markdown stringlengths 0 37k | code stringlengths 1 33.3k | path stringlengths 8 215 | repo_name stringlengths 6 77 | license stringclasses 15
values |
|---|---|---|---|---|
Under the Hood
This section contains additional information about the functions we've used and pointers to their documentation.
You don't need to know anything in these sections, so if you are already feeling overwhelmed, you might want to skip them. But if you are curious, read on.
State and TimeSeries objects are ba... | source_code(flip) | python/soln/chap02.ipynb | AllenDowney/ModSim | gpl-2.0 |
HLT2 nbody classification
did preselections:
any sv.n,
any sv.minpt
sv.nlt16 < 2
Training channels (read data)
We will use just 11114001, 11296013, 11874042, 12103035, 13246001, 13264021 | sig_train_modes_names = [11114001, 11296013, 11874042, 12103035, 13246001, 13264021]
bck_train_mode_name = 30000000
sig_train_files = ['mod_{}.csv'.format(name) for name in sig_train_modes_names]
bck_train_files = 'mod_30000000.csv'
folder = "datasets/prepared_hlt_body/"
# concat all signal data
if not os.path.exists(... | HLT2-TreesPruning.ipynb | yandexdataschool/LHCb-topo-trigger | apache-2.0 |
Counting events and svrs,
that passed L0 and GoodGenB preselection (this data was generated by skim) | print 'Signal', statistic_length(signal_data)
print 'Bck', statistic_length(bck_data)
total_bck_events = statistic_length(bck_data)['Events'] + empty_events[bck_train_mode_name]
total_signal_events_by_mode = dict()
for mode in sig_train_modes_names:
total_signal_events_by_mode[mode] = statistic_length(signal_data[... | HLT2-TreesPruning.ipynb | yandexdataschool/LHCb-topo-trigger | apache-2.0 |
events distribution by mode | print 'Bck:', total_bck_events
'Signal:', total_signal_events_by_mode | HLT2-TreesPruning.ipynb | yandexdataschool/LHCb-topo-trigger | apache-2.0 |
Define variables | variables = ["n", "mcor", "chi2", "eta", "fdchi2", "minpt", "nlt16", "ipchi2", "n1trk", "sumpt"] | HLT2-TreesPruning.ipynb | yandexdataschool/LHCb-topo-trigger | apache-2.0 |
Counting events and svrs,
which passed pass_nbody (equivalent Mike's preselections for nbody selection) | # hlt2 nbody selection
signal_data = signal_data[(signal_data['pass_nbody'] == 1) & (signal_data['mcor'] <= 10e3)]
bck_data = bck_data[(bck_data['pass_nbody'] == 1) & (bck_data['mcor'] <= 10e3)]
print 'Signal', statistic_length(signal_data)
print 'Bck', statistic_length(bck_data)
total_signal_events_by_mode_presel = ... | HLT2-TreesPruning.ipynb | yandexdataschool/LHCb-topo-trigger | apache-2.0 |
events distribution by mode | print 'Bck:', total_bck_events_presel
'Signal:', total_signal_events_by_mode_presel
signal_data.head() | HLT2-TreesPruning.ipynb | yandexdataschool/LHCb-topo-trigger | apache-2.0 |
Prepare train/test splitting
Divide events which passed alll preselections into two equal parts randomly | ds_train_signal, ds_train_bck, ds_test_signal, ds_test_bck = prepare_data(signal_data, bck_data, 'unique') | HLT2-TreesPruning.ipynb | yandexdataschool/LHCb-topo-trigger | apache-2.0 |
train: counting events and svrs | print 'Signal', statistic_length(ds_train_signal)
print 'Bck', statistic_length(ds_train_bck)
train = pandas.concat([ds_train_bck, ds_train_signal]) | HLT2-TreesPruning.ipynb | yandexdataschool/LHCb-topo-trigger | apache-2.0 |
test: counting events and svrs | print 'Signal', statistic_length(ds_test_signal)
print 'Bck', statistic_length(ds_test_bck)
test = pandas.concat([ds_test_bck, ds_test_signal]) | HLT2-TreesPruning.ipynb | yandexdataschool/LHCb-topo-trigger | apache-2.0 |
Define all total events in test samples
(which passed just l0 and goodgenB) using also empty events. Suppose that events which didn't pass pass_nboby also were equal randomly divided into training and test samples | total_test_bck_events = (total_bck_events - total_bck_events_presel) // 2 + statistic_length(ds_test_bck)['Events']
total_test_signal_events = dict()
for mode in sig_train_modes_names:
total_not_passed_signal = total_signal_events_by_mode[mode] - total_signal_events_by_mode_presel[mode]
total_test_signal_events... | HLT2-TreesPruning.ipynb | yandexdataschool/LHCb-topo-trigger | apache-2.0 |
Matrixnet training | from rep_ef.estimators import MatrixNetSkyGridClassifier | HLT2-TreesPruning.ipynb | yandexdataschool/LHCb-topo-trigger | apache-2.0 |
Base model with 5000 trees | ef_base = MatrixNetSkyGridClassifier(train_features=variables, user_name='antares',
connection='skygrid',
iterations=5000, sync=False)
ef_base.fit(train, train['signal']) | HLT2-TreesPruning.ipynb | yandexdataschool/LHCb-topo-trigger | apache-2.0 |
Base BBDT model | special_b = {
'n': [2.5, 3.5],
'mcor': [2000,3000,4000,5000,7500], # I want to remove splits too close the the B mass as I was looking in simulation and this could distort the mass peak (possibly)
'chi2': [1,2.5,5,7.5,10,100], # I also propose we add a cut to the pre-selection of chi2 < 1000. I don't want to put in ... | HLT2-TreesPruning.ipynb | yandexdataschool/LHCb-topo-trigger | apache-2.0 |
BBDT-5, 6 | ef_base_bbdt5 = MatrixNetSkyGridClassifier(train_features=variables, user_name='antares',
connection='skygrid',
iterations=5000, sync=False, intervals=5)
ef_base_bbdt5.fit(train, train['signal'])
ef_base_bbdt6 = MatrixNetSkyGridClassif... | HLT2-TreesPruning.ipynb | yandexdataschool/LHCb-topo-trigger | apache-2.0 |
Pruning | from rep.data import LabeledDataStorage
from rep.report import ClassificationReport
report = ClassificationReport({'base': ef_base}, LabeledDataStorage(test, test['signal']))
report.roc() | HLT2-TreesPruning.ipynb | yandexdataschool/LHCb-topo-trigger | apache-2.0 |
Minimize log_loss
он же BinomialDeviance | %run pruning.py | HLT2-TreesPruning.ipynb | yandexdataschool/LHCb-topo-trigger | apache-2.0 |
Training sample is cut to be aliquot 8 | new_trainlen = (len(train) // 8) * 8
trainX = train[ef_base.features][:new_trainlen].values
trainY = train['signal'][:new_trainlen].values
trainW = numpy.ones(len(trainY))
trainW[trainY == 0] *= sum(trainY) / sum(1 - trainY)
new_features, new_formula_mx, new_classifier = select_trees(trainX, trainY, sample_weight=tra... | HLT2-TreesPruning.ipynb | yandexdataschool/LHCb-topo-trigger | apache-2.0 |
Calculate thresholds on classifiers | thresholds = dict()
test_bck = test[test['signal'] == 0]
RATE = [2500., 4000.]
