markdown stringlengths 0 37k | code stringlengths 1 33.3k | path stringlengths 8 215 | repo_name stringlengths 6 77 | license stringclasses 15
values |
|---|---|---|---|---|
Made a mistake on the last loop above. The penultimate batch -- the last full 4096-image batch -- was added onto the end of the predictions array twice. The final 2194 image predictions were never run.
Easy enough to fix: modify the above code to work perfectly. Then either:
* create entirely new predictions from scrat... | print(81920 - 79726)
print(79726 % 4096)
print(81920 % 4096) # <-- that's yeh problem right there, kid
x = preds[len(preds) - 4096]
print(preds[-1])
print(x)
preds[0]
# ??image.ImageDataGenerator.flow_from_directory
# ??Sequential.predict() | FAI_old/conv_test_Asus.ipynb | WNoxchi/Kaukasos | mit |
Redoing predictions here: | fname = path + 'results/conv_test_feat.dat'
idx, inc = 4096, 4096
preds = []
conv_test_feat = bcolz.open(fname)[:idx]
preds = bn_model.predict(conv_test_feat, batch_size=batch_size, verbose=0)
while idx < test_batches.n - inc:
conv_test_feat = bcolz.open(fname)[idx:idx+inc]
idx += inc
next_preds = bn_model... | FAI_old/conv_test_Asus.ipynb | WNoxchi/Kaukasos | mit |
Oh I forgot, predictions through a FC NN are fast. CNNs are where it takes a long time.
This is just quick testing that it works. Full/polished will be in the reworked statefarm-codealong (or just statefarm) JNB: | def do_clip(arr, mx): return np.clip(arr, (1-mx)/9, mx)
subm = do_clip(preds, 0.93)
subm_name = path + 'results/subm01.gz'
trn_batches = get_batches(path + 'train', batch_size=batch_size, shuffle=False)
# make sure training batches defined before this:
classes = sorted(trn_batches.class_indices, key=trn_batches.cla... | FAI_old/conv_test_Asus.ipynb | WNoxchi/Kaukasos | mit |
Configuring logging: save DEBUG to a file and INFO to stdout | log = logging.getLogger()
log.setLevel(logging.DEBUG)
formatter = logging.Formatter('%(asctime)s - %(name)s - %(levelname)s - %(message)s')
fileHandler = logging.FileHandler("debug.log", 'a')
fileHandler.setLevel(logging.DEBUG)
fileHandler.setFormatter(formatter)
log.addHandler(fileHandler)
cformat = logging.Formatte... | docs/tutorials/iotlab/tuto_iotlab_g5k_ipv6.ipynb | BeyondTheClouds/enoslib | gpl-3.0 |
Getting resources
IoT-LAB provider configuration: reserve M3 nodes in saclay site
Note: It uses the following M3 images: border-router.iotlab-m3 and er-example-server.iotlab-m3.
More details on how to generate these images in: https://www.iot-lab.info/legacy/tutorials/contiki-coap-m3/index.html | job_name="iotlab_g5k-ipv6"
iotlab_dict = {
"walltime": "01:00",
"job_name": job_name,
"resources": {
"machines": [
{
"roles": ["border_router"],
"archi": "m3:at86rf231",
"site": "saclay",
"number": 1,
"image"... | docs/tutorials/iotlab/tuto_iotlab_g5k_ipv6.ipynb | BeyondTheClouds/enoslib | gpl-3.0 |
Grid'5000 provider configuration: reserve nodes in grenoble | g5k_dict = {
"job_type": "allow_classic_ssh",
"job_name": job_name,
"resources": {
"machines": [
{
"roles": ["client"],
"cluster": "yeti",
"nodes": 1,
"primary_network": "default",
"secondary_networks": [],
... | docs/tutorials/iotlab/tuto_iotlab_g5k_ipv6.ipynb | BeyondTheClouds/enoslib | gpl-3.0 |
We still need a Static provider to interact with the IoT-LAB frontend machine | import iotlabcli.auth
iotlab_user, _ = iotlabcli.auth.get_user_credentials()
iotlab_frontend_conf = (
StaticConf()
.add_machine(
roles=["frontend"],
address="saclay.iot-lab.info",
alias="saclay",
user=iotlab_user
)
.finalize()
) | docs/tutorials/iotlab/tuto_iotlab_g5k_ipv6.ipynb | BeyondTheClouds/enoslib | gpl-3.0 |
IoT-LAB: getting resources | iotlab_provider = Iotlab(iotlab_conf)
iotlab_roles, _ = iotlab_provider.init()
print(iotlab_roles) | docs/tutorials/iotlab/tuto_iotlab_g5k_ipv6.ipynb | BeyondTheClouds/enoslib | gpl-3.0 |
Grid'5000: getting resources | g5k_provider = G5k(g5k_conf)
g5k_roles, g5knetworks = g5k_provider.init()
print(g5k_roles) | docs/tutorials/iotlab/tuto_iotlab_g5k_ipv6.ipynb | BeyondTheClouds/enoslib | gpl-3.0 |
Static: getting resources | frontend_provider = Static(iotlab_frontend_conf)
frontend_roles, _ = frontend_provider.init()
print(frontend_roles) | docs/tutorials/iotlab/tuto_iotlab_g5k_ipv6.ipynb | BeyondTheClouds/enoslib | gpl-3.0 |
Configuring network connectivity
Enabling IPv6 on Grid'5000 nodes (https://www.grid5000.fr/w/IPv6) | result=run_command("dhclient -6 br0", roles=g5k_roles)
result = run_command("ip address show dev br0", roles=g5k_roles)
print(result['ok']) | docs/tutorials/iotlab/tuto_iotlab_g5k_ipv6.ipynb | BeyondTheClouds/enoslib | gpl-3.0 |
Starting tunslip command in frontend.
