markdown stringlengths 0 1.02M | code stringlengths 0 832k | output stringlengths 0 1.02M | license stringlengths 3 36 | path stringlengths 6 265 | repo_name stringlengths 6 127 |
|---|---|---|---|---|---|
TASK 4, Part 2**Instructions**The advertising campaign mentioned in Task 2 was successful as such the team wants more information on the average prices. Use a pivot table to look at the average prices for different room types within each neighbourhood. https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.... | # YOUR CODE FOR TASK 4, PART 2
#Your code here
#Pivot = ...
Pivot = df_brooklyn.pivot_table(index=['neighbourhood'], values=['price'], columns=['room_type'], aggfunc=np.mean, fill_value = 0)
Pivot | _____no_output_____ | MIT | Regressions/AirBnB Price prediction/.ipynb_checkpoints/Apprentice_Challenge_2021_Answers-Copy1-checkpoint.ipynb | shobhit009/Machine_Learning_Projects |
TASK 5, Part 1**Instructions**The Airbnb analysts want to know the factors influencing the price. Before proceedeing with Correlation analysis, you need to perform some feature engineering tasks such as converting the categorical columns, dropping descriptive columns.1. Encode the categorical variable `room_type` and ... | # YOUR CODE FOR TASK 5, PART 1
# encode the columnns room_type and neighbourhood
# df_brooklyn_rt = ...
df_brooklyn_rt = pd.get_dummies(df_brooklyn, columns=['room_type','neighbourhood'])
df_brooklyn_rt
#drop the descriptive columns from the dataframe
# df_brooklyn_rt = ...
df_brooklyn_rt = df_brooklyn_rt.drop(['name'... | df is correct
| MIT | Regressions/AirBnB Price prediction/.ipynb_checkpoints/Apprentice_Challenge_2021_Answers-Copy1-checkpoint.ipynb | shobhit009/Machine_Learning_Projects |
TASK 5, Part 2**Instructions**We will now study the correlation of the features in the dataset with `price`. Use Pandas dataframe.corr() to find the pairwise correlation of all columns in the dataframe. Use pandas corr() function to create correlation dataframe.Function syntax : new_dataframe = Dataframe.corr()Visuali... | # YOUR CODE FOR TASK 5, PART 2
# create a correlation matix
# corr = ...
corr = df_brooklyn_rt.corr()
# plot the heatmap
# sns.heatmap(...)
sns.heatmap(corr, xticklabels=corr.columns, yticklabels=corr.columns)
| _____no_output_____ | MIT | Regressions/AirBnB Price prediction/.ipynb_checkpoints/Apprentice_Challenge_2021_Answers-Copy1-checkpoint.ipynb | shobhit009/Machine_Learning_Projects |
Multicollinearity occurs when your data includes multiple attributes that are correlated not just to your target variable, but also to each other. **Based on the correlation matrix, answer the following:1. Which columns would you drop to prevent multicollinearity? Sample Answer: brooklyn_whole or number_of_reviews2. Wh... | ## RUN THIS CELL AS-IS TO CHECK IF YOUR OUTPUTS ARE CORRECT. IF THEY ARE NOT,
## THE APPROPRIATE OBJECTS WILL BE LOADED IN TO ENSURE THAT YOU CAN CONTINUE
## WITH THE ASSESSMENT.
task_5_part_2_check = data_load_files.TASK_5_PART_2_OUTPUT
corr_shape_check = (corr.shape == task_5_part_2_check.shape)
if corr_shape... | df is correct
| MIT | Regressions/AirBnB Price prediction/.ipynb_checkpoints/Apprentice_Challenge_2021_Answers-Copy1-checkpoint.ipynb | shobhit009/Machine_Learning_Projects |
TASK 6Property Hosts are expected to set their own prices for their listings. Although Airbnb provide some general guidance, there are currently no services which help hosts price their properties using range of data points.Airbnb pricing is important to get right, particularly in big cities like New York where there ... | #Your code here
#X =
#Y =
#Solution
X = df_brooklyn_rt.drop(['price', 'last_review', 'brooklyn_whole','price_category'], axis = 1)
Y = df_brooklyn_rt['price']
#Your code here
#Please don't change the test_size value it should remain 0.2
#X_train, X_test, Y_train, Y_test = train_test_split(<....your X value here...>,... | dfs are correct
| MIT | Regressions/AirBnB Price prediction/.ipynb_checkpoints/Apprentice_Challenge_2021_Answers-Copy1-checkpoint.ipynb | shobhit009/Machine_Learning_Projects |
Task 6 Part 2**Instructions**Training the modelWe use scikit-learn’s LinearRegression to train our model. Using the fit() method, we will pass the training datasets X_train and Y_train as arguments to the linear regression model. Testing the modelThe model has learnt about the dataset. We will now use the trained mode... | #Run this cell
lin_model = LinearRegression()
#Training the model
#Your code here
#lin_model.fit(X_argument, Y_argument)
#Solution
lin_model.fit(X_train, Y_train)
#Testing the model
#Your code here
#y_test_predict = lin_model.predict(...X test Dataset...)
#Solution
y_test_predict = lin_model.predict(X_test)
#Run th... | Model evaluation is correct
| MIT | Regressions/AirBnB Price prediction/.ipynb_checkpoints/Apprentice_Challenge_2021_Answers-Copy1-checkpoint.ipynb | shobhit009/Machine_Learning_Projects |
Task 6 Part 3**Instructions**Now we will compare the actual output values for X_test with the predicted values using a bar chart.- 1(pt) Create a new dataframe lr_pred_df using the Y_test and y_test_predict- 1(pt) Use first 20 records from the dataframe lr_pred_df and plot a bar graph showing comparision of actual and... | #Actual Vs Predicted for Linear Regression
#Your code here
#lr_pred_df =
#Solution
lr_pred_df = pd.DataFrame({
'actual_values': np.array(Y_test).flatten(),
'y_test_predict': y_test_predict.flatten()})
#Your code here
#lr_pred_df.plot()
#Solution
lr_pred_df = lr_pred_df.head(20)
plt = lr_pred_df.plot... | df is correct
| MIT | Regressions/AirBnB Price prediction/.ipynb_checkpoints/Apprentice_Challenge_2021_Answers-Copy1-checkpoint.ipynb | shobhit009/Machine_Learning_Projects |
Lab: Connect to Db2 database on Cloud using Python IntroductionThis notebook illustrates how to access a DB2 database on Cloud using Python by following the steps below:1. Import the `ibm_db` Python library1. Enter the database connection credentials1. Create the database connection1. Close the database connection__No... | import ibm_db | _____no_output_____ | MIT | 5. Databases_SQL/1-1-Connecting-v4-py.ipynb | naquech/IBM_Watson_Studio |