events_pass = dict()
for name, cl in estimators.items():
prob = cl.predict_proba(test_bck)
thr, result = calculate_thresholds(test_bck, prob, total_test_bck_events, rates=RATE)
for rate, val in result.items():
events_pass... | HLT2-TreesPruning.ipynb | yandexdataschool/LHCb-topo-trigger | apache-2.0 |
Final efficiencies for each mode | train_modes_eff, statistic = result_statistic(estimators, sig_train_modes_names,
test[test['signal'] == 1],
thresholds, RATE, total_test_signal_events)
from rep.plotting import BarComparePlot
xticks_labels = ['$B^0 \\to K^*\mu... | HLT2-TreesPruning.ipynb | yandexdataschool/LHCb-topo-trigger | apache-2.0 |
Classification report using events | plots = OrderedDict()
for key, value in estimators.items():
plots[key] = plot_roc_events(value, test[test['signal'] == 1], test[test['signal'] == 0], key)
bbdt_plots = plots.copy()
bbdt_plots.pop('Prunned MN')
bbdt_plots.pop('Prunned MN + forest')
from rep.plotting import FunctionsPlot
FunctionsPlot(bbdt_plots).p... | HLT2-TreesPruning.ipynb | yandexdataschool/LHCb-topo-trigger | apache-2.0 |
all channels efficiencies | from collections import defaultdict
all_channels = []
efficiencies = defaultdict(OrderedDict)
for mode in empty_events.keys():
if mode in set(sig_train_modes_names) or mode == bck_train_mode_name:
continue
df = pandas.read_csv(os.path.join(folder , 'mod_{}.csv'.format(mode)), sep='\t')
if len(df) <=... | HLT2-TreesPruning.ipynb | yandexdataschool/LHCb-topo-trigger | apache-2.0 |
DIfferent rates | thresholds = OrderedDict()
RATE = [2000., 2500., 3000., 3500., 4000.]
for name, cl in estimators.items():
prob = cl.predict_proba(ds_test_bck)
thr, result = calculate_thresholds(ds_test_bck, prob, total_test_bck_events, rates=RATE)
thresholds[name] = thr
print name, result
train_modes_eff, statistic = ... | HLT2-TreesPruning.ipynb | yandexdataschool/LHCb-topo-trigger | apache-2.0 |
Now let's set the integrator to whfast, and sacrificing accuracy for speed, set the timestep for the integration to about $10\%$ of Jupiter's orbital period. | sim.integrator = "whfast"
sim.dt = 1. # in years. About 10% of Jupiter's period
sim.move_to_com() | ipython_examples/FourierSpectrum.ipynb | dchandan/rebound | gpl-3.0 |
The last line (moving to the center of mass frame) is important to take out the linear drift in positions due to the constant COM motion. Without it we would erase some of the signal at low frequencies.
Now let's run the integration, storing time series for the two planets' eccentricities (for plotting) and x-position... | Nout = 100000
tmax = 3.e5
Nplanets = 2
x = np.zeros((Nplanets,Nout))
ecc = np.zeros((Nplanets,Nout))
longitude = np.zeros((Nplanets,Nout))
varpi = np.zeros((Nplanets,Nout))
times = np.linspace(0.,tmax,Nout)
ps = sim.particles
for i,time in enumerate(times):
sim.integrate(time)
os = sim.calculate_orbits()
... | ipython_examples/FourierSpectrum.ipynb | dchandan/rebound | gpl-3.0 |
Let's see what the eccentricity evolution looks like with matplotlib: | %matplotlib inline
labels = ["Jupiter", "Saturn"]
import matplotlib.pyplot as plt
fig = plt.figure(figsize=(12,5))
ax = plt.subplot(111)
plt.plot(times,ecc[0],label=labels[0])
plt.plot(times,ecc[1],label=labels[1])
ax.set_xlabel("Time (yrs)", fontsize=20)
ax.set_ylabel("Eccentricity", fontsize=20)
ax.tick_params(labels... | ipython_examples/FourierSpectrum.ipynb | dchandan/rebound | gpl-3.0 |
Now let's try to analyze the periodicities in this signal. Here we have a uniformly spaced time series, so we could run a Fast Fourier Transform, but as an example of the wider array of tools available through scipy, let's run a Lomb-Scargle periodogram (which allows for non-uniform time series). This could also be us... | from scipy import signal
Npts = 3000
logPmin = np.log10(10.)
logPmax = np.log10(1.e5)
Ps = np.logspace(logPmin,logPmax,Npts)
ws = np.asarray([2*np.pi/P for P in Ps])
periodogram = signal.lombscargle(times,x[0],ws)
fig = plt.figure(figsize=(12,5))
ax = plt.subplot(111)
ax.plot(Ps,np.sqrt(4*periodogram/Nout))
ax.set_xs... | ipython_examples/FourierSpectrum.ipynb | dchandan/rebound | gpl-3.0 |
We pick out the obvious signal in the eccentricity plot with a period of $\approx 45000$ yrs, which is due to secular interactions between the two planets. There is quite a bit of power aliased into neighboring frequencies due to the short integration duration, with contributions from the second secular timescale, whi... | fig = plt.figure(figsize=(12,5))
ax = plt.subplot(111)
ax.plot(Ps,np.sqrt(4*periodogram/Nout))
ax.set_xscale('log')
ax.set_xlim([600,1600])
ax.set_ylim([0,0.003])
ax.set_xlabel("Period (yrs)", fontsize=20)
ax.set_ylabel("Power", fontsize=20)
ax.tick_params(labelsize=20) | ipython_examples/FourierSpectrum.ipynb | dchandan/rebound | gpl-3.0 |
This is the right timescale to be due to resonant perturbations between giant planets ($\sim 100$ orbits). In fact, Jupiter and Saturn are close to a 5:2 mean-motion resonance. This is the famous great inequality that Laplace showed was responsible for slight offsets in the predicted positions of the two giant planet... | def zeroTo360(val):
while val < 0:
val += 2*np.pi
while val > 2*np.pi:
val -= 2*np.pi
return val*180/np.pi | ipython_examples/FourierSpectrum.ipynb | dchandan/rebound | gpl-3.0 |
Now we construct $\phi_{5:2}$ and plot it over the first 5000 yrs. | phi = [zeroTo360(5.*longitude[1][i] - 2.*longitude[0][i] - 3.*varpi[0][i]) for i in range(Nout)]
fig = plt.figure(figsize=(12,5))
ax = plt.subplot(111)
ax.plot(times,phi)
ax.set_xlim([0,5.e3])
ax.set_ylim([0,360.])
