def longest_path(self):
if self._model is None:
raise RuntimeError(
"Can't compute longest path of model without creating a model."
)
nx_graph = nx.drawing.nx_pydot.from_pydot(model_to_dot(self._model))
return nx.algorithms.dag.dag_longest_path(nx_graph)
def test_model(self):
"""Test that it can be called with a model."""
model = Sequential()
model.add(Conv2D(10, (5, 5), input_shape=(28, 28, 1), activation="relu"))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Flatten())
model.add(Dense(8, activation="relu"))
with patch("streamlit.graphviz_chart") as graphviz_chart:
st.write(model)
dot = vis_utils.model_to_dot(model)
graphviz_chart.assert_called_once_with(dot.to_string())
def plot_model_recursive(model, title=None, exclude_models_by_name=None):
# model.summary()
if type(exclude_models_by_name) is list:
if model.name in exclude_models_by_name:
return
else:
exclude_models_by_name = []
exclude_models_by_name.append(model.name)
if title is None:
title = 'Model %s' % model.name
# render pydot by calling dot, no file saved to disk
png_str = model_to_dot(model, show_shapes=True).create_png(prog='dot')
# treat the dot output string as an image file
sio = BytesIO()
sio.write(png_str)
sio.seek(0)
img = mpimg.imread(sio)
# set actual size of image on plot
dpi = 80
height, width, depth = img.shape
figsize = width / float(dpi), height / float(dpi)
fig = plt.figure(figsize=figsize)
# plot the image
plt.imshow(img)
plt.axis('off')
plt.title(title)
plt.show(block=False)
for layer in model.layers:
if isinstance(layer, Model):
plot_model_recursive(layer,
exclude_models_by_name=exclude_models_by_name)
def write(*args, **kwargs):
"""Write arguments to the app.
This is the swiss-army knife of Streamlit commands. It does different
things depending on what you throw at it.
Unlike other Streamlit commands, write() has some unique properties:
1. You can pass in multiple arguments, all of which will be written.
2. Its behavior depends on the input types as follows.
3. It returns None, so it's "slot" in the App cannot be reused.
Parameters
----------
*args : any
One or many objects to print to the App.
Arguments are handled as follows:
- write(string) : Prints the formatted Markdown string.
- write(data_frame) : Displays the DataFrame as a table.
- write(error) : Prints an exception specially.
- write(func) : Displays information about a function.
- write(module) : Displays information about the module.
- write(dict) : Displays dict in an interactive widget.
- write(obj) : The default is to print str(obj).
- write(mpl_fig) : Displays a Matplotlib figure.
- write(altair) : Displays an Altair chart.
- write(keras) : Displays a Keras model.
- write(graphviz) : Displays a Graphviz graph.
- write(plotly_fig) : Displays a Plotly figure.
- write(bokeh_fig) : Displays a Bokeh figure.
unsafe_allow_html : bool
This is a keyword-only argument that defaults to False.
By default, any HTML tags found in strings will be escaped and
therefore treated as pure text. This behavior may be turned off by
setting this argument to True.
That said, *we strongly advise* against it*. It is hard to write secure
HTML, so by using this argument you may be compromising your users'
security. For more information, see:
https://github.com/streamlit/streamlit/issues/152
*Also note that `unsafe_allow_html` is a temporary measure and may be
removed from Streamlit at any time.*
If you decide to turn on HTML anyway, we ask you to please tell us your
exact use case here:
https://discuss.streamlit.io/t/96
This will help us come up with safe APIs that allow you to do what you
want.
