Datasets:

bigcode

/

jupyter-parsed

like 4

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "<a href=\"https://colab.research.google.com/github/rwarnung/datacrunch-notebooks/blob/master/dcrunch_R_example.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>", "_____no_output_____" ], [ "**Data crunch example R script**\n\n\n---\n\n\nauthor: sweet-richard\n\ndate: Jan 30, 2022\n\nrequired packages:\n\n* `tidyverse` for data handling\n* `feather` for efficient loading of data\n* `xgboost` for predictive modelling\n* `httr` for the automatic upload.\n\n", "_____no_output_____" ] ], [ [ "library(tidyverse)\nlibrary(feather)", "_____no_output_____" ] ], [ [ "First, we set some **parameters**.\n\n\n* `is_download` controls whether you want to download data or just read prevously downloaded data\n* `is_upload` set this to TRUE for automatic upload.\n* `nrounds` is a parameter for `xgboost` that we set to 100 for illustration. You might want to adjust the paramters of xgboost.\n\n\n\n\n\n\n\n\n\n", "_____no_output_____" ] ], [ [ "#' ## Parameters\nfile_name_train = \"train_data.feather\"\nfile_name_test =\"test_data.feather\"\nis_download = TRUE # set this to true to download new data or to FALSE to load data in feather format\nis_upload = FALSE # set this to true to upload a submission\nnrounds = 300 # you might want to adjust this one and other parameters of xgboost", "_____no_output_____" ] ], [ [ "In the **functions** section we defined the correlation measure that we use to measure performance.", "_____no_output_____" ] ], [ [ "#' ## Functions\n#+ \n\ngetCorrMeasure = function(actual, predicted) {\n cor_measure = cor(actual, predicted, method=\"spearman\")\n return(cor_measure)\n}", "_____no_output_____" ] ], [ [ "Now, we either **download** the current data from the servers or load them in feather format. Furthermore, we define the features that we actually want to use. In this illustration we use all of them but `id` and `Moons`.", "_____no_output_____" ] ], [ [ "#' ## Download data\n#' after the download, data is stored in feather format to be read on demand quickly. Data is stored in integer format to save memory.\n#+\n\nif( is_download ) {\n \n cat(\"\\n start download\")\n train_datalink_X = 'https://tournament.datacrunch.com/data/X_train.csv' \n train_datalink_y = 'https://tournament.datacrunch.com/data/y_train.csv' \n hackathon_data_link = 'https://tournament.datacrunch.com/data/X_test.csv' \n \n train_dataX = read_csv(url(train_datalink_X))\n train_dataY = read_csv(url(train_datalink_y))\n test_data = read_csv(url(hackathon_data_link))\n \n train_data = \n bind_cols( train_dataX, train_dataY)\n \n train_data = train_data %>% mutate_at(vars(starts_with(\"feature_\")), list(~as.integer(.*100)))\n feather::write_feather(train_data, path = paste0(\"./\", file_name_train))\n \n test_data = test_data %>% mutate_at(vars(starts_with(\"feature_\")), list(~as.integer(.*100)))\n feather::write_feather(test_data, path = paste0(\"./\", file_name_test))\n \n names(train_data)\n \n nrow(train_data)\n nrow(test_data)\n \n cat(\"\\n data is downloaded\")\n} else {\n train_data = feather::read_feather(path = paste0(\"./\", file_name_train))\n test_data = feather::read_feather(path = paste0(\"./\", file_name_test))\n}\n\n## set vars used for modelling \nmodel_vars = setdiff(names(test_data), c(\"id\",\"Moons\"))", "\n start download" ] ], [ [ "Next we fit our go-to algorithm **xgboost** with mainly default parameters, only `eta` and `max_depth` are set.", "_____no_output_____" ] ], [ [ "#' ## Fit xgboost\n#+ cache = TRUE\n\nlibrary(xgboost, warn.conflicts = FALSE)\n\n# custom loss function for eval \ncorrmeasure <- function(preds, dtrain) {\n labels <- getinfo(dtrain, \"label\")\n corrm <- as.numeric(cor(labels, preds, method=\"spearman\"))\n return(list(metric = \"corr\", value = corrm))\n}\n \neval_metric_string = \"rmse\"\nmy_objective = \"reg:squarederror\"\n\ntree.params = list(\n booster = \"gbtree\", eta = 0.01, max_depth = 5,\n tree_method = \"hist\", # tree_method = \"auto\",\n objective = my_objective)\n\ncat(\"\\n starting xgboost \\n\")\n", "\n starting xgboost \n" ] ], [ [ "**First target** `target_r`", "_____no_output_____" ] ], [ [ "# first target target_r then g and b\n################\ncurrent_target = \"target_r\"\n \ndtrain = xgb.DMatrix(train_data %>% select(one_of(model_vars)) %>% as.matrix(), label = train_data %>% select(one_of(current_target)) %>% as.matrix())\n \nxgb.model.tree = xgb.train(data = dtrain, \n params = tree.params, nrounds = nrounds, verbose = 1, \n print_every_n = 50L, eval_metric = corrmeasure) \n \nxgboost_tree_train_pred1 = predict(xgb.model.tree, train_data %>% select(one_of(model_vars)) %>% as.matrix())\nxgboost_tree_live_pred1 = predict(xgb.model.tree, test_data %>% select(one_of(model_vars)) %>% as.matrix())\n \ncor_train = getCorrMeasure(train_data %>% select(one_of(current_target)), xgboost_tree_train_pred1)\n \ncat(\"\\n : metric: \", eval_metric_string, \"\\n\")\nprint(paste0(\"Corrm on train: \", round(cor_train,4)))\n\nprint(paste(\"xgboost\", current_target, \"ready\"))", "\n : metric: rmse \n[1] \"Corrm on train: 0.0867\"\n[1] \"xgboost target_r ready\"\n" ] ], [ [ "**Second target** `target_g`", "_____no_output_____" ] ], [ [ "# second target target_g\n################\ncurrent_target = \"target_g\"\n \ndtrain = xgb.DMatrix(train_data %>% select(one_of(model_vars)) %>% as.matrix(), label = train_data %>% select(one_of(current_target)) %>% as.matrix())\n\nxgb.model.tree = xgb.train(data = dtrain, \n params = tree.params, nrounds = nrounds, verbose = 1, \n print_every_n = 50L, eval_metric = corrmeasure) \n\nxgboost_tree_train_pred2 = predict(xgb.model.tree, train_data %>% select(one_of(model_vars)) %>% as.matrix())\nxgboost_tree_live_pred2 = predict(xgb.model.tree, test_data %>% select(one_of(model_vars)) %>% as.matrix())\n\ncor_train = getCorrMeasure(train_data %>% select(one_of(current_target)), xgboost_tree_train_pred2)\n\ncat(\"\\n : metric: \", eval_metric_string, \"\\n\")\nprint(paste0(\"Corrm on train: \", round(cor_train,4)))\n\nprint(paste(\"xgboost\", current_target, \"ready\"))", "\n : metric: rmse \n[1] \"Corrm on train: 0.1099\"\n[1] \"xgboost target_g ready\"\n" ] ], [ [ "**Third target** `target_b`", "_____no_output_____" ] ], [ [ "# third target target_b\n################\ncurrent_target = \"target_b\"\n\ndtrain = xgb.DMatrix(train_data %>% select(one_of(model_vars)) %>% as.matrix(), label = train_data %>% select(one_of(current_target)) %>% as.matrix())\n\nxgb.model.tree = xgb.train(data = dtrain, \n params = tree.params, nrounds = nrounds, verbose = 1, \n print_every_n = 50L, eval_metric = corrmeasure) \n\nxgboost_tree_train_pred3 = predict(xgb.model.tree, train_data %>% select(one_of(model_vars)) %>% as.matrix())\nxgboost_tree_live_pred3 = predict(xgb.model.tree, test_data %>% select(one_of(model_vars)) %>% as.matrix())\n\ncor_train = getCorrMeasure(train_data %>% select(one_of(current_target)), xgboost_tree_train_pred3)\n\ncat(\"\\n : metric: \", eval_metric_string, \"\\n\")\nprint(paste0(\"Corrm on train: \", round(cor_train,4)))\n\nprint(paste(\"xgboost\", current_target, \"ready\"))", "\n : metric: rmse \n[1] \"Corrm on train: 0.1232\"\n[1] \"xgboost target_b ready\"\n" ] ], [ [ "Then we produce simply histogram plots to see whether the predictions are plausible and prepare a **submission file**:", "_____no_output_____" ] ], [ [ "#' ## Submission\n#' simple histograms to check the submissions\n#+\nhist(xgboost_tree_live_pred1)\nhist(xgboost_tree_live_pred2)\nhist(xgboost_tree_live_pred3)\n \n#' create submission file\n#+\nsub_df = tibble(target_r = xgboost_tree_live_pred1, \n target_g = xgboost_tree_live_pred2, \n target_b = xgboost_tree_live_pred3)\n \nfile_name_submission = paste0(\"gbTree_\", gsub(\"-\",\"\",Sys.Date()), \".csv\")\n\nsub_df %>% readr::write_csv(file = paste0(\"./\", file_name_submission))\nnrow(sub_df)\n\ncat(\"\\n submission file written\")", "_____no_output_____" ] ], [ [ "Finally, we can **automatically upload** the file to the server:", "_____no_output_____" ] ], [ [ "#' ## Upload submission\n#+\nif( is_upload ) {\n library(httr)\n \n API_KEY = \"YourKeyHere\"\n \n response <- POST(\n url = \"https://tournament.crunchdao.com/api/v2/submissions\",\n query = list(apiKey = API_KEY),\n body = list(\n file = upload_file(path = paste0(\"./\", file_name_submission))\n ),\n encode = c(\"multipart\")\n );\n \n status <- status_code(response)\n \n if (status == 200) {\n print(\"Submission submitted :)\")\n } else if (status == 400) {\n print(\"ERR: The file must not be empty\")\n print(\"You have send a empty file.\")\n } else if (status == 401) {\n print(\"ERR: Your email hasn't been verified\")\n print(\"Please verify your email or contact a cruncher.\")\n } else if (status == 403) {\n print(\"ERR: Not authentified\")\n print(\"Is the API Key valid?\")\n } else if (status == 404) {\n print(\"ERR: Unknown API Key\")\n print(\"You should check that the provided API key is valid and is the same as the one you've received by email.\")\n } else if (status == 409) {\n print(\"ERR: Duplicate submission\")\n print(\"Your work has already been submitted with the same exact results, if you think that this a false positive, contact a cruncher.\")\n print(\"MD5 collision probability: 1/2^128 (source: https://stackoverflow.com/a/288519/7292958)\")\n } else if (status == 422) {\n print(\"ERR: API Key is missing or empty\")\n print(\"Did you forget to fill the API_KEY variable?\")\n } else if (status == 423) {\n print(\"ERR: Submissions are close\")\n print(\"You can only submit during rounds eg: Friday 7pm GMT+1 to Sunday midnight GMT+1.\")\n print(\"Or the server is currently crunching the submitted files, please wait some time before retrying.\")\n } else if (status == 423) {\n print(\"ERR: Too many submissions\")\n } else {\n content <- httr::content(response)\n print(\"ERR: Server returned: \" + toString(status))\n print(\"Ouch! It seems that we were not expecting this kind of result from the server, if the probleme persist, contact a cruncher.\")\n print(paste(\"Message:\", content$message, sep=\" \"))\n }\n \n # DEVELOPER WARNING:\n # THE API ERROR CODE WILL BE HANDLER DIFFERENTLY IN A NEAR FUTURE!\n # PLEASE STAY UPDATED BY JOINING THE DISCORD (https://discord.gg/veAtzsYn3M) AND READING THE NEWSLETTER EMAIL\n}", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "<table class=\"ee-notebook-buttons\" align=\"left\">\n <td><a target=\"_blank\" href=\"https://github.com/giswqs/earthengine-py-notebooks/tree/master/Datasets/Terrain/srtm_mtpi.ipynb\"><img width=32px src=\"https://www.tensorflow.org/images/GitHub-Mark-32px.png\" /> View source on GitHub</a></td>\n <td><a target=\"_blank\" href=\"https://nbviewer.jupyter.org/github/giswqs/earthengine-py-notebooks/blob/master/Datasets/Terrain/srtm_mtpi.ipynb\"><img width=26px src=\"https://upload.wikimedia.org/wikipedia/commons/thumb/3/38/Jupyter_logo.svg/883px-Jupyter_logo.svg.png\" />Notebook Viewer</a></td>\n <td><a target=\"_blank\" href=\"https://mybinder.org/v2/gh/giswqs/earthengine-py-notebooks/master?filepath=Datasets/Terrain/srtm_mtpi.ipynb\"><img width=58px src=\"https://mybinder.org/static/images/logo_social.png\" />Run in binder</a></td>\n <td><a target=\"_blank\" href=\"https://colab.research.google.com/github/giswqs/earthengine-py-notebooks/blob/master/Datasets/Terrain/srtm_mtpi.ipynb\"><img src=\"https://www.tensorflow.org/images/colab_logo_32px.png\" /> Run in Google Colab</a></td>\n</table>", "_____no_output_____" ], [ "## Install Earth Engine API\nInstall the [Earth Engine Python API](https://developers.google.com/earth-engine/python_install) and [geehydro](https://github.com/giswqs/geehydro). The **geehydro** Python package builds on the [folium](https://github.com/python-visualization/folium) package and implements several methods for displaying Earth Engine data layers, such as `Map.addLayer()`, `Map.setCenter()`, `Map.centerObject()`, and `Map.setOptions()`.\nThe magic command `%%capture` can be used to hide output from a specific cell. Uncomment these lines if you are running this notebook for the first time.", "_____no_output_____" ] ], [ [ "# %%capture\n# !pip install earthengine-api\n# !pip install geehydro", "_____no_output_____" ] ], [ [ "Import libraries", "_____no_output_____" ] ], [ [ "import ee\nimport folium\nimport geehydro", "_____no_output_____" ] ], [ [ "Authenticate and initialize Earth Engine API. You only need to authenticate the Earth Engine API once. Uncomment the line `ee.Authenticate()` \nif you are running this notebook for the first time or if you are getting an authentication error. ", "_____no_output_____" ] ], [ [ "# ee.Authenticate()\nee.Initialize()", "_____no_output_____" ] ], [ [ "## Create an interactive map \nThis step creates an interactive map using [folium](https://github.com/python-visualization/folium). The default basemap is the OpenStreetMap. Additional basemaps can be added using the `Map.setOptions()` function. \nThe optional basemaps can be `ROADMAP`, `SATELLITE`, `HYBRID`, `TERRAIN`, or `ESRI`.", "_____no_output_____" ] ], [ [ "Map = folium.Map(location=[40, -100], zoom_start=4)\nMap.setOptions('HYBRID')", "_____no_output_____" ] ], [ [ "## Add Earth Engine Python script ", "_____no_output_____" ] ], [ [ "dataset = ee.Image('CSP/ERGo/1_0/Global/SRTM_mTPI')\nsrtmMtpi = dataset.select('elevation')\nsrtmMtpiVis = {\n 'min': -200.0,\n 'max': 200.0,\n 'palette': ['0b1eff', '4be450', 'fffca4', 'ffa011', 'ff0000'],\n}\nMap.setCenter(-105.8636, 40.3439, 11)\nMap.addLayer(srtmMtpi, srtmMtpiVis, 'SRTM mTPI')\n", "_____no_output_____" ] ], [ [ "## Display Earth Engine data layers ", "_____no_output_____" ] ], [ [ "Map.setControlVisibility(layerControl=True, fullscreenControl=True, latLngPopup=True)\nMap", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "#Introduction to Data Science\n\nSee [Lesson 1](https://www.udacity.com/course/intro-to-data-analysis--ud170)\n\nYou should run it in local Jupyter env as this notebook refers to local dataset", "_____no_output_____" ] ], [ [ "import unicodecsv\nfrom datetime import datetime as dt\n\nenrollments_filename = 'dataset/enrollments.csv'\nengagement_filename = 'dataset/daily_engagement.csv'\nsubmissions_filename = 'dataset/project_submissions.csv'\n\n## Longer version of code (replaced with shorter, equivalent version below)\n\ndef read_csv(filename):\n with open(filename, 'rb') as f:\n reader = unicodecsv.DictReader(f)\n return list(reader)\n\nenrollments = read_csv(enrollments_filename)\ndaily_engagement = read_csv(engagement_filename)\nproject_submissions = read_csv(submissions_filename)\n\n\ndef renameKey(data, fromKey, toKey):\n for rec in data:\n if fromKey in rec:\n rec[toKey] = rec[fromKey]\n del rec[fromKey]\nrenameKey(daily_engagement, 'acct', 'account_key')\n\ndef cleanDataTypes():\n def fixIntFloat(data, field):\n if field not in data:\n print(f'WARNING : Field {field} is not in {data}')\n value = data[field]\n\n if value == '':\n data[field] = None\n else:\n data[field] = int(float(value))\n\n def fixFloat(data, field):\n if field not in data:\n print(f'WARNING : Field {field} is not in {data}')\n value = data[field]\n\n if value == '':\n data[field] = None\n else:\n data[field] = float(value)\n\n def fixDate(data, field):\n if field not in data:\n print(f'WARNING : Field {field} is not in {data}')\n value = data[field]\n\n if value == '':\n data[field] = None\n else:\n data[field] = dt.strptime(value, '%Y-%m-%d')\n\n def fixBool(data, field):\n if field not in data:\n print(f'WARNING : Field {field} is not in {data}')\n value = data[field]\n\n if value == 'True':\n data[field] = True\n elif value == 'False':\n data[field] = False\n else:\n print(f\"WARNING: invalid boolean '{value}' value converted to False in {data}\")\n data[field] = False\n\n def fixInt(data, field):\n if field not in data:\n print(f'WARNING : Field {field} is not in {data}')\n value = data[field]\n if value == '':\n data[field] = None\n else:\n data[field] = int(value)\n\n #clean data types\n for rec in enrollments:\n fixInt(rec, 'days_to_cancel')\n fixDate(rec, 'join_date')\n fixDate(rec, 'cancel_date')\n fixBool(rec, 'is_udacity')\n fixBool(rec, 'is_canceled')\n\n for rec in daily_engagement:\n fixDate(rec, 'utc_date')\n fixIntFloat(rec, 'num_courses_visited')\n fixFloat(rec, 'total_minutes_visited')\n fixIntFloat(rec, 'lessons_completed')\n fixIntFloat(rec, 'projects_completed')\n\n for rec in project_submissions:\n fixDate(rec, 'creation_date')\n fixDate(rec, 'completion_date')\n\ncleanDataTypes()\n\nprint(f\"enrollments[0] = {enrollments[0]}\\n\")\nprint(f\"daily_engagement[0] = {daily_engagement[0]}\\n\")\nprint(f\"project_submissions[0] = {project_submissions[0]}\\n\")", "enrollments[0] = {'account_key': '448', 'status': 'canceled', 'join_date': datetime.datetime(2014, 11, 10, 0, 0), 'cancel_date': datetime.datetime(2015, 1, 14, 0, 0), 'days_to_cancel': 65, 'is_udacity': True, 'is_canceled': True}\n\ndaily_engagement[0] = {'utc_date': datetime.datetime(2015, 1, 9, 0, 0), 'num_courses_visited': 1, 'total_minutes_visited': 11.6793745, 'lessons_completed': 0, 'projects_completed': 0, 'account_key': '0'}\n\nproject_submissions[0] = {'creation_date': datetime.datetime(2015, 1, 14, 0, 0), 'completion_date': datetime.datetime(2015, 1, 16, 0, 0), 'assigned_rating': 'UNGRADED', 'account_key': '256', 'lesson_key': '3176718735', 'processing_state': 'EVALUATED'}\n\n" ], [ "from collections import defaultdict\n\ndef getUniqueAccounts(data):\n accts = defaultdict(list)\n i = 0\n for record in data:\n accountKey = record['account_key']\n accts[accountKey].append(i)\n i+=1\n \n return accts\n\nenrollment_num_rows = len(enrollments)\nenrollment_unique_students = getUniqueAccounts(enrollments)\nenrollment_num_unique_students = len(enrollment_unique_students)\n\nengagement_num_rows = len(daily_engagement)\nengagement_unique_students = getUniqueAccounts(daily_engagement)\nengagement_num_unique_students = len(engagement_unique_students)\n\nsubmission_num_rows = len(project_submissions)\nsubmission_unique_students = getUniqueAccounts(project_submissions)\nsubmission_num_unique_students = len(submission_unique_students)\n\nprint(f\"enrollments total={enrollment_num_rows}, unique={enrollment_num_unique_students}\")\n\nprint(f\"engagements total={engagement_num_rows}, unique={engagement_num_unique_students}\")\n\nprint(f\"submissions total={submission_num_rows} unique={submission_num_unique_students}\")\n", "enrollments total=1640, unique=1302\nengagements total=136240, unique=1237\nsubmissions total=3642 unique=743\n" ], [ "for enrollment_acct in enrollment_unique_students:\n if enrollment_acct not in engagement_unique_students:\n #print(enrollment_unique_students[enrollment])\n enrollment_id = enrollment_unique_students[enrollment_acct][0]\n enrollment = enrollments[enrollment_id]\n print(f\"Strange student : enrollment={enrollment}\")\n break\n", "Strange student : enrollment={'account_key': '1219', 'status': 'canceled', 'join_date': datetime.datetime(2014, 11, 12, 0, 0), 'cancel_date': datetime.datetime(2014, 11, 12, 0, 0), 'days_to_cancel': 0, 'is_udacity': False, 'is_canceled': True}\n" ], [ "strange_enrollments_num_by_different_date = 0\n\nfor enrollment_acct in enrollment_unique_students:\n if enrollment_acct not in engagement_unique_students:\n for enrollment_id in enrollment_unique_students[enrollment_acct]:\n enrollment = enrollments[enrollment_id]\n if enrollment['join_date'] != enrollment['cancel_date']:\n strange_enrollments_num_by_different_date += 1\n #print(f\"Strange student with different dates : enrollments[{enrollment_id}]={enrollment}\\n\")\n\nprint(f\"Number of enrolled and cancelled at different dates but not engaged (problemactic accounts) : {strange_enrollments_num_by_different_date}\\n\")", "Number of enrolled and cancelled at different dates but not engaged (problemactic accounts) : 3\n\n" ], [ "num_problems = 0\nfor enrollment in enrollments:\n student = enrollment['account_key']\n if student not in engagement_unique_students and enrollment['join_date'] != enrollment['cancel_date']:\n num_problems += 1\n #print(enrollment)\n\nprint(f'Number of problematic account records : {num_problems}')", "Number of problematic account records : 3\n" ], [ "def getRealAccounts(enrollmentData):\n result = []\n for rec in enrollmentData:\n if not rec['is_udacity']:\n result.append(rec)\n return result\n\nreal_enrollments = getRealAccounts(enrollments)\nprint(f'Real account : {len(real_enrollments)}')", "Real account : 1622\n" ], [ "def getPaidStudents(enrollmentData):\n freePeriodDays = 7\n result = {}\n #result1 = {}\n for rec in enrollmentData:\n if rec['cancel_date'] == None or rec['days_to_cancel'] > freePeriodDays:\n accountKey = rec['account_key']\n joinDate = rec['join_date']\n if accountKey not in result or joinDate > result[accountKey]:\n result[accountKey] = joinDate\n #result1[accountKey] = joinDate\n '''\n for accountKey, joinDate in result.items():\n joinDate1 = result1[accountKey]\n if joinDate != joinDate1:\n print(f\"{accountKey} : {joinDate} != {joinDate1}\")\n '''\n return result\n\npaid_students = getPaidStudents(real_enrollments)\n\nprint(f'Paid students : {len(paid_students)}')", "Paid students : 995\n" ], [ "def isEngagementWithingOneWeek(joinDate, engagementDate):\n #if joinDate > engagementDate:\n # print(f'WARNING: join date is after engagement date')\n timeDelta = engagementDate - joinDate\n return 0 <= timeDelta.days and timeDelta.days < 7\n\n\ndef collectPaidEnagagementsInTheFirstWeek():\n result = []\n i = 0\n for engagement in daily_engagement:\n accountKey = engagement['account_key']\n if accountKey in paid_students:\n joinDate = paid_students[accountKey]\n engagementDate = engagement['utc_date']\n if isEngagementWithingOneWeek(joinDate, engagementDate):\n result.append(i)\n i+=1\n return result\n\npaid_engagement_in_first_week = collectPaidEnagagementsInTheFirstWeek()\n\nprint(f'Number of paid engagements in the first week : {len(paid_engagement_in_first_week)}')", "Number of paid engagements in the first week : 6919\n" ], [ "from collections import defaultdict\nimport numpy as np\n\ndef groupEngagementsByAccounts(engagements):\n result = defaultdict(list)\n for engagementId in engagements:\n engagement = daily_engagement[engagementId]\n accountKey = engagement['account_key']\n result[accountKey].append(engagementId)\n return result\n\nfirst_week_paid_engagements_by_account = groupEngagementsByAccounts(paid_engagement_in_first_week)\n\ndef sumEngagementsStatByAccount(engagements, getStatValue):\n result = {}\n for accountKey, engagementIds in engagements.items():\n stat_sum = 0\n for engagementId in engagementIds:\n engagement = daily_engagement[engagementId]\n stat_sum += getStatValue(engagement)\n result[accountKey] = stat_sum\n return result\n\ndef printStats(getStatValue, statLabel):\n first_week_paid_engagements_sum_stat_by_account = sumEngagementsStatByAccount(first_week_paid_engagements_by_account, getStatValue)\n first_week_paid_engagements_sum_stat = list(first_week_paid_engagements_sum_stat_by_account.values())\n\n print(f'Average {statLabel} spent by paid accounts during the first week : {np.mean(first_week_paid_engagements_sum_stat)}')\n print(f'StdDev {statLabel} spent by paid accounts during the first week : {np.std(first_week_paid_engagements_sum_stat)}')\n print(f'Min {statLabel} spent by paid accounts during the first week : {np.min(first_week_paid_engagements_sum_stat)}')\n print(f'Max {statLabel} spent by paid accounts during the first week : {np.max(first_week_paid_engagements_sum_stat)}')\n\n print('\\n')\n\nprintStats((lambda data : data['total_minutes_visited']), 'minutes')\nprintStats((lambda data : data['lessons_completed']), 'lessons')\nprintStats((lambda data : 1 if data['num_courses_visited'] > 0 else 0), 'days')\n", "Average minutes spent by paid accounts during the first week : 306.70832675342825\nStdDev minutes spent by paid accounts during the first week : 412.99693340852957\nMin minutes spent by paid accounts during the first week : 0.0\nMax minutes spent by paid accounts during the first week : 3564.7332644989997\n\n\nAverage lessons spent by paid accounts during the first week : 1.636180904522613\nStdDev lessons spent by paid accounts during the first week : 3.002561299829423\nMin lessons spent by paid accounts during the first week : 0\nMax lessons spent by paid accounts during the first week : 36\n\n\nAverage days spent by paid accounts during the first week : 2.8673366834170855\nStdDev days spent by paid accounts during the first week : 2.2551980029196814\nMin days spent by paid accounts during the first week : 0\nMax days spent by paid accounts during the first week : 7\n\n\n" ], [ "######################################\n# 11 #\n######################################\n\n## Create two lists of engagement data for paid students in the first week.\n## The first list should contain data for students who eventually pass the\n## subway project, and the second list should contain data for students\n## who do not.\n\nsubway_project_lesson_keys = {'746169184', '3176718735'}\npassing_grades = {'DISTINCTION', 'PASSED'} #{'', 'INCOMPLETE', 'DISTINCTION', 'PASSED', 'UNGRADED'}\n#passing_grades = {'PASSED'} #{'', 'INCOMPLETE', 'DISTINCTION', 'PASSED', 'UNGRADED'}\n\n\npassing_engagement = []\nnon_passing_engagement = []\n\nfor accountKey, engagementIds in first_week_paid_engagements_by_account.items():\n if accountKey in submission_unique_students:\n submissionIds = submission_unique_students[accountKey]\n isPassed = False\n for submissionId in submissionIds:\n submission = project_submissions[submissionId]\n if submission['assigned_rating'] in passing_grades and submission['lesson_key'] in subway_project_lesson_keys:\n isPassed = True\n break\n if isPassed:\n passing_engagement += engagementIds\n else:\n non_passing_engagement += engagementIds\n else:\n non_passing_engagement += engagementIds\n\nprint(f'First week engagements with passing grade : {len(passing_engagement)}')\nprint(f'First week engagements with non-passing grade : {len(non_passing_engagement)}')\n", "First week engagements with passing grade : 4527\nFirst week engagements with non-passing grade : 2392\n" ], [ "######################################\n# 12 #\n######################################\n\n## Compute some metrics you're interested in and see how they differ for\n## students who pass the subway project vs. students who don't. A good\n## starting point would be the metrics we looked at earlier (minutes spent\n## in the classroom, lessons completed, and days visited).\n\npassing_engagement_by_account = groupEngagementsByAccounts(passing_engagement)\nnon_passing_engagement_by_account = groupEngagementsByAccounts(non_passing_engagement)\n\ndef getArgStatEngagements(engagementIds, getStatValue):\n stat_sum = 0\n stat_num = 0\n for engagementId in engagementIds:\n engagement = daily_engagement[engagementId]\n stat_sum += getStatValue(engagement)\n stat_num += 1\n if stat_num > 0:\n return stat_sum / stat_num\n else:\n return 0\n\n#sumEngagementsStatByAccount(first_week_paid_engagements_by_account, getStatValue)\npassed_minutes = list(sumEngagementsStatByAccount(passing_engagement_by_account, (lambda data : data['total_minutes_visited'])).values())\nnon_passed_minutes = list(sumEngagementsStatByAccount(non_passing_engagement_by_account, (lambda data : data['total_minutes_visited'])).values())\npassed_lessons = list(sumEngagementsStatByAccount(passing_engagement_by_account, (lambda data : data['lessons_completed'])).values())\nnon_passed_lessons = list(sumEngagementsStatByAccount(non_passing_engagement_by_account, (lambda data : data['lessons_completed'])).values())\npassed_days = list(sumEngagementsStatByAccount(passing_engagement_by_account, (lambda data : 1 if data['num_courses_visited'] > 0 else 0)).values())\nnon_passed_days = list(sumEngagementsStatByAccount(non_passing_engagement_by_account, (lambda data : 1 if data['num_courses_visited'] > 0 else 0)).values())\n\n\nprint(f'Passed Avg Minutes = {np.mean(passed_minutes)}')\nprint(f'Non passed Avg Minutes = {np.mean(non_passed_minutes)}')\nprint(f'Passed Avg Lessons = {np.mean(passed_lessons)}')\nprint(f'Non passed Avg Lessons = {np.mean(non_passed_lessons)}')\nprint(f'Passed Avg Days = {np.mean(passed_days)}')\nprint(f'Non passed Avg Days = {np.mean(non_passed_days)}')\n", "Passed Avg Minutes = 394.58604648350865\nNon passed Avg Minutes = 143.32647426675584\nPassed Avg Lessons = 2.052550231839258\nNon passed Avg Lessons = 0.8620689655172413\nPassed Avg Days = 3.384853168469861\nNon passed Avg Days = 1.9051724137931034\n" ], [ "######################################\n# 13 #\n######################################\n\n## Make histograms of the three metrics we looked at earlier for both\n## students who passed the subway project and students who didn't. You\n## might also want to make histograms of any other metrics you examined.\n\n%matplotlib inline\n\nimport matplotlib.pyplot as plt\nimport seaborn as sns\n\nplt.hist(passed_minutes, color ='green')\nplt.hist(non_passed_minutes, color ='lightblue')\nplt.xlabel('Number of minutes')\nplt.title('Passed (green) VS Non-passed (light-blue) students')\n#sns.displot(passed_minutes, color ='green')\n#sns.displot(non_passed_minutes, color ='lightblue')", "_____no_output_____" ], [ "plt.hist(passed_lessons, color ='green')\nplt.hist(non_passed_lessons, color ='lightblue')\nplt.xlabel('Number of lessons')\nplt.title('Passed (green) VS Non-passed (light-blue) students')", "_____no_output_____" ], [ "plt.hist(passed_days, color ='green', bins = 8)\nplt.xlabel('Number of days')\nplt.title('Passed students')", "_____no_output_____" ], [ "plt.hist(non_passed_days, color ='lightblue', bins = 8)\nplt.xlabel('Number of days')\nplt.title('Non-passed students')", "_____no_output_____" ] ] ]

[ "markdown", "code" ]

[ [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "import pandas as pd\nimport numpy as np\nimport keras\nfrom keras.models import Sequential,Model\nfrom keras.layers import Dense, Dropout,BatchNormalization,Input\nfrom keras.optimizers import RMSprop\nfrom keras.regularizers import l2,l1\nfrom keras.optimizers import Adam\n\nfrom sklearn.model_selection import LeaveOneOut\nfrom sklearn.metrics import roc_curve, auc\nimport matplotlib.pyplot as plt\nfrom keras.callbacks import EarlyStopping", "Using TensorFlow backend.\n" ], [ "df = pd.read_csv(\"../../out_data/MLDB.csv\")", "_____no_output_____" ], [ "first_gene_index = df.columns.get_loc(\"rrrD\")", "_____no_output_____" ], [ "X, Y = np.split(df, [first_gene_index], axis=1)\nX = X.values\nX = X-0.5\nY1 = Y.values[:,1]\nY2 = Y.values[:,1]", "_____no_output_____" ], [ "X.shape", "_____no_output_____" ], [ "import collections\nModel_setting = collections.namedtuple('Model_setting','num_layers num_node alpha drop_rate act_method lr regularization \\\npatience')", "_____no_output_____" ], [ "setting_ = [1,100, 0.5, 0.2, 'tanh', 0.01, 'l2', 3]\nsetting = Model_setting(*setting_)\nsetting = setting._asdict()", "_____no_output_____" ], [ "setting", "_____no_output_____" ], [ "def getModel(setting,num_input=84):\n regularizer = l1(setting['alpha']) if setting['regularization']=='l1' else l2(setting['alpha'])\n \n model = Sequential()\n for i in range(setting['num_layers']):\n if i==0:\n model.add(Dense(setting['num_node'], input_shape=(num_input,), activation=setting['act_method'],\\\n kernel_regularizer = regularizer))\n model.add(Dropout(setting['drop_rate']))\n else:\n model.add(Dense(setting['num_node']//(2**i), activation=setting['act_method']))\n model.add(Dropout(setting['drop_rate']))\n model.add(Dense(1, activation='sigmoid'))\n model.compile(loss='binary_crossentropy', optimizer=Adam(lr=setting['lr']), metrics=['accuracy'])\n return model", "_____no_output_____" ], [ "num_output_ = 3\ndef create_model(num_input = 84,num_output = num_output_):\n X_input = Input(shape=(num_input,))\n\n X = Dense(64)(X_input)\n X = Dropout(0.2)(X)\n X = Dense(32)(X)\n Ys= []\n for i in range(num_output):\n Ys.append(Dense(1, activation = 'sigmoid')(X))\n model = Model(inputs=[X_input],outputs = Ys)\n model.compile(loss=['binary_crossentropy']*num_output,loss_weights=[1.]*num_output,optimizer=Adam(lr=setting['lr']), metrics=['accuracy'])\n return model", "_____no_output_____" ], [ "model = create_model()\ncallbacks = [EarlyStopping(monitor='loss',min_delta=0,patience=setting['patience'])]", "_____no_output_____" ], [ "ys = [*((Y.values).T[:num_output_])]", "_____no_output_____" ], [ "model.fit(X,ys,epochs = 50, verbose = 1,callbacks =callbacks)", "Epoch 1/50\n178/178 [==============================] - 0s - loss: 1.1467 - dense_8_loss: 0.5328 - dense_9_loss: 0.2510 - dense_10_loss: 0.3629 - dense_8_acc: 0.8258 - dense_9_acc: 0.8876 - dense_10_acc: 0.8371 \nEpoch 2/50\n178/178 [==============================] - 0s - loss: 0.6560 - dense_8_loss: 0.3028 - dense_9_loss: 0.1767 - dense_10_loss: 0.1766 - dense_8_acc: 0.8820 - dense_9_acc: 0.9775 - dense_10_acc: 0.9775 \nEpoch 3/50\n178/178 [==============================] - 0s - loss: 0.6066 - dense_8_loss: 0.3507 - dense_9_loss: 0.1281 - dense_10_loss: 0.1278 - dense_8_acc: 0.9213 - dense_9_acc: 0.9775 - dense_10_acc: 0.9719 \nEpoch 4/50\n178/178 [==============================] - 0s - loss: 0.4402 - dense_8_loss: 0.1729 - dense_9_loss: 0.1714 - dense_10_loss: 0.0959 - dense_8_acc: 0.9382 - dense_9_acc: 0.9775 - dense_10_acc: 0.9775 \nEpoch 5/50\n178/178 [==============================] - 0s - loss: 0.3534 - dense_8_loss: 0.1642 - dense_9_loss: 0.1119 - dense_10_loss: 0.0773 - dense_8_acc: 0.9270 - dense_9_acc: 0.9719 - dense_10_acc: 0.9831 \nEpoch 6/50\n178/178 [==============================] - 0s - loss: 0.3680 - dense_8_loss: 0.1766 - dense_9_loss: 0.0904 - dense_10_loss: 0.1010 - dense_8_acc: 0.9494 - dense_9_acc: 0.9775 - dense_10_acc: 0.9775 \nEpoch 7/50\n178/178 [==============================] - 0s - loss: 0.3434 - dense_8_loss: 0.1552 - dense_9_loss: 0.0953 - dense_10_loss: 0.0929 - dense_8_acc: 0.9551 - dense_9_acc: 0.9775 - dense_10_acc: 0.9775 \nEpoch 8/50\n178/178 [==============================] - 0s - loss: 0.3200 - dense_8_loss: 0.1422 - dense_9_loss: 0.0737 - dense_10_loss: 0.1041 - dense_8_acc: 0.9438 - dense_9_acc: 0.9775 - dense_10_acc: 0.9775 \nEpoch 9/50\n178/178 [==============================] - 0s - loss: 0.2566 - dense_8_loss: 0.1182 - dense_9_loss: 0.0744 - dense_10_loss: 0.0640 - dense_8_acc: 0.9607 - dense_9_acc: 0.9775 - dense_10_acc: 0.9888 \nEpoch 10/50\n178/178 [==============================] - 0s - loss: 0.2773 - dense_8_loss: 0.1246 - dense_9_loss: 0.0789 - dense_10_loss: 0.0738 - dense_8_acc: 0.9607 - dense_9_acc: 0.9775 - dense_10_acc: 0.9775 \nEpoch 11/50\n178/178 [==============================] - 0s - loss: 0.2452 - dense_8_loss: 0.1198 - dense_9_loss: 0.0804 - dense_10_loss: 0.0451 - dense_8_acc: 0.9663 - dense_9_acc: 0.9775 - dense_10_acc: 0.9719 \nEpoch 12/50\n178/178 [==============================] - 0s - loss: 0.2215 - dense_8_loss: 0.1145 - dense_9_loss: 0.0667 - dense_10_loss: 0.0403 - dense_8_acc: 0.9607 - dense_9_acc: 0.9719 - dense_10_acc: 0.9944 \nEpoch 13/50\n178/178 [==============================] - 0s - loss: 0.2623 - dense_8_loss: 0.1333 - dense_9_loss: 0.0784 - dense_10_loss: 0.0507 - dense_8_acc: 0.9494 - dense_9_acc: 0.9775 - dense_10_acc: 0.9775 \nEpoch 14/50\n178/178 [==============================] - 0s - loss: 0.2615 - dense_8_loss: 0.0934 - dense_9_loss: 0.0827 - dense_10_loss: 0.0854 - dense_8_acc: 0.9719 - dense_9_acc: 0.9775 - dense_10_acc: 0.9551 \nEpoch 15/50\n178/178 [==============================] - 0s - loss: 0.2276 - dense_8_loss: 0.1153 - dense_9_loss: 0.0632 - dense_10_loss: 0.0492 - dense_8_acc: 0.9663 - dense_9_acc: 0.9831 - dense_10_acc: 0.9775 \nEpoch 16/50\n178/178 [==============================] - 0s - loss: 0.2404 - dense_8_loss: 0.1128 - dense_9_loss: 0.0941 - dense_10_loss: 0.0335 - dense_8_acc: 0.9438 - dense_9_acc: 0.9775 - dense_10_acc: 0.9831 \n" ], [ "history = final_model.fit(X_train, [Y_train, Y_train2], \n nb_epoch = 100, \n batch_size = 256, \n verbose=1, \n validation_data=(X_test, [Y_test, Y_test2]),\n callbacks=[reduce_lr, checkpointer],\n shuffle=True)", "_____no_output_____" ], [ "callbacks = [EarlyStopping(monitor='loss',min_delta=0,patience=setting['patience'])]", "_____no_output_____" ], [ "def cross_validation(X,Y,setting,num_input):\n model = getModel(setting,num_input)\n preds = []\n for train, test in LeaveOneOut().split(X, Y):\n model.fit(X[train,:],Y[train],epochs=20,verbose=0, callbacks =callbacks)\n probas_ = model.predict(X[test,:])\n preds.append(probas_[0][0])\n # Compute ROC curve and area the curve\n fpr, tpr, thresholds = roc_curve(Y, preds)\n roc_auc = auc(fpr, tpr)\n if roc_auc < 0.5:\n roc_auc = 1 - roc_auc\n return roc_auc", "_____no_output_____" ], [ "\ndef backward_selection(X,Y,setting):\n survive_index=[i for i in range(X.shape[1])]\n best_perf=0 \n for i in range(len(survive_index)-1):\n perfs = []\n\n print(survive_index)\n for index in survive_index:\n print(index)\n survive_index_copy = [i for i in survive_index if i!=index]\n perfs.append(cross_validation(X[:,survive_index_copy],Y,setting,num_input = len(survive_index)-1))\n print(\"best_perf\",best_perf)\n max_index = np.argmax(perfs)\n current_best = np.max(perfs)\n\n print(\"current_best\",current_best)\n if current_best > best_perf:\n best_perf = current_best\n survive_index.remove(survive_index[max_index])\n else:\n break\n return (survive_index,best_perf)", "_____no_output_____" ], [ "backward_selection(X[:,0:10],Y,setting)", "[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]\n0\n1\n2\n3\n4\n5\n6\n7\n8\n9\nbest_perf 0\ncurrent_best 1.0\n[1, 2, 3, 4, 5, 6, 7, 8, 9]\n1\n" ], [ "fpr, tpr, thresholds = roc_curve(Y, preds)\nroc_auc = auc(fpr, tpr)\nplt.plot(fpr, tpr, lw=1, alpha=0.3)\nplt.title('(AUC = %0.2f)' % (roc_auc))\nplt.show()", "_____no_output_____" ], [ "def cross_validation(X=X,Y=Y,epochs_=20,num_input_ = 84):\n model = getModel(num_input=num_input_)\n preds = []\n for train, test in LeaveOneOut().split(X, Y):\n model.fit(X,Y,epochs=epochs_,verbose=0)\n # print(test)\n probas_ = model.predict(X[test,:])\n preds.append(probas_[0][0])\n # Compute ROC curve and area the curve\n fpr, tpr, thresholds = roc_curve(Y, preds)\n roc_auc = auc(fpr, tpr)\n return roc_auc", "_____no_output_____" ], [ "survive_index=[i for i in range(4)]\ndef backward_selection(survive_index):\n for i in range(len(survive_index)-1):\n perfs = []\n best_perf=0\n for index in survive_index:\n print(index,\"\\n\")\n survive_index_copy = [i for i in survive_index if i!=index]\n perfs.append(cross_validation(X=X[:,survive_index_copy],Y=Y,epochs_=20,num_input_ = len(survive_index)-1))\n \n max_index = np.argmax(perfs)\n current_best = np.max(perfs)\n print(current_best)\n if current_best > best_perf:\n best_perf = current_best\n survive_index.remove(survive_index[max_index])\n else:\n break\n return survive_index", "_____no_output_____" ], [ "backward_selection(survive_index)", "0 \n\n1 \n\n2 \n\n3 \n\n0.807926829268\n0 \n\n1 \n\n2 \n\n0.81881533101\n0 \n\n" ], [ "max_index = np.argmax(perfs)\nsurvive_index[max_index]", "_____no_output_____" ], [ "fpr, tpr, thresholds = roc_curve(Y, preds)\nroc_auc = auc(fpr, tpr)\nplt.plot(fpr, tpr, lw=1, alpha=0.3)\nplt.title('(AUC = %0.2f)' % (roc_auc))\nplt.show()", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Semana 14\n\n## Módulo 4b: Programação para História\n\n**Período**: 24/01/2022 a 28/01/2022\n\n**CH**: 2h30", "_____no_output_____" ], [ "## Live Coding 2 (AS)\n\n**Tema**: Python do Zero II\n\n**Data**: 02/02/2022, às 19h\n\n**CH**: 2h30\n\n**Plataforma**: `Google Meet` - link enviado por e-mail.\n\n```{Attention}\n[_Clique aqui para acessar a apresentação da aula_](https://ericbrasiln.github.io/intro-historia-digital/mod4b/sem13_ap.html).\n```", "_____no_output_____" ] ] ]

[ "markdown" ]

