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 |
|---|---|---|---|---|---|
HugeCTR Continuous Training and Inference Demo (Part I) OverviewIn HugeCTR version 3.3, we finished the whole pipeline of parameter server, including 1. The parameter dumping interface from training to kafka.2. CPU cache(Redis Cluster / Hash Map / Parallel Hash Map).3. RocksDB as a persistence storage.4. Embedding ca... | !mkdir criteo_data
!mkdir criteo_script | _____no_output_____ | BSD-3-Clause | samples/hierarchical_deployment/hps_e2e_demo/Continuous_Training.ipynb | miguelusque/hugectr_backend |
1.2 Download Criteo Dataset | !wget http://azuremlsampleexperiments.blob.core.windows.net/criteo/day_1.gz | _____no_output_____ | BSD-3-Clause | samples/hierarchical_deployment/hps_e2e_demo/Continuous_Training.ipynb | miguelusque/hugectr_backend |
**NOTE**: Replace `1` with a value from [0, 23] to use a different day.During preprocessing, the amount of data, which is used to speed up the preprocessing, fill missing values, and remove the feature values that are considered rare, is further reduced. 1.3 Write the preprocessing the script. | %%writefile preprocess.sh
#!/bin/bash
if [[ $# -lt 3 ]]; then
echo "Usage: preprocess.sh [DATASET_NO.] [DST_DATA_DIR] [SCRIPT_TYPE] [SCRIPT_TYPE_SPECIFIC_ARGS...]"
exit 2
fi
DST_DATA_DIR=$2
echo "Warning: existing $DST_DATA_DIR is erased"
rm -rf $DST_DATA_DIR
if [[ $3 == "nvt" ]]; then
if [[ $# -ne 6 ]]; the... | Overwriting preprocess.sh
| BSD-3-Clause | samples/hierarchical_deployment/hps_e2e_demo/Continuous_Training.ipynb | miguelusque/hugectr_backend |
**NOTE**: Here we only read the first 500000 lines of the data to do the demo. | %%writefile criteo_script/preprocess.py
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
from __future__ import unicode_literals
import argparse
import sys
import tempfile
from six.moves import urllib
import urllib.request
import sys
import os
import math... | Overwriting criteo_script/preprocess.py
| BSD-3-Clause | samples/hierarchical_deployment/hps_e2e_demo/Continuous_Training.ipynb | miguelusque/hugectr_backend |
1.4 Run the preprocess script | !bash preprocess.sh 0 criteo_data pandas 1 1 1 | _____no_output_____ | BSD-3-Clause | samples/hierarchical_deployment/hps_e2e_demo/Continuous_Training.ipynb | miguelusque/hugectr_backend |
**IMPORTANT NOTES**: Arguments may vary depend on your setting:- The first argument represents the dataset postfix. For instance, if `day_1` is used, the postfix is `1`.- The second argument, `criteo_data`, is where the preprocessed data is stored. 1.5 Generate data sample for inference | import pandas as pd
import numpy as np
df = pd.read_table("criteo_data/train/train.txt", header = None, sep= ' ', \
names = ['label'] + ['I'+str(i) for i in range(1, 14)] + \
['C1_C2', 'C3_C4'] + ['C'+str(i) for i in range(1, 27)])[:5]
left = df.iloc[:,:14].astype(np.float32)
right... | _____no_output_____ | BSD-3-Clause | samples/hierarchical_deployment/hps_e2e_demo/Continuous_Training.ipynb | miguelusque/hugectr_backend |
2. Start the Kafka Broker **Please refer to the README to start the Kafka Broker properly.** 3. Wide&Deep Model Demo | !rm -r *model
%%writefile wdl_demo.py
import hugectr
from mpi4py import MPI
solver = hugectr.CreateSolver(model_name = "wdl",
max_eval_batches = 5000,
batchsize_eval = 1024,
batchsize = 1024,
lr = 0.0... | [HUGECTR][03:34:23][INFO][RANK0]: Empty embedding, trained table will be stored in ./wdl_0_sparse_model
[HUGECTR][03:34:23][INFO][RANK0]: Empty embedding, trained table will be stored in ./wdl_1_sparse_model
HugeCTR Version: 3.2
====================================================Model Init=============================... | BSD-3-Clause | samples/hierarchical_deployment/hps_e2e_demo/Continuous_Training.ipynb | miguelusque/hugectr_backend |
4. WDL Inference 4.1 Inference using HugeCTR python API | #Create a folder for RocksDB
!mkdir /wdl_infer
!mkdir /wdl_infer/rocksdb | mkdir: cannot create directory β/wdl_inferβ: File exists
mkdir: cannot create directory β/wdl_infer/rocksdbβ: File exists
| BSD-3-Clause | samples/hierarchical_deployment/hps_e2e_demo/Continuous_Training.ipynb | miguelusque/hugectr_backend |
**Please make sure you have started Redis cluster following the README before you start doing inference.** | %%writefile 'wdl_predict.py'
from hugectr.inference import InferenceParams, CreateInferenceSession
import hugectr
import pandas as pd
import numpy as np
import sys
from mpi4py import MPI
def wdl_inference(model_name='wdl', network_file='wdl.json', dense_file='wdl_dense_0.model', \
embedding_file_list=... | [HUGECTR][03:36:30][INFO][RANK0]: default_emb_vec_value is not specified using default: 0.000000
[HUGECTR][03:36:30][INFO][RANK0]: default_emb_vec_value is not specified using default: 0.000000
[HUGECTR][03:36:30][INFO][RANK0]: Creating RedisCluster backend...
[HUGECTR][03:36:30][INFO][RANK0]: Connecting to Redis clust... | BSD-3-Clause | samples/hierarchical_deployment/hps_e2e_demo/Continuous_Training.ipynb | miguelusque/hugectr_backend |
4.2 Inference using Triton **Please refer to the [Triton_Inference.ipynb](./Triton_Inference.ipynb) notebook to start Triton and do the inference.** 5. Continue Training WDL Model | %%writefile wdl_continue.py
import hugectr
from mpi4py import MPI
solver = hugectr.CreateSolver(model_name = "wdl",
max_eval_batches = 5000,
batchsize_eval = 1024,
batchsize = 1024,
lr = 0.001,
... | [HUGECTR][03:37:25][INFO][RANK0]: Use existing embedding: ./wdl_0_sparse_model
[HUGECTR][03:37:25][INFO][RANK0]: Use existing embedding: ./wdl_1_sparse_model
HugeCTR Version: 3.2
====================================================Model Init=====================================================
[HUGECTR][03:37:25][INFO]... | BSD-3-Clause | samples/hierarchical_deployment/hps_e2e_demo/Continuous_Training.ipynb | miguelusque/hugectr_backend |
6. Inference with new model 6.1 Continuous inference using Python API | !python wdl_predict.py | [HUGECTR][03:38:09][INFO][RANK0]: default_emb_vec_value is not specified using default: 0.000000
[HUGECTR][03:38:09][INFO][RANK0]: default_emb_vec_value is not specified using default: 0.000000
[HUGECTR][03:38:09][INFO][RANK0]: Creating RedisCluster backend...
[HUGECTR][03:38:09][INFO][RANK0]: Connecting to Redis clust... | BSD-3-Clause | samples/hierarchical_deployment/hps_e2e_demo/Continuous_Training.ipynb | miguelusque/hugectr_backend |
Observations and Insights | # Dependencies and Setup
import matplotlib.pyplot as plt
import pandas as pd
import scipy.stats as st
# Study data files
mouse_metadata_path = "data/Mouse_metadata.csv"
study_results_path = "data/Study_results.csv"
# Read the mouse data and the study results
mouse_metadata = pd.read_csv(mouse_metadata_path)
study_res... | _____no_output_____ | ADSL | Pymaceuticals/pymaceuticals_starter.ipynb | tomlip/Matplotlib-challenge |
Summary Statistics | # Generate a summary statistics table of mean, median, variance, standard deviation, and SEM of the tumor volume for each regimen
# This method is the most straightforward, creating multiple series and putting them all together at the end.
drug_df = clean_df.groupby("Drug Regimen")
# calculating the statistics
mean_... | _____no_output_____ | ADSL | Pymaceuticals/pymaceuticals_starter.ipynb | tomlip/Matplotlib-challenge |
Bar Plots | # Generate a bar plot showing the number of mice per time point for each treatment throughout the course of the study using pandas.