Example
-------
Its simplest use case is to draw Markdown-formatted text, whenever the
input is a string:
>>> write('Hello, *World!*')
.. output::
https://share.streamlit.io/0.25.0-2JkNY/index.html?id=DUJaq97ZQGiVAFi6YvnihF
height: 50px
As mentioned earlier, `st.write()` also accepts other data formats, such as
numbers, data frames, styled data frames, and assorted objects:
>>> st.write(1234)
>>> st.write(pd.DataFrame({
... 'first column': [1, 2, 3, 4],
... 'second column': [10, 20, 30, 40],
... }))
.. output::
https://share.streamlit.io/0.25.0-2JkNY/index.html?id=FCp9AMJHwHRsWSiqMgUZGD
height: 250px
Finally, you can pass in multiple arguments to do things like:
>>> st.write('1 + 1 = ', 2)
>>> st.write('Below is a DataFrame:', data_frame, 'Above is a dataframe.')
.. output::
https://share.streamlit.io/0.25.0-2JkNY/index.html?id=DHkcU72sxYcGarkFbf4kK1
height: 300px
Oh, one more thing: `st.write` accepts chart objects too! For example:
>>> import pandas as pd
>>> import numpy as np
>>> import altair as alt
>>>
>>> df = pd.DataFrame(
... np.random.randn(200, 3),
... columns=['a', 'b', 'c'])
...
>>> c = alt.Chart(df).mark_circle().encode(
... x='a', y='b', size='c', color='c')
>>>
>>> st.write(c)
.. output::
https://share.streamlit.io/0.25.0-2JkNY/index.html?id=8jmmXR8iKoZGV4kXaKGYV5
height: 200px
"""
# Python2 doesn't support this syntax
# def write(*args, unsafe_allow_html=False)
# so we do this instead:
unsafe_allow_html = kwargs.get('unsafe_allow_html', False)
try:
string_buffer = []
def flush_buffer():
if string_buffer:
markdown(" ".join(string_buffer),
unsafe_allow_html=unsafe_allow_html) # noqa: F821
string_buffer[:] = []
for arg in args:
# Order matters!
if isinstance(arg, string_types): # noqa: F821
string_buffer.append(arg)
elif type(arg).__name__ in _DATAFRAME_LIKE_TYPES:
flush_buffer()
if len(_np.shape(arg)) > 2:
text(arg)
else:
dataframe(arg) # noqa: F821
elif isinstance(arg, Exception):
flush_buffer()
exception(arg) # noqa: F821
elif isinstance(arg, _HELP_TYPES):
flush_buffer()
help(arg)
elif _util.is_altair_chart(arg):
flush_buffer()
altair_chart(arg)
elif _util.is_type(arg, "matplotlib.figure.Figure"):
flush_buffer()
pyplot(arg)
elif _util.is_plotly_chart(arg):
flush_buffer()
plotly_chart(arg)
elif _util.is_type(arg, "bokeh.plotting.figure.Figure"):
flush_buffer()
bokeh_chart(arg)
elif _util.is_graphviz_chart(arg):
flush_buffer()
graphviz_chart(arg)
elif _util.is_keras_model(arg):
from tensorflow.python.keras.utils import vis_utils
flush_buffer()
dot = vis_utils.model_to_dot(arg)
graphviz_chart(dot.to_string())
elif (type(arg) in dict_types) or (isinstance(arg,
list)): # noqa: F821
flush_buffer()
json(arg)
elif _util.is_namedtuple(arg):
flush_buffer()
json(_json.dumps(arg._asdict()))
else:
string_buffer.append("`%s`" % str(arg).replace("`", "\\`"))
flush_buffer()
except Exception:
_, exc, exc_tb = _sys.exc_info()
exception(exc, exc_tb) # noqa: F821
def write(*args):
"""Write arguments to the report.
This is the swiss-army knife of Streamlit commands. It does different
things depending on what you throw at it.
Unlike other Streamlit commands, write() has some unique properties:
1. You can pass in multiple arguments, all of which will be written.
2. Its behavior depends on the input types as follows.
3. It returns None, so it's "slot" in the report cannot be reused.
Parameters
----------
*args : any
One or many objects to print to the Report.