[ [ "markdown", "markdown" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "!pip install torch torchvision\nimport torch\nimport torch.nn as nn\nimport numpy as np\nimport matplotlib.pyplot as plt", "Requirement already satisfied: torch in /usr/local/lib/python3.6/dist-packages (1.6.0+cu101)\nRequirement already satisfied: torchvision in /usr/local/lib/python3.6/dist-packages (0.7.0+cu101)\nRequirement already satisfied: numpy in /usr/local/lib/python3.6/dist-packages (from torch) (1.18.5)\nRequirement already satisfied: future in /usr/local/lib/python3.6/dist-packages (from torch) (0.16.0)\nRequirement already satisfied: pillow>=4.1.1 in /usr/local/lib/python3.6/dist-packages (from torchvision) (7.0.0)\n" ], [ "# Hyper-parameters\ninput_size = 1\noutput_size = 1\nnum_epochs = 60\nlearning_rate = 0.001", "_____no_output_____" ], [ "\n# Toy dataset\nx_train = np.array([[3.3], [4.4], [5.5], [6.71], [6.93], [4.168], \n [9.779], [6.182], [7.59], [2.167], [7.042], \n [10.791], [5.313], [7.997], [3.1]], dtype=np.float32)\n\ny_train = np.array([[1.7], [2.76], [2.09], [3.19], [1.694], [1.573], \n [3.366], [2.596], [2.53], [1.221], [2.827], \n [3.465], [1.65], [2.904], [1.3]], dtype=np.float32)", "_____no_output_____" ], [ "# Linear regression model\nmodel = nn.Linear(input_size, output_size)", "_____no_output_____" ], [ "# Loss and optimizer\ncriterion = nn.MSELoss()\noptimizer = torch.optim.SGD(model.parameters(), lr=learning_rate) ", "_____no_output_____" ], [ "# Train the model\nfor epoch in range(num_epochs):\n # Convert numpy arrays to torch tensors\n inputs = torch.from_numpy(x_train)\n targets = torch.from_numpy(y_train)\n\n # Forward pass\n outputs = model(inputs)\n loss = criterion(outputs, targets)\n \n # Backward and optimize\n optimizer.zero_grad()\n loss.backward()\n optimizer.step()\n \n if (epoch+1) % 5 == 0:\n print ('Epoch [{}/{}], Loss: {:.4f}'.format(epoch+1, num_epochs, loss.item()))\n", "Epoch [5/60], Loss: 1.3239\nEpoch [10/60], Loss: 0.6392\nEpoch [15/60], Loss: 0.3618\nEpoch [20/60], Loss: 0.2494\nEpoch [25/60], Loss: 0.2038\nEpoch [30/60], Loss: 0.1854\nEpoch [35/60], Loss: 0.1779\nEpoch [40/60], Loss: 0.1749\nEpoch [45/60], Loss: 0.1736\nEpoch [50/60], Loss: 0.1731\nEpoch [55/60], Loss: 0.1729\nEpoch [60/60], Loss: 0.1728\n" ], [ "# Plot the graph\npredicted = model(torch.from_numpy(x_train)).detach().numpy()\nplt.plot(x_train, y_train, 'ro', label='Original data')\nplt.plot(x_train, predicted, label='Fitted line')\nplt.legend()\nplt.show()", "_____no_output_____" ], [ "# Save the model checkpoint\ntorch.save(model.state_dict(), 'model.ckpt')", "_____no_output_____" ], [ "", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "<a href=\"https://colab.research.google.com/github/SarahRebulado/OOP1_2/blob/main/GUI_Application.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>", "_____no_output_____" ] ], [ [ "#@title Students Grade in OOP\n\nStudent_Name1= \"Enter the student name\"#@param{type: \"string\"}\nprelim= 90#@param{type: \"number\"}\nmidterm= 95#@param{type: \"number\"}\nfinal= 100#@param{type: \"number\"}\nsemestral_grade=(prelim+midterm+final)/3\n\nprint(\"The prelim grade of student 1 is\"+\" \"+ str(prelim))\nprint(\"The midterm grade of student 1 is\"+\" \"+ str(midterm))\nprint(\"The final grade of student 1 is\"+\" \"+ str(final))\nprint(\"The semestral grade of student1 is \" + str(semestral_grade))\n\n#@title Gender\nGender=\"Female\" #@param[\"Male\", \"Female\"]\nprint(Gender)\n\nBirthdate= '2003-02-12' #@param{type: \"date\"}", "The prelim grade of student 1 is 90\nThe midterm grade of student 1 is 95\nThe final grade of student 1 is 100\nThe semestral grade of student1 is 95.0\nFemale\n" ] ] ]

[ "markdown", "code" ]

[ [ "markdown" ], [ "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "This notebook contains code for model comparison. Optimal hyperparameters for models are supposed to be already found.", "_____no_output_____" ], [ "# Imports", "_____no_output_____" ] ], [ [ "#imports\n!pip install scipydirect\n\nimport math\nimport numpy as np\nimport pandas as pd\nfrom sklearn.model_selection import train_test_split\nfrom sklearn.model_selection import cross_val_score\n\nfrom sklearn import preprocessing\nfrom sklearn.preprocessing import normalize\n\nfrom sklearn.ensemble import AdaBoostClassifier\nfrom sklearn.linear_model import LogisticRegression\nfrom sklearn.neighbors import NearestNeighbors\nfrom sklearn.tree import DecisionTreeClassifier\n\nfrom sklearn.metrics import precision_recall_curve\nfrom sklearn.metrics import f1_score\nfrom sklearn.metrics import accuracy_score\nfrom sklearn.metrics import average_precision_score\nfrom sklearn.metrics import confusion_matrix\nfrom sklearn.metrics import balanced_accuracy_score\n\nimport collections\nfrom collections import Counter\nfrom imblearn.over_sampling import SMOTE\n\n\n%matplotlib inline\nimport matplotlib\nimport matplotlib.pyplot as plt\nfrom matplotlib.colors import ListedColormap\n\nfrom IPython import display\n\nfrom scipydirect import minimize # for DIvided RECTangles (DIRECT) method\nimport time", "Collecting scipydirect\n\u001b[?25l Downloading https://files.pythonhosted.org/packages/c2/dd/657e6c53838b3ff50e50bda4e905c8ec7e4b715f966f33d0566088391d75/scipydirect-1.3.tar.gz (49kB)\n\r\u001b[K |██████▋ | 10kB 15.9MB/s eta 0:00:01\r\u001b[K |█████████████▏ | 20kB 21.2MB/s eta 0:00:01\r\u001b[K |███████████████████▊ | 30kB 24.8MB/s eta 0:00:01\r\u001b[K |██████████████████████████▍ | 40kB 24.2MB/s eta 0:00:01\r\u001b[K |████████████████████████████████| 51kB 4.7MB/s \n\u001b[?25hRequirement already satisfied: numpy in /usr/local/lib/python3.7/dist-packages (from scipydirect) (1.19.5)\nBuilding wheels for collected packages: scipydirect\n Building wheel for scipydirect (setup.py) ... \u001b[?25l\u001b[?25hdone\n Created wheel for scipydirect: filename=scipydirect-1.3-cp37-cp37m-linux_x86_64.whl size=119893 sha256=201aad9d5036ecc1328b332fbd9c6adc620fa9b0d03010fc0ccea07c5bea8a47\n Stored in directory: /root/.cache/pip/wheels/b7/de/2b/c5550296e6dac87737c622568c6a10bc87ae76e98b52ff210e\nSuccessfully built scipydirect\nInstalling collected packages: scipydirect\nSuccessfully installed scipydirect-1.3\n" ] ], [ [ "# 1) Describe classes of the following models: AdaFair, SMOTEBoost, ASR", "_____no_output_____" ], [ "#### 1.1) AdaFair", "_____no_output_____" ] ], [ [ "#AdaFair\nclass AdaFairClassifier(AdaBoostClassifier):\n def __init__(self,\n base_estimator=None, *,\n n_estimators=50,\n learning_rate=1,\n algorithm='SAMME',\n random_state=42,\n protected=None,\n epsilon = 0):\n\n super().__init__(\n base_estimator=base_estimator,\n n_estimators=n_estimators,\n algorithm = algorithm,\n learning_rate=learning_rate,\n random_state=random_state)\n\n self.protected = np.array(protected)\n self.algorithm = algorithm\n self.epsilon = epsilon\n \n def _boost_discrete(self, iboost, X, y, sample_weight, random_state):\n \"\"\"Implement a single boost using the SAMME discrete algorithm.\"\"\"\n estimator = self._make_estimator(random_state=random_state)\n\n estimator.fit(X, y, sample_weight=sample_weight)\n\n y_predict = estimator.predict(X)\n\n if iboost == 0:\n self.classes_ = getattr(estimator, 'classes_', None)\n self.n_classes_ = len(self.classes_)\n\n # Instances incorrectly classified\n incorrect = y_predict != y\n\n # Error fraction\n estimator_error = np.mean(\n np.average(incorrect, weights=sample_weight, axis=0))\n\n # Stop if classification is perfect\n if estimator_error <= 0:\n return sample_weight, 1., 0.\n\n n_classes = self.n_classes_\n\n # Stop if the error is at least as bad as random guessing\n if estimator_error >= 1. - (1. / n_classes):\n self.estimators_.pop(-1)\n if len(self.estimators_) == 0:\n print (\"BaseClassifier in AdaBoostClassifier ensemble is worse than random, ensemble can not be fit.\")\n raise ValueError('BaseClassifier in AdaBoostClassifier '\n 'ensemble is worse than random, ensemble '\n 'can not be fit.')\n return None, None, None\n\n if len(self.protected) != len(y):\n print (\"Error: not given or given incorrect list of protected objects\")\n return None, None, None\n\n #Compute fairness-related costs\n #CUMULATIVE prediction \n \n y_cumulative_pred = list(self.staged_predict(X))[0]\n\n u = np.array(self.get_fairness_related_costs(y, y_cumulative_pred, self.protected))\n\n # Boost weight using multi-class AdaBoost SAMME alg\n estimator_weight = self.learning_rate * (\n np.log((1. - estimator_error) / estimator_error) +\n np.log(n_classes - 1.))\n\n # Only boost the weights if I will fit again\n if not iboost == self.n_estimators - 1:\n # Only boost positive weights\n\n sample_weight = sample_weight * np.exp(estimator_weight * incorrect * (sample_weight > 0)) * (np.ones(len(u)) + u)\n\n return sample_weight, estimator_weight, estimator_error\n\n def get_fairness_related_costs(self, y, y_pred_t, protected):\n\n y_true_protected, y_pred_protected, y_true_non_protected, y_pred_non_protected = separate_protected_from_non_protected(y, y_pred_t, protected)\n #Rates for non protected group\n tp, tn, fp, fn = tp_tn_fp_fn(y_true_non_protected, y_pred_non_protected)\n FPR_non_protected = fp / (fp + tn)\n FNR_non_protected = fn / (fn + tp)\n #Rates for protected group\n tp, tn, fp, fn = tp_tn_fp_fn(y_true_protected, y_pred_protected)\n FPR_protected = fp / (fp + tn)\n FNR_protected = fn / (fn + tp)\n \n delta_FPR = - FPR_non_protected + FPR_protected \n delta_FNR = - FNR_non_protected + FNR_protected\n\n self.delta_FPR = delta_FPR\n self.delta_FNR = delta_FNR\n #Compute fairness related costs\n u = []\n for y_i, y_pred_t__i, protection in zip(y, y_pred_t, protected):\n if y_i == 1 and y_pred_t__i == 0 and abs(delta_FNR) > self.epsilon and protection == 1 and delta_FNR > 0:\n u.append(abs(delta_FNR))\n elif y_i == 1 and y_pred_t__i == 0 and abs(delta_FNR) > self.epsilon and protection == 0 and delta_FNR < 0:\n u.append(abs(delta_FNR))\n elif y_i == 0 and y_pred_t__i == 1 and abs(delta_FPR) > self.epsilon and protection == 1 and delta_FPR > 0:\n u.append(abs(delta_FPR))\n elif y_i == 0 and y_pred_t__i == 1 and abs(delta_FPR) > self.epsilon and protection == 0 and delta_FPR < 0:\n u.append(abs(delta_FPR))\n else: u.append(0)\n return u", "_____no_output_____" ] ], [ [ "#### 1.2 SMOTEBoost", "_____no_output_____" ] ], [ [ "#SMOTEBoost\n\nrandom_state = 42\n\n#4, y - true label\ndef ada_boost_eps(y, y_pred_t, distribution):\n eps = np.sum((1 - (y == y_pred_t) + (np.logical_not(y) == y_pred_t)) * distribution)\n return eps\n\n#5\ndef ada_boost_betta(eps):\n betta = eps/(1 - eps)\n return betta\n\ndef ada_boost_w(y, y_pred_t):\n w = 0.5 * (1 + (y == y_pred_t) - (np.logical_not(y) == y_pred_t))\n return w\n\n#6\ndef ada_boost_distribution(distribution, betta, w):\n distribution = distribution * betta ** w / np.sum(distribution)\n return distribution\n\ndef min_target(y):\n minority_target = min(Counter(y), key=Counter(y).get)\n return minority_target\n\nclass SMOTEBoost():\n \n def __init__(self,\n n_samples = 100,\n k_neighbors = 5, \n n_estimators = 50, #n_estimators = T\n base_classifier = None,\n random_state = 42,\n get_eps = ada_boost_eps,\n get_betta = ada_boost_betta,\n get_w = ada_boost_w,\n update_distribution=ada_boost_distribution):\n self.n_samples = n_samples\n self.k_neighbors = k_neighbors\n self.n_estimators = n_estimators\n self.base_classifier = base_classifier\n self.random_state = random_state\n self.get_eps = get_eps\n self.get_betta = get_betta\n self.get_w = get_w\n self.update_distribution = update_distribution\n \n def fit(self, X, y):\n X = np.array(X)\n distribution = np.ones(X.shape[0], dtype=float) / X.shape[0]\n self.classifiers = []\n self.betta = []\n y = np.array(y)\n for i in range(self.n_estimators):\n minority_class = min_target(y)\n X_min = X[np.where(y == minority_class)]\n\n # create a new classifier\n self.classifiers.append(self.base_classifier())\n\n # SMOTE\n self.smote = SMOTE(n_samples=self.n_samples,\n k_neighbors=self.k_neighbors,\n random_state=self.random_state)\n self.smote.fit(X_min)\n X_syn = self.smote.sample()\n y_syn = np.full(X_syn.shape[0], fill_value=minority_class, dtype=np.int64)\n\n # Modify distribution\n distribution_syn = np.empty(X_syn.shape[0], dtype=np.float64)\n distribution_syn[:] = 1. / X.shape[0]\n mod_distribution = np.append(distribution, distribution_syn).reshape(1, -1)\n mod_distribution = np.squeeze(normalize(mod_distribution, axis=1, norm='l1'))\n # Concatenate original and synthetic datasets for training a weak learner\n mod_X = np.vstack((X, X_syn))\n mod_y = np.append(y, y_syn) \n\n # Train a weak lerner\n self.classifiers[-1].fit(mod_X, mod_y, sample_weight=mod_distribution)\n\n # Make a prediction for the original dataset\n y_pred_t = self.classifiers[-1].predict(X)\n \n # Compute the pseudo-loss of hypothesis\n eps_t = ada_boost_eps(y, y_pred_t, distribution)\n betta_t = ada_boost_betta(eps_t)\n w_t = ada_boost_w(y, y_pred_t)\n self.betta.append(betta_t)\n \n # Update distribution and normalize\n distribution = ada_boost_distribution(distribution, betta_t, w_t)\n \n def predict(self, X):\n final_predictions_0 = np.zeros(X.shape[0])\n final_predictions_1 = np.zeros(X.shape[0])\n y_pred = np.empty(X.shape[0])\n # get the weighted votes of the classifiers\n for i in range(len(self.betta)):\n h_i = self.classifiers[i].predict(X)\n final_predictions_0 = final_predictions_0 + math.log(1/self.betta[i])*(h_i == 0)\n final_predictions_1 = final_predictions_1 + math.log(1/self.betta[i])*(h_i == 1)\n for i in range(len(final_predictions_0)):\n if final_predictions_0[i] > final_predictions_1[i]:\n y_pred[i] = 0\n else:\n y_pred[i] = 1\n return y_pred\n\n\nclass SMOTE():\n\n def __init__(self, n_samples, k_neighbors=5, random_state=None):\n self.n_samples = n_samples\n self.k = k_neighbors\n self.random_state = random_state\n\n def fit(self, X):\n self.X = X\n self.n_features = self.X.shape[1]\n self.neigh = NearestNeighbors(n_neighbors=self.k) #k + 1\n self.neigh.fit(self.X)\n return self\n\n def sample(self):\n np.random.seed(seed=self.random_state)\n S = np.zeros(shape=(self.n_samples, self.n_features))\n\n for i in range(self.n_samples):\n j = np.random.randint(0, self.X.shape[0])\n nn = self.neigh.kneighbors(self.X[j].reshape(1, -1),\n return_distance=False)[:, 1:]\n nn_index = np.random.choice(nn[0])\n print (self.X[nn_index], self.X[j])\n dif = self.X[nn_index] - self.X[j]\n gap = np.random.random() \n S[i, :] = self.X[j, :] + gap * dif[:]\n return S", "_____no_output_____" ] ], [ [ "#### 1.3 Adaptive sensitive reweighting", "_____no_output_____" ] ], [ [ "#Adaptive sensitive reweighting\nclass ReweightedClassifier:\n def __init__(self, baze_clf, alpha, beta, params = {}):\n \"\"\"\n Input: \n baze_clf - object from sklearn with methods .fit(sample_weight=), .predict(), .predict_proba()\n alpha - list of alphas for sensitive and non-sensitive objects [alpha, alpha']\n beta - list of betss for sensitive and non-sensitive objects [beta, beta']\n params - **kwargs compatible with baze_clf\n \n \"\"\"\n self.baze_clf = baze_clf \n self.model = None\n self.alpha = np.array(alpha)\n self.alphas = None\n self.beta = np.array(beta)\n self.betas = None\n self.weights = None\n self.prev_weights = None\n self.params = params\n \n def reweight_dataset(self, length, error, minority_idx):\n \"\"\"\n This function recalculates values of weights and saves their previous values\n \"\"\"\n if self.alphas is None or self.betas is None:\n # If alpha_0, alpha_1 or beta_0, beta_1 are not defined,\n # then define alpha_0 and beta_0 to every object from non-sensitive class,\n # and alpha_1 and beta_1 to every object from sensitive class (minority).\n self.alphas = np.ones(length) * self.alpha[0]\n self.betas = np.ones(length) * self.beta[0]\n self.alphas[minority_idx] = self.alpha[1]\n self.betas[minority_idx] = self.beta[1]\n \n # w_i_prev <- w_i for all i in dataset\n self.prev_weights = self.weights.copy()\n \n # w_i = alpha_i * L_{beta_i} (P'(y_pred_i =! y_true_i)) \n # + (1 - alpha_i) * L_{beta_i} (-P'(y_pred_i =! y_true_i)),\n # where \n # L_{beta_i} (x) = exp(beta_i * x)\n self.weights = self.alphas * np.exp(self.betas * error) \\\n + (1 - self.alphas) * np.exp(- self.betas * error)\n \n def pRule(self, prediction, minority_idx):\n \"\"\"\n This function calculates \n | P(y_pred_i = 1 | i in S) P(y_pred_i = 1 | i not in S) |\n pRule = min { ---------------------------- , ---------------------------- }\n | P(y_pred_i = 1 | i not in S) P(y_pred_i = 1 | i in S) |\n \n S - the group of sensitive objects\n ---------\n Input:\n prediction - labels ({0,1}) of a sample for which pRule is calculated\n minority_idx - indexes of objects from a sensitive group\n \"\"\"\n # majority indexes = set of all indexes / set of minority indexes,\n # where set of all indexes = all numbers from 0 to size of sample (=len(prediction))\n majority_idx = set(np.linspace(0, len(prediction) - 1, len(prediction), dtype = int)).difference(minority_idx)\n \n # minority = P(y_pred_i = 1 | i in minority)\n # majority = P(y_pred_i = 1 | i in majority)\n minority = prediction[minority_idx].mean()\n majority = prediction[list(majority_idx)].mean()\n \n minority = np.clip(minority, 1e-10, 1 - 1e-10)\n majority = np.clip(majority, 1e-10, 1 - 1e-10)\n \n return min(minority/majority, majority/minority) \n \n def fit(self, X_train, y_train, X_test, y_test, minority_idx, verbose=True, max_iter=30):\n # Initialize equal weights w_i = 1\n self.weights = np.ones(len(y_train))\n self.prev_weights = np.zeros(len(y_train))\n \n # Lists for saving metrics\n accuracys = []\n pRules = []\n differences = []\n accuracy_plus_prule = []\n \n # Adaptive Sensitive Reweighting\n iteration = 0\n \n while ((self.prev_weights - self.weights) ** 2).mean() > 10**(-6) and iteration < max_iter:\n \n iteration += 1\n # Train classifier on X_train with weights w_i\n self.model = self.baze_clf(**self.params)\n self.model.fit(X_train, y_train, sample_weight = self.weights)\n \n # Use classifier to obtain P`(y_pred_i =! y_pred) (here it is called 'error')\n prediction_proba = self.model.predict_proba(X_train)[:, 1]\n error = (y_train == 1) * (1 - prediction_proba) + (y_train == 0) * prediction_proba \n \n # Update weights\n self.reweight_dataset(len(y_train), error, minority_idx)\n \n # Get metrics on X_train\n prediction = self.model.predict(X_train)\n accuracys.append(accuracy_score(prediction, y_train))\n pRules.append(self.pRule(prediction, minority_idx))\n accuracy_plus_prule.append(accuracys[-1] + pRules[-1])\n differences.append(((self.prev_weights - self.weights)**2).mean()**0.5)\n \n # Visualize metrics if it's needed\n if verbose:\n display.clear_output(True)\n fig, axes = plt.subplots(ncols=2, nrows=2, figsize=(16, 7))\n\n metrics = [accuracys, pRules, accuracy_plus_prule, differences]\n metrics_names = [\"Accuracy score\", \"pRule\", \"Accuracy + pRule\", \"Mean of weight edits\"]\n for name, metric, ax in zip(metrics_names, metrics, axes.flat):\n ax.plot(metric, label='train')\n ax.set_title(f'{name}, iteration {iteration}')\n ax.legend()\n if name == \"Mean of weight edits\":\n ax.set_yscale('log')\n plt.show()\n \n return accuracys, pRules, accuracy_plus_prule\n \n \n def predict(self, X):\n return self.model.predict(X)\n \n def predict_proba(self, X):\n return self.model.predict_proba(X)\n \n def get_metrics_test(self, X_test, y_test, minority_idx_test):\n \"\"\"\n Obtain pRule and accuracy for trained model\n \"\"\"\n # Obtain predictions on X_test to calculate metrics\n prediction_test = self.model.predict(X_test)\n\n # Get metrics on test\n accuracy_test = accuracy_score(prediction_test, y_test)\n pRule_test = self.pRule(prediction_test, minority_idx_test)\n\n return accuracy_test, pRule_test \n\ndef prep_train_model(X_train, y_train, X_test, y_test, minority_idx):\n \n def train_model(a):\n \"\"\"\n Function of 4 variables (a[0], a[1], a[2], a[3]) that will be minimized by DIRECT method.\n a[0], a[1] = alpha, alpha'\n a[2], a[3] = beta, beta'\n \"\"\"\n model = ReweightedClassifier(LogisticRegression, [a[0], a[1]], [a[2], a[3]], params = {\"max_iter\": 4000})\n _, _, accuracy_plus_prule = model.fit(X_train, y_train, X_test, y_test, minority_idx)\n\n # We'll maximize [acc + pRule] which we get at the last iteration of Adaptive Sensitive Reweighting\n return - accuracy_plus_prule[-1]\n \n return train_model # return function for optimization", "_____no_output_____" ] ], [ [ "#### 1.4 Some functions used for fitting models, calculating metrics, and data separation", "_____no_output_____" ] ], [ [ "#This function returns binary list of whether the corresponding feature is protected (1) or not (0)\ndef get_protected_instances(X, feature, label):\n protected = []\n for i in range(len(X)):\n if X.iloc[i][feature] == label:\n protected.append(1)\n else: protected.append(0)\n return protected\n\n#To calculate TRP and TNR for protected and non-protected groups, first separate them\ndef separate_protected_from_non_protected(y_true, y_pred, protected):\n y_true_protected = []\n y_pred_protected = []\n y_true_non_protected = []\n y_pred_non_protected = []\n for true_label, pred_label, is_protected in zip(y_true, y_pred, protected):\n if is_protected == 1:\n y_true_protected.append(true_label)\n y_pred_protected.append(pred_label)\n elif is_protected == 0:\n y_true_non_protected.append(true_label)\n y_pred_non_protected.append(pred_label)\n else:\n print(\"Error: invalid value of in protected array \", is_protected)\n return 0,0,0,0\n return (np.array(y_true_protected), np.array(y_pred_protected), np.array(y_true_non_protected), np.array(y_pred_non_protected))\n\ndef tp_tn_fp_fn(y_true, y_pred):\n matrix = confusion_matrix(y_true, y_pred)\n tp = matrix[1][1]\n tn = matrix[0][0]\n fp = matrix[0][1]\n fn = matrix[1][0]\n return (tp, tn, fp, fn)\n\n#same pRule as in ASR, but used for calculating this metric in others classifiers' predictions\ndef pRule(prediction, minority_idx):\n \"\"\"\n This function calculates \n | P(y_pred_i = 1 | i in S) P(y_pred_i = 1 | i not in S) |\n pRule = min { ---------------------------- , ---------------------------- }\n | P(y_pred_i = 1 | i not in S) P(y_pred_i = 1 | i in S) |\n \n S - the group of sensitive objects\n ---------\n Input:\n prediction - labels ({0,1}) of a sample for which pRule is calculated\n minority_idx - indexes of objects from a sensitive group\n \"\"\"\n # majority indexes = set of all indexes / set of minority indexes,\n # where set of all indexes = all numbers from 0 to size of sample (=len(prediction))\n majority_idx = set(np.linspace(0, len(prediction) - 1, len(prediction), dtype = int)).difference(minority_idx)\n \n # minority = P(y_pred_i = 1 | i in minority)\n # majority = P(y_pred_i = 1 | i in majority)\n minority = prediction[minority_idx].mean()\n majority = prediction[list(majority_idx)].mean()\n \n minority = np.clip(minority, 1e-10, 1 - 1e-10)\n majority = np.clip(majority, 1e-10, 1 - 1e-10)\n \n return min(minority/majority, majority/minority) ", "_____no_output_____" ] ], [ [ "# 2) Download datasets (run one cell for one dataset)", "_____no_output_____" ], [ "Adult census", "_____no_output_____" ] ], [ [ "#Adult census\n#adult_census_names = ['old_id' ,'age','workclass','fnlwgt','education','education_num','marital_status','occupation','relationship','race','sex','capital_gain','capital_loss','hours_per_week','native_country']\nX_train = pd.read_csv(\"splits/X_train_preprocessed_adult.csv\").drop(\"Unnamed: 0\", axis = 1)#, names = adult_census_names).iloc[1:]\nX_test = pd.read_csv(\"splits/X_test_preprocessed_adult.csv\").drop(\"Unnamed: 0\", axis = 1)#, names = adult_census_names).iloc[1:]\n\ny_train = pd.read_csv(\"splits/y_train_preprocessed_adult.csv\")['income']\ny_test = pd.read_csv(\"splits/y_test_preprocessed_adult.csv\")['income']\nreweight_prediction = \"/content/predictions/y_pred_test_adult.csv\"\n\n#X_test, X_train = preprocess_adult_census(X_test.drop('old_id', axis = 1)), preprocess_adult_census(X_train.drop('old_id', axis = 1))\n#y_test, y_train = adult_label_transform(y_test)['income'], adult_label_transform(y_train)['income']\n#Obtain protected group (used in AdaFair)\nprotected_test = get_protected_instances(X_test, 'gender', 1)\nprotected_train = get_protected_instances(X_train, 'gender', 1)\n\n# Obtain indexes of sensitive class (for regression algorithm)\nminority_idx = X_train.reset_index(drop=True).index.values[X_train[\"gender\"] == 1]\nminority_idx_test = X_test.reset_index(drop=True).index.values[X_test[\"gender\"] == 1]\n\n#best hyperparameters for AdaFair\nadafair_max_depth = 2\nadafair_n_estimators = 20\n\n#result of ASR optimizing\na_1 = [0.01851852, 0.99382716, 1.16666667, 2.94444444]\n#print(len(X_train),len(y_train))\n#X_train.head()", "_____no_output_____" ] ], [ [ "Bank", "_____no_output_____" ] ], [ [ "X_train = pd.read_csv(\"splits/X_train_preprocessed_bank.csv\").drop(\"Unnamed: 0\", axis = 1)#, names = adult_census_names).iloc[1:]\nX_test = pd.read_csv(\"splits/X_test_preprocessed_bank.csv\").drop(\"Unnamed: 0\", axis = 1)#, names = adult_census_names).iloc[1:]\n\ny_train = pd.read_csv(\"splits/y_train_preprocessed_bank.csv\")['y']\ny_test = pd.read_csv(\"splits/y_test_preprocessed_bank.csv\")['y']\nreweight_prediction = \"/content/predictions/y_pred_test_bank.csv\"\n\n#X_test, X_train = preprocess_adult_census(X_test.drop('old_id', axis = 1)), preprocess_adult_census(X_train.drop('old_id', axis = 1))\n#y_test, y_train = adult_label_transform(y_test)['income'], adult_label_transform(y_train)['income']\n#Obtain protected group (used in AdaFair)\nprotected_test = get_protected_instances(X_test, 'age', 1)\nprotected_train = get_protected_instances(X_train, 'age', 1)\n\n# Obtain indexes of sensitive class (for regression algorithm)\nminority_idx = X_train.reset_index(drop=True).index.values[X_train[\"age\"] == 1]\nminority_idx_test = X_test.reset_index(drop=True).index.values[X_test[\"age\"] == 1]\n\n#best hyperparameters for AdaFair\nadafair_max_depth = 2\nadafair_n_estimators = 9\n\n#result of ASR optimizing\na_1 = [0.87037037, 0.01851852, 2.72222222, 1.57407407] \n#print(len(X_train),len(y_train))\n#X_train.head()", "_____no_output_____" ] ], [ [ "Compass", "_____no_output_____" ] ], [ [ "X_train = pd.read_csv(\"splits/X_train_preprocessed_compas.csv\").drop(\"Unnamed: 0\", axis = 1)#, names = adult_census_names).iloc[1:]\nX_test = pd.read_csv(\"splits/X_test_preprocessed_compas.csv\").drop(\"Unnamed: 0\", axis = 1)#, names = adult_census_names).iloc[1:]\n\ny_train = pd.read_csv(\"splits/y_train_preprocessed_compas.csv\")['two_year_recid']\ny_test = pd.read_csv(\"splits/y_test_preprocessed_compas.csv\")['two_year_recid']\nreweight_prediction = \"/content/predictions/y_pred_test_compas.csv\"\n\n#X_test, X_train = preprocess_adult_census(X_test.drop('old_id', axis = 1)), preprocess_adult_census(X_train.drop('old_id', axis = 1))\n#y_test, y_train = adult_label_transform(y_test)['income'], adult_label_transform(y_train)['income']\n#Obtain protected group (used in AdaFair)\nprotected_test = get_protected_instances(X_test, 'race', 0)\nprotected_train = get_protected_instances(X_train, 'race', 0)\n\n# Obtain indexes of sensitive class (for regression algorithm)\nminority_idx = X_train.reset_index(drop=True).index.values[X_train[\"race\"] == 0]\nminority_idx_test = X_test.reset_index(drop=True).index.values[X_test[\"race\"] == 0]\n\n#best hyperparameters for AdaFair\nadafair_max_depth = 4\nadafair_n_estimators = 5\n\n#result of ASR optimizing\na_1 = [0.0308642, 0.72222222, 0.5, 0.45061728]\n#print(len(X_train),len(y_train))\n#y_train.head()", "_____no_output_____" ] ], [ [ "KDD Census", "_____no_output_____" ] ], [ [ "#adult_census_names = ['old_id' ,'age','workclass','fnlwgt','education','education_num','marital_status','occupation','relationship','race','sex','capital_gain','capital_loss','hours_per_week','native_country']\nX_train = pd.read_csv(\"splits/X_train_preprocessed_kdd.csv\").drop(\"Unnamed: 0\", axis = 1)#, names = adult_census_names).iloc[1:]\nX_test = pd.read_csv(\"splits/X_test_preprocessed_kdd.csv\").drop(\"Unnamed: 0\", axis = 1)#, names = adult_census_names).iloc[1:]\n\ny_train = pd.read_csv(\"splits/y_train_preprocessed_kdd.csv\")['income_50k']\ny_test = pd.read_csv(\"splits/y_test_preprocessed_kdd.csv\")['income_50k']\nreweight_prediction = \"/content/predictions/y_pred_test_kdd.csv\"\n\n#X_test, X_train = preprocess_adult_census(X_test.drop('old_id', axis = 1)), preprocess_adult_census(X_train.drop('old_id', axis = 1))\n#y_test, y_train = adult_label_transform(y_test)['income'], adult_label_transform(y_train)['income']\n#Obtain protected group (used in AdaFair)\nprotected_test = get_protected_instances(X_test, 'sex', 1)\nprotected_train = get_protected_instances(X_train, 'sex', 1)\n\n# Obtain indexes of sensitive class (for regression algorithm)\nminority_idx = X_train.reset_index(drop=True).index.values[X_train[\"sex\"] == 1]\nminority_idx_test = X_test.reset_index(drop=True).index.values[X_test[\"sex\"] == 1]\n\n#best hyperparameters for AdaFair\nadafair_max_depth = 5\nadafair_n_estimators = 11 \n\n#result of ASR optimizing\na_1 = [0.01851852, 0.99382716, 1.16666667, 2.94444444]\n#print(len(X_train),len(y_train))\n#X_train.head()", "98147 98147\n" ] ], [ [ "# 3) Create models, train classifiers", "_____no_output_____" ] ], [ [ "#Regression\n# Create model with obtained hyperparameters alpha, alpha', beta, beta'\nmodel_reweighted_classifier = ReweightedClassifier(LogisticRegression, [a_1[0], a_1[1]], [a_1[2], a_1[3]], params = {\"max_iter\": 4})\n\n# Train model on X_train \nmodel_reweighted_classifier.fit(X_train, y_train, X_test, y_test, minority_idx, verbose=False)\n\n# Calculate metrics (pRule, accuracy) on X_test\naccuracy_test, pRule_test = model_reweighted_classifier.get_metrics_test(X_test, y_test, minority_idx_test)\n\n\n#print('ASR+CULEP for X_test')\n#print(f\"prule = {pRule_test:.6}, accuracy = {accuracy_test:.6}\")\n#print(f\"prule + accuracy = {(pRule_test + accuracy_test):.6}\")", "/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\nSTOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n\nIncrease the number of iterations (max_iter) or scale the data as shown in:\n https://scikit-learn.org/stable/modules/preprocessing.html\nPlease also refer to the documentation for alternative solver options:\n https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\nSTOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n\nIncrease the number of iterations (max_iter) or scale the data as shown in:\n https://scikit-learn.org/stable/modules/preprocessing.html\nPlease also refer to the documentation for alternative solver options:\n https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\nSTOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n\nIncrease the number of iterations (max_iter) or scale the data as shown in:\n https://scikit-learn.org/stable/modules/preprocessing.html\nPlease also refer to the documentation for alternative solver options:\n https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\nSTOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n\nIncrease the number of iterations (max_iter) or scale the data as shown in:\n https://scikit-learn.org/stable/modules/preprocessing.html\nPlease also refer to the documentation for alternative solver options:\n https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\nSTOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n\nIncrease the number of iterations (max_iter) or scale the data as shown in:\n https://scikit-learn.org/stable/modules/preprocessing.html\nPlease also refer to the documentation for alternative solver options:\n https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\nSTOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n\nIncrease the number of iterations (max_iter) or scale the data as shown in:\n https://scikit-learn.org/stable/modules/preprocessing.html\nPlease also refer to the documentation for alternative solver options:\n https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\nSTOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n\nIncrease the number of iterations (max_iter) or scale the data as shown in:\n https://scikit-learn.org/stable/modules/preprocessing.html\nPlease also refer to the documentation for alternative solver options:\n https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\nSTOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n\nIncrease the number of iterations (max_iter) or scale the data as shown in:\n https://scikit-learn.org/stable/modules/preprocessing.html\nPlease also refer to the documentation for alternative solver options:\n https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\nSTOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n\nIncrease the number of iterations (max_iter) or scale the data as shown in:\n https://scikit-learn.org/stable/modules/preprocessing.html\nPlease also refer to the documentation for alternative solver options:\n https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\nSTOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n\nIncrease the number of iterations (max_iter) or scale the data as shown in:\n https://scikit-learn.org/stable/modules/preprocessing.html\nPlease also refer to the documentation for alternative solver options:\n https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\nSTOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n\nIncrease the number of iterations (max_iter) or scale the data as shown in:\n https://scikit-learn.org/stable/modules/preprocessing.html\nPlease also refer to the documentation for alternative solver options:\n https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\nSTOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n\nIncrease the number of iterations (max_iter) or scale the data as shown in:\n https://scikit-learn.org/stable/modules/preprocessing.html\nPlease also refer to the documentation for alternative solver options:\n https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\nSTOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n\nIncrease the number of iterations (max_iter) or scale the data as shown in:\n https://scikit-learn.org/stable/modules/preprocessing.html\nPlease also refer to the documentation for alternative solver options:\n https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\nSTOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n\nIncrease the number of iterations (max_iter) or scale the data as shown in:\n https://scikit-learn.org/stable/modules/preprocessing.html\nPlease also refer to the documentation for alternative solver options:\n https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\nSTOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n\nIncrease the number of iterations (max_iter) or scale the data as shown in:\n https://scikit-learn.org/stable/modules/preprocessing.html\nPlease also refer to the documentation for alternative solver options:\n https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\nSTOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n\nIncrease the number of iterations (max_iter) or scale the data as shown in:\n https://scikit-learn.org/stable/modules/preprocessing.html\nPlease also refer to the documentation for alternative solver options:\n https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\nSTOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n\nIncrease the number of iterations (max_iter) or scale the data as shown in:\n https://scikit-learn.org/stable/modules/preprocessing.html\nPlease also refer to the documentation for alternative solver options:\n https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\nSTOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n\nIncrease the number of iterations (max_iter) or scale the data as shown in:\n https://scikit-learn.org/stable/modules/preprocessing.html\nPlease also refer to the documentation for alternative solver options:\n https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\nSTOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n\nIncrease the number of iterations (max_iter) or scale the data as shown in:\n https://scikit-learn.org/stable/modules/preprocessing.html\nPlease also refer to the documentation for alternative solver options:\n https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\nSTOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n\nIncrease the number of iterations (max_iter) or scale the data as shown in:\n https://scikit-learn.org/stable/modules/preprocessing.html\nPlease also refer to the documentation for alternative solver options:\n https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\nSTOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n\nIncrease the number of iterations (max_iter) or scale the data as shown in:\n https://scikit-learn.org/stable/modules/preprocessing.html\nPlease also refer to the documentation for alternative solver options:\n https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\nSTOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n\nIncrease the number of iterations (max_iter) or scale the data as shown in:\n https://scikit-learn.org/stable/modules/preprocessing.html\nPlease also refer to the documentation for alternative solver options:\n https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\nSTOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n\nIncrease the number of iterations (max_iter) or scale the data as shown in:\n https://scikit-learn.org/stable/modules/preprocessing.html\nPlease also refer to the documentation for alternative solver options:\n https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\nSTOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n\nIncrease the number of iterations (max_iter) or scale the data as shown in:\n https://scikit-learn.org/stable/modules/preprocessing.html\nPlease also refer to the documentation for alternative solver options:\n https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\nSTOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n\nIncrease the number of iterations (max_iter) or scale the data as shown in:\n https://scikit-learn.org/stable/modules/preprocessing.html\nPlease also refer to the documentation for alternative solver options:\n https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\nSTOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n\nIncrease the number of iterations (max_iter) or scale the data as shown in:\n https://scikit-learn.org/stable/modules/preprocessing.html\nPlease also refer to the documentation for alternative solver options:\n https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\nSTOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n\nIncrease the number of iterations (max_iter) or scale the data as shown in:\n https://scikit-learn.org/stable/modules/preprocessing.html\nPlease also refer to the documentation for alternative solver options:\n https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\nSTOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n\nIncrease the number of iterations (max_iter) or scale the data as shown in:\n https://scikit-learn.org/stable/modules/preprocessing.html\nPlease also refer to the documentation for alternative solver options:\n https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\nSTOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n\nIncrease the number of iterations (max_iter) or scale the data as shown in:\n https://scikit-learn.org/stable/modules/preprocessing.html\nPlease also refer to the documentation for alternative solver options:\n https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\nSTOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n\nIncrease the number of iterations (max_iter) or scale the data as shown in:\n https://scikit-learn.org/stable/modules/preprocessing.html\nPlease also refer to the documentation for alternative solver options:\n https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n" ], [ "#SMOTEBOOST\nmax_depth = 2\nn_samples = 100\nk_neighbors = 5\nn_estimators = 5 # T\nrandom_state = 42\n\nget_base_clf = lambda: DecisionTreeClassifier(max_depth = max_depth)\nsmoteboost1 = SMOTEBoost(n_samples = n_samples,\n k_neighbors = k_neighbors,\n n_estimators = n_estimators,\n base_classifier = get_base_clf,\n random_state = random_state)\nsmoteboost1.fit(X_train, y_train)\n#smote_boost = smoteboost1.predict(X_test)", "\u001b[1;30;43mВыходные данные были обрезаны до нескольких последних строк (5000).\u001b[0m\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. ]\n[1.34556873 2.75584724 0.17405731 0. 0. 0.\n 0. 0.21544598 2.0119666 0. 1.30023581 2.20770341\n 0.99471607 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [1.54597258 2.75584724 0.87028654 1. 0. 0.\n 0. 0.59709709 1.50897495 0. 1.30023581 1.69823339\n 0.99471607 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[1.37419785 2.24313148 1.47948712 1. 0. 0.\n 0. 0.66401166 1.50897495 0. 1.30023581 2.20770341\n 0.99471607 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [1.40282698 2.179042 1.47948712 1. 0. 0.\n 0. 1.5090856 1.50897495 0. 1.30023581 2.20770341\n 0.99471607 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[1.173794 1.98677359 2.26274501 1. 0. 0.\n 0. 0.79234537 3.01794989 0. 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [1.34556873 1.98677359 2.26274501 0. 0. 0.\n 0. 0.50076076 3.01794989 0. 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[0.94476102 2.75584724 0.78325789 0. 0. 0.\n 0.24910027 0.84522308 3.01794989 0. 1.30023581 1.65577755\n 0.99471607 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. ] [1.34556873 2.75584724 0.78325789 0. 0. 0.\n 0. 0.14819297 1.0059833 0. 1.30023581 2.20770341\n 0.99471607 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. ]\n[1.25968137e+00 1.40996836e+00 1.74057309e-01 0.00000000e+00\n 2.26308624e+02 0.00000000e+00 0.00000000e+00 1.17111109e+00\n 2.01196660e+00 0.00000000e+00 1.30023581e+00 2.20770341e+00\n 1.00529816e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 1.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 1.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 1.00000000e+00 0.00000000e+00\n 1.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 1.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 1.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 1.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 1.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 1.00000000e+00\n 0.00000000e+00 0.00000000e+00 1.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 1.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 1.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 1.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 1.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 1.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 1.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 1.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 1.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 1.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 1.00000000e+00\n 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00\n 0.00000000e+00 1.00000000e+00 1.00000000e+00 0.00000000e+00] [ 1.23105224 1.5381473 0.43514327 0. 226.30862398\n 0. 0. 0.63024748 1.0059833 0.\n 1.30023581 2.20770341 1.00529816 0. 0.\n 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 1. 0. 0. 1. 0.\n 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 1.\n 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0.\n 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0.\n 0. 0. 1. 1. 0. ]\n[1.48871434 2.69175777 0.17405731 1. 0. 0.\n 0. 0.6275055 3.01794989 0. 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [1.68911819 2.81993671 0.17405731 1. 0. 0.\n 0.12455013 0.88471214 3.01794989 0. 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[ 2.03266766 0. 0. 0. 0. 0.\n 24.91002675 1.76102626 0. 0. 1.30023581 0.\n 1.00529816 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [ 2.29032976 0. 0. 0. 0. 0.\n 27.1868032 1.41254849 0. 0. 1.30023581 0.\n 1.00529816 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[1.4314561 2.11495253 0.17405731 1. 0. 0.\n 0. 0.74757886 0.50299165 0. 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [1.05927751 2.11495253 0.17405731 1. 0. 0.\n 0. 0.38319418 1.0059833 0. 1.30023581 1.1038517\n 1.00529816 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[ 1.86089293 2.43539989 0.17405731 0. 0. 0.\n 57.29306153 0.59575478 0.50299165 0. 1.30023581 2.20770341\n 0.99471607 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [ 2.14718415 0. 0. 1. 0. 0.\n 56.31658848 1.44690925 0. 0. 1.30023581 0.\n 1.00529816 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[0.83024454 1.60223677 3.13303155 0. 0. 0.\n 0. 1.22688566 3.01794989 5.6084 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [1.37419785 1.85859465 2.87194559 0. 0. 0.\n 0. 1.20763446 3.01794989 5.6084 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[ 1.91815117 0.25635788 0.17405731 0. 45.37759597 0.\n 0. 0.84879683 2.0119666 0. 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [ 1.88952205 0.76907365 0.34811462 0. 45.37759597 0.\n 0. 0.6755532 3.01794989 0. 1.30023581 2.20770341\n 0.99471607 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[ 1.28831049 2.75584724 0.78325789 1. 0. 0.\n 0. 0.74453286 3.01794989 11.2168 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [ 1.68911819 3.07629459 0.87028654 1. 0. 0.\n 0. 0.67416501 3.01794989 11.2168 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[1.173794 0.25635788 2.95897425 0. 0. 0.\n 0.49820054 1.17411694 0.50299165 0. 1.30023581 2.0803359\n 1.00529816 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [1.14516488 0.25635788 2.95897425 0. 0. 0.\n 0. 1.3862301 0.50299165 0. 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[ 1.34556873 2.11495253 0.69622923 0. 0. 0.\n 11.95681284 0.51027163 3.01794989 0. 1.30023581 2.20770341\n 0.99471607 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [ 1.66048907 2.88402618 0.52217193 0. 0. 0.\n 12.45501338 0.16380159 3.01794989 0. 0.65011791 2.20770341\n 0.99471607 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 0. 0. ]\n[1.11653576 0.25635788 0.17405731 0. 0. 0.\n 0. 1.74132762 0.50299165 0. 1.30023581 2.12279174\n 0.99471607 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [1.20242312 0.25635788 0.17405731 0. 0. 0.\n 0. 0.44438961 0.50299165 0. 1.30023581 2.20770341\n 0.99471607 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[0.97339015 2.6276683 0.69622923 1. 0. 0.\n 0. 0.55421784 2.51495825 0. 1.30023581 2.20770341\n 0.99471607 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [0.97339015 2.6276683 0.60920058 1. 0. 0.\n 0. 0.6991583 3.01794989 0. 1.30023581 2.03788007\n 0.99471607 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[1.63185995 2.75584724 0.78325789 0. 0. 0.\n 0.24910027 1.31320058 3.01794989 0. 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [1.54597258 2.75584724 0.17405731 0. 0. 0.\n 0. 1.86499768 3.01794989 0. 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[ 1.14516488 0.70498418 0.17405731 0. 34.00094768 0.\n 0.57293062 0.57411152 2.0119666 0. 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [ 1.03064839 0.25635788 0.17405731 0. 34.00094768 0.\n 0.49820054 0.65458109 1.50897495 0. 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[1.57460171 0.25635788 2.95897425 0. 0. 0.\n 4.98200535 0.77034645 3.01794989 0. 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [1.48871434 0.57680524 3.0460029 0. 0. 0.\n 4.98200535 0.43494757 3.01794989 0. 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. ]\n[ 1.25968137 2.37131042 0.17405731 0. 0. 63.67395845\n 7.47300803 0.67673489 2.0119666 0. 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [ 1.46008522 2.69175777 0.60920058 0. 0. 63.67395845\n 9.9640107 0.23941821 0.50299165 0. 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[ 1.23105224 2.6276683 0.69622923 1. 0. 0.\n 0. 1.02482833 3.01794989 11.2168 1.30023581 2.20770341\n 0.99471607 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [ 1.25968137 0.76907365 0.34811462 0. 0. 0.\n 0. 1.09903954 3.01794989 11.2168 1.30023581 2.20770341\n 0.99471607 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[ 1.00201927 0.19226841 0.17405731 0. 34.00094768 0.\n 2.49100268 1.43771962 3.01794989 0. 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [ 1.51734346 0.25635788 0.17405731 1. 34.00094768 0.\n 2.49100268 1.75781391 2.51495825 0. 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[ 0.83024454 2.11495253 1.39245847 0. 62.97779366 0.\n 0. 1.49897814 0.50299165 0. 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 0. 0.\n 1. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [ 0.88750278 2.11495253 1.39245847 1. 62.97779366 0.\n 0. 0.93582306 0.50299165 0. 1.30023581 2.20770341\n 0.99471607 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[ 1.40282698 2.75584724 0.17405731 1. 0. 59.534492\n 3.98560428 1.56896739 2.51495825 11.2168 1.30023581 2.20770341\n 0.99471607 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [ 2.0899259 2.05086306 1.39245847 0. 0. 57.31974562\n 6.81040131 0.58146552 0.50299165 11.2168 1.30023581 2.12279174\n 1.00529816 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[ 1.46008522 2.179042 1.47948712 0. 0. 0.\n 8.96760963 0.93351131 3.01794989 11.2168 1.30023581 2.20770341\n 0.99471607 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [ 1.71774732 2.05086306 1.39245847 0. 0. 0.\n 9.9640107 0.84560742 1.50897495 11.2168 1.30023581 2.20770341\n 0.99471607 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[1.48871434 2.05086306 1.56651578 0. 0. 0.\n 0. 1.30072401 0.50299165 0. 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [1.68911819 2.11495253 1.39245847 0. 0. 0.\n 0. 0.62126435 0.50299165 0. 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[1.08790663 1.02543153 1.04434385 1. 0. 0.\n 0. 1.03931262 2.51495825 0. 1.30023581 2.20770341\n 0.99471607 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [1.03064839 0.83316312 0.34811462 0. 0. 0.\n 0. 0.52084947 2.51495825 0. 1.30023581 2.20770341\n 0.99471607 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[1.37419785 0.83316312 3.13303155 0. 0. 0.\n 0.23913626 0.74674709 3.01794989 5.6084 1.30023581 2.20770341\n 0.99471607 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [1.14516488 0.51271577 3.13303155 0. 0. 0.\n 0. 0.98278659 2.0119666 5.6084 1.30023581 2.20770341\n 0.99471607 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[1.28831049 2.69175777 0.60920058 0. 0. 0.\n 0. 0.64587331 0.50299165 0. 1.30023581 2.20770341\n 0.99471607 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [1.48871434 2.69175777 0.60920058 0. 0. 0.\n 0.09964011 0.23620011 0.50299165 0. 1.30023581 2.20770341\n 0.99471607 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[1.57460171 3.20447354 0.43514327 0. 0. 0.\n 5.97840642 1.07606544 3.01794989 0. 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [1.91815117 2.6276683 0.17405731 0. 0. 0.\n 7.57264813 1.13681917 3.01794989 0. 1.30023581 1.35858671\n 0.99471607 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 1. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[1.00201927 2.6276683 0.69622923 1. 0. 0.\n 0. 0.61352601 2.51495825 0. 1.30023581 2.20770341\n 0.99471607 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [1.48871434 2.6276683 0.69622923 1. 0. 0.\n 0. 1.6987524 3.01794989 0. 1.30023581 2.20770341\n 0.99471607 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[1.34556873 0.19226841 2.87194559 0. 0. 0.\n 0. 0.41329283 3.01794989 0. 1.30023581 2.20770341\n 0.99471607 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [1.23105224 0.19226841 2.95897425 0. 0. 0.\n 0. 0.90305127 3.01794989 0. 1.30023581 1.18876337\n 0.99471607 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[ 1.03064839 1.60223677 0.17405731 0. 62.97779366 0.\n 19.9280214 0.702646 3.01794989 0. 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. ] [ 1.34556873 2.24313148 1.39245847 0. 62.97779366 0.\n 19.9280214 1.19402209 3.01794989 0. 1.30023581 2.20770341\n 0.99471607 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[1.40282698 0.51271577 0.17405731 0. 0. 0.\n 0. 0.95311243 2.0119666 0. 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [1.60323083 0.70498418 0.17405731 0. 0. 0.\n 0. 1.17533878 2.51495825 0. 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[ 1.54597258 1.73041571 0.17405731 0. 16.51616854 0.\n 0.87185094 1.853829 1.0059833 0. 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [ 1.28831049 0.25635788 0.17405731 0. 16.51616854 0.\n 3.14862738 0.1793299 0.50299165 0. 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[1.31693961 2.75584724 0.87028654 1. 0. 0.\n 0. 0.94258048 2.51495825 0. 1.30023581 2.20770341\n 0.99471607 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [1.4314561 2.75584724 0.87028654 1. 0. 0.\n 0. 1.23000624 3.01794989 0. 1.30023581 2.20770341\n 0.99471607 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[1.34556873 1.98677359 2.26274501 0. 0. 0.\n 0. 0.50076076 3.01794989 0. 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [0.97339015 1.98677359 2.26274501 0. 0. 0.\n 0.05978406 0.65976676 3.01794989 0. 1.30023581 2.20770341\n 1.00529816 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 1. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[ 1.60323083 1.85859465 3.30708886 0. 0. 0.\n 19.9280214 0.66031745 3.01794989 0. 1.30023581 2.20770341\n 0.99471607 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [ 1.28831049 1.60223677 0.52217193 0. 0. 0.\n 20.92442247 0.78800295 3.01794989 0. 1.30023581 2.20770341\n 0.99471607 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n[ 1.25968137 2.69175777 1.04434385 1. 0. 0.\n 0. 1.45286933 2.0119666 11.2168 1.30023581 2.20770341\n 0.99471607 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ] [ 1.20242312 2.69175777 1.82760174 1. 0. 0.\n 0. 1.18653615 1.50897495 11.2168 1.30023581 2.20770341\n 0.99471607 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 1. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 0. 0. 0. 1. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 1. 0. 0. 0.\n 1. 0. 1. 0. 0. 0.\n 0. 1. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 0.\n 0. 0. 0. 0. 0. 1.\n 0. 0. 0. 0. 0. 1.\n 1. 0. ]\n" ], [ "#AdaFair\n\n#Tolerance to unfairness\nepsilon = 0\n\n\nget_base_clf = lambda: DecisionTreeClassifier(max_depth=adafair_max_depth)\n\nada_fair = AdaFairClassifier(DecisionTreeClassifier(max_depth=adafair_max_depth),\n algorithm=\"SAMME\",\n n_estimators=adafair_n_estimators,\n protected = protected_train,\n epsilon = epsilon)\n\nada_fair.fit(X_train, y_train)", "/usr/local/lib/python3.7/dist-packages/sklearn/ensemble/_weight_boosting.py:758: RuntimeWarning: invalid value encountered in true_divide\n yield (tmp_pred / norm).sum(axis=1)\n" ], [ "#Adaboost\n\n\nada_boost_sklearn = AdaBoostClassifier(DecisionTreeClassifier(max_depth=max_depth),\n algorithm=\"SAMME\",\n n_estimators=n_estimators)\n\nada_boost_sklearn.fit(X_train, y_train)", "_____no_output_____" ] ], [ [ "# 4) Compute and plot metrics", "_____no_output_____" ], [ "#### 4.1 Compute", "_____no_output_____" ] ], [ [ "names = ['ada_fair','ada_boost_sklearn', 'smoteboost', \"reweighted_classifier\"]\nclassifiers = [ada_fair, ada_boost_sklearn, smoteboost1, model_reweighted_classifier]\n\naccuracy = {}\nbal_accuracy = {}\nTPR = {}\nTNR = {}\neq_odds = {}\np_rule = {}\n#DELETA\n#y_test = y_test[:][1]\n\nfor i, clf in enumerate(classifiers):\n print(names[i])\n prediction = clf.predict(X_test)\n if i == 3:\n \n prediction = np.array(pd.read_csv(reweight_prediction, names = ['idx', 'pred'])['pred'][1:])\n print((prediction), (y_test))\n print(len(prediction), len(y_test))\n\n accuracy[names[i]] = (prediction == y_test).sum() * 1. / len(y_test)\n bal_accuracy[names[i]] = balanced_accuracy_score(y_test, prediction)\n print('accuracy {}: {}'.format(names[i], (prediction == y_test).sum() * 1. / len(y_test))) \n print('balanced accuracy {}: {}'.format(names[i], balanced_accuracy_score(y_test, prediction))) \n \n y_true_protected, y_pred_protected, y_true_non_protected, y_pred_non_protected = separate_protected_from_non_protected(y_test, prediction, protected_test)\n \n #TPR for protected group\n tp, tn, fp, fn = tp_tn_fp_fn(y_true_protected, y_pred_protected)\n TPR_protected = tp / (tp + fn)\n TNR_protected = tn / (tn + fp)\n FPR_protected = fp / (fp + tn)\n FNR_protected = fn / (fn + tp)\n print('TPR protected {}: {}'.format(names[i], TPR_protected))\n print('TNR protected {}: {}'.format(names[i], TNR_protected)) \n\n TPR[names[i] + ' protected'] = TPR_protected\n TNR[names[i] + ' protected'] = TNR_protected\n\n #TPR for non protected group\n tp, tn, fp, fn = tp_tn_fp_fn(y_true_non_protected, y_pred_non_protected)\n TPR_non_protected = tp / (tp + fn)\n TNR_non_protected = tn / (tn + fp)\n FPR_non_protected = fp / (fp + tn)\n FNR_non_protected = fn / (fn + tp)\n print('TPR non protected {}: {}'.format(names[i], TPR_non_protected))\n print('TNR non protected {}: {}'.format(names[i], TNR_non_protected))\n\n delta_FPR = -FPR_non_protected + FPR_protected\n delta_FNR = -FNR_non_protected + FNR_protected\n eq_odds[names[i]] = abs(delta_FPR) + abs(delta_FNR)\n\n TPR[names[i] + ' non protected'] = TPR_non_protected\n TNR[names[i] + ' non protected'] = TNR_non_protected\n\n p_rule[names[i]] = pRule(prediction, minority_idx_test)\n print('pRule {}: {}'.format(names[i],p_rule[names[i]]))\n ", "ada_fair\naccuracy ada_fair: 0.9228809846454807\nbalanced accuracy ada_fair: 0.7176952582693732\nTPR protected ada_fair: 0.6626686656671664\nTNR protected ada_fair: 0.9330514588008977\nTPR non protected ada_fair: 0.4335117332235488\nTNR non protected ada_fair: 0.9756010558607405\npRule ada_fair: 0.8097128558491885\nada_boost_sklearn\naccuracy ada_boost_sklearn: 0.946417109030332\nbalanced accuracy ada_boost_sklearn: 0.6529104164181531\nTPR protected ada_boost_sklearn: 0.08320839580209895\nTNR protected ada_boost_sklearn: 0.9992385379929465\nTPR non protected ada_boost_sklearn: 0.3812268423219432\nTNR non protected ada_boost_sklearn: 0.976409597869254\npRule ada_boost_sklearn: 0.047964583961589244\nsmoteboost\naccuracy smoteboost: 0.9368192609045616\nbalanced accuracy smoteboost: 0.5606524639130167\nTPR protected smoteboost: 0.13793103448275862\nTNR protected smoteboost: 0.984550336646361\nTPR non protected smoteboost: 0.12803622890078223\nTNR non protected smoteboost: 0.9989536515183943\npRule smoteboost: 0.7617401588170212\nreweighted_classifier\n[0 0 0 ... 0 0 0] 0 0\n1 0\n2 0\n3 0\n4 0\n ..\n98142 0\n98143 0\n98144 0\n98145 0\n98146 0\nName: income_50k, Length: 98147, dtype: int64\n98147 98147\naccuracy reweighted_classifier: 0.9402528859771567\nbalanced accuracy reweighted_classifier: 0.5368035200272799\nTPR protected reweighted_classifier: 0.08095952023988005\nTNR protected reweighted_classifier: 0.9981965373517153\nTPR non protected reweighted_classifier: 0.07348703170028818\nTNR non protected reweighted_classifier: 0.9988823095764666\npRule reweighted_classifier: 0.4486914878692678\n" ] ], [ [ "#### 4.2 Plot", "_____no_output_____" ] ], [ [ "labels = ['Accuracy', 'Bal. accuracy', 'Eq. odds','TPR prot.', 'TPR non-prot', 'TNR prot.', 'TNR non-prot.', 'pRule']\nadaFair_metrics = [accuracy['ada_fair'], bal_accuracy['ada_fair'], eq_odds['ada_fair'], TPR['ada_fair protected'], TPR['ada_fair non protected'], TNR['ada_fair protected'], TNR['ada_fair non protected'], p_rule['ada_fair']]\nadaBoost_metrics = [accuracy['ada_boost_sklearn'], bal_accuracy['ada_boost_sklearn'], eq_odds['ada_boost_sklearn'], TPR['ada_boost_sklearn protected'], TPR['ada_boost_sklearn non protected'], TNR['ada_boost_sklearn protected'], TNR['ada_boost_sklearn non protected'], p_rule['ada_boost_sklearn']]\nSMOTEBoost_metrics = [accuracy['smoteboost'], bal_accuracy['smoteboost'], eq_odds['smoteboost'], TPR['smoteboost protected'], TPR['smoteboost non protected'], TNR['smoteboost protected'], TNR['smoteboost non protected'], p_rule['smoteboost']]\nreweighted_class_metrics = [accuracy['reweighted_classifier'], bal_accuracy['reweighted_classifier'], eq_odds['reweighted_classifier'], TPR['reweighted_classifier protected'], TPR['reweighted_classifier non protected'], TNR['reweighted_classifier protected'], TNR['reweighted_classifier non protected'],p_rule['reweighted_classifier']]\n\nx = np.arange(len(labels)) # the label locations\nwidth = 0.20 # the width of the bars\n\nfont = {'family' : 'normal',\n 'weight' : 'normal',\n 'size' : 19}\n\nmatplotlib.rc('font', **font)\n\nfig, ax = plt.subplots(figsize = [15, 5])\nrects1 = ax.bar(x - 1.5*width, adaBoost_metrics, width, label='AdaBoost', color='gray')\nrects2 = ax.bar(x - width/2 , SMOTEBoost_metrics, width, label='SMOTEBoost', color='red')\nrects3 = ax.bar(x + width/2, adaFair_metrics, width, label='AdaFair', color='green')\nrects4 = ax.bar(x + 1.5*width, reweighted_class_metrics, width, label='ASR', color='blue')\n\nax.set_ylabel('Scores')\n#ax.set_title('Scores by algoritms')\nax.set_xticks(x)\nax.set_xticklabels(labels)\nax.legend(loc=9, ncol = 4, bbox_to_anchor=(0.5, 1.15))\nax.grid()\n\nfig.tight_layout()\n\nplt.show()", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code", "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "import json\nimport pandas as pd\nfrom bs4 import BeautifulSoup", "_____no_output_____" ], [ "with open('pahang.json') as fopen:\n pahang = json.load(fopen)", "_____no_output_____" ], [ "res = []\n\nfor j in pahang:\n s = BeautifulSoup(j)\n s.tbody.table.decompose()\n table_rows = s.tbody.find_all('tr')\n for tr in table_rows:\n td = tr.find_all('td')\n row = [tr.text.strip() for tr in td if tr.text.strip()]\n if row:\n res.append(row)", "_____no_output_____" ], [ "df = pd.DataFrame(res)\ndf", "_____no_output_____" ], [ "df.to_csv('pahang.csv', index = False)", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Apple and Tesla Split on 8/31", "_____no_output_____" ] ], [ [ "import numpy as np\nimport pandas as pd\nimport matplotlib.pyplot as plt\nimport seaborn as sns\nimport math\n\nimport warnings\nwarnings.filterwarnings(\"ignore\")\n\n# for fetching data\nimport yfinance as yf", "_____no_output_____" ], [ "# input\n# Coronavirus 2nd Wave\ntitle = \"Apple and Tesla\"\nsymbols = ['AAPL', 'TSLA']\nstart = '2020-01-01'\nend = '2020-08-31'", "_____no_output_____" ], [ "df = pd.DataFrame()\nfor symbol in symbols:\n df[symbol] = yf.download(symbol, start, end)['Adj Close']", "[*********************100%***********************] 1 of 1 completed\n[*********************100%***********************] 1 of 1 completed\n" ], [ "from datetime import datetime\nfrom dateutil import relativedelta\n\nd1 = datetime.strptime(start, \"%Y-%m-%d\")\nd2 = datetime.strptime(end, \"%Y-%m-%d\")\ndelta = relativedelta.relativedelta(d2, d1)\nprint('How many years of investing?')\nprint('%s years' % delta.years)", "How many years of investing?\n0 years\n" ], [ "number_of_years = delta.years", "_____no_output_____" ], [ "days = (df.index[-1] - df.index[0]).days\ndays", "_____no_output_____" ], [ "df.head()", "_____no_output_____" ], [ "df.tail()", "_____no_output_____" ], [ "df.min()", "_____no_output_____" ], [ "df.max()", "_____no_output_____" ], [ "df.describe()", "_____no_output_____" ], [ "plt.figure(figsize=(12,8))\nplt.plot(df)\nplt.title(title + ' Closing Price')\nplt.legend(labels=df.columns)\nplt.show()", "_____no_output_____" ], [ "# Normalize the data\nnormalize = (df - df.min())/ (df.max() - df.min())", "_____no_output_____" ], [ "plt.figure(figsize=(18,12))\nplt.plot(normalize)\nplt.title(title + ' Stocks Normalize')\nplt.legend(labels=normalize.columns)\nplt.show()", "_____no_output_____" ], [ "stock_rets = df.pct_change().dropna()", "_____no_output_____" ], [ "plt.figure(figsize=(12,8))\nplt.plot(stock_rets)\nplt.title(title + ' Stocks Returns')\nplt.legend(labels=stock_rets.columns)\nplt.show()", "_____no_output_____" ], [ "plt.figure(figsize=(12,8))\nplt.plot(stock_rets.cumsum())\nplt.title(title + ' Stocks Returns Cumulative Sum')\nplt.legend(labels=stock_rets.columns)\nplt.show()", "_____no_output_____" ], [ "sns.set(style='ticks')\nax = sns.pairplot(stock_rets, diag_kind='hist')\n\nnplot = len(stock_rets.columns)\nfor i in range(nplot) :\n for j in range(nplot) :\n ax.axes[i, j].locator_params(axis='x', nbins=6, tight=True)", "_____no_output_____" ], [ "ax = sns.PairGrid(stock_rets)\nax.map_upper(plt.scatter, color='purple')\nax.map_lower(sns.kdeplot, color='blue')\nax.map_diag(plt.hist, bins=30)\nfor i in range(nplot) :\n for j in range(nplot) :\n ax.axes[i, j].locator_params(axis='x', nbins=6, tight=True)", "_____no_output_____" ], [ "plt.figure(figsize=(10,10))\ncorr = stock_rets.corr()\n\n# plot the heatmap\nsns.heatmap(corr, \n xticklabels=corr.columns,\n yticklabels=corr.columns,\n cmap=\"Reds\")", "_____no_output_____" ], [ "# Box plot\nstock_rets.plot(kind='box',figsize=(24,8))", "_____no_output_____" ], [ "rets = stock_rets.dropna()\n\nplt.figure(figsize=(16,8))\nplt.scatter(rets.std(), rets.mean(),alpha = 0.5)\n\nplt.title('Stocks Risk & Returns')\nplt.xlabel('Risk')\nplt.ylabel('Expected Returns')\nplt.grid(which='major')\n\nfor label, x, y in zip(rets.columns, rets.std(), rets.mean()):\n plt.annotate(\n label, \n xy = (x, y), xytext = (50, 50),\n textcoords = 'offset points', ha = 'right', va = 'bottom',\n arrowprops = dict(arrowstyle = '-', connectionstyle = 'arc3,rad=-0.3'))", "_____no_output_____" ], [ "rets = stock_rets.dropna()\narea = np.pi*20.0\n\nsns.set(style='darkgrid')\nplt.figure(figsize=(16,8))\nplt.scatter(rets.std(), rets.mean(), s=area)\nplt.xlabel(\"Risk\", fontsize=15)\nplt.ylabel(\"Expected Return\", fontsize=15)\nplt.title(\"Return vs. Risk for Stocks\", fontsize=20)\n\nfor label, x, y in zip(rets.columns, rets.std(), rets.mean()) : \n plt.annotate(label, xy=(x,y), xytext=(50, 0), textcoords='offset points',\n arrowprops=dict(arrowstyle='-', connectionstyle='bar,angle=180,fraction=-0.2'),\n bbox=dict(boxstyle=\"round\", fc=\"w\"))", "_____no_output_____" ], [ "def annual_risk_return(stock_rets):\n tradeoff = stock_rets.agg([\"mean\", \"std\"]).T\n tradeoff.columns = [\"Return\", \"Risk\"]\n tradeoff.Return = tradeoff.Return*252\n tradeoff.Risk = tradeoff.Risk * np.sqrt(252)\n return tradeoff", "_____no_output_____" ], [ "tradeoff = annual_risk_return(stock_rets)\ntradeoff", "_____no_output_____" ], [ "import itertools\n\ncolors = itertools.cycle([\"r\", \"b\", \"g\"])\n\ntradeoff.plot(x = \"Risk\", y = \"Return\", kind = \"scatter\", figsize = (13,9), s = 20, fontsize = 15, c='g')\nfor i in tradeoff.index:\n plt.annotate(i, xy=(tradeoff.loc[i, \"Risk\"]+0.002, tradeoff.loc[i, \"Return\"]+0.002), size = 15)\nplt.xlabel(\"Annual Risk\", fontsize = 15)\nplt.ylabel(\"Annual Return\", fontsize = 15)\nplt.title(\"Return vs. Risk for \" + title + \" Stocks\", fontsize = 20)\nplt.show()", "_____no_output_____" ], [ "rest_rets = rets.corr()\npair_value = rest_rets.abs().unstack()\npair_value.sort_values(ascending = False)", "_____no_output_____" ], [ "# Normalized Returns Data\nNormalized_Value = ((rets[:] - rets[:].min()) /(rets[:].max() - rets[:].min()))\nNormalized_Value.head()", "_____no_output_____" ], [ "Normalized_Value.corr()", "_____no_output_____" ], [ "normalized_rets = Normalized_Value.corr()\nnormalized_pair_value = normalized_rets.abs().unstack()\nnormalized_pair_value.sort_values(ascending = False)", "_____no_output_____" ], [ "print(\"Stock returns: \")\nprint(rets.mean())\nprint('-' * 50)\nprint(\"Stock risks:\")\nprint(rets.std())", "Stock returns: \nAAPL 0.003572\nTSLA 0.011590\ndtype: float64\n--------------------------------------------------\nStock risks:\nAAPL 0.031211\nTSLA 0.059151\ndtype: float64\n" ], [ "table = pd.DataFrame()\ntable['Returns'] = rets.mean()\ntable['Risk'] = rets.std()\ntable.sort_values(by='Returns')", "_____no_output_____" ], [ "table.sort_values(by='Risk')", "_____no_output_____" ], [ "rf = 0.01\ntable['Sharpe Ratio'] = (table['Returns'] - rf) / table['Risk']\ntable", "_____no_output_____" ], [ "table['Max Returns'] = rets.max()", "_____no_output_____" ], [ "table['Min Returns'] = rets.min()", "_____no_output_____" ], [ "table['Median Returns'] = rets.median()", "_____no_output_____" ], [ "total_return = stock_rets[-1:].transpose()\ntable['Total Return'] = 100 * total_return\ntable", "_____no_output_____" ], [ "table['Average Return Days'] = (1 + total_return)**(1 / days) - 1\ntable", "_____no_output_____" ], [ "initial_value = df.iloc[0]\nending_value = df.iloc[-1]\ntable['CAGR'] = ((ending_value / initial_value) ** (252.0 / days)) -1\ntable", "_____no_output_____" ], [ "table.sort_values(by='Average Return Days')", "_____no_output_____" ] ] ]

