markdown stringlengths 0 37k | code stringlengths 1 33.3k | path stringlengths 8 215 | repo_name stringlengths 6 77 | license stringclasses 15
values |
|---|---|---|---|---|
scipy
implementation using scipy.ndimage.generic_filter—the custom callback function is just-in-time compiled by numba | @jit(nopython=True)
def filter_denoise(neighborhood):
if neighborhood.mean() < 10:
return neighborhood.min()
else:
return neighborhood[13]
def denoise_scipy(a, b):
for channel in range(2):
b[channel] = generic_filter(input=a[channel], function=filter_denoise,
... | profiling/Denoise algorithm.ipynb | jacobdein/alpine-soundscapes | mit |
numba
implementation of a universal function via numba.guvectorize | # just removed return statement
def denoise_guvectorize(a, b):
for channel in range(2):
for f_band in range(4, a.shape[1] - 4):
for t_step in range(1, a.shape[2] - 1):
neighborhood = a[channel, f_band - 4:f_band + 5, t_step - 1:t_step + 2]
if neighborhood.mean() <... | profiling/Denoise algorithm.ipynb | jacobdein/alpine-soundscapes | mit |
serial version | denoise_numba = guvectorize('float64[:,:,:], float64[:,:,:]', '(c,f,t)->(c,f,t)',
nopython=True)(denoise_guvectorize) | profiling/Denoise algorithm.ipynb | jacobdein/alpine-soundscapes | mit |
parallel version | denoise_parallel = guvectorize('float64[:,:,:], float64[:,:,:]', '(c,f,t)->(c,f,t)',
nopython=True, target='parallel')(denoise_guvectorize) | profiling/Denoise algorithm.ipynb | jacobdein/alpine-soundscapes | mit |
check results
test the implementations on a randomly generated dataset and verfiy that all the results are the same | size = 100
data = np.random.rand(2, size, int(size*1.5))
data[:, int(size/4):int(size/2), int(size/4):int(size/2)] = 27
result_python = denoise(data, np.zeros_like(data))
result_scipy = denoise_scipy(data, np.zeros_like(data))
result_numba = denoise_numba(data, np.zeros_like(data))
result_parallel = denoise_parallel(d... | profiling/Denoise algorithm.ipynb | jacobdein/alpine-soundscapes | mit |
check if the different implementations produce the same result | assert np.allclose(result_python, result_scipy) and np.allclose(result_python, result_numba) | profiling/Denoise algorithm.ipynb | jacobdein/alpine-soundscapes | mit |
plot results | fig, ax = plt.subplots(2,2)
fig.set_figheight(8)
fig.set_figwidth(12)
im1 = ax[0, 0].imshow(data[0], cmap='viridis', interpolation='none', vmax=1)
t1 = ax[0, 0].set_title('data')
im2 = ax[0, 1].imshow(result_python[0], cmap='viridis', interpolation='none', vmax=1)
t1 = ax[0, 1].set_title('pure python')
im3 = ax[1, 0].i... | profiling/Denoise algorithm.ipynb | jacobdein/alpine-soundscapes | mit |
profile for different data sizes
time the different implementations on different dataset sizes | sizes = [30, 50, 100, 200, 400, 800, 1600]
progress_bar = pyprind.ProgBar(iterations=len(sizes), track_time=True, stream=1, monitor=True)
t_python = np.empty_like(sizes, dtype=np.float64)
t_scipy = np.empty_like(sizes, dtype=np.float64)
t_numba = np.empty_like(sizes, dtype=np.float64)
t_parallel = np.empty_like(sizes,... | profiling/Denoise algorithm.ipynb | jacobdein/alpine-soundscapes | mit |
plot profile results | fig, ax = plt.subplots(figsize=(15,5))
p1 = ax.loglog(sizes, t_python, color='black', marker='.', label='python')
p2 = ax.loglog(sizes, t_scipy, color='blue', marker='.', label='scipy')
p3 = ax.loglog(sizes, t_numba, color='green', marker='.', label='numba')
p4 = ax.loglog(sizes, t_parallel, color='red', marker='.', la... | profiling/Denoise algorithm.ipynb | jacobdein/alpine-soundscapes | mit |
Download the sequence data
Sequence data for this study are archived on the NCBI sequence read archive (SRA). Below I read in SraRunTable.txt for this project which contains all of the information we need to download the data.
SRA link: http://trace.ncbi.nlm.nih.gov/Traces/study/?acc=SRP021469 | %%bash
## make a new directory for this analysis
mkdir -p empirical_10/fastq/ | emp_nb_Pedicularis.ipynb | dereneaton/RADmissing | mit |
For each ERS (individuals) get all of the ERR (sequence file accessions). | ## IPython code
import pandas as pd
import numpy as np
import urllib2
import os
## open the SRA run table from github url
url = "https://raw.githubusercontent.com/"+\
"dereneaton/RADmissing/master/empirical_10_SraRunTable.txt"
intable = urllib2.urlopen(url)
indata = pd.read_table(intable, sep="\t")
## print fir... | emp_nb_Pedicularis.ipynb | dereneaton/RADmissing | mit |
Here we pass the SRR number and the sample name to the wget_download function so that the files are saved with their sample names. | for ID, SRR in zip(indata.Sample_Name_s, indata.Run_s):
wget_download(SRR, "empirical_10/fastq/", ID)
%%bash
## convert sra files to fastq using fastq-dump tool
## output as gzipped into the fastq directory
fastq-dump --gzip -O empirical_10/fastq/ empirical_10/fastq/*.sra
## remove .sra files
rm empirical_10/fast... | emp_nb_Pedicularis.ipynb | dereneaton/RADmissing | mit |
Note:
The data here are from Illumina Casava <1.8, so the phred scores are offset by 64 instead of 33, so we use that in the params file below. | %%bash
## substitute new parameters into file
sed -i '/## 1. /c\empirical_10/ ## 1. working directory ' params.txt
sed -i '/## 6. /c\TGCAG ## 6. cutters ' params.txt
sed -i '/## 7. /c\20 ## 7. N processors ' params.txt
sed -i '/## 9. /c\6 ## 9. NQu... | emp_nb_Pedicularis.ipynb | dereneaton/RADmissing | mit |
Assemble in pyrad | %%bash
pyrad -p params.txt -s 234567 >> log.txt 2>&1
%%bash
sed -i '/## 12./c\2 ## 12. MinCov ' params.txt
sed -i '/## 14./c\empirical_10_m2 ## 14. output name ' params.txt
%%bash
pyrad -p params.txt -s 7 >> log.txt 2>&1 | emp_nb_Pedicularis.ipynb | dereneaton/RADmissing | mit |
Results
We are interested in the relationship between the amount of input (raw) data between any two samples, the average coverage they recover when clustered together, and the phylogenetic distances separating samples.