Redirect tunslip command output to a file to read it later. | iotlab_ipv6_net="2001:660:3207:4c0::"
tun_cmd = "sudo tunslip6.py -v2 -L -a %s -p 20000 %s1/64 > tunslip.output 2>&1" % (iotlab_roles["border_router"][0].alias, iotlab_ipv6_net)
result=run_command(tun_cmd, roles=frontend_roles, asynch=3600, poll=0) | docs/tutorials/iotlab/tuto_iotlab_g5k_ipv6.ipynb | BeyondTheClouds/enoslib | gpl-3.0 |
Reseting border router | iotlab_roles["border_router"][0].reset() | docs/tutorials/iotlab/tuto_iotlab_g5k_ipv6.ipynb | BeyondTheClouds/enoslib | gpl-3.0 |
Get the Border Router IPv6 address from tunslip output | result = run_command("cat tunslip.output", roles=frontend_roles)
print(result['ok'])
import re
out = result['ok']['saclay']['stdout']
print(out)
match = re.search(rf'Server IPv6 addresses:\n.+({iotlab_ipv6_net}\w{{4}})', out, re.MULTILINE|re.DOTALL)
br_ipv6 = match.groups()[0]
print("Border Router IPv6 address from tu... | docs/tutorials/iotlab/tuto_iotlab_g5k_ipv6.ipynb | BeyondTheClouds/enoslib | gpl-3.0 |
Checking ping from Grid'5000 to border router node | result = run_command("ping6 -c3 %s" % br_ipv6, pattern_hosts="client*", roles=g5k_roles)
print(result['ok']) | docs/tutorials/iotlab/tuto_iotlab_g5k_ipv6.ipynb | BeyondTheClouds/enoslib | gpl-3.0 |
Installing and using CoAP clients
Install aiocoap client and lynx on grid'5000 nodes | with play_on(roles=g5k_roles) as p:
p.apt(name=["python3-aiocoap", "lynx"], state="present") | docs/tutorials/iotlab/tuto_iotlab_g5k_ipv6.ipynb | BeyondTheClouds/enoslib | gpl-3.0 |
Grab the CoAP server node’s IPv6 address from the BR’s web interface | result = run_command("lynx -dump http://[%s]" % br_ipv6, roles=g5k_roles)
print(result['ok']) | docs/tutorials/iotlab/tuto_iotlab_g5k_ipv6.ipynb | BeyondTheClouds/enoslib | gpl-3.0 |
For a CoAP server, GET light sensor | out = result['ok'][g5k_roles["client"][0].address]['stdout']
print(out)
match = re.search(r'fe80::(\w{4})', out, re.MULTILINE|re.DOTALL)
node_uid = match.groups()[0]
print(node_uid)
result = run_command("aiocoap-client coap://[%s%s]:5683/sensors/light" % (iotlab_ipv6_net, node_uid), roles=g5k_roles)
print(result['ok']... | docs/tutorials/iotlab/tuto_iotlab_g5k_ipv6.ipynb | BeyondTheClouds/enoslib | gpl-3.0 |
GET pressure for the same sensor | result = run_command("aiocoap-client coap://[%s%s]:5683/sensors/pressure" % (iotlab_ipv6_net, node_uid), roles=g5k_roles)
print(result['ok']) | docs/tutorials/iotlab/tuto_iotlab_g5k_ipv6.ipynb | BeyondTheClouds/enoslib | gpl-3.0 |
Clean-up phase
Stop tunslip in frontend node | result = run_command("pgrep tunslip6 | xargs kill", roles=frontend_roles) | docs/tutorials/iotlab/tuto_iotlab_g5k_ipv6.ipynb | BeyondTheClouds/enoslib | gpl-3.0 |
Destroy jobs in testbeds | g5k_provider.destroy()
iotlab_provider.destroy() | docs/tutorials/iotlab/tuto_iotlab_g5k_ipv6.ipynb | BeyondTheClouds/enoslib | gpl-3.0 |
Now we make some plots and calculate some figures. For example, here's how the correlation (Pearson's r) changes by game number: | # Drop short season and current season, and get correlations
data2 = data[(data.Season != 2012) & (data.Season != 2017)].dropna()
corrs = data2.drop({'Season', 'Team'}, axis=1).groupby('GameNum').corr().drop('ROY_CF%', axis=1)
corrs = corrs[corrs['YTD_CF%'] < 1]
corrs = corrs.reset_index().drop('level_1', axis=1).renam... | examples/YTD and ROY CF%.ipynb | muneebalam/scrapenhl2 | mit |
Here's how the slope changes by game: | # Now look at the slope
from scipy.stats import linregress
def get_slope(df):
m, b, r, p, e = linregress(df['YTD_CF%'], df['ROY_CF%'])
return m
plot(data2.groupby('GameNum').apply(get_slope), label = 'Slope')
xlabel('Game Number')
plot(corrs.r, label = 'R')
title('Correlation between YTD CF% and ROY CF%')
lege... | examples/YTD and ROY CF%.ipynb | muneebalam/scrapenhl2 | mit |
You'll note that although the predictivity (as measured by r) is best at 40 games, the slope is still not 1--meaning we expect some regression still.
You can see that in the scatterplot below (with the best-fit and 1:1 lines included for reference): | tmp = data2[data2.GameNum == 40]
x = 'YTD_CF%'
y = 'ROY_CF%'
scatter(tmp[x], tmp[y], label='_nolegend')
xlabel('Season to date CF%')
ylabel('Rest of season CF%')
title('YTD and ROY CF% at game 40')
m, b, r, p, e = linregress(tmp[x], tmp[y])
xs = arange(0, 1, 0.01)
ys = m * xs + b
xlimits = xlim()
ylimits = ylim()
pl... | examples/YTD and ROY CF%.ipynb | muneebalam/scrapenhl2 | mit |
Vertex SDK: AutoML training text classification model for online prediction
<table align="left">
<td>
<a href="https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/tree/master/notebooks/official/automl/sdk_automl_text_classification_online.ipynb">
<img src="https://cloud.google.com... | import os
# Google Cloud Notebook
if os.path.exists("/opt/deeplearning/metadata/env_version"):
USER_FLAG = "--user"
else:
USER_FLAG = ""
! pip3 install --upgrade google-cloud-aiplatform $USER_FLAG | notebooks/community/sdk/sdk_automl_text_classification_online.ipynb | GoogleCloudPlatform/vertex-ai-samples | apache-2.0 |
Send a online prediction request
Send a online prediction to your deployed model.