[ "markdown", "code" ]

[ [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import numpy as np\nimport matplotlib.pyplot as plt\n%matplotlib inline\nimport pandas as pd\nfrom matplotlib.pylab import rcParams\nrcParams['figure.figsize'] = 15, 6\nimport seaborn as sns\nimport warnings\nwarnings.filterwarnings('ignore')", "_____no_output_____" ], [ "ori=pd.read_csv('website_data_20190225.csv')\nori.drop(['STATE','DISTRICT','WLCODE','SITE_TYPE','TEH_NAME'],axis=1,inplace=True)\nori.replace(to_replace=\"'0\",value=0,inplace=True)\nori.head()", "_____no_output_____" ], [ "dataset=pd.DataFrame().reindex_like(ori)\ndataset1=pd.DataFrame().reindex_like(ori)\ndataset.dropna(inplace=True)\ndataset1.dropna(inplace=True)\n\n# j=0\n# for i in range(0,ori.shape[0]):\n# if ori['STATE'][i]=='RJ':\n# dataset.loc[j] = ori.iloc[i]\n# j+=1\n# dataset.drop(['STATE'],axis=1,inplace=True)\n\n# j=0\n# for i in range(0,ori.shape[0]):\n# if ori['DISTRICT'][i]=='Ajmer':\n# dataset.loc[j] = ori.iloc[i]\n# j+=1\n# dataset.drop(['DISTRICT'],axis=1,inplace=True)\n\nj=0\nfor i in range(0,ori.shape[0]):\n if ori['BLOCK_NAME'][i]=='Arain':\n dataset1.loc[j] = ori.iloc[i]\n j+=1\ndataset1.drop(['BLOCK_NAME'],axis=1,inplace=True)\ndataset.drop(['BLOCK_NAME'],axis=1,inplace=True)\n\nj=0\nfor i in range(0,dataset1.shape[0]):\n if dataset1['SITE_NAME'][i]=='Sanpla':\n dataset.loc[j] = dataset1.iloc[i]\n j+=1\nlat=dataset[\"LAT\"][0]\nlon=dataset[\"LON\"][0]\ndataset.drop(['SITE_NAME','LAT','LON'],axis=1,inplace=True)\n \ndataset", "_____no_output_____" ], [ "for i in range(0,dataset.shape[0]):\n dataset['MONSOON'][i]=float(dataset['MONSOON'][i])\n dataset['POMRB'][i]=float(dataset['POMRB'][i])\n dataset['POMKH'][i]=float(dataset['POMKH'][i])\n dataset['PREMON'][i]=float(dataset['PREMON'][i])\n dataset['YEAR_OBS'][i]=int(dataset['YEAR_OBS'][i])", "_____no_output_____" ], [ "first=list(dataset['MONSOON'])\nsecond=list(dataset['POMRB'])\nthird=list(dataset['POMKH'])\nfourth=list(dataset['PREMON'])\ndataset['MONSOON']=pd.core.frame.DataFrame(x+y+z+w for x, y,z,w in zip(first, second, third, fourth))\ndataset.drop(['POMRB','POMKH','PREMON'],axis=1,inplace=True)\ndataset = dataset.iloc[::-1]\ndataset", "_____no_output_____" ], [ "dataset['YEAR_OBS']=(dataset['YEAR_OBS']).apply(np.int64)", "_____no_output_____" ], [ "dataset['YEAR_OBS']=pd.to_datetime(dataset['YEAR_OBS'],yearfirst=True,format='%Y',infer_datetime_format=True)\nindexedDataset=dataset.set_index(['YEAR_OBS'])", "_____no_output_____" ], [ "from datetime import datetime\nindexedDataset.head(50)", "_____no_output_____" ], [ "plt.xlabel('Years')\nplt.ylabel('Water-Level')\nplt.plot(indexedDataset)", "_____no_output_____" ] ], [ [ "- A stationary time series is one whose statistical properties such as mean, variance, autocorrelation, etc. are all constant over time. Most statistical forecasting methods are based on the assumption that the time series can be rendered approximately stationary (i.e., \"stationarized\") through the use of mathematical transformations. A stationarized series is relatively easy to predict: you simply predict that its statistical properties will be the same in the future as they have been in the past!", "_____no_output_____" ], [ "- We can check stationarity using the following:\n\n- - Plotting Rolling Statistics: We can plot the moving average or moving variance and see if it varies with time. This is more of a visual technique.\n- - Dickey-Fuller Test: This is one of the statistical tests for checking stationarity. Here the null hypothesis is that the TimeSeries is non-stationary. The test results comprise of a Test Statistic and some Critical Values for difference confidence levels. If the ‘Test Statistic’ is less than the ‘Critical Value’, we can reject the null hypothesis and say that the series is stationary.", "_____no_output_____" ] ], [ [ "from statsmodels.tsa.stattools import adfuller\ndef test_stationary(timeseries):\n \n #Determing rolling statistics\n moving_average=timeseries.rolling(window=12).mean()\n standard_deviation=timeseries.rolling(window=12).std()\n \n #Plot rolling statistics:\n plt.plot(timeseries,color='blue',label=\"Original\")\n plt.plot(moving_average,color='red',label='Mean')\n plt.plot(standard_deviation,color='black',label='Standard Deviation')\n plt.legend(loc='best') #best for axes\n plt.title('Rolling Mean & Deviation')\n# plt.show()\n plt.show(block=False)\n \n #Perform Dickey-Fuller test:\n print('Results Of Dickey-Fuller Test')\n tstest=adfuller(timeseries['MONSOON'],autolag='AIC')\n tsoutput=pd.Series(tstest[0:4],index=['Test Statistcs','P-value','#Lags used',\"#Obs. used\"])\n #Test Statistics should be less than the Critical Value for Stationarity\n #lesser the p-value, greater the stationarity\n # print(list(dftest))\n for key,value in tstest[4].items():\n tsoutput['Critical Value (%s)'%key]=value\n print((tsoutput))", "_____no_output_____" ], [ "test_stationary(indexedDataset)", "_____no_output_____" ] ], [ [ "- There are 2 major reasons behind non-stationaruty of a TS:\n- - Trend – varying mean over time. For eg, in this case we saw that on average, the number of passengers was growing over time.\n- - Seasonality – variations at specific time-frames. eg people might have a tendency to buy cars in a particular month because of pay increment or festivals.", "_____no_output_____" ], [ "## Indexed Dataset Logscale", "_____no_output_____" ] ], [ [ "indexedDataset_logscale=np.log(indexedDataset)\ntest_stationary(indexedDataset_logscale)", "_____no_output_____" ] ], [ [ "## Dataset Log Minus Moving Average (dl_ma)", "_____no_output_____" ] ], [ [ "rolmeanlog=indexedDataset_logscale.rolling(window=12).mean()\ndl_ma=indexedDataset_logscale-rolmeanlog\ndl_ma.head(12)", "_____no_output_____" ], [ "dl_ma.dropna(inplace=True)\ndl_ma.head(12)", "_____no_output_____" ], [ "test_stationary(dl_ma)", "_____no_output_____" ] ], [ [ "## Exponential Decay Weighted Average (edwa)", "_____no_output_____" ] ], [ [ "edwa=indexedDataset_logscale.ewm(halflife=12,min_periods=0,adjust=True).mean()\nplt.plot(indexedDataset_logscale)\nplt.plot(edwa,color='red')", "_____no_output_____" ] ], [ [ "## Dataset Logscale Minus Moving Exponential Decay Average (dlmeda)", "_____no_output_____" ] ], [ [ "dlmeda=indexedDataset_logscale-edwa\ntest_stationary(dlmeda)", "_____no_output_____" ] ], [ [ "## Eliminating Trend and Seasonality", "_____no_output_____" ], [ "- Differencing – taking the differece with a particular time lag\n- Decomposition – modeling both trend and seasonality and removing them from the model.", "_____no_output_____" ], [ "# Differencing", "_____no_output_____" ], [ "## Dataset Log Div Shifting (dlds)", "_____no_output_____" ] ], [ [ "#Before Shifting\nindexedDataset_logscale.head()", "_____no_output_____" ], [ "#After Shifting\nindexedDataset_logscale.shift().head()", "_____no_output_____" ], [ "dlds=indexedDataset_logscale-indexedDataset_logscale.shift()\ndlds.dropna(inplace=True)\ntest_stationary(dlds)", "_____no_output_____" ] ], [ [ "# Decomposition", "_____no_output_____" ] ], [ [ "from statsmodels.tsa.seasonal import seasonal_decompose\ndecompostion= seasonal_decompose(indexedDataset_logscale,freq=10)\n\ntrend=decompostion.trend\nseasonal=decompostion.seasonal\nresidual=decompostion.resid\n\nplt.subplot(411)\nplt.plot(indexedDataset_logscale,label='Original')\nplt.legend(loc='best')\n\nplt.subplot(412)\nplt.plot(trend,label='Trend')\nplt.legend(loc='best')\n\nplt.subplot(413)\nplt.plot(seasonal,label='Seasonal')\nplt.legend(loc='best')\n\nplt.subplot(414)\nplt.plot(residual,label='Residual')\nplt.legend(loc='best')\n\nplt.tight_layout() #To Show Multiple Grpahs In One Output, Use plt.subplot(abc)", "_____no_output_____" ] ], [ [ "- Here trend, seasonality are separated out from data and we can model the residuals. Lets check stationarity of residuals:", "_____no_output_____" ] ], [ [ "decomposedlogdata=residual\ndecomposedlogdata.dropna(inplace=True)\ntest_stationary(decomposedlogdata)", "_____no_output_____" ] ], [ [ "# Forecasting a Time Series", "_____no_output_____" ], [ "- ARIMA stands for Auto-Regressive Integrated Moving Averages. The ARIMA forecasting for a stationary time series is nothing but a linear (like a linear regression) equation. The predictors depend on the parameters (p,d,q) of the ARIMA model:\n\n- - Number of AR (Auto-Regressive) terms (p): AR terms are just lags of dependent variable. For instance if p is 5, the predictors for x(t) will be x(t-1)….x(t-5).\n- - Number of MA (Moving Average) terms (q): MA terms are lagged forecast errors in prediction equation. For instance if q is 5, the predictors for x(t) will be e(t-1)….e(t-5) where e(i) is the difference between the moving average at ith instant and actual value.\n- - Number of Differences (d): These are the number of nonseasonal differences, i.e. in this case we took the first order difference. So either we can pass that variable and put d=0 or pass the original variable and put d=1. Both will generate same results.\n\n\n- An importance concern here is how to determine the value of ‘p’ and ‘q’. We use two plots to determine these numbers.\n\n- - Autocorrelation Function (ACF): It is a measure of the correlation between the the TS with a lagged version of itself-. For instance at lag 5, ACF would compare series at time instant ‘t1’…’t2’ with series at instant ‘t1-5’…’t2-5’ (t1-5 and t2 being end points).\n- - Partial Autocorrelation Function (PACF): This measures the correlation between the TS with a lagged version of itself but after eliminating the variations already explained by the intervening comparisons. Eg at lag 5, it will check the correlation but remove the effects already explained by lags 1 to 4.", "_____no_output_____" ], [ "## ACF & PACF Plots", "_____no_output_____" ] ], [ [ "from statsmodels.tsa.stattools import acf,pacf\nlag_acf=acf(dlds,nlags=20)\nlag_pacf=pacf(dlds,nlags=20,method='ols')\n\nplt.subplot(121)\nplt.plot(lag_acf)\nplt.axhline(y=0, linestyle='--',color='gray')\nplt.axhline(y=1.96/np.sqrt(len(dlds)),linestyle='--',color='gray')\nplt.axhline(y=-1.96/np.sqrt(len(dlds)),linestyle='--',color='gray')\nplt.title('AutoCorrelation Function')\n\nplt.subplot(122)\nplt.plot(lag_pacf)\nplt.axhline(y=0, linestyle='--',color='gray')\nplt.axhline(y=1.96/np.sqrt(len(dlds)),linestyle='--',color='gray')\nplt.axhline(y=-1.96/np.sqrt(len(dlds)),linestyle='--',color='gray')\nplt.title('PartialAutoCorrelation Function')\n\nplt.tight_layout()", "_____no_output_____" ] ], [ [ "- In this plot, the two dotted lines on either sides of 0 are the confidence interevals. These can be used to determine the ‘p’ and ‘q’ values as:\n\n- - p – The lag value where the PACF chart crosses the upper confidence interval for the first time. If we notice closely, in this case p=2.\n- - q – The lag value where the ACF chart crosses the upper confidence interval for the first time. If we notice closely, in this case q=2.", "_____no_output_____" ] ], [ [ "from statsmodels.tsa.arima_model import ARIMA\n\nmodel=ARIMA(indexedDataset_logscale,order=(5,1,0))\nresults_AR=model.fit(disp=-1)\nplt.plot(dlds)\nplt.plot(results_AR.fittedvalues,color='red')\nplt.title('RSS: %.4f'%sum((results_AR.fittedvalues-dlds['MONSOON'])**2))\nprint('Plotting AR Model')", "Plotting AR Model\n" ], [ "model = ARIMA(indexedDataset_logscale, order=(0, 1, 2)) #0,1,2\nresults_MA = model.fit(disp=-1) \nplt.plot(dlds)\nplt.plot(results_MA.fittedvalues, color='red')\nplt.title('RSS: %.4f'%sum((results_MA.fittedvalues-dlds['MONSOON'])**2))\nprint('Plotting MA Model')", "Plotting MA Model\n" ], [ "import itertools\na={}\np=d=q=range(0,6)\npdq=list(itertools.product(p,d,q))\nfor i in pdq:\n try:\n model_arima=(ARIMA(indexedDataset_logscale,order=i)).fit()\n a[i]=model_arima.aic\n except:\n continue\n# min(a, key=a.get)", "_____no_output_____" ], [ "RSS=[]\nRSS1=[]\nfor i in a.keys(): \n model = ARIMA(indexedDataset_logscale, order=i)\n results_ARIMA = model.fit(disp=-1)\n RSS.append(sum((results_ARIMA.fittedvalues-dlds['MONSOON'])**2))\nfor i in range(0,len(RSS)):\n if(~np.isnan(RSS[i])):\n RSS1.append(RSS[i])\nmin(RSS1)", "_____no_output_____" ], [ "model = ARIMA(indexedDataset_logscale, order=(5, 1, 2)) \nresults_ARIMA = model.fit(disp=-1) \nplt.plot(dlds)\nplt.plot(results_ARIMA.fittedvalues, color='red')\nplt.title('RSS: %.4f'%sum((results_ARIMA.fittedvalues-dlds['MONSOON'])**2))\nprint('Plotting Combined Model')", "Plotting Combined Model\n" ] ], [ [ "# Taking it back to original scale from residual scale", "_____no_output_____" ] ], [ [ "#storing the predicted results as a separate series\npredictions_ARIMA_diff = pd.Series(results_ARIMA.fittedvalues, copy=True)\npredictions_ARIMA_diff.head()", "_____no_output_____" ] ], [ [ "- Notice that these start from ‘1949-02-01’ and not the first month. Why? This is because we took a lag by 1 and first element doesn’t have anything before it to subtract from. The way to convert the differencing to log scale is to add these differences consecutively to the base number. An easy way to do it is to first determine the cumulative sum at index and then add it to the base number.", "_____no_output_____" ] ], [ [ "#convert to cummuative sum\npredictions_ARIMA_diff_cumsum = predictions_ARIMA_diff.cumsum()\npredictions_ARIMA_diff_cumsum", "_____no_output_____" ], [ "predictions_ARIMA_log = pd.Series(indexedDataset_logscale['MONSOON'].ix[0], index=indexedDataset_logscale.index)\npredictions_ARIMA_log", "_____no_output_____" ], [ "predictions_ARIMA_log = predictions_ARIMA_log.add(predictions_ARIMA_diff_cumsum,fill_value=0)\npredictions_ARIMA_log", "_____no_output_____" ] ], [ [ "- Here the first element is base number itself and from there on the values cumulatively added.", "_____no_output_____" ] ], [ [ "#Last step is to take the exponent and compare with the original series.\npredictions_ARIMA = np.exp(predictions_ARIMA_log)\nplt.plot(indexedDataset)\nplt.plot(predictions_ARIMA)\nplt.title('RMSE: %.4f'% np.sqrt(sum((predictions_ARIMA-indexedDataset['MONSOON'])**2)/len(indexedDataset)))", "_____no_output_____" ] ], [ [ "- Finally we have a forecast at the original scale.", "_____no_output_____" ] ], [ [ "results_ARIMA.plot_predict(1,26)\n\n#start = !st month\n#end = 10yrs forcasting = 144+12*10 = 264th month\n\n#Two models corresponds to AR & MA", "_____no_output_____" ], [ "x=results_ARIMA.forecast(steps=5)\nprint(x)\n#values in residual equivalent", "(array([2.61731503, 2.74811248, 2.79018171, 2.72974134, 2.71475412]), array([0.23507659, 0.33764008, 0.33764014, 0.3376707 , 0.33861065]), array([[2.15657338, 3.07805668],\n [2.08635009, 3.40987487],\n [2.12841919, 3.45194423],\n [2.06791892, 3.39156376],\n [2.05108943, 3.3784188 ]]))\n" ], [ "for i in range(0,5):\n print(x[0][i],end='')\n print('\\t',x[1][i],end='')\n print('\\t',x[2][i])", "2.617315031254026\t 0.2350765910977012\t [2.15657338 3.07805668]\n2.7481124785660804\t 0.3376400750662718\t [2.08635009 3.40987487]\n2.790181712476766\t 0.3376401428052289\t [2.12841919 3.45194423]\n2.7297413397725365\t 0.3376707032327619\t [2.06791892 3.39156376]\n2.7147541154092676\t 0.3386106520100188\t [2.05108943 3.3784188 ]\n" ], [ "np.exp(results_ARIMA.forecast(steps=5)[0])", "_____no_output_____" ], [ "predictions_ARIMA_diff = pd.Series(results_ARIMA.forecast(steps=5)[0], copy=True)\npredictions_ARIMA_diff.head()", "_____no_output_____" ], [ "predictions_ARIMA_diff_cumsum = predictions_ARIMA_diff.cumsum()\npredictions_ARIMA_diff_cumsum.head()", "_____no_output_____" ], [ "predictions_ARIMA_log=[]\nfor i in range(0,len(predictions_ARIMA_diff_cumsum)):\n predictions_ARIMA_log.append(predictions_ARIMA_diff_cumsum[i]+3.411478)\npredictions_ARIMA_log", "_____no_output_____" ], [ "#Last step is to take the exponent and compare with the original series.\npredictions_ARIMA = np.exp(predictions_ARIMA_log)\nplt.subplot(121)\nplt.plot(indexedDataset)\nplt.subplot(122)\nplt.plot(predictions_ARIMA)\nplt.tight_layout()\n# plt.title('RMSE: %.4f'% np.sqrt(sum((predictions_ARIMA-indexedDataset['MONSOON'])**2)/len(indexedDataset)))", "_____no_output_____" ], [ "np.exp(predictions_ARIMA_log)", "_____no_output_____" ] ] ]

[ "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code", "code", "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown", "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "## 1. Create an assert statement that throws an AssertionError if the variable spam is a negative integer.", "_____no_output_____" ] ], [ [ "import pyinputplus as pyip\n\ndef is_pos_integer():\n n = pyip.inputInt(prompt='Enter a positive integer: ')\n assert n > 0, 'This is a negative integer'", "_____no_output_____" ], [ "is_pos_integer()", "Enter a positive integer: -1\n" ] ], [ [ "## 2. Write an assert statement that triggers an AssertionError if the variables eggs and bacon contain strings that are the same as each other, even if their cases are different (that is, &#39;hello&#39; and &#39;hello&#39; are considered the same, and &#39;goodbye&#39; and &#39;GOODbye&#39; are also considered the same).", "_____no_output_____" ] ], [ [ "import re\nimport pyinputplus as pyip\n\ndef same_or_different():\n eggs = pyip.inputStr(prompt='What is eggs: ')\n bacon = pyip.inputStr(prompt='What is bacon: ')\n \n assert eggs.lower() != bacon.lower(), 'Strings are the same!'", "_____no_output_____" ], [ "same_or_different()", "What is eggs: hello\nWhat is bacon: Hello\n" ] ], [ [ "## 3. Create an assert statement that throws an AssertionError every time.", "_____no_output_____" ] ], [ [ "assert 0 !=0 , 'Assertion Error every time!'", "_____no_output_____" ] ], [ [ "## 4. What are the two lines that must be present in your software in order to call logging.debug()?\n > import logging\n > logging.basicConfig(level=logging.DEBUG, format='%(asctime)s %(levelname)s %(message)s')\n\n## 5. What are the two lines that your program must have in order to have logging.debug() send a logging message to a file named programLog.txt?\n > import logging\n > logging.basicConfig(filename='programLog.txt', level=logging.DEBUG, format='%(asctime)s %(levelname)s %(message)s')\n\n## 6. What are the five levels of logging?\n > DEBUG, INFO, WARNING, ERROR, CRITICAL\n\n## 7. What line of code would you add to your software to disable all logging messages?\n > lg.disable(lg.CRITICAL)\n\n## 8.Why is using logging messages better than using print() to display the same message?\n > Because of timestamps, we have a log of erros, without the messages in console. \n\n## 9. What are the differences between the Step Over, Step In, and Step Out buttons in the debugger?\n * Step - executes the current line\n * Over - quickly executes function without stepping into it\n * Out - executes the rest of the code until it steps out to the place it is currently\n\n## 10.After you click Continue, when will the debugger stop ?\n > Upon breakpoint.\n\n## 11. What is the concept of a breakpoint?\n > Debugger stops at breakpoint. And this is the main concept behind it. ", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown" ]