ax.set_xlabel("time (yrs)", fontsize=20)
ax.set_ylabel(r"$\phi_{5:2}$", fontsize=20)
ax.tick_params(label... | ipython_examples/FourierSpectrum.ipynb | dchandan/rebound | gpl-3.0 |
We see that the resonant angle $\phi_{5:2}$ circulates, but with a long period of $\approx 900$ yrs (compared to the orbital periods of $\sim 10$ yrs), which precisely matches the blip we saw in the Lomb-Scargle periodogram. This is approximately the same oscillation period observed in the Solar System, despite our si... | phi2 = [zeroTo360(2*longitude[1][i] - longitude[0][i] - varpi[0][i]) for i in range(Nout)]
fig = plt.figure(figsize=(12,5))
ax = plt.subplot(111)
ax.plot(times,phi2)
ax.set_xlim([0,5.e3])
ax.set_ylim([0,360.])
ax.set_xlabel("time (yrs)", fontsize=20)
ax.set_ylabel(r"$\phi_{2:1}$", fontsize=20)
ax.tick_params(labelsize=... | ipython_examples/FourierSpectrum.ipynb | dchandan/rebound | gpl-3.0 |
Using interact for animation with data
A soliton is a constant velocity wave that maintains its shape as it propagates. They arise from non-linear wave equations, such has the Korteweg–de Vries equation, which has the following analytical solution:
$$
\phi(x,t) = \frac{1}{2} c \mathrm{sech}^2 \left[ \frac{\sqrt{c}}{2} ... | def soliton(x, t, c, a):
"""Return phi(x, t) for a soliton wave with constants c and a."""
return (0.5)*c*((1/np.cosh(((np.sqrt(c)*0.5)*(x-c*t-a)))**2))
assert np.allclose(soliton(np.array([0]),0.0,1.0,0.0), np.array([0.5])) | assignments/assignment05/InteractEx03.ipynb | brettavedisian/phys202-2015-work | mit |
Compute a 2d NumPy array called phi:
It should have a dtype of float.
It should have a shape of (xpoints, tpoints).
phi[i,j] should contain the value $\phi(x[i],t[j])$. | phi=np.zeros((xpoints,tpoints)) #worked with Hunter Thomas
for i in x:
for j in t:
phi[i,j]=soliton(x[i],t[j],c,a)
assert phi.shape==(xpoints, tpoints)
assert phi.ndim==2
assert phi.dtype==np.dtype(float)
assert phi[0,0]==soliton(x[0],t[0],c,a) | assignments/assignment05/InteractEx03.ipynb | brettavedisian/phys202-2015-work | mit |
Write a plot_soliton_data(i) function that plots the soliton wave $\phi(x, t[i])$. Customize your plot to make it effective and beautiful. | def plot_soliton_data(i=0):
"""Plot the soliton data at t[i] versus x."""
plt.plot(soliton(x,t[i],c,a))
plt.xlabel('Time')
plt.ylabel('Phi')
plt.title('Solition wave vs. Time')
plt.tick_params(axis='x', top='off', direction='out')
plt.tick_params(axis='y', right='off', direction='out')
plot... | assignments/assignment05/InteractEx03.ipynb | brettavedisian/phys202-2015-work | mit |
Use interact to animate the plot_soliton_data function versus time. | interact(plot_soliton_data, i=(0.0,50.0,0.1));
assert True # leave this for grading the interact with plot_soliton_data cell | assignments/assignment05/InteractEx03.ipynb | brettavedisian/phys202-2015-work | mit |
画像分類器を再トレーニングする
<table class="tfo-notebook-buttons" align="left">
<td><a target="_blank" href="https://www.tensorflow.org/hub/tutorials/tf2_image_retraining"><img src="https://www.tensorflow.org/images/tf_logo_32px.png"> TensorFlow.orgで表示</a></td>
<td> <a target="_blank" href="https://colab.research.google.com/... | import itertools
import os
import matplotlib.pylab as plt
import numpy as np
import tensorflow as tf
import tensorflow_hub as hub
print("TF version:", tf.__version__)
print("Hub version:", hub.__version__)
print("GPU is", "available" if tf.config.list_physical_devices('GPU') else "NOT AVAILABLE") | site/ja/hub/tutorials/tf2_image_retraining.ipynb | tensorflow/docs-l10n | apache-2.0 |
使用する TF2 SavedModel モジュールを選択する
手始めに、https://tfhub.dev/google/imagenet/mobilenet_v2_100_224/feature_vector/4 を使用します。同じ URL を、SavedModel を識別するコードに使用できます。またブラウザで使用すれば、そのドキュメントを表示することができます。（ここでは TF1 Hub 形式のモデルは機能しないことに注意してください。）
画像特徴量ベクトルを生成するその他の TF2 モデルは、こちらをご覧ください。
試すことのできるモデルはたくさんあります。下のセルから別のモデルを選択し、ノートブックの指示に従ってください。 | model_name = "efficientnetv2-xl-21k" # @param ['efficientnetv2-s', 'efficientnetv2-m', 'efficientnetv2-l', 'efficientnetv2-s-21k', 'efficientnetv2-m-21k', 'efficientnetv2-l-21k', 'efficientnetv2-xl-21k', 'efficientnetv2-b0-21k', 'efficientnetv2-b1-21k', 'efficientnetv2-b2-21k', 'efficientnetv2-b3-21k', 'efficientnetv2-... | site/ja/hub/tutorials/tf2_image_retraining.ipynb | tensorflow/docs-l10n | apache-2.0 |
Flowers データセットをセットアップする
入力は、選択されたモジュールに合わせてサイズ変更されます。データセットを拡張することで（読み取られるたびに画像をランダムに歪みを加える）、特にファインチューニング時のトレーニングが改善されます。 | data_dir = tf.keras.utils.get_file(
'flower_photos',
'https://storage.googleapis.com/download.tensorflow.org/example_images/flower_photos.tgz',
untar=True)
def build_dataset(subset):
return tf.keras.preprocessing.image_dataset_from_directory(
data_dir,
validation_split=.20,
subset=subse... | site/ja/hub/tutorials/tf2_image_retraining.ipynb | tensorflow/docs-l10n | apache-2.0 |
モデルを定義する
Hub モジュールを使用して、線形分類器を feature_extractor_layer の上に配置するだけで定義できます。
高速化するため、トレーニング不可能な feature_extractor_layer から始めますが、ファインチューニングを実施して精度を高めることもできます。 | do_fine_tuning = False #@param {type:"boolean"}
print("Building model with", model_handle)
model = tf.keras.Sequential([
# Explicitly define the input shape so the model can be properly
# loaded by the TFLiteConverter
tf.keras.layers.InputLayer(input_shape=IMAGE_SIZE + (3,)),