Get test item
You will use an arbitrary example out of the dataset as a test item. Don't be concerned that the example was likely used in training the model -- we just want to demonstrate how to make a prediction. | test_item = ! gsutil cat $IMPORT_FILE | head -n1
if len(test_item[0]) == 3:
_, test_item, test_label = str(test_item[0]).split(",")
else:
test_item, test_label = str(test_item[0]).split(",")
print(test_item, test_label) | notebooks/community/sdk/sdk_automl_text_classification_online.ipynb | GoogleCloudPlatform/vertex-ai-samples | apache-2.0 |
We're going to use one convolutional layer. To avoid config-fu with caffe, the data will go into LMDB store. But 1st, let's cleanup the environment to avoid costly debugging. | rm -r digits_train_lmdb digits_test_lmdb | Classification_With_NN_1.ipynb | marcino239/notebooks | gpl-2.0 |
Caffe uses fairly simple kev / value format. The code below performs conversion from sklearn dataset into LMDB. | import lmdb
import sys
sys.path.append( './caffe/python' )
import caffe
env = lmdb.open( 'digits_train_lmdb', map_size=1000000000L )
k = 0
for i in range( X_train.shape[0] ):
x_in = X_train[ i ].reshape( (8,8) ).astype( np.uint8 )
# add original digit
datum = caffe.proto.caffe_pb2.Datum()
datum.c... | Classification_With_NN_1.ipynb | marcino239/notebooks | gpl-2.0 |
We're going to define here solver parameters and network structure. The original MNIST reference net included with caffe, had two convolutional layers. Given the small size of the input picture (8x8), we're only going to use one convolutional layer. | solver_txt = """
# The train/test net protocol buffer definition
net: "digits_train_test.prototxt"
# test_iter specifies how many forward passes the test should carry out.
# the batch is 64 images
test_iter: 100
# Carry out testing every 100 training iterations.
test_interval: 100
# The base learning rate, momentum and... | Classification_With_NN_1.ipynb | marcino239/notebooks | gpl-2.0 |
Let's start the training using GPU. | !./caffe/build/tools/caffe train -solver digits_solver.prototxt 2> digits_train.log
!python ./caffe/tools/extra/parse_log.py digits_train.log .
import pandas as pd
# columns: NumIters,Seconds,TrainingLoss,LearningRate
df_train = pd.read_csv( 'digits_train.log.train' )
# columns: NumIters,Seconds,TestAccuracy,TestLos... | Classification_With_NN_1.ipynb | marcino239/notebooks | gpl-2.0 |
Time-varying Convex Optimization
Copyright 2018 Google LLC.
Install Dependencies
Time Varying Semi-definite Programs
Some Polynomial Tools
Examples: To Add.
Install Dependencies | !pip install cvxpy
!pip install sympy
import numpy as np
import scipy as sp | time_varying_optimization/tvsdp.ipynb | google-research/google-research | apache-2.0 |
Time Varying Semi-definite Programs
The TV-SDP framework for CVXPY for imposing constraints of the form:
$$A(t) \succeq 0 \; \forall t \in [0, 1],$$
where $$A(t)$$ is a polynomial symmetric matrix, i.e. a symmetric matrix
whose entries are polynomial functions of time, and $$A(t) \succeq 0$$
means that all the eigen va... | def _mult_poly_matrix_poly(p, mat_y):
"""Multiplies the polynomial matrix mat_y by the polynomial p entry-wise.
Args:
p: list of size d1+1 representation the polynomial sum p[i] t^i.
mat_y: (m, m, d2+1) tensor representing a polynomial
matrix Y_ij(t) = sum mat_y[i, j, k] t^k.
Returns:
(m, m, d... | time_varying_optimization/tvsdp.ipynb | google-research/google-research | apache-2.0 |
Some Polynomial Tools | def integ_poly_0_1(p):
"""Return the integral of p(t) between 0 and 1."""
return np.array(p).dot(1 / np.linspace(1, len(p), len(p)))
def spline_regression(x, y, num_parts, deg=3, alpha=.01, smoothness=1):
"""Fits splines with `num_parts` to data `(x, y)`.
Finds a piecewise polynomial function `p` of degree ... | time_varying_optimization/tvsdp.ipynb | google-research/google-research | apache-2.0 |
Paste existing code into model.py
A Python package requires our code to be in a .py file, as opposed to notebook cells. So, we simply copy and paste our existing code for the previous notebook into a single file.
In the cell below, we write the contents of the cell into model.py packaging the model we
developed in the... | %%writefile ./taxifare/trainer/model.py
#TODO 1
import datetime
import logging
import os
import shutil
import numpy as np
import tensorflow as tf
from tensorflow.keras import activations
from tensorflow.keras import callbacks
from tensorflow.keras import layers
from tensorflow.keras import models
from tensorflow imp... | courses/machine_learning/deepdive2/building_production_ml_systems/labs/1_training_at_scale.ipynb | turbomanage/training-data-analyst | apache-2.0 |
Modify code to read data from and write checkpoint files to GCS
If you look closely above, you'll notice a new function, train_and_evaluate that wraps the code that actually trains the model. This allows us to parametrize the training by passing a dictionary of parameters to this function (e.g, batch_size, num_examples... | %%writefile taxifare/trainer/task.py
# TODO 1
import argparse
from trainer import model
if __name__ == '__main__':
parser = argparse.ArgumentParser()
# TODO: Your code goes here
# TODO: Your code goes here
# TODO: Your code goes here
parser.add_argument(
"--eval_data_path",
... | courses/machine_learning/deepdive2/building_production_ml_systems/labs/1_training_at_scale.ipynb | turbomanage/training-data-analyst | apache-2.0 |
Run your training package on Cloud AI Platform
Once the code works in standalone mode locally, you can run it on Cloud AI Platform. To submit to the Cloud we use gcloud ai-platform jobs submit training [jobname] and simply specify some additional parameters for AI Platform Training Service:
- jobid: A unique identifier... | %%bash
# TODO 2
# Output directory and jobID
OUTDIR=gs://${BUCKET}/taxifare/trained_model_$(date -u +%y%m%d_%H%M%S)
JOBID=taxifare_$(date -u +%y%m%d_%H%M%S)
echo ${OUTDIR} ${REGION} ${JOBID}
gsutil -m rm -rf ${OUTDIR}
# Model and training hyperparameters
BATCH_SIZE=50
NUM_EXAMPLES_TO_TRAIN_ON=100
NUM_EVALS=100
NBUCKE... | courses/machine_learning/deepdive2/building_production_ml_systems/labs/1_training_at_scale.ipynb | turbomanage/training-data-analyst | apache-2.0 |
2) Setup the Request
Extract the Models from the Cache with this request and upload them object files to the configured S3 Bucket.
Please make sure the environment variables are set correctly and the S3 Bucket exists:
ENV_AWS_KEY=<AWS API Key>
ENV_AWS_SECRET=<AWS API Secret>
For docker containers make su... | ds_name = "iris_regressor" | examples/ML-IRIS-Redis-Labs-Extract-From-Cache.ipynb | jay-johnson/sci-pype | apache-2.0 |
3) Build and Run the Extract + Upload Request | cache_req = {
"RAName" : "CACHE", # Redis endpoint name holding the models
"DSName" : str(ds_name), # Dataset name for pulling out of the cache
"S3Loc" : str(s3_loc), # S3 location to store the model file
... | examples/ML-IRIS-Redis-Labs-Extract-From-Cache.ipynb | jay-johnson/sci-pype | apache-2.0 |
The right hand side of ODEs can be implemented in either python or in C. Although not absolutely necessary for this example, we here show how to implement the RHS in C. This is often significantly faster than using a python callback function.