[ [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "empty" ] ] ]

[ "empty" ]

[ [ "empty" ] ]

[ "OML" ]

[ "OML" ]

[ "OML" ]

[ [ [ "# 2 Dead reckoning\n\n*Dead reckoning* is a means of navigation that does not rely on external observations. Instead, a robot’s position is estimated by summing its incremental movements relative to a known starting point.\n\nEstimates of the distance traversed are usually obtained from measuring how many times the wheels have turned, and how many times they have turned in relation to each other. For example, the wheels of the robot could be attached to an odometer, similar to the device that records the mileage of a car.\n\nIn RoboLab we will calculate the position of a robot from how long it moves in a straight line or rotates about its centre. We will assume that the length of time for which the motors are switched on is directly related to the distance travelled by the wheels.", "_____no_output_____" ], [ "*By design, the simulator does not provide the robot with access to any magical GPS-style service. In principle, we could create a magical ‘simulated-GPS’ sensor that would allow the robot to identify its location from the simulator’s point of view; but in the real world we can’t always guarantee that external location services are available. For example, GPS doesn’t work indoors or underground, or even in many cities where line-of-sight access to four or more GPS satellites is not available.*\n\n*Furthermore, the robot cannot magically teleport itself to a new location from within a program. Only the magics can teleport the robot to a specific location...*\n\n*Although the simulator is omniscient and does keep track of where the robot is, the robot must figure out for itself where it is based on things like how far the motors have turned, or from its own sensor readings (ultrasound-based distance to a target, for example, or gyroscope heading); you will learn how to make use of sensors for navigation in later notebooks.*", "_____no_output_____" ], [ "## 2.1 Activity – Dead reckoning\n\nAn environment for the simulated robot to navigate is shown below, based on the 2018 First Lego League ‘Into Orbit’ challenge.\n\nThe idea is that the robot must get to the target satellite from its original starting point by avoiding the obstacles in its direct path.\n\n", "_____no_output_____" ], [ "The [First Lego League (FLL)](https://www.firstlegoleague.org/) is a friendly international youth-based robot competition in which teams compete at national and international level on an annual basis. School teams are often coached by volunteers. In the UK, volunteers often coach teams under the auspices of the [STEM Ambassadors Scheme](https://www.stem.org.uk/stem-ambassadors). Many companies run volunteering schemes that allow employees to volunteer their skills in company time using schemes such as STEM Ambassadors.", "_____no_output_____" ], [ "Load in the simulator in the usual way:", "_____no_output_____" ] ], [ [ "from nbev3devsim.load_nbev3devwidget import roboSim, eds\n\n%load_ext nbev3devsim", "_____no_output_____" ] ], [ [ "To navigate the environment, we will use a small robot configuration within the simulator. The robot configuration can be set via the simulator user interface, or by passing the `-r Small_Robot` parameter setting in the simulator magic.\n\nThe following program should drive the robot from its starting point to the target, whilst avoiding the obstacles. We define the obstacle as being avoided if it is not crossed by the robot’s *pen down* trail.\n\nLoad the *FLL_2018_Into_Orbit* background into the simulator. Run the following code cell to download the program to the simulator and then, with the *pen down*, run the program in the simulator.", "_____no_output_____" ], [ "Remember, you can use the `-P / --pencolor` flag to change the pen colour and the `-C / --clear` option to clear the pen trace.", "_____no_output_____" ], [ "Does the robot reach the target satellite without encountering any obstacles?", "_____no_output_____" ] ], [ [ "%%sim_magic_preloaded -b FLL_2018_Into_Orbit -p -r Small_Robot\n\n# Turn on the spot to the right\ntank_turn.on_for_rotations(100, SpeedPercent(70), 1.7 )\n\n# Go forwards\ntank_drive.on_for_rotations(SpeedPercent(30), SpeedPercent(30), 20)\n\n# Slight graceful turn to left\ntank_drive.on_for_rotations(SpeedPercent(35), SpeedPercent(50), 8.5)\n\n# Turn on the spot to the left\ntank_turn.on_for_rotations(-100, SpeedPercent(75), 0.8)\n\n# Forwards a bit\ntank_drive.on_for_rotations(SpeedPercent(30), SpeedPercent(30), 2.0)\n\n# Turn on the spot a bit more to the right\ntank_turn.on_for_rotations(100, SpeedPercent(60), 0.4 )\n\n# Go forwards a bit more and dock on the satellite\ntank_drive.on_for_rotations(SpeedPercent(30), SpeedPercent(30), 1.0)\n\nsay(\"Hopefully I have docked with the satellite...\")", "_____no_output_____" ] ], [ [ "*Add your notes on how well the simulated robot performed the task here.*", "_____no_output_____" ], [ "To set the speeds and times, I used a bit of trial and error.\n\nIf the route had been much more complex, then I would have been tempted to comment out the steps up I had already run and add new steps that would be applied from wherever the robot was currently located.\n\nNote that the robot could have taken other routes to get to the satellite – I just thought I should avoid the asteroid!", "_____no_output_____" ], [ "### 2.1.1 Using motor tacho counts to identify how far the robot has travelled\n\nIn the above example, the motors were turned on for a specific amount of time to move the robot on each leg of its journey. This would not be an appropriate control strategy if we wanted to collect sensor data along the route, because the `on_for_X()` motor commands are blocking commands.\n\nHowever, suppose we replaced the forward driving `tank_drive.on_for_rotations()` commands with commands of the form:\n\n```python\nfrom time import sleep\n\ntank_drive.on(SPEED)\nwhile int(tank_drive.left_motor.position) < DISTANCE:\n # We need something that takes a finite time\n # to run in the loop or the program will hang\n sleep(0.1)\n```\n\nNow we could drive the robot forwards until the motor tacho count exceeds a specified `DISTANCE` and at the same time, optionally include additional commands, such as sensor data-logging commands, inside the body of each `while` loop.\n\n*As well as `tank_drive.left_motor.position` we can also refer to `tank_drive.right_motor.position`. Also note that these values are returned as strings and need to be cast to integers for numerical comparisons.*", "_____no_output_____" ], [ "### 2.1.2 Activity – Dead reckoning over distances (optional)\n\nUse the `.left_motor.position` and/or `.right_motor.position` motor tacho counts in a program that allows the robot to navigate from its home base to the satellite rendezvous.", "_____no_output_____" ], [ "*Your design notes here.*", "_____no_output_____" ] ], [ [ "# YOUR CODE HERE", "_____no_output_____" ] ], [ [ "*Your notes and observations here.*", "_____no_output_____" ], [ "## 2.2 Challenge – Reaching the moon base", "_____no_output_____" ], [ "In the following code cell, write a program to move the simulated robot from its location servicing the satellite to the moon base identified as the circular area marked on the moon in the top right-hand corner of the simulated world.\n\nIn the simulator, set the robot’s *x* location to `1250` and *y* location to `450`.\n\nUse the following code cell to write your own dead-reckoning program to drive the robot to the moon base at location `(2150, 950)`.", "_____no_output_____" ] ], [ [ "%%sim_magic_preloaded\n\n# YOUR CODE HERE\n", "_____no_output_____" ] ], [ [ "## 2.3 Dead reckoning with noise\n\nThe robot traverses its path using timing information for dead reckoning. In principle, if the simulated robot had a map then it could calculate all the distances and directions for itself, convert these to times, and dead reckon its way to the target. However, there is a problem with dead reckoning: *noise*.\n\nIn many physical systems, a perfect intended behaviour is subject to *noise* – random perturbations that arise within the system as time goes on as a side effect of its operation. In a robot, noise might arise in the behaviour of the motors, the transmission or the wheels. The result is that the robot does not execute its motion without error. We can model noise effects in the mobility system of our robot by adding a small amount of noise to the motor speeds as the simulator runs. This noise component may speed up or slow down the speed of each motor, in a random way. As with real systems, the noise represents slight random deviations from the theoretical, ideal behaviour.\n\nFor the following experiment, create a new, empty background cleared of pen traces.", "_____no_output_____" ] ], [ [ "%sim_magic -b Empty_Map --clear", "_____no_output_____" ] ], [ [ "Run the following code cell to download the program to the simulator using an empty background (select the *Empty_Map*) and the *Pen Down* mode selected. Also reset the initial location of the robot to an *x* value of `150` and *y* value of `400`.\n\nRun the program in the simulator and observe what happens.", "_____no_output_____" ] ], [ [ "%%sim_magic_preloaded -b Empty_Map -p -x 150 -y 400 -r Small_Robot --noisecontrols\n\n\ntank_drive.on_for_rotations(SpeedPercent(30),\n SpeedPercent(30), 10)", "_____no_output_____" ] ], [ [ "*Record your observations here describing what happens when you run the program.*", "_____no_output_____" ], [ "When you run the program, you should see the robot drive forwards a short way in a straight line, leaving a straight line trail behind it.\n\nReset the location of the robot. Within the simulator, use the *Noise controls* to increase the *Wheel noise* value from zero by dragging the slider to the right a little way. Alternatively, add noise in the range `0...500` using the `--motornoise / -M` magic flag. \n\nRun the program in the simulator again.\n\nYou should notice this time that the robot does not travel in a straight line. Instead, it drifts from side to side, although possibly to one side of the line.\n\nMove the robot back to the start position, or rerun the previous code cell to do so, and run the program in the simulator again. This time, you should see it follows yet another different path.\n\nDepending on how severe the noise setting is, the robot will travel closer (low noise) to the original straight line, or follow an ever-more erratic path (high noise).", "_____no_output_____" ], [ "*Record your own notes and observations here describing the behaviour of the robot for different levels of motor noise.*", "_____no_output_____" ], [ "Clear the pen traces from the simulator by running the following line magic:", "_____no_output_____" ] ], [ [ "%sim_magic -C", "_____no_output_____" ] ], [ [ "Now run the original satellite-finding dead-reckoning program again, using the *FLL_2018_Into_Orbit* background, but in the presence of *Wheel noise*. How well does it perform this time compared to previously?", "_____no_output_____" ] ], [ [ "%%sim_magic_preloaded -b FLL_2018_Into_Orbit -p -r Small_Robot\n\n# Turn on the spot to the right\ntank_turn.on_for_rotations(100, SpeedPercent(70), 1.7 )\n\n# Go forwards\ntank_drive.on_for_rotations(SpeedPercent(30), SpeedPercent(30), 20)\n\n# Slight graceful turn to left\ntank_drive.on_for_rotations(SpeedPercent(35), SpeedPercent(50), 8.5)\n\n# Turn on the spot to the left\ntank_turn.on_for_rotations(-100, SpeedPercent(75), 0.8)\n\n# Forwards a bit\ntank_drive.on_for_rotations(SpeedPercent(30), SpeedPercent(30), 2.0)\n\n# Turn on the spot a bit more to the right\ntank_turn.on_for_rotations(100, SpeedPercent(60), 0.4 )\n\n# Go forwards a bit more and dock on the satellite\ntank_drive.on_for_rotations(SpeedPercent(30), SpeedPercent(30), 1.0)\n\nsay(\"Did I avoid crashing and dock with the satellite?\")", "_____no_output_____" ] ], [ [ "Reset the robot to its original location and run the program in the simulator again. Even with the same level of motor noise as on the previous run, how does the path followed by the robot this time compare with the previous run?", "_____no_output_____" ], [ "*Add your own notes and observations here.*", "_____no_output_____" ], [ "## 2.4 Summary\n\nIn this notebook, you have seen how we can use dead reckoning to move the robot along a specified path. Using the robot’s motor speeds and by monitoring how long the motors are switched on for, we can use distance–time calculations to estimate the robot’s path. If we add in accurate measurements regarding how far we want the robot to travel, and in what direction, this provides one way of helping the robot to navigate to a particular waypoint.\n\nHowever, in the presence of noise, this approach is likely to be very unreliable: whilst the robot may think it is following one path, as determined by how long it has turned its motors on, and at what speed, it may in fact be following another path. In a real robot, the noise may be introduced in all sorts of ways, including from friction in the motor bearings, the time taken to accelerate from a standing start and get up to speed, and loss of traction effects such as wheel spin and slip as the robot’s wheels turn. \n\nWhilst in some cases it may reach the target safely, in others it may end somewhere completely different, or encounter an obstacle along the way.\n\n<!-- JD: should we say what's coming up in the next notebook? -->", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown" ]

[ [ "markdown", "markdown", "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import datetime\ndob = '20110101'\n\n\ntoday = datetime.datetime.now()\nyyyy = int(dob[0:4])\nmm = int(dob[4:6])\ndd = int(dob[6:8])\ndob = datetime.datetime(yyyy,mm,dd)\nage_in_days = (today - dob).days\nage_in_years = age_in_days/365\nprint(int(age_in_years))\n\n", "7\n" ], [ "# Formatt date\ndob1 = datetime.date(2011,1,1)\nprint(dob1)\n\nfrm = 'year {:%A, %B, %d, %Y}'\nprint(dob1.strftime('%A, %B, %d, %Y')) # old method\nprint(frm.format(dob1))\n\ndob3 = datetime.datetime.strptime('Jun 1 2005 1:33PM', '%b %d %Y %I:%M%p')\nprint(dob3)", "2011-01-01\nSaturday, January, 01, 2011\nyear Saturday, January, 01, 2011\n2005-06-01 13:33:00\n" ] ] ]

[ "code" ]

[ [ "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "## What this code does\nIn short, it is a reverse meme search, that identifies the source of the meme. It takes an image copypasta, extracts the individual *subimages* and compares it with a database of pictures (the database should be made up of copypastas, which is in TODO)\n\n### TODO\n\n### Clean up the code\nThere are many repetitive import statements. <br\\>\nThe code is saving the picture as file so that you can load it into model. <br\\>\nAnything that you cannot explain in this code <br\\>\nChange VGG16 to Xception (because I can't upgrade both TF and keras for reasons)\n\n#### Feature vector robustness check\nTo what extent the following transformations affects the feature vector?\n- crop (a little, add bounding boxes)\n- photoshop - e.g. cropping a face onto a body\n- rotate the image (a little, a lot)\n- add text (different sizes)\n- vandalised - scribbling markers over \n- add noise (Gaussian etc)\n- compression changes\n- recoloring - grey-scale\n- picture effects - e.g. twisted picture meme\n- special effects - e.g. shining eyes meme\n\n#### Image separation testing\nWe need to ensure the individual pictures are separated correctly.\n- pictures now don't have borders\n- pictures are no longer rectangular\n- whether does it identify the source of the cropped face\n\n#### Database management\nWe need to do preprocessing of the database. Currently the feature vector is only calculated when you start this notebook. \n\nMoreover, since the database of copypasta will not be single images, we need to process that aspect as well. From the copypastas we need to identify its subimages and then calculate their feature vector. There also needs to be some way to associate the feature vector and the location of the subimages to the image copypasta, together with its metadata - in a manner that is scalable.", "_____no_output_____" ], [ "### import imagenet model", "_____no_output_____" ] ], [ [ "%run image_database_helper.ipynb", "Using TensorFlow backend.\n" ], [ "model = init_model()", "_____no_output_____" ] ], [ [ "### making a list of all the files", "_____no_output_____" ] ], [ [ "!rm 'imgs/.DS_Store'", "rm: imgs/.DS_Store: No such file or directory\r\n" ], [ "images = findfiles(\"new/\")\nprint(len(images))", "24\n" ] ], [ [ "### Processing pictures", "_____no_output_____" ] ], [ [ "from PIL import Image\nfrom matplotlib.pyplot import imshow\nimport matplotlib.pyplot as plt\nimport cv2", "_____no_output_____" ], [ "import csv\n\nfieldnames = ['img_file_name',\n 'number_of_subimages',\n 'subimage_number',\n 'x',\n 'y',\n 'w',\n 'h',\n 'feature_vector']\n\nimport os\nif not os.path.exists('index_subimage.csv'):\n with open('index_subimage.csv', 'w') as csvfile:\n db = csv.DictWriter(csvfile, fieldnames=fieldnames)\n db.writeheader()", "_____no_output_____" ], [ "import subprocess\nimport csv\n\n\nfor img_name in images:\n path_image_to_analyse = \"./new/\"+img_name\n print(img_name)\n \n img = cv2.imread(path_image_to_analyse)\n output_boxes = get_bounding_boxes(img)\n\n for i, box in enumerate(output_boxes):\n [x,y,w,h] = box\n output_img = np.array(img[y:y+h, x:x+w])\n cv2.imwrite(\"temp.jpg\",output_img)\n feature_vector = calc_feature_vector(model, \"temp.jpg\")\n\n dict_to_write = {'img_file_name':img_name,\n 'number_of_subimages':len(output_boxes),\n 'subimage_number':i,\n 'x':x,\n 'y':y,\n 'w':w,\n 'h':h,\n 'feature_vector':feature_vector}\n\n with open('index_subimage.csv', 'a') as csvfile:\n db = csv.DictWriter(csvfile, fieldnames=fieldnames)\n db.writerow(dict_to_write)\n \n subprocess.run(\"mv ./new/{} ./database/{}\".format(img_name,img_name),shell=True)", "our-singapore-malaysian-durian-fake-768x791.jpg\ndog-meat-satay-post.jpg\nScreen Shot 2018-04-02 at 05.19.52 AM.png\nScreen-Shot-2017-10-12-at-5.33.37-PM.png\nge14fakenews.png\npm-lee-gst-meme.jpg\ncasket-1-768x1366.jpg\n17626490_10155249769684175_8841223787064818741_n.jpg\ncity-harvest-church-headline-changed.jpg\n17951666_1758817064340674_3343590308811084675_n-768x1365.jpg\nsexdoll-768x603.jpg\npmo-lky-hoax.jpg\nBear-hoax-singapore-768x432.jpg\nnotice-sammyboy-e1520150158500-768x620.jpg\nntucfairpricewarning.jpg\nIMG_7EF6C6D23BCD-1-768x1366.jpg\nntuc27e_2x.jpg\nflyer-breast-cancer-yishun-fake.jpg\nPunggo_Waterway_Terraces_roof_top_floors_collapse_1-768x801.jpg\ncross-with-phone-fine-fake-768x930.jpg\nfake copy.jpg\nScreen-Shot-2018-02-09-at-10.26.30.png\nASS-Prakash-2-768x787.jpg\nnude-2.jpg\n" ], [ "!cp ./database/* ./new/", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Computer Vision Nanodegree\n\n## Project: Image Captioning\n\n---\n\nIn this notebook, you will use your trained model to generate captions for images in the test dataset.\n\nThis notebook **will be graded**. \n\nFeel free to use the links below to navigate the notebook:\n- [Step 1](#step1): Get Data Loader for Test Dataset \n- [Step 2](#step2): Load Trained Models\n- [Step 3](#step3): Finish the Sampler\n- [Step 4](#step4): Clean up Captions\n- [Step 5](#step5): Generate Predictions!", "_____no_output_____" ], [ "<a id='step1'></a>\n## Step 1: Get Data Loader for Test Dataset\n\nBefore running the code cell below, define the transform in `transform_test` that you would like to use to pre-process the test images. \n\nMake sure that the transform that you define here agrees with the transform that you used to pre-process the training images (in **2_Training.ipynb**). For instance, if you normalized the training images, you should also apply the same normalization procedure to the test images.", "_____no_output_____" ] ], [ [ "import sys\nsys.path.append('/opt/cocoapi/PythonAPI')\nfrom pycocotools.coco import COCO\nfrom data_loader import get_loader\nfrom torchvision import transforms\n\n# TODO #1: Define a transform to pre-process the testing images.\ntransform_test = transforms.Compose([ \n transforms.Resize(256), # smaller edge of image resized to 256\n transforms.RandomCrop(224), # get 224x224 crop from random location\n transforms.RandomHorizontalFlip(), # horizontally flip image with probability=0.5\n transforms.ToTensor(), # convert the PIL Image to a tensor\n transforms.Normalize((0.485, 0.456, 0.406), # normalize image for pre-trained model\n (0.229, 0.224, 0.225))])\n\n#-#-#-# Do NOT modify the code below this line. #-#-#-#\n\n# Create the data loader.\ndata_loader = get_loader(transform=transform_test, \n mode='test')", "Vocabulary successfully loaded from vocab.pkl file!\n" ] ], [ [ "Run the code cell below to visualize an example test image, before pre-processing is applied.", "_____no_output_____" ] ], [ [ "import numpy as np\nimport matplotlib.pyplot as plt\n%matplotlib inline\n\n# Obtain sample image before and after pre-processing.\norig_image, image = next(iter(data_loader))\n\n# Visualize sample image, before pre-processing.\nplt.imshow(np.squeeze(orig_image))\nplt.title('example image')\nplt.show()", "_____no_output_____" ] ], [ [ "<a id='step2'></a>\n## Step 2: Load Trained Models\n\nIn the next code cell we define a `device` that you will use move PyTorch tensors to GPU (if CUDA is available). Run this code cell before continuing.", "_____no_output_____" ] ], [ [ "import torch\n\ndevice = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")", "_____no_output_____" ] ], [ [ "Before running the code cell below, complete the following tasks.\n\n### Task #1\n\nIn the next code cell, you will load the trained encoder and decoder from the previous notebook (**2_Training.ipynb**). To accomplish this, you must specify the names of the saved encoder and decoder files in the `models/` folder (e.g., these names should be `encoder-5.pkl` and `decoder-5.pkl`, if you trained the model for 5 epochs and saved the weights after each epoch). \n\n### Task #2\n\nPlug in both the embedding size and the size of the hidden layer of the decoder corresponding to the selected pickle file in `decoder_file`.", "_____no_output_____" ] ], [ [ "# Watch for any changes in model.py, and re-load it automatically.\n% load_ext autoreload\n% autoreload 2\n\nimport os\nimport torch\nfrom model import EncoderCNN, DecoderRNN\n\n# TODO #2: Specify the saved models to load.\nencoder_file = \"encoder-1.pkl\" \ndecoder_file = \"decoder-1.pkl\"\n\n# TODO #3: Select appropriate values for the Python variables below.\nembed_size = 256 #512 #300\nhidden_size = 512\n\n# The size of the vocabulary.\nvocab_size = len(data_loader.dataset.vocab)\n\n# Initialize the encoder and decoder, and set each to inference mode.\nencoder = EncoderCNN(embed_size)\nencoder.eval()\ndecoder = DecoderRNN(embed_size, hidden_size, vocab_size)\ndecoder.eval()\n\n# Load the trained weights.\nencoder.load_state_dict(torch.load(os.path.join('./models', encoder_file)))\ndecoder.load_state_dict(torch.load(os.path.join('./models', decoder_file)))\n\n# Move models to GPU if CUDA is available.\nencoder.to(device)\ndecoder.to(device)", "Downloading: \"https://download.pytorch.org/models/resnet50-19c8e357.pth\" to /root/.torch/models/resnet50-19c8e357.pth\n100%|██████████| 102502400/102502400 [00:01<00:00, 58944421.27it/s]\n" ] ], [ [ "<a id='step3'></a>\n## Step 3: Finish the Sampler\n\nBefore executing the next code cell, you must write the `sample` method in the `DecoderRNN` class in **model.py**. This method should accept as input a PyTorch tensor `features` containing the embedded input features corresponding to a single image.\n\nIt should return as output a Python list `output`, indicating the predicted sentence. `output[i]` is a nonnegative integer that identifies the predicted `i`-th token in the sentence. The correspondence between integers and tokens can be explored by examining either `data_loader.dataset.vocab.word2idx` (or `data_loader.dataset.vocab.idx2word`).\n\nAfter implementing the `sample` method, run the code cell below. If the cell returns an assertion error, then please follow the instructions to modify your code before proceeding. Do **not** modify the code in the cell below. ", "_____no_output_____" ] ], [ [ "# Move image Pytorch Tensor to GPU if CUDA is available.\nimage = image.to(device)\n\n# Obtain the embedded image features.\nfeatures = encoder(image).unsqueeze(1)\n\n# Pass the embedded image features through the model to get a predicted caption.\noutput = decoder.sample(features)\nprint('example output:', output)\n\nassert (type(output)==list), \"Output needs to be a Python list\" \nassert all([type(x)==int for x in output]), \"Output should be a list of integers.\" \nassert all([x in data_loader.dataset.vocab.idx2word for x in output]), \"Each entry in the output needs to correspond to an integer that indicates a token in the vocabulary.\"", "example output: [0, 3, 2436, 170, 77, 3, 204, 21, 3, 769, 77, 32, 297, 18, 1, 1, 18, 1, 1, 18]\n" ] ], [ [ "<a id='step4'></a>\n## Step 4: Clean up the Captions\n\nIn the code cell below, complete the `clean_sentence` function. It should take a list of integers (corresponding to the variable `output` in **Step 3**) as input and return the corresponding predicted sentence (as a single Python string). ", "_____no_output_____" ] ], [ [ "# TODO #4: Complete the function.\ndef clean_sentence(output):\n \n seperator = \" \"\n word_list = [];\n for word_index in output:\n if word_index not in [0,2]: # 0: '<start>', 1: '<end>', 2: '<unk>', 18: '.'\n if word_index == 1:\n break\n word = data_loader.dataset.vocab.idx2word[word_index]\n word_list.append(word)\n \n sentence = seperator.join(word_list) \n return sentence ", "_____no_output_____" ] ], [ [ "After completing the `clean_sentence` function above, run the code cell below. If the cell returns an assertion error, then please follow the instructions to modify your code before proceeding.", "_____no_output_____" ] ], [ [ "sentence = clean_sentence(output)\nprint('example sentence:', sentence)\n\nassert type(sentence)==str, 'Sentence needs to be a Python string!'", "example sentence: a giraffe standing in a field with a tree in the background .\n" ] ], [ [ "<a id='step5'></a>\n## Step 5: Generate Predictions!\n\nIn the code cell below, we have written a function (`get_prediction`) that you can use to use to loop over images in the test dataset and print your model's predicted caption.", "_____no_output_____" ] ], [ [ "def get_prediction():\n orig_image, image = next(iter(data_loader))\n plt.imshow(np.squeeze(orig_image))\n plt.title('Sample Image')\n plt.show()\n image = image.to(device)\n features = encoder(image).unsqueeze(1)\n output = decoder.sample(features) \n sentence = clean_sentence(output)\n print(sentence)", "_____no_output_____" ] ], [ [ "Run the code cell below (multiple times, if you like!) to test how this function works.", "_____no_output_____" ] ], [ [ "get_prediction()", "_____no_output_____" ] ], [ [ "As the last task in this project, you will loop over the images until you find four image-caption pairs of interest:\n- Two should include image-caption pairs that show instances when the model performed well.\n- Two should highlight image-caption pairs that highlight instances where the model did not perform well.\n\nUse the four code cells below to complete this task.", "_____no_output_____" ], [ "### The model performed well!\n\nUse the next two code cells to loop over captions. Save the notebook when you encounter two images with relatively accurate captions.", "_____no_output_____" ] ], [ [ "get_prediction()", "_____no_output_____" ], [ "get_prediction()", "_____no_output_____" ] ], [ [ "### The model could have performed better ...\n\nUse the next two code cells to loop over captions. Save the notebook when you encounter two images with relatively inaccurate captions.", "_____no_output_____" ] ], [ [ "get_prediction()", "_____no_output_____" ], [ "get_prediction()", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Purpose: To run the full segmentation using the best scored method from 2_compare_auto_to_manual_threshold", "_____no_output_____" ], [ "Date Created: January 7, 2022", "_____no_output_____" ], [ "Dates Edited: January 26, 2022 - changed the ogd severity study to be the otsu data as the yen data did not run on all samples.", "_____no_output_____" ], [ "*Step 1: Import Necessary Packages*", "_____no_output_____" ] ], [ [ "# import major packages\nimport numpy as np\nimport matplotlib.pyplot as plt\nimport skimage\nimport PIL as Image\nimport os\nimport pandas as pd\n\n# import specific package functions\nfrom skimage.filters import threshold_otsu\nfrom skimage import morphology\nfrom scipy import ndimage\nfrom skimage.measure import label\nfrom skimage import io\nfrom skimage import measure", "_____no_output_____" ] ], [ [ "__OGD Severity Study__", "_____no_output_____" ] ], [ [ "im_folder_location = '/Users/hhelmbre/Desktop/ogd_severity_undergrad/10_4_21_redownload/'", "_____no_output_____" ], [ "im_paths = []\nfiles = []\nfor file in os.listdir(im_folder_location):\n if file.endswith(\".tif\"):\n file_name = os.path.join(im_folder_location, file)\n files.append(file)\n im_paths.append(file_name)", "_____no_output_____" ], [ "files", "_____no_output_____" ], [ "properties_list = ('area', 'bbox_area', 'centroid', 'convex_area', \n 'eccentricity', 'equivalent_diameter', 'euler_number', \n 'extent', 'filled_area', 'major_axis_length', \n 'minor_axis_length', 'orientation', 'perimeter', 'solidity')", "_____no_output_____" ], [ "source_dir = '/Users/hhelmbre/Desktop/microfiber/ogd_severity_segmentations/'", "_____no_output_____" ], [ "j = 0\nfor image in im_paths:\n \n short_im_name = image.rsplit('/', 1)\n short_im_name = short_im_name[1]\n \n im = io.imread(image)\n microglia_im = im[:,:,1]\n #otsu threshold\n thresh_otsu = skimage.filters.threshold_otsu(microglia_im)\n binary_otsu = microglia_im > thresh_otsu\n new_binary_otsu = morphology.remove_small_objects(binary_otsu, min_size=71)\n new_binary_otsu = ndimage.binary_fill_holes(new_binary_otsu)\n label_image = measure.label(new_binary_otsu)\n props = measure.regionprops_table(label_image, properties=(properties_list))\n \n np.save(str(source_dir + short_im_name[:-4] + '_otsu_thresh'), new_binary_otsu)\n\n if j == 0:\n df = pd.DataFrame(props)\n df['filename'] = image\n else:\n df2 = pd.DataFrame(props)\n df2['filename'] = image\n df = df.append(df2)\n\n j = 1\n \n ", "_____no_output_____" ], [ "df['circularity'] = 4*np.pi*df.area/df.perimeter**2", "_____no_output_____" ], [ "df['aspect_ratio'] = df.major_axis_length/df.minor_axis_length", "_____no_output_____" ], [ "df", "_____no_output_____" ], [ "df.to_csv('/Users/hhelmbre/Desktop/microfiber/ogd_severity_study_features_otsu.csv' )", "_____no_output_____" ], [ "%load_ext watermark\n\n%watermark -v -m -p numpy,pandas,scipy,skimage,matplotlib,wget\n\n%watermark -u -n -t -z", "Python implementation: CPython\nPython version : 3.7.4\nIPython version : 7.8.0\n\nnumpy : 1.21.5\npandas : 1.3.5\nscipy : 1.3.1\nskimage : 0.17.2\nmatplotlib: 3.1.1\nwget : 3.2\n\nCompiler : Clang 4.0.1 (tags/RELEASE_401/final)\nOS : Darwin\nRelease : 20.6.0\nMachine : x86_64\nProcessor : i386\nCPU cores : 8\nArchitecture: 64bit\n\nLast updated: Wed Jan 26 2022 14:39:49PST\n\n" ] ] ]

[ "markdown", "code", "markdown", "code" ]

[ [ "markdown", "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# TensorFlow pix2pix implementation", "_____no_output_____" ], [ "from __future__ import absolute_import, division, print_function, unicode_literals\n\ntry:\n # %tensorflow_version only exists in Colab.\n %tensorflow_version 2.x\nexcept Exception:\n pass\nimport tensorflow as tf\n\nimport os\nimport time\n\nfrom matplotlib import pyplot as plt\nfrom IPython import display", "_____no_output_____" ], [ "print(\"Num GPUs Available: \", len(tf.config.experimental.list_physical_devices('GPU')))", "Num GPUs Available: 0\n" ], [ "PATH = \"/Volumes/Data/projects/cs230/Project/RenderGAN/pix2pix/data/train_data/10-10000/AB/\"\n\nBUFFER_SIZE = 400\nBATCH_SIZE = 1\nIMG_WIDTH = 256\nIMG_HEIGHT = 256", "_____no_output_____" ], [ "def load(image_file):\n image = tf.io.read_file(image_file)\n image = tf.image.decode_png(image)\n\n w = tf.shape(image)[1]\n\n w = w // 2\n real_image = image[:, w:, :]\n input_image = image[:, :w, :]\n\n input_image = tf.cast(input_image, tf.float32)\n real_image = tf.cast(real_image, tf.float32)\n\n return input_image, real_image", "_____no_output_____" ], [ "inp, re = load(PATH+'train/8.png')\n# casting to int for matplotlib to show the image\nplt.figure()\nplt.imshow(inp/255.0)\nplt.figure()\nplt.imshow(re/255.0)", "_____no_output_____" ], [ "def resize(input_image, real_image, height, width):\n input_image = tf.image.resize(input_image, [height, width],\n method=tf.image.ResizeMethod.NEAREST_NEIGHBOR)\n real_image = tf.image.resize(real_image, [height, width],\n method=tf.image.ResizeMethod.NEAREST_NEIGHBOR)\n\n return input_image, real_image\n\ndef random_crop(input_image, real_image):\n stacked_image = tf.stack([input_image, real_image], axis=0)\n cropped_image = tf.image.random_crop(\n stacked_image, size=[2, IMG_HEIGHT, IMG_WIDTH, 3])\n\n return cropped_image[0], cropped_image[1]\n\n# normalizing the images to [-1, 1]\n\ndef normalize(input_image, real_image):\n input_image = (input_image / 127.5) - 1\n real_image = (real_image / 127.5) - 1\n\n return input_image, real_image\n\n@tf.function()\ndef random_jitter(input_image, real_image):\n # resizing to 286 x 286 x 3\n input_image, real_image = resize(input_image, real_image, 286, 286)\n\n # randomly cropping to 256 x 256 x 3\n input_image, real_image = random_crop(input_image, real_image)\n\n if tf.random.uniform(()) > 0.5:\n # random mirroring\n input_image = tf.image.flip_left_right(input_image)\n real_image = tf.image.flip_left_right(real_image)\n\n return input_image, real_image\n", "_____no_output_____" ], [ "plt.figure(figsize=(6, 6))\nfor i in range(4):\n rj_inp, rj_re = random_jitter(inp, re)\n plt.subplot(2, 2, i+1)\n plt.imshow(rj_inp/255.0)\n plt.axis('off')\nplt.show()", "_____no_output_____" ], [ "def load_image_train(image_file):\n input_image, real_image = load(image_file)\n input_image, real_image = random_jitter(input_image, real_image)\n input_image, real_image = normalize(input_image, real_image)\n\n return input_image, real_image", "_____no_output_____" ], [ "def load_image_test(image_file):\n input_image, real_image = load(image_file)\n input_image, real_image = resize(input_image, real_image,\n IMG_HEIGHT, IMG_WIDTH)\n input_image, real_image = normalize(input_image, real_image)\n\n return input_image, real_image", "_____no_output_____" ], [ "train_dataset = tf.data.Dataset.list_files(PATH+'train/*.png')\ntrain_dataset = train_dataset.map(load_image_train,\n num_parallel_calls=tf.data.experimental.AUTOTUNE)\ntrain_dataset = train_dataset.shuffle(BUFFER_SIZE)\ntrain_dataset = train_dataset.batch(BATCH_SIZE)", "_____no_output_____" ], [ "test_dataset = tf.data.Dataset.list_files(PATH+'test/*.png')\ntest_dataset = test_dataset.map(load_image_test)\ntest_dataset = test_dataset.batch(BATCH_SIZE)", "_____no_output_____" ], [ "OUTPUT_CHANNELS = 3", "_____no_output_____" ], [ "def downsample(filters, size, apply_batchnorm=True):\n initializer = tf.random_normal_initializer(0., 0.02)\n\n result = tf.keras.Sequential()\n result.add(\n tf.keras.layers.Conv2D(filters, size, strides=2, padding='same',\n kernel_initializer=initializer, use_bias=False))\n\n if apply_batchnorm:\n result.add(tf.keras.layers.BatchNormalization())\n\n result.add(tf.keras.layers.LeakyReLU())\n\n return result", "_____no_output_____" ], [ "down_model = downsample(3, 4)\ndown_result = down_model(tf.expand_dims(inp, 0))\nprint (down_result.shape)", "(1, 256, 256, 3)\n" ], [ "def upsample(filters, size, apply_dropout=False):\n initializer = tf.random_normal_initializer(0., 0.02)\n\n result = tf.keras.Sequential()\n result.add(\n tf.keras.layers.Conv2DTranspose(filters, size, strides=2,\n padding='same',\n kernel_initializer=initializer,\n use_bias=False))\n\n result.add(tf.keras.layers.BatchNormalization())\n\n if apply_dropout:\n result.add(tf.keras.layers.Dropout(0.5))\n\n result.add(tf.keras.layers.ReLU())\n\n return result", "_____no_output_____" ], [ "up_model = upsample(3, 4)\nup_result = up_model(down_result)\nprint (up_result.shape)", "(1, 512, 512, 3)\n" ], [ "def Generator():\n inputs = tf.keras.layers.Input(shape=[256,256,3])\n\n down_stack = [\n downsample(64, 4, apply_batchnorm=False), # (bs, 128, 128, 64)\n downsample(128, 4), # (bs, 64, 64, 128)\n downsample(256, 4), # (bs, 32, 32, 256)\n downsample(512, 4), # (bs, 16, 16, 512)\n downsample(512, 4), # (bs, 8, 8, 512)\n downsample(512, 4), # (bs, 4, 4, 512)\n downsample(512, 4), # (bs, 2, 2, 512)\n downsample(512, 4), # (bs, 1, 1, 512)\n ]\n\n up_stack = [\n upsample(512, 4, apply_dropout=True), # (bs, 2, 2, 1024)\n upsample(512, 4, apply_dropout=True), # (bs, 4, 4, 1024)\n upsample(512, 4, apply_dropout=True), # (bs, 8, 8, 1024)\n upsample(512, 4), # (bs, 16, 16, 1024)\n upsample(256, 4), # (bs, 32, 32, 512)\n upsample(128, 4), # (bs, 64, 64, 256)\n upsample(64, 4), # (bs, 128, 128, 128)\n ]\n\n initializer = tf.random_normal_initializer(0., 0.02)\n last = tf.keras.layers.Conv2DTranspose(OUTPUT_CHANNELS, 4,\n strides=2,\n padding='same',\n kernel_initializer=initializer,\n activation='tanh') # (bs, 256, 256, 3)\n\n x = inputs\n\n # Downsampling through the model\n skips = []\n for down in down_stack:\n x = down(x)\n skips.append(x)\n\n skips = reversed(skips[:-1])\n\n # Upsampling and establishing the skip connections\n for up, skip in zip(up_stack, skips):\n x = up(x)\n x = tf.keras.layers.Concatenate()([x, skip])\n\n x = last(x)\n\n return tf.keras.Model(inputs=inputs, outputs=x)", "_____no_output_____" ], [ "generator = Generator()\ntf.keras.utils.plot_model(generator, show_shapes=True, dpi=64)", "_____no_output_____" ], [ "gen_output = generator(inp[tf.newaxis,...], training=False)\nplt.imshow(gen_output[0,...])", "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n" ], [ "LAMBDA = 100", "_____no_output_____" ], [ "def generator_loss(disc_generated_output, gen_output, target):\n gan_loss = loss_object(tf.ones_like(disc_generated_output), disc_generated_output)\n\n # mean absolute error\n l1_loss = tf.reduce_mean(tf.abs(target - gen_output))\n\n total_gen_loss = gan_loss + (LAMBDA * l1_loss)\n\n return total_gen_loss, gan_loss, l1_loss", "_____no_output_____" ], [ "def Discriminator():\n initializer = tf.random_normal_initializer(0., 0.02)\n\n inp = tf.keras.layers.Input(shape=[256, 256, 3], name='input_image')\n tar = tf.keras.layers.Input(shape=[256, 256, 3], name='target_image')\n\n x = tf.keras.layers.concatenate([inp, tar]) # (bs, 256, 256, channels*2)\n\n down1 = downsample(64, 4, False)(x) # (bs, 128, 128, 64)\n down2 = downsample(128, 4)(down1) # (bs, 64, 64, 128)\n down3 = downsample(256, 4)(down2) # (bs, 32, 32, 256)\n\n zero_pad1 = tf.keras.layers.ZeroPadding2D()(down3) # (bs, 34, 34, 256)\n conv = tf.keras.layers.Conv2D(512, 4, strides=1,\n kernel_initializer=initializer,\n use_bias=False)(zero_pad1) # (bs, 31, 31, 512)\n\n batchnorm1 = tf.keras.layers.BatchNormalization()(conv)\n\n leaky_relu = tf.keras.layers.LeakyReLU()(batchnorm1)\n\n zero_pad2 = tf.keras.layers.ZeroPadding2D()(leaky_relu) # (bs, 33, 33, 512)\n\n last = tf.keras.layers.Conv2D(1, 4, strides=1,\n kernel_initializer=initializer)(zero_pad2) # (bs, 30, 30, 1)\n\n return tf.keras.Model(inputs=[inp, tar], outputs=last)", "_____no_output_____" ], [ "discriminator = Discriminator()\ntf.keras.utils.plot_model(discriminator, show_shapes=True, dpi=64)", "_____no_output_____" ], [ "disc_out = discriminator([inp[tf.newaxis,...], gen_output], training=False)\nplt.imshow(disc_out[0,...,-1], vmin=-20, vmax=20, cmap='RdBu_r')\nplt.colorbar()", "_____no_output_____" ], [ "loss_object = tf.keras.losses.BinaryCrossentropy(from_logits=True)", "_____no_output_____" ], [ "def discriminator_loss(disc_real_output, disc_generated_output):\n real_loss = loss_object(tf.ones_like(disc_real_output), disc_real_output)\n\n generated_loss = loss_object(tf.zeros_like(disc_generated_output), disc_generated_output)\n\n total_disc_loss = real_loss + generated_loss\n\n return total_disc_loss", "_____no_output_____" ], [ "generator_optimizer = tf.keras.optimizers.Adam(2e-4, beta_1=0.5)\ndiscriminator_optimizer = tf.keras.optimizers.Adam(2e-4, beta_1=0.5)", "_____no_output_____" ], [ "checkpoint_dir = './training_checkpoints'\ncheckpoint_prefix = os.path.join(checkpoint_dir, \"ckpt\")\ncheckpoint = tf.train.Checkpoint(generator_optimizer=generator_optimizer,\n discriminator_optimizer=discriminator_optimizer,\n generator=generator,\n discriminator=discriminator)", "_____no_output_____" ], [ "def generate_images(model, test_input, tar):\n prediction = model(test_input, training=True)\n plt.figure(figsize=(15,15))\n\n display_list = [test_input[0], tar[0], prediction[0]]\n title = ['Input Image', 'Ground Truth', 'Predicted Image']\n\n for i in range(3):\n plt.subplot(1, 3, i+1)\n plt.title(title[i])\n # getting the pixel values between [0, 1] to plot it.\n plt.imshow(display_list[i] * 0.5 + 0.5)\n plt.axis('off')\n plt.show()", "_____no_output_____" ], [ "for example_input, example_target in test_dataset.take(1):\n generate_images(generator, example_input, example_target)", "_____no_output_____" ], [ "EPOCHS = 10", "_____no_output_____" ], [ "import datetime\nlog_dir=\"logs/\"\n\nsummary_writer = tf.summary.create_file_writer(\n log_dir + \"fit/\" + datetime.datetime.now().strftime(\"%Y%m%d-%H%M%S\"))", "_____no_output_____" ], [ "@tf.function\ndef train_step(input_image, target, epoch):\n with tf.GradientTape() as gen_tape, tf.GradientTape() as disc_tape:\n gen_output = generator(input_image, training=True)\n\n disc_real_output = discriminator([input_image, target], training=True)\n disc_generated_output = discriminator([input_image, gen_output], training=True)\n\n gen_total_loss, gen_gan_loss, gen_l1_loss = generator_loss(disc_generated_output, gen_output, target)\n disc_loss = discriminator_loss(disc_real_output, disc_generated_output)\n\n generator_gradients = gen_tape.gradient(gen_total_loss,\n generator.trainable_variables)\n discriminator_gradients = disc_tape.gradient(disc_loss,\n discriminator.trainable_variables)\n\n generator_optimizer.apply_gradients(zip(generator_gradients,\n generator.trainable_variables))\n discriminator_optimizer.apply_gradients(zip(discriminator_gradients,\n discriminator.trainable_variables))\n\n with summary_writer.as_default():\n tf.summary.scalar('gen_total_loss', gen_total_loss, step=epoch)\n tf.summary.scalar('gen_gan_loss', gen_gan_loss, step=epoch)\n tf.summary.scalar('gen_l1_loss', gen_l1_loss, step=epoch)\n tf.summary.scalar('disc_loss', disc_loss, step=epoch)", "_____no_output_____" ], [ "def fit(train_ds, epochs, test_ds):\n for epoch in range(epochs):\n start = time.time()\n\n display.clear_output(wait=True)\n\n for example_input, example_target in test_ds.take(1):\n generate_images(generator, example_input, example_target)\n print(\"Epoch: \", epoch)\n\n # Train\n for n, (input_image, target) in train_ds.enumerate():\n print('.', end='')\n if (n+1) % 100 == 0:\n print()\n train_step(input_image, target, epoch)\n print()\n\n # saving (checkpoint) the model every 20 epochs\n if (epoch + 1) % 20 == 0:\n checkpoint.save(file_prefix = checkpoint_prefix)\n\n print ('Time taken for epoch {} is {} sec\\n'.format(epoch + 1,\n time.time()-start))\n checkpoint.save(file_prefix = checkpoint_prefix)", "_____no_output_____" ], [ "%load_ext tensorboard\n%tensorboard --logdir {log_dir}", "The tensorboard extension is already loaded. To reload it, use:\n %reload_ext tensorboard\n" ], [ "fit(train_dataset, EPOCHS, test_dataset)", "_____no_output_____" ], [ "!ls {checkpoint_dir}", "ls: ./training_checkpoints: No such file or directory\r\n" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Basic Workflow", "_____no_output_____" ] ], [ [ "# Always have your imports at the top\nimport pandas as pd\nfrom sklearn.pipeline import make_pipeline\nfrom sklearn.impute import SimpleImputer\nfrom sklearn.ensemble import RandomForestClassifier\nfrom sklearn.base import TransformerMixin\n\nfrom hashlib import sha1 # just for grading purposes\nimport json # just for grading purposes\n\ndef _hash(obj, salt='none'):\n if type(obj) is not str:\n obj = json.dumps(obj)\n to_encode = obj + salt\n return sha1(to_encode.encode()).hexdigest()", "_____no_output_____" ] ], [ [ "# Workflow steps\n\nWhat are the basic workflow steps?\n\nIt's incredibly obvious what the steps are since you can see them graded in plain text. However we deem it worth actually making you type each one of the steps and take a moment to think about it and internalize them.\n\nPlease do actually type them rather than just copy-pasting as fast as you can. Type it out character by character and internalize.", "_____no_output_____" ] ], [ [ "# step_1 = ...\n# step_2 = ...\n# step_2_a = ...\n# step_2_b = ...\n# step_2_c = ...\n# step_2_d = ...\n# step_3 = ...\n# step_4 = ...\n# step_5 = ...\n\n# YOUR CODE HERE\nraise NotImplementedError()\n", "_____no_output_____" ], [ "### BEGIN TESTS\nassert step_1 == 'Get the data'\nassert step_2 == 'Data analysis and preparation'\nassert step_2_a == 'Data analysis'\nassert step_2_b == 'Dealing with data problems'\nassert step_2_c == 'Feature engineering'\nassert step_2_d == 'Feature selection'\nassert step_3 == 'Train model'\nassert step_4 == 'Evaluate results'\nassert step_5 == 'Iterate'\n### END TESTS", "_____no_output_____" ] ], [ [ "# Specific workflow questions\n\nHere are some more specific questions about individual workflow steps.", "_____no_output_____" ] ], [ [ "# True or False, it's super easy to gather your dataset in a production environment\n# real_world_dataset_gathering_easy = ...\n\n# True or False, it's super easy to gather your dataset in the context of the academy\n# academy_dataset_gathering_easy = ...\n\n# True or False, you should try as hard as you can to get the best possible score\n# on your test set by iterating until you can't get your test set score any higher\n# by any means possible\n# test_set_optimization_is_good = ...\n\n# True or False, you should choose one metric by which to evaluate your model and\n# never consider using another one\n# one_metric_should_rule_them_all = ...\n\n# YOUR CODE HERE\nraise NotImplementedError()", "_____no_output_____" ], [ "### BEGIN TESTS\nassert _hash(real_world_dataset_gathering_easy, 'salt1') == '63b5b9a8f2d359e1fc175c3b01b907ef87590484'\nassert _hash(academy_dataset_gathering_easy, 'salt2') == 'dd7dee495a153c95d28c7aa95289c0415242f5d8'\nassert _hash(test_set_optimization_is_good, 'salt3') == 'f24a294afb4a09f7f9df9ee13eb18e7d341c439d'\nassert _hash(one_metric_should_rule_them_all, 'salt4') == '2360691a582e4f0fbefa238ab6ced1cbfbfe8a50'\n### END TESTS", "_____no_output_____" ] ], [ [ "# scikit pipelines\n\nMake a simple pipeline that\n1. Drops all columns that start with the string `evil`\n1. Fills all nulls with the median", "_____no_output_____" ] ], [ [ "# Create a pipeline step called RemoveEvilColumns the removed any\n# column whose name starts with the string 'evil'\n\n# YOUR CODE HERE\nraise NotImplementedError()\n\n\n# Create an pipeline using make_pipeline\n# 1. removes evil columns\n# 2. imputes with the mean\n# 3. has a random forest classifier as the last step\n\n# YOUR CODE HERE\nraise NotImplementedError()", "_____no_output_____" ], [ "X = pd.DataFrame({\n 'evil_1': ['a'] * 100,\n 'evil_2': ['b'] * 100,\n 'not_so_evil': list(range(0, 100))\n})\ny = pd.Series([x % 2 for x in range(0, 100)])\n\npipeline.fit(X, y)\n\n### BEGIN TESTS\nassert pipeline.steps[0][0] == 'removeevilcolumns', pipeline.steps[0][0]\nassert pipeline.steps[1][0] == 'simpleimputer', pipeline.steps[1][0]\nassert pipeline.steps[2][0] == 'randomforestclassifier', pipeline.steps[2][0]\n### END TESTS", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ] ]