Raw data amounts
The average number of raw reads per sample is 1.36M. | import pandas as pd
## read in the data
sdat = pd.read_table("empirical_10/stats/s2.rawedit.txt", header=0, nrows=14)
## print summary stats
print sdat["passed.total"].describe()
## find which sample has the most raw data
maxraw = sdat["passed.total"].max()
print "\nmost raw data in sample:"
print sdat['sample '][sda... | emp_nb_Pedicularis.ipynb | dereneaton/RADmissing | mit |
Look at distributions of coverage
pyrad v.3.0.63 outputs depth information for each sample which I read in here and plot. First let's ask which sample has the highest depth of coverage. The std of coverages is pretty low in this data set compared to several others. | ## read in the s3 results
sdat = pd.read_table("empirical_10/stats/s3.clusters.txt", header=0, nrows=14)
## print summary stats
print "summary of means\n=================="
print sdat['dpt.me'].describe()
## print summary stats
print "\nsummary of std\n=================="
print sdat['dpt.sd'].describe()
## print sum... | emp_nb_Pedicularis.ipynb | dereneaton/RADmissing | mit |
Plot the coverage for the sample with highest mean coverage
Green shows the loci that were discarded and orange the loci that were retained. The majority of data were discarded for being too low of coverage. | import toyplot
import toyplot.svg
import numpy as np
## read in the depth information for this sample
with open("empirical_10/clust.85/38362_rex.depths", 'rb') as indat:
depths = np.array(indat.read().strip().split(","), dtype=int)
## make a barplot in Toyplot
canvas = toyplot.Canvas(width=350, height=300)
ax... | emp_nb_Pedicularis.ipynb | dereneaton/RADmissing | mit |
Print final stats table | cat empirical_10/stats/empirical_10_m4.stats
%%bash
head -n 10 empirical_10/stats/empirical_10_m2.stats | emp_nb_Pedicularis.ipynb | dereneaton/RADmissing | mit |
Infer ML phylogeny in raxml as an unrooted tree | %%bash
## raxml argumement w/ ...
raxmlHPC-PTHREADS-AVX -f a -m GTRGAMMA -N 100 -x 12345 -p 12345 -T 20 \
-w /home/deren/Documents/RADmissing/empirical_10/ \
-n empirical_10_m4 -s empirical_10/outfiles/empirical_10_m4.phy
%%bash
## raxml argumement w/ ... | emp_nb_Pedicularis.ipynb | dereneaton/RADmissing | mit |
Plot the tree in R using ape | %load_ext rpy2.ipython
%%R -h 800 -w 800
library(ape)
tre <- read.tree("empirical_10/RAxML_bipartitions.empirical_10_m4")
ltre <- ladderize(tre)
par(mfrow=c(1,2))
plot(ltre, use.edge.length=F)
nodelabels(ltre$node.label)
plot(ltre, type='u') | emp_nb_Pedicularis.ipynb | dereneaton/RADmissing | mit |
Using simple normalisation: f(z)/f(z0) (equation 2 from first draft of GZH paper)
$\frac{f}{f_{0}}=1 - {\zeta} * (z-z_{0})$
$\zeta = constant$ | fit_and_plot(x, yn, mu, fzeta_lin_mu_none, 2) | python/notebooks/zeta_models_compared.ipynb | willettk/gzhubble | mit |
$\frac{f}{f_{0}}=1 - {\zeta} * (z-z_{0})$
$\zeta = \zeta[0]+\zeta[1] * \mu$ | fit_and_plot(x, yn, mu, fzeta_lin_mu_lin, 3) | python/notebooks/zeta_models_compared.ipynb | willettk/gzhubble | mit |
$\frac{f}{f_{0}}=e^{\frac{-(z-z_0)}{\zeta}}$
$\zeta = constant$ | fit_and_plot(x, yn, mu, fzeta_exp_mu_none, 2) | python/notebooks/zeta_models_compared.ipynb | willettk/gzhubble | mit |
$\frac{f}{f_{0}}=e^{\frac{-(z-z_0)}{\zeta}}$
$\zeta = \zeta[0]+\zeta[1] * \mu$ | p = fit_and_plot(x, yn, mu, fzeta_exp_mu_lin, 3) | python/notebooks/zeta_models_compared.ipynb | willettk/gzhubble | mit |
$\frac{f}{f_{0}}= 1 - \zeta_{a}(z-z_{0}) + \zeta_{b}(z-z_{0})^2$
$\zeta_{a}, \zeta_{b} = constant$ | fit_and_plot(x, yn, mu, fzeta_qud_mu_none, 3) | python/notebooks/zeta_models_compared.ipynb | willettk/gzhubble | mit |
$\frac{f}{f_{0}}= 1 - \zeta_{a}(z-z_{0}) + \zeta_{b}(z-z_{0})^2$
$\zeta_{a} = \zeta_{a}[0] + \zeta_{a}[1] * \mu $
$\zeta_{b} = \zeta_{b}[0] + \zeta_{b}[1] * \mu $ | fit_and_plot(x, yn, mu, fzeta_qud_mu_lin, 5) | python/notebooks/zeta_models_compared.ipynb | willettk/gzhubble | mit |
$\frac{f}{f_{0}}= 1 - \zeta_{a}(z-z_{0}) + \zeta_{b}(z-z_{0})^2 + \zeta_{c}*(z-z_{0})^3$
$\zeta_{a}, \zeta_{b}, \zeta_{c} = constant$ | fit_and_plot(x, yn, mu, fzeta_cub_mu_none, 4) | python/notebooks/zeta_models_compared.ipynb | willettk/gzhubble | mit |
$\frac{f}{f_{0}}= 1 - \zeta_{a}(z-z_{0}) + \zeta_{b}(z-z_{0})^2 + \zeta_{c}*(z-z_{0})^3$