We use a simple spin model which is one second order ODE, or a set of two co... | %%writefile rhs.c
#include "rebound.h"
void derivatives(struct reb_ode* const ode, double* const yDot, const double* const y, const double t){
struct reb_orbit o = reb_tools_particle_to_orbit(ode->r->G, ode->r->particles[1], ode->r->particles[0]);
double omega2 = 3.*0.26;
yDot[0] = y[1];
yDot[1] =... | ipython_examples/ChaoticHyperion.ipynb | hannorein/rebound | gpl-3.0 |
We now compile this into a shared library. We need the REBOUND headers and library for this. The following is a bit of hack: we just copy the files into the current folder. This works if you've installed REBOUND from the git repository. Otherwise, you'll need to find these files manually (which might depend on your pyt... | !cp ../src/librebound.so .
!cp ../src/rebound.h .
!gcc -c -O3 -fPIC rhs.c -o rhs.o
!gcc -L. -shared rhs.o -o rhs.so -lrebound | ipython_examples/ChaoticHyperion.ipynb | hannorein/rebound | gpl-3.0 |
Using ctypes, we can load the library into python | from ctypes import cdll
clibrhs = cdll.LoadLibrary("rhs.so") | ipython_examples/ChaoticHyperion.ipynb | hannorein/rebound | gpl-3.0 |
The following function is setting up the N-body simulation as well as the ODE system that governs the spin evolution. Note that we set the derivatives function pointer to the C function we've just compiled. You could also set this function pointer to a python function and avoid all the C complications. | def setup():
sim = rebound.Simulation()
sim.add(m=1) # Saturn
sim.add(a=1, e=0.123233) # Hyperion, massless, semi-major axis of 1
sim.integrator = "BS"
sim.ri_bs.eps_rel = 1e-12 # tolerance
sim.ri_bs.eps_abs = 1e-12
ode_spin = sim.create_ode(length=2, needs_nbody=True)
... | ipython_examples/ChaoticHyperion.ipynb | hannorein/rebound | gpl-3.0 |
We will create two simulations that are slightly offset from each other. | sim, ode_spin = setup()
sim2, ode_spin2 = setup()
ode_spin2.y[0] += 1e-8 # small perturbation | ipython_examples/ChaoticHyperion.ipynb | hannorein/rebound | gpl-3.0 |
With these two simulations, we can measure the growing divergence of nearby trajectories, a key feature of chaos. | times = 2.*np.pi*np.linspace(0,30,100) # a couple of orbits
obliq = np.zeros((len(times)))
obliq2 = obliq.copy()
for i, t in enumerate(times):
sim.integrate(t, exact_finish_time=1)
sim2.integrate(t, exact_finish_time=1)
obliq[i] = ode_spin.y[0]
obliq2[i] = ode_spin2.y[0] | ipython_examples/ChaoticHyperion.ipynb | hannorein/rebound | gpl-3.0 |
Finally, let us plot the divergence as a function of time. | fig, ax = plt.subplots(1,1)
ax.set_xlabel("time [orbits]")
ax.set_ylabel("obliquity difference [degrees]")
ax.set_yscale("log")
ax.set_ylim([1e-8,1e2])
o1 = np.remainder((obliq-obliq2)*180/np.pi,360.)
o2 = np.remainder((obliq2-obliq)*180/np.pi,360.)
ax.scatter(times/np.pi/2.,np.minimum(o1,o2))
ax.plot(times/np.pi/2.,1e... | ipython_examples/ChaoticHyperion.ipynb | hannorein/rebound | gpl-3.0 |
Next, import all the dependencies we'll use in this exercise, which include Fairness Indicators, TensorFlow Model Analysis (tfma), and the What-If tool (WIT): | %tensorflow_version 2.x
import os
import tempfile
import apache_beam as beam
import numpy as np
import pandas as pd
from datetime import datetime
import tensorflow_hub as hub
import tensorflow as tf
import tensorflow_model_analysis as tfma
from tensorflow_model_analysis.addons.fairness.post_export_metrics import fairn... | ml/pc/exercises/fairness_text_toxicity_part2.ipynb | google/eng-edu | apache-2.0 |
Run the following code to download and import the training and validation datasets. By default, the following code will load the preprocessed data (see Fairness Exercise 1: Explore the Model for more details). If you prefer, you can enable the download_original_data checkbox at right to download the original dataset an... | download_original_data = False #@param {type:"boolean"}
if download_original_data:
train_tf_file = tf.keras.utils.get_file('train_tf.tfrecord',
'https://storage.googleapis.com/civil_comments_dataset/train_tf.tfrecord')
validate_tf_file = tf.keras.utils.get_file('validate_t... | ml/pc/exercises/fairness_text_toxicity_part2.ipynb | google/eng-edu | apache-2.0 |
Next, train the original model from Fairness Exercise 1: Explore the Model, which we'll use as the baseline model for this exercise: | #@title Run this cell to train the baseline model from Exercise 1
TEXT_FEATURE = 'comment_text'
LABEL = 'toxicity'
FEATURE_MAP = {
# Label:
LABEL: tf.io.FixedLenFeature([], tf.float32),
# Text:
TEXT_FEATURE: tf.io.FixedLenFeature([], tf.string),
# Identities:
'sexual_orientation':tf.io.VarLen... | ml/pc/exercises/fairness_text_toxicity_part2.ipynb | google/eng-edu | apache-2.0 |
In the next section, we'll apply bias-remediation techniques on our data and then train a revised model on the updated data.