hub.KerasLayer(model_handl... | site/ja/hub/tutorials/tf2_image_retraining.ipynb | tensorflow/docs-l10n | apache-2.0 |
モデルをトレーニングする | model.compile(
optimizer=tf.keras.optimizers.SGD(learning_rate=0.005, momentum=0.9),
loss=tf.keras.losses.CategoricalCrossentropy(from_logits=True, label_smoothing=0.1),
metrics=['accuracy'])
steps_per_epoch = train_size // BATCH_SIZE
validation_steps = valid_size // BATCH_SIZE
hist = model.fit(
train_ds,
... | site/ja/hub/tutorials/tf2_image_retraining.ipynb | tensorflow/docs-l10n | apache-2.0 |
検証データの画像でモデルが機能するか試してみましょう。 | x, y = next(iter(val_ds))
image = x[0, :, :, :]
true_index = np.argmax(y[0])
plt.imshow(image)
plt.axis('off')
plt.show()
# Expand the validation image to (1, 224, 224, 3) before predicting the label
prediction_scores = model.predict(np.expand_dims(image, axis=0))
predicted_index = np.argmax(prediction_scores)
print("... | site/ja/hub/tutorials/tf2_image_retraining.ipynb | tensorflow/docs-l10n | apache-2.0 |
最後に次のようにして、トレーニングされたモデルを、TF Serving または TF Lite（モバイル）用に保存することができます。 | saved_model_path = f"/tmp/saved_flowers_model_{model_name}"
tf.saved_model.save(model, saved_model_path) | site/ja/hub/tutorials/tf2_image_retraining.ipynb | tensorflow/docs-l10n | apache-2.0 |
オプション: TensorFlow Lite にデプロイする
TensorFlow Lite では、TensorFlow モデルをモバイルおよび IoT デバイスにデプロイすることができます。以下のコードには、トレーニングされたモデルを TF Lite に変換して、TensorFlow Model Optimization Toolkit のポストトレーニングツールを適用する方法が示されています。最後に、結果の質を調べるために、変換したモデルを TF Lite Interpreter で実行しています。
最適化せずに変換すると、前と同じ結果が得られます（丸め誤差まで）。
データなしで最適化して変換すると、モデルの重みを 8 ビット... | #@title Optimization settings
optimize_lite_model = False #@param {type:"boolean"}
#@markdown Setting a value greater than zero enables quantization of neural network activations. A few dozen is already a useful amount.
num_calibration_examples = 60 #@param {type:"slider", min:0, max:1000, step:1}
representative_data... | site/ja/hub/tutorials/tf2_image_retraining.ipynb | tensorflow/docs-l10n | apache-2.0 |
Review: Forward / Backward
Here are solutions to last hands-on lecture's coding problems along with example uses with a pre-defined A and B matrices.
$\alpha_t(z_t) = B_{z_t,x_t} \sum_{z_{t-1}} \alpha_{t-1}(z_{t-1}) A_{z_{t-1}, z_t} $
$\beta(z_t) = \sum_{z_{t+1}} A_{z_t, z_{t+1}} B_{z_{t+1}, x_{t+1}} \beta_{t+1}(z_{t+1... | def forward(params, observations):
pi, A, B = params
N = len(observations)
S = pi.shape[0]
alpha = np.zeros((N, S))
# base case
alpha[0, :] = pi * B[observations[0], :]
# recursive case
for i in range(1, N):
for s2 in range(S):
for s1 in range(S):
... | handsOn_lecture19_baum-welch-pgm-inference/handsOn_lecture19_baum-welch-pgm-inference-template.ipynb | eecs445-f16/umich-eecs445-f16 | mit |
Implementing Baum-welch
With the forward and backward algorithm implementions ready, let's use them to implement baum-welch, EM for HMMs.
In the M step, here's the parameters are updated:
$ p(z_{t-1}, z_t | \X, \theta) = \frac{\alpha_{t-1}(z_{t-1}) \beta_t(z_t) A_{z_{t-1}, z_t} B_{z_t, x_t}}{\sum_k \alpha_t(k)\beta_t(k... | # Some utitlities for tracing our implementation below
def left_pad(i, s):
return "\n".join(["{}{}".format(' '*i, l) for l in s.split("\n")])
def pad_print(i, s):
print(left_pad(i, s))
def pad_print_args(i, **kwargs):
pad_print(i, "\n".join(["{}:\n{}".format(k, kwargs[k]) for k in sorted(kwargs.keys(... | handsOn_lecture19_baum-welch-pgm-inference/handsOn_lecture19_baum-welch-pgm-inference-template.ipynb | eecs445-f16/umich-eecs445-f16 | mit |
Training with examples
Let's try producing updated parameters to our HMM using a few examples. How did the A and B matrixes get updated with data? Was any confidence gained in the emission probabilities of nouns? Verbs? | pi2, A2, B2 = baum_welch([
[THE, DOG, WALKED, IN, THE, PARK, END, END], # END -> END needs at least one transition example
[THE, DOG, RAN, IN, THE, PARK, END],
[THE, CAT, WALKED, IN, THE, PARK, END],
[THE, DOG, RAN, IN, THE, PARK, END]], pi, A, B, 10, trace=False)
print("original A")
pr... | handsOn_lecture19_baum-welch-pgm-inference/handsOn_lecture19_baum-welch-pgm-inference-template.ipynb | eecs445-f16/umich-eecs445-f16 | mit |
Tracing through the implementation
Let's look at a trace of one iteration. Study the steps carefully and make sure you understand how we are updating the parameters, corresponding to these updates:
$ p(z_{t-1}, z_t | \X, \theta) = \frac{\alpha_{t-1}(z_{t-1}) \beta_t(z_t) A_{z_{t-1}, z_t} B_{z_t, x_t}}{\sum_k \alpha_t(k... | pi3, A3, B3 = baum_welch([
[THE, DOG, WALKED, IN, THE, PARK, END, END],
[THE, CAT, RAN, IN, THE, PARK, END, END]], pi, A, B, 1, trace=True)
print("\n\n")
print_A(A3)
print_B(B3) | handsOn_lecture19_baum-welch-pgm-inference/handsOn_lecture19_baum-welch-pgm-inference-template.ipynb | eecs445-f16/umich-eecs445-f16 | mit |
first just checking that the flattening and reshaping works as expected | test = np.random.randn(11,11,4,100)
test.shape
test_flat = test.flatten()
test_flat.shape
np.savetxt('test.txt', test_flat)
test_back = np.loadtxt('test.txt').reshape((11,11,4,100))
test_back.shape
np.mean(test - test_back) | bin/calculations/human_data/old/saving.ipynb | michaelneuder/image_quality_analysis | mit |
Time Series Database
This notebook demonstrates the persistent behavior of the database.