Arguments are handled as follows:
- write(string) : Prints the formatted Markdown string.
- write(data_frame) : Displays the DataFrame as a table.
- write(error) : Prints an exception specially.
- write(func) : Displays information about a function.
- write(module) : Displays information about the module.
- write(dict) : Displays dict in an interactive widget.
- write(obj) : The default is to print str(obj).
- write(mpl_fig) : Displays a Matplotlib figure.
- write(altair) : Displays an Altair chart.
- write(keras) : Displays a Keras model.
- write(graphviz) : Displays a Graphviz graph.
- write(plotly_fig) : Displays a Plotly figure.
- write(bokeh_fig) : Displays a Bokeh figure.
Example
-------
Its simplest use case is to draw Markdown-formatted text, whenever the
input is a string:
>>> write('Hello, *World!*')
.. output::
https://share.streamlit.io/0.25.0-2JkNY/index.html?id=DUJaq97ZQGiVAFi6YvnihF
height: 50px
As mentioned earlier, `st.write()` also accepts other data formats, such as
numbers, data frames, styled data frames, and assorted objects:
>>> st.write(1234)
>>> st.write(pd.DataFrame({
... 'first column': [1, 2, 3, 4],
... 'second column': [10, 20, 30, 40],
... }))
.. output::
https://share.streamlit.io/0.25.0-2JkNY/index.html?id=FCp9AMJHwHRsWSiqMgUZGD
height: 250px
Finally, you can pass in multiple arguments to do things like:
>>> st.write('1 + 1 = ', 2)
>>> st.write('Below is a DataFrame:', data_frame, 'Above is a dataframe.')
.. output::
https://share.streamlit.io/0.25.0-2JkNY/index.html?id=DHkcU72sxYcGarkFbf4kK1
height: 300px
Oh, one more thing: `st.write` accepts chart objects too! For example:
>>> import pandas as pd
>>> import numpy as np
>>> import altair as alt
>>>
>>> df = pd.DataFrame(
... np.random.randn(200, 3),
... columns=['a', 'b', 'c'])
...
>>> c = alt.Chart(df).mark_circle().encode(
... x='a', y='b', size='c', color='c')
>>>
>>> st.write(c)
.. output::
https://share.streamlit.io/0.25.0-2JkNY/index.html?id=8jmmXR8iKoZGV4kXaKGYV5
height: 200px
"""
try:
string_buffer = []
def flush_buffer():
if string_buffer:
markdown(' '.join(string_buffer)) # noqa: F821
string_buffer[:] = []
for arg in args:
# Order matters!
if isinstance(arg, string_types): # noqa: F821
string_buffer.append(arg)
elif type(arg).__name__ in _DATAFRAME_LIKE_TYPES:
flush_buffer()
dataframe(arg) # noqa: F821
elif isinstance(arg, Exception):
flush_buffer()
exception(arg) # noqa: F821
elif isinstance(arg, _HELP_TYPES):
flush_buffer()
help(arg)
elif _util.is_altair_chart(arg):
flush_buffer()
altair_chart(arg)
elif _util.is_type(arg, 'matplotlib.figure.Figure'):
flush_buffer()
pyplot(arg)
elif _util.is_plotly_chart(arg):
flush_buffer()
plotly_chart(arg)
elif _util.is_type(arg, 'bokeh.plotting.figure.Figure'):
flush_buffer()
bokeh_chart(arg)
elif _util.is_graphviz_chart(arg):
flush_buffer()
graphviz_chart(arg)
elif util.is_keras_model(arg):
from tensorflow.python.keras.utils import vis_utils
flush_buffer()
dot = vis_utils.model_to_dot(arg)
graphviz_chart(dot.to_string())
elif type(arg) in dict_types: # noqa: F821
flush_buffer()
json(arg)
else:
string_buffer.append('`%s`' % str(arg).replace('`', '\\`'))
flush_buffer()
except Exception:
_, exc, exc_tb = _sys.exc_info()
exception(exc, exc_tb) # noqa: F821
def write(self, *args, **kwargs):
"""Write arguments to the app.
