Model restored.")
  # Check the values of the variables
  print("v1 : %s" % v1.eval())
  print("v2 : %s" % v2.eval())

`simple_save`

Many good answer, for completeness I'll add my 2 cents: simple_save. Also a standalone code example using the tf.data.Dataset API.

Python 3 ; Tensorflow 1.14

import tensorflow as tf
from tensorflow.saved_model import tag_constants

with tf.Graph().as_default():
    with tf.Session() as sess:
        ...

        # Saving
        inputs = {
            "batch_size_placeholder": batch_size_placeholder,
            "features_placeholder": features_placeholder,
            "labels_placeholder": labels_placeholder,
        }
        outputs = {"prediction": model_output}
        tf.saved_model.simple_save(
            sess, 'path/to/your/location/', inputs, outputs
        )

Restoring:

graph = tf.Graph()
with restored_graph.as_default():
    with tf.Session() as sess:
        tf.saved_model.loader.load(
            sess,
            [tag_constants.SERVING],
            'path/to/your/location/',
        )
        batch_size_placeholder = graph.get_tensor_by_name('batch_size_placeholder:0')
        features_placeholder = graph.get_tensor_by_name('features_placeholder:0')
        labels_placeholder = graph.get_tensor_by_name('labels_placeholder:0')
        prediction = restored_graph.get_tensor_by_name('dense/BiasAdd:0')

        sess.run(prediction, feed_dict={
            batch_size_placeholder: some_value,
            features_placeholder: some_other_value,
            labels_placeholder: another_value
        })

Standalone example

Original blog post

The following code generates random data for the sake of the demonstration.

We start by creating the placeholders. They will hold the data at runtime. From them, we create the Dataset and then its Iterator. We get the iterator's generated tensor, called input_tensor which will serve as input to our model.
The model itself is built from input_tensor: a GRU-based bidirectional RNN followed by a dense classifier. Because why not.
The loss is a softmax_cross_entropy_with_logits, optimized with Adam. After 2 epochs (of 2 batches each), we save the "trained" model with tf.saved_model.simple_save. If you run the code as is, then the model will be saved in a folder called simple/ in your current working directory.
In a new graph, we then restore the saved model with tf.saved_model.loader.load. We grab the placeholders and logits with graph.get_tensor_by_name and the Iterator initializing operation with graph.get_operation_by_name.
Lastly we run an inference for both batches in the dataset, and check that the saved and restored model both yield the same values. They do!

Code:

import os
import shutil
import numpy as np
import tensorflow as tf
from tensorflow.python.saved_model import tag_constants

def model(graph, input_tensor):
    """Create the model which consists of
    a bidirectional rnn (GRU(10)) followed by a dense classifier

    Args:
        graph (tf.Graph): Tensors' graph
        input_tensor (tf.Tensor): Tensor fed as input to the model

    Returns:
        tf.Tensor: the model's output layer Tensor
    """
    cell = tf.nn.rnn_cell.GRUCell(10)
    with graph.as_default():
        ((fw_outputs, bw_outputs), (fw_state, bw_state)) = tf.nn.bidirectional_dynamic_rnn(
            cell_fw=cell,
            cell_bw=cell,
            inputs=input_tensor,
            sequence_length=[10] * 32,
            dtype=tf.float32,
            swap_memory=True,
            scope=None)
        outputs = tf.concat((fw_outputs, bw_outputs), 2)
        mean = tf.reduce_mean(outputs, axis=1)
        dense = tf.layers.dense(mean, 5, activation=None)

        return dense

def get_opt_op(graph, logits, labels_tensor):
    """Create optimization operation from model's logits and labels

    Args:
        graph (tf.Graph): Tensors' graph
        logits (tf.Tensor): The model's output without activation
        labels_tensor (tf.Tensor): Target labels

    Returns:
        tf.Operation: the operation performing a stem of Adam optimizer
    """
    with graph.as_default():
        with tf.variable_scope('loss'):
            loss = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(
                    logits=logits, labels=labels_tensor, name='xent'),
                    name="mean-xent"
                    )
        with tf.variable_scope('optimizer'):
            opt_op = tf.train.AdamOptimizer(1e-2).minimize(loss)
        return opt_op

if __name__ == '__main__':
    # Set random seed for reproducibility
    # and create synthetic data
    np.random.seed(0)
    features = np.random.randn(64, 10, 30)
    labels = np.eye(5)[np.random.randint(0, 5, (64,))]

    graph1 = tf.Graph()
    with graph1.as_default():
        # Random seed for reproducibility
        tf.set_random_seed(0)
        # Placeholders
        batch_size_ph = tf.placeholder(tf.int64, name='batch_size_ph')
        features_data_ph = tf.placeholder(tf.float32, [None, None, 30], 'features_data_ph')
        labels_data_ph = tf.placeholder(tf.int32, [None, 5], 'labels_data_ph')
        # Dataset
        dataset = tf.data.Dataset.from_tensor_slices((features_data_ph, labels_data_ph))
        dataset = dataset.batch(batch_size_ph)
        iterator = tf.data.Iterator.from_structure(dataset.output_types, dataset.output_shapes)
        dataset_init_op = iterator.make_initializer(dataset, name='dataset_init')
        input_tensor, labels_tensor = iterator.get_next()