$\zeta_{a} = \zeta_{a}[0] + \zeta_{a}[1] * \mu $
$\zeta_{b} = \zeta_{b}[0] + \zeta_{b}[1] * \mu $
$\zeta_{c} = \zeta_{c}[0] + \zeta_{c}[1] * \mu $ | fit_and_plot(x, yn, mu, fzeta_cub_mu_lin, 7) | python/notebooks/zeta_models_compared.ipynb | willettk/gzhubble | mit |
Using alternative normalisation, as in eqn. 4: (f(z0)-1) / (f(z)-1)
$\frac{1-f_{0}}{1-f}=1 - {\zeta} * (z-z_{0})$
$\zeta = constant$ | fit_and_plot(x, ym, mu, fzeta_lin_mu_none, 2) | python/notebooks/zeta_models_compared.ipynb | willettk/gzhubble | mit |
$\frac{1-f_{0}}{1-f}=1 - {\zeta} * (z-z_{0})$
$\zeta = \zeta[0]+\zeta[1] * \mu$ | fit_and_plot(x, ym, mu, fzeta_lin_mu_lin, 3) | python/notebooks/zeta_models_compared.ipynb | willettk/gzhubble | mit |
$\frac{1-f_{0}}{1-f}=e^{\frac{-(z-z_0)}{\zeta}}$
$\zeta = constant$ | fit_and_plot(x, ym, mu, fzeta_exp_mu_none, 2) | python/notebooks/zeta_models_compared.ipynb | willettk/gzhubble | mit |
$\frac{1-f_{0}}{1-f}=e^{\frac{-(z-z_0)}{\zeta}}$
$\zeta = \zeta[0]+\zeta[1] * \mu$ | fit_and_plot(x, ym, mu, fzeta_exp_mu_lin, 3) | python/notebooks/zeta_models_compared.ipynb | willettk/gzhubble | mit |
$\frac{1-f_{0}}{1-f}= 1 - \zeta_{a}(z-z_{0}) + \zeta_{b}(z-z_{0})^2$
$\zeta_{a}, \zeta_{b} = constant$ | fit_and_plot(x, ym, mu, fzeta_qud_mu_none, 3) | python/notebooks/zeta_models_compared.ipynb | willettk/gzhubble | mit |
$\frac{1-f_{0}}{1-f}= 1 - \zeta_{a}(z-z_{0}) + \zeta_{b}(z-z_{0})^2$
$\zeta_{a} = \zeta_{a}[0] + \zeta_{a}[1] * \mu $
$\zeta_{b} = \zeta_{b}[0] + \zeta_{b}[1] * \mu $ | fit_and_plot(x, ym, mu, fzeta_qud_mu_lin, 5) | python/notebooks/zeta_models_compared.ipynb | willettk/gzhubble | mit |
$\frac{1-f_{0}}{1-f}= 1 - \zeta_{a}(z-z_{0}) + \zeta_{b}(z-z_{0})^2 + \zeta_{c}*(z-z_{0})^3$
$\zeta_{a}, \zeta_{b}, \zeta_{c} = constant$ | fit_and_plot(x, ym, mu, fzeta_cub_mu_none, 4) | python/notebooks/zeta_models_compared.ipynb | willettk/gzhubble | mit |
$\frac{1-f_{0}}{1-f}= 1 - \zeta_{a}(z-z_{0}) + \zeta_{b}(z-z_{0})^2 + \zeta_{c}*(z-z_{0})^3$
$\zeta_{a} = \zeta_{a}[0] + \zeta_{a}[1] * \mu $
$\zeta_{b} = \zeta_{b}[0] + \zeta_{b}[1] * \mu $
$\zeta_{c} = \zeta_{c}[0] + \zeta_{c}[1] * \mu $ | fit_and_plot(x, ym, mu, fzeta_cub_mu_lin, 7) | python/notebooks/zeta_models_compared.ipynb | willettk/gzhubble | mit |
Plotting the phase diagram
To plot a phase diagram, we send our phase diagram object into the PDPlotter class. | #Let's show all phases, including unstable ones
plotter = PDPlotter(pd, show_unstable=0.2, backend="matplotlib")
plotter.show() | notebooks/2013-01-01-Plotting and Analyzing a Phase Diagram using the Materials API.ipynb | materialsvirtuallab/matgenb | bsd-3-clause |
Calculating energy above hull and other phase equilibria properties | import collections
data = collections.defaultdict(list)
for e in entries:
decomp, ehull = pd.get_decomp_and_e_above_hull(e)
data["Materials ID"].append(e.entry_id)
data["Composition"].append(e.composition.reduced_formula)
data["Ehull"].append(ehull)
data["Decomposition"].append(" + ".join(["%.2... | notebooks/2013-01-01-Plotting and Analyzing a Phase Diagram using the Materials API.ipynb | materialsvirtuallab/matgenb | bsd-3-clause |
USE-CASE: Testing Proportions
Is coin biased ?
We toss coin 250 times, 140 heads, 120 tails. | # se have:
n_h = 140
n_t = 110
observations = (n_h, n_t)
n_observations = n_h + n_t
print observations, n_observations,
# We define the null hypothesis and the test statistic
def run_null_hypothesis(n_observations):
"""the model of Null hypothesis"""
sample = [random.choice('HT') for _ in range(n_observation... | core/Hypothesis_Testing.ipynb | tsarouch/python_minutes | gpl-2.0 |
In the example above, like most of what will follow, we used the MC way to evaluate the p-value.
Nevertheless, in many cases we can analytically, with the frequentist approach have an evaluation of the p-value.
Below is show the way of getting a p-value using the Probability Mass Function (pmf) of the binomial distribu... | p = 0
for i in range(140, 250):
p += stats.distributions.binom.pmf(k, 250, 0.5)
pval = 1-p
print "The p-value using the frequentist approach is: " , pval | core/Hypothesis_Testing.ipynb | tsarouch/python_minutes | gpl-2.0 |
Is dice crooked ?