When the command above completes, the `ibm_db` library is loaded in your notebook. Identify the database connection credentialsConnecting to dashDB or DB2 database requires the following information:* Driver Name* Database name * Host DNS name or IP address * Host port* Connection protocol* User ID (or username)* User... | #Replace the placeholder values with your actual Db2 hostname, username, and password:
dsn_hostname = "YourDb2Hostname" # e.g.: "dashdb-txn-sbox-yp-dal09-04.services.dal.bluemix.net"
dsn_uid = "YourDb2Username" # e.g. "abc12345"
dsn_pwd = "YoueDb2Password" # e.g. "7dBZ3wWt9XN6$o0J"
dsn_driver = "{IBM DB2 O... | _____no_output_____ | MIT | 5. Databases_SQL/1-1-Connecting-v4-py.ipynb | naquech/IBM_Watson_Studio |
Create the DB2 database connectionIbm_db API uses the IBM Data Server Driver for ODBC and CLI APIs to connect to IBM DB2 and Informix.Lets build the dsn connection string using the credentials you entered above | #DO NOT MODIFY THIS CELL. Just RUN it with Shift + Enter
#Create the dsn connection string
dsn = (
"DRIVER={0};"
"DATABASE={1};"
"HOSTNAME={2};"
"PORT={3};"
"PROTOCOL={4};"
"UID={5};"
"PWD={6};").format(dsn_driver, dsn_database, dsn_hostname, dsn_port, dsn_protocol, dsn_uid, dsn_pwd)
#print... | DRIVER=DATABASE=BLUDB;HOSTNAME=dashdb-txn-sbox-yp-dal09-03.services.dal.bluemix.net;PORT=50000;PROTOCOL=TCPIP;UID=wvb91528;PWD=tm^1nlbn4dj3j04b;;DATABASE=BLUDB;HOSTNAME=dashdb-txn-sbox-yp-dal09-03.services.dal.bluemix.net;PORT=50000;PROTOCOL=TCPIP;UID=wvb91528;PWD=tm^1nlbn4dj3j04b;
| MIT | 5. Databases_SQL/1-1-Connecting-v4-py.ipynb | naquech/IBM_Watson_Studio |
Now establish the connection to the database | #DO NOT MODIFY THIS CELL. Just RUN it with Shift + Enter
#Create database connection
try:
conn = ibm_db.connect(dsn, "", "")
print ("Connected to database: ", dsn_database, "as user: ", dsn_uid, "on host: ", dsn_hostname)
except:
print ("Unable to connect: ", ibm_db.conn_errormsg() )
| Connected to database: BLUDB as user: wvb91528 on host: dashdb-txn-sbox-yp-dal09-03.services.dal.bluemix.net
| MIT | 5. Databases_SQL/1-1-Connecting-v4-py.ipynb | naquech/IBM_Watson_Studio |
Congratulations if you were able to connect successfuly. Otherwise check the error and try again. | #Retrieve Metadata for the Database Server
server = ibm_db.server_info(conn)
print ("DBMS_NAME: ", server.DBMS_NAME)
print ("DBMS_VER: ", server.DBMS_VER)
print ("DB_NAME: ", server.DB_NAME)
#Retrieve Metadata for the Database Client / Driver
client = ibm_db.client_info(conn)
print ("DRIVER_NAME: ", clien... | DRIVER_NAME: libdb2.a
DRIVER_VER: 11.01.0404
DATA_SOURCE_NAME: BLUDB
DRIVER_ODBC_VER: 03.51
ODBC_VER: 03.01.0000
ODBC_SQL_CONFORMANCE: EXTENDED
APPL_CODEPAGE: 1208
CONN_CODEPAGE: 1208
| MIT | 5. Databases_SQL/1-1-Connecting-v4-py.ipynb | naquech/IBM_Watson_Studio |
Close the ConnectionWe free all resources by closing the connection. Remember that it is always important to close connections so that we can avoid unused connections taking up resources. | ibm_db.close(conn) | _____no_output_____ | MIT | 5. Databases_SQL/1-1-Connecting-v4-py.ipynb | naquech/IBM_Watson_Studio |
import numpy as np
import pandas as pd
import seaborn as sns
sns.set()
np.__version__
def fetch_financial_data(company='AMZN'):
import pandas_datareader.data as web
return web.DataReader(name=company, data_source='stooq')
google = fetch_financial_data(company='GOOGL')
google
google.info()
pd.set_option('prec... | _____no_output_____ | MIT | Pandas121.ipynb | mariuszkr33/dw_matrix | |
Setup environment | from monai.utils import first, set_determinism
from monai.transforms import (
AsDiscrete,
AsDiscreted,
EnsureChannelFirstd,
Compose,
CropForegroundd,
LoadImaged,
Orientationd,
RandCropByPosNegLabeld,
ScaleIntensityRanged,
Spacingd,
EnsureTyped,
EnsureType,
Invertd,
)
... | _____no_output_____ | Apache-2.0 | tutorials/segmentation/spleen_segmentation_3d.ipynb | YipengHu/MPHY0043 |
Setup imports | # Copyright 2020 MONAI Consortium
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
# http://www.apache.org/licenses/LICENSE-2.0
# Unless required by applicable law or agreed to in writing, s... | _____no_output_____ | Apache-2.0 | tutorials/segmentation/spleen_segmentation_3d.ipynb | YipengHu/MPHY0043 |
Setup data directoryYou can specify a directory with the `MONAI_DATA_DIRECTORY` environment variable. This allows you to save results and reuse downloads. If not specified a temporary directory will be used. | directory = os.environ.get("MONAI_DATA_DIRECTORY")
root_dir = tempfile.mkdtemp() if directory is None else directory
print(root_dir) | _____no_output_____ | Apache-2.0 | tutorials/segmentation/spleen_segmentation_3d.ipynb | YipengHu/MPHY0043 |
Download datasetDownloads and extracts the dataset. The dataset comes from http://medicaldecathlon.com/. | resource = "https://msd-for-monai.s3-us-west-2.amazonaws.com/Task09_Spleen.tar"
md5 = "410d4a301da4e5b2f6f86ec3ddba524e"
compressed_file = os.path.join(root_dir, "Task09_Spleen.tar")
data_dir = os.path.join(root_dir, "Task09_Spleen")
if not os.path.exists(data_dir):
download_and_extract(resource, compressed_file, ... | _____no_output_____ | Apache-2.0 | tutorials/segmentation/spleen_segmentation_3d.ipynb | YipengHu/MPHY0043 |
Set MSD Spleen dataset path | train_images = sorted(
glob.glob(os.path.join(data_dir, "imagesTr", "*.nii.gz")))
train_labels = sorted(
glob.glob(os.path.join(data_dir, "labelsTr", "*.nii.gz")))
data_dicts = [
{"image": image_name, "label": label_name}
for image_name, label_name in zip(train_images, train_labels)
]
train_files, val_f... | _____no_output_____ | Apache-2.0 | tutorials/segmentation/spleen_segmentation_3d.ipynb | YipengHu/MPHY0043 |
Set deterministic training for reproducibility | set_determinism(seed=0) | _____no_output_____ | Apache-2.0 | tutorials/segmentation/spleen_segmentation_3d.ipynb | YipengHu/MPHY0043 |