#creating count dataframe
bar_df = clean_df.groupby("Drug Regimen").count()
#creating bar chart dataframe
bar_df = bar_df.sort_values("Timepoint", ascending=False)
bars_df = bar_df["Time... | _____no_output_____ | ADSL | Pymaceuticals/pymaceuticals_starter.ipynb | tomlip/Matplotlib-challenge |
Pie Plots | # Generate a pie plot showing the distribution of female versus male mice using pandas
pie_chart = mouse_metadata["Sex"].value_counts()
pie_chart.plot(kind='pie', subplots=True, autopct="%0.1f%%")
# Generate a pie plot showing the distribution of female versus male mice using pyplot
f_vs_m = mouse_metadata["Sex"].value... | _____no_output_____ | ADSL | Pymaceuticals/pymaceuticals_starter.ipynb | tomlip/Matplotlib-challenge |
Quartiles, Outliers and Boxplots | # Calculate the final tumor volume of each mouse across four of the most promising treatment regimens. Calculate the IQR and quantitatively determine if there are any potential outliers.
# returning the max timepoints for Capomulin
temp_df = clean_df.loc[clean_df["Drug Regimen"] == "Capomulin", :]
max_capo = temp_df.... | _____no_output_____ | ADSL | Pymaceuticals/pymaceuticals_starter.ipynb | tomlip/Matplotlib-challenge |
Line and Scatter Plots | # Generate a line plot of time point versus tumor volume for a mouse treated with Capomulin
# Generate a scatter plot of mouse weight versus average tumor volume for the Capomulin regimen
| _____no_output_____ | ADSL | Pymaceuticals/pymaceuticals_starter.ipynb | tomlip/Matplotlib-challenge |
Correlation and Regression | # Calculate the correlation coefficient and linear regression model
# for mouse weight and average tumor volume for the Capomulin regimen
| _____no_output_____ | ADSL | Pymaceuticals/pymaceuticals_starter.ipynb | tomlip/Matplotlib-challenge |
Create Dataset | def create_dataset(df, name, bonds):
print(f"Creating Dataset and Saving to {drive_path}/data/{name}.pkl")
data = df.sample(frac=1)
data = data.reset_index(drop=True)
data['mol'] = data['smiles'].apply(lambda x: create_dgl_features(x, bonds))
data.to_pickle(f"{drive_path}/data/{name}.pkl")
retur... | time: 148 ms (started: 2021-11-30 11:59:08 +00:00)
| MIT | Prototypical Nets Tox21 ECFP.ipynb | danielvlla/Few-Shot-Learning-for-Low-Data-Drug-Discovery |
Create Episode | def create_episode(n_support_pos, n_support_neg, n_query, data, test=False, train_balanced=True):
"""
n_query = per class data points
Xy = dataframe dataset in format [['y', 'mol']]
"""
support = []
query = []
n_query_pos = n_query
n_query_neg = n_query
support_neg = data[data['y'] == 0].sample(... | time: 2.25 ms (started: 2021-11-27 16:25:20 +00:00)
| MIT | Prototypical Nets Tox21 ECFP.ipynb | danielvlla/Few-Shot-Learning-for-Low-Data-Drug-Discovery |
Graph Embedding | class GCN(nn.Module):
def __init__(self, in_channels, out_channels=128):
super(GCN, self).__init__()
self.conv1 = GraphConv(in_channels, 64)
self.conv2 = GraphConv(64, 128)
self.conv3 = GraphConv(128, 64)
self.sum_pool = SumPooling()
self.dense = nn.Linear(64, out_channels)
... | time: 9.12 ms (started: 2021-11-30 11:46:53 +00:00)
| MIT | Prototypical Nets Tox21 ECFP.ipynb | danielvlla/Few-Shot-Learning-for-Low-Data-Drug-Discovery |
Distance Function | def euclidean_dist(x, y):
# x: N x D
# y: M x D
n = x.size(0)
m = y.size(0)
d = x.size(1)
assert d == y.size(1)
x = x.unsqueeze(1).expand(n, m, d)
y = y.unsqueeze(0).expand(n, m, d)
return torch.pow(x - y, 2).sum(2) | time: 4.05 ms (started: 2021-11-30 11:44:08 +00:00)
| MIT | Prototypical Nets Tox21 ECFP.ipynb | danielvlla/Few-Shot-Learning-for-Low-Data-Drug-Discovery |
LSTM | def cos(x, y):
transpose_shape = tuple(list(range(len(y.shape)))[::-1])
x = x.float()
denom = (
torch.sqrt(torch.sum(torch.square(x)) *
torch.sum(torch.square(y))) + torch.finfo(torch.float32).eps)
return torch.matmul(x, torch.permute(y, transpose_shape)) / denom
class ResiLSTMEmb... | time: 75.5 ms (started: 2021-11-30 11:44:11 +00:00)
| MIT | Prototypical Nets Tox21 ECFP.ipynb | danielvlla/Few-Shot-Learning-for-Low-Data-Drug-Discovery |
Protonet https://colab.research.google.com/drive/1QDYIwg2-iiUpVU8YyAh0lOgFgFPhVgvxscrollTo=BnLOgECOKG_y | class ProtoNet(nn.Module):
def __init__(self, with_bonds=False):
"""
Prototypical Network
"""
super(ProtoNet, self).__init__()
def forward(self, X_support, X_query, n_support_pos, n_support_neg):
n_support = len(X_support)
# prototypes
z_dim = X_support.size(-1) # size of the embe... | time: 18.5 ms (started: 2021-11-30 11:44:14 +00:00)
| MIT | Prototypical Nets Tox21 ECFP.ipynb | danielvlla/Few-Shot-Learning-for-Low-Data-Drug-Discovery |
Training Loop | def train(train_tasks, train_dfs, balanced_queries, k_pos, k_neg, n_query, episodes, lr):
writer = SummaryWriter()
start_time = time.time()
node_feat_size = 177
embedding_size = 128
encoder = Net()
resi_lstm = ResiLSTMEmbedding(k_pos+k_neg)
proto_net = ProtoNet()
loss_fn = nn.NLLLoss()
if torch.cuda... | time: 171 ms (started: 2021-11-30 11:57:42 +00:00)
| MIT | Prototypical Nets Tox21 ECFP.ipynb | danielvlla/Few-Shot-Learning-for-Low-Data-Drug-Discovery |
Testing Loop | def test(encoder, lstm, proto_net, test_tasks, test_dfs, k_pos, k_neg, rounds):
encoder.eval()
lstm.eval()
proto_net.eval()
test_info = {}
with torch.no_grad():
for task in test_tasks:
Xy = test_dfs[task]
running_loss = []
running_acc = []
running_roc = [0]
running_prc = ... | time: 161 ms (started: 2021-11-30 11:57:33 +00:00)
| MIT | Prototypical Nets Tox21 ECFP.ipynb | danielvlla/Few-Shot-Learning-for-Low-Data-Drug-Discovery |
Initiate Training and Testing | from google.colab import drive
drive.mount('/content/drive')
# PATHS
drive_path = "/content/drive/MyDrive/Colab Notebooks/MSC_21"
method_dir = "ProtoNets"
log_path = f"{drive_path}/{method_dir}/logs/"
# PARAMETERS
dataset = 'tox21'
with_bonds = False
test_rounds = 20
n_query = 64 # per class
episodes = 10000
lr = 0.0... | _____no_output_____ | MIT | Prototypical Nets Tox21 ECFP.ipynb | danielvlla/Few-Shot-Learning-for-Low-Data-Drug-Discovery |
Normalize text | herod_fp = '/Users/kyle/cltk_data/greek/text/tlg/plaintext/TLG0016.txt'
with open(herod_fp) as fo:
herod_raw = fo.read()
print(herod_raw[2000:2500]) # What do we notice needs help?