we have the frequencies {1:8, 2:9, 3:19, 4:5, 6:8, 6:11} | observations = {1:8, 2:9, 3:19, 4:5, 5:8, 6:11}
observations_frequencies = np.array(observations.values())
n_dice_drops = np.sum(observations_frequencies)
print n_dice_drops
def run_null_hypothesis(n_dice_drops):
"""the model of Null hypothesis"""
dice_values = [1, 2, 3, 4, 5, 6]
rolls = np.random.choice(... | core/Hypothesis_Testing.ipynb | tsarouch/python_minutes | gpl-2.0 |
USE-CASE: Testing Difference in Means | d1 = np.random.normal(38.601, 1.42, 1000)
d2 = np.random.normal(38.523, 1.42, 1000)
plt.figure(1)
plt.subplot(211)
count, bins, ignored = plt.hist(d1, 30, normed=True)
plt.figure(1)
plt.subplot(211)
count, bins, ignored = plt.hist(d2, 30, normed=True)
# plt.plot(bins, 1/(sigma * np.sqrt(2 * np.pi)) *
# ... | core/Hypothesis_Testing.ipynb | tsarouch/python_minutes | gpl-2.0 |
USE-CASE: Testing a Correlation | data = np.random.multivariate_normal([0, 0], [[1, .75],[.75, 1]], 1000)
x = data[:, 0]
y = data[:, 1]
plt.scatter(x, y)
# we can make the null hypothesis model just by shuffling the data of one variable
x2 = x.copy()
np.random.shuffle(x2)
plt.scatter(x2, y)
def run_null_hypothesis(x, y):
"""the model of Null hypo... | core/Hypothesis_Testing.ipynb | tsarouch/python_minutes | gpl-2.0 |
USE-CASE: Testing Proportions with chi2 test
Above we used total deviation as test statistic Sum(abs(observed-expected))
It is more common to use chi2 statistic. Sum((observed-expected)^2/expected)
Lets see how what results we get having chi2 statistic | observations = {1:8, 2:9, 3:19, 4:5, 5:8, 6:11}
observations_frequencies = np.array(observations.values())
n_dice_drops = np.sum(observations_frequencies)
print n_dice_drops
def run_null_hypothesis(n_dice_drops):
"""the model of Null hypothesis"""
dice_values = [1, 2, 3, 4, 5, 6]
rolls = np.random.choice(... | core/Hypothesis_Testing.ipynb | tsarouch/python_minutes | gpl-2.0 |
we see that the p-value is smaller using the chi2 statistic as test statistic.
=> This is very important point since we see that the chioice of t-statistic affects quite a lot the p-value
USE-CASE: Testing Structures in Histograms
e.g.understand if we have signal over background | # Lets say we have already a histogram with the bins values below:
x_obs = {1:1, 2:2, 3:2, 4:0, 5:3, 6:1, 7:1, 8:2, 9:5, 10:6,
11:1, 12:0, 13:1, 14:2, 15:1, 16:3, 17:1, 18:0, 19:1, 20:0}
x_bgr = {1:1.2, 2:1.8, 3:1.8, 4:1.9, 5:1.9, 6:2, 7:2, 8:2, 9:1.8, 10:1.8,
11:1.7, 12:1.7, 13:1.7, 14:1.6, 15:1.6,... | core/Hypothesis_Testing.ipynb | tsarouch/python_minutes | gpl-2.0 |
lets focus only in the bin 9 with signal value = 5
How likelie it is to find nobs = 5 while a backgronud is 1.8 ?
The number of entries $n_s$ in a bar can be treated as a Poisson variable with mean $\nu_s$.
In this scenario we can calculate the p-value as
$P(n>= n_{obs}) = \Sigma_{n=n_{obs}}^{\infty} pmf_{poisson}(n;... | from scipy import stats
pmf_values = []
N_obs = 5
N_bgr = 1.8
for i in range(0, N_obs-1):
pmf_values.append(stats.distributions.poisson.pmf(i, N_bgr))
pval = 1-np.sum(pmf_values)
print 'The p-value is ', pval | core/Hypothesis_Testing.ipynb | tsarouch/python_minutes | gpl-2.0 |
a point to keep in mind is that the background comes with uncertainty so we eventually have a range of p-values
In principle we can apply the procedure above to the number of entries in a subset of bins. E.g. in the two bings with large peak we have $n_{obs}=11$ with expected $\nu_b=3.2$. | from scipy import stats
pmf_values = []
N_obs = 11
N_bgr = 3.2
for i in range(0, N_obs-1):
pmf_values.append(stats.distributions.poisson.pmf(i, N_bgr))
pval = 1-np.sum(pmf_values)
print 'The p-value is ', pval | core/Hypothesis_Testing.ipynb | tsarouch/python_minutes | gpl-2.0 |
In Caffe models get specified in separate protobuf files.
Additionally a solver has to be specified, that determines training parameters.
Instantiate the solver and train the network. | solver = caffe.SGDSolver('mnist_solver.prototxt')
solver.net.forward()
niter = 2500
test_interval = 100
# losses will also be stored in the log
train_loss = np.zeros(niter)
test_acc = np.zeros(int(np.ceil(niter / test_interval)))
output = np.zeros((niter, 8, 10))
# the main solver loop
for it in range(niter):
sol... | notebooks/caffe/train.ipynb | Petr-By/qtpyvis | mit |
The weights are saved in a .caffemodel file. | solver.net.save('mnist.caffemodel') | notebooks/caffe/train.ipynb | Petr-By/qtpyvis | mit |
Este pequeño script muestra algunos aspectos importantes de la sintaxis de Python.
Comentarios
Los comentarios en Python empiezan con un "pound", "hash" o numeral # y cualquier cosa que lo siga hasta el final de la línea es ignorada por el intérprete. Es decir, pueden tener comentarios que toman toda la línea, o sólo p... | print(2*(3+4))
print(2*3+4)
print((2*3)+4) | clases/02-Sintaxis-de-Python.ipynb | leoferres/prograUDD | mit |
Los parentesis también se usan para pasar parámetros a una función cuando se llama. En el siguiente snippet de código, la función print() se usa para mostrar, por ej, los contenidos de una variable. La función se "llama" con un par de parentesis con los argumentos de la función adentro. | x = 3
print('first value:', x)
print('second value:', 2) | clases/02-Sintaxis-de-Python.ipynb | leoferres/prograUDD | mit |
Algunas funciones se llaman sin argumentos y actuan sobre el objeto que evalúan. Los parentesis deben ser usados igual, aunque la función tenga argumentos. | L = [4,2,3,1]
L.sort()
print(L) | clases/02-Sintaxis-de-Python.ipynb | leoferres/prograUDD | mit |
Make some data | n = 1000
p = 10
X = np.random.standard_normal((n,p))
X.shape
A = np.random.random((p,1))
A
y = X @ A
y.shape | notebooks/linear_model.ipynb | cbare/Etudes | apache-2.0 |
Too easy | model = LinearRegression().fit(X, y)
model
X_test = np.random.standard_normal((n,p))
y_test = X_test @ A
from sklearn.metrics import r2_score
y_pred = model.predict(X_test)
r2_score(y_test, y_pred)
plt.scatter(y_test, y_pred, color='#3033ff30')
plt.show() | notebooks/linear_model.ipynb | cbare/Etudes | apache-2.0 |
Adding Noise | noise = 1/2
X_train = X + np.random.normal(loc=0, scale=noise, size=(n,p))
y_train = y + np.random.normal(loc=0, scale=noise, size=(n,1))
model = LinearRegression().fit(X_train, y_train)
X_test_noisy = X_test + np.random.normal(loc=0, scale=noise, size=(n,p))
y_test_noisy = y_test + np.random.normal(loc=0, scale=noi... | notebooks/linear_model.ipynb | cbare/Etudes | apache-2.0 |
Transformed features
Let's make the problem harder. Let's say there are 10 true features that are linearly related with our target variable. We don't necessarily get to observe those, but we can measure 10 other features. These might be combinations of the original features with more or less noise added. Some variable ... | def tr(v, extra_noise):
a,b,c,d,e,f,g,h,i,j = v
super_noisy = np.random.normal(loc=0, scale=extra_noise, size=None)
return (a+b, b*c, (c + d + e)/3, d + i/10, e, f, g+super_noisy, h + i/5, h + c/3, 0)
noise = 1/5
X_tr_train = np.apply_along_axis(tr, axis=1, arr=X, extra_noise=2) + np.random.normal(loc=0, s... | notebooks/linear_model.ipynb | cbare/Etudes | apache-2.0 |
1.1 Required Module
numpy: NumPy is the fundamental package for scientific computing in Python.