Remediate Bias
To remediate bias in our model, we'll first need to define the remediation metrics we'll use to gauge success and choose an appropriate remediation technique. Then we'll retrain th... | def train_input_fn_with_remediation():
def parse_function(serialized):
parsed_example = tf.io.parse_single_example(
serialized=serialized, features=FEATURE_MAP)
# Adds a weight column to deal with unbalanced classes.
parsed_example['weight'] = tf.add(parsed_example[LABEL], 0.1)
# BEGIN U... | ml/pc/exercises/fairness_text_toxicity_part2.ipynb | google/eng-edu | apache-2.0 |
Retrain the model
Now, let's retrain the model with our upweighted examples: | BASE_DIR = tempfile.gettempdir()
model_dir_with_remediation = os.path.join(BASE_DIR, 'train', datetime.now().strftime(
"%Y%m%d-%H%M%S"))
embedded_text_feature_column = hub.text_embedding_column(
key=TEXT_FEATURE,
module_spec='https://tfhub.dev/google/nnlm-en-dim128/1')
classifier_with_remediation = tf.... | ml/pc/exercises/fairness_text_toxicity_part2.ipynb | google/eng-edu | apache-2.0 |
Recompute fairness metrics
Now that we've retrained the model, let's recompute our fairness metrics. First, export the model: | def eval_input_receiver_fn():
serialized_tf_example = tf.compat.v1.placeholder(
dtype=tf.string, shape=[None], name='input_example_placeholder')
receiver_tensors = {'examples': serialized_tf_example}
features = tf.io.parse_example(serialized_tf_example, FEATURE_MAP)
features['weight'] = tf.ones_like(fea... | ml/pc/exercises/fairness_text_toxicity_part2.ipynb | google/eng-edu | apache-2.0 |
Next, run the fairness evaluation using TFMA: | tfma_eval_result_path_with_remediation = os.path.join(BASE_DIR, 'tfma_eval_result_with_remediation')
slice_selection = 'gender'
compute_confidence_intervals = False
# Define slices that you want the evaluation to run on.
slice_spec = [
tfma.slicer.SingleSliceSpec(), # Overall slice
tfma.slicer.SingleSliceSpec... | ml/pc/exercises/fairness_text_toxicity_part2.ipynb | google/eng-edu | apache-2.0 |
Load evaluation results
Run the following two cells to load results in the What-If tool and Fairness Indicators.
In the What-If tool, we'll load 1,000 examples with the corresponding predictions returned from both the baseline model and the remediated model.
WARNING: When you launch the What-If tool widget below, the ... | DEFAULT_MAX_EXAMPLES = 1000
# Load 100000 examples in memory. When first rendered, What-If Tool only
# displays 1000 of these examples to ensure data loads successfully for most
# browser/machine configurations.
def wit_dataset(file, num_examples=100000):
dataset = tf.data.TFRecordDataset(
filenames=[train_tf... | ml/pc/exercises/fairness_text_toxicity_part2.ipynb | google/eng-edu | apache-2.0 |
In Fairness Indicators, we'll display the remediated model's evaluation results on the validation set. | # Link Fairness Indicators widget with WIT widget above,
# so that clicking a slice in FI below will load its data in WIT above.
event_handlers={'slice-selected':
wit.create_selection_callback(wit_data, DEFAULT_MAX_EXAMPLES)}
widget_view.render_fairness_indicator(eval_result=eval_result_with_remediation,
... | ml/pc/exercises/fairness_text_toxicity_part2.ipynb | google/eng-edu | apache-2.0 |
We can instantiate an instance of the tool using an API key. | cii = CancerImageInterface(api_key=YOUR_API_KEY_HERE) | tutorials/2_cancer_imaging_archive.ipynb | TariqAHassan/BioVida | bsd-3-clause |
Note:
The number of studies (collections) provided here is somewhat shorter than what is provided on The Cancer Imaging Archive's website. This is because certain studies, such as those with restricted access, are excluded.
With this notice out of the way, let's go ahead and perform a search. | cii.search(cancer_type='breast') | tutorials/2_cancer_imaging_archive.ipynb | TariqAHassan/BioVida | bsd-3-clause |
Next, we can easily download this data. | import numpy as np
def simplify_df(df):
"""This function simplifies dataframes
for the purposes of this tutorial."""
data_frame = df.copy()
for c in ('cached_dicom_images_path', 'cached_images_path'):
data_frame[c] = data_frame[c].map(
lambda x: tuple(['path_to_image'] * len(x)) if i... | tutorials/2_cancer_imaging_archive.ipynb | TariqAHassan/BioVida | bsd-3-clause |
The code below will download the data in our search results, but with two noteworthy restrictions. <br>
First, patient_limit=5 will limit the number of patients/subjects downloaded to the first 5. <br>
Second, collections_limit will limit the number of collections downloaded to one (in this case, 'TCGA-COAD'). <br>
Thi... | pull_df = cii.pull(patient_limit=5, collections_limit=1, session_limit=1) | tutorials/2_cancer_imaging_archive.ipynb | TariqAHassan/BioVida | bsd-3-clause |
Let's take a look at the data we've downloaded. We could view the pull_df object above, or the identical records_db attribute of cii, e.g., cii.records_db. However, both of those DataFrames contain several column which are not typically relevant for every data use. So, instead, we can view an abbreviated DataFrame, rec... | simplify_df(cii.records_db_short) | tutorials/2_cancer_imaging_archive.ipynb | TariqAHassan/BioVida | bsd-3-clause |
Notes:
The 'cached_dicom_images_path' and 'cached_images_path' columns refer to multiple images.
The number of converted images may differ from the number of raw DICOM images because 3D DICOM images are saved as individual frames when they are converted to PNG. The 'image_count_converted_cache' column provides an acco... | import dicom # in the future you will have to use `import pydicom as dicom` | tutorials/2_cancer_imaging_archive.ipynb | TariqAHassan/BioVida | bsd-3-clause |
We can also go ahead an import matplotlib to allow us to visualize the ndarrays we extract.
We can start by extracting a list of DICOMs from the images we downloaded above. | sample_dicoms = cii.records_db['cached_dicom_images_path'].iloc[1] | tutorials/2_cancer_imaging_archive.ipynb | TariqAHassan/BioVida | bsd-3-clause |
We can load these images in as ndarrays using the dicom (pydicom) library. | dicoms_arrs = [dicom.read_file(f).pixel_array for f in sample_dicoms] | tutorials/2_cancer_imaging_archive.ipynb | TariqAHassan/BioVida | bsd-3-clause |
DICOM represents imaging sessions with a tag known as SeriesInstanceUID. That is, the unique ID of the series.
If multiple DICOM files/images share the same SeriesInstanceUID, it means they are part of the same 'series'.
If we get the length of dicoms_arrs we see that multiple DICOM files share the same SeriesInstance... | len(dicoms_arrs)
# cii.records_db['series_instance_uid'].iloc[1] | tutorials/2_cancer_imaging_archive.ipynb | TariqAHassan/BioVida | bsd-3-clause |
Thus suggesting that this particular series is either a 3D volume or a time-series. So we can go ahead stack these images on top of one another as a way of representing this relationship between the images. | stacked_dicoms = np.stack(dicoms_arrs) | tutorials/2_cancer_imaging_archive.ipynb | TariqAHassan/BioVida | bsd-3-clause |
If we check the shape of stacked_dicoms can see that we have indeed stacked 15 256x256 images on top of one another. | stacked_dicoms.shape | tutorials/2_cancer_imaging_archive.ipynb | TariqAHassan/BioVida | bsd-3-clause |
We can also go ahead and define a small function which will enable us to visualize this 'stack' of images. | import matplotlib.pyplot as plt
def sample_stack(stack, rows=3, cols=5, start_with=0, show_every=1):
"""Function to display stacked ndarray.