from cltk.corpus.utils.formatter import tlg_plaintext_cleanup
herod_clean = tlg_plaintext_cleanup(herod_raw, rm_punctuation=True, rm_... | α½³ΟαΏ Ξ΄α½² Ξ»α½³Ξ³ΞΏΟ
ΟΞΉ γΡνΡῠμΡΟα½° ΟΞ±αΏ¦ΟΞ± αΌΞ»α½³ΞΎΞ±Ξ½Ξ΄ΟΞΏΞ½ Οα½ΈΞ½ Ξ ΟΞΉα½±ΞΌΞΏΟ
αΌΞΊΞ·ΞΊΞΏα½ΉΟΞ± ΟΞ±αΏ¦ΟΞ± αΌΞΈΞ΅Ξ»αΏΟΞ±α½· ΞΏαΌ± αΌΞΊ ΟαΏΟ αΌΞ»Ξ»α½±Ξ΄ΞΏΟ δι αΌΟΟΞ±Ξ³αΏΟ γΡνέΟΞΈΞ±ΞΉ Ξ³Ο
Ξ½Ξ±αΏΞΊΞ± αΌΟΞΉΟΟάμΡνον Οα½±Ξ½ΟΟΟ α½
ΟΞΉ ΞΏα½ Ξ΄α½½ΟΡι Ξ΄α½·ΞΊΞ±Ο ΞΏα½Ξ΄α½² Ξ³α½°Ο αΌΞΊΞ΅α½·Ξ½ΞΏΟ
Ο Ξ΄ΞΉΞ΄α½ΉΞ½Ξ±ΞΉ. Ξα½ΟΟ Ξ΄α½΄ αΌΟΟα½±ΟΞ±Ξ½ΟΞΏΟ Ξ±α½ΟΞΏαΏ¦ αΌΞ»α½³Ξ½Ξ·Ξ½ ΟΞΏαΏΟΞΉ αΌΞ»Ξ»Ξ·ΟΞΉ Ξ΄α½ΉΞΎΞ±ΞΉ ΟΟαΏΆΟΞΏΞ½ Οα½³ΞΌΟΞ±Ξ½ΟΞ±Ο αΌΞ³Ξ³α½³Ξ»ΞΏΟ
Ο αΌΟΞ±ΞΉΟέΡιν ΟΞ΅ αΌΞ»α½³Ξ½Ξ·Ξ½ ΞΊΞ±α½Ά Ξ΄α½·ΞΊΞ±Ο ΟαΏΟ αΌΟΟΞ±Ξ³αΏΟ Ξ±αΌ°ΟέΡιν. ΀ο... | MIT | public_talks/2015_11_17_nyu/3 Normalization, tokenization, tagging.ipynb | kylepjohnson/ipython_notebooks |
Tokenize sentences | from cltk.tokenize.sentence import TokenizeSentence
tokenizer = TokenizeSentence('greek')
herod_sents = tokenizer.tokenize_sentences(herod_clean)
print(herod_sents[:5])
for sent in herod_sents:
print(sent)
print()
input() | αΌ©ΟΞΏΞ΄α½ΉΟΞΏΟ
ΞΞΏΟ
Οα½·ΞΏΟ
αΌ±ΟΟΞΏΟα½·Ξ·Ο αΌΟα½ΉΞ΄Ξ΅ΞΎΞΉΟ αΌ₯δΡ α½‘Ο ΞΌα½΅ΟΞ΅ Οα½° γΡνόμΡνα αΌΞΎ αΌΞ½ΞΈΟα½½ΟΟΞ½ ΟαΏ· ΟΟα½ΉΞ½αΏ³ αΌΞΎα½·Οηλα Ξ³α½³Ξ½Ξ·ΟΞ±ΞΉ μὡΟΞ΅ αΌΟΞ³Ξ± μΡγάλα ΟΞ΅ ΞΊΞ±α½Ά ΞΈΟΞΌΞ±ΟΟα½± Οα½° ΞΌα½²Ξ½ αΌΞ»Ξ»Ξ·ΟΞΉ Οα½° Ξ΄α½² Ξ²Ξ±ΟΞ²α½±ΟΞΏΞΉΟΞΉ αΌΟοδΡΟΞΈα½³Ξ½ΟΞ± αΌΞΊΞ»α½³Ξ± Ξ³α½³Ξ½Ξ·ΟΞ±ΞΉ Οα½± ΟΞ΅ αΌΞ»Ξ»Ξ± ΞΊΞ±α½Ά δι αΌ£Ξ½ Ξ±αΌ°Οα½·Ξ·Ξ½ αΌΟΞΏΞ»α½³ΞΌΞ·ΟΞ±Ξ½ αΌΞ»Ξ»α½΅Ξ»ΞΏΞΉΟΞΉ.
Ξ Ξ΅ΟΟα½³ΟΞ½ ΞΌα½³Ξ½ Ξ½Ο
Ξ½ ΞΏαΌ± Ξ»α½ΉΞ³ΞΉΞΏΞΉ Ξ¦ΞΏα½·Ξ½ΞΉΞΊΞ±Ο Ξ±αΌ°Οα½·ΞΏΟ
Ο ΟΞ±Οα½Ά γΡνέΟΞΈΞ±ΞΉ ΟαΏΟ διαΟΞΏΟαΏΟ ΟΞΏα½»ΟΞΏΟ
Ο Ξ³α½±Ο ... | MIT | public_talks/2015_11_17_nyu/3 Normalization, tokenization, tagging.ipynb | kylepjohnson/ipython_notebooks |
Make word tokens | from cltk.tokenize.word import nltk_tokenize_words
for sent in herod_sents:
words = nltk_tokenize_words(sent)
print(words)
input() | ['αΌ©ΟΞΏΞ΄α½ΉΟΞΏΟ
', 'ΞΞΏΟ
Οα½·ΞΏΟ
', 'αΌ±ΟΟΞΏΟα½·Ξ·Ο', 'αΌΟόδΡξιΟ', 'αΌ₯δΡ', 'ὑΟ', 'μὡΟΞ΅', 'Οα½°', 'γΡνόμΡνα', 'αΌΞΎ', 'αΌΞ½ΞΈΟα½½ΟΟΞ½', 'ΟαΏ·', 'ΟΟα½ΉΞ½αΏ³', 'αΌΞΎα½·Οηλα', 'Ξ³α½³Ξ½Ξ·ΟΞ±ΞΉ', 'μὡΟΞ΅', 'αΌΟΞ³Ξ±', 'μΡγάλα', 'ΟΞ΅', 'ΞΊΞ±α½Ά', 'ΞΈΟΞΌΞ±ΟΟα½±', 'Οα½°', 'ΞΌα½²Ξ½', 'αΌΞ»Ξ»Ξ·ΟΞΉ', 'Οα½°', 'Ξ΄α½²', 'Ξ²Ξ±ΟΞ²α½±ΟΞΏΞΉΟΞΉ', 'αΌΟοδΡΟΞΈα½³Ξ½ΟΞ±', 'αΌΞΊΞ»α½³Ξ±', 'Ξ³α½³Ξ½Ξ·ΟΞ±ΞΉ', 'Οα½±', 'ΟΞ΅', 'αΌΞ»Ξ»Ξ±', 'ΞΊΞ±α½Ά', 'δι', 'αΌ£Ξ½', 'Ξ±αΌ°Οα½·... | MIT | public_talks/2015_11_17_nyu/3 Normalization, tokenization, tagging.ipynb | kylepjohnson/ipython_notebooks |
Tokenize Latin enclitics | from cltk.corpus.utils.formatter import phi5_plaintext_cleanup
from cltk.tokenize.word import WordTokenizer
# 'LAT0474': 'Marcus Tullius Cicero, Cicero, Tully',
cicero_fp = '/Users/kyle/cltk_data/latin/text/phi5/plaintext/LAT0474.TXT'
with open(cicero_fp) as fo:
cicero_raw = fo.read()
cicero_clean = phi5_plaintex... | ['Quae', 'res', 'in', 'civitate', 'duae', 'plurimum', 'possunt', 'eae', 'contra', 'nos', 'ambae', 'faciunt', 'in', 'hoc', 'tempore', 'summa', 'gratia', 'et', 'eloquentia', 'quarum', 'alteram', 'C.', 'Aquili', 'vereor', 'alteram', 'metuo.']