This is the Swiss Army knife of Streamlit commands: it does different
things depending on what you throw at it. Unlike other Streamlit commands,
write() has some unique properties:
1. You can pass in multiple arguments, all of which will be written.
2. Its behavior depends on the input types as follows.
3. It returns None, so it's "slot" in the App cannot be reused.
Parameters
----------
*args : any
One or many objects to print to the App.
Arguments are handled as follows:
- write(string) : Prints the formatted Markdown string, with
support for LaTeX expression and emoji shortcodes.
See docs for st.markdown for more.
- write(data_frame) : Displays the DataFrame as a table.
- write(error) : Prints an exception specially.
- write(func) : Displays information about a function.
- write(module) : Displays information about the module.
- write(dict) : Displays dict in an interactive widget.
- write(obj) : The default is to print str(obj).
- write(mpl_fig) : Displays a Matplotlib figure.
- write(altair) : Displays an Altair chart.
- write(keras) : Displays a Keras model.
- write(graphviz) : Displays a Graphviz graph.
- write(plotly_fig) : Displays a Plotly figure.
- write(bokeh_fig) : Displays a Bokeh figure.
- write(sympy_expr) : Prints SymPy expression using LaTeX.
unsafe_allow_html : bool
This is a keyword-only argument that defaults to False.
By default, any HTML tags found in strings will be escaped and
therefore treated as pure text. This behavior may be turned off by
setting this argument to True.
That said, *we strongly advise* against it*. It is hard to write secure
HTML, so by using this argument you may be compromising your users'
security. For more information, see:
https://github.com/streamlit/streamlit/issues/152
**Also note that `unsafe_allow_html` is a temporary measure and may be
removed from Streamlit at any time.**
If you decide to turn on HTML anyway, we ask you to please tell us your
exact use case here:
https://discuss.streamlit.io/t/96 .
This will help us come up with safe APIs that allow you to do what you
want.
Example
-------
Its basic use case is to draw Markdown-formatted text, whenever the
input is a string:
>>> write('Hello, *World!* :sunglasses:')
.. output::
https://static.streamlit.io/0.50.2-ZWk9/index.html?id=Pn5sjhgNs4a8ZbiUoSTRxE
height: 50px
As mentioned earlier, `st.write()` also accepts other data formats, such as
numbers, data frames, styled data frames, and assorted objects:
>>> st.write(1234)
>>> st.write(pd.DataFrame({
... 'first column': [1, 2, 3, 4],
... 'second column': [10, 20, 30, 40],
... }))
.. output::
https://static.streamlit.io/0.25.0-2JkNY/index.html?id=FCp9AMJHwHRsWSiqMgUZGD
height: 250px
Finally, you can pass in multiple arguments to do things like:
>>> st.write('1 + 1 = ', 2)
>>> st.write('Below is a DataFrame:', data_frame, 'Above is a dataframe.')
.. output::
https://static.streamlit.io/0.25.0-2JkNY/index.html?id=DHkcU72sxYcGarkFbf4kK1
height: 300px
Oh, one more thing: `st.write` accepts chart objects too! For example:
>>> import pandas as pd
>>> import numpy as np
>>> import altair as alt
>>>
>>> df = pd.DataFrame(
... np.random.randn(200, 3),
... columns=['a', 'b', 'c'])
...