Setup transforms for training and validationHere we use several transforms to augment the dataset:1. `LoadImaged` loads the spleen CT images and labels from NIfTI format files.1. `AddChanneld` as the original data doesn't have channel dim, add 1 dim to construct "channel first" shape.1. `Spacingd` adjusts the spacing ... | train_transforms = Compose(
[
LoadImaged(keys=["image", "label"]),
EnsureChannelFirstd(keys=["image", "label"]),
Spacingd(keys=["image", "label"], pixdim=(
1.5, 1.5, 2.0), mode=("bilinear", "nearest")),
Orientationd(keys=["image", "label"], axcodes="RAS"),
ScaleIn... | _____no_output_____ | Apache-2.0 | tutorials/segmentation/spleen_segmentation_3d.ipynb | YipengHu/MPHY0043 |
Check transforms in DataLoader | check_ds = Dataset(data=val_files, transform=val_transforms)
check_loader = DataLoader(check_ds, batch_size=1)
check_data = first(check_loader)
image, label = (check_data["image"][0][0], check_data["label"][0][0])
print(f"image shape: {image.shape}, label shape: {label.shape}")
# plot the slice [:, :, 80]
plt.figure("c... | _____no_output_____ | Apache-2.0 | tutorials/segmentation/spleen_segmentation_3d.ipynb | YipengHu/MPHY0043 |
Define CacheDataset and DataLoader for training and validationHere we use CacheDataset to accelerate training and validation process, it's 10x faster than the regular Dataset. To achieve best performance, set `cache_rate=1.0` to cache all the data, if memory is not enough, set lower value. Users can also set `cache_... | train_ds = CacheDataset(
data=train_files, transform=train_transforms,
cache_rate=1.0, num_workers=4)
# train_ds = monai.data.Dataset(data=train_files, transform=train_transforms)
# use batch_size=2 to load images and use RandCropByPosNegLabeld
# to generate 2 x 4 images for network training
train_loader = Dat... | _____no_output_____ | Apache-2.0 | tutorials/segmentation/spleen_segmentation_3d.ipynb | YipengHu/MPHY0043 |
Create Model, Loss, Optimizer | # standard PyTorch program style: create UNet, DiceLoss and Adam optimizer
device = torch.device("cpu")
model = UNet(
dimensions=3,
in_channels=1,
out_channels=2,
channels=(16, 32, 64, 128, 256),
strides=(2, 2, 2, 2),
num_res_units=2,
norm=Norm.BATCH,
).to(device)
loss_function = DiceLoss(to... | _____no_output_____ | Apache-2.0 | tutorials/segmentation/spleen_segmentation_3d.ipynb | YipengHu/MPHY0043 |
Execute a typical PyTorch training process | max_epochs = 600
val_interval = 2
best_metric = -1
best_metric_epoch = -1
epoch_loss_values = []
metric_values = []
post_pred = Compose([EnsureType(), AsDiscrete(argmax=True, to_onehot=True, n_classes=2)])
post_label = Compose([EnsureType(), AsDiscrete(to_onehot=True, n_classes=2)])
for epoch in range(max_epochs):
... | _____no_output_____ | Apache-2.0 | tutorials/segmentation/spleen_segmentation_3d.ipynb | YipengHu/MPHY0043 |
Plot the loss and metric | plt.figure("train", (12, 6))
plt.subplot(1, 2, 1)
plt.title("Epoch Average Loss")
x = [i + 1 for i in range(len(epoch_loss_values))]
y = epoch_loss_values
plt.xlabel("epoch")
plt.plot(x, y)
plt.subplot(1, 2, 2)
plt.title("Val Mean Dice")
x = [val_interval * (i + 1) for i in range(len(metric_values))]
y = metric_values
... | _____no_output_____ | Apache-2.0 | tutorials/segmentation/spleen_segmentation_3d.ipynb | YipengHu/MPHY0043 |
Check best model output with the input image and label | model.load_state_dict(torch.load(
os.path.join(root_dir, "best_metric_model.pth")))
model.eval()
with torch.no_grad():
for i, val_data in enumerate(val_loader):
roi_size = (160, 160, 160)
sw_batch_size = 4
val_outputs = sliding_window_inference(
val_data["image"].to(device), ... | _____no_output_____ | Apache-2.0 | tutorials/segmentation/spleen_segmentation_3d.ipynb | YipengHu/MPHY0043 |
Evaluation on original image spacings | val_org_transforms = Compose(
[
LoadImaged(keys=["image", "label"]),
EnsureChannelFirstd(keys=["image", "label"]),
Spacingd(keys=["image"], pixdim=(
1.5, 1.5, 2.0), mode="bilinear"),
Orientationd(keys=["image"], axcodes="RAS"),
ScaleIntensityRanged(
ke... | _____no_output_____ | Apache-2.0 | tutorials/segmentation/spleen_segmentation_3d.ipynb | YipengHu/MPHY0043 |
Cleanup data directoryRemove directory if a temporary was used. | if directory is None:
shutil.rmtree(root_dir) | _____no_output_____ | Apache-2.0 | tutorials/segmentation/spleen_segmentation_3d.ipynb | YipengHu/MPHY0043 |
View source on GitHub Notebook Viewer Run in Google Colab Install Earth Engine API and geemapInstall the [Earth Engine Python API](https://developers.google.com/earth-engine/python_install) and [geemap](https://github.com/giswqs/geemap). The **geemap** Python package is built upon the [ipyleaflet](https://... | # Installs geemap package
import subprocess
try:
import geemap
except ImportError:
print('geemap package not installed. Installing ...')
subprocess.check_call(["python", '-m', 'pip', 'install', 'geemap'])
# Checks whether this notebook is running on Google Colab
try:
import google.colab
import gee... | _____no_output_____ | MIT | FeatureCollection/set_properties.ipynb | c11/earthengine-py-notebooks |
Create an interactive map The default basemap is `Google Satellite`. [Additional basemaps](https://github.com/giswqs/geemap/blob/master/geemap/geemap.pyL13) can be added using the `Map.add_basemap()` function. | Map = emap.Map(center=[40,-100], zoom=4)
Map.add_basemap('ROADMAP') # Add Google Map
Map | _____no_output_____ | MIT | FeatureCollection/set_properties.ipynb | c11/earthengine-py-notebooks |
Add Earth Engine Python script | # Add Earth Engine dataset
# Make a feature and set some properties.
feature = ee.Feature(ee.Geometry.Point([-122.22599, 37.17605])) \
.set('genus', 'Sequoia').set('species', 'sempervirens')
# Get a property from the feature.
species = feature.get('species')
print(species.getInfo())
# Set a new property.
feature = ... | _____no_output_____ | MIT | FeatureCollection/set_properties.ipynb | c11/earthengine-py-notebooks |
Display Earth Engine data layers | Map.addLayerControl() # This line is not needed for ipyleaflet-based Map.