['Eloquentia', 'Q.', 'Hortensi', 'ne', 'me', 'in', 'dicendo', 'impediat', 'non'... | MIT | public_talks/2015_11_17_nyu/3 Normalization, tokenization, tagging.ipynb | kylepjohnson/ipython_notebooks |
POS Tagging | from cltk.tag.pos import POSTag
tagger = POSTag('greek')
# Heordotus again
for sent in herod_sents:
tagged_text = tagger.tag_unigram(sent)
print(tagged_text)
input() | [('αΌ©ΟΞΏΞ΄α½ΉΟΞΏΟ
', None), ('ΞΞΏΟ
Οα½·ΞΏΟ
', None), ('αΌ±ΟΟΞΏΟα½·Ξ·Ο', None), ('αΌΟόδΡξιΟ', None), ('αΌ₯δΡ', 'P-S---FN-'), ('ὑΟ', 'D--------'), ('μὡΟΞ΅', None), ('Οα½°', 'L-P---NA-'), ('γΡνόμΡνα', None), ('αΌΞΎ', 'R--------'), ('αΌΞ½ΞΈΟα½½ΟΟΞ½', None), ('ΟαΏ·', 'P-S---MD-'), ('ΟΟα½ΉΞ½αΏ³', None), ('αΌΞΎα½·Οηλα', None), ('Ξ³α½³Ξ½Ξ·ΟΞ±ΞΉ', None), ('μὡΟΞ΅', None), ('αΌΟΞ³Ξ±'... | MIT | public_talks/2015_11_17_nyu/3 Normalization, tokenization, tagging.ipynb | kylepjohnson/ipython_notebooks |
NER | ## Latin -- decent, but see M, P, etc
from cltk.tag import ner
# Heordotus again
for sent in cicero_sents:
ner_tags = ner.tag_ner('latin', input_text=sent, output_type=list)
print(ner_tags)
input()
# Greek -- not as good!
from cltk.tag import ner
# Heordotus again
for sent in herod_sents:
ner_tags = n... | [('αΌ©ΟΞΏΞ΄α½ΉΟΞΏΟ
',), ('ΞΞΏΟ
Οα½·ΞΏΟ
',), ('αΌ±ΟΟΞΏΟα½·Ξ·Ο',), ('αΌΟόδΡξιΟ',), ('αΌ₯δΡ',), ('ὑΟ',), ('μὡΟΞ΅',), ('Οα½°',), ('γΡνόμΡνα',), ('αΌΞΎ',), ('αΌΞ½ΞΈΟα½½ΟΟΞ½',), ('ΟαΏ·',), ('ΟΟα½ΉΞ½αΏ³',), ('αΌΞΎα½·Οηλα',), ('Ξ³α½³Ξ½Ξ·ΟΞ±ΞΉ',), ('μὡΟΞ΅',), ('αΌΟΞ³Ξ±',), ('μΡγάλα',), ('ΟΞ΅',), ('ΞΊΞ±α½Ά',), ('ΞΈΟΞΌΞ±ΟΟα½±',), ('Οα½°',), ('ΞΌα½²Ξ½',), ('αΌΞ»Ξ»Ξ·ΟΞΉ', 'Entity'), ('Οα½°',), ('Ξ΄α½²',), ('Ξ²Ξ±ΟΞ²... | MIT | public_talks/2015_11_17_nyu/3 Normalization, tokenization, tagging.ipynb | kylepjohnson/ipython_notebooks |
Stopword filtering | from cltk.stop.greek.stops import STOPS_LIST
#p = PunktLanguageVars()
for sent in herod_sents:
words = nltk_tokenize_words(sent)
print('W/ STOPS', words)
words = [w for w in words if not w in STOPS_LIST]
print('W/O STOPS', words)
input() | W/ STOPS ['αΌ©ΟΞΏΞ΄α½ΉΟΞΏΟ
', 'ΞΞΏΟ
Οα½·ΞΏΟ
', 'αΌ±ΟΟΞΏΟα½·Ξ·Ο', 'αΌΟόδΡξιΟ', 'αΌ₯δΡ', 'ὑΟ', 'μὡΟΞ΅', 'Οα½°', 'γΡνόμΡνα', 'αΌΞΎ', 'αΌΞ½ΞΈΟα½½ΟΟΞ½', 'ΟαΏ·', 'ΟΟα½ΉΞ½αΏ³', 'αΌΞΎα½·Οηλα', 'Ξ³α½³Ξ½Ξ·ΟΞ±ΞΉ', 'μὡΟΞ΅', 'αΌΟΞ³Ξ±', 'μΡγάλα', 'ΟΞ΅', 'ΞΊΞ±α½Ά', 'ΞΈΟΞΌΞ±ΟΟα½±', 'Οα½°', 'ΞΌα½²Ξ½', 'αΌΞ»Ξ»Ξ·ΟΞΉ', 'Οα½°', 'Ξ΄α½²', 'Ξ²Ξ±ΟΞ²α½±ΟΞΏΞΉΟΞΉ', 'αΌΟοδΡΟΞΈα½³Ξ½ΟΞ±', 'αΌΞΊΞ»α½³Ξ±', 'Ξ³α½³Ξ½Ξ·ΟΞ±ΞΉ', 'Οα½±', 'ΟΞ΅', 'αΌΞ»Ξ»Ξ±', 'ΞΊΞ±α½Ά', 'δι', 'αΌ£... | MIT | public_talks/2015_11_17_nyu/3 Normalization, tokenization, tagging.ipynb | kylepjohnson/ipython_notebooks |
Concordance | from cltk.utils.philology import Philology
p = Philology()
herod_fp = '/Users/kyle/cltk_data/greek/text/tlg/plaintext/TLG0016.txt'
p.write_concordance_from_file(herod_fp, 'kyle_herod') | INFO:CLTK:Wrote concordance to '/Users/kyle/cltk_data/user_data/concordance_kyle_herod.txt'.
| MIT | public_talks/2015_11_17_nyu/3 Normalization, tokenization, tagging.ipynb | kylepjohnson/ipython_notebooks |
Word count | from nltk.text import Text
words = nltk_tokenize_words(herod_clean)
print(words[:15])
t = Text(words)
vocabulary_count = t.vocab()
vocabulary_count['αΌ±ΟΟΞΏΟα½·Ξ·Ο']
vocabulary_count['μὡΟΞ΅']
vocabulary_count['αΌΞ½ΞΈΟα½½ΟΟΞ½'] | _____no_output_____ | MIT | public_talks/2015_11_17_nyu/3 Normalization, tokenization, tagging.ipynb | kylepjohnson/ipython_notebooks |
Word frequency | from cltk.utils.frequency import Frequency
freq = Frequency()
herod_frequencies = freq.counter_from_str(herod_clean)
herod_frequencies.most_common() | _____no_output_____ | MIT | public_talks/2015_11_17_nyu/3 Normalization, tokenization, tagging.ipynb | kylepjohnson/ipython_notebooks |
Fuel type | # remove the 'Type de carburant:' string from the carburant_type feature
df.fuel_type = df.fuel_type.map(lambda x: x.lstrip('Type de carburant:')) | _____no_output_____ | MIT | cars-price-dataset.ipynb | jaselnik/Car-Price-Predictor-Django |
Mark & Model | # remove the 'Marque:' string from the mark feature
df['mark'] = df['mark'].map(lambda x: x.replace('Marque:', ''))
df = df[df.mark != '-']
# remove the 'Modèle:' string from model feature
df['model'] = df['model'].map(lambda x: x.replace('Modèle:', '')) | _____no_output_____ | MIT | cars-price-dataset.ipynb | jaselnik/Car-Price-Predictor-Django |
fiscal powerFor the fiscal power we can see that there is exactly 5728 rows not announced, so we will fill them by the mean of the other columns, since it is an important feature in cars price prediction so we can not drop it. | # remove the 'Puissance fiscale:' from the fiscal_power feature
df.fiscal_power = df.fiscal_power.map(lambda x: x.lstrip('Puissance fiscale:Plus de').rstrip(' CV'))
# replace the - with NaN values and convert them to integer values
df.fiscal_power = df.fiscal_power.str.replace("-","0")
# convert all fiscal_power values... | _____no_output_____ | MIT | cars-price-dataset.ipynb | jaselnik/Car-Price-Predictor-Django |
fuel type | # remove those lines having the fuel_type not set
df = df[df.fuel_type != '-'] | _____no_output_____ | MIT | cars-price-dataset.ipynb | jaselnik/Car-Price-Predictor-Django |
drop unwanted columns | df = df.drop(columns=['sector', 'type'])
df = df[['price', 'year_model', 'mileage', 'fiscal_power', 'fuel_type', 'mark']]
df.to_csv('data/car_dataset.csv')
df.head()
from car_price.wsgi import application
from api.models import Car
for x in df.values[5598:]:
car = Car(
price=x[0],
year_model=x[1],
... | _____no_output_____ | MIT | cars-price-dataset.ipynb | jaselnik/Car-Price-Predictor-Django |
Credit Risk ClassificationCredit risk poses a classification problem thatβs inherently imbalanced. This is because healthy loans easily outnumber risky loans. In this Challenge, youβll use various techniques to train and evaluate models with imbalanced classes. Youβll use a dataset of historical lending activity from ... | # Import the modules
import numpy as np
import pandas as pd
from pathlib import Path
from sklearn.metrics import balanced_accuracy_score
from sklearn.metrics import confusion_matrix
from imblearn.metrics import classification_report_imbalanced
import warnings
warnings.filterwarnings('ignore') | _____no_output_____ | MIT | Starter_Code/credit_risk_resampling.ipynb | LeoHarada/Challenge_12 |
--- Split the Data into Training and Testing Sets Step 1: Read the `lending_data.csv` data from the `Resources` folder into a Pandas DataFrame. | # Read the CSV file from the Resources folder into a Pandas DataFrame
lending_data_df = pd.read_csv(Path("../Starter_Code/Resources/lending_data.csv"))
# Review the DataFrame
display(lending_data_df.head()) | _____no_output_____ | MIT | Starter_Code/credit_risk_resampling.ipynb | LeoHarada/Challenge_12 |
Step 2: Create the labels set (`y`) from the βloan_statusβ column, and then create the features (`X`) DataFrame from the remaining columns. | # Separate the data into labels and features
# Separate the y variable, the labels
y = lending_data_df["loan_status"]
# Separate the X variable, the features
X = lending_data_df.drop(columns=["loan_status"])
# Review the y variable Series
y.head()
# Review the X variable DataFrame
X.head() | _____no_output_____ | MIT | Starter_Code/credit_risk_resampling.ipynb | LeoHarada/Challenge_12 |