pytorch: End-to-end deep learning platform.
torchvision: This package consists of popular datasets, model architectures, and common image transformations for computer vision.
tensorflow: An open source machine learning frame... | # Load all necessary modules here, for clearness
import torch
import numpy as np
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
# from torchvision.datasets import MNIST
import torchvision
from torchvision import transforms
from torch.optim import lr_scheduler
# from tensorboardX impor... | Homework/Principles of Artificial Neural Networks/Week 4 Training Issues 1/TrainingIssues.ipynb | MegaShow/college-programming | mit |
2. Classfication Model
Ww would define a simple Convolutional Neural Network to classify MNIST
2.1 Short indroduction of MNIST
The MNIST database (Modified National Institute of Standards and Technology database) is a large database of handwritten digits that is commonly used for training various image processing syste... | class FeedForwardNeuralNetwork(nn.Module):
"""
Inputs Linear/Function Output
[128, 1, 28, 28] -> Linear(28*28, 100) -> [128, 100] # first hidden lyaer
-> ReLU -> [128, 100] # relu activation function, may sigmoid
-> Linear... | Homework/Principles of Artificial Neural Networks/Week 4 Training Issues 1/TrainingIssues.ipynb | MegaShow/college-programming | mit |
3. Training
We would define training function here. Additionally, hyper-parameters, loss function, metric would be included here too.
3.1 Pre-set hyper-parameters
setting hyperparameters like below
hyper paprameters include following part
learning rate: usually we start from a quite bigger lr like 1e-1, 1e-2, 1e-3, a... | ### Hyper parameters
batch_size = 128 # batch size is 128
n_epochs = 5 # train for 5 epochs
learning_rate = 0.01 # learning rate is 0.01
input_size = 28*28 # input image has size 28x28
hidden_size = 100 # hidden neurons is 100 for each layer
output_size = 10 # classes of prediction
l2_norm = 0 # not to use l2 penalty
... | Homework/Principles of Artificial Neural Networks/Week 4 Training Issues 1/TrainingIssues.ipynb | MegaShow/college-programming | mit |
3.2 Initialize model parameters
Pytorch provide default initialization (uniform intialization) for linear layer. But there is still some useful intialization method.
Read more about initialization from this link
torch.nn.init.normal_
torch.nn.init.uniform_
torch.nn.init.constant_
torch.nn.init.eye_
torc... | def show_weight_bias(model):
"""Show some weights and bias distribution every layers in model.
!!YOU CAN READ THIS CODE LATER!!
"""
# Create a figure and a set of subplots
fig, axs = plt.subplots(2,3, sharey=False, tight_layout=True)
# weight and bias for every hidden layer
h1_w = ... | Homework/Principles of Artificial Neural Networks/Week 4 Training Issues 1/TrainingIssues.ipynb | MegaShow/college-programming | mit |
作业1
使用 torch.nn.init.constant_, torch.nn.init.xavier_uniform_, torch.nn.kaiming_uniform_去重写初始化函数，使用对应函数初始化模型，并且使用show_weight_bias显示模型隐藏层的参数分布。此处应该有6个cell作答。 | def weight_bias_reset_constant(model):
"""Constant initalization
"""
for m in model.modules():
if isinstance(m, nn.Linear):
val = 0.1
torch.nn.init.constant_(m.weight, val)
torch.nn.init.constant_(m.bias, val)
weight_bias_reset_constant(model)
show_weight_bias(mo... | Homework/Principles of Artificial Neural Networks/Week 4 Training Issues 1/TrainingIssues.ipynb | MegaShow/college-programming | mit |
3.3 Repeat over certain numbers of epoch
Shuffle whole training data
shuffle
train_loader = torch.utils.data.DataLoader(train_dataset, batch_size=batch_size, shuffle=True, **kwargs)
* For each mini-batch data
* load mini-batch data
```
for batch_idx, (data, target) in enumerate(train_loader): \
...
```
* compu... | # define method of preprocessing data for evaluating
train_transform = transforms.Compose([
transforms.ToTensor(), # Convert a PIL Image or numpy.ndarray to tensor.
# Normalize a tensor image with mean 0.1307 and standard deviation 0.3081
transforms.Normalize((0.1307,), (0.3081,))
])
test_transform = tran... | Homework/Principles of Artificial Neural Networks/Week 4 Training Issues 1/TrainingIssues.ipynb | MegaShow/college-programming | mit |
3.3.2 & 3.3.3 compute gradient of loss over parameters & update parameters with gradient descent | def train(train_loader, model, loss_fn, optimizer, get_grad=False):
"""train model using loss_fn and optimizer. When thid function is called, model trains for one epoch.