Map | _____no_output_____ | MIT | FeatureCollection/set_properties.ipynb | c11/earthengine-py-notebooks |
DEMO_TOY_IMAGES Simple illustration of GLFM pipeline, replicating the example of the IBP linear-Gaussian model in (Griffiths and Ghahramani, 2011). | # ---------------------------------------------
# Import necessary libraries
# ---------------------------------------------
import numpy as np # import numpy matrix for calculus with matrices
import matplotlib.pyplot as plt # import plotting library
import time # import time to be able to measure iteration spee... | Print inferred latent features...
| MIT | demos/python/demo_toy_images.ipynb | ferjorosa/test-glfm |
Codecademy Completion This problem will be used for verifying that you have completed the Python course on http://www.codecademy.com/.Here are the steps to do this verification:1. Go to the page on http://www.codecademy.com/ that shows your percent completion.2. Take a screen shot of that page.3. Name the file `codeca... | from IPython.display import Image
Image(filename='codecademy.png', width='100%') | _____no_output_____ | MIT | assignments/assignment01/Codecademy.ipynb | edwardd1/phys202-2015-work |
*Copyright 2020 Google LLC**Licensed under the Apache License, Version 2.0 (the "License")* | # Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# https://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the L... | _____no_output_____ | Apache-2.0 | retrain_detection_qat_tf1.ipynb | KeithAzzopardi1998/tutorials |
Retrain a detection model for Edge TPU with quant-aware training (TF 1.12) This notebook uses a set of TensorFlow training scripts to perform transfer-learning on a quantization-aware object detection model and then convert it for compatibility with the [Edge TPU](https://coral.ai/products/).Specifically, this tutoria... | ! pip uninstall tensorflow -y
! pip install tensorflow==1.12
import tensorflow as tf
print(tf.__version__) | _____no_output_____ | Apache-2.0 | retrain_detection_qat_tf1.ipynb | KeithAzzopardi1998/tutorials |
Clone the model and training repos | ! git clone https://github.com/tensorflow/models.git
! cd models && git checkout f788046ca876a8820e05b0b48c1fc2e16b0955bc
! git clone https://github.com/google-coral/tutorials.git
! cp -r tutorials/docker/object_detection/scripts/* models/research/ | _____no_output_____ | Apache-2.0 | retrain_detection_qat_tf1.ipynb | KeithAzzopardi1998/tutorials |
Import dependencies For details, see https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/installation.md | ! apt-get install -y python python-tk
! pip install Cython contextlib2 pillow lxml jupyter matplotlib
# Get protoc 3.0.0, rather than the old version already in the container
! wget https://www.github.com/google/protobuf/releases/download/v3.0.0/protoc-3.0.0-linux-x86_64.zip
! unzip protoc-3.0.0-linux-x86_64.zip -d pro... | _____no_output_____ | Apache-2.0 | retrain_detection_qat_tf1.ipynb | KeithAzzopardi1998/tutorials |
Just to verify everything is correctly set up: | ! python object_detection/builders/model_builder_test.py | _____no_output_____ | Apache-2.0 | retrain_detection_qat_tf1.ipynb | KeithAzzopardi1998/tutorials |
Convert training data to TFRecord To train with different images, read [how to configure your own training data](https://coral.ai/docs/edgetpu/retrain-detection/configure-your-own-training-data). | ! ./prepare_checkpoint_and_dataset.sh --network_type mobilenet_v1_ssd --train_whole_model false | _____no_output_____ | Apache-2.0 | retrain_detection_qat_tf1.ipynb | KeithAzzopardi1998/tutorials |
Perform transfer-learning The following script takes several hours to finish in Colab. (You can shorten by reducing the steps, but that reduces the final accuracy.)If you didn't already select "Run all" then you should run all remaining cells now. That will ensure the rest of the notebook completes while you are away,... | %env NUM_TRAINING_STEPS=500
%env NUM_EVAL_STEPS=100
# If you're retraining the whole model, we suggest thes values:
# %env NUM_TRAINING_STEPS=50000
# %env NUM_EVAL_STEPS=2000
! ./retrain_detection_model.sh --num_training_steps $NUM_TRAINING_STEPS --num_eval_steps $NUM_EVAL_STEPS | _____no_output_____ | Apache-2.0 | retrain_detection_qat_tf1.ipynb | KeithAzzopardi1998/tutorials |
As training progresses, you can see new checkpoint files appear in the `models/research/learn_pet/train/` directory. Compile for the Edge TPU | ! ./convert_checkpoint_to_edgetpu_tflite.sh --checkpoint_num $NUM_TRAINING_STEPS
! curl https://packages.cloud.google.com/apt/doc/apt-key.gpg | sudo apt-key add -
! echo "deb https://packages.cloud.google.com/apt coral-edgetpu-stable main" | sudo tee /etc/apt/sources.list.d/coral-edgetpu.list
! sudo apt-get update
!... | _____no_output_____ | Apache-2.0 | retrain_detection_qat_tf1.ipynb | KeithAzzopardi1998/tutorials |
Download the files: | from google.colab import files
files.download('output_tflite_graph_edgetpu.tflite')
files.download('labels.txt') | _____no_output_____ | Apache-2.0 | retrain_detection_qat_tf1.ipynb | KeithAzzopardi1998/tutorials |
Initial Processing & EDABy: Aditya Mengani, Ognjen Sosa, Sanjay Elangovan, Song Park, Sophia Skowronski | '''Importing basic data analysis packages'''
import numpy as np
import pandas as pd
import csv
import warnings
import os
import time
import math
warnings.filterwarnings('ignore')
'''Plotting packages'''
import matplotlib.pyplot as plt
%matplotlib inline
import seaborn as sns
sns.set(font_scale=1.3) | _____no_output_____ | MIT | 1_4_EDA.ipynb | aditya-mengani/network_analysis_cbase_p1_ml |
Function: memory reduction of dataframe | def reduce_mem_usage(df, verbose=True):
numerics = ['int16', 'int32', 'int64', 'float16', 'float32', 'float64']
start_mem = df.memory_usage().sum() / 1024**2
for col in df.columns:
col_type = df[col].dtypes
if col_type in numerics:
c_min = df[col].min()
c_max = df... | _____no_output_____ | MIT | 1_4_EDA.ipynb | aditya-mengani/network_analysis_cbase_p1_ml |
Load DataUse `tar -xvzf 20200908_bulk_export.tar.gz` to unzip Crunchbase export (for Windows)Check out summary of data from Crunchbase export here. | ###########################
# Pledge 1% Company UUIDs #
###########################
print('='*100)
p1 = pd.read_csv('~/Desktop/w207/Project/Data/p1.csv')
print('PLEDGE 1%/p1 cols: {}\nSHAPE: {}'.format(p1.columns.to_list(), p1.shape))
p1 = reduce_mem_usage(p1)
#################
# Organizations #
#################
pri... | _____no_output_____ | MIT | 1_4_EDA.ipynb | aditya-mengani/network_analysis_cbase_p1_ml |
Explore Degree Data | df2 = pd.merge(df.copy(),degrees.copy(),how='outer',on='uuid')
pledge1_2 = df2[df2['p1_tag'] == 1].sort_values('p1_date')
pledge1_2.head(10)
# Exclude rows that have NaN institution_uuid
pledge1_2_degrees = pledge1_2[~pledge1_2['institution_name'].isna()]
df2_degrees = df2[~df2['institution_name'].isna()]
# Create cou... | _____no_output_____ | MIT | 1_4_EDA.ipynb | aditya-mengani/network_analysis_cbase_p1_ml |
Can't create plots with degree data because degree data is missing for all P1 companies...:( | # Barplots
_, ax = plt.subplots(nrows=1, ncols=2, figsize=(12, 12), sharey=True)
#sns.barplot(x='p1_tag', y='institution_name', data=pledge1_2_degrees, orient='h', ax=ax[0])
sns.barplot(x='count', y='institution_name', data=df2_degrees, orient='h', ax=ax[1])
# Labels
ax[0].set_title('Pledge Companies by Employee Count... | _____no_output_____ | MIT | 1_4_EDA.ipynb | aditya-mengani/network_analysis_cbase_p1_ml |
Project DescriptionThis is a chatbot that provides live daily news headlines from the Wall Street Journal, with main three categories of news : political, deals, and economy. First, it gives you a brief view about titles after you ask for certain type of news, then it can provide more details in a specific news based o... | import string
import random
import nltk
import requests
import json
news_request = requests.get(url='https://newsapi.org/v2/top-headlines?sources=the-wall-street-journal&apiKey=df21f07e419c41feb602fb9ba2a8456c')
news_dict = news_request.json()['articles']
def is_question(input_string):
"""Check if the input is a ... | _____no_output_____ | MIT | Wu, You-Final Project-vF.ipynb | fionacandicewu/cogs18-finalproject-chatbot |
Matrizeak: zerrenden zerrendakMatrize baten moduko egitura sor daiteke zerrenden zerrendekin: | A = [[1,2,3],[4,5,6],[7,8,9]]
print(A)
print(A[1][2])
# Hau okerra litzateke
#print(A[1,2])