Step 3: Check the balance of the labels variable (`y`) by using the `value_counts` function. | # Check the balance of our target values
y.value_counts() | _____no_output_____ | MIT | Starter_Code/credit_risk_resampling.ipynb | LeoHarada/Challenge_12 |
Step 4: Split the data into training and testing datasets by using `train_test_split`. | # Import the train_test_learn module
from sklearn.model_selection import train_test_split
# Split the data using train_test_split
# Assign a random_state of 1 to the function
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=1) | _____no_output_____ | MIT | Starter_Code/credit_risk_resampling.ipynb | LeoHarada/Challenge_12 |
--- Create a Logistic Regression Model with the Original Data Step 1: Fit a logistic regression model by using the training data (`X_train` and `y_train`). | # Import the LogisticRegression module from SKLearn
from sklearn.linear_model import LogisticRegression
# Instantiate the Logistic Regression model
# Assign a random_state parameter of 1 to the model
model = LogisticRegression(random_state=1)
# Fit the model using training data
model.fit(X_train, y_train) | _____no_output_____ | MIT | Starter_Code/credit_risk_resampling.ipynb | LeoHarada/Challenge_12 |
Step 2: Save the predictions on the testing data labels by using the testing feature data (`X_test`) and the fitted model. | # Make a prediction using the testing data
y_pred = model.predict(X_test)
y_pred | _____no_output_____ | MIT | Starter_Code/credit_risk_resampling.ipynb | LeoHarada/Challenge_12 |
Step 3: Evaluate the modelβs performance by doing the following:* Calculate the accuracy score of the model.* Generate a confusion matrix.* Print the classification report. | # Print the balanced_accuracy score of the model
BAS = balanced_accuracy_score(y_test, y_pred)
print(BAS)
# Generate a confusion matrix for the model
print(confusion_matrix(y_test, y_pred))
# Print the classification report for the model
print(classification_report_imbalanced(y_test, y_pred)) | pre rec spe f1 geo iba sup
0 1.00 0.99 0.91 1.00 0.95 0.91 18765
1 0.85 0.91 0.99 0.88 0.95 0.90 619
avg / total 0.99 0.99 0.91 0.99 0.95 0... | MIT | Starter_Code/credit_risk_resampling.ipynb | LeoHarada/Challenge_12 |
Step 4: Answer the following question. **Question:** How well does the logistic regression model predict both the `0` (healthy loan) and `1` (high-risk loan) labels?**Answer:** The regression models predicts both healthy loans and high-risk loans, for the most part, accurately. We have an average 99% for our F1 score,... | # Import the RandomOverSampler module form imbalanced-learn
from imblearn.over_sampling import RandomOverSampler
# Instantiate the random oversampler model
# # Assign a random_state parameter of 1 to the model
random_oversampler = RandomOverSampler(random_state=1)
# Fit the original training data to the random_oversa... | _____no_output_____ | MIT | Starter_Code/credit_risk_resampling.ipynb | LeoHarada/Challenge_12 |
Step 2: Use the `LogisticRegression` classifier and the resampled data to fit the model and make predictions. | # Instantiate the Logistic Regression model
# Assign a random_state parameter of 1 to the model
resampled_model = LogisticRegression(random_state=1)
# Fit the model using the resampled training data
resampled_model.fit(X_resampled, y_resampled)
# Make a prediction using the testing data
y_pred = resampled_model.predi... | _____no_output_____ | MIT | Starter_Code/credit_risk_resampling.ipynb | LeoHarada/Challenge_12 |
Step 3: Evaluate the modelβs performance by doing the following:* Calculate the accuracy score of the model.* Generate a confusion matrix.* Print the classification report. | # Print the balanced_accuracy score of the model
print(balanced_accuracy_score(y_test, y_pred))
# Generate a confusion matrix for the model
confusion_matrix(y_test, y_pred)
# Print the classification report for the model
print(classification_report_imbalanced(y_test, y_pred)) | pre rec spe f1 geo iba sup
0 1.00 0.99 0.99 1.00 0.99 0.99 18765
1 0.84 0.99 0.99 0.91 0.99 0.99 619
avg / total 0.99 0.99 0.99 0.99 0.99 0... | MIT | Starter_Code/credit_risk_resampling.ipynb | LeoHarada/Challenge_12 |
Average | from sklearn.metrics import accuracy_score, roc_auc_score
from sklearn.preprocessing import StandardScaler
oof_preds = np.mean(df_norm[all_feat_names].to_numpy(),1)
oof_gts = df_norm['MGMT_value']
cv_preds = [np.mean(df_norm[df_norm.fold==fold][all_feat_names].to_numpy(),1) for fold in range(5)]
cv_gts = [df_norm[df_n... | OOF acc 0.5944540727902946, OOF auc 0.6514516827964754, CV AUC 0.6504285580435163 (std 0.02232524533384981)
| MIT | notebooks/7-Ensemble.ipynb | jpjuvo/RSNA-MICCAI-Brain-Tumor-Classification |
2nd level models | import xgboost as xgb
def get_data(fold, features):
df = df_norm.dropna(inplace=False)
scaler = StandardScaler()
df_train = df[df.fold != fold]
df_val = df[df.fold == fold]
if len(df_val) == 0:
df_val = df[df.fold == 0]
# shuffle train
df_train = df_train.sample(frac=1)
... | _____no_output_____ | MIT | notebooks/7-Ensemble.ipynb | jpjuvo/RSNA-MICCAI-Brain-Tumor-Classification |
Introduction to \LaTeX Math ModeJupyter notebooks integrate the MathJax Javascript library in order to render mathematical formulas and symbols in the same way as one would in \LaTeX (often used to typeset textbooks, research papers, or other technical documents).First, we will take a look at a couple of rendered expr... | | One | Two | Three | Four |
| --- | --- | --- | --- |
| 10% | Something | Else | 40% |
| 90% | To | Do | 50% | | _____no_output_____ | MIT | Introductions/LaTeX and Markdown Intro.ipynb | mtr3t/notebook-examples |
Interacting with models November 2014, by Max Zwiessele with edits by James HensmanThe GPy model class has a set of features which are designed to make it simple to explore the parameter space of the model. By default, the scipy optimisers are used to fit GPy models (via model.optimize()), for which we provide mechani... | m = GPy.examples.regression.sparse_GP_regression_1D(plot=False, optimize=False) | _____no_output_____ | MIT | tests/GPy/models_basic.ipynb | gopala-kr/ds-notebooks |
Examining the model using printTo see the current state of the model parameters, and the modelβs (marginal) likelihood just print the model print mThe first thing displayed on the screen is the log-likelihood value of the model with its current parameters. Below the log-likelihood, a table with all the modelβs para... | m | _____no_output_____ | MIT | tests/GPy/models_basic.ipynb | gopala-kr/ds-notebooks |
In this case the kernel parameters (`bf.variance`, `bf.lengthscale`) as well as the likelihood noise parameter (`Gaussian_noise.variance`), are constrained to be positive, while the inducing inputs have no constraints associated. Also there are no ties or prior defined.You can also print all subparts of the model, by p... | m.rbf | _____no_output_____ | MIT | tests/GPy/models_basic.ipynb | gopala-kr/ds-notebooks |
When you want to get a closer look into multivalue parameters, print them directly: | m.inducing_inputs
m.inducing_inputs[0] = 1 | _____no_output_____ | MIT | tests/GPy/models_basic.ipynb | gopala-kr/ds-notebooks |