Initialization
Clear the file system for demonstration purposes. | # database parameters
ts_length = 100
data_dir = '../db_files'
db_name = 'default'
dir_path = data_dir + '/' + db_name + '/'
# clear file system for testing
if not os.path.exists(dir_path):
os.makedirs(dir_path)
filelist = [dir_path + f for f in os.listdir(dir_path)]
for f in filelist:
os.remove(f) | docs/persistence_demo.ipynb | Mynti207/cs207project | mit |
Load the database server. | # when running from the terminal
# python go_server_persistent.py --ts_length 100 --db_name 'demo'
# here we load the server as a subprocess for demonstration purposes
server = subprocess.Popen(['python', '../go_server_persistent.py',
'--ts_length', str(ts_length), '--data_dir', data_dir, '-... | docs/persistence_demo.ipynb | Mynti207/cs207project | mit |
Load the database webserver. | # when running from the terminal
# python go_webserver.py
# here we load the server as a subprocess for demonstration purposes
webserver = subprocess.Popen(['python', '../go_webserver.py'])
time.sleep(5) # make sure it loads completely | docs/persistence_demo.ipynb | Mynti207/cs207project | mit |
Import the web interface and initialize it. | from webserver import *
web_interface = WebInterface() | docs/persistence_demo.ipynb | Mynti207/cs207project | mit |
Generate Data
Let's create some dummy data to aid in our demonstration. You will need to import the timeseries package to work with the TimeSeries format.
Note: the database is persistent, so can store data between sessions, but we will start with an empty database here for demonstration purposes. | from timeseries import *
def tsmaker(m, s, j):
'''
Helper function: randomly generates a time series for testing.
Parameters
----------
m : float
Mean value for generating time series data
s : float
Standard deviation value for generating time series data
j : float
... | docs/persistence_demo.ipynb | Mynti207/cs207project | mit |
Insert Data
Let's start by loading the data into the database, using the REST API web interface. | # check that the database is empty
web_interface.select()
# add stats trigger
web_interface.add_trigger('stats', 'insert_ts', ['mean', 'std'], None)
# insert the time series
for k in tsdict:
web_interface.insert_ts(k, tsdict[k])
# upsert the metadata
for k in tsdict:
web_interface.upsert_meta(k, metadict[k])... | docs/persistence_demo.ipynb | Mynti207/cs207project | mit |
Inspect Data
Let's inspect the data, to make sure that all the previous operations were successful. | # select all database entries; all metadata fields
results = web_interface.select(fields=[])
# we have the right number of database entries
assert len(results) == num_ts
# we have all the right primary keys
assert sorted(results.keys()) == ts_keys
# check that all the time series and metadata matches
for k in tsdict... | docs/persistence_demo.ipynb | Mynti207/cs207project | mit |
Let's generate an additional time series for similarity searches. We'll store the time series and the results of the similarity searches, so that we can compare against them after reloading the database. | _, query = tsmaker(np.random.uniform(low=0.0, high=1.0),
np.random.uniform(low=0.05, high=0.4),
np.random.uniform(low=0.05, high=0.2))
results_vp = web_interface.vp_similarity_search(query, 1)
results_vp
results_isax = web_interface.isax_similarity_search(query)
results_isax | docs/persistence_demo.ipynb | Mynti207/cs207project | mit |
Finally, let's store our iSAX tree representation. | results_tree = web_interface.isax_tree()
print(results_tree) | docs/persistence_demo.ipynb | Mynti207/cs207project | mit |
Terminate and Reload Database
Now that we know that everything is loaded, let's close the database and re-open it. | os.kill(server.pid, signal.SIGINT)
time.sleep(5) # give it time to terminate
os.kill(webserver.pid, signal.SIGINT)
time.sleep(5) # give it time to terminate
web_interface = None
server = subprocess.Popen(['python', '../go_server_persistent.py',
'--ts_length', str(ts_length), '--data_dir', ... | docs/persistence_demo.ipynb | Mynti207/cs207project | mit |
Inspect Data
Let's repeat the previous tests to check whether our persistence architecture worked. | # select all database entries; all metadata fields
results = web_interface.select(fields=[])
# we have the right number of database entries
assert len(results) == num_ts
# we have all the right primary keys
assert sorted(results.keys()) == ts_keys
# check that all the time series and metadata matches
for k in tsdict... | docs/persistence_demo.ipynb | Mynti207/cs207project | mit |
We have successfully reloaded all of the database components from disk! | # terminate processes before exiting
os.kill(server.pid, signal.SIGINT)
time.sleep(5) # give it time to terminate
web_interface = None
webserver.terminate() | docs/persistence_demo.ipynb | Mynti207/cs207project | mit |
Above, we see 5 points. These indicate the centroids of astrocytes. We will use the origin, i.e. $(0,0)$, as the target astrocyte. Thus we will construct the voronoi region around the origin. | plt.scatter((0),(0), color="red")
plt.scatter(xpoints, ypoints)
for pt in zip(xpoints,ypoints):
plt.plot((0,pt[0]),(0,pt[1]),color="black")
plt.xlim((-5,5)); plt.ylim((-5,5)) | Voronoi The Astrocytes.ipynb | mathnathan/notebooks | mit |
The target astrocyte is drawn in red. To help us visualize how the voronoi mesh is constructed we drew a line connecting the target astrocyte to each of its neighbors. Next we will mark the midpoint of each line with a green dot. These will form the vertices of the voronoi mesh. | plt.scatter((0),(0), color="red")
plt.scatter(xpoints, ypoints)
vpts = []
for pt in zip(xpoints,ypoints):
plt.plot((0,pt[0]),(0,pt[1]),color="black")
plt.scatter((pt[0]/2),(pt[1]/2),color="green")
plt.xlim((-5,5)); plt.ylim((-5,5)) | Voronoi The Astrocytes.ipynb | mathnathan/notebooks | mit |
Lastly, by connecting all of these vertices with green lines we will have the voronoi region around the origin marked off by the perimeter of green lines. | plt.scatter((0),(0), color="red")
plt.scatter(xpoints, ypoints)
for i,pt in enumerate(zip(xpoints,ypoints)):
plt.plot((0,pt[0]),(0,pt[1]),color="black")
plt.scatter((pt[0]/2),(pt[1]/2),color="green")
plt.plot((pt[0]/2,xpoints[i-1]/2),(pt[1]/2,ypoints[i-1]/2),color="green")
plt.xlim((-5,5)); plt.ylim((-5,5)) | Voronoi The Astrocytes.ipynb | mathnathan/notebooks | mit |
Questões
1. Rode o mesmo programa nos dados contendo anos de escolaridade (primeira coluna) versus salário (segunda coluna). Baixe os dados aqui. Esse exemplo foi trabalhado em sala de aula em várias ocasiões. Os itens a seguir devem ser respondidos usando esses dados.