# Lerro berriak gehituz, argiagoa.
A = [
[1,2,3],
[4,5,6],
[7,8,9]
]
print(A)
print(A[1][2])
# Baina... ez da matrize bat!
A[1].append(100)
print(A)
def matrizea_idatzi(M):
ilara = len(M)
zutabe = len(M[0])
... | 1 2 3
4 5 6 100
7 8 9
| MIT | Gardenkiak/Programazioa/.ipynb_checkpoints/ZerrendenZerrendak-checkpoint.ipynb | mpenagar/Konputaziorako-Sarrera |
Kontuz espresio literal oso trinkoekin... | print([0]*3)
print([[0]*3])
print([[0]*3]*3)
A = [[0]*3]*3
matrizea_idatzi(A)
A[1][1] = 5
matrizea_idatzi(A)
print(A[0] is A[1], A[0] is A[2], A[1] is A[2]) | True True True
| MIT | Gardenkiak/Programazioa/.ipynb_checkpoints/ZerrendenZerrendak-checkpoint.ipynb | mpenagar/Konputaziorako-Sarrera |
Match Analysis | import numpy as np
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
%matplotlib inline | _____no_output_____ | MIT | KKR VS MI/Match Analysis KKR VS MI.ipynb | tacklesta/WPL |
Data Cleaning and Exploring | matches = pd.read_csv("matches.csv" , index_col = "id")
matches = matches.iloc[:,:-3]
matches.head()
matches.shape
matches.winner.unique() | _____no_output_____ | MIT | KKR VS MI/Match Analysis KKR VS MI.ipynb | tacklesta/WPL |
Taking in consideration only KKR VS MI matches | KM =matches[np.logical_or(np.logical_and(matches['team1']=='Kolkata Knight Riders',matches['team2']=='Mumbai Indians'),
np.logical_and(matches['team2']=='Kolkata Knight Riders',matches['team1']=='Mumbai Indians'))]
KM.head()
KM.shape
KM.season.unique()
KM.isnull().sum()
KM.describe().iloc[:,... | _____no_output_____ | MIT | KKR VS MI/Match Analysis KKR VS MI.ipynb | tacklesta/WPL |
HEAD TO HEAD | KM.groupby("winner")["winner"].count()
sns.countplot(KM["winner"])
plt.text(-0.09,17,str(KM['winner'].value_counts()['Mumbai Indians']),size=20,color='white')
plt.text(0.95,4,str(KM['winner'].value_counts()['Kolkata Knight Riders']),size=20,color='white')
plt.xlabel('Winner',fontsize=15)
plt.ylabel('No. of Matches',fon... | Season wise winner of matches between KKR and MI :
| MIT | KKR VS MI/Match Analysis KKR VS MI.ipynb | tacklesta/WPL |
Winning Percentage | Winning_Percentage = KM['winner'].value_counts()/len(KM['winner'])
print(" MI winning percentage against KKR(overall) : {}%".format(int(round(Winning_Percentage[0]*100))))
print("KKR winning percentage against MI(overall) : {}%".format(int(round(Winning_Percentage[1]*100)))) | MI winning percentage against KKR(overall) : 76%
KKR winning percentage against MI(overall) : 24%
| MIT | KKR VS MI/Match Analysis KKR VS MI.ipynb | tacklesta/WPL |
Performance Based Analysis | def performance( team_name , given_df ):
for value in given_df.groupby('winner'):
if value[0] == team_name:
total_win_by_runs = sum(list(value[1]['win_by_runs']))
total_win_by_wickets = sum(list(value[1]['win_by_wickets']))
if 0 in list(value[1]['win_by_runs... | Number of times given team win while defending : 8
Number of times given team win while chasing : 11
Average runs by which a given team wins while defending : 40.0
Average wickets by which a given team wins while chasing : 6.0
| MIT | KKR VS MI/Match Analysis KKR VS MI.ipynb | tacklesta/WPL |
Toss Analysis | Toss_Decision_based_Winner = pd.DataFrame(KM.groupby(['toss_winner',"toss_decision","winner"])["winner"].count())
print(" No of times toss winning decision leading to match winning : ")
Toss_Decision_based_Winner
Toss_Decision = pd.DataFrame(KM.groupby(['toss_winner',"toss_decision"])["toss_decision"].count())
print ... | _____no_output_____ | MIT | KKR VS MI/Match Analysis KKR VS MI.ipynb | tacklesta/WPL |
From the above analysis we can see that mostly both the teams prefer chasing the score after winning the toss | sns.set(style='whitegrid')
plt.figure(figsize = (18,9))
sns.countplot(KM['toss_winner'],hue=KM['winner'])
plt.title('Match Winner vs Toss Winner statistics for both team',fontsize=15)
plt.yticks(fontsize=15)
plt.xticks(fontsize=15)
plt.xlabel('Toss winner',fontsize=15)
plt.ylabel('Match Winner',fontsize=15)
plt.legend(... | _____no_output_____ | MIT | KKR VS MI/Match Analysis KKR VS MI.ipynb | tacklesta/WPL |
Toss Decision based Analysis of both the teams seperately : | KKR = KM[KM["toss_winner"]=="Kolkata Knight Riders"]
MI = KM[KM["toss_winner"]=="Mumbai Indians"]
sns.set(style='whitegrid')
plt.figure(figsize = (18,9))
sns.countplot(KKR['toss_decision'],hue=KKR['winner'])
plt.title('Match Winner vs Toss Winner statistics for KKR',fontsize=15)
plt.yticks(fontsize=15)
plt.xticks(fonts... | Man of the match :
| MIT | KKR VS MI/Match Analysis KKR VS MI.ipynb | tacklesta/WPL |
Recent Year Performance Analysis | cond1 = KM["season"] == 2015
cond2 = KM["season"] == 2016
cond3 = KM["season"] == 2017
cond4 = KM["season"] == 2018
cond5 = KM["season"] == 2019
final = KM[cond1 | cond2 | cond3 | cond4 | cond5]
final
final.shape
player = pd.DataFrame(final.player_of_match.value_counts())
print("Man of the match :")
player
plt.figure(... | _____no_output_____ | MIT | KKR VS MI/Match Analysis KKR VS MI.ipynb | tacklesta/WPL |
Interpolation & Fitting 1. Libraries | import matplotlib.pyplot as plt
import numpy as np
import statsmodels.api as sm
from matplotlib.cm import colors
from scipy.interpolate import interp1d, lagrange
from scipy.optimize import curve_fit
from statsmodels.nonparametric.kernel_regression import KernelReg | _____no_output_____ | MIT | plotting/ex_interpolation.ipynb | nathanielng/python-snippets |