[ "BSD-3-Clause" ]

[ "BSD-3-Clause" ]

[ "BSD-3-Clause" ]

[ [ [ "import geopandas", "_____no_output_____" ], [ "help(geopandas)", "Help on package geopandas:\n\nNAME\n geopandas\n\nPACKAGE CONTENTS\n _compat\n _config\n _vectorized\n _version\n array\n base\n conftest\n datasets (package)\n geodataframe\n geoseries\n io (package)\n plotting\n sindex\n testing\n tests (package)\n tools (package)\n\nDATA\n options = Options(\n display_precision: None [default: Non...USE_PYGEO...\n\nVERSION\n 0.9.0\n\nFILE\n /home/daniel/.local/lib/python3.9/site-packages/geopandas/__init__.py\n\n\n" ], [ "mosquito_points = geopandas.read_file(\"../../data/interim/mosquito_points.shp\")\n", "_____no_output_____" ], [ "mosquito_points", "_____no_output_____" ], [ "mosquito_points.plot", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "<a href=\"https://colab.research.google.com/github/Sarbajit097/Assignment/blob/main/Assignment_3.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>", "_____no_output_____" ] ], [ [ "import pandas as pd\npath=\"https://raw.githubusercontent.com/Sarbajit097/Assignment/main/Toyota.csv\"\ndata =pd.read_csv(path)\ndata ", "_____no_output_____" ], [ "type(data)", "_____no_output_____" ], [ "data.shape", "_____no_output_____" ], [ "data.info()", "<class 'pandas.core.frame.DataFrame'>\nRangeIndex: 1436 entries, 0 to 1435\nData columns (total 11 columns):\n # Column Non-Null Count Dtype \n--- ------ -------------- ----- \n 0 Unnamed: 0 1436 non-null int64 \n 1 Price 1436 non-null int64 \n 2 Age 1336 non-null float64\n 3 KM 1436 non-null object \n 4 FuelType 1336 non-null object \n 5 HP 1436 non-null object \n 6 MetColor 1286 non-null float64\n 7 Automatic 1436 non-null int64 \n 8 CC 1436 non-null int64 \n 9 Doors 1436 non-null object \n 10 Weight 1436 non-null int64 \ndtypes: float64(2), int64(5), object(4)\nmemory usage: 123.5+ KB\n" ], [ "data.index", "_____no_output_____" ], [ "data.columns", "_____no_output_____" ], [ "data.head()", "_____no_output_____" ], [ "data.tail()", "_____no_output_____" ], [ "data.head(5)", "_____no_output_____" ], [ "data[['Price',\"Age\"]].head(10)", "_____no_output_____" ], [ "data.isnull().sum()", "_____no_output_____" ], [ "data.dropna(inplace=True)\ndata.isnull().sum()", "_____no_output_____" ], [ "data.shape", "_____no_output_____" ], [ "data.head(10)", "_____no_output_____" ], [ "data['MetColor'].mean()", "_____no_output_____" ], [ "data['MetColor'].head()", "_____no_output_____" ], [ "import numpy as np\ndata['MetColor'].replace(np.NaN,data['MetColor'].mean()).head()", "_____no_output_____" ], [ "data.head(10)", "_____no_output_____" ], [ "data['CC'].mean()", "_____no_output_____" ], [ "data['CC'].head()", "_____no_output_____" ], [ "data[['Age',\"KM\"]].head(20)", "_____no_output_____" ], [ "", "_____no_output_____" ] ] ]

[ "markdown", "code" ]

[ [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "BSD-3-Clause" ]

[ "BSD-3-Clause" ]

[ "BSD-3-Clause" ]

[ [ [ "empty" ] ] ]

[ "empty" ]

[ [ "empty" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "<a href=\"https://colab.research.google.com/github/kathleenmei/CPEN-21A-ECE-2-1/blob/main/Loop_Statement.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>", "_____no_output_____" ], [ "##For Loop\n", "_____no_output_____" ] ], [ [ "week = [\"Sunday\",\"Monday\",\"Tuesday\",\"Wednesday\",\"Thursday\",\"Friday\",\"Saturday\"]\nfor x in week:\n print(x)", "Sunday\nMonday\nTuesday\nWednesday\nThursday\nFriday\nSaturday\n" ] ], [ [ "The Break Statement", "_____no_output_____" ] ], [ [ "week = [\"Sunday\",\"Monday\",\"Tuesday\",\"Wednesday\",\"Thursday\",\"Friday\",\"Saturday\"]\nfor x in week:\n print(x)\n if x==\"Thursday\":\n break\n ", "Sunday\nMonday\nTuesday\nWednesday\nThursday\n" ], [ "week = [\"Sunday\",\"Monday\",\"Tuesday\",\"Wednesday\",\"Thursday\",\"Friday\",\"Saturday\"]\nfor x in week:\n if x==\"Thursday\":\n break\n print(x)", "Sunday\nMonday\nTuesday\nWednesday\n" ] ], [ [ "Looping through string", "_____no_output_____" ] ], [ [ "for x in \"Python Programming\":\n print(x)", "P\ny\nt\nh\no\nn\n \nP\nr\no\ng\nr\na\nm\nm\ni\nn\ng\n" ] ], [ [ "The range () function", "_____no_output_____" ] ], [ [ "for x in range(16):\n print(x)", "0\n1\n2\n3\n4\n5\n6\n7\n8\n9\n10\n11\n12\n13\n14\n15\n" ], [ "for x in range(2,16):\n print(x)", "2\n3\n4\n5\n6\n7\n8\n9\n10\n11\n12\n13\n14\n15\n" ] ], [ [ "Nested Loops", "_____no_output_____" ] ], [ [ "adjective=[\"red\",\"big\",\"tasty\"]\nfruits = [\"apple\",\"banana\",\"cherry\"]\nfor x in adjective:\n for y in fruits:\n print(x,y)", "red apple\nred banana\nred cherry\nbig apple\nbig banana\nbig cherry\ntasty apple\ntasty banana\ntasty cherry\n" ] ], [ [ "##While Loop", "_____no_output_____" ] ], [ [ "i=1\nwhile i<=6:\n print(i)\n i+=1 #Assignment operator for addition", "1\n2\n3\n4\n5\n6\n" ] ], [ [ "The break statement", "_____no_output_____" ] ], [ [ "i=1\nwhile i<6:\n print(i)\n if i==3:\n break\n i+=1", "1\n2\n3\n" ] ], [ [ "The continue statement", "_____no_output_____" ] ], [ [ "i = 0\nwhile i<6:\n i+=1 #Assignment operator for addition\n if i==3:\n continue\n print(i)", "1\n2\n4\n5\n6\n" ] ], [ [ "The else statement", "_____no_output_____" ] ], [ [ "i = 1\nwhile i<=6:\n print(i)\n i+=1\nelse:\n print(\"i is no longer less than 6\")", "1\n2\n3\n4\n5\n6\ni is no longer less than 6\n" ] ], [ [ "Application 1", "_____no_output_____" ] ], [ [ "#Create a Python program that displays Hello 0 to Hello 10 in vertical sequence\n\nhello=[\"Hello\"]\nnum=[\"0\",\"1\",\"2\",\"3\",\"4\",\"5\",\"6\",\"7\",\"8\",\"9\",\"10\"]\n\n#for loop\nfor x in hello:\n for y in num:\n print(x,y)\n\n#while loop\ni=0\nwhile i<=10:\n print(\"Hello\",i)\n i+=1 #Assignment operator to increment i", "Hello 0\nHello 1\nHello 2\nHello 3\nHello 4\nHello 5\nHello 6\nHello 7\nHello 8\nHello 9\nHello 10\n" ] ], [ [ "Application 2", "_____no_output_____" ] ], [ [ "#Create a Python program that displays integers less than 10 but not less than 3\n\ni = 0\nwhile i<10:\n i+=1 #Assignment operator to increment i\n if i<3:\n continue\n if i==10:\n break\n print(i)", "3\n4\n5\n6\n7\n8\n9\n" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Lab 11: MLP -- exercise\n\n# Understanding the training loop ", "_____no_output_____" ] ], [ [ "import torch\nimport torch.nn as nn\nimport torch.nn.functional as F\nimport torch.optim as optim\nfrom random import randint\nimport utils", "_____no_output_____" ] ], [ [ "### Download the data and print the sizes", "_____no_output_____" ] ], [ [ "train_data=torch.load('../data/fashion-mnist/train_data.pt')\n\nprint(train_data.size())", "torch.Size([60000, 28, 28])\n" ], [ "train_label=torch.load('../data/fashion-mnist/train_label.pt')\n\nprint(train_label.size())", "torch.Size([60000])\n" ], [ "test_data=torch.load('../data/fashion-mnist/test_data.pt')\n\nprint(test_data.size())", "torch.Size([10000, 28, 28])\n" ] ], [ [ "### Make a ONE layer net class. The network output are the scores! No softmax needed! You have only one line to write in the forward function", "_____no_output_____" ] ], [ [ "class one_layer_net(nn.Module):\n\n def __init__(self, input_size, output_size):\n super(one_layer_net , self).__init__()\n self.linear_layer = nn.Linear(input_size, output_size, bias=False)# complete here\n \n def forward(self, x):\n scores = self.linear_layer(x) # complete here\n return scores", "_____no_output_____" ] ], [ [ "### Build the net", "_____no_output_____" ] ], [ [ "net= one_layer_net(784,10)# complete here\nprint(net)", "one_layer_net(\n (linear_layer): Linear(in_features=784, out_features=10, bias=False)\n)\n" ] ], [ [ "### Choose the criterion and the optimizer: use the CHEAT SHEET to see the correct syntax. \n\n### Remember that the optimizer need to have access to the parameters of the network (net.parameters()).\n\n### Set the batchize and learning rate to be:\n### batchize = 50\n### learning rate = 0.01\n\n\n\n\n\n", "_____no_output_____" ] ], [ [ "# make the criterion\ncriterion = nn.CrossEntropyLoss()# complete here\n\n# make the SGD optimizer. \noptimizer=torch.optim.SGD(net.parameters(), lr=0.01) #complete here )\n\n# set up the batch size \nbs=50", "_____no_output_____" ] ], [ [ "### Complete the training loop", "_____no_output_____" ] ], [ [ "for iter in range(1,5000):\n \n # Set dL/dU, dL/dV, dL/dW to be filled with zeros\n optimizer.zero_grad()\n\n # create a minibatch\n indices = torch.LongTensor(bs).random_(0,60000)\n minibatch_data = train_data[indices]\n minibatch_label = train_label[indices]\n\n # reshape the minibatch\n inputs = minibatch_data.view(bs, 784)\n\n # tell Pytorch to start tracking all operations that will be done on \"inputs\"\n inputs.requires_grad_()\n \n # forward the minibatch through the net \n scores = net(inputs)\n\n # Compute the average of the losses of the data points in the minibatch\n loss = criterion(scores, minibatch_label)\n\n # backward pass to compute dL/dU, dL/dV and dL/dW \n loss.backward()\n\n # do one step of stochastic gradient descent: U=U-lr(dL/dU), V=V-lr(dL/dU), ...\n optimizer.step()", "_____no_output_____" ] ], [ [ "### Choose image at random from the test set and see how good/bad are the predictions", "_____no_output_____" ] ], [ [ "# choose a picture at random\nidx=randint(0, 10000-1)\nim=test_data[idx]\n\n# diplay the picture\nutils.show(im)\n\n# feed it to the net and display the confidence scores\nscores = net( im.view(1,784)) \nprobs= F.softmax(scores, dim=1)\nutils.show_prob_fashion_mnist(probs)", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "# Copyright 2020 NVIDIA Corporation. All Rights Reserved.\n#\n# Licensed under the Apache License, Version 2.0 (the \"License\");\n# you may not use this file except in compliance with the License.\n# You may obtain a copy of the License at\n#\n# http://www.apache.org/licenses/LICENSE-2.0\n#\n# Unless required by applicable law or agreed to in writing, software\n# distributed under the License is distributed on an \"AS IS\" BASIS,\n# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n# See the License for the specific language governing permissions and\n# limitations under the License.\n# ==============================================================================", "_____no_output_____" ] ], [ [ "<img src=\"http://developer.download.nvidia.com/compute/machine-learning/frameworks/nvidia_logo.png\" style=\"width: 90px; float: right;\">\n\n# NVTabular demo on Rossmann data - TensorFlow\n\n## Overview\n\nNVTabular is a feature engineering and preprocessing library for tabular data designed to quickly and easily manipulate terabyte scale datasets used to train deep learning based recommender systems. It provides a high level abstraction to simplify code and accelerates computation on the GPU using the RAPIDS cuDF library.", "_____no_output_____" ], [ "### Learning objectives\n\nIn the previous notebooks ([rossmann-store-sales-preproc.ipynb](https://github.com/NVIDIA/NVTabular/blob/main/examples/rossmann/rossmann-store-sales-preproc.ipynb) and [rossmann-store-sales-feature-engineering.ipynb](https://github.com/NVIDIA/NVTabular/blob/main/examples/rossmann/rossmann-store-sales-feature-engineering.ipynb)), we downloaded, preprocessed and created features for the dataset. Now, we are ready to train our deep learning model on the dataset. In this notebook, we use **TensorFlow** with the NVTabular data loader for TensorFlow to accelereate the training pipeline.", "_____no_output_____" ] ], [ [ "import os\nimport math\nimport json\nimport nvtabular as nvt\nimport glob", "_____no_output_____" ] ], [ [ "## Loading NVTabular workflow\nThis time, we only need to define our data directories. We can load the data schema from the NVTabular workflow.", "_____no_output_____" ] ], [ [ "DATA_DIR = os.environ.get(\"OUTPUT_DATA_DIR\", \"./data\")\nINPUT_DATA_DIR = os.environ.get(\"INPUT_DATA_DIR\", \"./data\")\nPREPROCESS_DIR = os.path.join(INPUT_DATA_DIR, 'ross_pre/')\nPREPROCESS_DIR_TRAIN = os.path.join(PREPROCESS_DIR, 'train')\nPREPROCESS_DIR_VALID = os.path.join(PREPROCESS_DIR, 'valid')", "_____no_output_____" ] ], [ [ "What files are available to train on in our directories?", "_____no_output_____" ] ], [ [ "!ls $PREPROCESS_DIR", "stats.json train valid\r\n" ], [ "!ls $PREPROCESS_DIR_TRAIN", "0.274616a5186f4ae6b9688cad2a1ac1e5.parquet _metadata\r\n_file_list.txt\t\t\t\t _metadata.json\r\n" ], [ "!ls $PREPROCESS_DIR_VALID", "_metadata part.0.parquet\n" ] ], [ [ "We load the data schema and statistic information from `stats.json`. We created the file in the previous notebook `rossmann-store-sales-feature-engineering`.", "_____no_output_____" ] ], [ [ "stats = json.load(open(PREPROCESS_DIR + \"/stats.json\", \"r\"))", "_____no_output_____" ], [ "CATEGORICAL_COLUMNS = stats['CATEGORICAL_COLUMNS']\nCONTINUOUS_COLUMNS = stats['CONTINUOUS_COLUMNS']\nLABEL_COLUMNS = stats['LABEL_COLUMNS']\nCOLUMNS = CATEGORICAL_COLUMNS + CONTINUOUS_COLUMNS + LABEL_COLUMNS", "_____no_output_____" ] ], [ [ "The embedding table shows the cardinality of each categorical variable along with its associated embedding size. Each entry is of the form `(cardinality, embedding_size)`.", "_____no_output_____" ] ], [ [ "EMBEDDING_TABLE_SHAPES = stats['EMBEDDING_TABLE_SHAPES']\nEMBEDDING_TABLE_SHAPES", "_____no_output_____" ] ], [ [ "## Training a Network\n\nNow that our data is preprocessed and saved out, we can leverage `dataset`s to read through the preprocessed parquet files in an online fashion to train neural networks.\n\nWe'll start by setting some universal hyperparameters for our model and optimizer. These settings will be the same across all of the frameworks that we explore in the different notebooks.", "_____no_output_____" ], [ "If you're interested in contributing to NVTabular, feel free to take this challenge on and submit a pull request if successful. 12% RMSPE is achievable using the Novograd optimizer, but we know of no Novograd implementation for TensorFlow that supports sparse gradients, and so we are not including that solution below.", "_____no_output_____" ] ], [ [ "EMBEDDING_DROPOUT_RATE = 0.04\nDROPOUT_RATES = [0.001, 0.01]\nHIDDEN_DIMS = [1000, 500]\nBATCH_SIZE = 65536\nLEARNING_RATE = 0.001\nEPOCHS = 25\n\n# TODO: Calculate on the fly rather than recalling from previous analysis.\nMAX_SALES_IN_TRAINING_SET = 38722.0\nMAX_LOG_SALES_PREDICTION = 1.2 * math.log(MAX_SALES_IN_TRAINING_SET + 1.0)\n\nTRAIN_PATHS = sorted(glob.glob(os.path.join(PREPROCESS_DIR_TRAIN, '*.parquet')))\nVALID_PATHS = sorted(glob.glob(os.path.join(PREPROCESS_DIR_VALID, '*.parquet')))", "_____no_output_____" ] ], [ [ "## TensorFlow\n<a id=\"TensorFlow\"></a>\n", "_____no_output_____" ], [ "### TensorFlow: Preparing Datasets\n\n`KerasSequenceLoader` wraps a lightweight iterator around a `dataset` object to handle chunking, shuffling, and application of any workflows (which can be applied online as a preprocessing step). For column names, can use either a list of string names or a list of TensorFlow `feature_columns` that will be used to feed the network", "_____no_output_____" ] ], [ [ "import tensorflow as tf\n\n# we can control how much memory to give tensorflow with this environment variable\n# IMPORTANT: make sure you do this before you initialize TF's runtime, otherwise\n# it's too late and TF will have claimed all free GPU memory\nos.environ['TF_MEMORY_ALLOCATION'] = \"8192\" # explicit MB\nos.environ['TF_MEMORY_ALLOCATION'] = \"0.5\" # fraction of free memory\nfrom nvtabular.loader.tensorflow import KerasSequenceLoader, KerasSequenceValidater\n\n# cheap wrapper to keep things some semblance of neat\ndef make_categorical_embedding_column(name, dictionary_size, embedding_dim):\n return tf.feature_column.embedding_column(\n tf.feature_column.categorical_column_with_identity(name, dictionary_size),\n embedding_dim\n )\n\n# instantiate our columns\ncategorical_columns = [\n make_categorical_embedding_column(name, *EMBEDDING_TABLE_SHAPES[name]) for\n name in CATEGORICAL_COLUMNS\n]\ncontinuous_columns = [\n tf.feature_column.numeric_column(name, (1,)) for name in CONTINUOUS_COLUMNS\n]\n\n# feed them to our datasets\ntrain_dataset = KerasSequenceLoader(\n TRAIN_PATHS, # you could also use a glob pattern\n feature_columns=categorical_columns+continuous_columns,\n batch_size=BATCH_SIZE,\n label_names=LABEL_COLUMNS,\n shuffle=True,\n buffer_size=0.06 # amount of data, as a fraction of GPU memory, to load at once\n)\n\nvalid_dataset = KerasSequenceLoader(\n VALID_PATHS, # you could also use a glob pattern\n feature_columns=categorical_columns+continuous_columns,\n batch_size=BATCH_SIZE*4,\n label_names=LABEL_COLUMNS,\n shuffle=False,\n buffer_size=0.06 # amount of data, as a fraction of GPU memory, to load at once\n)", "_____no_output_____" ] ], [ [ "### TensorFlow: Defining a Model\n\nUsing Keras, we can define the layers of our model and their parameters explicitly. Here, for the sake of consistency, we'll mimic fast.ai's [TabularModel](https://docs.fast.ai/tabular.learner.html).", "_____no_output_____" ] ], [ [ "# DenseFeatures layer needs a dictionary of {feature_name: input}\ncategorical_inputs = {}\nfor column_name in CATEGORICAL_COLUMNS:\n categorical_inputs[column_name] = tf.keras.Input(name=column_name, shape=(1,), dtype=tf.int64)\ncategorical_embedding_layer = tf.keras.layers.DenseFeatures(categorical_columns)\ncategorical_x = categorical_embedding_layer(categorical_inputs)\ncategorical_x = tf.keras.layers.Dropout(EMBEDDING_DROPOUT_RATE)(categorical_x)\n\n# Just concatenating continuous, so can use a list\ncontinuous_inputs = []\nfor column_name in CONTINUOUS_COLUMNS:\n continuous_inputs.append(tf.keras.Input(name=column_name, shape=(1,), dtype=tf.float32))\ncontinuous_embedding_layer = tf.keras.layers.Concatenate(axis=1)\ncontinuous_x = continuous_embedding_layer(continuous_inputs)\ncontinuous_x = tf.keras.layers.BatchNormalization(epsilon=1e-5, momentum=0.1)(continuous_x)\n\n# concatenate and build MLP\nx = tf.keras.layers.Concatenate(axis=1)([categorical_x, continuous_x])\nfor dim, dropout_rate in zip(HIDDEN_DIMS, DROPOUT_RATES):\n x = tf.keras.layers.Dense(dim, activation='relu')(x)\n x = tf.keras.layers.BatchNormalization(epsilon=1e-5, momentum=0.1)(x)\n x = tf.keras.layers.Dropout(dropout_rate)(x)\nx = tf.keras.layers.Dense(1, activation='linear')(x)\n\n# TODO: Initialize model weights to fix saturation issues.\n# For now, we'll just scale the output of our model directly before\n# hitting the sigmoid.\nx = 0.1 * x\n\nx = MAX_LOG_SALES_PREDICTION * tf.keras.activations.sigmoid(x)\n\n# combine all our inputs into a single list\n# (note that you can still use .fit, .predict, etc. on a dict\n# that maps input tensor names to input values)\ninputs = list(categorical_inputs.values()) + continuous_inputs\ntf_model = tf.keras.Model(inputs=inputs, outputs=x)", "_____no_output_____" ] ], [ [ "### TensorFlow: Training", "_____no_output_____" ] ], [ [ "def rmspe_tf(y_true, y_pred):\n # map back into \"true\" space by undoing transform\n y_true = tf.exp(y_true) - 1\n y_pred = tf.exp(y_pred) - 1\n\n percent_error = (y_true - y_pred) / y_true\n return tf.sqrt(tf.reduce_mean(percent_error**2))", "_____no_output_____" ], [ "%%time\nfrom time import time\n\noptimizer = tf.keras.optimizers.Adam(learning_rate=LEARNING_RATE)\ntf_model.compile(optimizer, 'mse', metrics=[rmspe_tf])\n\nvalidation_callback = KerasSequenceValidater(valid_dataset)\nstart = time()\nhistory = tf_model.fit(\n train_dataset,\n callbacks=[validation_callback],\n epochs=EPOCHS,\n)\nt_final = time() - start\ntotal_rows = train_dataset.num_rows_processed + valid_dataset.num_rows_processed\nprint(f\"run_time: {t_final} - rows: {total_rows} - epochs: {EPOCHS} - dl_thru: { (EPOCHS * total_rows) / t_final}\")", "Epoch 1/25\n13/13 [==============================] - 7s 168ms/step - loss: 6.3708 - rmspe_tf: 0.8916\nEpoch 2/25\n13/13 [==============================] - 2s 166ms/step - loss: 5.3491 - rmspe_tf: 0.8906\nEpoch 3/25\n13/13 [==============================] - 2s 168ms/step - loss: 4.7029 - rmspe_tf: 0.8801\nEpoch 4/25\n13/13 [==============================] - 2s 164ms/step - loss: 3.9542 - rmspe_tf: 0.8585\nEpoch 5/25\n13/13 [==============================] - 2s 168ms/step - loss: 3.0444 - rmspe_tf: 0.8195\nEpoch 6/25\n13/13 [==============================] - 2s 165ms/step - loss: 2.0530 - rmspe_tf: 0.7533\nEpoch 7/25\n13/13 [==============================] - 2s 167ms/step - loss: 1.1581 - rmspe_tf: 0.6474\nEpoch 8/25\n13/13 [==============================] - 2s 166ms/step - loss: 0.5232 - rmspe_tf: 0.5006\nEpoch 9/25\n13/13 [==============================] - 2s 164ms/step - loss: 0.1878 - rmspe_tf: 0.3450\nEpoch 10/25\n13/13 [==============================] - 2s 164ms/step - loss: 0.0650 - rmspe_tf: 0.2355\nEpoch 11/25\n13/13 [==============================] - 2s 166ms/step - loss: 0.0372 - rmspe_tf: 0.2073\nEpoch 12/25\n13/13 [==============================] - 2s 166ms/step - loss: 0.0329 - rmspe_tf: 0.2094\nEpoch 13/25\n13/13 [==============================] - 2s 165ms/step - loss: 0.0317 - rmspe_tf: 0.2090\nEpoch 14/25\n13/13 [==============================] - 2s 168ms/step - loss: 0.0301 - rmspe_tf: 0.2035\nEpoch 15/25\n13/13 [==============================] - 2s 170ms/step - loss: 0.0292 - rmspe_tf: 0.1987\nEpoch 16/25\n13/13 [==============================] - 2s 168ms/step - loss: 0.0283 - rmspe_tf: 0.1952\nEpoch 17/25\n13/13 [==============================] - 2s 166ms/step - loss: 0.0276 - rmspe_tf: 0.1905\nEpoch 18/25\n13/13 [==============================] - 2s 165ms/step - loss: 0.0274 - rmspe_tf: 0.1877\nEpoch 19/25\n13/13 [==============================] - 2s 165ms/step - loss: 0.0270 - rmspe_tf: 0.1956\nEpoch 20/25\n13/13 [==============================] - 2s 165ms/step - loss: 0.0248 - rmspe_tf: 0.1789\nEpoch 21/25\n13/13 [==============================] - 2s 166ms/step - loss: 0.0244 - rmspe_tf: 0.1792\nEpoch 22/25\n13/13 [==============================] - 2s 167ms/step - loss: 0.0239 - rmspe_tf: 0.1785\nEpoch 23/25\n13/13 [==============================] - 2s 167ms/step - loss: 0.0234 - rmspe_tf: 0.1775\nEpoch 24/25\n13/13 [==============================] - 2s 166ms/step - loss: 0.0233 - rmspe_tf: 0.1747\nEpoch 25/25\n13/13 [==============================] - 2s 168ms/step - loss: 0.0228 - rmspe_tf: 0.1725\nCPU times: user 2min 52s, sys: 13 s, total: 3min 5s\nWall time: 1min 8s\n" ] ] ]

[ "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import os\nimport shutil\n\ndef raw2processed(filename):\n shutil.copy(os.path.join('data_raw',filename),os.path.join('data_processed',filename))", "_____no_output_____" ], [ "raw2processed('FlexibilityEvents.HH1.json')\nraw2processed('FlexibilityEvents2.HH1.json')", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "## Main points\n* Solution should be reasonably simple because the contest is only 24 hours long \n* Metric is based on the prediction of clicked pictures one week ahead, so clicks are the most important information\n* More recent information is more important\n* Only pictures that were shown to a user could be clicked, so pictures popularity is important\n* Metric is MAPK@100\n* Link https://contest.yandex.ru/contest/12899/problems (Russian)", "_____no_output_____" ], [ "## Plan\n* Build a classic recommending system based on user click history\n* Only use recent days of historical data\n* Take into consideration projected picture popularity", "_____no_output_____" ], [ "## Magic constants\n### ALS recommending system:", "_____no_output_____" ] ], [ [ "# Factors for ALS\nfactors_count=100\n\n# Last days of click history used\ntrail_days=14 \n\n# number of best candidates generated by ALS \noutput_candidates_count=2000 \n\n# Last days of history with more weight\nlast_days=1\n\n# Coefficient for additional weight\nlast_days_weight=4", "_____no_output_____" ] ], [ [ "## Popular pictures prediction model:", "_____no_output_____" ] ], [ [ "import lightgbm\n\nlightgbm.__version__", "_____no_output_____" ], [ "popularity_model = lightgbm.LGBMRegressor(seed=0)\nheuristic_alpha = 0.2", "_____no_output_____" ], [ "import datetime\nimport tqdm\nimport pandas as pd\nfrom scipy.sparse import coo_matrix\n\nimport implicit\nimplicit.__version__", "_____no_output_____" ], [ "test_users = pd.read_csv('Blitz/test_users.csv')\ndata = pd.read_csv('Blitz/train_clicks.csv', parse_dates=['day'])", "_____no_output_____" ] ], [ [ "## Split last 7 days to calculate clicks similar to test set\n", "_____no_output_____" ] ], [ [ "train, target_week = (\n data[data.day <= datetime.datetime(2019, 3, 17)].copy(),\n data[data.day > datetime.datetime(2019, 3, 17)],\n)\ntrain.day.nunique(), target_week.day.nunique()", "_____no_output_____" ], [ "last_date = train.day.max()\ntrain.loc[:, 'delta_days'] = 1 + (last_date - train.day).apply(lambda d: d.days)\n\nlast_date = data.day.max()\ndata.loc[:, 'delta_days'] = 1 + (last_date - data.day).apply(lambda d: d.days)", "_____no_output_____" ], [ "def picture_features(data):\n \"\"\"Generating clicks count for every picture in last days\"\"\"\n days = range(1, 3)\n features = []\n names = []\n for delta_days in days:\n features.append(\n data[(data.delta_days == delta_days)].groupby(['picture_id'])['user_id'].count()\n )\n names.append('%s_%d' % ('click', delta_days))\n \n features = pd.concat(features, axis=1).fillna(0)\n features.columns = names\n features = features.reindex(data.picture_id.unique())\n return features.fillna(0)", "_____no_output_____" ], [ "X = picture_features(train)\nX.mean(axis=0)", "_____no_output_____" ], [ "def clicks_count(data, index):\n return data.groupby('picture_id')['user_id'].count().reindex(index).fillna(0)\n \ny = clicks_count(target_week, X.index)\ny.shape, y.mean()", "_____no_output_____" ] ], [ [ "## Train a model predicting popular pictures next week", "_____no_output_____" ] ], [ [ "popularity_model.fit(X, y)", "_____no_output_____" ], [ "X_test = picture_features(data)\nX_test.mean(axis=0)", "_____no_output_____" ], [ "X_test['p'] = popularity_model.predict(X_test)\nX_test.loc[X_test['p'] < 0, 'p'] = 0\nX_test['p'].mean()", "_____no_output_____" ] ], [ [ "## Generate dict with predicted clicks for every picture\n", "_____no_output_____" ] ], [ [ "# This prediction would be used to correct recommender score\npicture = dict(X_test['p'])", "_____no_output_____" ] ], [ [ "# Recommender part", "_____no_output_____" ], [ "## Generate prediction using ALS approach", "_____no_output_____" ] ], [ [ "import os\nos.environ['OPENBLAS_NUM_THREADS'] = \"1\"\n\ndef als_baseline(\n train, test_users, \n factors_n, last_days, trail_days, output_candidates_count, last_days_weight\n):\n train = train[train.delta_days <= trail_days].drop_duplicates([\n 'user_id', 'picture_id'\n ])\n \n users = train.user_id\n items = train.picture_id\n weights = 1 + last_days_weight * (train.delta_days <= last_days)\n \n user_item = coo_matrix((weights, (users, items)))\n model = implicit.als.AlternatingLeastSquares(factors=factors_n, iterations=factors_n)\n model.fit(user_item.T.tocsr())\n \n user_item_csr = user_item.tocsr()\n \n rows = []\n for user_id in tqdm.tqdm_notebook(test_users.user_id.values):\n items = [(picture_id, score) for picture_id, score in model.recommend(user_id, user_item_csr, N=output_candidates_count)]\n rows.append(items)\n\n test_users['predictions_full'] = [\n p\n for p, user_id in zip(\n rows,\n test_users.user_id.values\n )\n ]\n test_users['predictions'] = [\n [x[0] for x in p]\n for p, user_id in zip(\n rows,\n test_users.user_id.values\n )\n ]\n return test_users", "_____no_output_____" ], [ "test_users = als_baseline(\n data, test_users, factors_count, last_days, trail_days, output_candidates_count, last_days_weight)", "100%|██████████| 100.0/100 [11:00<00:00, 6.78s/it]\n" ] ], [ [ "## Calculate history clicks to exclude them from results. Such clicks are excluded from test set according to task", "_____no_output_____" ] ], [ [ "clicked = data.groupby('user_id').agg({'picture_id': set})\n\ndef substract_clicked(p, c):\n filtered = [picture for picture in p if picture not in c][:100]\n return filtered", "_____no_output_____" ] ], [ [ "## Heuristical approach to reweight ALS score according to picture predicted popularity", "_____no_output_____" ], [ "Recommender returns (picture, score) pairs sorted decreasing for every user.\n\nFor every user we replace picture $score_p$ with $score_p \\cdot (1 + popularity_{p})^{0.2}$\n\n$popularity_{p}$ - popularity predicted for this picture for next week\n\nThis slightly moves popular pictures to the top of list for every user", "_____no_output_____" ] ], [ [ "import math\n\nrows = test_users['predictions_full']\n\ndef correct_with_popularity(items, picture, alpha):\n return sorted([\n (score * (1 + picture.get(picture_id, 0)) ** alpha, picture_id, score, picture.get(picture_id, 0)) \n for picture_id, score in items], reverse=True\n )\n\ncorrected_rows = [\n [x[1] for x in correct_with_popularity(items, picture, heuristic_alpha)]\n for items in rows\n]", "_____no_output_____" ] ], [ [ "## Submission formatting", "_____no_output_____" ] ], [ [ "test_users['predictions'] = [\n ' '.join(map(str,\n substract_clicked(p, {} if user_id not in clicked.index else clicked.loc[user_id][0])\n ))\n for p, user_id in zip(\n corrected_rows,\n test_users.user_id.values\n )\n]", "_____no_output_____" ], [ "test_users[['user_id', 'predictions']].to_csv('submit.csv', index=False)", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "empty" ] ] ]

[ "empty" ]