RESOLUÇÃO: Arquivo baixado, encontra-se no diretór... | ## Show figure
# @param data Data to show in the graphic.
# @param xlabel Text to be shown in abscissa axis.
# @param ylabel Text to be shown in ordinate axis.
def show_figure(data, xlabel, ylabel):
plt.plot(data)
plt.xlabel(xlabel)
plt.ylabel(ylabel) | Tarefas/Regressão-Linear-Simples-do-Zero/Task 01.ipynb | italoPontes/Machine-learning | lgpl-3.0 |
3. O que acontece com o RSS ao longo das iterações (aumenta ou diminui) se você usar 1000 iterações e um learning_rate (tamanho do passo do gradiente) de 0.001? Por que você acha que isso acontece? | points = genfromtxt("income.csv", delimiter=",")
x = points[:,0]
y = points[:,1]
starting_w0 = 0
starting_w1 = 0
learning_rate = 0.001
iterations_number = 50
[w0, w1, iter_number, rss_total] = gradient_descent_runner(x, y, starting_w0, starting_w1, learning_rate, iterations_number)
show_figure(rss_total, "Iteraction... | Tarefas/Regressão-Linear-Simples-do-Zero/Task 01.ipynb | italoPontes/Machine-learning | lgpl-3.0 |
Com esse gráfico é possível observar que:
Quanto maior o Learning Rate, maior o número de iterações necessárias para se atingir um mesmo erro. | learning_rate = 0.001
iterations_number = 1000
[w0, w1, iter_number, rss_total] = gradient_descent_runner(x, y, starting_w0, starting_w1, learning_rate, iterations_number)
print("RSS na última iteração: %.2f" % rss_total[-1])
iterations_number = 10000
[w0, w1, iter_number, rss_total] = gradient_descent_runner(x, y, st... | Tarefas/Regressão-Linear-Simples-do-Zero/Task 01.ipynb | italoPontes/Machine-learning | lgpl-3.0 |
Ao observar os valores de RSS calculados quando o número de iterações aumenta, é possível observar que o RSS obtido diminui cada vez mais.
4. Teste valores diferentes do número de iterações e learning_rate até que w0 e w1 sejam aproximadamente iguais a -39 e 5 respectivamente. Reporte os valores do número de iterações ... | learning_rate = 0.0025
iterations_number = 20000
[w0, w1, iter_number, rss_total] = gradient_descent_runner(x, y, starting_w0, starting_w1, learning_rate, iterations_number)
print("W0: %.2f" % w0)
print("W1: %.2f" % w1)
print("RSS na última iteração: %.2f" % rss_total[-1]) | Tarefas/Regressão-Linear-Simples-do-Zero/Task 01.ipynb | italoPontes/Machine-learning | lgpl-3.0 |
5. O algoritmo do vídeo usa o número de iterações como critério de parada. Mude o algoritmo para considerar um critério de tolerância que é comparado ao tamanho do gradiente (como no algoritmo dos slides apresentados em sala).
A metodologia aplicada foi a seguinte: quando não se fornece o número de iterações por parâme... | learning_rate = 0.0025
iterations_number = 0
[w0, w1, iter_number, rss_total] = gradient_descent_runner(x, y, starting_w0, starting_w1, learning_rate, iterations_number)
print("W0: %.2f" % w0)
print("W1: %.2f" % w1)
print("RSS na última iteração: %.2f" % rss_total[-1]) | Tarefas/Regressão-Linear-Simples-do-Zero/Task 01.ipynb | italoPontes/Machine-learning | lgpl-3.0 |
6. Ache um valor de tolerância que se aproxime dos valores dos parâmetros do item 4 acima. Que valor foi esse?
O valor utilizado, conforme descrito na questão anterior, foi 0,001. Ou seja, quando o tamanho do gradiente for menor que 0,001, então, o algoritmo entenderá que a aproximação convergiu e terminará o processam... | import time
start_time = time.time()
[w0, w1, iter_number, rss_total] = gradient_descent_runner(x, y, starting_w0, starting_w1, learning_rate, iterations_number)
gradient_time = float(time.time()-start_time)
print("Tempo para calcular os coeficientes pelo gradiente descendente: %.2f s." % gradient_time)
start_time = ... | Tarefas/Regressão-Linear-Simples-do-Zero/Task 01.ipynb | italoPontes/Machine-learning | lgpl-3.0 |
Import pandas
We are using the very handy pandas library for dataframes. | import pandas as pd | oldsitejekyll/markdown_generator/publications.ipynb | manparvesh/manparvesh.github.io | mit |
Import TSV
Pandas makes this easy with the read_csv function. We are using a TSV, so we specify the separator as a tab, or \t.
I found it important to put this data in a tab-separated values format, because there are a lot of commas in this kind of data and comma-separated values can get messed up. However, you can mod... | publications = pd.read_csv("publications.tsv", sep="\t", header=0)
publications
| oldsitejekyll/markdown_generator/publications.ipynb | manparvesh/manparvesh.github.io | mit |
Escape special characters
YAML is very picky about how it takes a valid string, so we are replacing single and double quotes (and ampersands) with their HTML encoded equivilents. This makes them look not so readable in raw format, but they are parsed and rendered nicely. | html_escape_table = {
"&": "&",
'"': """,
"'": "'"
}
def html_escape(text):
"""Produce entities within text."""
return "".join(html_escape_table.get(c,c) for c in text) | oldsitejekyll/markdown_generator/publications.ipynb | manparvesh/manparvesh.github.io | mit |
Creating the markdown files
This is where the heavy lifting is done. This loops through all the rows in the TSV dataframe, then starts to concatentate a big string (md) that contains the markdown for each type. It does the YAML metadata first, then does the description for the individual page. | import os
for row, item in publications.iterrows():
md_filename = str(item.pub_date) + "-" + item.url_slug + ".md"
html_filename = str(item.pub_date) + "-" + item.url_slug
year = item.pub_date[:4]
## YAML variables
md = "---\ntitle: \"" + item.title + '"\n'
md += """collect... | oldsitejekyll/markdown_generator/publications.ipynb | manparvesh/manparvesh.github.io | mit |
These files are in the publications directory, one directory below where we're working from. | !ls ../_publications/
!cat ../_publications/2009-10-01-paper-title-number-1.md | oldsitejekyll/markdown_generator/publications.ipynb | manparvesh/manparvesh.github.io | mit |
Aim
Motive of the notebook is to give a brief overview as to how to use the evolutionary sampling powered ensemble models as part of the EvoML research project.