2. Calculations 2.1 Original data | # Original data points
x = np.array([1.0, 2.0, 3.0, 4.0, 5.0])
y = np.array([0.1, 1.2, 3.0, 4.2, 3.8])
# Extra data points for drawing the curves
x1 = np.linspace(-0.9, 6.7, 50)
x2 = np.linspace(x.min(), x.max(), 50) | _____no_output_____ | MIT | plotting/ex_interpolation.ipynb | nathanielng/python-snippets |
2.2 Calculate interpolating functions | lg = lagrange(x, y)
linear = interp1d(x, y, kind='linear')
spline0 = interp1d(x, y, kind='zero')
spline1 = interp1d(x, y, kind='slinear')
spline2 = interp1d(x, y, kind='quadratic')
spline3 = interp1d(x, y, kind='cubic') | _____no_output_____ | MIT | plotting/ex_interpolation.ipynb | nathanielng/python-snippets |
3. Plots - Lagrange vs Splines | fig, (ax0, ax1) = plt.subplots(2, 1, figsize=(7,13))
ax0.plot(x, y, 'bo')
ax0.plot(x1, lg(x1), label='Lagrange')
ax0.plot(x2, linear(x2), label='linear')
ax0.legend(loc='best', frameon=False)
ax1.plot(x, y, 'bo')
ax1.plot(x2, lg(x2), label='Lagrange', color='black')
ax1.plot(x2, spline0(x2), label='spline (0th order)'... | _____no_output_____ | MIT | plotting/ex_interpolation.ipynb | nathanielng/python-snippets |
4. Comparison - Lagrange, LOWESS, Kernel, Cubic Splines 4.1 Function to fit the data | def get_interpolated_data(x_train, y_train, x_new, kind, frac=0.1):
if kind == 'lagrange':
fn = lagrange(x_train, y_train)
x_pred = x_new
y_pred = fn(x_new)
elif kind == 'lowess':
xy = sm.nonparametric.lowess(y_train, x_train, frac=frac)
x_pred = xy[:, 0]
y_pred =... | _____no_output_____ | MIT | plotting/ex_interpolation.ipynb | nathanielng/python-snippets |
4.2 Data | n_pts = 10
n_all = 50
x = np.linspace(0, 2*np.pi, n_pts)
y = np.sin(x) + 0.1*(np.random.uniform(0, 1, n_pts) - 0.5)
x_actual = np.linspace(0, 2*np.pi, n_all)
y_actual = np.sin(x_actual)
x2 = np.linspace(x.min(), x.max(), 50) | _____no_output_____ | MIT | plotting/ex_interpolation.ipynb | nathanielng/python-snippets |
4.3 Plots | fig, axs = plt.subplots(2, 2, figsize=(12,8))
kinds = ['lagrange', 'lowess', 'kernel', 'cubic']
cmap = plt.get_cmap("tab10")
i = 0
for row in range(2):
for col in range(2):
kind = kinds[i]
x_p, y_p = get_interpolated_data(x, y, x2, kind)
axs[row][col].plot(x, y, 'bo')
axs[row][col].... | _____no_output_____ | MIT | plotting/ex_interpolation.ipynb | nathanielng/python-snippets |
모두를 위한 딥러닝 week 2 Tensorflow for logistic classifiersJimin Sun | import tensorflow as tf | /usr/local/anaconda3/lib/python3.6/importlib/_bootstrap.py:219: RuntimeWarning: compiletime version 3.5 of module 'tensorflow.python.framework.fast_tensor_util' does not match runtime version 3.6
return f(*args, **kwds)
| MIT | week-2/index.ipynb | PluVian/deep-learning-study-2017-winter |
Example from lab video | x_data = [[1,2], [2,3], [3,1], [4,4], [5,3], [6,2]]
y_data = [[0],[0],[0],[1],[1],[1]]
X = tf.placeholder(tf.float32, shape = [None, 2])
Y = tf.placeholder(tf.float32, shape = [None, 1]) # Shape에 주의!
W = tf.Variable(tf.random_normal([2,1]), name = 'weight')
# 들어오는 값 2개, 나가는 값 1개.
b = tf.Variable(tf.random_normal([1]), ... | _____no_output_____ | MIT | week-2/index.ipynb | PluVian/deep-learning-study-2017-winter |
What is 'reduce_mean'?* tf.reduce_mean 은 평균을 구해주는 operation을 한다. (np.mean과 같은 기능!)* 둘의 차이점은, numpy operation은 파이썬 어디서든 사용할 수 있지만, tensorflow operation은 tensorflow **Session** 내에서만 동작한다는 데에 있다.* But why **reduce**_mean?> The key here is the word reduce, a concept from functional programming, which makes it possible for ... | with tf.Session() as sess:
sess.run(tf.global_variables_initializer())
for step in range(10001):
cost_val, _ = sess.run([cost, train], feed_dict = {X: x_data, Y: y_data})
if step % 1000 == 0:
print(step, cost_val)
# Accuracy
h, c, a = sess.run([hypothesis, predicted... | 0 0.585211
1000 0.387248
2000 0.313725
3000 0.261586
4000 0.223239
5000 0.194146
6000 0.17146
7000 0.153347
8000 0.13859
9000 0.126357
10000 0.116064
Hypothesis:
[[ 0.0268787 ]
[ 0.16871433]
[ 0.21446182]
[ 0.86153394]
[ 0.93646842]
[ 0.97216052]]
Correct (Y):
[[ 0.]
[ 0.]
[ 0.]
[ 1.]
[ 1.]
[ 1.]]
Accu... | MIT | week-2/index.ipynb | PluVian/deep-learning-study-2017-winter |
Let's apply this classifier to another example!The same dataset from JaeYoung's example last week :)Where to get it : https://www.kaggle.com/c/uci-wine-quality-dataset/data | import numpy as np
import pandas as pd
data = pd.read_csv('winequality-data.csv', dtype = 'float32', header=0)
data.head() | _____no_output_____ | MIT | week-2/index.ipynb | PluVian/deep-learning-study-2017-winter |
Binary Classifier Here, you can spot the **'quality'** column, where the quality of wine is classified to 7 categories. | data['quality'].unique() | _____no_output_____ | MIT | week-2/index.ipynb | PluVian/deep-learning-study-2017-winter |
We'll start from a binary classifier, so we label * wines of quality 3.0, 4.0, 5.0 as class 0,* and wines of quality 6.0, 7.0, 8.0, 9.0 as class 1. | data.info()
# You can easily check if there are any missing values in your data.
grade = data['quality'] > 5.0
data['grade'] = grade.astype(np.float32)
data.head()
# new column 'grade' is added at the end
y_data = data.values[:,[-1]]
x_data = data.values[:,:-3] # columns quality, id, grade are excluded
y_data.shape, x... | 0 nan
100 nan
200 nan
300 nan
400 nan
500 nan
600 nan
700 nan
800 nan
900 nan
1000 nan
Hypothesis:
[[ nan]
[ nan]
[ nan]
...,
[ nan]
[ nan]
[ nan]]
Correct (Y)
: [[ 0.]