Source: https://www.raddq.com/dicom-processing-segmentation-visualization-in-python
Note: this code has been slightly modified."""
if rows*cols != stack.shape[0]:
... | tutorials/2_cancer_imaging_archive.ipynb | TariqAHassan/BioVida | bsd-3-clause |
If you're curious to see what these images look like, you can uncomment the line below to view the stack of images. | # sample_stack(stacked_dicoms, rows=3) | tutorials/2_cancer_imaging_archive.ipynb | TariqAHassan/BioVida | bsd-3-clause |
Note:
Ordering DICOM images in space is tricky. Currently, this class uses a somewhat reliable, but far from ideal, means of ordering images. <br> Errors are possible. <br>
<small>
For the Medical Imaging Folks: <br>
Images in the 'series_instance_uid' column are ordered against the InstanceNumber tag instead of actual... | from biovida.images import image_divvy | tutorials/2_cancer_imaging_archive.ipynb | TariqAHassan/BioVida | bsd-3-clause |
Next, we can define a 'divvy_rule'. | def my_divvy_rule(row):
if row['modality_full'] == 'Magnetic Resonance Imaging (MRI)':
return 'mri'
if row['modality_full'] == 'Segmentation':
return 'seg' | tutorials/2_cancer_imaging_archive.ipynb | TariqAHassan/BioVida | bsd-3-clause |
This rule will select only those images which are MRIs. All other images will be excluded. | train_test = image_divvy(instance=cii,
divvy_rule=my_divvy_rule,
db_to_extract='records_db',
action='ndarray',
train_val_test_dict={'train': 0.8, 'test': 0.2})
train_mri, test_mri = train_test['train']['mri'], train_tes... | tutorials/2_cancer_imaging_archive.ipynb | TariqAHassan/BioVida | bsd-3-clause |
One important thing to point out is that some of the image arrays returned will, in fact, be stacked arrays of images. <br>
For example: | train_seg[10].shape | tutorials/2_cancer_imaging_archive.ipynb | TariqAHassan/BioVida | bsd-3-clause |
Loading FFT routines | gridDIM = 64
size = gridDIM*gridDIM
axes0 = 0
axes1 = 1
makeC2C = 0
makeR2C = 1
makeC2R = 1
axesSplit_0 = 0
axesSplit_1 = 1
m = size
segment_axes0 = 0
segment_axes1 = 0
DIR_BASE = "/home/robert/Documents/new1/FFT/mycode/"
# FAFT
_faft128_2D = ctypes.cdll.LoadLibrary( DIR_BASE+'FAFT128_2D_R2C.so' )
_faft128_2D.... | FAFT_64-points_R2C/nbFAFT128_offset_xy_2D.ipynb | robertclf/FAFT | bsd-3-clause |
Initializing Data
Gaussian | def Gaussian(x,mu,sigma):
return np.exp( - (x-mu)**2/sigma**2/2. )/(sigma*np.sqrt( 2*np.pi ))
def fftGaussian(p,mu,sigma):
return np.exp(-1j*mu*p)*np.exp( - p**2*sigma**2/2. )
# Gaussian parameters
mu_x = 1.5
sigma_x = 1.
mu_y = 1.5
sigma_y = 1.
# Grid parameters
x_amplitude = 7.
p_amplitude = 5. ... | FAFT_64-points_R2C/nbFAFT128_offset_xy_2D.ipynb | robertclf/FAFT | bsd-3-clause |
Forward Transform | # Executing FFT
cuda_faft( int(f_gpu.gpudata), int(f33_gpu.gpudata), dx, delta, segment_axes0, axes0, makeR2C, axesSplit_0 )
cuda_faft( int(f_gpu.gpudata), int(f33_gpu.gpudata), dx, delta, segment_axes1, axes1, makeC2C, axesSplit_0 )
plt.imshow(
f_gpu.get()
)
plt.plot( f33_gpu.get().real.reshape(64) )
def Recons... | FAFT_64-points_R2C/nbFAFT128_offset_xy_2D.ipynb | robertclf/FAFT | bsd-3-clause |
Central Section: $p_y =0$ | plt.figure(figsize=(10,10))
plt.plot( p_range,
ReconstructFFT2D_axesSplit_0( f_gpu.get() , f33_gpu.get() )[32,:].real/float(size),
'o-' , label='Real')
plt.plot( p_range,
ReconstructFFT2D_axesSplit_0( f_gpu.get() , f33_gpu.get() )[32,:].imag/float(size),
'ro-' , label='Imag')
plt.xlabel('$p_x$',**axis_f... | FAFT_64-points_R2C/nbFAFT128_offset_xy_2D.ipynb | robertclf/FAFT | bsd-3-clause |
Inverse Transform | # Executing iFFT
cuda_ifaft( int(f_gpu.gpudata), int(f33_gpu.gpudata), dx, -delta, segment_axes1, axes1, makeC2C, axesSplit_0 )
cuda_ifaft( int(f_gpu.gpudata), int(f33_gpu.gpudata), dx, -delta, segment_axes0, axes0, makeC2R, axesSplit_0 )
plt.imshow( f_gpu.get()/(float(size*size)) ,
extent=[-x_amplitude ,... | FAFT_64-points_R2C/nbFAFT128_offset_xy_2D.ipynb | robertclf/FAFT | bsd-3-clause |
$W$ TRANSFORM FROM AXES-1
After the transfom, f_gpu[:, :32] contains real values and f_gpu[:, 32:] contains imaginary values. f33_gpu contains the 33th. complex values | f = Gaussian(x,mu_x,sigma_x)*Gaussian(y,mu_y,sigma_y)
plt.imshow( f, extent=[-x_amplitude , x_amplitude-dx, -x_amplitude , x_amplitude-dx] , origin='lower')
f33 = np.zeros( [64, 1], dtype = np.complex64 )