>>> c = alt.Chart(df).mark_circle().encode(
... x='a', y='b', size='c', color='c', tooltip=['a', 'b', 'c'])
>>>
>>> st.write(c)
.. output::
https://static.streamlit.io/0.25.0-2JkNY/index.html?id=8jmmXR8iKoZGV4kXaKGYV5
height: 200px
"""
string_buffer = [] # type: List[str]
unsafe_allow_html = kwargs.get("unsafe_allow_html", False)
# This bans some valid cases like: e = st.empty(); e.write("a", "b").
# BUT: 1) such cases are rare, 2) this rule is easy to understand,
# and 3) this rule should be removed once we have st.container()
if not self.dg._is_top_level and len(args) > 1:
raise StreamlitAPIException(
"Cannot replace a single element with multiple elements.\n\n"
"The `write()` method only supports multiple elements when "
"inserting elements rather than replacing. That is, only "
"when called as `st.write()` or `st.sidebar.write()`.")
def flush_buffer():
if string_buffer:
self.dg.markdown(
" ".join(string_buffer),
unsafe_allow_html=unsafe_allow_html,
)
string_buffer[:] = []
for arg in args:
# Order matters!
if isinstance(arg, str):
string_buffer.append(arg)
elif type_util.is_dataframe_like(arg):
flush_buffer()
if len(np.shape(arg)) > 2:
self.dg.text(arg)
else:
self.dg.dataframe(arg)
elif isinstance(arg, Exception):
flush_buffer()
self.dg.exception(arg)
elif isinstance(arg, HELP_TYPES):
flush_buffer()
self.dg.help(arg)
elif type_util.is_altair_chart(arg):
flush_buffer()
self.dg.altair_chart(arg)
elif type_util.is_type(arg, "matplotlib.figure.Figure"):
flush_buffer()
self.dg.pyplot(arg)
elif type_util.is_plotly_chart(arg):
flush_buffer()
self.dg.plotly_chart(arg)
elif type_util.is_type(arg, "bokeh.plotting.figure.Figure"):
flush_buffer()
self.dg.bokeh_chart(arg)
elif type_util.is_graphviz_chart(arg):
flush_buffer()
self.dg.graphviz_chart(arg)
elif type_util.is_sympy_expession(arg):
flush_buffer()
self.dg.latex(arg)
elif type_util.is_keras_model(arg):
from tensorflow.python.keras.utils import vis_utils
flush_buffer()
dot = vis_utils.model_to_dot(arg)
self.dg.graphviz_chart(dot.to_string())
elif isinstance(arg, (dict, list)):
flush_buffer()
self.dg.json(arg)
elif type_util.is_namedtuple(arg):
flush_buffer()
self.dg.json(json.dumps(arg._asdict()))
elif type_util.is_pydeck(arg):
flush_buffer()
self.dg.pydeck_chart(arg)
else:
string_buffer.append("`%s`" % str(arg).replace("`", "\\`"))
flush_buffer()
kernel_regularizer=tf.keras.regularizers.l2(1e-4),
model_name="fashion_mnist_test")
# Get keras fit data
OPTIMIZER = tf.keras.optimizers.SGD(lr=MAX_LR, momentum=0.9, nesterov=True)
LOSS = tf.keras.losses.categorical_crossentropy
METRICS = [
tf.keras.metrics.categorical_accuracy,
]
CALLBACKS = []
# get visualisation
SVG(
model_to_dot(keras_model, show_shapes=True).create(prog='dot',
format='svg'))
# compile
keras_model.compile(optimizer=OPTIMIZER, loss=LOSS, metrics=METRICS)
# fit
history = keras_model.fit(train_dataset_fn(),
epochs=NUM_EPOCHS,
verbose=2,
callbacks=CALLBACKS,
validation_data=test_dataset_fn(),
initial_epoch=0,
steps_per_epoch=NUM_TRAIN_STEPS,
validation_steps=NUM_TEST_STEPS)
# Plot training & validation accuracy values
# preprocessing sentences into sentence vectors
sentence = Input(shape=(T, embedding_size), name='Sentences') # batch, 50, 300