Interacting with Parameters:The preferred way of interacting with parameters is to act on the parameter handle itself. Interacting with parameter handles is simple. The names, printed by print m are accessible interactively and programatically. For example try to set the kernel's `lengthscale` to 0.2 and print the res... | m.rbf.lengthscale = 0.2
print m |
Name : sparse_gp
Objective : 563.178096129
Number of Parameters : 8
Number of Optimization Parameters : 8
Updates : True
Parameters:
[1msparse_gp. [0;0m | value | constraints | priors
[1minducing_inputs [0;0m | (5, 1) | |
[1mrbf.variance [0;0m |... | MIT | tests/GPy/models_basic.ipynb | gopala-kr/ds-notebooks |
This will already have updated the modelβs inner state: note how the log-likelihood has changed. YOu can immediately plot the model or see the changes in the posterior (`m.posterior`) of the model. Regular expressionsThe modelβs parameters can also be accessed through regular expressions, by βindexingβ the model with ... | print m['.*var']
#print "variances as a np.array:", m['.*var'].values()
#print "np.array of rbf matches: ", m['.*rbf'].values() | [1mindex[0;0m | sparse_gp.rbf.variance | constraints | priors
[1m[0] [0;0m | 1.00000000 | +ve |
[1m-----[0;0m | sparse_gp.Gaussian_noise.variance | ----------- | ------
[1m[0] [0;0m | 1.00000000 | +ve ... | MIT | tests/GPy/models_basic.ipynb | gopala-kr/ds-notebooks |
There is access to setting parameters by regular expression, as well. Here are a few examples of how to set parameters by regular expression. Note that each time the values are set, computations are done internally to compute the log likeliood of the model. | m['.*var'] = 2.
print m
m['.*var'] = [2., 3.]
print m |
Name : sparse_gp
Objective : 680.058219518
Number of Parameters : 8
Number of Optimization Parameters : 8
Updates : True
Parameters:
[1msparse_gp. [0;0m | value | constraints | priors
[1minducing_inputs [0;0m | (5, 1) | |
[1mrbf.variance [0;0m |... | MIT | tests/GPy/models_basic.ipynb | gopala-kr/ds-notebooks |
A handy trick for seeing all of the parameters of the model at once is to regular-expression match every variable: | print m[''] | [1mindex[0;0m | sparse_gp.inducing_inputs | constraints | priors
[1m[0 0][0;0m | 1.00000000 | |
[1m[1 0][0;0m | -1.51676820 | |
[1m[2 0][0;0m | -2.23387110 | ... | MIT | tests/GPy/models_basic.ipynb | gopala-kr/ds-notebooks |
Setting and fetching parameters parameter_arrayAnother way to interact with the modelβs parameters is through the parameter_array. The Parameter array holds all the parameters of the model in one place and is editable. It can be accessed through indexing the model for example you can set all the parameters through thi... | new_params = np.r_[[-4,-2,0,2,4], [.1,2], [.7]]
print new_params
m[:] = new_params
print m | [-4. -2. 0. 2. 4. 0.1 2. 0.7]
Name : sparse_gp
Objective : 322.428807303
Number of Parameters : 8
Number of Optimization Parameters : 8
Updates : True
Parameters:
[1msparse_gp. [0;0m | value | constraints | priors
[1minducing_inputs [0;0m | (5, 1) | | ... | MIT | tests/GPy/models_basic.ipynb | gopala-kr/ds-notebooks |
Parameters themselves (leafs of the hierarchy) can be indexed and used the same way as numpy arrays. First let us set a slice of the inducing_inputs: | m.inducing_inputs[2:, 0] = [1,3,5]
print m.inducing_inputs | [1mindex[0;0m | sparse_gp.inducing_inputs | constraints | priors
[1m[0 0][0;0m | -4.00000000 | |
[1m[1 0][0;0m | -2.00000000 | |
[1m[2 0][0;0m | 1.00000000 | |
[1m[3 0][0;0m |... | MIT | tests/GPy/models_basic.ipynb | gopala-kr/ds-notebooks |
Or you use the parameters as normal numpy arrays for calculations: | precision = 1./m.Gaussian_noise.variance
print precision | [ 1.42857143]
| MIT | tests/GPy/models_basic.ipynb | gopala-kr/ds-notebooks |
Getting the model parameterβs gradientsThe gradients of a model can shed light on understanding the (possibly hard) optimization process. The gradients of each parameter handle can be accessed through their gradient field.: | print "all gradients of the model:\n", m.gradient
print "\n gradients of the rbf kernel:\n", m.rbf.gradient | all gradients of the model:
[ 2.1054468 3.67055686 1.28382016 -0.36934978 -0.34404866
99.49876932 -12.83697274 -268.02492615]
gradients of the rbf kernel:
[ 99.49876932 -12.83697274]
| MIT | tests/GPy/models_basic.ipynb | gopala-kr/ds-notebooks |
If we optimize the model, the gradients (should be close to) zero | m.optimize()
print m.gradient | [ -4.62140715e-04 -2.13365576e-04 9.60255226e-05 4.82744982e-04
8.56445996e-05 -5.25465293e-06 -6.89058756e-06 -9.34850797e-02]
| MIT | tests/GPy/models_basic.ipynb | gopala-kr/ds-notebooks |
Adjusting the modelβs constraintsWhen we initially call the example, it was optimized and hence the log-likelihood gradients were close to zero. However, since we have been changing the parameters, the gradients are far from zero now. Next we are going to show how to optimize the model setting different restrictions o... | m.rbf.variance.unconstrain()
print m
m.unconstrain()
print m |
Name : sparse_gp
Objective : -613.999681976
Number of Parameters : 8
Number of Optimization Parameters : 8
Updates : True
Parameters:
[1msparse_gp. [0;0m | value | constraints | priors
[1minducing_inputs [0;0m | (5, 1) | |
[1mrbf.varianc... | MIT | tests/GPy/models_basic.ipynb | gopala-kr/ds-notebooks |
If you want to unconstrain only a specific constraint, you can call the respective method, such as `unconstrain_fixed()` (or `unfix()`) to only unfix fixed parameters: | m.inducing_inputs[0].fix()
m.rbf.constrain_positive()
print m
m.unfix()
print m |
Name : sparse_gp
Objective : -613.999681976
Number of Parameters : 8
Number of Optimization Parameters : 7
Updates : True
Parameters:
[1msparse_gp. [0;0m | value | constraints | priors
[1minducing_inputs [0;0m | (5, 1) | {fixed} |
[1mrbf.varianc... | MIT | tests/GPy/models_basic.ipynb | gopala-kr/ds-notebooks |
Tying ParametersNot yet implemented for GPy version 0.8.0 Optimizing the modelOnce we have finished defining the constraints, we can now optimize the model with the function optimize.: | m.Gaussian_noise.constrain_positive()
m.rbf.constrain_positive()
m.optimize() | No handlers could be found for logger "rbf"
| MIT | tests/GPy/models_basic.ipynb | gopala-kr/ds-notebooks |
By deafult, GPy uses the lbfgsb optimizer.Some optional parameters may be discussed here. * `optimizer`: which optimizer to use, currently there are lbfgsb, fmin_tnc, scg, simplex or any unique identifier uniquely identifying an optimizer.Thus, you can say m.optimize('bfgs') for using the `lbfgsb` optimizer * `messages... | fig = m.plot() | /home/nbuser/anaconda2_501/lib/python2.7/site-packages/matplotlib/figure.py:1999: UserWarning:This figure includes Axes that are not compatible with tight_layout, so results might be incorrect.
| MIT | tests/GPy/models_basic.ipynb | gopala-kr/ds-notebooks |
We can even change the backend for plotting and plot the model using a different backend. | GPy.plotting.change_plotting_library('plotly')
fig = m.plot(plot_density=True)
GPy.plotting.show(fig, filename='gpy_sparse_gp_example') | This is the format of your plot grid:
[ (1,1) x1,y1 ]
Aw, snap! We don't have an account for ''. Want to try again? You can authenticate with your email address or username. Sign in is not case sensitive.