[ [ "empty" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "!pip install pydantic", "Requirement already satisfied: pydantic in /usr/local/lib/python3.8/dist-packages (1.6.1)\n\u001b[33mWARNING: You are using pip version 20.2.2; however, version 20.2.4 is available.\nYou should consider upgrading via the '/usr/bin/python3 -m pip install --upgrade pip' command.\u001b[0m\n" ], [ "from tzlocal import get_localzone # $ pip install tzlocal\n\n# get local timezone \nlocal_tz = get_localzone()\n\nprint(local_tz)", "UTC\n" ], [ "from typing import Union\n\ndef func(name: str, /, *, age: Union[int, str] = \"a\", parent: str = \"\"):\n print(name)\n\n# func.__defaults__ = (\"bob\",) # デフォルトを変更できる\nfunc.__kwdefaults__\n\nprint(func.__defaults__)\nprint(func.__kwdefaults__)\nprint(func.__annotations__)\n", "None\n{'age': 'a', 'parent': ''}\n{'name': <class 'str'>, 'age': typing.Union[int, str], 'parent': <class 'str'>}\n" ], [ "from pydantic import BaseModel\n\nclass Sample(BaseModel):\n name: str = \"bob\"\n age: int = 20\n\n \n\ndef create_func(model: BaseModel):\n \n __annotations__ = {}\n kwdefaults = {}\n \n fields_required = []\n fields_non_required = []\n \n # sort required first\n for item in model.__fields__.values():\n if item.required:\n fields_required.append(item)\n else:\n fields_non_required.append(item)\n \n fields = fields_required\n fields += fields_non_required\n \n for item in fields:\n __annotations__[item.name] = item.type_\n if item.required:\n kwdefaults[item.name] = item.default\n \n # kwdefaults = tuple(kwdefaults)\n codes = []\n codes.append(\"def overload(*args\")\n \n for item in fields:\n codes.append(f\", {item.name} = None\")\n \n codes.append(\"):\\n if len(args) > 1: Exception()\")\n codes.append(\"\\n return args[0] if len(args) else model(\")\n \n for item in fields:\n codes.append(f\"{item.name}={item.name},\")\n\n codes.append(\")\")\n codes.append(\"\\n\")\n code = \"\".join(codes)\n print(code)\n namespace = {\"model\": model}\n \n exec(code, namespace, namespace)\n func = eval(\"overload\", namespace, namespace)\n func.__annotations__ = __annotations__\n func.__kwdefaults__ = kwdefaults\n return func\n\n \nfunc = create_func(Sample)\n\n# print(func())\nprint(func(name=\"a\", age=20))\n\n\ndef func(model):\n if model:\n return model\n else:\n return mymodel(\n name=name,\n age=age\n )\n ", "def overload(*args, name, age):\n if len(args) > 1: Exception()\n return args[0] if len(args) else model(name=name,age=age,)\n\nname='a' age=20\n" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Load predictor", "_____no_output_____" ] ], [ [ "%matplotlib inline\nimport os\nimport matplotlib\nimport numpy as np\nimport matplotlib.pyplot as plt", "_____no_output_____" ], [ "matplotlib.use(\"Agg\")", "/home/del/anaconda3/envs/allennmt/lib/python3.6/site-packages/ipykernel/__main__.py:1: UserWarning: \nThis call to matplotlib.use() has no effect because the backend has already\nbeen chosen; matplotlib.use() must be called *before* pylab, matplotlib.pyplot,\nor matplotlib.backends is imported for the first time.\n\nThe backend was *originally* set to 'module://ipykernel.pylab.backend_inline' by the following code:\n File \"/home/del/anaconda3/envs/allennmt/lib/python3.6/runpy.py\", line 193, in _run_module_as_main\n \"__main__\", mod_spec)\n File \"/home/del/anaconda3/envs/allennmt/lib/python3.6/runpy.py\", line 85, in _run_code\n exec(code, run_globals)\n File \"/home/del/anaconda3/envs/allennmt/lib/python3.6/site-packages/ipykernel/__main__.py\", line 3, in <module>\n app.launch_new_instance()\n File \"/home/del/anaconda3/envs/allennmt/lib/python3.6/site-packages/traitlets/config/application.py\", line 658, in launch_instance\n app.start()\n File \"/home/del/anaconda3/envs/allennmt/lib/python3.6/site-packages/ipykernel/kernelapp.py\", line 486, in start\n self.io_loop.start()\n File \"/home/del/anaconda3/envs/allennmt/lib/python3.6/site-packages/tornado/ioloop.py\", line 888, in start\n handler_func(fd_obj, events)\n File \"/home/del/anaconda3/envs/allennmt/lib/python3.6/site-packages/tornado/stack_context.py\", line 277, in null_wrapper\n return fn(*args, **kwargs)\n File \"/home/del/anaconda3/envs/allennmt/lib/python3.6/site-packages/zmq/eventloop/zmqstream.py\", line 450, in _handle_events\n self._handle_recv()\n File \"/home/del/anaconda3/envs/allennmt/lib/python3.6/site-packages/zmq/eventloop/zmqstream.py\", line 480, in _handle_recv\n self._run_callback(callback, msg)\n File \"/home/del/anaconda3/envs/allennmt/lib/python3.6/site-packages/zmq/eventloop/zmqstream.py\", line 432, in _run_callback\n callback(*args, **kwargs)\n File \"/home/del/anaconda3/envs/allennmt/lib/python3.6/site-packages/tornado/stack_context.py\", line 277, in null_wrapper\n return fn(*args, **kwargs)\n File \"/home/del/anaconda3/envs/allennmt/lib/python3.6/site-packages/ipykernel/kernelbase.py\", line 283, in dispatcher\n return self.dispatch_shell(stream, msg)\n File \"/home/del/anaconda3/envs/allennmt/lib/python3.6/site-packages/ipykernel/kernelbase.py\", line 233, in dispatch_shell\n handler(stream, idents, msg)\n File \"/home/del/anaconda3/envs/allennmt/lib/python3.6/site-packages/ipykernel/kernelbase.py\", line 399, in execute_request\n user_expressions, allow_stdin)\n File \"/home/del/anaconda3/envs/allennmt/lib/python3.6/site-packages/ipykernel/ipkernel.py\", line 208, in do_execute\n res = shell.run_cell(code, store_history=store_history, silent=silent)\n File \"/home/del/anaconda3/envs/allennmt/lib/python3.6/site-packages/ipykernel/zmqshell.py\", line 537, in run_cell\n return super(ZMQInteractiveShell, self).run_cell(*args, **kwargs)\n File \"/home/del/anaconda3/envs/allennmt/lib/python3.6/site-packages/IPython/core/interactiveshell.py\", line 2739, in run_cell\n self.events.trigger('post_run_cell')\n File \"/home/del/anaconda3/envs/allennmt/lib/python3.6/site-packages/IPython/core/events.py\", line 73, in trigger\n func(*args, **kwargs)\n File \"/home/del/anaconda3/envs/allennmt/lib/python3.6/site-packages/ipykernel/pylab/backend_inline.py\", line 160, in configure_once\n activate_matplotlib(backend)\n File \"/home/del/anaconda3/envs/allennmt/lib/python3.6/site-packages/IPython/core/pylabtools.py\", line 308, in activate_matplotlib\n matplotlib.pyplot.switch_backend(backend)\n File \"/home/del/anaconda3/envs/allennmt/lib/python3.6/site-packages/matplotlib/pyplot.py\", line 231, in switch_backend\n matplotlib.use(newbackend, warn=False, force=True)\n File \"/home/del/anaconda3/envs/allennmt/lib/python3.6/site-packages/matplotlib/__init__.py\", line 1400, in use\n reload(sys.modules['matplotlib.backends'])\n File \"/home/del/anaconda3/envs/allennmt/lib/python3.6/importlib/__init__.py\", line 166, in reload\n _bootstrap._exec(spec, module)\n File \"/home/del/anaconda3/envs/allennmt/lib/python3.6/site-packages/matplotlib/backends/__init__.py\", line 16, in <module>\n line for line in traceback.format_stack()\n\n\n if __name__ == '__main__':\n" ], [ "os.getcwd()", "_____no_output_____" ], [ "os.chdir('/home/del/research/span_ae')", "_____no_output_____" ], [ "import span_ae\nfrom allennlp.models.archival import load_archive\nfrom allennlp.service.predictors import Predictor", "/home/del/anaconda3/envs/allennmt/lib/python3.6/site-packages/h5py/__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.\n from ._conv import register_converters as _register_converters\n" ], [ "archive = load_archive(\"models/baseline/model.tar.gz\")", "_____no_output_____" ], [ "predictor = Predictor.from_archive(archive, 'span_ae')", "_____no_output_____" ] ], [ [ "## Func", "_____no_output_____" ] ], [ [ "def predict_plot(sentence):\n # predict\n result = predictor.predict_json(sentence)\n attention_matrix = result['attention_matrix']\n predicted_tokens = result['predicted_tokens']\n survived_span_ids = result['top_spans']\n input_sentence = ['BOS'] + sentence['src'].split() + ['EOS']\n predicted_tokens = predicted_tokens + ['EOS']\n survived_spans = []\n for span_id in survived_span_ids:\n ind_from = span_id[0]\n ind_to = span_id[1] + 1\n survived_spans.append(\" \".join(input_sentence[ind_from:ind_to]))\n attention_matrix_local = attention_matrix[0:len(predicted_tokens)]\n att_matrix_np = np.array([np.array(xi) for xi in attention_matrix_local])\n \n \n #print\n print('ORIGINAL :', \" \".join(input_sentence))\n #print('TOP SPANs:', \" \\n \".join(survived_spans))\n print('PREDICTED:', \" \".join(predicted_tokens))\n #print('span scores:', result['top_spans_scores'])\n print('\\nAttnetion matrix:')\n \n # plot\n plt.figure(figsize=(9, 9), dpi= 80, facecolor='w', edgecolor='k')\n plt.imshow(att_matrix_np.transpose(), interpolation=\"nearest\", cmap=\"Greys\")\n plt.xlabel(\"target\")\n plt.ylabel(\"source\")\n plt.gca().set_xticks([i for i in range(0, len(predicted_tokens))])\n plt.gca().set_yticks([i for i in range(0, len(survived_spans))])\n plt.gca().set_xticklabels(predicted_tokens, rotation='vertical')\n plt.gca().set_yticklabels(survived_spans)\n plt.tight_layout()", "_____no_output_____" ] ], [ [ "## Inference", "_____no_output_____" ] ], [ [ "# change it\nsentence = \"to school\"\n\n# do not change it\npredict_plot({'src': sentence})", "ORIGINAL : BOS to school EOS\nPREDICTED: to studies EOS\n\nAttnetion matrix:\n" ], [ "# change it\nsentence = \"school\"\n\n# do not change it\npredict_plot({'src': sentence})", "ORIGINAL : BOS school EOS\nPREDICTED: poverty EOS\n\nAttnetion matrix:\n" ], [ "# change it\nsentence = \"it is spring already , but there are a lot of snow out there\"\n\n# do not change it\npredict_plot({'src': sentence})", "ORIGINAL : BOS it is spring already , but there are a lot of snow out there EOS\nPREDICTED: it is already spring , but there are a lot of snow there there EOS\n\nAttnetion matrix:\n" ], [ "b", "ORIGINAL : BOS state of the art EOS\nPREDICTED: state of the EOS\n\nAttnetion matrix:\n" ], [ "# change it\nsentence = \"let us discard our entire human knowledge\"\n\n# do not change it\npredict_plot({'src': sentence})", "ORIGINAL : BOS let us discard our entire human knowledge EOS\nPREDICTED: let us discard our entire development knowledge EOS\n\nAttnetion matrix:\n" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "This challenge implements an instantiation of OTR based on AES block cipher with modified version 1.0. OTR, which stands for Offset Two-Round, is a blockcipher mode of operation to realize an authenticated encryption with associated data (see [[1]](#1)). AES-OTR algorithm is a campaign of CAESAR competition, it has successfully entered the third round of screening by virtue of its unique advantages, you can see the whole algorithms and structure of AES-OTR from the design document (see [[2]](#2)).\n\nHowever, the first version is vulnerable to forgery attacks in the known plaintext conditions and association data and public message number are reused, many attacks can be applied here to forge an excepted ciphertext with a valid tag (see [[3]](#3)).\n\nFor example, in this challenge we can build the following three plaintexts:", "_____no_output_____" ] ], [ [ "M_0 = [b'Uid=16112\\xffUserNa', b'me=AdministratoR', b'\\xffT=111111111111\\xff', b'Cmd=Give_Me_FlaG', b'\\xff???????????????']\nM_1 = [b'Uid=16111\\xffUserNa', b'me=Administrator', b'r\\xffT=11111111111\\xff', b'Cmd=Give_Me_FlaG', b'\\xff???????????????']\nM_2 = [b'Uid=16112\\xffUserNa', b'me=AdministratoR', b'\\xffT=111111111111\\xff', b'Cmd=Give_Me_Flag', b'g\\xff??????????????']", "_____no_output_____" ] ], [ [ "Here `'111111111111'` can represent any value since the server won't check whether the message and its corresponding hash value match, so we just need to make sure that they are at the right length. If you look closely, you will find that none of the three plaintexts contains illegal fields, so we can use the encrypt Oracle provided by the server to get their corresponding ciphertexts easily. Next, noticed that these plaintexts satisfied:", "_____no_output_____" ] ], [ [ "from Crypto.Util.strxor import strxor\n\nM_0 = [b'Uid=16112\\xffUserNa', b'me=AdministratoR', b'\\xffT=111111111111\\xff', b'Cmd=Give_Me_FlaG', b'\\xff???????????????']\nM_1 = [b'Uid=16111\\xffUserNa', b'me=Administrator', b'r\\xffT=11111111111\\xff', b'Cmd=Give_Me_FlaG', b'\\xff???????????????']\nM_2 = [b'Uid=16112\\xffUserNa', b'me=AdministratoR', b'\\xffT=111111111111\\xff', b'Cmd=Give_Me_Flag', b'g\\xff??????????????']\n\nstrxor(M_0[1], M_0[3]) == strxor(M_1[1], M_2[3])", "_____no_output_____" ] ], [ [ "So according to the forgery attacks described in [[3]](#3), suppose their corresponding ciphertexts are `C_0`, `C_1` and `C_2`, then we can forge a valid ciphertext and tag using:", "_____no_output_____" ] ], [ [ "from Toy_AE import Toy_AE\n\ndef unpack(r):\n data = r.split(b\"\\xff\")\n uid, uname, token, cmd, appendix = int(data[0][4:]), data[1][9:], data[2][2:], data[3][4:], data[4]\n return (uid, uname, token, cmd, appendix)\n\nae = Toy_AE()\n\nM_0 = [b'Uid=16112\\xffUserNa', b'me=AdministratoR', b'\\xffT=111111111111\\xff', b'Cmd=Give_Me_FlaG', b'\\xff???????????????']\nM_1 = [b'Uid=16111\\xffUserNa', b'me=Administrator', b'r\\xffT=11111111111\\xff', b'Cmd=Give_Me_FlaG', b'\\xff???????????????']\nM_2 = [b'Uid=16112\\xffUserNa', b'me=AdministratoR', b'\\xffT=111111111111\\xff', b'Cmd=Give_Me_Flag', b'g\\xff??????????????']\n\nC_0, T_0 = ae.encrypt(b''.join(M_0))\nC_1, T_1 = ae.encrypt(b''.join(M_1))\nC_2, T_2 = ae.encrypt(b''.join(M_2))\nC_forge = C_1[:32] + C_2[32:64] + C_0[64:]\n\nT_forge = T_0\n\n_, uname, _, cmd, _ = unpack(ae.decrypt(C_forge, T_forge))\nuname == b\"Administrator\" and cmd == b\"Give_Me_Flag\"", "_____no_output_____" ] ], [ [ "Here is my final exp:", "_____no_output_____" ] ], [ [ "import string\nfrom pwn import *\nfrom hashlib import sha256\nfrom Crypto.Util.strxor import strxor\nfrom Crypto.Util.number import long_to_bytes, bytes_to_long\n\ndef bypass_POW(io):\n chall = io.recvline()\n post = chall[14:30]\n tar = chall[38:-2]\n io.recvuntil(':')\n found = iters.bruteforce(lambda x:sha256((x + post.decode()).encode()).hexdigest() == tar.decode(), string.ascii_letters + string.digits, 4)\n io.sendline(found.encode())\n\nC = []\nT = []\n\nio = remote(\"123.57.4.93\", 45216)\nbypass_POW(io)\n\nio.sendlineafter(b\"Your option:\", '1')\nio.sendlineafter(b\"Set up your user id:\", '16108')\nio.sendlineafter(b\"Your username:\", 'AdministratoR')\nio.sendlineafter(b\"Your command:\", 'Give_Me_FlaG')\nio.sendlineafter(b\"Any Appendix?\", \"???????????????\")\n\n_ = io.recvuntil(b\"Your ticket:\")\nC.append(long_to_bytes(int(io.recvline().strip(), 16)))\n_ = io.recvuntil(b\"With my Auth:\")\nT.append(long_to_bytes(int(io.recvline().strip(), 16)))\n\nio.sendlineafter(b\"Your option:\", '1')\nio.sendlineafter(b\"Set up your user id:\", '16107')\nio.sendlineafter(b\"Your username:\", 'Administratorr')\nio.sendlineafter(b\"Your command:\", 'Give_Me_FlaG')\nio.sendlineafter(b\"Any Appendix?\", \"???????????????\")\n\n_ = io.recvuntil(b\"Your ticket:\")\nC.append(long_to_bytes(int(io.recvline().strip(), 16)))\n_ = io.recvuntil(b\"With my Auth:\")\nT.append(long_to_bytes(int(io.recvline().strip(), 16)))\n\nio.sendlineafter(b\"Your option:\", '1')\nio.sendlineafter(b\"Set up your user id:\", '16108')\nio.sendlineafter(b\"Your username:\", 'AdministratoR')\nio.sendlineafter(b\"Your command:\", 'Give_Me_Flagg')\nio.sendlineafter(b\"Any Appendix?\", \"??????????????\")\n\n_ = io.recvuntil(b\"Your ticket:\")\nC.append(long_to_bytes(int(io.recvline().strip(), 16)))\n_ = io.recvuntil(b\"With my Auth:\")\nT.append(long_to_bytes(int(io.recvline().strip(), 16)))\n\nct = (C[1][:32] + C[2][32:64] + C[0][64:]).hex()\nte = T[0].hex()\n\nio.sendlineafter(b\"Your option:\", '2')\nio.sendlineafter(b\"Ticket:\", ct)\nio.sendlineafter(b\"Auth:\", te)\nflag = io.recvline().strip().decode()\nprint(flag)", "_____no_output_____" ] ], [ [ "b'X-NUCA{Gentlem3n_as_0f_th1s_mOment_I aM_th4t_sec0nd_mouse}'", "_____no_output_____" ], [ "**P.S.**\n\n* The version used in this challenge is v 1.0, some vulnerabilities have been fixed in subsequent versions(v 2.0, v 3.0 and v 3.1), you can see the final version at [[4]](#4). Also, for some attacks on the new version, see [[5]](#5) and [[6]](#6).\n\n* The content of the FLAG is a quote from movie *Catch Me If You Can* \"Two little mice fell in a bucket of cream. The first mouse quickly gave up and drowned. The second mouse, wouldn't quit. He struggled so hard that eventually he churned that cream into butter and crawled out. Gentlemen, as of this moment, I am that second mouse.\"\n\n**References**\n\n<a id=\"1\" href=\"https://eprint.iacr.org/2013/628.pdf\"> [1] Minematsu K. Parallelizable rate-1 authenticated encryption from pseudorandom functions[C]//Annual International Conference on the Theory and Applications of Cryptographic Techniques. Springer, Berlin, Heidelberg, 2014: 275-292.</a>\n\n<a id=\"2\" href=\"https://competitions.cr.yp.to/round1/aesotrv1.pdf\"> [2] Minematsu K. AES-OTR v1 design document.</a>\n\n<a id=\"3\" href=\"http://www.shcas.net/jsjyup/pdf/2017/10/对认证加密算法AES-OTR的伪造攻击.pdf\"> [3] Xiulin Zheng, Yipeng Fu, Haiyan Song. Forging attacks on authenticated encryption algorithm AES-OTR[J]. Computer Applications and Software, 2017, 034(010):320-324,329.</a>\n\n<a id=\"4\" href=\"https://competitions.cr.yp.to/round1/aesotrv1.pdf\"> [4] Minematsu K. AES-OTR v3.1 design document.</a>\n\n<a id=\"5\" href=\"https://eprint.iacr.org/2017/332.pdf\">[5] Forler, Christian, et al. \"Reforgeability of authenticated encryption schemes.\" Australasian Conference on Information Security and Privacy. Springer, Cham, 2017.</a>\n\n<a id=\"6\" href=\"https://eprint.iacr.org/2017/1147.pdf\">[6] Vaudenay, Serge, and Damian Vizár. \"Under Pressure: Security of Caesar Candidates beyond their Guarantees.\" IACR Cryptol. ePrint Arch. 2017 (2017): 1147.</a>", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown" ]

[ [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import os\nimport numpy as np\nimport pandas as pd\nimport seaborn as sns\nimport matplotlib.pyplot as plt\n%matplotlib inline\nimport torch\nfrom sklearn.model_selection import train_test_split\nfrom sklearn.ensemble import IsolationForest\n", "_____no_output_____" ], [ "cubism_path = \"/home/hopkinsl/Downloads/wikiart/wikiart/Cubism\"\nlistdir = os.listdir(cubism_path)\nimage_names = []\nlabels = []\nfor file in listdir:\n if file.startswith('pablo-picasso'):\n image_names.append(file)\n labels.append(0)\n elif file.startswith('marevna'):\n image_names.append(file)\n labels.append(1)\n elif file.startswith('fernand-leger'):\n image_names.append(file)\n labels.append(2)\nprint(labels.count(0))\nprint(labels.count(1))\nprint(labels.count(2))\nprint(listdir[0])", "203\n36\n141\npablo-picasso_las-meninas-velazquez-1957-1.jpg\n" ], [ "#Reading, reshaping, and normalizing images\n#images = plt.imread(data)\n#labels = pandas.read_csv()\n\nimages = torch.zeros\nfor file in image_names:\n \n\nimages = images.reshape(len(images),-1)\nimages = (images - images.mean()) / images.std())\nimages_train, images_test, labels_train, labels_test = train_test_split(images, labels, test_size=0.33, random_state=42)", "1\n" ], [ "\n\nclass MLP(torch.nn.Module):\n # this defines the model\n def __init__(self, sizes):\n super(MLP, self).__init__()\n print(sizes)\n self.sizes = sizes\n self.layers = [torch.nn.Linear(sizes[i], sizes[i+1]) for i in range(len(sizes)-1)]\n self.sigmoid = torch.nn.Sigmoid()\n self.relu = torch.nn.ReLU()\n self.softmax = torch.nn.Softmax()\n def forward(self, x):\n temp_activation = x\n temp_layer = self.layers[0](temp_activation)\n for i in range(1, len(layers)):\n temp_activation = self.sigmoid(temp_layer)\n temp_layer = self.layers[i](temp_activation)\n return self.softmax(temp_layer)\n\n", "_____no_output_____" ], [ "def train_model(training_data, test_data,training_labels,test_labels, model):\n # define the optimization\n criterion = torch.nn.BCELoss()\n optimizer = torch.optim.SGD(model.parameters(), lr=0.007,momentum=0.9)\n for epoch in range(10):\n # clear the gradient\n optimizer.zero_grad()\n # compute the model output\n myoutput = model(training_data)\n # calculate loss\n loss = criterion(myoutput, training_labels)\n # credit assignment\n loss.backward()\n # update model weights\n optimizer.step()\n\n # STUDENTS ADD THIS PART\n output_test = model(test_data)\n loss_test = criterion(output_test, test_labels)\n plt.plot(epoch,loss.detach().numpy(),'ko')\n plt.plot(epoch,loss_test.detach().numpy(),'ro')\n print(epoch,loss.detach().numpy())\n plt.show()\n\nsizes = [len(images[0,:]), 30,30,23]\ntrain_model(images_train, images_test, labels_train, labels_test, MLP())", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "empty" ] ] ]

[ "empty" ]

[ [ "empty" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Testing different SA methods 4/5\n## Textblob\n", "_____no_output_____" ] ], [ [ "import csv\nimport re\nimport random\n\nfrom textblob import TextBlob\n\n# Ugly hackery, but necessary: stackoverflow.com/questions/4383571/importing-files-from-different-folder\nimport sys\nsys.path.append('../../../')\n \nfrom src.streaming import spark_functions", "_____no_output_____" ], [ "preprocess = spark_functions.preprocessor()\ntokenize = spark_functions.tokenizer()\n\nwith open('./../../../data/interim/sanders_hydrated.csv') as csv_file:\n iterator = csv.reader(csv_file, delimiter=',')\n # Load the parts we need and preprocess as well as tokenize the text\n tweets = [(text, sentiment) for (topic, sentiment, id, text) in iterator if sentiment=='positive' or sentiment=='negative']\n # Shuffle for good measure\n random.shuffle(tweets)", "_____no_output_____" ], [ "import matplotlib.pyplot as plt\n\nresults = {\n \"positive\":{\"color\":\"green\",\"x\":[],\"y\":[]},\n \"neutral\":{\"color\":\"orange\",\"x\":[],\"y\":[]},\n \"negative\":{\"color\":\"red\",\"x\":[],\"y\":[]}\n}\n \n# Create plot\nfig = plt.figure()\nax = fig.add_subplot(1, 1, 1)\n \nfor (tweet,sentiment) in tweets[:200]:\n analysis = TextBlob(preprocess(tweet)).sentiment\n results[sentiment][\"x\"].append(analysis.polarity)\n results[sentiment][\"y\"].append(analysis.subjectivity)\n \nfor key in results:\n ax.scatter(results[key][\"x\"], results[key][\"y\"], alpha=0.8, c=results[key][\"color\"], edgecolors='none', label=key)\n \nplt.xlabel('polarity')\nplt.ylabel('subjectivity')\n\nplt.legend(loc=2)\nplt.savefig('textblob.pdf', format='pdf')\nplt.show()", "_____no_output_____" ], [ "# This is fucking hopeless\nlabeled_correctly = 0\nfor (tweet,sentiment) in tweets:\n analysis = TextBlob(preprocess(tweet)).sentiment\n if (analysis.polarity < 0 and sentiment == 'negative') \\\n or (analysis.polarity >= 0 and sentiment == 'positive'):\n labeled_correctly += 1\n\nprint(\"Labeled correctly: %d/%d = %.2d percent\" % (labeled_correctly, len(tweets), labeled_correctly/len(tweets)*100))", "Labeled correctly: 580/946 = 61 percent\n" ] ] ]

[ "markdown", "code" ]

[ [ "markdown" ], [ "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "### Easy string manipulation", "_____no_output_____" ] ], [ [ "x = 'a string'\ny = \"a string\"\nif x == y:\n print(\"they are the same\")", "they are the same\n" ], [ "fox = \"tHe qUICk bROWn fOx.\"", "_____no_output_____" ] ], [ [ "To convert the entire string into upper-case or lower-case, you can use the ``upper()`` or ``lower()`` methods respectively:", "_____no_output_____" ] ], [ [ "fox.upper()", "_____no_output_____" ], [ "fox.lower()", "_____no_output_____" ] ], [ [ "A common formatting need is to capitalize just the first letter of each word, or perhaps the first letter of each sentence.\nThis can be done with the ``title()`` and ``capitalize()`` methods:", "_____no_output_____" ] ], [ [ "fox.title()", "_____no_output_____" ], [ "fox.capitalize()", "_____no_output_____" ] ], [ [ "The cases can be swapped using the ``swapcase()`` method:", "_____no_output_____" ] ], [ [ "fox.swapcase()", "_____no_output_____" ], [ "line = ' this is the content '\nline.strip()", "_____no_output_____" ] ], [ [ "To remove just space to the right or left, use ``rstrip()`` or ``lstrip()`` respectively:", "_____no_output_____" ] ], [ [ "line.rstrip()", "_____no_output_____" ], [ "line.lstrip()", "_____no_output_____" ] ], [ [ "To remove characters other than spaces, you can pass the desired character to the ``strip()`` method:", "_____no_output_____" ] ], [ [ "num = \"000000000000435\"\nnum.strip('0')", "_____no_output_____" ], [ "line = 'the quick brown fox jumped over a lazy dog'\nline.find('fox')", "_____no_output_____" ], [ "line.index('fox')", "_____no_output_____" ], [ "line[16:21]", "_____no_output_____" ] ], [ [ "The only difference between ``find()`` and ``index()`` is their behavior when the search string is not found; ``find()`` returns ``-1``, while ``index()`` raises a ``ValueError``:", "_____no_output_____" ] ], [ [ "line.find('bear')", "_____no_output_____" ], [ "line.index('bear')", "_____no_output_____" ], [ "line.partition('fox')", "_____no_output_____" ] ], [ [ "The ``rpartition()`` method is similar, but searches from the right of the string.\n\nThe ``split()`` method is perhaps more useful; it finds *all* instances of the split-point and returns the substrings in between.\nThe default is to split on any whitespace, returning a list of the individual words in a string:", "_____no_output_____" ] ], [ [ "line_list = line.split()\nprint(line_list)", "['the', 'quick', 'brown', 'fox', 'jumped', 'over', 'a', 'lazy', 'dog']\n" ], [ "print(line_list[1])", "quick\n" ] ], [ [ "A related method is ``splitlines()``, which splits on newline characters.\nLet's do this with a Haiku, popularly attributed to the 17th-century poet Matsuo Bashō:", "_____no_output_____" ] ], [ [ "haiku = \"\"\"matsushima-ya\naah matsushima-ya\nmatsushima-ya\"\"\"\n\nhaiku.splitlines()", "_____no_output_____" ] ], [ [ "Note that if you would like to undo a ``split()``, you can use the ``join()`` method, which returns a string built from a splitpoint and an iterable:", "_____no_output_____" ] ], [ [ "'--'.join(['1', '2', '3'])", "_____no_output_____" ] ], [ [ "A common pattern is to use the special character ``\"\\n\"`` (newline) to join together lines that have been previously split, and recover the input:", "_____no_output_____" ] ], [ [ "print(\"\\n\".join(['matsushima-ya', 'aah matsushima-ya', 'matsushima-ya']))", "matsushima-ya\naah matsushima-ya\nmatsushima-ya\n" ], [ "pi = 3.14159\nstr(pi)", "_____no_output_____" ], [ "print (\"The value of pi is \" + pi)", "_____no_output_____" ] ], [ [ "Pi is a float number so it must be transform to sting.", "_____no_output_____" ] ], [ [ "print( \"The value of pi is \" + str(pi))", "The value of pi is 3.14159\n" ] ], [ [ "A more flexible way to do this is to use *format strings*, which are strings with special markers (noted by curly braces) into which string-formatted values will be inserted.\nHere is a basic example:", "_____no_output_____" ] ], [ [ "\"The value of pi is {}\".format(pi)", "_____no_output_____" ] ], [ [ "### Easy regex manipulation!", "_____no_output_____" ] ], [ [ "import re", "_____no_output_____" ], [ "line = 'the quick brown fox jumped over a lazy dog'", "_____no_output_____" ] ], [ [ "With this, we can see that the ``regex.search()`` method operates a lot like ``str.index()`` or ``str.find()``:", "_____no_output_____" ] ], [ [ "line.index('fox')", "_____no_output_____" ], [ "regex = re.compile('fox')\nmatch = regex.search(line)\nmatch.start()", "_____no_output_____" ] ], [ [ "Similarly, the ``regex.sub()`` method operates much like ``str.replace()``:", "_____no_output_____" ] ], [ [ "line.replace('fox', 'BEAR')", "_____no_output_____" ], [ "regex.sub('BEAR', line)", "_____no_output_____" ] ], [ [ "The following is a table of the repetition markers available for use in regular expressions:\n\n| Character | Description | Example |\n|-----------|-------------|---------|\n| ``?`` | Match zero or one repetitions of preceding | ``\"ab?\"`` matches ``\"a\"`` or ``\"ab\"`` |\n| ``*`` | Match zero or more repetitions of preceding | ``\"ab*\"`` matches ``\"a\"``, ``\"ab\"``, ``\"abb\"``, ``\"abbb\"``... |\n| ``+`` | Match one or more repetitions of preceding | ``\"ab+\"`` matches ``\"ab\"``, ``\"abb\"``, ``\"abbb\"``... but not ``\"a\"`` |\n| ``.`` | Any character | ``.*`` matches everything | \n| ``{n}`` | Match ``n`` repetitions of preeeding | ``\"ab{2}\"`` matches ``\"abb\"`` |\n| ``{m,n}`` | Match between ``m`` and ``n`` repetitions of preceding | ``\"ab{2,3}\"`` matches ``\"abb\"`` or ``\"abbb\"`` |", "_____no_output_____" ] ], [ [ "bool(re.search(r'ab', \"Boabab\"))", "_____no_output_____" ], [ "bool(re.search(r'.*ma.*', \"Ala ma kota\"))", "_____no_output_____" ], [ "bool(re.search(r'.*(psa|kota).*', \"Ala ma kota\"))", "_____no_output_____" ], [ "bool(re.search(r'.*(psa|kota).*', \"Ala ma psa\"))", "_____no_output_____" ], [ "bool(re.search(r'.*(psa|kota).*', \"Ala ma chomika\"))", "_____no_output_____" ], [ "zdanie = \"Ala ma kota.\"\nwzor = r'.*' #pasuje do każdego zdania\nzamiennik = r\"Ala ma psa.\"", "_____no_output_____" ], [ "re.sub(wzor, zamiennik, zdanie)", "_____no_output_____" ], [ "wzor = r'(.*)kota.'\nzamiennik = r\"\\1 psa.\"", "_____no_output_____" ], [ "re.sub(wzor, zamiennik, zdanie)", "_____no_output_____" ], [ "wzor = r'(.*)ma(.*)'\nzamiennik = r\"\\1 posiada \\2\"", "_____no_output_____" ], [ "re.sub(wzor, zamiennik, zdanie)", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "import re\nimport os\nimport keras.backend as K\nimport numpy as np\nimport pandas as pd\nfrom keras import layers, models, utils\nimport json", "Using TensorFlow backend.\n" ], [ "def reset_everything():\n import tensorflow as tf\n %reset -f in out dhist\n tf.reset_default_graph()\n K.set_session(tf.InteractiveSession())", "_____no_output_____" ], [ "# Constants for our networks. We keep these deliberately small to reduce training time.\n\nVOCAB_SIZE = 250000\nEMBEDDING_SIZE = 100\nMAX_DOC_LEN = 128\nMIN_DOC_LEN = 12", "_____no_output_____" ], [ "def extract_stackexchange(filename, limit=1000000):\n json_file = filename + 'limit=%s.json' % limit\n\n rows = []\n for i, line in enumerate(os.popen('7z x -so \"%s\" Posts.xml' % filename)):\n line = str(line)\n if not line.startswith(' <row'):\n continue\n \n if i % 1000 == 0:\n print('\\r%05d/%05d' % (i, limit), end='', flush=True)\n\n parts = line[6:-5].split('\"')\n record = {}\n for i in range(0, len(parts), 2):\n k = parts[i].replace('=', '').strip()\n v = parts[i+1].strip()\n record[k] = v\n rows.append(record)\n \n if len(rows) > limit:\n break\n \n with open(json_file, 'w') as fout:\n json.dump(rows, fout)\n \n return rows\n\n\nxml_7z = utils.get_file(\n fname='travel.stackexchange.com.7z',\n origin='https://ia800107.us.archive.org/27/items/stackexchange/travel.stackexchange.com.7z',\n)\nprint()\n\nrows = extract_stackexchange(xml_7z)", "\n87000/1000000" ] ], [ [ "# Data Exploration\n\nNow that we have extracted our data, let's clean it up and take a look at what we have to work with.", "_____no_output_____" ] ], [ [ "df = pd.DataFrame.from_records(rows) \ndf = df.set_index('Id', drop=False)\ndf['Title'] = df['Title'].fillna('').astype('str')\ndf['Tags'] = df['Tags'].fillna('').astype('str')\ndf['Body'] = df['Body'].fillna('').astype('str')\ndf['Id'] = df['Id'].astype('int')\ndf['PostTypeId'] = df['PostTypeId'].astype('int')\ndf['ViewCount'] = df['ViewCount'].astype('float')\n\ndf.head()", "_____no_output_____" ], [ "list(df[df['ViewCount'] > 250000]['Title'])", "_____no_output_____" ], [ "from keras.preprocessing.text import Tokenizer\nfrom keras.preprocessing.sequence import pad_sequences\n\ntokenizer = Tokenizer(num_words=VOCAB_SIZE)\ntokenizer.fit_on_texts(df['Body'] + df['Title'])", "_____no_output_____" ], [ "# Compute TF/IDF Values\n\ntotal_count = sum(tokenizer.word_counts.values())\nidf = { k: np.log(total_count/v) for (k,v) in tokenizer.word_counts.items() }", "_____no_output_____" ], [ "# Download pre-trained word2vec embeddings\n\nimport gensim\n\nglove_100d = utils.get_file(\n fname='glove.6B.100d.txt',\n origin='https://storage.googleapis.com/deep-learning-cookbook/glove.6B.100d.txt',\n)\n\nw2v_100d = glove_100d + '.w2v'\nfrom gensim.scripts.glove2word2vec import glove2word2vec\nglove2word2vec(glove_100d, w2v_100d)\nw2v_model = gensim.models.KeyedVectors.load_word2vec_format(w2v_100d)\n\nw2v_weights = np.zeros((VOCAB_SIZE, w2v_model.syn0.shape[1]))\nidf_weights = np.zeros((VOCAB_SIZE, 1))\n\nfor k, v in tokenizer.word_index.items():\n if v >= VOCAB_SIZE:\n continue\n \n if k in w2v_model:\n w2v_weights[v] = w2v_model[k]\n \n idf_weights[v] = idf[k]\n \ndel w2v_model", "_____no_output_____" ], [ "df['title_tokens'] = tokenizer.texts_to_sequences(df['Title'])\ndf['body_tokens'] = tokenizer.texts_to_sequences(df['Body'])", "_____no_output_____" ], [ "import random\n\n# We can create a data generator that will randomly title and body tokens for questions. We'll use random text\n# from other questions as a negative example when necessary.\ndef data_generator(batch_size, negative_samples=1):\n questions = df[df['PostTypeId'] == 1]\n all_q_ids = list(questions.index)\n \n batch_x_a = []\n batch_x_b = []\n batch_y = []\n \n def _add(x_a, x_b, y):\n batch_x_a.append(x_a[:MAX_DOC_LEN])\n batch_x_b.append(x_b[:MAX_DOC_LEN])\n batch_y.append(y)\n \n while True:\n questions = questions.sample(frac=1.0)\n \n for i, q in questions.iterrows():\n _add(q['title_tokens'], q['body_tokens'], 1)\n \n negative_q = random.sample(all_q_ids, negative_samples)\n for nq_id in negative_q:\n _add(q['title_tokens'], df.at[nq_id, 'body_tokens'], 0) \n \n if len(batch_y) >= batch_size:\n yield ({\n 'title': pad_sequences(batch_x_a, maxlen=None),\n 'body': pad_sequences(batch_x_b, maxlen=None),\n }, np.asarray(batch_y))\n \n batch_x_a = []\n batch_x_b = []\n batch_y = []\n\n# dg = data_generator(1, 2)\n# next(dg)\n# next(dg)", "_____no_output_____" ] ], [ [ "# Embedding Lookups\n\nLet's define a helper class for looking up our embedding results. We'll use it\nto verify our models.", "_____no_output_____" ] ], [ [ "questions = df[df['PostTypeId'] == 1]['Title'].reset_index(drop=True)\nquestion_tokens = pad_sequences(tokenizer.texts_to_sequences(questions))\n\nclass EmbeddingWrapper(object):\n def __init__(self, model):\n self._r = questions\n self._i = {i:s for (i, s) in enumerate(questions)}\n self._w = model.predict({'title': question_tokens}, verbose=1, batch_size=1024)\n self._model = model\n self._norm = np.sqrt(np.sum(self._w * self._w + 1e-5, axis=1))\n\n def nearest(self, sentence, n=10):\n x = tokenizer.texts_to_sequences([sentence])\n if len(x[0]) < MIN_DOC_LEN:\n x[0] += [0] * (MIN_DOC_LEN - len(x))\n e = self._model.predict(np.asarray(x))[0]\n norm_e = np.sqrt(np.dot(e, e))\n dist = np.dot(self._w, e) / (norm_e * self._norm)\n\n top_idx = np.argsort(dist)[-n:]\n return pd.DataFrame.from_records([\n {'question': self._r[i], 'dist': float(dist[i])}\n for i in top_idx\n ])", "_____no_output_____" ], [ "# Our first model will just sum up the embeddings of each token.\n# The similarity between documents will be the dot product of the final embedding.\n\nimport tensorflow as tf\n\ndef sum_model(embedding_size, vocab_size, embedding_weights=None, idf_weights=None):\n title = layers.Input(shape=(None,), dtype='int32', name='title')\n body = layers.Input(shape=(None,), dtype='int32', name='body')\n\n def make_embedding(name):\n if embedding_weights is not None:\n embedding = layers.Embedding(mask_zero=True, input_dim=vocab_size, output_dim=w2v_weights.shape[1], \n weights=[w2v_weights], trainable=False, \n name='%s/embedding' % name)\n else:\n embedding = layers.Embedding(mask_zero=True, input_dim=vocab_size, output_dim=embedding_size,\n name='%s/embedding' % name)\n\n if idf_weights is not None:\n idf = layers.Embedding(mask_zero=True, input_dim=vocab_size, output_dim=1, \n weights=[idf_weights], trainable=False,\n name='%s/idf' % name)\n else:\n idf = layers.Embedding(mask_zero=True, input_dim=vocab_size, output_dim=1,\n name='%s/idf' % name)\n \n return embedding, idf\n \n embedding_a, idf_a = make_embedding('a')\n embedding_b, idf_b = embedding_a, idf_a\n# embedding_b, idf_b = make_embedding('b')\n\n mask = layers.Masking(mask_value=0)\n def _combine_and_sum(args):\n [embedding, idf] = args\n return K.sum(embedding * K.abs(idf), axis=1)\n\n sum_layer = layers.Lambda(_combine_and_sum, name='combine_and_sum')\n\n sum_a = sum_layer([mask(embedding_a(title)), idf_a(title)])\n sum_b = sum_layer([mask(embedding_b(body)), idf_b(body)])\n\n sim = layers.dot([sum_a, sum_b], axes=1, normalize=True)\n sim_model = models.Model(\n inputs=[title, body],\n outputs=[sim],\n )\n sim_model.compile(loss='binary_crossentropy', optimizer='nadam', metrics=['accuracy'])\n sim_model.summary()\n\n embedding_model = models.Model(\n inputs=[title],\n outputs=[sum_a]\n )\n return sim_model, embedding_model", "_____no_output_____" ], [ "# Try using our model with pretrained weights from word2vec\n\nsum_model_precomputed, sum_embedding_precomputed = sum_model(\n embedding_size=EMBEDDING_SIZE, vocab_size=VOCAB_SIZE,\n embedding_weights=w2v_weights, idf_weights=idf_weights\n)\n\nx, y = next(data_generator(batch_size=4096))\nsum_model_precomputed.evaluate(x, y)", "_____no_output_____" ], [ "SAMPLE_QUESTIONS = [\n 'Roundtrip ticket versus one way',\n 'Shinkansen from Kyoto to Hiroshima',\n 'Bus tour of Germany',\n]\n\ndef evaluate_sample(lookup):\n pd.set_option('display.max_colwidth', 100)\n results = []\n for q in SAMPLE_QUESTIONS:\n print(q)\n q_res = lookup.nearest(q, n=4)\n q_res['result'] = q_res['question']\n q_res['question'] = q\n results.append(q_res)\n\n return pd.concat(results)\n\nlookup = EmbeddingWrapper(model=sum_embedding_precomputed)\nevaluate_sample(lookup)", "_____no_output_____" ] ], [ [ "# Training our own network\n\nThe results are okay but not great... instead of using the word2vec embeddings, what happens if we train our network end-to-end?", "_____no_output_____" ] ], [ [ "sum_model_trained, sum_embedding_trained = sum_model(\n embedding_size=EMBEDDING_SIZE, vocab_size=VOCAB_SIZE, \n embedding_weights=None,\n idf_weights=None\n)\nsum_model_trained.fit_generator(\n data_generator(batch_size=128),\n epochs=10,\n steps_per_epoch=1000\n)", "_____no_output_____" ], [ "lookup = EmbeddingWrapper(model=sum_embedding_trained)\nevaluate_sample(lookup)", "_____no_output_____" ] ], [ [ "## CNN Model\n\nUsing a sum-of-embeddings model works well. What happens if we try to make a simple CNN model?", "_____no_output_____" ] ], [ [ "def cnn_model(embedding_size, vocab_size):\n title = layers.Input(shape=(None,), dtype='int32', name='title')\n body = layers.Input(shape=(None,), dtype='int32', name='body')\n\n embedding = layers.Embedding(\n mask_zero=False,\n input_dim=vocab_size,\n output_dim=embedding_size,\n )\n\n\n def _combine_sum(v):\n return K.sum(v, axis=1)\n\n cnn_1 = layers.Convolution1D(256, 3)\n cnn_2 = layers.Convolution1D(256, 3)\n cnn_3 = layers.Convolution1D(256, 3)\n \n global_pool = layers.GlobalMaxPooling1D()\n local_pool = layers.MaxPooling1D(strides=2, pool_size=3)\n\n def forward(input):\n embed = embedding(input)\n return global_pool(\n cnn_2(local_pool(cnn_1(embed))))\n\n sum_a = forward(title)\n sum_b = forward(body)\n\n sim = layers.dot([sum_a, sum_b], axes=1, normalize=False)\n sim_model = models.Model(\n inputs=[title, body],\n outputs=[sim],\n )\n sim_model.compile(loss='binary_crossentropy', optimizer='rmsprop', metrics=['accuracy'])\n\n embedding_model = models.Model(\n inputs=[title],\n outputs=[sum_a]\n )\n return sim_model, embedding_model", "_____no_output_____" ], [ "cnn, cnn_embedding = cnn_model(embedding_size=25, vocab_size=VOCAB_SIZE)\ncnn.summary()\ncnn.fit_generator(\n data_generator(batch_size=128),\n epochs=10,\n steps_per_epoch=1000,\n)", "_____no_output_____" ], [ "lookup = EmbeddingWrapper(model=cnn_embedding)\nevaluate_sample(lookup)", "_____no_output_____" ] ], [ [ "## LSTM Model\n\nWe can also make an LSTM model. Warning, this will be very slow to train and evaluate unless you have a relatively fast GPU to run it on!", "_____no_output_____" ] ], [ [ "def lstm_model(embedding_size, vocab_size):\n title = layers.Input(shape=(None,), dtype='int32', name='title')\n body = layers.Input(shape=(None,), dtype='int32', name='body')\n\n embedding = layers.Embedding(\n mask_zero=True,\n input_dim=vocab_size,\n output_dim=embedding_size,\n# weights=[w2v_weights],\n# trainable=False\n )\n\n lstm_1 = layers.LSTM(units=512, return_sequences=True)\n lstm_2 = layers.LSTM(units=512, return_sequences=False)\n \n sum_a = lstm_2(lstm_1(embedding(title)))\n sum_b = lstm_2(lstm_1(embedding(body)))\n\n sim = layers.dot([sum_a, sum_b], axes=1, normalize=True)\n# sim = layers.Activation(activation='sigmoid')(sim)\n sim_model = models.Model(\n inputs=[title, body],\n outputs=[sim],\n )\n sim_model.compile(loss='binary_crossentropy', optimizer='rmsprop')\n\n embedding_model = models.Model(\n inputs=[title],\n outputs=[sum_a]\n )\n return sim_model, embedding_model", "_____no_output_____" ], [ "lstm, lstm_embedding = lstm_model(embedding_size=EMBEDDING_SIZE, vocab_size=VOCAB_SIZE)\nlstm.summary()\nlstm.fit_generator(\n data_generator(batch_size=128),\n epochs=10,\n steps_per_epoch=100,\n)", "_____no_output_____" ], [ "lookup = EmbeddingWrapper(model=lstm_embedding)\nevaluate_sample(lookup)", "_____no_output_____" ] ] ]