Will make the notebook more verbose if time permits. Priority will be to showcase the flexible API of the new estimators which encourage research and tinkerin... | from evoml.subsampling import BasicSegmenter_FEMPO, BasicSegmenter_FEGT, BasicSegmenter_FEMPT
df = pd.read_csv('datasets/ozone.csv')
df.head(2)
X, y = df.iloc[:,:-1], df['output']
print(BasicSegmenter_FEGT.__doc__)
from sklearn.tree import DecisionTreeRegressor
clf_dt = DecisionTreeRegressor(max_depth=3)
clf = Bas... | EvoML - Example Usage.ipynb | EvoML/EvoML | gpl-3.0 |
2. Subspacing - sampling in the domain of features - evolving and mutating columns | from evoml.subspacing import FeatureStackerFEGT, FeatureStackerFEMPO
print(FeatureStackerFEGT.__doc__)
clf = FeatureStackerFEGT(ngen=30)
clf.fit(X, y)
clf.score(X, y)
## Get the Hall of Fame individual
hof = clf.segment[0]
sampled_datasets = [eg.get_data() for eg in hof]
[data.columns.tolist() for data in sample... | EvoML - Example Usage.ipynb | EvoML/EvoML | gpl-3.0 |
Set up training data
In the next cell, we set up the training data for this example. For each task we'll be using 50 random points on [0,1), which we evaluate the function on and add Gaussian noise to get the training labels. Note that different inputs are used for each task.
We'll have two functions - a sine function ... | train_x1 = torch.rand(50)
train_x2 = torch.rand(50)
train_y1 = torch.sin(train_x1 * (2 * math.pi)) + torch.randn(train_x1.size()) * 0.2
train_y2 = torch.cos(train_x2 * (2 * math.pi)) + torch.randn(train_x2.size()) * 0.2 | examples/03_Multitask_Exact_GPs/Hadamard_Multitask_GP_Regression.ipynb | jrg365/gpytorch | mit |
Set up a Hadamard multitask model
The model should be somewhat similar to the ExactGP model in the simple regression example.
The differences:
The model takes two input: the inputs (x) and indices. The indices indicate which task the observation is for.
Rather than just using a RBFKernel, we're using that in conjuncti... | class MultitaskGPModel(gpytorch.models.ExactGP):
def __init__(self, train_x, train_y, likelihood):
super(MultitaskGPModel, self).__init__(train_x, train_y, likelihood)
self.mean_module = gpytorch.means.ConstantMean()
self.covar_module = gpytorch.kernels.RBFKernel()
# We lear... | examples/03_Multitask_Exact_GPs/Hadamard_Multitask_GP_Regression.ipynb | jrg365/gpytorch | mit |
Training the model
In the next cell, we handle using Type-II MLE to train the hyperparameters of the Gaussian process.
See the simple regression example for more info on this step. | # this is for running the notebook in our testing framework
import os
smoke_test = ('CI' in os.environ)
training_iterations = 2 if smoke_test else 50
# Find optimal model hyperparameters
model.train()
likelihood.train()
# Use the adam optimizer
optimizer = torch.optim.Adam(model.parameters(), lr=0.1) # Includes Gau... | examples/03_Multitask_Exact_GPs/Hadamard_Multitask_GP_Regression.ipynb | jrg365/gpytorch | mit |
Make predictions with the model | # Set into eval mode
model.eval()
likelihood.eval()
# Initialize plots
f, (y1_ax, y2_ax) = plt.subplots(1, 2, figsize=(8, 3))
# Test points every 0.02 in [0,1]
test_x = torch.linspace(0, 1, 51)
tast_i_task1 = torch.full_like(test_x, dtype=torch.long, fill_value=0)
test_i_task2 = torch.full_like(test_x, dtype=torch.lo... | examples/03_Multitask_Exact_GPs/Hadamard_Multitask_GP_Regression.ipynb | jrg365/gpytorch | mit |
La función
Esta ecuación por Glass y Pasternack (1978) sirve para modelar redes neuronales y de interacción génica.
$$x_{t+1}=\frac{\alpha x_{t}}{1+\beta x_{t}}$$
Donde $\alpha$ y $\beta$ son números positivos y $x_{t}\geq0$. | def g(x, alpha, beta):
assert alpha >= 0 and beta >= 0
return (alpha*x)/(1 + (beta * x))
def plot_cobg(x, alpha, beta):
y = np.linspace(x[0],x[1],300)
g_y = g(y, alpha, beta)
cobweb(lambda x: g(x, alpha, beta), y, g_y)
# configura gráfica interactiva
interact(plot_cobg,
x=widgets.FloatR... | Glass_Pasternack.ipynb | rgarcia-herrera/sistemas-dinamicos | gpl-3.0 |
Búsqueda algebráica de puntos fijos
A continuación sustituiremos f(x) en x reiteradamente hasta obtener la cuarta iterada de f. | # primera iterada
f0 = (alpha*x)/(1+beta*x)
Eq(f(x),f0)
# segunda iterada
# subs-tituye f0 en la x de f0 para generar f1
f1 = simplify(f0.subs(x, f0))
Eq(f(f(x)), f1)
# tercera iterada
f2 = simplify(f1.subs(x, f1))
Eq(f(f(f(x))), f2)
# cuarta iterada
f3 = simplify(f2.subs(x, f2))
Eq(f(f(f(f(x)))), f3)
# puntos fij... | Glass_Pasternack.ipynb | rgarcia-herrera/sistemas-dinamicos | gpl-3.0 |
Punto fijo oscilatorio
Al configurar $$\alpha, \beta$$ de modo que haya un punto fijo la serie de tiempo revela una oscilación entre cero y el punto fijo. | def solve_g(a, b):
y = list(np.linspace(0,float(list(solveset(Eq(f1.subs(alpha, a).subs(beta, b), x), x)).pop()),2))
for t in range(30):
y.append(g(y[t], a, b))
zoom = plt.plot(y)
print("ultimos 15 de la serie:")
pprint(y[-15:])
print("\npuntos fijos:")
return solveset(Eq(f1.subs(a... | Glass_Pasternack.ipynb | rgarcia-herrera/sistemas-dinamicos | gpl-3.0 |
¿Qué pasará con infinitas iteraciones?
Todo parece indicar que la función converge a 1 si $\alpha=1$ y $\beta=1$.
Si no, converge a $\frac{\alpha}{\beta}$ | # con alfa=1 y beta=1
Eq(collect(f3, x), x/(x+1))
def plot_g(x, alpha, beta):
pprint(x)
y = np.linspace(x[0],x[1],300)
g_y = g(y, alpha, beta)
fig1 = plt.plot(y, g_y)
fig1 = plt.plot(y, y, color='red')
plt.axis('equal')
interact(plot_g,
x=widgets.FloatRangeSlider(min=0, max=30, step... | Glass_Pasternack.ipynb | rgarcia-herrera/sistemas-dinamicos | gpl-3.0 |
<a id='step1'></a>
1. Making Collectors for each number panel and xgap case | x = 2
y = 1
ygap = 0.1524 # m = 6 in
zgap = 0.002 # m, veyr little gap to torquetube.