Don't have an account? plot.ly
Questions? support@plot.ly
| MIT | tests/GPy/models_basic.ipynb | gopala-kr/ds-notebooks |
Partial Dependence PlotsSigurd Carlsen Feb 2019Holger Nahrstaedt 2020.. currentmodule:: skoptPlot objective now supports optional use of partial dependence as well asdifferent methods of defining parameter values for dependency plots. | print(__doc__)
import sys
from skopt.plots import plot_objective
from skopt import forest_minimize
import numpy as np
np.random.seed(123)
import matplotlib.pyplot as plt | _____no_output_____ | BSD-3-Clause | dev/notebooks/auto_examples/plots/partial-dependence-plot.ipynb | scikit-optimize/scikit-optimize.github.io |
Objective functionPlot objective now supports optional use of partial dependence as well asdifferent methods of defining parameter values for dependency plots | # Here we define a function that we evaluate.
def funny_func(x):
s = 0
for i in range(len(x)):
s += (x[i] * i) ** 2
return s | _____no_output_____ | BSD-3-Clause | dev/notebooks/auto_examples/plots/partial-dependence-plot.ipynb | scikit-optimize/scikit-optimize.github.io |
Optimisation using decision treesWe run forest_minimize on the function | bounds = [(-1, 1.), ] * 3
n_calls = 150
result = forest_minimize(funny_func, bounds, n_calls=n_calls,
base_estimator="ET",
random_state=4) | _____no_output_____ | BSD-3-Clause | dev/notebooks/auto_examples/plots/partial-dependence-plot.ipynb | scikit-optimize/scikit-optimize.github.io |
Partial dependence plotHere we see an example of using partial dependence. Even when settingn_points all the way down to 10 from the default of 40, this method isstill very slow. This is because partial dependence calculates 250 extrapredictions for each point on the plots. | _ = plot_objective(result, n_points=10) | _____no_output_____ | BSD-3-Clause | dev/notebooks/auto_examples/plots/partial-dependence-plot.ipynb | scikit-optimize/scikit-optimize.github.io |
It is possible to change the location of the red dot, which normally showsthe position of the found minimum. We can set it 'expected_minimum',which is the minimum value of the surrogate function, obtained by aminimum search method. | _ = plot_objective(result, n_points=10, minimum='expected_minimum') | _____no_output_____ | BSD-3-Clause | dev/notebooks/auto_examples/plots/partial-dependence-plot.ipynb | scikit-optimize/scikit-optimize.github.io |
Plot without partial dependenceHere we plot without partial dependence. We see that it is a lot faster.Also the values for the other parameters are set to the default "result"which is the parameter set of the best observed value so far. In the caseof funny_func this is close to 0 for all parameters. | _ = plot_objective(result, sample_source='result', n_points=10) | _____no_output_____ | BSD-3-Clause | dev/notebooks/auto_examples/plots/partial-dependence-plot.ipynb | scikit-optimize/scikit-optimize.github.io |
Modify the shown minimumHere we try with setting the `minimum` parameters to something other than"result". First we try with "expected_minimum" which is the set ofparameters that gives the miniumum value of the surrogate function,using scipys minimum search method. | _ = plot_objective(result, n_points=10, sample_source='expected_minimum',
minimum='expected_minimum') | _____no_output_____ | BSD-3-Clause | dev/notebooks/auto_examples/plots/partial-dependence-plot.ipynb | scikit-optimize/scikit-optimize.github.io |
"expected_minimum_random" is a naive way of finding the minimum of thesurrogate by only using random sampling: | _ = plot_objective(result, n_points=10, sample_source='expected_minimum_random',
minimum='expected_minimum_random') | _____no_output_____ | BSD-3-Clause | dev/notebooks/auto_examples/plots/partial-dependence-plot.ipynb | scikit-optimize/scikit-optimize.github.io |
We can also specify how many initial samples are used for the two different"expected_minimum" methods. We set it to a low value in the next examplesto showcase how it affects the minimum for the two methods. | _ = plot_objective(result, n_points=10, sample_source='expected_minimum_random',
minimum='expected_minimum_random',
n_minimum_search=10)
_ = plot_objective(result, n_points=10, sample_source="expected_minimum",
minimum='expected_minimum', n_minimum_search=2) | _____no_output_____ | BSD-3-Clause | dev/notebooks/auto_examples/plots/partial-dependence-plot.ipynb | scikit-optimize/scikit-optimize.github.io |
Set a minimum locationLastly we can also define these parameters ourself by parsing a listas the minimum argument: | _ = plot_objective(result, n_points=10, sample_source=[1, -0.5, 0.5],
minimum=[1, -0.5, 0.5]) | _____no_output_____ | BSD-3-Clause | dev/notebooks/auto_examples/plots/partial-dependence-plot.ipynb | scikit-optimize/scikit-optimize.github.io |
Training ModelsThe central goal of machine learning is to train predictive models that can be used by applications. In Azure Machine Learning, you can use scripts to train models leveraging common machine learning frameworks like Scikit-Learn, Tensorflow, PyTorch, SparkML, and others. You can run these training scrip... | import azureml.core
from azureml.core import Workspace
# Load the workspace from the saved config file
ws = Workspace.from_config()
print('Ready to use Azure ML {} to work with {}'.format(azureml.core.VERSION, ws.name)) | Ready to use Azure ML 1.17.0 to work with ml-sdk
| MIT | 02-Training_Models.ipynb | djanie1/mslearn-aml-labs |
Create a Training ScriptYou're going to use a Python script to train a machine learning model based on the diabates data, so let's start by creating a folder for the script and data files. | import os, shutil
# Create a folder for the experiment files
training_folder = 'diabetes-training'
os.makedirs(training_folder, exist_ok=True)
# Copy the data file into the experiment folder
shutil.copy('data/diabetes.csv', os.path.join(training_folder, "diabetes.csv")) | _____no_output_____ | MIT | 02-Training_Models.ipynb | djanie1/mslearn-aml-labs |
Now you're ready to create the training script and save it in the folder. | %%writefile $training_folder/diabetes_training.py
# Import libraries
from azureml.core import Run
import pandas as pd
import numpy as np
import joblib
import os
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import roc_auc_score
from sklearn... | Overwriting diabetes-training/diabetes_training.py
| MIT | 02-Training_Models.ipynb | djanie1/mslearn-aml-labs |
Use an Estimator to Run the Script as an ExperimentYou can run experiment scripts using a **RunConfiguration** and a **ScriptRunConfig**, or you can use an **Estimator**, which abstracts both of these configurations in a single object.In this case, we'll use a generic **Estimator** object to run the training experimen... | from azureml.train.estimator import Estimator
from azureml.core import Experiment
# Create an estimator
estimator = Estimator(source_directory=training_folder,
entry_script='diabetes_training.py',
compute_target='local',
conda_packages=['scikit-learn']
... | WARNING - If 'script' has been provided here and a script file name has been specified in 'run_config', 'script' provided in ScriptRunConfig initialization will take precedence.