# One gpu array.
f_gpu = gpuarray.to_gpu( np.ascontiguousarray( f , dtype = np.float32 ) )
f33_gpu = gpuarr... | FAFT_64-points_R2C/nbFAFT128_offset_xy_2D.ipynb | robertclf/FAFT | bsd-3-clause |
Forward Transform | # Executing FFT
cuda_faft( int(f_gpu.gpudata), int(f33_gpu.gpudata), dx, delta, segment_axes1, axes1, makeR2C, axesSplit_1 )
cuda_faft( int(f_gpu.gpudata), int(f33_gpu.gpudata), dx, delta, segment_axes0, axes0, makeC2C, axesSplit_1 )
plt.imshow(
f_gpu.get()
)
plt.plot( f33_gpu.get().real.reshape(64) )
def Reconst... | FAFT_64-points_R2C/nbFAFT128_offset_xy_2D.ipynb | robertclf/FAFT | bsd-3-clause |
Inverse Transform | # Executing iFFT
cuda_ifaft( int(f_gpu.gpudata), int(f33_gpu.gpudata), dx, -delta, segment_axes0, axes0, makeC2C, axesSplit_1 )
cuda_ifaft( int(f_gpu.gpudata), int(f33_gpu.gpudata), dx, -delta, segment_axes1, axes1, makeC2R, axesSplit_1 )
plt.imshow( f_gpu.get()/float(size)**2 ,
extent=[-x_amplitude , x_a... | FAFT_64-points_R2C/nbFAFT128_offset_xy_2D.ipynb | robertclf/FAFT | bsd-3-clause |
User
~~Products purchased (total_distinct_items)~~
~~Orders made (nb_orders)~~
~~frequency and recency of orders (average_days_between_orders)~~
Aisle purchased from
Department purchased from
~~frequency and recency of reorders~~
tenure
~~mean order size (average basket)~~
etc. | usr = pd.DataFrame()
o_grouped = orders.groupby('user_id')
p_grouped = priors.groupby('user_id')
usr['average_days_between_orders'] = o_grouped.days_since_prior_order.mean()
usr['max_days_between_orders'] = o_grouped.days_since_prior_order.max()
usr['min_days_between_orders'] = o_grouped.days_since_prior_order.min()
u... | instacart/Model.ipynb | bataeves/kaggle | unlicense |
Product
~~users~~
~~orders~~
~~order frequency~~
~~reorder rate~~
recency
~~mean/std add_to_cart_order~~
Срок годности/время перепокупки
etc. | prods = pd.DataFrame()
p_grouped = priors.groupby(priors.product_id)
prods['orders'] = p_grouped.size().astype(np.float32)
prods['order_freq'] = (prods['orders'] / len(priors.order_id.unique()))
prods['users'] = p_grouped.user_id.unique().apply(len)
prods['add_to_cart_order_mean'] = p_grouped.add_to_cart_order.mean().... | instacart/Model.ipynb | bataeves/kaggle | unlicense |
Aisle
~~users~~
~~orders~~
~~order frequency~~
~~reorder rate~~
recency
~~mean/std add_to_cart_order~~
Срок годности/время перепокупки
etc. | prods = pd.DataFrame()
p_grouped = priors.groupby(priors.aisle_id)
prods['orders'] = p_grouped.size().astype(np.float32)
prods['order_freq'] = (prods['orders'] / len(priors.order_id.unique())).astype(np.float32)
prods['users'] = p_grouped.user_id.unique().apply(len).astype(np.float32)
prods['add_to_cart_order_mean'] =... | instacart/Model.ipynb | bataeves/kaggle | unlicense |
Department
~~users~~
~~orders~~
~~order frequency~~
~~reorder rate~~
recency
~~mean add_to_cart_order~~
Срок годности/время перепокупки
etc. | prods = pd.DataFrame()
p_grouped = priors.groupby(priors.department_id)
prods['orders'] = p_grouped.size().astype(np.float32)
prods['order_freq'] = (prods['orders'] / len(priors.order_id.unique())).astype(np.float32)
prods['users'] = p_grouped.user_id.unique().apply(len).astype(np.float32)
prods['add_to_cart_order_mea... | instacart/Model.ipynb | bataeves/kaggle | unlicense |
User Product Interaction (UP)
~~purchases (nb_orders)~~
~~purchases ratio~~
~~reorders~~
~~Average position in cart~~
day since last purchase
average/min/max days between purchase
day since last purchase - срок годности
~~order since last purchase (UP_orders_since_last)~~
Latest one/two/three/four week features
etc. | %%cache userXproduct.pkl userXproduct
priors['z'] = priors.product_id + priors.user_id * 100000
d = dict()
for row in tqdm_notebook(priors.itertuples(), total=len(priors)):
z = row.z
if z not in d:
d[z] = (
1,
(row.order_number, row.order_id),
row.add_to_cart_order,... | instacart/Model.ipynb | bataeves/kaggle | unlicense |
User aisle interaction (UA)
purchases
reorders
day since last purchase
average/min/max days between purchase
order since last purchase
etc.
User department interaction (UD)
purchases
reorders
day since last purchase
average days between purchase
order since last purchase
etc.
User time interaction (UT)
user prefer... | ### build list of candidate products to reorder, with features ###
train_index = set(op_train.index)
def features(selected_orders, labels_given=False):
order_list = []
product_list = []
labels = []
for row in tqdm_notebook(
selected_orders.itertuples(),
total=len(selected_orders)
)... | instacart/Model.ipynb | bataeves/kaggle | unlicense |
Train | f_to_use = [
'user_total_orders', 'user_total_items', 'user_total_distinct_items',
'user_average_days_between_orders', 'user_average_basket',
'order_hour_of_day', 'days_since_prior_order', 'days_since_ratio',
'aisle_id', 'department_id', 'product_orders', 'product_reorders',
'product_reorder_rate', ... | instacart/Model.ipynb | bataeves/kaggle | unlicense |
Predict | def predict(model, df_test, TRESHOLD=0.19):
### build candidates list for test ###
df_test['pred'] = model.predict(feature_select(df_test))
# Нужно добавить https://www.kaggle.com/mmueller/f1-score-expectation-maximization-in-o-n/code
d = dict()
for row in df_test.itertuples():
# Вот тут мо... | instacart/Model.ipynb | bataeves/kaggle | unlicense |
CV
https://www.kaggle.com/happycube/validation-demo-325-cv-3276-lb/notebook | lgb.cv(params, d_train, ROUNDS, nfold=5, verbose_eval=10)
%%cache df_train_gt.pkl df_train_gt
from functools import partial
products_raw = pd.read_csv(IDIR + 'products.csv')
# combine aisles, departments and products (left joined to products)
goods = pd.merge(
left=pd.merge(
left=products_raw,
r... | instacart/Model.ipynb | bataeves/kaggle | unlicense |
Подбираем threshold | for th in np.arange(0.18, 0.22, 0.01):
print th
print cv(threshold=th)
print | instacart/Model.ipynb | bataeves/kaggle | unlicense |
Модель определения кол-ва покупок | prior_orders_count = priors[["order_id", "product_id"]].groupby("order_id").count()
prior_orders_count = prior_orders_count.rename(columns={"product_id": "product_counts"})
train_orders_count = op_train.drop(["product_id", "order_id"], axis=1, errors="ignore")
train_orders_count = train_orders_count.reset_index()[["or... | instacart/Model.ipynb | bataeves/kaggle | unlicense |
Let's look at the tables and columns we have for analysis. Please query the INFORMATION_SCHEMA.