[ "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import os, sys\n# os.environ['TF_FORCE_GPU_ALLOW_GROWTH'] = 'true'\nos.environ['CUDA_VISIBLE_DEVICES'] = '1'\nimport tensorflow as tf\ntf.compat.v1.enable_eager_execution()\n\nimport numpy as np\nimport imageio\nimport json\nimport random\nimport time\nimport pprint\n\nimport matplotlib.pyplot as plt\n\nimport run_nerf\n\nfrom load_llff import load_llff_data\nfrom load_deepvoxels import load_dv_data\nfrom load_blender import load_blender_data\n", "_____no_output_____" ], [ "basedir = './logs'\nexpname = 'fern_example'\n\nconfig = os.path.join(basedir, expname, 'config.txt')\nprint('Args:')\nprint(open(config, 'r').read())\nparser = run_nerf.config_parser()\n\nargs = parser.parse_args('--config {} --ft_path {}'.format(config, os.path.join(basedir, expname, 'model_200000.npy')))\nprint('loaded args')\n\nimages, poses, bds, render_poses, i_test = load_llff_data(args.datadir, args.factor, \n recenter=True, bd_factor=.75, \n spherify=args.spherify)\nH, W, focal = poses[0,:3,-1].astype(np.float32)\n\nH = int(H)\nW = int(W)\nhwf = [H, W, focal]\n\nimages = images.astype(np.float32)\nposes = poses.astype(np.float32)\n\nif args.no_ndc:\n near = tf.reduce_min(bds) * .9\n far = tf.reduce_max(bds) * 1.\nelse:\n near = 0.\n far = 1.\n\n", "Args:\ndatadir = ./data/nerf_llff_data/fern\ndataset_type = llff\nfactor = 4\n\nno_batching = False\ni_embed = 0\nN_samples = 64\nN_importance = 128\nuse_viewdirs = True\nlrate_decay = 250\n\nllffhold = 8\n\nN_rand = 4096\nloaded args\nLoaded image data (756, 1008, 3, 20) [ 756. 1008. 815.13158322]\nLoaded ./data/nerf_llff_data/fern 16.985296178676084 80.00209740336334\nrecentered (3, 5)\n[[ 1.0000000e+00 0.0000000e+00 0.0000000e+00 1.4901161e-09]\n [ 0.0000000e+00 1.0000000e+00 -1.8730975e-09 -9.6857544e-09]\n [-0.0000000e+00 1.8730975e-09 1.0000000e+00 0.0000000e+00]]\nData:\n(20, 3, 5) (20, 756, 1008, 3) (20, 2)\nHOLDOUT view is 12\n" ], [ "# Create nerf model\n_, render_kwargs_test, start, grad_vars, models = run_nerf.create_nerf(args)\n\nbds_dict = {\n 'near' : tf.cast(near, tf.float32),\n 'far' : tf.cast(far, tf.float32),\n}\nrender_kwargs_test.update(bds_dict)\n\nprint('Render kwargs:')\npprint.pprint(render_kwargs_test)\n\n\ndown = 4\nrender_kwargs_fast = {k : render_kwargs_test[k] for k in render_kwargs_test}\nrender_kwargs_fast['N_importance'] = 0\n\nc2w = np.eye(4)[:3,:4].astype(np.float32) # identity pose matrix\ntest = run_nerf.render(H//down, W//down, focal/down, c2w=c2w, **render_kwargs_fast)\nimg = np.clip(test[0],0,1)\nplt.imshow(img)\nplt.show()", "2021-10-24 23:28:33.855478: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcuda.so.1\n2021-10-24 23:28:33.898606: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1618] Found device 0 with properties: \nname: NVIDIA GeForce GTX 1080 Ti major: 6 minor: 1 memoryClockRate(GHz): 1.6575\npciBusID: 0000:03:00.0\n2021-10-24 23:28:33.898823: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudart.so.10.0\n2021-10-24 23:28:33.900023: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcublas.so.10.0\n2021-10-24 23:28:33.901057: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcufft.so.10.0\n2021-10-24 23:28:33.901293: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcurand.so.10.0\n2021-10-24 23:28:33.902667: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcusolver.so.10.0\n2021-10-24 23:28:33.904100: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcusparse.so.10.0\n2021-10-24 23:28:33.907445: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudnn.so.7\n2021-10-24 23:28:33.909616: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1746] Adding visible gpu devices: 0\n2021-10-24 23:28:33.910383: I tensorflow/core/platform/cpu_feature_guard.cc:142] Your CPU supports instructions that this TensorFlow binary was not compiled to use: SSE4.1 SSE4.2 AVX AVX2 FMA\n2021-10-24 23:28:33.939562: I tensorflow/core/platform/profile_utils/cpu_utils.cc:94] CPU Frequency: 3491830000 Hz\n2021-10-24 23:28:33.941574: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x55967b42abf0 initialized for platform Host (this does not guarantee that XLA will be used). Devices:\n2021-10-24 23:28:33.941629: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): Host, Default Version\n2021-10-24 23:28:34.097235: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x55967b298ef0 initialized for platform CUDA (this does not guarantee that XLA will be used). Devices:\n2021-10-24 23:28:34.097277: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): NVIDIA GeForce GTX 1080 Ti, Compute Capability 6.1\n2021-10-24 23:28:34.099082: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1618] Found device 0 with properties: \nname: NVIDIA GeForce GTX 1080 Ti major: 6 minor: 1 memoryClockRate(GHz): 1.6575\npciBusID: 0000:03:00.0\n2021-10-24 23:28:34.099153: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudart.so.10.0\n2021-10-24 23:28:34.099182: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcublas.so.10.0\n2021-10-24 23:28:34.099208: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcufft.so.10.0\n2021-10-24 23:28:34.099254: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcurand.so.10.0\n2021-10-24 23:28:34.099283: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcusolver.so.10.0\n2021-10-24 23:28:34.099309: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcusparse.so.10.0\n2021-10-24 23:28:34.099335: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudnn.so.7\n2021-10-24 23:28:34.103922: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1746] Adding visible gpu devices: 0\n2021-10-24 23:28:34.104020: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudart.so.10.0\n2021-10-24 23:28:34.105825: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1159] Device interconnect StreamExecutor with strength 1 edge matrix:\n2021-10-24 23:28:34.105852: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1165] 0 \n2021-10-24 23:28:34.105864: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1178] 0: N \n2021-10-24 23:28:34.108282: W tensorflow/core/common_runtime/gpu/gpu_bfc_allocator.cc:39] Overriding allow_growth setting because the TF_FORCE_GPU_ALLOW_GROWTH environment variable is set. Original config value was 0.\n2021-10-24 23:28:34.108311: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1304] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 10482 MB memory) -> physical GPU (device: 0, name: NVIDIA GeForce GTX 1080 Ti, pci bus id: 0000:03:00.0, compute capability: 6.1)\n" ], [ "down = 8 # trade off resolution+aliasing for render speed to make this video faster\nframes = []\nfor i, c2w in enumerate(render_poses):\n if i%8==0: print(i)\n test = run_nerf.render(H//down, W//down, focal/down, c2w=c2w[:3,:4], **render_kwargs_fast)\n frames.append((255*np.clip(test[0],0,1)).astype(np.uint8))\n \nprint('done, saving')\nf = 'logs/fern_example/video.mp4'\nimageio.mimwrite(f, frames, fps=30, quality=8)\n\nfrom IPython.display import Video\nVideo(f, height=320)", "0\n8\n16\n24\n32\n40\n48\n56\n64\n72\n80\n88\n96\n104\n112\n" ], [ "%matplotlib inline\nfrom ipywidgets import interactive, widgets\nimport matplotlib.pyplot as plt\nimport numpy as np\n\n\ndef f(x, y, z):\n \n c2w = tf.convert_to_tensor([\n [1,0,0,x],\n [0,1,0,y],\n [0,0,1,z],\n [0,0,0,1],\n ], dtype=tf.float32)\n \n test = run_nerf.render(H//down, W//down, focal/down, c2w=c2w, **render_kwargs_fast)\n img = np.clip(test[0],0,1)\n \n plt.figure(2, figsize=(20,6))\n plt.imshow(img)\n plt.show()\n \n\nsldr = lambda : widgets.FloatSlider(\n value=0.,\n min=-1.,\n max=1.,\n step=.01,\n)\n\nnames = ['x', 'y', 'z']\n \ninteractive_plot = interactive(f, **{n : sldr() for n in names})\ninteractive_plot", "_____no_output_____" ], [ "!conda install -c conda-forge ipywidgets", "Collecting package metadata (current_repodata.json): | " ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "from platform import python_version\nimport tensorflow as tf\n\nprint(tf.test.is_gpu_available())\nprint(python_version())", "True\n3.7.5\n" ], [ "import os\nimport numpy as np\nfrom os import listdir\nfrom PIL import Image\nimport time\nimport tensorflow as tf\nfrom tensorflow.keras import layers,models,optimizers\nfrom keras import backend as K\nimport matplotlib.pyplot as plt", "_____no_output_____" ], [ "path1=\"datasets/ofg_family/\"\npath2=\"datasets/TSKinFace_Data/TSKinFace_cropped/\"\n\nrandomiser = np.random.RandomState(123)", "_____no_output_____" ], [ "img_size = 64\nmean = 0.0009\nstd_dev = 0.009\nlr = 0.0005\nb1 = 0.875\nb2 = 0.975\nsd_random_normal_init = 0.02\n\nEPOCHS = 10\nbatch = 10", "_____no_output_____" ], [ "def generate_image_1(family_dir):\n dic={}\n sub=[a for a in listdir(path1+\"/\"+family_dir)]\n \n for ele in sub:\n if ele == '.DS_Store':\n continue;\n mypath = path1+\"/\"+family_dir+\"/\"+ele+\"/\"\n onlyfiles = [mypath+f for f in listdir(mypath)]\n \n addr = randomiser.choice(onlyfiles)\n original_img = np.array(Image.open(addr).resize((64,64),Image.ANTIALIAS))\n if ele[0].lower()=='f':\n dic['father'] = original_img\n elif ele[0].lower()=='m':\n dic['mother'] = original_img\n elif ele.lower()=='child_male':\n dic['child'] = original_img \n dic['gender']=np.zeros((original_img.shape))\n elif ele.lower()=='child_female':\n dic['child'] = original_img \n dic['gender'] = np.ones((original_img.shape))\n return [dic['father'],dic['mother'],dic['gender'],dic['child']]\n\ndef generate_image_2(family_dir, family_number, gender):\n dic={}\n sub = [\"F\" , \"M\", gender]\n family_pth = path2+\"/\"+family_dir+\"/\" + family_dir + \"-\" + str(family_number) + \"-\"\n for ele in sub:\n addr = family_pth+ele+\".jpg\"\n original_img = np.array(Image.open(addr).resize((64,64),Image.ANTIALIAS))\n if ele =='F':\n dic['father'] = original_img\n elif ele == 'M':\n dic['mother'] = original_img\n elif ele == 'S':\n dic['child'] = original_img \n dic['gender']=np.zeros((original_img.shape))\n elif ele == 'D':\n dic['child'] = original_img \n dic['gender'] = np.ones((original_img.shape))\n return [dic['father'],dic['mother'],dic['gender'],dic['child']]\n\ndef generate_batch(families_batch):\n np_images=[]\n for family in families_batch:\n if(len(family) == 3):\n res = generate_image_2(family[0], family[1], family[2])\n elif(len(family) == 1):\n res = generate_image_1(family[0])\n if( res != None):\n np_images.append(res)\n \n return np_images", "_____no_output_____" ], [ "for r, d, f in os.walk(path1):\n all_families = d\n break \nall_families = [[family] for family in all_families] \n\nfor i in range(285):\n all_families.append(['FMS', i+1, 'S'])\nfor i in range(274):\n all_families.append(['FMD', i+1, 'D'])\n \nfor i in range(228):\n all_families.append(['FMSD', i+1, 'D']) \n all_families.append(['FMSD', i+1, 'S']) \n\n\nrandomiser.shuffle(all_families)\n\ntrain_families = all_families[:-100]\ntest_families = all_families[-100:]", "_____no_output_____" ], [ "OUTPUT_CHANNELS = 3", "_____no_output_____" ], [ "def gen_downsample_parent(filters, size, apply_batchnorm=True, apply_dropout=False):\n initializer = tf.random_normal_initializer(mean, std_dev) \n \n\n result = tf.keras.Sequential()\n result.add(\n tf.keras.layers.Conv2D(filters, size, strides=2, padding='same',\n kernel_initializer=initializer,\n use_bias=False))\n\n if apply_batchnorm:\n result.add(tf.keras.layers.BatchNormalization())\n \n result.add(tf.keras.layers.ELU())\n \n if apply_dropout:\n result.add(tf.keras.layers.Dropout(rate = 0.5))\n\n return result", "_____no_output_____" ] ], [ [ "def gen_downsample_noise(filters, size, apply_batchnorm=True):\n initializer = tf.random_normal_initializer(mean, std_dev) \n \n\n result = tf.keras.Sequential()\n result.add(\n tf.keras.layers.Conv2DTranspose(filters, size, strides=2, padding='same',\n kernel_initializer=initializer,\n use_bias=False))\n\n if apply_batchnorm:\n result.add(tf.keras.layers.BatchNormalization())\n \n result.add(tf.keras.layers.ELU())\n\n return result", "_____no_output_____" ] ], [ [ "def gen_upsample(filters, size,apply_batchnorm = False):\n initializer = tf.random_normal_initializer(mean, std_dev)\n\n result = tf.keras.Sequential()\n result.add(\n tf.keras.layers.Conv2DTranspose(filters, size, strides=2,\n padding='same',\n kernel_initializer=initializer,\n use_bias=False))\n if apply_batchnorm:\n result.add(tf.keras.layers.BatchNormalization())\n result.add(tf.keras.layers.ELU())\n\n return result", "_____no_output_____" ], [ "def EncoderNN():\n down_stack_parent = [\n gen_downsample_parent(32,4,apply_batchnorm=True, apply_dropout=False),\n gen_downsample_parent(64,4,apply_batchnorm=True, apply_dropout=False)\n ]\n \n# down_stack_noise =[\n# # z = 4x4x64\n# gen_downsample_noise(64,4,apply_batchnorm=True), #8x8x64\n# gen_downsample_noise(32,4,apply_batchnorm=True) #16x16x32 \n# ]\n \n final_conv =[\n gen_upsample(32,4 ,apply_batchnorm = True)\n ]\n \n initializer = tf.random_normal_initializer(mean, sd_random_normal_init)\n last = tf.keras.layers.Conv2DTranspose(OUTPUT_CHANNELS, 4,\n strides=2,\n padding='same',\n kernel_initializer=initializer,\n activation='tanh')\n\n concat = tf.keras.layers.Concatenate()\n\n father = tf.keras.layers.Input(shape=(img_size,img_size,3))\n mother = tf.keras.layers.Input(shape=(img_size,img_size,3))\n\n \n \n x1 = father\n for down in down_stack_parent:\n x1 = down(x1)\n \n# print(x1.shape)\n \n x2 = mother\n for down in down_stack_parent:\n x2 = down(x2) \n \n# print(x2.shape)\n \n final = concat([x1,x2])\n# print(final.shape)\n final = final_conv[0](final)\n \n final = last(final)\n# print(final.shape)\n return tf.keras.Model(inputs=[father, mother], outputs=final)", "_____no_output_____" ], [ "encoder_optimizer = tf.keras.optimizers.Adam(learning_rate = lr, beta_1=b1)", "_____no_output_____" ], [ "def tensor_to_array(tensor1):\n return tensor1.numpy()", "_____no_output_____" ], [ "def train_encoder(father_batch, mother_batch, target_batch, b_size):\n with tf.GradientTape() as enc_tape:\n gen_outputs = encoder([father_batch, mother_batch], training=True)\n \n diff = tf.abs(target_batch - gen_outputs)\n flatten_diff = tf.reshape(diff, (b_size, img_size*img_size*3))\n \n encoder_loss_batch = tf.reduce_mean(flatten_diff, axis=1)\n encoder_loss = tf.reduce_mean(encoder_loss_batch)\n \n print(\"ENCODER_LOSS: \",tensor_to_array(encoder_loss))\n #calculate gradients\n encoder_gradients = enc_tape.gradient(encoder_loss,encoder.trainable_variables)\n\n #apply gradients on optimizer\n encoder_optimizer.apply_gradients(zip(encoder_gradients,encoder.trainable_variables))\n \n", "_____no_output_____" ], [ "def fit_encoder(train_ds, epochs, test_ds, batch):\n losses=np.array([])\n for epoch in range(epochs):\n print(\"______________________________EPOCH %d_______________________________\"%(epoch+1))\n start = time.time()\n for i in range(len(train_ds)//batch):\n batch_data = np.asarray(generate_batch(train_ds[i*batch:(i+1)*batch]))\n batch_data = batch_data / 255 * 2 -1\n \n \n print(\"Generated batch\", batch_data.shape)\n\n X_Father_train = tf.convert_to_tensor(batch_data[:,0],dtype =tf.float32)\n X_Mother_train = tf.convert_to_tensor(batch_data[:,1],dtype =tf.float32)\n Y_train = tf.convert_to_tensor(batch_data[:,3],dtype =tf.float32)\n \n train_encoder(X_Father_train, X_Mother_train, Y_train,batch)\n \n print(\"Trained for batch %d/%d\"%(i+1,(len(train_ds)//batch)))\n print(\"______________________________TRAINING COMPLETED_______________________________\")", "_____no_output_____" ], [ "train_dataset = all_families[:-100]\ntest_dataset = all_families[-100:]\nencoder = EncoderNN()\n\nwith tf.device('/gpu:0'):\n fit_encoder(train_dataset, EPOCHS, test_dataset,batch)", "______________________________EPOCH 1_______________________________\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.40341973\nTrained for batch 1/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.3747058\nTrained for batch 2/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.4229235\nTrained for batch 3/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.4062442\nTrained for batch 4/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.37259284\nTrained for batch 5/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.40570754\nTrained for batch 6/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.43097487\nTrained for batch 7/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.37538522\nTrained for batch 8/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.3661008\nTrained for batch 9/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.36800843\nTrained for batch 10/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.3724699\nTrained for batch 11/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.43448296\nTrained for batch 12/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.35709864\nTrained for batch 13/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.3950496\nTrained for batch 14/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.3688807\nTrained for batch 15/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.3860542\nTrained for batch 16/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.44605786\nTrained for batch 17/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.34664398\nTrained for batch 18/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.39454833\nTrained for batch 19/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.33471403\nTrained for batch 20/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.3614666\nTrained for batch 21/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.40210095\nTrained for batch 22/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.42147255\nTrained for batch 23/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.35217696\nTrained for batch 24/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.38657412\nTrained for batch 25/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.41433\nTrained for batch 26/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.35280803\nTrained for batch 27/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.32902092\nTrained for batch 28/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.365082\nTrained for batch 29/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.36655417\nTrained for batch 30/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.40149218\nTrained for batch 31/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.35119835\nTrained for batch 32/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.33154935\nTrained for batch 33/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.38874537\nTrained for batch 34/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.3770574\nTrained for batch 35/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.37216097\nTrained for batch 36/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.30604583\nTrained for batch 37/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.34849322\nTrained for batch 38/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.34970385\nTrained for batch 39/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.35399386\nTrained for batch 40/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.37763038\nTrained for batch 41/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.3642118\nTrained for batch 42/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.32215607\nTrained for batch 43/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.34084243\nTrained for batch 44/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.3115418\nTrained for batch 45/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.405502\nTrained for batch 46/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.33786473\nTrained for batch 47/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.34356505\nTrained for batch 48/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.34688368\nTrained for batch 49/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.35725087\nTrained for batch 50/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.40099317\nTrained for batch 51/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.31783652\nTrained for batch 52/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.3588978\nTrained for batch 53/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.3546448\nTrained for batch 54/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.38290676\nTrained for batch 55/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.35456428\nTrained for batch 56/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.4008767\nTrained for batch 57/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.39036384\nTrained for batch 58/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.3118421\nTrained for batch 59/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.40766144\nTrained for batch 60/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.33152086\nTrained for batch 61/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.318183\nTrained for batch 62/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.3289036\nTrained for batch 63/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.30764475\nTrained for batch 64/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.2862015\nTrained for batch 65/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.3151431\nTrained for batch 66/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.35122877\nTrained for batch 67/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.38102263\nTrained for batch 68/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.31507578\nTrained for batch 69/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.3562093\nTrained for batch 70/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.34080186\nTrained for batch 71/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.34523338\nTrained for batch 72/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.3719592\nTrained for batch 73/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.35613218\nTrained for batch 74/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.36692142\nTrained for batch 75/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.3916842\nTrained for batch 76/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.31589094\nTrained for batch 77/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.3126195\nTrained for batch 78/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.36948365\nTrained for batch 79/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.40351337\nTrained for batch 80/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.33806187\nTrained for batch 81/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.31995544\nTrained for batch 82/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.3503887\nTrained for batch 83/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.3623436\nTrained for batch 84/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.35348386\nTrained for batch 85/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.39553097\nTrained for batch 86/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.3288309\nTrained for batch 87/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.40606794\nTrained for batch 88/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.32926646\nTrained for batch 89/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.34279853\nTrained for batch 90/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.3104426\nTrained for batch 91/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.38569975\nTrained for batch 92/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.38053983\nTrained for batch 93/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.34726876\nTrained for batch 94/297\nGenerated batch (10, 4, 64, 64, 3)\nENCODER_LOSS: 0.3098844\nTrained for batch 95/297\n" ], [ "f_no = 1106\nfamily_data = generate_batch([all_families[f_no]])\ninp = [family_data[0][0],family_data[0][1]]\ninp = tf.cast(inp, tf.float32)\n\nfather_inp = inp[0][tf.newaxis,...]\nmother_inp = inp[1][tf.newaxis,...]\n\nwith tf.device('/cpu:0'):\n gen_output = encoder([father_inp, mother_inp], training=True)\ntemp = gen_output.numpy()\nplt.imshow(np.squeeze(temp))\n# print(temp)\nprint(np.amin(temp))\nprint(np.amax(temp))", "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n" ], [ "target = family_data[0][3]\nplt.imshow(target)", "_____no_output_____" ] ], [ [ "###############################################################################################################################", "_____no_output_____" ] ], [ [ "def disc_downsample_parent_target(filters, size, apply_batchnorm=True):\n initializer = tf.random_normal_initializer(mean, std_dev) \n \n\n result = tf.keras.Sequential()\n result.add(\n tf.keras.layers.Conv2D(filters, size, strides=2, padding='same',\n kernel_initializer=initializer,\n use_bias=False))\n\n if apply_batchnorm:\n result.add(tf.keras.layers.BatchNormalization())\n \n result.add(tf.keras.layers.LeakyReLU(alpha = 0.2))\n\n return result", "_____no_output_____" ], [ "def disc_loss(filters, size,apply_batchnorm = False):\n initializer = tf.random_normal_initializer(mean, std_dev)\n\n result = tf.keras.Sequential()\n result.add(\n tf.keras.layers.Conv2D(filters, size, strides=2,\n padding='same',\n kernel_initializer=initializer,\n use_bias=False))\n if apply_batchnorm:\n result.add(tf.keras.layers.BatchNormalization())\n \n result.add(tf.keras.layers.LeakyReLU(alpha = 0.2))\n \n return result", "_____no_output_____" ], [ "def Discriminator():\n\n father = tf.keras.layers.Input(shape=(img_size,img_size,3))\n mother = tf.keras.layers.Input(shape=(img_size,img_size,3))\n target = tf.keras.layers.Input(shape=(img_size,img_size,3))\n \n down_stack_parent_target = [\n disc_downsample_parent_target(32,4,apply_batchnorm=False), #32x32x32\n disc_downsample_parent_target(64,4,apply_batchnorm=True) #16x16x64\n ]\n \n down_stack_combined =[\n disc_loss(192,4,apply_batchnorm=True),\n disc_loss(256,4,apply_batchnorm=False)\n ]\n \n initializer = tf.random_normal_initializer(mean, sd_random_normal_init)\n last = tf.keras.layers.Conv2D(1, 4, strides=1,padding='same',\n kernel_initializer=initializer) # linear layer\n \n \n concat = tf.keras.layers.Concatenate()\n x1 = father\n for down in down_stack_parent_target:\n x1 = down(x1)\n \n x2 = mother\n for down in down_stack_parent_target:\n x2 = down(x2)\n \n x3 = target\n for down in down_stack_parent_target:\n x3 = down(x3)\n \n combined = concat([x1,x2,x3])\n # combined is Batchx16x16x192\n \n x4 = combined\n for down in down_stack_combined:\n x4 = down(x4)\n# print(x4.shape)\n \n output = last(x4) #4X4 \n print(output.shape)\n\n return tf.keras.Model(inputs=[father,mother,target], outputs=output)", "_____no_output_____" ], [ "discriminator = Discriminator()", "(None, 4, 4, 1)\n" ], [ "# family_data = generate_image(all_families[126])\n# p1 = tf.cast(family_data[0], tf.float32)\n# p2 = tf.cast(family_data[1], tf.float32)\n# c = tf.cast(family_data[2], tf.float32)\n\n# discriminator = Discriminator()\n# with tf.device('/cpu:0'):\n# disc_out = discriminator(inputs = [p1,p2,c], training=True)", "_____no_output_____" ], [ "LAMBDA = 1", "_____no_output_____" ], [ "def tensor_to_array(tensor1):\n return tensor1.numpy()", "_____no_output_____" ], [ "def discriminator_loss(disc_real_output, disc_generated_output,b_size):\n real_loss_diff = tf.abs(tf.ones_like(disc_real_output) - disc_real_output)\n real_flatten_diff = tf.reshape(real_loss_diff, (b_size, 4*4*1))\n real_loss_batch = tf.reduce_mean(real_flatten_diff, axis=1)\n real_loss = tf.reduce_mean(real_loss_batch)\n \n gen_loss_diff = tf.abs(tf.zeros_like(disc_generated_output) - disc_generated_output)\n gen_flatten_diff = tf.reshape(gen_loss_diff, (b_size, 4*4*1))\n gen_loss_batch = tf.reduce_mean(gen_flatten_diff, axis=1)\n gen_loss = tf.reduce_mean(gen_loss_batch)\n\n total_disc_loss = real_loss + gen_loss\n return total_disc_loss", "_____no_output_____" ], [ "def generator_loss(disc_generated_output, gen_output, target,b_size):\n gen_loss_diff = tf.abs(tf.ones_like(disc_generated_output) - disc_generated_output)\n gen_flatten_diff = tf.reshape(gen_loss_diff, (b_size, 4*4*1))\n gen_loss_batch = tf.reduce_mean(gen_flatten_diff, axis=1)\n gen_loss = tf.reduce_mean(gen_loss_batch)\n \n l1_loss_diff = tf.abs(target - gen_output)\n l1_flatten_diff = tf.reshape(l1_loss_diff, (b_size, img_size*img_size*3))\n l1_loss_batch = tf.reduce_mean(l1_flatten_diff, axis=1)\n l1_loss = tf.reduce_mean(l1_loss_batch)\n \n total_gen_loss = gen_loss + LAMBDA * l1_loss \n# print(\"Reconstruction loss: {}, GAN loss: {}\".format(l1_loss, gen_loss))\n return total_gen_loss", "_____no_output_____" ], [ "generator_optimizer = tf.keras.optimizers.Adam(lr, beta_1=b1 ,beta_2 = b2)\ndiscriminator_optimizer = tf.keras.optimizers.Adam(lr, beta_1=b1, beta_2 = b2)", "_____no_output_____" ], [ "def train_step(father_batch, mother_batch, target_batch,b_size):\n with tf.GradientTape() as gen_tape, tf.GradientTape() as disc_tape:\n \n gen_outputs = encoder([father_batch, mother_batch], training=True)\n# print(\"Generated outputs\",gen_outputs.shape)\n \n disc_real_output = discriminator([father_batch, mother_batch, target_batch], training=True)\n# print(\"disc_real_output \", disc_real_output.shape)\n \n disc_generated_output = discriminator([father_batch, mother_batch, gen_outputs], training=True)\n# print(\"disc_generated_output \", disc_generated_output.shape)\n \n gen_loss = generator_loss(disc_generated_output, gen_outputs, target_batch,b_size)\n disc_loss = discriminator_loss(disc_real_output, disc_generated_output,b_size)\n \n \n print(\"GEN_LOSS\",tensor_to_array(gen_loss))\n print(\"DISC_LOSS\",tensor_to_array(disc_loss))\n\n generator_gradients = gen_tape.gradient(gen_loss,encoder.trainable_variables)\n discriminator_gradients = disc_tape.gradient(disc_loss,discriminator.trainable_variables)\n\n generator_optimizer.apply_gradients(zip(generator_gradients,encoder.trainable_variables))\n discriminator_optimizer.apply_gradients(zip(discriminator_gradients,discriminator.trainable_variables))", "_____no_output_____" ], [ "def fit(train_ds, epochs, test_ds,batch):\n for epoch in range(epochs):\n print(\"______________________________EPOCH %d_______________________________\"%(epoch))\n start = time.time()\n for i in range(len(train_ds)//batch):\n batch_data = np.asarray(generate_batch(train_ds[i*batch:(i+1)*batch]))\n batch_data = batch_data / 255 * 2 -1\n \n print(\"Generated batch\", batch_data.shape)\n\n X_father_train = tf.convert_to_tensor(batch_data[:,0],dtype =tf.float32)\n X_mother_train = tf.convert_to_tensor(batch_data[:,1],dtype =tf.float32)\n# print(\"Xtrain\",X_train.shape)\n# print(\"Batch converted to tensor\")\n\n Y_train = tf.convert_to_tensor(batch_data[:,3],dtype =tf.float32)\n train_step(X_father_train, X_mother_train, Y_train, batch)\n print(\"Trained for batch %d/%d\"%(i+1,(len(train_ds)//batch)))\n \n# family_no = 400\n# family_data = generate_image(all_families[family_no][0], all_families[family_no][1], all_families[family_no][2])\n# inp = [family_data[0],family_data[1]]\n# inp = tf.cast(inp, tf.float32)\n# father_inp = inp[0][tf.newaxis,...]\n# mother_inp = inp[1][tf.newaxis,...]\n# gen_output = encoder([father_inp, mother_inp], training=True)\n# print(tf.reduce_min(gen_output))\n# print(tf.reduce_max(gen_output))\n# plt.figure()\n# plt.imshow(gen_output[0,...])\n# plt.show()\n \n print(\"______________________________TRAINING COMPLETED_______________________________\")\n checkpoint.save(file_prefix = checkpoint_prefix)", "_____no_output_____" ], [ "concat = tf.keras.layers.Concatenate()", "_____no_output_____" ], [ "train_dataset = all_families[:-10]\ntest_dataset = all_families[-10:]\nencoder = EncoderNN()\ndiscriminator = Discriminator()", "_____no_output_____" ], [ "img_size = 64\nmean = 0.\nstd_dev = 0.02\nlr = 0.0005\nb1 = 0.9\nb2 = 0.999\nsd_random_normal_init = 0.02\n\nEPOCHS = 5\nbatch = 25", "_____no_output_____" ], [ "checkpoint_dir = './checkpoint'\ncheckpoint_prefix = os.path.join(checkpoint_dir, \"ckpt\")\ncheckpoint = tf.train.Checkpoint(generator_optimizer=generator_optimizer,\n discriminator_optimizer=discriminator_optimizer,\n generator=encoder,\n discriminator=discriminator)", "_____no_output_____" ], [ "with tf.device('/gpu:0'):\n fit(train_dataset, EPOCHS, test_dataset,batch)", "______________________________EPOCH 0_______________________________\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.8714736\nDISC_LOSS 1.86838\nTrained for batch 1/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 2.4612353\nDISC_LOSS 4.9009223\nTrained for batch 2/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.5188591\nDISC_LOSS 3.105586\nTrained for batch 3/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 0.97126555\nDISC_LOSS 1.2524204\nTrained for batch 4/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.5256444\nDISC_LOSS 1.3195959\nTrained for batch 5/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.7478199\nDISC_LOSS 1.5654196\nTrained for batch 6/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.615851\nDISC_LOSS 1.3376899\nTrained for batch 7/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.3071386\nDISC_LOSS 1.0237919\nTrained for batch 8/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.033449\nDISC_LOSS 0.9461256\nTrained for batch 9/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 0.8963132\nDISC_LOSS 0.9979107\nTrained for batch 10/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 0.8280088\nDISC_LOSS 1.0079626\nTrained for batch 11/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 0.7844893\nDISC_LOSS 0.98840153\nTrained for batch 12/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 0.8130939\nDISC_LOSS 0.9575056\nTrained for batch 13/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 0.80703056\nDISC_LOSS 0.9344727\nTrained for batch 14/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 0.9170277\nDISC_LOSS 0.909178\nTrained for batch 15/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 0.9479166\nDISC_LOSS 0.89190656\nTrained for batch 16/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 0.96813387\nDISC_LOSS 0.91292846\nTrained for batch 17/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 0.9825418\nDISC_LOSS 0.9106399\nTrained for batch 18/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.0002934\nDISC_LOSS 0.89057213\nTrained for batch 19/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 0.9695563\nDISC_LOSS 0.9175471\nTrained for batch 20/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 0.9578241\nDISC_LOSS 0.9103563\nTrained for batch 21/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 0.9494082\nDISC_LOSS 0.93349683\nTrained for batch 22/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 0.96414757\nDISC_LOSS 0.96248746\nTrained for batch 23/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 0.9864141\nDISC_LOSS 0.9575828\nTrained for batch 24/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.0093925\nDISC_LOSS 0.9762958\nTrained for batch 25/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.0861416\nDISC_LOSS 0.92497194\nTrained for batch 26/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.0958909\nDISC_LOSS 0.9291233\nTrained for batch 27/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.0585023\nDISC_LOSS 0.9400269\nTrained for batch 28/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.1250031\nDISC_LOSS 0.9135808\nTrained for batch 29/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.0957966\nDISC_LOSS 0.941985\nTrained for batch 30/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.0928779\nDISC_LOSS 0.9060876\nTrained for batch 31/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.0167775\nDISC_LOSS 0.97104776\nTrained for batch 32/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 0.9991605\nDISC_LOSS 0.9424188\nTrained for batch 33/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 0.97426724\nDISC_LOSS 0.9429213\nTrained for batch 34/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 0.93371403\nDISC_LOSS 0.9179907\nTrained for batch 35/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 0.9309088\nDISC_LOSS 0.8584677\nTrained for batch 36/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 0.9956986\nDISC_LOSS 0.86251956\nTrained for batch 37/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.0101424\nDISC_LOSS 0.82430947\nTrained for batch 38/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.0703489\nDISC_LOSS 0.8889017\nTrained for batch 39/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.1007025\nDISC_LOSS 0.86110675\nTrained for batch 40/40\n______________________________EPOCH 1_______________________________\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.0465988\nDISC_LOSS 0.8305098\nTrained for batch 1/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 0.98515004\nDISC_LOSS 0.8789594\nTrained for batch 2/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.0371296\nDISC_LOSS 0.82892394\nTrained for batch 3/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.1436797\nDISC_LOSS 0.76781416\nTrained for batch 4/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 0.98828745\nDISC_LOSS 0.8327844\nTrained for batch 5/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 0.96988016\nDISC_LOSS 0.8151218\nTrained for batch 6/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.0995154\nDISC_LOSS 0.7874242\nTrained for batch 7/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.1080232\nDISC_LOSS 0.8091966\nTrained for batch 8/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.1456304\nDISC_LOSS 0.7980537\nTrained for batch 9/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.0870317\nDISC_LOSS 0.7926892\nTrained for batch 10/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.1679136\nDISC_LOSS 0.73891556\nTrained for batch 11/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.0935614\nDISC_LOSS 0.7004093\nTrained for batch 12/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.1414446\nDISC_LOSS 0.6916524\nTrained for batch 13/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.1969086\nDISC_LOSS 0.6280055\nTrained for batch 14/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.1256477\nDISC_LOSS 0.6662531\nTrained for batch 15/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.2153999\nDISC_LOSS 0.67171395\nTrained for batch 16/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.2723532\nDISC_LOSS 0.6211302\nTrained for batch 17/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.3242301\nDISC_LOSS 0.5451824\nTrained for batch 18/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.3218682\nDISC_LOSS 0.5655677\nTrained for batch 19/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.3787348\nDISC_LOSS 0.50307274\nTrained for batch 20/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.3548858\nDISC_LOSS 0.5147311\nTrained for batch 21/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.2932595\nDISC_LOSS 0.4773554\nTrained for batch 22/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.3981586\nDISC_LOSS 0.45202476\nTrained for batch 23/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.3732232\nDISC_LOSS 0.4811604\nTrained for batch 24/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.42959\nDISC_LOSS 0.4987492\nTrained for batch 25/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.3369378\nDISC_LOSS 0.48916674\nTrained for batch 26/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.3274763\nDISC_LOSS 0.49758807\nTrained for batch 27/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.3826015\nDISC_LOSS 0.49969247\nTrained for batch 28/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.232553\nDISC_LOSS 0.5970494\nTrained for batch 29/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.1694779\nDISC_LOSS 0.61546373\nTrained for batch 30/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.174477\nDISC_LOSS 0.6335112\nTrained for batch 31/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.2373971\nDISC_LOSS 0.6561134\nTrained for batch 32/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.2919565\nDISC_LOSS 0.58528715\nTrained for batch 33/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.2439705\nDISC_LOSS 0.5333974\nTrained for batch 34/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.1827548\nDISC_LOSS 0.5410408\nTrained for batch 35/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.4111747\nDISC_LOSS 0.48652083\nTrained for batch 36/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.1749924\nDISC_LOSS 0.59631115\nTrained for batch 37/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.3081505\nDISC_LOSS 0.59527826\nTrained for batch 38/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.4181799\nDISC_LOSS 0.5750201\nTrained for batch 39/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.279568\nDISC_LOSS 0.5244153\nTrained for batch 40/40\n______________________________EPOCH 2_______________________________\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.3625059\nDISC_LOSS 0.49883476\nTrained for batch 1/40\nGenerated batch (25, 4, 64, 64, 3)\nGEN_LOSS 1.4587033\nDISC_LOSS 0.4600417\nTrained for batch 2/40\nGenerated batch (25, 4, 64, 64, 3)\n" ], [ "family_no = 1011\nfamily_data = generate_image(all_families[family_no][0], all_families[family_no][1], all_families[family_no][2])\ninp = [family_data[0],family_data[1]]\ninp = tf.cast(inp, tf.float32)\nfather_inp = inp[0][tf.newaxis,...]\nmother_inp = inp[1][tf.newaxis,...]\nwith tf.device('/gpu:0'):\n gen_output = encoder([father_inp, mother_inp], training=True)\ntemp = gen_output.numpy()\nplt.imshow(np.squeeze(temp))\nprint(np.amin(temp))\n+print(np.amax(temp))", "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n" ], [ "family_no = 1011\nfamily_data = generate_image(all_families[family_no][0], all_families[family_no][1], all_families[family_no][2])\ninp = [family_data[0],family_data[1]]\ninp = tf.cast(inp, tf.float32)\nfather_inp = inp[0][tf.newaxis,...]\nmother_inp = inp[1][tf.newaxis,...]\nwith tf.device('/gpu:0'):\n gen_output = encoder([father_inp, mother_inp], training=True)\ntemp = gen_output.numpy()\nplt.imshow(np.squeeze(temp))\nprint(np.amin(temp))\nprint(np.amax(temp))", "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n" ] ] ]

[ "code", "raw", "code", "markdown", "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code", "code" ], [ "raw" ], [ "code", "code", "code", "code", "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "empty" ] ] ]

[ "empty" ]

[ [ "empty" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "# AWS Marketplace Product Usage Demonstration - Algorithms\n\n## Using Algorithm ARN with Amazon SageMaker APIs\n\nThis sample notebook demonstrates two new functionalities added to Amazon SageMaker:\n1. Using an Algorithm ARN to run training jobs and use that result for inference\n2. Using an AWS Marketplace product ARN - we will use [Scikit Decision Trees](https://aws.amazon.com/marketplace/pp/prodview-ha4f3kqugba3u?qid=1543169069960&sr=0-1&ref_=srh_res_product_title)\n\n## Overall flow diagram\n<img src=\"images/AlgorithmE2EFlow.jpg\">\n\n## Compatibility\nThis notebook is compatible only with [Scikit Decision Trees](https://aws.amazon.com/marketplace/pp/prodview-ha4f3kqugba3u?qid=1543169069960&sr=0-1&ref_=srh_res_product_title) sample algorithm published to AWS Marketplace. \n\n***Pre-Requisite:*** Please subscribe to this free product before proceeding with this notebook", "_____no_output_____" ], [ "## Set up the environment", "_____no_output_____" ] ], [ [ "import sagemaker as sage\nfrom sagemaker import get_execution_role\n\nrole = get_execution_role()\n\n# S3 prefixes\ncommon_prefix = \"DEMO-scikit-byo-iris\"\ntraining_input_prefix = common_prefix + \"/training-input-data\"\nbatch_inference_input_prefix = common_prefix + \"/batch-inference-input-data\"", "_____no_output_____" ] ], [ [ "### Create the session\n\nThe session remembers our connection parameters to Amazon SageMaker. We'll use it to perform all of our Amazon SageMaker operations.", "_____no_output_____" ] ], [ [ "sagemaker_session = sage.Session()", "_____no_output_____" ] ], [ [ "## Upload the data for training\n\nWhen training large models with huge amounts of data, you'll typically use big data tools, like Amazon Athena, AWS Glue, or Amazon EMR, to create your data in S3. For the purposes of this example, we're using some the classic [Iris dataset](https://en.wikipedia.org/wiki/Iris_flower_data_set), which we have included. \n\nWe can use use the tools provided by the Amazon SageMaker Python SDK to upload the data to a default bucket. ", "_____no_output_____" ] ], [ [ "TRAINING_WORKDIR = \"data/training\"\n\ntraining_input = sagemaker_session.upload_data(TRAINING_WORKDIR, key_prefix=training_input_prefix)\nprint(\"Training Data Location \" + training_input)", "_____no_output_____" ] ], [ [ "## Creating Training Job using Algorithm ARN\n\nPlease put in the algorithm arn you want to use below. This can either be an AWS Marketplace algorithm you subscribed to (or) one of the algorithms you created in your own account.\n\nThe algorithm arn listed below belongs to the [Scikit Decision Trees](https://aws.amazon.com/marketplace/pp/prodview-ha4f3kqugba3u?qid=1543169069960&sr=0-1&ref_=srh_res_product_title) product.", "_____no_output_____" ] ], [ [ "from src.scikit_product_arns import ScikitArnProvider\n\nalgorithm_arn = ScikitArnProvider.get_algorithm_arn(sagemaker_session.boto_region_name)", "_____no_output_____" ], [ "import json\nimport time\nfrom sagemaker.algorithm import AlgorithmEstimator\n\nalgo = AlgorithmEstimator(\n algorithm_arn=algorithm_arn,\n role=role,\n train_instance_count=1,\n train_instance_type=\"ml.c4.xlarge\",\n base_job_name=\"scikit-from-aws-marketplace\",\n)", "_____no_output_____" ] ], [ [ "## Run Training Job", "_____no_output_____" ] ], [ [ "print(\n \"Now run the training job using algorithm arn %s in region %s\"\n % (algorithm_arn, sagemaker_session.boto_region_name)\n)\nalgo.fit({\"training\": training_input})", "_____no_output_____" ] ], [ [ "## Automated Model Tuning (optional)\n\nSince this algorithm supports tunable hyperparameters with a tuning objective metric, we can run a Hyperparameter Tuning Job to obtain the best training job hyperparameters and its corresponding model artifacts. \n\n<img src=\"images/HPOFlow.jpg\">\n", "_____no_output_____" ] ], [ [ "from sagemaker.tuner import HyperparameterTuner, IntegerParameter\n\n## This demo algorithm supports max_leaf_nodes as the only tunable hyperparameter.\nhyperparameter_ranges = {\"max_leaf_nodes\": IntegerParameter(1, 100000)}\n\ntuner = HyperparameterTuner(\n estimator=algo,\n base_tuning_job_name=\"some-name\",\n objective_metric_name=\"validation:accuracy\",\n hyperparameter_ranges=hyperparameter_ranges,\n max_jobs=2,\n max_parallel_jobs=2,\n)\n\ntuner.fit({\"training\": training_input}, include_cls_metadata=False)\ntuner.wait()", "_____no_output_____" ] ], [ [ "## Batch Transform Job\n\nNow let's use the model built to run a batch inference job and verify it works.\n\n### Batch Transform Input Preparation\n\nThe snippet below is removing the \"label\" column (column indexed at 0) and retaining the rest to be batch transform's input. \n\n***NOTE:*** This is the same training data, which is a no-no from a ML science perspective. But the aim of this notebook is to demonstrate how things work end-to-end.", "_____no_output_____" ] ], [ [ "import pandas as pd\n\n## Remove first column that contains the label\nshape = pd.read_csv(TRAINING_WORKDIR + \"/iris.csv\", header=None).drop([0], axis=1)\n\nTRANSFORM_WORKDIR = \"data/transform\"\nshape.to_csv(TRANSFORM_WORKDIR + \"/batchtransform_test.csv\", index=False, header=False)\n\ntransform_input = (\n sagemaker_session.upload_data(TRANSFORM_WORKDIR, key_prefix=batch_inference_input_prefix)\n + \"/batchtransform_test.csv\"\n)\nprint(\"Transform input uploaded to \" + transform_input)", "_____no_output_____" ], [ "transformer = algo.transformer(1, \"ml.m4.xlarge\")\ntransformer.transform(transform_input, content_type=\"text/csv\")\ntransformer.wait()\n\nprint(\"Batch Transform output saved to \" + transformer.output_path)", "_____no_output_____" ] ], [ [ "#### Inspect the Batch Transform Output in S3", "_____no_output_____" ] ], [ [ "from urllib.parse import urlparse\n\nparsed_url = urlparse(transformer.output_path)\nbucket_name = parsed_url.netloc\nfile_key = \"{}/{}.out\".format(parsed_url.path[1:], \"batchtransform_test.csv\")\n\ns3_client = sagemaker_session.boto_session.client(\"s3\")\n\nresponse = s3_client.get_object(Bucket=sagemaker_session.default_bucket(), Key=file_key)\nresponse_bytes = response[\"Body\"].read().decode(\"utf-8\")\nprint(response_bytes)", "_____no_output_____" ] ], [ [ "## Live Inference Endpoint\n\nFinally, we demonstrate the creation of an endpoint for live inference using this AWS Marketplace algorithm generated model\n", "_____no_output_____" ] ], [ [ "from sagemaker.predictor import csv_serializer\n\npredictor = algo.deploy(1, \"ml.m4.xlarge\", serializer=csv_serializer)", "_____no_output_____" ] ], [ [ "### Choose some data and use it for a prediction\n\nIn order to do some predictions, we'll extract some of the data we used for training and do predictions against it. This is, of course, bad statistical practice, but a good way to see how the mechanism works.\n", "_____no_output_____" ] ], [ [ "shape = pd.read_csv(TRAINING_WORKDIR + \"/iris.csv\", header=None)\n\nimport itertools\n\na = [50 * i for i in range(3)]\nb = [40 + i for i in range(10)]\nindices = [i + j for i, j in itertools.product(a, b)]\n\ntest_data = shape.iloc[indices[:-1]]\ntest_X = test_data.iloc[:, 1:]\ntest_y = test_data.iloc[:, 0]", "_____no_output_____" ] ], [ [ "Prediction is as easy as calling predict with the predictor we got back from deploy and the data we want to do predictions with. The serializers take care of doing the data conversions for us.", "_____no_output_____" ] ], [ [ "print(predictor.predict(test_X.values).decode(\"utf-8\"))", "_____no_output_____" ] ], [ [ "### Cleanup the endpoint", "_____no_output_____" ] ], [ [ "algo.delete_endpoint()", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ] ]

[ "BSD-3-Clause" ]

[ "BSD-3-Clause" ]

[ "BSD-3-Clause" ]