tubeParams = {'diameter':0.15,
'tubetype':'square',
'material':'Metal_Grey',
'axisofrotation':True,
'visible': True}
ft2m = 0.3048
xgaps = [3, 4, 6, 9, 12, 15, 18, 21]
numpane... | docs/tutorials/16 - AgriPV - 3-up and 4-up collector optimization.ipynb | NREL/bifacial_radiance | bsd-3-clause |
<a id='step2'></a>
2. Build the Scene so it can be viewed with rvu | xgaps = np.round(np.array([3, 4, 6, 9, 12, 15, 18, 21]) * ft2m,1)
numpanelss = [3, 4]
sensorsxs = np.array(list(range(0, 201)))
# Select CASE:
xgap = np.round(xgaps[-1],1)
numpanels = 4
# All the rest
ft2m = 0.3048
hub_height = 8.0 * ft2m
y = 1
pitch = 0.001 # If I recall, it doesn't like when pitch is 0 even if ... | docs/tutorials/16 - AgriPV - 3-up and 4-up collector optimization.ipynb | NREL/bifacial_radiance | bsd-3-clause |
To View the generated Scene, you can navigate to the testfolder on a terminal and use:
<b>front view:<b>
rvu -vf views\front.vp -e .0265652 -vp 2 -21 2.5 -vd 0 1 0 makemod.oct
<b> top view: </b>
rvu -vf views\front.vp -e .0265652 -vp 5 0 70 -vd 0 0.0001 -1 makemod.oct
Or run it directly from Jupyter by removing the... |
## Comment the ! line below to run rvu from the Jupyter notebook instead of your terminal.
## Simulation will stop until you close the rvu window
#!rvu -vf views\front.vp -e .0265652 -vp 2 -21 2.5 -vd 0 1 0 makemod.oct
#!rvu -vf views\front.vp -e .0265652 -vp 5 0 70 -vd 0 0.0001 -1 makemod.oct
| docs/tutorials/16 - AgriPV - 3-up and 4-up collector optimization.ipynb | NREL/bifacial_radiance | bsd-3-clause |
Load cloudmlmagic extension | %load_ext cloudmlmagic | examples/Keras_Fine_Tuning.ipynb | hayatoy/cloudml-magic | mit |
Initialize and setup ML Engine parameters.
<font color="red">Change PROJECTID and BUCKET</font>
Following dict will be written in setup.py of your package,
so list up neccesary packages of your code. | %%ml_init -projectId PROJECTID -bucket BUCKET -scaleTier BASIC_GPU -region asia-east1 -runtimeVersion 1.2
{'install_requires': ['keras', 'h5py', 'Pillow']} | examples/Keras_Fine_Tuning.ipynb | hayatoy/cloudml-magic | mit |
Load InceptionV3 model | %%ml_code
from keras.applications.inception_v3 import InceptionV3
model = InceptionV3(weights='imagenet') | examples/Keras_Fine_Tuning.ipynb | hayatoy/cloudml-magic | mit |
Load dataset | %%ml_code
from keras.preprocessing import image
from keras.applications.inception_v3 import preprocess_input, decode_predictions
from io import BytesIO
import numpy as np
import pandas as pd
import requests
url = 'https://github.com/hayatoy/deep-learning-datasets/releases/download/v0.1/tl_opera_capitol.npz'
response ... | examples/Keras_Fine_Tuning.ipynb | hayatoy/cloudml-magic | mit |
Split dataset for train and test | %%ml_code
from keras.utils import np_utils
from sklearn.model_selection import train_test_split
X_dataset = preprocess_input(X_dataset)
y_dataset = np_utils.to_categorical(y_dataset)
X_train, X_test, y_train, y_test = train_test_split(
X_dataset, y_dataset, test_size=0.2, random_state=42) | examples/Keras_Fine_Tuning.ipynb | hayatoy/cloudml-magic | mit |
The code cell above won't be included in the package being deployed on ML Engine.
Just to clarify that normal InceptionV3 model cannot predict correctly with the Opera/Capitol dataset. | x = X_dataset[0]
x = np.expand_dims(x, axis=0)
preds = model.predict(x)
print('Predicted:')
for p in decode_predictions(preds, top=5)[0]:
print("Score {}, Label {}".format(p[2], p[1])) | examples/Keras_Fine_Tuning.ipynb | hayatoy/cloudml-magic | mit |
Visualize last layers of InceptionV3 | pd.DataFrame(model.layers).tail()
%ml_code
from keras.models import Model
# Intermediate layer
intermediate_layer_model = Model(inputs=model.input, outputs=model.layers[311].output) | examples/Keras_Fine_Tuning.ipynb | hayatoy/cloudml-magic | mit |
Extract intermediate features | x = np.expand_dims(X_dataset[0], axis=0)
feature = intermediate_layer_model.predict(x)
pd.DataFrame(feature.reshape(-1,1)).plot(figsize=(12, 3)) | examples/Keras_Fine_Tuning.ipynb | hayatoy/cloudml-magic | mit |
Append dense layer at the last | %%ml_code
from keras.layers import Dense
# Append dense layer
x = intermediate_layer_model.output
x = Dense(1024, activation='relu')(x)
predictions = Dense(2, activation='softmax')(x)
# Transfer learning model, all layers are trainable at this moment
transfer_model = Model(inputs=intermediate_layer_model.input, outp... | examples/Keras_Fine_Tuning.ipynb | hayatoy/cloudml-magic | mit |
Create Model and Version for Online Prediction | # !gcloud ml-engine models create OperaCapitol
!gcloud ml-engine versions create v1 --model OperaCapitol --runtime-version 1.2 --origin gs://BUCKET/keras-mlengine/savedmodel | examples/Keras_Fine_Tuning.ipynb | hayatoy/cloudml-magic | mit |
Let's classify this image! This must be class 0..
<img src="opera.jpg"> | from oauth2client.client import GoogleCredentials
from googleapiclient import discovery
from googleapiclient import errors
PROJECTID = 'PROJECTID'
projectID = 'projects/{}'.format(PROJECTID)
modelName = 'OperaCapitol'
modelID = '{}/models/{}'.format(projectID, modelName)
credentials = GoogleCredentials.get_applicatio... | examples/Keras_Fine_Tuning.ipynb | hayatoy/cloudml-magic | mit |
Loading the EIA Data, the path may need to be updated...
This will take a few minutes to run. | #Iterate through the directory to find all the files to import
#Modified so that it also works on macs
path = os.path.join('EIA Data', '923-No_Header')
full_path = os.path.join(path, '*.*')
eiaNames = os.listdir(path)
#Rename the keys for easier merging later
fileNameMap = {'EIA923 SCHEDULES 2_3_4_5 Final 2010.xls':... | Raw Data/Merging EPA and EIA.ipynb | gschivley/ERCOT_power | mit |
The excel documents have different column names so we need to standardize them all | #Dict of values to replace to standardize column names across all dataframes
monthDict = {"JANUARY":"JAN",
"FEBRUARY":"FEB",
"MARCH":"MAR",
"APRIL":"APR",
"MAY":"MAY",
"JUNE":"JUN",
"JULY":"JUL",
"AUGUST":"AUG",
"SEPTEMBER":"SEP",
... | Raw Data/Merging EPA and EIA.ipynb | gschivley/ERCOT_power | mit |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.