Learning objective 1. | %%with_globals
%%bigquery --project {PROJECT}
SELECT table_name, column_name, data_type
FROM `asl-ml-immersion.stock_src.INFORMATION_SCHEMA.COLUMNS`
ORDER BY table_name, ordinal_position | courses/machine_learning/deepdive2/time_series_prediction/labs/optional_1_data_exploration.ipynb | turbomanage/training-data-analyst | apache-2.0 |
Price History
TODO: Visualize stock symbols from the dataset. | %%with_globals
%%bigquery --project {PROJECT}
SELECT *
FROM `asl-ml-immersion.stock_src.price_history`
LIMIT 10
def query_stock(symbol):
return bq.query('''
# TODO: query a specific stock
'''.format(symbol)).to_dataframe()
df_stock = query_stock('GOOG')
df_stock.Date = pd.to_datetime(df_stock.Date)
ax = ... | courses/machine_learning/deepdive2/time_series_prediction/labs/optional_1_data_exploration.ipynb | turbomanage/training-data-analyst | apache-2.0 |
TODO 2: Compare individual stocks to the S&P 500. | SP500_SYMBOL = gspc
df_sp = query_stock(SP500_SYMBOL)
# TODO: visualize S&P 500 price | courses/machine_learning/deepdive2/time_series_prediction/labs/optional_1_data_exploration.ipynb | turbomanage/training-data-analyst | apache-2.0 |
Let's see how the price of stocks change over time on a yearly basis. Using the LAG function we can compute the change in stock price year-over-year.
Let's compute average close difference for each year. This line could, of course, be done in Pandas. Often times it's useful to use some combination of BigQuery and Pand... | %%with_globals
%%bigquery --project {PROJECT} df
WITH
with_year AS
(
SELECT symbol,
EXTRACT(YEAR FROM date) AS year,
close
FROM `asl-ml-immersion.stock_src.price_history`
WHERE symbol in (SELECT symbol FROM `asl-ml-immersion.stock_src.snp500`)
),
year_aggregated AS
(
SELECT year, s... | courses/machine_learning/deepdive2/time_series_prediction/labs/optional_1_data_exploration.ipynb | turbomanage/training-data-analyst | apache-2.0 |
Stock splits can also impact our data - causing a stock price to rapidly drop. In practice, we would need to clean all of our stock data to account for this. This would be a major effort! Fortunately, in the case of IBM, for example, all stock splits occurred before the year 2000.
Learning objective 2
TODO: Query the I... | stock_symbol = 'IBM'
%%with_globals
%%bigquery --project {PROJECT} df
SELECT date, close
FROM `asl-ml-immersion.stock_src.price_history`
WHERE symbol='{stock_symbol}'
ORDER BY date
# TODO: can you visualize when the major stock splits occured? | courses/machine_learning/deepdive2/time_series_prediction/labs/optional_1_data_exploration.ipynb | turbomanage/training-data-analyst | apache-2.0 |
Making Predictions
First let's generate our own sentences to see how the model classifies them. | def string_to_array(s, separator=' '):
return s.split(separator)
def generate_data_row(sentence, label, max_length):
sequence = np.zeros((max_length), dtype='int32')
for i, word in enumerate(string_to_array(sentence)):
sequence[i] = word_list.index(word)
return sequence, label, max_length
... | code_samples/RNN/sentiment_analysis/.ipynb_checkpoints/SentimentAnalysis-New-checkpoint.ipynb | mari-linhares/tensorflow-workshop | apache-2.0 |
Data
CoTeDe comes with a few datasets for demonstration.
Here we will use a CTD cast from the PIRATA hydrographic collection, i.e. measurements from the Tropical Atlantic Ocean.
If curious about this data, check CoTeDe's documentation for more details.
Let's start by loading this dataset. | data = cotede.datasets.load_ctd()
print("There is a total of {} observed levels.\n".format(len(data["TEMP"])))
print("The variables are: ", list(data.keys())) | docs/notebooks/anomaly_detection/anomaly_detection_profile.ipynb | castelao/CoTeDe | bsd-3-clause |
This CTD was equipped with backup sensors to provide more robustness. Measurements from the secondary sensor are identified by a 2 in the end of the name, such as "TEMP2" for the secondary temperature sensor. Let's focus on the primary sensors.
To visualize this profile we will use Bokeh so you can explore the data and... | p1 = figure(plot_width=420, plot_height=600)
p1.circle(
data['TEMP'], -data['PRES'], size=8, line_color="seagreen",
fill_color="mediumseagreen", fill_alpha=0.3)
p1.xaxis.axis_label = "Temperature [C]"
p1.yaxis.axis_label = "Depth [m]"
p2 = figure(plot_width=420, plot_height=600)
p2.y_range = p1.y_range
p2.circ... | docs/notebooks/anomaly_detection/anomaly_detection_profile.ipynb | castelao/CoTeDe | bsd-3-clause |
Considering the unusual magnitudes and variability near the bottom, there are clearly bad measurements on this profile.
Let's start with a traditional QC and then we'll include the Anomaly Detection.
Traditional QC with CoTeDe framework
NOTE: If you are not familiar with CoTeDe, it might be helpfull to check first the ... | pqc = cotede.ProfileQC(data, cfg="eurogoos") | docs/notebooks/anomaly_detection/anomaly_detection_profile.ipynb | castelao/CoTeDe | bsd-3-clause |
That's it, the temperature and salinity from the primary and secondary sensors were all evaluated.
Which criteria were flagged for the primary sensors? | print("Flags for temperature:\n {}\n".format(list(pqc.flags["TEMP"].keys())))
print("Flags for salinity:\n {}\n".format(list(pqc.flags["PSAL"].keys()))) | docs/notebooks/anomaly_detection/anomaly_detection_profile.ipynb | castelao/CoTeDe | bsd-3-clause |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.