[ 0.]
[ 0.]
...,
[ 0.]
[ 0.]
[ 0.]]
Accuracy
: 0.335375
| MIT | week-2/index.ipynb | PluVian/deep-learning-study-2017-winter |
??????? Data normalization was needed! There are many ways to normalize (= change scales to $[0,1]$) data, but this time we'll use **min_max_normalization**. The idea here is to apply this formula below.$$min\_max(x_{ij}) = \dfrac{x_{ij} - min_{1 \leq i \leq n}(x_{ij})}{(max_{1 \leq i \leq n}(x_{ij})-min_{1 \leq i \leq... | def min_max_normalized(data):
col_max = np.max(data, axis=0) # axis=0 : 열, axis=1 : 행
col_min = np.min(data, axis=0)
return np.divide(data - col_min, col_max - col_min)
x_data = min_max_normalized(x_data)
with tf.Session() as sess:
sess.run(tf.global_variables_initializer())
feed = {X: x_data, ... | 0 1.53646
2000 0.611673
4000 0.596641
6000 0.585185
8000 0.576219
10000 0.569024
12000 0.563115
14000 0.558161
16000 0.553931
18000 0.550264
20000 0.547041
Hypothesis:
[[ 0.83623403]
[ 0.87260079]
[ 0.53597355]
...,
[ 0.79522443]
[ 0.75276494]
[ 0.38663006]]
Correct (Y)
: [[ 1.]
[ 1.]
[ 1.]
...,
[ 1.]
... | MIT | week-2/index.ipynb | PluVian/deep-learning-study-2017-winter |
Even after 20,000 steps, the accuracy doesn't look high enough. $(\approx 70\%)$ Why?Let's plot the data, and see if we can find a reason there. | %matplotlib inline
import matplotlib.pyplot as plt
import seaborn as sns
data.head()
data.describe() | _____no_output_____ | MIT | week-2/index.ipynb | PluVian/deep-learning-study-2017-winter |
*The variance of most columns seem extremely low.* | sns.FacetGrid(data, hue = 'grade', size=6).map(plt.scatter, 'free.sulfur.dioxide', 'total.sulfur.dioxide').add_legend()
plt.show()
sns.FacetGrid(data, hue = 'grade', size=6).map(plt.scatter, 'fixed.acidity', 'residual.sugar').add_legend()
plt.show()
sns.FacetGrid(data, hue = 'grade', size=6).map(plt.scatter, 'density',... | _____no_output_____ | MIT | week-2/index.ipynb | PluVian/deep-learning-study-2017-winter |
The data itself doesn't really seem linearly separable :(This kind of problem will lead us to our next session in the lecture, such as Neural Networks! Multiclass classifier Example from lab videoA quick review of the example dealt in the lecture. | xy = np.loadtxt('data-04-zoo.csv', delimiter=',', dtype=np.float32)
xy
x_data = xy[:, 0:-1]
y_data = xy[:, [-1]]
# nb_classes = 7
# For cases when you don't want to set a specific number to 'nb_classes',
# or if you don't exactly know the number of categories ,
# this might be a more generalized method.
np.unique(y_... | _____no_output_____ | MIT | week-2/index.ipynb | PluVian/deep-learning-study-2017-winter |
Back to our wine quality dataset! | x_data = data.values[:,:-3]
x_data = min_max_normalized(x_data)
x_data
y_data = (data.values[:,[-3]]).astype(np.int32) # quality column으로 다시 설정
x_data.shape, y_data.shape
num_class = len(data['quality'].unique())
num_class
X = tf.placeholder(tf.float32, [None, 11])
Y = tf.placeholder(tf.int32, [None, 1])
Y_one_hot = tf... | _____no_output_____ | MIT | week-2/index.ipynb | PluVian/deep-learning-study-2017-winter |
It seems that the classifier has assigned most data points to the class 5 and 6, since these two took up 75% of all classes. :( | alist = ['a1', 'a2', 'a3']
blist = ['b1', 'b2', 'b3']
print(set(zip(alist, blist))) | {('a2', 'b2'), ('a3', 'b3'), ('a1', 'b1')}
| MIT | week-2/index.ipynb | PluVian/deep-learning-study-2017-winter |
Magic MethodsBelow you'll find the same code from the previous exercise except two more methods have been added: an __add__ method and a __repr__ method. Your task is to fill out the code and get all of the unit tests to pass. You'll find the code cell with the unit tests at the bottom of this Jupyter notebook.As in p... | import math
import matplotlib.pyplot as plt
class Gaussian():
""" Gaussian distribution class for calculating and
visualizing a Gaussian distribution.
Attributes:
mean (float) representing the mean value of the distribution
stdev (float) representing the standard deviation of the dist... | _____no_output_____ | FTL | Software_Engineering_Practices/magic_methods.ipynb | tthoraldson/MachineLearningNanodegree |
FastTreeSHAP in Census Income Data This notebook contains usages and detailed comparisons of FastTreeSHAP v1, FastTreeSHAP v2 and the original TreeSHAP in **binary classification** problems using scikit-learn, XGBoost and LightGBM. It also contains the discussions of automatic algorithm selection. It may take a few mi... | import numpy as np
import pandas as pd
from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import roc_auc_score, accuracy_score
import xgboost as xgb
import lightgbm as lgb
import fasttreeshap
import os
import time | _____no_output_____ | BSD-2-Clause | notebooks/FastTreeSHAP_Census_Income.ipynb | linkedin/FastTreeSHAP |
Pre-process training and testing data | # source of data: https://archive.ics.uci.edu/ml/datasets/census+income
train = pd.read_csv("../data/adult_data.txt", sep = ",\s+", header = None, engine = "python")
test = pd.read_csv("../data/adult_test.txt", sep = ",\s+", header = None, skiprows = 1, engine = "python")
label_train = train[14].map({"<=50K": 0, ">50K"... | _____no_output_____ | BSD-2-Clause | notebooks/FastTreeSHAP_Census_Income.ipynb | linkedin/FastTreeSHAP |
Train a random forest model using scikit-learn and compute SHAP values | n_estimators = 200 # number of trees in random forest model
max_depth = 8 # maximum depth of any trees in random forest model
# train a random forest model
rf_model = RandomForestClassifier(n_estimators = n_estimators, max_depth = max_depth, random_state = 0)
rf_model.fit(train, label_train)
print("AUC on testing set... | _____no_output_____ | BSD-2-Clause | notebooks/FastTreeSHAP_Census_Income.ipynb | linkedin/FastTreeSHAP |
Compute SHAP values via different versions of TreeSHAP | num_sample = 10000 # number of samples to be explained
n_jobs = -1 # number of parallel threads (-1 means utilizing all available cores)
# compute SHAP values via FastTreeSHAP v0 (i.e., original TreeSHAP)
# parallel computing is not enabled in original TreeSHAP in SHAP package, but here we enable it for a fair compar... | _____no_output_____ | BSD-2-Clause | notebooks/FastTreeSHAP_Census_Income.ipynb | linkedin/FastTreeSHAP |
Compare running times of different versions of TreeSHAP in computing SHAP values | # compute SHAP values/SHAP interaction values via TreeSHAP algorithm with version "algorithm_version"
# (parallel on "n_jobs" threads)
def run_fasttreeshap(model, sample, interactions, algorithm_version, n_jobs, num_round, num_sample, shortcut = False):
shap_explainer = fasttreeshap.TreeExplainer(
model, al... | _____no_output_____ | BSD-2-Clause | notebooks/FastTreeSHAP_Census_Income.ipynb | linkedin/FastTreeSHAP |