Args:
train_loader: train data
model: prediction model
loss_fn: loss function to judge the distance between target and... | Homework/Principles of Artificial Neural Networks/Week 4 Training Issues 1/TrainingIssues.ipynb | MegaShow/college-programming | mit |
Define function fit and use train_epoch and test_epoch | def fit(train_loader, val_loader, model, loss_fn, optimizer, n_epochs, get_grad=False):
"""train and val model here, we use train_epoch to train model and
val_epoch to val model prediction performance
Args:
train_loader: train data
val_loader: validation data
model: prediction mode... | Homework/Principles of Artificial Neural Networks/Week 4 Training Issues 1/TrainingIssues.ipynb | MegaShow/college-programming | mit |
作业 2
运行一下fit函数，根据结束时候训练集的accuracy，回答：模型是否训练到过拟合。
使用提供的show_curve函数，画出训练的时候loss和accuracy的变化
Hints: 因为jupyter对变量有上下文关系，模型，优化器需要重新声明。可以使用以下代码进行重新定义模型和优化器。注意到此处用的是默认初始化。 | ### Hyper parameters
batch_size = 128 # batch size is 128
n_epochs = 5 # train for 5 epochs
learning_rate = 0.01 # learning rate is 0.01
input_size = 28*28 # input image has size 28x28
hidden_size = 100 # hidden neurons is 100 for each layer
output_size = 10 # classes of prediction
l2_norm = 0 # not to use l2 penalty
... | Homework/Principles of Artificial Neural Networks/Week 4 Training Issues 1/TrainingIssues.ipynb | MegaShow/college-programming | mit |
模型没有训练到过拟合，观察上面训练数据，随着代数增多，测试集的正确率并没有下降。 | show_curve(train_accs, 'accuracy')
show_curve(train_losses, 'loss') | Homework/Principles of Artificial Neural Networks/Week 4 Training Issues 1/TrainingIssues.ipynb | MegaShow/college-programming | mit |
作业 3
将n_epochs设为10，观察模型是否能在训练集上达到过拟合， 使用show_curve作图。
当希望模型在5个epoch内在训练集上达到过拟合，可以通过适当调整learning rate来实现。选择一个合适的learing rate，训练模型，并且使用show_curve作图， 验证你的learning rate
Hints: 因为jupyter对变量有上下文关系，模型，优化器需要重新声明。可以使用以下代码进行重新定义模型和优化器。注意到此处用的是默认初始化。 | ### Hyper parameters
batch_size = 128 # batch size is 128
n_epochs = 5 # train for 5 epochs
learning_rate = 0.01 # learning rate is 0.01
input_size = 28*28 # input image has size 28x28
hidden_size = 100 # hidden neurons is 100 for each layer
output_size = 10 # classes of prediction
l2_norm = 0 # not to use l2 penalty
... | Homework/Principles of Artificial Neural Networks/Week 4 Training Issues 1/TrainingIssues.ipynb | MegaShow/college-programming | mit |
观察数据可以发现其实10代也没有过拟合。 | # 3.1 show_curve
show_curve(train_accs, 'accuracy')
show_curve(train_losses, 'loss')
# 3.2 Train
batch_size = 128 # batch size is 128
n_epochs = 5 # train for 5 epochs
learning_rate = 0.7
input_size = 28*28 # input image has size 28x28
hidden_size = 100 # hidden neurons is 100 for each layer
output_size = 10 # classes... | Homework/Principles of Artificial Neural Networks/Week 4 Training Issues 1/TrainingIssues.ipynb | MegaShow/college-programming | mit |
3.4 save model
Pytorch provide two kinds of method to save model. We recommmend the method which only saves parameters. Because it's more feasible and dont' rely on fixed model.
When saving parameters, we not only save learnable parameters in model, but also learnable parameters in optimizer.
A common PyTorch convent... | # show parameters in model
# Print model's state_dict
print("Model's state_dict:")
for param_tensor in model.state_dict():
print(param_tensor, "\t", model.state_dict()[param_tensor].size())
# Print optimizer's state_dict
print("\nOptimizer's state_dict:")
for var_name in optimizer.state_dict():
print(var_name... | Homework/Principles of Artificial Neural Networks/Week 4 Training Issues 1/TrainingIssues.ipynb | MegaShow/college-programming | mit |
作业 4
使用 test_epoch 函数，预测new_model在test_loader上的accuracy和loss | # test your model prediction performance
new_test_loss, new_test_accuracy = evaluate(test_loader, new_model, loss_fn)
message = 'Average loss: {:.4f}, Accuracy: {:.4f}'.format(new_test_loss, new_test_accuracy)
print(message) | Homework/Principles of Artificial Neural Networks/Week 4 Training Issues 1/TrainingIssues.ipynb | MegaShow/college-programming | mit |
4. Training Advanced
4.1 l2_norm
we could minimize the regularization term below by use $weight_decay$ in SGD optimizer
\begin{equation}
L_norm = {\sum_{i=1}^{m}{\theta_{i}^{2}}}
\end{equation}
set l2_norm=0.01, let's train and see | ### Hyper parameters
batch_size = 128
n_epochs = 5
learning_rate = 0.01
input_size = 28*28
hidden_size = 100
output_size = 10
l2_norm = 0.01 # use l2 penalty
get_grad = False
# declare a model
model = FeedForwardNeuralNetwork(input_size=input_size, hidden_size=hidden_size, output_size=output_size)
# Cross entropy
loss... | Homework/Principles of Artificial Neural Networks/Week 4 Training Issues 1/TrainingIssues.ipynb | MegaShow/college-programming | mit |
作业 5
思考正则项在loss中占比的影响。使用 l2_norm = 1, 训练模型
Hints: 因为jupyter对变量有上下文关系，模型，优化器需要重新声明。可以使用以下代码进行重新定义模型和优化器。注意到此处用的是默认初始化。 | # Hyper parameters
batch_size = 128
n_epochs = 5
learning_rate = 0.01
input_size = 28*28
hidden_size = 100
output_size = 10
l2_norm = 1 # use l2 penalty
get_grad = False
# declare a model
model = FeedForwardNeuralNetwork(input_size=input_size, hidden_size=hidden_size, output_size=output_size)
# Cross entropy
loss_fn =... | Homework/Principles of Artificial Neural Networks/Week 4 Training Issues 1/TrainingIssues.ipynb | MegaShow/college-programming | mit |
4.2 dropout
During training, randomly zeroes some of the elements of the input tensor with probability p using samples from a Bernoulli distribution.
Each channel will be zeroed out independently on every forward call.