sentence_vec = Bidirectional(CuDNNGRU(units=n_a, return_sequences=False), name='Sentence_Vectors')(sentence) # batch, 300
# dot
#product = Dot(axes=-1, normalize=False, name='Matrix')([word_vec, sentence_vec])
product = tf.matmul(word_vec, sentence_vec, transpose_b = True, name = 'Matrix')
key_matrix = K.transpose(product)
model = Model(inputs= sentence, outputs=key_matrix)
return model
# create a model
Factorize = model(embedding_size, n_a)
# get a summary of the model
Factorize.summary()
SVG(model_to_dot(Factorize, show_shapes=True, show_layer_names=True, rankdir='HB').create(prog='dot', format='svg'))
plot_model(Factorize, to_file='model.png', show_shapes=True, show_layer_names=True)
#%% Run
Factorize.compile(loss='mean_squared_error', optimizer = Adam(lr=0.001), metrics=['mean_squared_error', 'mean_absolute_error'])
hist = Factorize.fit(X_train_padded, key_matrix_train, batch_size=256, epochs=500, verbose=1, validation_data=(X_test_padded, key_matrix_test))
#%% Loss plot
%matplotlib inline
import matplotlib.pyplot as plt
# accuracy
np.newaxis],
to_categorical(y_dataset),
test_size=0.2,
random_state=42)
hid_dim = 10
# SimpleRNNにDenseを接続し、分類
model = Sequential()
model.add(SimpleRNN(hid_dim, input_shape=x_train.shape[1:])
) # input_shape=(系列長T, x_tの次元), output_shape=(units(=hid_dim),)
model.add(Dense(y_train.shape[1], activation='softmax'))
model.compile(loss='categorical_crossentropy',
optimizer='adam',
metrics=['accuracy'])
model.fit(x_train,
y_train,
epochs=50,
batch_size=100,
verbose=2,
validation_split=0.2)
score = model.evaluate(x_test, y_test, verbose=0)
print('test_loss:', score[0])
print('test_acc:', score[1])
SVG(model_to_dot(model).create(prog='dot', format='svg'))
#Model
input = Input(shape=(224, 224, 3))
x = conv_batchnorm_relu(input, filters=64, kernel_size=7, strides=2)
x = MaxPool2D(pool_size=3, strides=2)(x)
x = resnet_block(x, filters=64, reps=3, strides=1)
x = resnet_block(x, filters=128, reps=4, strides=2)
x = resnet_block(x, filters=256, reps=6, strides=2)
x = resnet_block(x, filters=512, reps=3, strides=2)
x = GlobalAvgPool2D()(x)
output = Dense(1000, activation='softmax')(x)
model = Model(inputs=input, outputs=output)
model.summary()
from tensorflow.python.keras.utils.vis_utils import model_to_dot
from IPython.display import SVG
import pydot
import graphviz
SVG(
model_to_dot(model,
show_shapes=True,
show_layer_names=True,
rankdir='TB',
expand_nested=False,
dpi=60,
subgraph=False).create(prog='dot', format='svg'))
# loss は損失関数
# 連続値の時は平均二乗誤差が使われることが多い loss='mean_squared_error'
# 離散値の時は交差エントロピーが使われることが多い
# 2クラス交差エントロピー loss='binary_crossentropy'
# 多クラス交差エントロピー loss='categorical_crossentropy'
model.compile(loss='categorical_crossentropy',
optimizer='sgd',
metrics=['accuracy'])
# 最後に fit でデータを入れる
model.fit(x_train,
y_train,
batch_size=1000,
epochs=10,
verbose=1,
validation_data=(x_test, y_test))
# モデルを評価する
score = model.evaluate(x_test, y_test, verbose=0)
print('Test loss:', score[0])
print('Test accuracy:', score[1])
# 新しいデータを予測する場合
# classes = model.predict(x_test, batch_size=128)
# print(classes)
# モデルの可視化
from IPython.display import SVG
from tensorflow.python.keras.utils.vis_utils import model_to_dot
print(model_to_dot(model, show_shapes=True))