        # Model
        logits = model(graph1, input_tensor)
        # Optimization
        opt_op = get_opt_op(graph1, logits, labels_tensor)

        with tf.Session(graph=graph1) as sess:
            # Initialize variables
            tf.global_variables_initializer().run(session=sess)
            for epoch in range(3):
                batch = 0
                # Initialize dataset (could feed epochs in Dataset.repeat(epochs))
                sess.run(
                    dataset_init_op,
                    feed_dict={
                        features_data_ph: features,
                        labels_data_ph: labels,
                        batch_size_ph: 32
                    })
                values = []
                while True:
                    try:
                        if epoch < 2:
                            # Training
                            _, value = sess.run([opt_op, logits])
                            print('Epoch {}, batch {} | Sample value: {}'.format(epoch, batch, value[0]))
                            batch += 1
                        else:
                            # Final inference
                            values.append(sess.run(logits))
                            print('Epoch {}, batch {} | Final inference | Sample value: {}'.format(epoch, batch, values[-1][0]))
                            batch += 1
                    except tf.errors.OutOfRangeError:
                        break
            # Save model state
            print('\nSaving...')
            cwd = os.getcwd()
            path = os.path.join(cwd, 'simple')
            shutil.rmtree(path, ignore_errors=True)
            inputs_dict = {
                "batch_size_ph": batch_size_ph,
                "features_data_ph": features_data_ph,
                "labels_data_ph": labels_data_ph
            }
            outputs_dict = {
                "logits": logits
            }
            tf.saved_model.simple_save(
                sess, path, inputs_dict, outputs_dict
            )
            print('Ok')
    # Restoring
    graph2 = tf.Graph()
    with graph2.as_default():
        with tf.Session(graph=graph2) as sess:
            # Restore saved values
            print('\nRestoring...')
            tf.saved_model.loader.load(
                sess,
                [tag_constants.SERVING],
                path
            )
            print('Ok')
            # Get restored placeholders
            labels_data_ph = graph2.get_tensor_by_name('labels_data_ph:0')
            features_data_ph = graph2.get_tensor_by_name('features_data_ph:0')
            batch_size_ph = graph2.get_tensor_by_name('batch_size_ph:0')
            # Get restored model output
            restored_logits = graph2.get_tensor_by_name('dense/BiasAdd:0')
            # Get dataset initializing operation
            dataset_init_op = graph2.get_operation_by_name('dataset_init')

            # Initialize restored dataset
            sess.run(
                dataset_init_op,
                feed_dict={
                    features_data_ph: features,
                    labels_data_ph: labels,
                    batch_size_ph: 32
                }

            )
            # Compute inference for both batches in dataset
            restored_values = []
            for i in range(2):
                restored_values.append(sess.run(restored_logits))
                print('Restored values: ', restored_values[i][0])

    # Check if original inference and restored inference are equal
    valid = all((v == rv).all() for v, rv in zip(values, restored_values))
    print('\nInferences match: ', valid)

This will print:

$ python3 save_and_restore.py

Epoch 0, batch 0 | Sample value: [-0.13851789 -0.3087595   0.12804556  0.20013677 -0.08229901]
Epoch 0, batch 1 | Sample value: [-0.00555491 -0.04339041 -0.05111827 -0.2480045  -0.00107776]
Epoch 1, batch 0 | Sample value: [-0.19321944 -0.2104792  -0.00602257  0.07465433  0.11674127]
Epoch 1, batch 1 | Sample value: [-0.05275984  0.05981954 -0.15913513 -0.3244143   0.10673307]
Epoch 2, batch 0 | Final inference | Sample value: [-0.26331693 -0.13013336 -0.12553    -0.04276478  0.2933622 ]
Epoch 2, batch 1 | Final inference | Sample value: [-0.07730117  0.11119192 -0.20817074 -0.35660955  0.16990358]

Saving...
INFO:tensorflow:Assets added to graph.
INFO:tensorflow:No assets to write.
INFO:tensorflow:SavedModel written to: b'/some/path/simple/saved_model.pb'
Ok

Restoring...
INFO:tensorflow:Restoring parameters from b'/some/path/simple/variables/variables'
Ok
Restored values:  [-0.26331693 -0.13013336 -0.12553    -0.04276478  0.2933622 ]
Restored values:  [-0.07730117  0.11119192 -0.20817074 -0.35660955  0.16990358]

Inferences match:  True

Import multiple MetaGraphDefs into a single graph and restore variables

Solved! The Saver adds ops to the graph, and since both Savers were in the same name_scope they were interfering with one another. You need to wrap each call to import_meta_graph in its own name_scope:

with self.graph.as_default(), tf.device(device):
    with tf.Session(graph=self.graph, config=self.tf_config) as sess:

        # Add inherited graphs to CenterNet's graph.
        with tf.name_scope(self.maskmaker.name):
            self.mm_saver = tf.train.import_meta_graph(self.maskmaker.model_ckpt + ".meta")
        with tf.name_scope(self.deepvar.name):
            self.dv_saver = tf.train.import_meta_graph(self.deepvar.model_ckpt + ".meta")

        # First saver can restore
        self.mm_saver.restore(sess, self.maskmaker.model_ckpt)
        # Second saver can also restore
        self.dv_saver.restore(sess, self.deepvar.model_ckpt)