Hints: 因为jupyter对变量有上下文关系，模型，优化器需要重新声明。可以使用以下代码进行重新定义模型和优化器。注意到此处用的是默认初始化。 | ### Hyper parameters
batch_size = 128
n_epochs = 5
learning_rate = 0.01
input_size = 28*28
hidden_size = 100
output_size = 10
l2_norm = 0 # without using l2 penalty
get_grad = False
# declare a model
model = FeedForwardNeuralNetwork(input_size=input_size, hidden_size=hidden_size, output_size=output_size)
# Cross entro... | Homework/Principles of Artificial Neural Networks/Week 4 Training Issues 1/TrainingIssues.ipynb | MegaShow/college-programming | mit |
4.3 batch_normalization
Batch normalization is a technique for improving the performance and stability of artificial neural networks
\begin{equation}
y=\frac{x-E[x]}{\sqrt{Var[x]+\epsilon}} * \gamma + \beta,
\end{equation}
$\gamma$ and $\beta$ are learnable parameters
Hints: 因为jupyter对变量有上下文关系，模型，优化器需要重新声明。可以使用以下代... | ### Hyper parameters
batch_size = 128
n_epochs = 5
learning_rate = 0.01
input_size = 28*28
hidden_size = 100
output_size = 10
l2_norm = 0 # without using l2 penalty
get_grad = False
# declare a model
model = FeedForwardNeuralNetwork(input_size=input_size, hidden_size=hidden_size, output_size=output_size)
# Cross entro... | Homework/Principles of Artificial Neural Networks/Week 4 Training Issues 1/TrainingIssues.ipynb | MegaShow/college-programming | mit |
4.4 data augmentation
data augmentation can be more complicated to gain a better generalization on test dataset | # only add random horizontal flip
train_transform_1 = transforms.Compose([
transforms.RandomHorizontalFlip(),
transforms.ToTensor(), # Convert a PIL Image or numpy.ndarray to tensor.
# Normalize a tensor image with mean and standard deviation
transforms.Normalize((0.1307,), (0.3081,))
])
# only add ran... | Homework/Principles of Artificial Neural Networks/Week 4 Training Issues 1/TrainingIssues.ipynb | MegaShow/college-programming | mit |
作业 6
使用提供的train_transform_2, train_transform_3，重新加载train_loader，并且使用fit进行训练
Hints: 因为jupyter对变量有上下文关系，模型，优化器需要重新声明。注意到此处用的是默认初始化。 | # train_transform_2
batch_size = 128
train_dataset_2 = torchvision.datasets.MNIST(root='./data',
train=True,
transform=train_transform_2,
download=False)
train_loader_2 = torch.utils.data.DataLoader(dataset=train_dataset_2,
... | Homework/Principles of Artificial Neural Networks/Week 4 Training Issues 1/TrainingIssues.ipynb | MegaShow/college-programming | mit |
5. Visualizatio of training and validation phase
We could use tensorboard to visualize our training and test phase.
You could find example here
6. Gradient explosion and vanishing
We have embedded code which shows grad for hidden2 and hidden3 layer. By observing their grad changes, we can
see whether gradient is norma... | ### Hyper parameters
batch_size = 128
n_epochs = 15
learning_rate = 0.01
input_size = 28*28
hidden_size = 100
output_size = 10
l2_norm = 0 # use l2 penalty
get_grad = True
# declare a model
model = FeedForwardNeuralNetwork(input_size=input_size, hidden_size=hidden_size, output_size=output_size)
# Cross entropy
loss_fn... | Homework/Principles of Artificial Neural Networks/Week 4 Training Issues 1/TrainingIssues.ipynb | MegaShow/college-programming | mit |
6.1.1 Gradient Vanishing
Set learning=e-10 | ### Hyper parameters
batch_size = 128
n_epochs = 15
learning_rate = 1e-10
input_size = 28*28
hidden_size = 100
output_size = 10
l2_norm = 0 # use l2 penalty
get_grad = True
# declare a model
model = FeedForwardNeuralNetwork(input_size=input_size, hidden_size=hidden_size, output_size=output_size)
# Cross entropy
loss_f... | Homework/Principles of Artificial Neural Networks/Week 4 Training Issues 1/TrainingIssues.ipynb | MegaShow/college-programming | mit |
6.1.2 Gradient Explosion
6.1.2.1 learning rate
set learning rate = 10 | ### Hyper parameters
batch_size = 128
n_epochs = 15
learning_rate = 10
input_size = 28*28
hidden_size = 100
output_size = 10
l2_norm = 0 # not to use l2 penalty
get_grad = True
# declare a model
model = FeedForwardNeuralNetwork(input_size=input_size, hidden_size=hidden_size, output_size=output_size)
# Cross entropy
lo... | Homework/Principles of Artificial Neural Networks/Week 4 Training Issues 1/TrainingIssues.ipynb | MegaShow/college-programming | mit |
6.1.2.2 normalization for input data
6.1.2.3 unsuitable weight initialization | ### Hyper parameters
batch_size = 128
n_epochs = 15
learning_rate = 1
input_size = 28*28
hidden_size = 100
output_size = 10
l2_norm = 0 # not to use l2 penalty
get_grad = True
# declare a model
model = FeedForwardNeuralNetwork(input_size=input_size, hidden_size=hidden_size, output_size=output_size)
# Cross entropy
los... | Homework/Principles of Artificial Neural Networks/Week 4 Training Issues 1/TrainingIssues.ipynb | MegaShow/college-programming | mit |
astroplan
Plan for everything but the clouds ☁️
Brett Morris
with Jazmin Berlanga Medina, Christoph Deil, Eric Jeschke, Adrian Price-Whelan, Erik Tollerud
Getting started
bash
pip install astropy astroplan
echo "Optionally:"
pip install wcsaxes astroquery
Outline
Background: astropy
astroplan basics
<img src="https:/... | # Altitude-azimuth frame:
from astropy.coordinates import SkyCoord, EarthLocation, AltAz
import astropy.units as u
from astropy.time import Time
# Specify location of Apache Point Observatory with astropy.coordinates.EarthLocation
apache_point = EarthLocation.from_geodetic(-105.82*u.deg, 32.78*u.deg, 2798*u.m)
# Spe... | presentation.ipynb | bmorris3/gsoc2015 | mit |
astropy: Get coordinates of the Sun | # Where is the sun right now?
from astropy.coordinates import get_sun
sun = get_sun(time)
print(sun) | presentation.ipynb | bmorris3/gsoc2015 | mit |
astroplan v0.1
Open source in Python
astropy powered
Get (alt/az) positions of targets at any time, from any observatory
Can I observe these targets given some constraints (airmass, moon separation, etc.)?