Compute SHAP interaction values via different versions of TreeSHAP | num_sample = 100 # number of samples to be explained
n_jobs = -1 # number of parallel threads (-1 means utilizing all available cores)
# compute SHAP interaction values via FastTreeSHAP v0 (i.e., original TreeSHAP)
# parallel computing is not enabled in original TreeSHAP in SHAP package, but here we enable it for a f... | _____no_output_____ | BSD-2-Clause | notebooks/FastTreeSHAP_Census_Income.ipynb | linkedin/FastTreeSHAP |
Compare running times of different versions of TreeSHAP in computing SHAP interaction values | num_sample = 100 # number of samples to be explained
num_round = 3 # number of rounds to record mean and standard deviation of running time
n_jobs = -1 # number of parallel threads (-1 means utilizing all available cores)
# run FastTreeSHAP v0 (i.e., original TreeSHAP) multiple times and record its average running t... | _____no_output_____ | BSD-2-Clause | notebooks/FastTreeSHAP_Census_Income.ipynb | linkedin/FastTreeSHAP |
Train an XGBoost model and compute SHAP values | n_estimators = 200 # number of trees in XGBoost model
max_depth = 8 # maximum depth of any trees in XGBoost model
# train an XGBoost model
xgb_model = xgb.XGBClassifier(
max_depth = max_depth, n_estimators = n_estimators, learning_rate = 0.1, n_jobs = -1,
use_label_encoder = False, eval_metric = "logloss", ra... | _____no_output_____ | BSD-2-Clause | notebooks/FastTreeSHAP_Census_Income.ipynb | linkedin/FastTreeSHAP |
Compute SHAP values via different versions of TreeSHAP | num_sample = 10000 # number of samples to be explained
n_jobs = -1 # number of parallel threads (-1 means utilizing all available cores)
# compute SHAP values via "shortcut" (i.e., original TreeSHAP in XGBoost package)
# by default, parallel computing on all available cores is enabled in "shortcut"
shap_explainer = f... | _____no_output_____ | BSD-2-Clause | notebooks/FastTreeSHAP_Census_Income.ipynb | linkedin/FastTreeSHAP |
Compare running times of different versions of TreeSHAP in computing SHAP values | num_sample = 10000 # number of samples to be explained
num_round = 3 # number of rounds to record mean and standard deviation of running time
n_jobs = -1 # number of parallel threads (-1 means utilizing all available cores)
# run "shortcut" version of TreeSHAP multiple times and record its average running time
# by ... | _____no_output_____ | BSD-2-Clause | notebooks/FastTreeSHAP_Census_Income.ipynb | linkedin/FastTreeSHAP |
Compute SHAP interaction values via different versions of TreeSHAP | num_sample = 100 # number of samples to be explained
n_jobs = -1 # number of parallel threads (-1 means utilizing all available cores)
# compute SHAP interaction values via "shortcut" (i.e., original TreeSHAP in XGBoost package)
# by default, parallel computing on all available cores is enabled in "shortcut"
shap_exp... | _____no_output_____ | BSD-2-Clause | notebooks/FastTreeSHAP_Census_Income.ipynb | linkedin/FastTreeSHAP |
Compare running times of different versions of TreeSHAP in computing SHAP interaction values | num_sample = 100 # number of samples to be explained
num_round = 3 # number of rounds to record mean and standard deviation of running time
n_jobs = -1 # number of parallel threads (-1 means utilizing all available cores)
# run "shortcut" version of TreeSHAP multiple times and record its average running time
# by de... | _____no_output_____ | BSD-2-Clause | notebooks/FastTreeSHAP_Census_Income.ipynb | linkedin/FastTreeSHAP |
Train a LightGBM model and compute SHAP values | n_estimators = 500 # number of trees in LightGBM model
max_depth = 8 # maximum depth of any trees in LightGBM model
# train a LightGBM model
lgb_model = lgb.LGBMClassifier(
max_depth = max_depth, n_estimators = n_estimators, learning_rate = 0.1, n_jobs = -1, random_state = 0)
lgb_model.fit(train, label_train)
pri... | _____no_output_____ | BSD-2-Clause | notebooks/FastTreeSHAP_Census_Income.ipynb | linkedin/FastTreeSHAP |
Compute SHAP values via different versions of TreeSHAP | num_sample = 10000 # number of samples to be explained
n_jobs = -1 # number of parallel threads (-1 means utilizing all available cores)
# compute SHAP values via "shortcut" (i.e., original TreeSHAP in LightGBM package)
# by default, parallel computing on all available cores is enabled in "shortcut"
shap_explainer = ... | _____no_output_____ | BSD-2-Clause | notebooks/FastTreeSHAP_Census_Income.ipynb | linkedin/FastTreeSHAP |
Compare running times of different versions of TreeSHAP in computing SHAP values | num_sample = 10000 # number of samples to be explained
num_round = 3 # number of rounds to record mean and standard deviation of running time
n_jobs = -1 # number of parallel threads (-1 means utilizing all available cores)
# run "shortcut" version of TreeSHAP multiple times and record its average running time
# by ... | _____no_output_____ | BSD-2-Clause | notebooks/FastTreeSHAP_Census_Income.ipynb | linkedin/FastTreeSHAP |
Compute SHAP interaction values via different versions of TreeSHAP | num_sample = 100 # number of samples to be explained
n_jobs = -1 # number of parallel threads (-1 means utilizing all available cores)
# compute SHAP interaction values via FastTreeSHAP v0 (i.e., original TreeSHAP in SHAP package)
# parallel computing is not enabled in original TreeSHAP in SHAP package, but here we e... | _____no_output_____ | BSD-2-Clause | notebooks/FastTreeSHAP_Census_Income.ipynb | linkedin/FastTreeSHAP |
Compare running times of different versions of TreeSHAP in computing SHAP interaction values | num_sample = 100 # number of samples to be explained
num_round = 3 # number of rounds to record mean and standard deviation of running time
n_jobs = -1 # number of parallel threads (-1 means utilizing all available cores)
# run FastTreeSHAP v0 (i.e., original TreeSHAP) multiple times and record its average running t... | _____no_output_____ | BSD-2-Clause | notebooks/FastTreeSHAP_Census_Income.ipynb | linkedin/FastTreeSHAP |
Deep dive into automatic algorithm selection The default value of the argument `algorithm` in the class `TreeExplainer` is `auto`, indicating that the TreeSHAP algorithm is automatically selected from `"v0"`, `"v1"` and `"v2"` according to the number of samples to be explained and the constraint on the allocated memor... | n_estimators = 200 # number of trees in random forest model
max_depth = 8 # maximum depth of any trees in random forest model
# train a random forest model
rf_model = RandomForestClassifier(n_estimators = n_estimators, max_depth = max_depth, random_state = 0)
rf_model.fit(train, label_train)
# estimated memory usage ... | _____no_output_____ | BSD-2-Clause | notebooks/FastTreeSHAP_Census_Income.ipynb | linkedin/FastTreeSHAP |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.