| MIT | 02-Training_Models.ipynb | djanie1/mslearn-aml-labs |
As with any experiment run, you can use the **RunDetails** widget to view information about the run and get a link to it in Azure Machine Learning studio. | from azureml.widgets import RunDetails
RunDetails(run).show() | _____no_output_____ | MIT | 02-Training_Models.ipynb | djanie1/mslearn-aml-labs |
You can also retrieve the metrics and outputs from the **Run** object. | # Get logged metrics
metrics = run.get_metrics()
for key in metrics.keys():
print(key, metrics.get(key))
print('\n')
for file in run.get_file_names():
print(file) | Regularization Rate 0.01
Accuracy 0.774
AUC 0.8484929598487486
azureml-logs/60_control_log.txt
azureml-logs/70_driver_log.txt
logs/azureml/8_azureml.log
outputs/diabetes_model.pkl
| MIT | 02-Training_Models.ipynb | djanie1/mslearn-aml-labs |
Register the Trained ModelNote that the outputs of the experiment include the trained model file (**diabetes_model.pkl**). You can register this model in your Azure Machine Learning workspace, making it possible to track model versions and retrieve them later. | from azureml.core import Model
# Register the model
run.register_model(model_path='outputs/diabetes_model.pkl', model_name='diabetes_model',
tags={'Training context':'Estimator'},
properties={'AUC': run.get_metrics()['AUC'], 'Accuracy': run.get_metrics()['Accuracy']})
# List regi... | diabetes_model version: 4
Training context : Estimator
AUC : 0.8484929598487486
Accuracy : 0.774
diabetes_mitigated_20 version: 1
diabetes_mitigated_19 version: 1
diabetes_mitigated_18 version: 1
diabetes_mitigated_17 version: 1
diabetes_mitigated_16 version: 1
diabetes_mitigated_15 version: 1
diabe... | MIT | 02-Training_Models.ipynb | djanie1/mslearn-aml-labs |
Create a Parameterized Training ScriptYou can increase the flexibility of your training experiment by adding parameters to your script, enabling you to repeat the same training experiment with different settings. In this case, you'll add a parameter for the regularization rate used by the Logistic Regression algorithm... | import os, shutil
# Create a folder for the experiment files
training_folder = 'diabetes-training-params'
os.makedirs(training_folder, exist_ok=True)
# Copy the data file into the experiment folder
shutil.copy('data/diabetes.csv', os.path.join(training_folder, "diabetes.csv")) | _____no_output_____ | MIT | 02-Training_Models.ipynb | djanie1/mslearn-aml-labs |
Now let's create a script containing a parameter for the regularization rate hyperparameter. | %%writefile $training_folder/diabetes_training.py
# Import libraries
from azureml.core import Run
import pandas as pd
import numpy as np
import joblib
import os
import argparse
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import roc_auc_sc... | Writing diabetes-training-params/diabetes_training.py
| MIT | 02-Training_Models.ipynb | djanie1/mslearn-aml-labs |
Use a Framework-Specific EstimatorYou used a generic **Estimator** class to run the training script, but you can also take advantage of framework-specific estimators that include environment definitions for common machine learning frameworks. In this case, you're using Scikit-Learn, so you can use the **SKLearn** esti... | from azureml.train.sklearn import SKLearn
from azureml.widgets import RunDetails
# Create an estimator
estimator = SKLearn(source_directory=training_folder,
entry_script='diabetes_training.py',
script_params = {'--reg_rate': 0.1},
compute_target='local'
... | WARNING - If 'script' has been provided here and a script file name has been specified in 'run_config', 'script' provided in ScriptRunConfig initialization will take precedence.
WARNING - If 'arguments' has been provided here and arguments have been specified in 'run_config', 'arguments' provided in ScriptRunConfig ini... | MIT | 02-Training_Models.ipynb | djanie1/mslearn-aml-labs |
Once again, you can get the metrics and outputs from the run. | # Get logged metrics
metrics = run.get_metrics()
for key in metrics.keys():
print(key, metrics.get(key))
print('\n')
for file in run.get_file_names():
print(file) | Regularization Rate 0.1
Accuracy 0.7736666666666666
AUC 0.8483904671874223
azureml-logs/60_control_log.txt
azureml-logs/70_driver_log.txt
logs/azureml/8_azureml.log
outputs/diabetes_model.pkl
| MIT | 02-Training_Models.ipynb | djanie1/mslearn-aml-labs |
Register A New Version of the ModelNow that you've trained a new model, you can register it as a new version in the workspace. | from azureml.core import Model
# Register the model
run.register_model(model_path='outputs/diabetes_model.pkl', model_name='diabetes_model',
tags={'Training context':'Parameterized SKLearn Estimator'},
properties={'AUC': run.get_metrics()['AUC'], 'Accuracy': run.get_metrics()['Acc... | diabetes_model version: 5
Training context : Parameterized SKLearn Estimator
AUC : 0.8483904671874223
Accuracy : 0.7736666666666666
diabetes_model version: 4
Training context : Estimator
AUC : 0.8484929598487486
Accuracy : 0.774
diabetes_mitigated_20 version: 1
diabetes_mitigated_19 version: 1
diabe... | MIT | 02-Training_Models.ipynb | djanie1/mslearn-aml-labs |
CHALLENGE TASKStats Challege notebook Fit multiple linear regression for the following data and check for the assumptions using pythonX1 22 22 25 26 24 28 29 27 24 33 39 42X2 15 14 18 13 12 11 11 10 5 9 7 3Y 55 56 55 59 66 65 69 70 75 75 78 79 | import numpy as np
import pandas as pd
import statsmodels.api as sm
import matplotlib.pyplot as plt
""" Convert the data values into DataFrames"""
stats_chal={"X1":[22, 22, 25, 26, 24, 28, 29, 27, 24, 33, 39, 42],
"X2":[15, 14, 18, 13, 12, 11, 11, 10, 5, 9, 7, 3],
"Y":[55, 56, 55, 59, 66, 65, 6... | Intercept:
74.59582972285749
Coefficients:
[ 0. 0.33138486 -1.61056402]
| MIT | Stats_Live_MLR_challenge.ipynb | krishnavizster/Statistics |
CHECKING FOR LINEAR REGRESSION ASSUMPTIONS 1.Linear Relationship Aims at finding linear relationship between the independent and dependent variables TESTA simple visual way of determining this is through the use of scatter plot2.Variables follow a normal DistributionThis assumption ensures that for each value of indepe... | #Multicollinearity test
corr =df.corr()
print(corr)
#Linearity and Normality Test
import seaborn as sns
sns.set(style="ticks", color_codes=True, font_scale=2)
g=sns.pairplot(df, height=3, diag_kind="hist",kind="reg")
g.fig.suptitle("Scatter Plot",y=1.08)
X_test = sm.add_constant(X_test)
X_test
y_pred=mlr_model.pr... | _____no_output_____ | MIT | Stats_Live_MLR_challenge.ipynb | krishnavizster/Statistics |
Prepare stimuli in stereo with sync tone in the L channelTo syncrhonize the recording systems, each stimulus file goes in stereo, the L channel has the stimulus, and the R channel has a pure tone (500-5Khz).This is done here, with the help of the rigmq.util.stimprep moduleIt uses (or creates) a dictionary of {stim_fil... | import socket
import os
import sys
import logging
import warnings
import numpy as np
import glob
from rigmq.util import stimprep as sp
# setup the logger
logger = logging.getLogger()
handler = logging.StreamHandler()
formatter = logging.Formatter(
'%(asctime)s %(name)-12s %(levelname)-8s %(message)s')
handler... | 2019-06-27 16:36:59,810 rigmq.util.stimprep INFO Processing /Users/zeke/experiment/birds/g3v3/SongData/acute_0/bos.wav
2019-06-27 16:36:59,813 rigmq.util.stimprep INFO tag_freq = 1000
2019-06-27 16:36:59,815 rigmq.util.stimprep INFO Will resample from 40414 to 60621 sampes
2019-06-27 16:36:59,831 rigmq.util... | MIT | rigmq/util/prepare_audio_stims_debug.ipynb | zekearneodo/rigmq |
Scaling and Normalization | import pandas as pd
import numpy as np
import seaborn as sns
import matplotlib.pyplot as plt
from sklearn.preprocessing import StandardScaler, MinMaxScaler, RobustScaler
from scipy.cluster.vq import whiten | _____no_output_____ | MIT | Scaling and Normalization.ipynb | dbl007/python-cheat-sheet |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.