astroplan basics
astroplan.Observer: contains information about an observer's location, environment on the Eart... | from astroplan import Observer
# Construct an astroplan.Observer at Apache Point Observatory
apache_point = Observer.at_site("Apache Point")
apache_point = Observer.at_site("APO") # also works
print(apache_point.location.to_geodetic()) | presentation.ipynb | bmorris3/gsoc2015 | mit |
astroplan basics
astroplan.FixedTarget: contains information about celestial objects with no (slow) proper motion | from astroplan import FixedTarget
# Construct an astroplan.FixedTarget for Vega
vega = FixedTarget.from_name("Vega") # (with internet access)
# # (without internet access)
# vega_icrs = SkyCoord(ra=279.235416*u.deg, dec=38.78516*u.deg)
# vega = FixedTarget(coord=vega_icrs, name="Vega")
vega_altaz = apache_point.alt... | presentation.ipynb | bmorris3/gsoc2015 | mit |
Convenience methods
Is it night at this observatory at time?
| Question | Answer |
|------------------|---------------|
| Is it nighttime? | observer.is_night(time) |
| Is Vega up? | observer.target_is_up(time, vega) |
| What is the LST? | observer.local_sidereal_time(time) |
| Hour angle of Vega... | apache_point.is_night(time) | presentation.ipynb | bmorris3/gsoc2015 | mit |
Is Vega above the horizon at time? | apache_point.target_is_up(time, vega) | presentation.ipynb | bmorris3/gsoc2015 | mit |
Make your own TUI window: | # Local Sidereal time
apache_point.local_sidereal_time(time)
# Hour angle
apache_point.target_hour_angle(time, vega)
# Parallactic angle
apache_point.parallactic_angle(time, vega) | presentation.ipynb | bmorris3/gsoc2015 | mit |
Rise/set times
Next sunset | sunset = apache_point.sun_set_time(time, which='next')
print("{0.jd} = {0.iso}".format(sunset)) | presentation.ipynb | bmorris3/gsoc2015 | mit |
Next rise of Vega | vega_rise = apache_point.target_rise_time(time, vega, which='next')
print(vega_rise.iso) | presentation.ipynb | bmorris3/gsoc2015 | mit |
Next astronomical (-18 deg) twilight | astronomical_twilight = apache_point.twilight_evening_astronomical(time, which='next')
print(astronomical_twilight.iso) | presentation.ipynb | bmorris3/gsoc2015 | mit |
What is that time in local Seattle time (PST)? | # Specify your time zone with `pytz`
import pytz
my_timezone = pytz.timezone('US/Pacific')
astronomical_twilight.to_datetime(my_timezone) | presentation.ipynb | bmorris3/gsoc2015 | mit |
Constraints
Can I observe target(s) given:
Time
of year
of night
at "night"
Telescope:
Altitude constraints, i.e. 15-80$^\circ$ altitude
Location on Earth
Moon separation, illumination
Constraints example
Let's read in a list of RA/Dec of our targets: | %%writefile targets.txt
# name ra_degrees dec_degrees
Polaris 37.95456067 89.26410897
Vega 279.234734787 38.783688956
Albireo 292.68033548 27.959680072
Algol 47.042218553 40.955646675
Rigel 78.634467067 -8.201638365
Regulus 152.092962438 11.967208776 | presentation.ipynb | bmorris3/gsoc2015 | mit |
Read in target file to a list of astroplan.FixedTarget objects: | # Read in the table of targets
from astropy.table import Table
target_table = Table.read('targets.txt', format='ascii')
# Create astroplan.FixedTarget objects for each one in the table
from astropy.coordinates import SkyCoord
import astropy.units as u
from astroplan import FixedTarget
targets = [FixedTarget(coord=S... | presentation.ipynb | bmorris3/gsoc2015 | mit |
Initialize astroplan.Observer, observing time window: | from astroplan import Observer
from astropy.time import Time
subaru = Observer.at_site("Subaru")
time_range = Time(["2015-08-01 06:00", "2015-08-01 12:00"]) | presentation.ipynb | bmorris3/gsoc2015 | mit |
Define and compute constraints: | from astroplan import (AltitudeConstraint, AirmassConstraint,
AtNightConstraint)
# Define constraints:
constraints = [AltitudeConstraint(10*u.deg, 80*u.deg),
AirmassConstraint(5), AtNightConstraint.twilight_civil()]
from astroplan import is_observable, is_always_observable
# Com... | presentation.ipynb | bmorris3/gsoc2015 | mit |
Plot celestial sphere positions | %matplotlib inline
from astroplan.plots import plot_sky, plot_airmass
import numpy as np
import matplotlib.pyplot as plt
plot_times = time_range[0] + np.linspace(0, 1, 10)*(time_range[1] - time_range[0])
fig = plt.figure(figsize=(12, 6))
ax0 = fig.add_subplot(121, projection='polar')
ax1 = fig.add_subplot(122)
# Pl... | presentation.ipynb | bmorris3/gsoc2015 | mit |
Plot finder charts | # This method requires astroquery, wcsaxes
from astroplan.plots import plot_finder_image
m1 = FixedTarget.from_name('M1')
plot_finder_image(m1, survey='DSS'); | presentation.ipynb | bmorris3/gsoc2015 | mit |
Actually it is even easier. There are special operators to add, subtract, multipy, and divide into the current variable.
+=
-=
*=
/= | print(count)
count += 5
print(count)
count *= 2
print(count) | Lesson04_Iteration/Iterations.ipynb | WomensCodingCircle/CodingCirclePython | mit |
TRY IT
Assign 5 to the variable x and then in a new line add 2 to x and store the result back in x.
While loops
Loops are one of the most powerful feature of programming. With a loop you can make a computer do some repetitive task over and over again (is it any wonder that computers are taking our jobs?)
The while loop... | feet_of_snow = 0
while (feet_of_snow < 3):
print("Snow is falling!")
feet_of_snow += 1
print("We got " + str(feet_of_snow) + " feet of snow :(") | Lesson04_Iteration/Iterations.ipynb | WomensCodingCircle/CodingCirclePython | mit |
TRY IT
Write out a while loop that prints out the values 1 - 5
Infinite loops
One thing you need to be careful of with while loops is the case where the condition never turns false. This means that the loop will keep going and going forever. You'll hear your computer's fan heat up and the program will hang. If you are ... | snowflakes = 0
# This condition is never false so this will run forever
while (snowflakes >= 0):
snowflakes += 1
# If you ran this by accident, press the square (stop) button to kill the loop
# The other common infinite loop is forgetting to update to a counting variable
count = 0
while (count < 3):
prin... | Lesson04_Iteration/Iterations.ipynb | WomensCodingCircle/CodingCirclePython | mit |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.