[ [ [ "# Detrending, Stylized Facts and the Business Cycle\n\nIn an influential article, Harvey and Jaeger (1993) described the use of unobserved components models (also known as \"structural time series models\") to derive stylized facts of the business cycle.\n\nTheir paper begins:\n\n \"Establishing the 'stylized facts' associated with a set of time series is widely considered a crucial step\n in macroeconomic research ... For such facts to be useful they should (1) be consistent with the stochastic\n properties of the data and (2) present meaningful information.\"\n \nIn particular, they make the argument that these goals are often better met using the unobserved components approach rather than the popular Hodrick-Prescott filter or Box-Jenkins ARIMA modeling techniques.\n\nStatsmodels has the ability to perform all three types of analysis, and below we follow the steps of their paper, using a slightly updated dataset.", "_____no_output_____" ] ], [ [ "%matplotlib inline\n\nimport numpy as np\nimport pandas as pd\nimport statsmodels.api as sm\nimport matplotlib.pyplot as plt\n\nfrom IPython.display import display, Latex", "_____no_output_____" ] ], [ [ "## Unobserved Components\n\nThe unobserved components model available in Statsmodels can be written as:\n\n$$\ny_t = \\underbrace{\\mu_{t}}_{\\text{trend}} + \\underbrace{\\gamma_{t}}_{\\text{seasonal}} + \\underbrace{c_{t}}_{\\text{cycle}} + \\sum_{j=1}^k \\underbrace{\\beta_j x_{jt}}_{\\text{explanatory}} + \\underbrace{\\varepsilon_t}_{\\text{irregular}}\n$$\n\nsee Durbin and Koopman 2012, Chapter 3 for notation and additional details. Notice that different specifications for the different individual components can support a wide range of models. The specific models considered in the paper and below are specializations of this general equation.\n\n### Trend\n\nThe trend component is a dynamic extension of a regression model that includes an intercept and linear time-trend.\n\n$$\n\\begin{align}\n\\underbrace{\\mu_{t+1}}_{\\text{level}} & = \\mu_t + \\nu_t + \\eta_{t+1} \\qquad & \\eta_{t+1} \\sim N(0, \\sigma_\\eta^2) \\\\\\\\\n\\underbrace{\\nu_{t+1}}_{\\text{trend}} & = \\nu_t + \\zeta_{t+1} & \\zeta_{t+1} \\sim N(0, \\sigma_\\zeta^2) \\\\\n\\end{align}\n$$\n\nwhere the level is a generalization of the intercept term that can dynamically vary across time, and the trend is a generalization of the time-trend such that the slope can dynamically vary across time.\n\nFor both elements (level and trend), we can consider models in which:\n\n- The element is included vs excluded (if the trend is included, there must also be a level included).\n- The element is deterministic vs stochastic (i.e. whether or not the variance on the error term is confined to be zero or not)\n\nThe only additional parameters to be estimated via MLE are the variances of any included stochastic components.\n\nThis leads to the following specifications:\n\n| | Level | Trend | Stochastic Level | Stochastic Trend |\n|----------------------------------------------------------------------|-------|-------|------------------|------------------|\n| Constant | ✓ | | | |\n| Local Level <br /> (random walk) | ✓ | | ✓ | |\n| Deterministic trend | ✓ | ✓ | | |\n| Local level with deterministic trend <br /> (random walk with drift) | ✓ | ✓ | ✓ | |\n| Local linear trend | ✓ | ✓ | ✓ | ✓ |\n| Smooth trend <br /> (integrated random walk) | ✓ | ✓ | | ✓ |\n\n### Seasonal\n\nThe seasonal component is written as:\n\n<span>$$\n\\gamma_t = - \\sum_{j=1}^{s-1} \\gamma_{t+1-j} + \\omega_t \\qquad \\omega_t \\sim N(0, \\sigma_\\omega^2)\n$$</span>\n\nThe periodicity (number of seasons) is `s`, and the defining character is that (without the error term), the seasonal components sum to zero across one complete cycle. The inclusion of an error term allows the seasonal effects to vary over time.\n\nThe variants of this model are:\n\n- The periodicity `s`\n- Whether or not to make the seasonal effects stochastic.\n\nIf the seasonal effect is stochastic, then there is one additional parameter to estimate via MLE (the variance of the error term).\n\n### Cycle\n\nThe cyclical component is intended to capture cyclical effects at time frames much longer than captured by the seasonal component. For example, in economics the cyclical term is often intended to capture the business cycle, and is then expected to have a period between \"1.5 and 12 years\" (see Durbin and Koopman).\n\nThe cycle is written as:\n\n<span>$$\n\\begin{align}\nc_{t+1} & = c_t \\cos \\lambda_c + c_t^* \\sin \\lambda_c + \\tilde \\omega_t \\qquad & \\tilde \\omega_t \\sim N(0, \\sigma_{\\tilde \\omega}^2) \\\\\\\\\nc_{t+1}^* & = -c_t \\sin \\lambda_c + c_t^* \\cos \\lambda_c + \\tilde \\omega_t^* & \\tilde \\omega_t^* \\sim N(0, \\sigma_{\\tilde \\omega}^2)\n\\end{align}\n$$</span>\n\nThe parameter $\\lambda_c$ (the frequency of the cycle) is an additional parameter to be estimated by MLE. If the seasonal effect is stochastic, then there is one another parameter to estimate (the variance of the error term - note that both of the error terms here share the same variance, but are assumed to have independent draws).\n\n### Irregular\n\nThe irregular component is assumed to be a white noise error term. Its variance is a parameter to be estimated by MLE; i.e.\n\n$$\n\\varepsilon_t \\sim N(0, \\sigma_\\varepsilon^2)\n$$\n\nIn some cases, we may want to generalize the irregular component to allow for autoregressive effects:\n\n$$\n\\varepsilon_t = \\rho(L) \\varepsilon_{t-1} + \\epsilon_t, \\qquad \\epsilon_t \\sim N(0, \\sigma_\\epsilon^2)\n$$\n\nIn this case, the autoregressive parameters would also be estimated via MLE.\n\n### Regression effects\n\nWe may want to allow for explanatory variables by including additional terms\n\n<span>$$\n\\sum_{j=1}^k \\beta_j x_{jt}\n$$</span>\n\nor for intervention effects by including\n\n<span>$$\n\\begin{align}\n\\delta w_t \\qquad \\text{where} \\qquad w_t & = 0, \\qquad t < \\tau, \\\\\\\\\n& = 1, \\qquad t \\ge \\tau\n\\end{align}\n$$</span>\n\nThese additional parameters could be estimated via MLE or by including them as components of the state space formulation.\n", "_____no_output_____" ], [ "## Data\n\nFollowing Harvey and Jaeger, we will consider the following time series:\n\n- US real GNP, \"output\", ([GNPC96](https://research.stlouisfed.org/fred2/series/GNPC96))\n- US GNP implicit price deflator, \"prices\", ([GNPDEF](https://research.stlouisfed.org/fred2/series/GNPDEF))\n- US monetary base, \"money\", ([AMBSL](https://research.stlouisfed.org/fred2/series/AMBSL))\n\nThe time frame in the original paper varied across series, but was broadly 1954-1989. Below we use data from the period 1948-2008 for all series. Although the unobserved components approach allows isolating a seasonal component within the model, the series considered in the paper, and here, are already seasonally adjusted.\n\nAll data series considered here are taken from [Federal Reserve Economic Data (FRED)](https://research.stlouisfed.org/fred2/). Conveniently, the Python library [Pandas](http://pandas.pydata.org/) has the ability to download data from FRED directly.", "_____no_output_____" ] ], [ [ "# Datasets\nfrom pandas.io.data import DataReader\n\n# Get the raw data\nstart = '1948-01'\nend = '2008-01'\nus_gnp = DataReader('GNPC96', 'fred', start=start, end=end)\nus_gnp_deflator = DataReader('GNPDEF', 'fred', start=start, end=end)\nus_monetary_base = DataReader('AMBSL', 'fred', start=start, end=end).resample('QS')\nrecessions = DataReader('USRECQ', 'fred', start=start, end=end).resample('QS', how='last').values[:,0]\n\n# Construct the dataframe\ndta = pd.concat(map(np.log, (us_gnp, us_gnp_deflator, us_monetary_base)), axis=1)\ndta.columns = ['US GNP','US Prices','US monetary base']\ndates = dta.index._mpl_repr()", "_____no_output_____" ] ], [ [ "To get a sense of these three variables over the timeframe, we can plot them:", "_____no_output_____" ] ], [ [ "# Plot the data\nax = dta.plot(figsize=(13,3))\nylim = ax.get_ylim()\nax.xaxis.grid()\nax.fill_between(dates, ylim[0]+1e-5, ylim[1]-1e-5, recessions, facecolor='k', alpha=0.1);", "_____no_output_____" ] ], [ [ "## Model\n\nSince the data is already seasonally adjusted and there are no obvious explanatory variables, the generic model considered is:\n\n$$\ny_t = \\underbrace{\\mu_{t}}_{\\text{trend}} + \\underbrace{c_{t}}_{\\text{cycle}} + \\underbrace{\\varepsilon_t}_{\\text{irregular}}\n$$\n\nThe irregular will be assumed to be white noise, and the cycle will be stochastic and damped. The final modeling choice is the specification to use for the trend component. Harvey and Jaeger consider two models:\n\n1. Local linear trend (the \"unrestricted\" model)\n2. Smooth trend (the \"restricted\" model, since we are forcing $\\sigma_\\eta = 0$)\n\nBelow, we construct `kwargs` dictionaries for each of these model types. Notice that rather that there are two ways to specify the models. One way is to specify components directly, as in the table above. The other way is to use string names which map to various specifications.", "_____no_output_____" ] ], [ [ "# Model specifications\n\n# Unrestricted model, using string specification\nunrestricted_model = {\n 'level': 'local linear trend', 'cycle': True, 'damped_cycle': True, 'stochastic_cycle': True\n}\n\n# Unrestricted model, setting components directly\n# This is an equivalent, but less convenient, way to specify a\n# local linear trend model with a stochastic damped cycle:\n# unrestricted_model = {\n# 'irregular': True, 'level': True, 'stochastic_level': True, 'trend': True, 'stochastic_trend': True,\n# 'cycle': True, 'damped_cycle': True, 'stochastic_cycle': True\n# }\n\n# The restricted model forces a smooth trend\nrestricted_model = {\n 'level': 'smooth trend', 'cycle': True, 'damped_cycle': True, 'stochastic_cycle': True\n}\n\n# Restricted model, setting components directly\n# This is an equivalent, but less convenient, way to specify a\n# smooth trend model with a stochastic damped cycle. Notice\n# that the difference from the local linear trend model is that\n# `stochastic_level=False` here.\n# unrestricted_model = {\n# 'irregular': True, 'level': True, 'stochastic_level': False, 'trend': True, 'stochastic_trend': True,\n# 'cycle': True, 'damped_cycle': True, 'stochastic_cycle': True\n# }", "_____no_output_____" ] ], [ [ "We now fit the following models:\n\n1. Output, unrestricted model\n2. Prices, unrestricted model\n3. Prices, restricted model\n4. Money, unrestricted model\n5. Money, restricted model", "_____no_output_____" ] ], [ [ "# Output\noutput_mod = sm.tsa.UnobservedComponents(dta['US GNP'], **unrestricted_model)\noutput_res = output_mod.fit(method='powell', disp=False)\n\n# Prices\nprices_mod = sm.tsa.UnobservedComponents(dta['US Prices'], **unrestricted_model)\nprices_res = prices_mod.fit(method='powell', disp=False)\n\nprices_restricted_mod = sm.tsa.UnobservedComponents(dta['US Prices'], **restricted_model)\nprices_restricted_res = prices_restricted_mod.fit(method='powell', disp=False)\n\n# Money\nmoney_mod = sm.tsa.UnobservedComponents(dta['US monetary base'], **unrestricted_model)\nmoney_res = money_mod.fit(method='powell', disp=False)\n\nmoney_restricted_mod = sm.tsa.UnobservedComponents(dta['US monetary base'], **restricted_model)\nmoney_restricted_res = money_restricted_mod.fit(method='powell', disp=False)", "_____no_output_____" ] ], [ [ "Once we have fit these models, there are a variety of ways to display the information. Looking at the model of US GNP, we can summarize the fit of the model using the `summary` method on the fit object.", "_____no_output_____" ] ], [ [ "print(output_res.summary())", "_____no_output_____" ] ], [ [ "For unobserved components models, and in particular when exploring stylized facts in line with point (2) from the introduction, it is often more instructive to plot the estimated unobserved components (e.g. the level, trend, and cycle) themselves to see if they provide a meaningful description of the data.\n\nThe `plot_components` method of the fit object can be used to show plots and confidence intervals of each of the estimated states, as well as a plot of the observed data versus the one-step-ahead predictions of the model to assess fit.", "_____no_output_____" ] ], [ [ "fig = output_res.plot_components(legend_loc='lower right', figsize=(15, 9));", "_____no_output_____" ] ], [ [ "Finally, Harvey and Jaeger summarize the models in another way to highlight the relative importances of the trend and cyclical components; below we replicate their Table I. The values we find are broadly consistent with, but different in the particulars from, the values from their table.", "_____no_output_____" ] ], [ [ "# Create Table I\ntable_i = np.zeros((5,6))\n\nstart = dta.index[0]\nend = dta.index[-1]\ntime_range = '%d:%d-%d:%d' % (start.year, start.quarter, end.year, end.quarter)\nmodels = [\n ('US GNP', time_range, 'None'),\n ('US Prices', time_range, 'None'),\n ('US Prices', time_range, r'$\\sigma_\\eta^2 = 0$'),\n ('US monetary base', time_range, 'None'),\n ('US monetary base', time_range, r'$\\sigma_\\eta^2 = 0$'),\n]\nindex = pd.MultiIndex.from_tuples(models, names=['Series', 'Time range', 'Restrictions'])\nparameter_symbols = [\n r'$\\sigma_\\zeta^2$', r'$\\sigma_\\eta^2$', r'$\\sigma_\\kappa^2$', r'$\\rho$',\n r'$2 \\pi / \\lambda_c$', r'$\\sigma_\\varepsilon^2$',\n]\n\ni = 0\nfor res in (output_res, prices_res, prices_restricted_res, money_res, money_restricted_res):\n if res.model.stochastic_level:\n (sigma_irregular, sigma_level, sigma_trend,\n sigma_cycle, frequency_cycle, damping_cycle) = res.params\n else:\n (sigma_irregular, sigma_level,\n sigma_cycle, frequency_cycle, damping_cycle) = res.params\n sigma_trend = np.nan\n period_cycle = 2 * np.pi / frequency_cycle\n \n table_i[i, :] = [\n sigma_level*1e7, sigma_trend*1e7,\n sigma_cycle*1e7, damping_cycle, period_cycle,\n sigma_irregular*1e7\n ]\n i += 1\n \npd.set_option('float_format', lambda x: '%.4g' % np.round(x, 2) if not np.isnan(x) else '-')\ntable_i = pd.DataFrame(table_i, index=index, columns=parameter_symbols)\ntable_i", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Logistic Regression with a Neural Network mindset\n\nWelcome to your first (required) programming assignment! You will build a logistic regression classifier to recognize cats. This assignment will step you through how to do this with a Neural Network mindset, and so will also hone your intuitions about deep learning.\n\n**Instructions:**\n- Do not use loops (for/while) in your code, unless the instructions explicitly ask you to do so.\n\n**You will learn to:**\n- Build the general architecture of a learning algorithm, including:\n - Initializing parameters\n - Calculating the cost function and its gradient\n - Using an optimization algorithm (gradient descent) \n- Gather all three functions above into a main model function, in the right order.", "_____no_output_____" ], [ "## 1 - Packages ##\n\nFirst, let's run the cell below to import all the packages that you will need during this assignment. \n- [numpy](www.numpy.org) is the fundamental package for scientific computing with Python.\n- [h5py](http://www.h5py.org) is a common package to interact with a dataset that is stored on an H5 file.\n- [matplotlib](http://matplotlib.org) is a famous library to plot graphs in Python.\n- [PIL](http://www.pythonware.com/products/pil/) and [scipy](https://www.scipy.org/) are used here to test your model with your own picture at the end.", "_____no_output_____" ] ], [ [ "import numpy as np\nimport matplotlib.pyplot as plt\nimport h5py\nimport scipy\nfrom PIL import Image\nfrom scipy import ndimage\nfrom lr_utils import load_dataset\n\n%matplotlib inline", "_____no_output_____" ] ], [ [ "## 2 - Overview of the Problem set ##\n\n**Problem Statement**: You are given a dataset (\"data.h5\") containing:\n - a training set of m_train images labeled as cat (y=1) or non-cat (y=0)\n - a test set of m_test images labeled as cat or non-cat\n - each image is of shape (num_px, num_px, 3) where 3 is for the 3 channels (RGB). Thus, each image is square (height = num_px) and (width = num_px).\n\nYou will build a simple image-recognition algorithm that can correctly classify pictures as cat or non-cat.\n\nLet's get more familiar with the dataset. Load the data by running the following code.", "_____no_output_____" ] ], [ [ "# Loading the data (cat/non-cat)\ntrain_set_x_orig, train_set_y, test_set_x_orig, test_set_y, classes = load_dataset()", "_____no_output_____" ] ], [ [ "We added \"_orig\" at the end of image datasets (train and test) because we are going to preprocess them. After preprocessing, we will end up with train_set_x and test_set_x (the labels train_set_y and test_set_y don't need any preprocessing).\n\nEach line of your train_set_x_orig and test_set_x_orig is an array representing an image. You can visualize an example by running the following code. Feel free also to change the `index` value and re-run to see other images. ", "_____no_output_____" ] ], [ [ "# Example of a picture\nindex = 25\nplt.imshow(train_set_x_orig[index])\nprint (\"y = \" + str(train_set_y[:, index]) + \", it's a '\" + classes[np.squeeze(train_set_y[:, index])].decode(\"utf-8\") + \"' picture.\")", "_____no_output_____" ] ], [ [ "Many software bugs in deep learning come from having matrix/vector dimensions that don't fit. If you can keep your matrix/vector dimensions straight you will go a long way toward eliminating many bugs. \n\n**Exercise:** Find the values for:\n - m_train (number of training examples)\n - m_test (number of test examples)\n - num_px (= height = width of a training image)\nRemember that `train_set_x_orig` is a numpy-array of shape (m_train, num_px, num_px, 3). For instance, you can access `m_train` by writing `train_set_x_orig.shape[0]`.", "_____no_output_____" ] ], [ [ "### START CODE HERE ### (≈ 3 lines of code)\nm_train = None\nm_test = None\nnum_px = None\n### END CODE HERE ###\n\nprint (\"Number of training examples: m_train = \" + str(m_train))\nprint (\"Number of testing examples: m_test = \" + str(m_test))\nprint (\"Height/Width of each image: num_px = \" + str(num_px))\nprint (\"Each image is of size: (\" + str(num_px) + \", \" + str(num_px) + \", 3)\")\nprint (\"train_set_x shape: \" + str(train_set_x_orig.shape))\nprint (\"train_set_y shape: \" + str(train_set_y.shape))\nprint (\"test_set_x shape: \" + str(test_set_x_orig.shape))\nprint (\"test_set_y shape: \" + str(test_set_y.shape))", "_____no_output_____" ] ], [ [ "**Expected Output for m_train, m_test and num_px**: \n<table style=\"width:15%\">\n <tr>\n <td>**m_train**</td>\n <td> 209 </td> \n </tr>\n \n <tr>\n <td>**m_test**</td>\n <td> 50 </td> \n </tr>\n \n <tr>\n <td>**num_px**</td>\n <td> 64 </td> \n </tr>\n \n</table>\n", "_____no_output_____" ], [ "For convenience, you should now reshape images of shape (num_px, num_px, 3) in a numpy-array of shape (num_px $*$ num_px $*$ 3, 1). After this, our training (and test) dataset is a numpy-array where each column represents a flattened image. There should be m_train (respectively m_test) columns.\n\n**Exercise:** Reshape the training and test data sets so that images of size (num_px, num_px, 3) are flattened into single vectors of shape (num\\_px $*$ num\\_px $*$ 3, 1).\n\nA trick when you want to flatten a matrix X of shape (a,b,c,d) to a matrix X_flatten of shape (b$*$c$*$d, a) is to use: \n```python\nX_flatten = X.reshape(X.shape[0], -1).T # X.T is the transpose of X\n```", "_____no_output_____" ] ], [ [ "# Reshape the training and test examples\n\n### START CODE HERE ### (≈ 2 lines of code)\ntrain_set_x_flatten = None\ntest_set_x_flatten = None\n### END CODE HERE ###\n\nprint (\"train_set_x_flatten shape: \" + str(train_set_x_flatten.shape))\nprint (\"train_set_y shape: \" + str(train_set_y.shape))\nprint (\"test_set_x_flatten shape: \" + str(test_set_x_flatten.shape))\nprint (\"test_set_y shape: \" + str(test_set_y.shape))\nprint (\"sanity check after reshaping: \" + str(train_set_x_flatten[0:5,0]))", "_____no_output_____" ] ], [ [ "**Expected Output**: \n\n<table style=\"width:35%\">\n <tr>\n <td>**train_set_x_flatten shape**</td>\n <td> (12288, 209)</td> \n </tr>\n <tr>\n <td>**train_set_y shape**</td>\n <td>(1, 209)</td> \n </tr>\n <tr>\n <td>**test_set_x_flatten shape**</td>\n <td>(12288, 50)</td> \n </tr>\n <tr>\n <td>**test_set_y shape**</td>\n <td>(1, 50)</td> \n </tr>\n <tr>\n <td>**sanity check after reshaping**</td>\n <td>[17 31 56 22 33]</td> \n </tr>\n</table>", "_____no_output_____" ], [ "To represent color images, the red, green and blue channels (RGB) must be specified for each pixel, and so the pixel value is actually a vector of three numbers ranging from 0 to 255.\n\nOne common preprocessing step in machine learning is to center and standardize your dataset, meaning that you substract the mean of the whole numpy array from each example, and then divide each example by the standard deviation of the whole numpy array. But for picture datasets, it is simpler and more convenient and works almost as well to just divide every row of the dataset by 255 (the maximum value of a pixel channel).\n\n<!-- During the training of your model, you're going to multiply weights and add biases to some initial inputs in order to observe neuron activations. Then you backpropogate with the gradients to train the model. But, it is extremely important for each feature to have a similar range such that our gradients don't explode. You will see that more in detail later in the lectures. !--> \n\nLet's standardize our dataset.", "_____no_output_____" ] ], [ [ "train_set_x = train_set_x_flatten/255.\ntest_set_x = test_set_x_flatten/255.", "_____no_output_____" ] ], [ [ "<font color='blue'>\n**What you need to remember:**\n\nCommon steps for pre-processing a new dataset are:\n- Figure out the dimensions and shapes of the problem (m_train, m_test, num_px, ...)\n- Reshape the datasets such that each example is now a vector of size (num_px \\* num_px \\* 3, 1)\n- \"Standardize\" the data", "_____no_output_____" ], [ "## 3 - General Architecture of the learning algorithm ##\n\nIt's time to design a simple algorithm to distinguish cat images from non-cat images.\n\nYou will build a Logistic Regression, using a Neural Network mindset. The following Figure explains why **Logistic Regression is actually a very simple Neural Network!**\n\n<img src=\"images/LogReg_kiank.png\" style=\"width:650px;height:400px;\">\n\n**Mathematical expression of the algorithm**:\n\nFor one example $x^{(i)}$:\n$$z^{(i)} = w^T x^{(i)} + b \\tag{1}$$\n$$\\hat{y}^{(i)} = a^{(i)} = sigmoid(z^{(i)})\\tag{2}$$ \n$$ \\mathcal{L}(a^{(i)}, y^{(i)}) = - y^{(i)} \\log(a^{(i)}) - (1-y^{(i)} ) \\log(1-a^{(i)})\\tag{3}$$\n\nThe cost is then computed by summing over all training examples:\n$$ J = \\frac{1}{m} \\sum_{i=1}^m \\mathcal{L}(a^{(i)}, y^{(i)})\\tag{6}$$\n\n**Key steps**:\nIn this exercise, you will carry out the following steps: \n - Initialize the parameters of the model\n - Learn the parameters for the model by minimizing the cost \n - Use the learned parameters to make predictions (on the test set)\n - Analyse the results and conclude", "_____no_output_____" ], [ "## 4 - Building the parts of our algorithm ## \n\nThe main steps for building a Neural Network are:\n1. Define the model structure (such as number of input features) \n2. Initialize the model's parameters\n3. Loop:\n - Calculate current loss (forward propagation)\n - Calculate current gradient (backward propagation)\n - Update parameters (gradient descent)\n\nYou often build 1-3 separately and integrate them into one function we call `model()`.\n\n### 4.1 - Helper functions\n\n**Exercise**: Using your code from \"Python Basics\", implement `sigmoid()`. As you've seen in the figure above, you need to compute $sigmoid( w^T x + b) = \\frac{1}{1 + e^{-(w^T x + b)}}$ to make predictions. Use np.exp().", "_____no_output_____" ] ], [ [ "# GRADED FUNCTION: sigmoid\n\ndef sigmoid(z):\n \"\"\"\n Compute the sigmoid of z\n\n Arguments:\n z -- A scalar or numpy array of any size.\n\n Return:\n s -- sigmoid(z)\n \"\"\"\n\n ### START CODE HERE ### (≈ 1 line of code)\n s = None\n ### END CODE HERE ###\n \n return s", "_____no_output_____" ], [ "print (\"sigmoid([0, 2]) = \" + str(sigmoid(np.array([0,2]))))", "_____no_output_____" ] ], [ [ "**Expected Output**: \n\n<table>\n <tr>\n <td>**sigmoid([0, 2])**</td>\n <td> [ 0.5 0.88079708]</td> \n </tr>\n</table>", "_____no_output_____" ], [ "### 4.2 - Initializing parameters\n\n**Exercise:** Implement parameter initialization in the cell below. You have to initialize w as a vector of zeros. If you don't know what numpy function to use, look up np.zeros() in the Numpy library's documentation.", "_____no_output_____" ] ], [ [ "# GRADED FUNCTION: initialize_with_zeros\n\ndef initialize_with_zeros(dim):\n \"\"\"\n This function creates a vector of zeros of shape (dim, 1) for w and initializes b to 0.\n \n Argument:\n dim -- size of the w vector we want (or number of parameters in this case)\n \n Returns:\n w -- initialized vector of shape (dim, 1)\n b -- initialized scalar (corresponds to the bias)\n \"\"\"\n \n ### START CODE HERE ### (≈ 1 line of code)\n w = None\n b = None\n ### END CODE HERE ###\n\n assert(w.shape == (dim, 1))\n assert(isinstance(b, float) or isinstance(b, int))\n \n return w, b", "_____no_output_____" ], [ "dim = 2\nw, b = initialize_with_zeros(dim)\nprint (\"w = \" + str(w))\nprint (\"b = \" + str(b))", "_____no_output_____" ] ], [ [ "**Expected Output**: \n\n\n<table style=\"width:15%\">\n <tr>\n <td> ** w ** </td>\n <td> [[ 0.]\n [ 0.]] </td>\n </tr>\n <tr>\n <td> ** b ** </td>\n <td> 0 </td>\n </tr>\n</table>\n\nFor image inputs, w will be of shape (num_px $\\times$ num_px $\\times$ 3, 1).", "_____no_output_____" ], [ "### 4.3 - Forward and Backward propagation\n\nNow that your parameters are initialized, you can do the \"forward\" and \"backward\" propagation steps for learning the parameters.\n\n**Exercise:** Implement a function `propagate()` that computes the cost function and its gradient.\n\n**Hints**:\n\nForward Propagation:\n- You get X\n- You compute $A = \\sigma(w^T X + b) = (a^{(0)}, a^{(1)}, ..., a^{(m-1)}, a^{(m)})$\n- You calculate the cost function: $J = -\\frac{1}{m}\\sum_{i=1}^{m}y^{(i)}\\log(a^{(i)})+(1-y^{(i)})\\log(1-a^{(i)})$\n\nHere are the two formulas you will be using: \n\n$$ \\frac{\\partial J}{\\partial w} = \\frac{1}{m}X(A-Y)^T\\tag{7}$$\n$$ \\frac{\\partial J}{\\partial b} = \\frac{1}{m} \\sum_{i=1}^m (a^{(i)}-y^{(i)})\\tag{8}$$", "_____no_output_____" ] ], [ [ "# GRADED FUNCTION: propagate\n\ndef propagate(w, b, X, Y):\n \"\"\"\n Implement the cost function and its gradient for the propagation explained above\n\n Arguments:\n w -- weights, a numpy array of size (num_px * num_px * 3, 1)\n b -- bias, a scalar\n X -- data of size (num_px * num_px * 3, number of examples)\n Y -- true \"label\" vector (containing 0 if non-cat, 1 if cat) of size (1, number of examples)\n\n Return:\n cost -- negative log-likelihood cost for logistic regression\n dw -- gradient of the loss with respect to w, thus same shape as w\n db -- gradient of the loss with respect to b, thus same shape as b\n \n Tips:\n - Write your code step by step for the propagation. np.log(), np.dot()\n \"\"\"\n \n m = X.shape[1]\n \n # FORWARD PROPAGATION (FROM X TO COST)\n ### START CODE HERE ### (≈ 2 lines of code)\n A = None # compute activation\n cost = None # compute cost\n ### END CODE HERE ###\n \n # BACKWARD PROPAGATION (TO FIND GRAD)\n ### START CODE HERE ### (≈ 2 lines of code)\n dw = None\n db = None\n ### END CODE HERE ###\n\n assert(dw.shape == w.shape)\n assert(db.dtype == float)\n cost = np.squeeze(cost)\n assert(cost.shape == ())\n \n grads = {\"dw\": dw,\n \"db\": db}\n \n return grads, cost", "_____no_output_____" ], [ "w, b, X, Y = np.array([[1],[2]]), 2, np.array([[1,2],[3,4]]), np.array([[1,0]])\ngrads, cost = propagate(w, b, X, Y)\nprint (\"dw = \" + str(grads[\"dw\"]))\nprint (\"db = \" + str(grads[\"db\"]))\nprint (\"cost = \" + str(cost))", "_____no_output_____" ] ], [ [ "**Expected Output**:\n\n<table style=\"width:50%\">\n <tr>\n <td> ** dw ** </td>\n <td> [[ 0.99993216]\n [ 1.99980262]]</td>\n </tr>\n <tr>\n <td> ** db ** </td>\n <td> 0.499935230625 </td>\n </tr>\n <tr>\n <td> ** cost ** </td>\n <td> 6.000064773192205</td>\n </tr>\n\n</table>", "_____no_output_____" ], [ "### d) Optimization\n- You have initialized your parameters.\n- You are also able to compute a cost function and its gradient.\n- Now, you want to update the parameters using gradient descent.\n\n**Exercise:** Write down the optimization function. The goal is to learn $w$ and $b$ by minimizing the cost function $J$. For a parameter $\\theta$, the update rule is $ \\theta = \\theta - \\alpha \\text{ } d\\theta$, where $\\alpha$ is the learning rate.", "_____no_output_____" ] ], [ [ "# GRADED FUNCTION: optimize\n\ndef optimize(w, b, X, Y, num_iterations, learning_rate, print_cost = False):\n \"\"\"\n This function optimizes w and b by running a gradient descent algorithm\n \n Arguments:\n w -- weights, a numpy array of size (num_px * num_px * 3, 1)\n b -- bias, a scalar\n X -- data of shape (num_px * num_px * 3, number of examples)\n Y -- true \"label\" vector (containing 0 if non-cat, 1 if cat), of shape (1, number of examples)\n num_iterations -- number of iterations of the optimization loop\n learning_rate -- learning rate of the gradient descent update rule\n print_cost -- True to print the loss every 100 steps\n \n Returns:\n params -- dictionary containing the weights w and bias b\n grads -- dictionary containing the gradients of the weights and bias with respect to the cost function\n costs -- list of all the costs computed during the optimization, this will be used to plot the learning curve.\n \n Tips:\n You basically need to write down two steps and iterate through them:\n 1) Calculate the cost and the gradient for the current parameters. Use propagate().\n 2) Update the parameters using gradient descent rule for w and b.\n \"\"\"\n \n costs = []\n \n for i in range(num_iterations):\n \n \n # Cost and gradient calculation (≈ 1-4 lines of code)\n ### START CODE HERE ### \n grads, cost = None\n ### END CODE HERE ###\n \n # Retrieve derivatives from grads\n dw = grads[\"dw\"]\n db = grads[\"db\"]\n \n # update rule (≈ 2 lines of code)\n ### START CODE HERE ###\n w = None\n b = None\n ### END CODE HERE ###\n \n # Record the costs\n if i % 100 == 0:\n costs.append(cost)\n \n # Print the cost every 100 training examples\n if print_cost and i % 100 == 0:\n print (\"Cost after iteration %i: %f\" %(i, cost))\n \n params = {\"w\": w,\n \"b\": b}\n \n grads = {\"dw\": dw,\n \"db\": db}\n \n return params, grads, costs", "_____no_output_____" ], [ "params, grads, costs = optimize(w, b, X, Y, num_iterations= 100, learning_rate = 0.009, print_cost = False)\n\nprint (\"w = \" + str(params[\"w\"]))\nprint (\"b = \" + str(params[\"b\"]))\nprint (\"dw = \" + str(grads[\"dw\"]))\nprint (\"db = \" + str(grads[\"db\"]))", "_____no_output_____" ] ], [ [ "**Expected Output**: \n\n<table style=\"width:40%\">\n <tr>\n <td> **w** </td>\n <td>[[ 0.1124579 ]\n [ 0.23106775]] </td>\n </tr>\n \n <tr>\n <td> **b** </td>\n <td> 1.55930492484 </td>\n </tr>\n <tr>\n <td> **dw** </td>\n <td> [[ 0.90158428]\n [ 1.76250842]] </td>\n </tr>\n <tr>\n <td> **db** </td>\n <td> 0.430462071679 </td>\n </tr>\n\n</table>", "_____no_output_____" ], [ "**Exercise:** The previous function will output the learned w and b. We are able to use w and b to predict the labels for a dataset X. Implement the `predict()` function. There is two steps to computing predictions:\n\n1. Calculate $\\hat{Y} = A = \\sigma(w^T X + b)$\n\n2. Convert the entries of a into 0 (if activation <= 0.5) or 1 (if activation > 0.5), stores the predictions in a vector `Y_prediction`. If you wish, you can use an `if`/`else` statement in a `for` loop (though there is also a way to vectorize this). ", "_____no_output_____" ] ], [ [ "# GRADED FUNCTION: predict\n\ndef predict(w, b, X):\n '''\n Predict whether the label is 0 or 1 using learned logistic regression parameters (w, b)\n \n Arguments:\n w -- weights, a numpy array of size (num_px * num_px * 3, 1)\n b -- bias, a scalar\n X -- data of size (num_px * num_px * 3, number of examples)\n \n Returns:\n Y_prediction -- a numpy array (vector) containing all predictions (0/1) for the examples in X\n '''\n \n m = X.shape[1]\n Y_prediction = np.zeros((1,m))\n w = w.reshape(X.shape[0], 1)\n \n # Compute vector \"A\" predicting the probabilities of a cat being present in the picture\n ### START CODE HERE ### (≈ 1 line of code)\n A = None\n ### END CODE HERE ###\n \n for i in range(A.shape[1]):\n \n # Convert probabilities A[0,i] to actual predictions p[0,i]\n ### START CODE HERE ### (≈ 4 lines of code)\n pass\n ### END CODE HERE ###\n \n assert(Y_prediction.shape == (1, m))\n \n return Y_prediction", "_____no_output_____" ], [ "print (\"predictions = \" + str(predict(w, b, X)))", "_____no_output_____" ] ], [ [ "**Expected Output**: \n\n<table style=\"width:30%\">\n <tr>\n <td>\n **predictions**\n </td>\n <td>\n [[ 1. 1.]]\n </td> \n </tr>\n\n</table>\n", "_____no_output_____" ], [ "<font color='blue'>\n**What to remember:**\nYou've implemented several functions that:\n- Initialize (w,b)\n- Optimize the loss iteratively to learn parameters (w,b):\n - computing the cost and its gradient \n - updating the parameters using gradient descent\n- Use the learned (w,b) to predict the labels for a given set of examples", "_____no_output_____" ], [ "## 5 - Merge all functions into a model ##\n\nYou will now see how the overall model is structured by putting together all the building blocks (functions implemented in the previous parts) together, in the right order.\n\n**Exercise:** Implement the model function. Use the following notation:\n - Y_prediction for your predictions on the test set\n - Y_prediction_train for your predictions on the train set\n - w, costs, grads for the outputs of optimize()", "_____no_output_____" ] ], [ [ "# GRADED FUNCTION: model\n\ndef model(X_train, Y_train, X_test, Y_test, num_iterations = 2000, learning_rate = 0.5, print_cost = False):\n \"\"\"\n Builds the logistic regression model by calling the function you've implemented previously\n \n Arguments:\n X_train -- training set represented by a numpy array of shape (num_px * num_px * 3, m_train)\n Y_train -- training labels represented by a numpy array (vector) of shape (1, m_train)\n X_test -- test set represented by a numpy array of shape (num_px * num_px * 3, m_test)\n Y_test -- test labels represented by a numpy array (vector) of shape (1, m_test)\n num_iterations -- hyperparameter representing the number of iterations to optimize the parameters\n learning_rate -- hyperparameter representing the learning rate used in the update rule of optimize()\n print_cost -- Set to true to print the cost every 100 iterations\n \n Returns:\n d -- dictionary containing information about the model.\n \"\"\"\n \n ### START CODE HERE ###\n \n # initialize parameters with zeros (≈ 1 line of code)\n w, b = None\n\n # Gradient descent (≈ 1 line of code)\n parameters, grads, costs = None\n \n # Retrieve parameters w and b from dictionary \"parameters\"\n w = parameters[\"w\"]\n b = parameters[\"b\"]\n \n # Predict test/train set examples (≈ 2 lines of code)\n Y_prediction_test = None\n Y_prediction_train = None\n\n ### END CODE HERE ###\n\n # Print train/test Errors\n print(\"train accuracy: {} %\".format(100 - np.mean(np.abs(Y_prediction_train - Y_train)) * 100))\n print(\"test accuracy: {} %\".format(100 - np.mean(np.abs(Y_prediction_test - Y_test)) * 100))\n\n \n d = {\"costs\": costs,\n \"Y_prediction_test\": Y_prediction_test, \n \"Y_prediction_train\" : Y_prediction_train, \n \"w\" : w, \n \"b\" : b,\n \"learning_rate\" : learning_rate,\n \"num_iterations\": num_iterations}\n \n return d", "_____no_output_____" ] ], [ [ "Run the following cell to train your model.", "_____no_output_____" ] ], [ [ "d = model(train_set_x, train_set_y, test_set_x, test_set_y, num_iterations = 2000, learning_rate = 0.005, print_cost = True)", "_____no_output_____" ] ], [ [ "**Expected Output**: \n\n<table style=\"width:40%\"> \n \n <tr>\n <td> **Train Accuracy** </td> \n <td> 99.04306220095694 % </td>\n </tr>\n\n <tr>\n <td>**Test Accuracy** </td> \n <td> 70.0 % </td>\n </tr>\n</table> \n\n\n", "_____no_output_____" ], [ "**Comment**: Training accuracy is close to 100%. This is a good sanity check: your model is working and has high enough capacity to fit the training data. Test error is 68%. It is actually not bad for this simple model, given the small dataset we used and that logistic regression is a linear classifier. But no worries, you'll build an even better classifier next week!\n\nAlso, you see that the model is clearly overfitting the training data. Later in this specialization you will learn how to reduce overfitting, for example by using regularization. Using the code below (and changing the `index` variable) you can look at predictions on pictures of the test set.", "_____no_output_____" ] ], [ [ "# Example of a picture that was wrongly classified.\nindex = 1\nplt.imshow(test_set_x[:,index].reshape((num_px, num_px, 3)))\nprint (\"y = \" + str(test_set_y[0,index]) + \", you predicted that it is a \\\"\" + classes[d[\"Y_prediction_test\"][0,index]].decode(\"utf-8\") + \"\\\" picture.\")", "_____no_output_____" ] ], [ [ "Let's also plot the cost function and the gradients.", "_____no_output_____" ] ], [ [ "# Plot learning curve (with costs)\ncosts = np.squeeze(d['costs'])\nplt.plot(costs)\nplt.ylabel('cost')\nplt.xlabel('iterations (per hundreds)')\nplt.title(\"Learning rate =\" + str(d[\"learning_rate\"]))\nplt.show()", "_____no_output_____" ] ], [ [ "**Interpretation**:\nYou can see the cost decreasing. It shows that the parameters are being learned. However, you see that you could train the model even more on the training set. Try to increase the number of iterations in the cell above and rerun the cells. You might see that the training set accuracy goes up, but the test set accuracy goes down. This is called overfitting. ", "_____no_output_____" ], [ "## 6 - Further analysis (optional/ungraded exercise) ##\n\nCongratulations on building your first image classification model. Let's analyze it further, and examine possible choices for the learning rate $\\alpha$. ", "_____no_output_____" ], [ "#### Choice of learning rate ####\n\n**Reminder**:\nIn order for Gradient Descent to work you must choose the learning rate wisely. The learning rate $\\alpha$ determines how rapidly we update the parameters. If the learning rate is too large we may \"overshoot\" the optimal value. Similarly, if it is too small we will need too many iterations to converge to the best values. That's why it is crucial to use a well-tuned learning rate.\n\nLet's compare the learning curve of our model with several choices of learning rates. Run the cell below. This should take about 1 minute. Feel free also to try different values than the three we have initialized the `learning_rates` variable to contain, and see what happens. ", "_____no_output_____" ] ], [ [ "learning_rates = [0.01, 0.001, 0.0001]\nmodels = {}\nfor i in learning_rates:\n print (\"learning rate is: \" + str(i))\n models[str(i)] = model(train_set_x, train_set_y, test_set_x, test_set_y, num_iterations = 1500, learning_rate = i, print_cost = False)\n print ('\\n' + \"-------------------------------------------------------\" + '\\n')\n\nfor i in learning_rates:\n plt.plot(np.squeeze(models[str(i)][\"costs\"]), label= str(models[str(i)][\"learning_rate\"]))\n\nplt.ylabel('cost')\nplt.xlabel('iterations')\n\nlegend = plt.legend(loc='upper center', shadow=True)\nframe = legend.get_frame()\nframe.set_facecolor('0.90')\nplt.show()", "_____no_output_____" ] ], [ [ "**Interpretation**: \n- Different learning rates give different costs and thus different predictions results.\n- If the learning rate is too large (0.01), the cost may oscillate up and down. It may even diverge (though in this example, using 0.01 still eventually ends up at a good value for the cost). \n- A lower cost doesn't mean a better model. You have to check if there is possibly overfitting. It happens when the training accuracy is a lot higher than the test accuracy.\n- In deep learning, we usually recommend that you: \n - Choose the learning rate that better minimizes the cost function.\n - If your model overfits, use other techniques to reduce overfitting. (We'll talk about this in later videos.) \n", "_____no_output_____" ], [ "## 7 - Test with your own image (optional/ungraded exercise) ##\n\nCongratulations on finishing this assignment. You can use your own image and see the output of your model. To do that:\n 1. Click on \"File\" in the upper bar of this notebook, then click \"Open\" to go on your Coursera Hub.\n 2. Add your image to this Jupyter Notebook's directory, in the \"images\" folder\n 3. Change your image's name in the following code\n 4. Run the code and check if the algorithm is right (1 = cat, 0 = non-cat)!", "_____no_output_____" ] ], [ [ "## START CODE HERE ## (PUT YOUR IMAGE NAME) \nmy_image = \"my_image.jpg\" # change this to the name of your image file \n## END CODE HERE ##\n\n# We preprocess the image to fit your algorithm.\nfname = \"images/\" + my_image\nimage = np.array(ndimage.imread(fname, flatten=False))\nmy_image = scipy.misc.imresize(image, size=(num_px,num_px)).reshape((1, num_px*num_px*3)).T\nmy_predicted_image = predict(d[\"w\"], d[\"b\"], my_image)\n\nplt.imshow(image)\nprint(\"y = \" + str(np.squeeze(my_predicted_image)) + \", your algorithm predicts a \\\"\" + classes[int(np.squeeze(my_predicted_image)),].decode(\"utf-8\") + \"\\\" picture.\")", "_____no_output_____" ] ], [ [ "<font color='blue'>\n**What to remember from this assignment:**\n1. Preprocessing the dataset is important.\n2. You implemented each function separately: initialize(), propagate(), optimize(). Then you built a model().\n3. Tuning the learning rate (which is an example of a \"hyperparameter\") can make a big difference to the algorithm. You will see more examples of this later in this course!", "_____no_output_____" ], [ "Finally, if you'd like, we invite you to try different things on this Notebook. Make sure you submit before trying anything. Once you submit, things you can play with include:\n - Play with the learning rate and the number of iterations\n - Try different initialization methods and compare the results\n - Test other preprocessings (center the data, or divide each row by its standard deviation)", "_____no_output_____" ], [ "Bibliography:\n- http://www.wildml.com/2015/09/implementing-a-neural-network-from-scratch/\n- https://stats.stackexchange.com/questions/211436/why-do-we-normalize-images-by-subtracting-the-datasets-image-mean-and-not-the-c", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown" ]

[ [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import pandas as pd\nimport numpy as np\nimport os\n\nos.chdir('/Users/gianni/Google Drive/Bas Zahy Gianni - Games/Data')", "_____no_output_____" ], [ "oc = [\n 'index', 'subject', 'color', 'gi', 'mi', \n 'status', 'bp', 'wp', 'response', 'rt',\n 'time', 'mouse_t', 'mouse_x'\n]\n\nfc = [\n 'subject', 'is_comp', 'color', 'status',\n 'bp', 'wp', 'response', 'rt', 'gi', 'mi',\n 'computer', 'human', 'time'\n]\n\nmc = ['subject', 'color', 'bp', 'wp', 'response', 'rt', 'condition']\n\nclass Data():\n \"\"\" Data is the primary object for holding experimental data. It also contains functions\n for the loading, cleaning, augmentation, and export of the data tables. \"\"\"\n\n def __init__(self, folder):\n self.data = self.load(folder)\n\n def load_file(self, folder, file_name, mouse=False):\n \"\"\" Initial preparation of data for individual files \"\"\"\n print(file_name[:-4])\n\n # load file, drop nuissance columns, remove non-observations\n drop_cols = ['index'] if mouse else ['index', 'mouse_t', 'mouse_x']\n data = pd.read_csv(folder + file_name, names=oc).drop(drop_cols, axis=1)\n drop_status = (data.status != 'dummy') & (data.status != 'ready') & (data.status != 'draw offer')\n data = data.loc[drop_status, :].copy().reset_index(drop=True)\n\n # assign unique subject label (from filename) and create separate cols for humans and computers\n sub_filter = data.rt > 0\n comp_filter = data.rt == 0\n first_move_filter = data.bp.map(lambda x: np.array(list(x)).astype(int).sum()==1) #(data.mi == 0) & (data.gi%2 == 0)\n second_move_filter = data.bp.map(lambda x: np.array(list(x)).astype(int).sum()==2) #(data.mi == 1) & (data.gi%2 == 0)\n condition_filter = (data.rt>0)&(data.status == 'playing')\n data.loc[condition_filter, 'condition'] = data.loc[condition_filter, 'subject'].map(lambda x: x[-1])\n data.loc[:, 'condition'] = data.loc[:, 'condition'].fillna(method='ffill')\n \n data.loc[data.rt > 0, 'subject'] = file_name[:-4]\n data.loc[:, 'human'] = file_name[:-4]\n data.loc[:, 'computer'] = np.nan\n data.loc[comp_filter, 'computer'] = data.loc[comp_filter, 'subject']\n data.loc[first_move_filter, 'computer'] = data.loc[second_move_filter, 'computer']\n data.loc[:, 'computer'] = data.loc[:, 'computer'].fillna(method='ffill')\n data.loc[0, 'computer'] = data.loc[1, 'computer']\n\n return data\n\n def load(self, folder):\n \"\"\" Calls other functions to corrale data and some support information \"\"\"\n self.exp_name = folder\n files = os.listdir(folder + '/Raw/')\n files = [f for f in files if f[-3:] == 'csv']\n# files =[f for f in files if f[:-4] != 'HH']\n self.subjects = [f[:-4] for f in files]\n self.subject_dict = dict(zip(self.subjects, np.arange(len(self.subjects))))\n data = pd.concat([self.load_file(folder + '/Raw/', f) for f in files])\n data = data.reset_index(drop=True)\n data = self.clean(data)\n\n return data\n\n def clean(self, df):\n \"\"\" Performs further cleaning that can be done on all data collectively \"\"\"\n\n # anonymize subjects\n sub_filter = df.rt > 0 # filter computers out\n df.loc[sub_filter, 'subject'] = df.loc[sub_filter, 'subject'].map(self.subject_dict)\n df.loc[:, 'human'] = df.loc[:, 'human'].map(self.subject_dict)\n\n # give computers identifiable names\n comp_filter = df.rt == 0\n df.loc[comp_filter, 'subject'] = df.loc[comp_filter, 'subject'].astype(int) + 1000\n df.loc[pd.notnull(df.computer), 'computer'] = df.loc[pd.notnull(df.computer), 'computer'].astype(int) + 1000\n\n # force remove response from board\n for i in df.loc[df.status != 'EVAL', :].index.values:\n if df.loc[i,\"color\"] == 0:\n l = list(df.loc[i,\"bp\"])\n l[df.loc[i, \"response\"]] = '0'\n df.loc[i,\"bp\"] = ''.join(l)\n else:\n l = list(df.loc[i,\"wp\"])\n l[df.loc[i,\"response\"]] = '0'\n df.loc[i,\"wp\"] = ''.join(l)\n\n # force correct colors\n count_pieces = lambda x: np.array([np.array(list(df.loc[i, x])).astype(int).sum() for i in df.index.values])\n df.loc[:, 'color'] = count_pieces('bp') - count_pieces('wp')\n df.loc[:, 'color'] = df.loc[:, 'color'].astype(int).astype(str)\n\n # add is_comp\n is_computer = lambda x: \"0\" if x > 0 else \"1\"\n df.loc[:, 'is_comp'] = df.loc[:, 'rt'].map(is_computer)\n\n # correct move index in games\n df.loc[df.status.isin(['playing', 'win', 'draw', 'timeout']), 'mi'] = df.loc[df.status.isin(['playing', 'win', 'draw']), 'mi'] - 1\n return df\n\n def export_individuals(self, folder):\n for s, i in self.subject_dict.items():\n c = self.data.human == i\n d = self.data.loc[c, :].reset_index(drop=True)\n d = d.reindex_axis(self.full_output_columns, axis=1)\n d.to_csv(folder + '/Clean/' + s + '.csv', index=False)\n\n return None\n\n def export(self, folder):\n f = folder + 'Clean/_summaries/'\n E = self.data.loc[self.data.status.isin(['playing', 'win', 'draw', 'timeout']), :]\n E.loc[:, fc].to_csv(f + 'all_fields.csv', index=False)\n E.loc[:, mc].to_csv(f + 'model_fields.csv', index=False)\n \n return None", "_____no_output_____" ], [ "D = Data('./5_tai')\n\nD.export('./5_tai/')", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# SLU10 - Metrics for regression: Example Notebook\n\nIn this notebook [some regression validation metrics offered by scikit-learn](http://scikit-learn.org/stable/modules/model_evaluation.html#common-cases-predefined-values) are presented.", "_____no_output_____" ] ], [ [ "import numpy as np\nimport pandas as pd\nfrom sklearn.datasets import load_boston\nfrom sklearn.linear_model import LinearRegression\n\n# some scikit-learn regression validation metrics\nfrom sklearn.metrics import mean_squared_error, mean_absolute_error, r2_score\n\nnp.random.seed(60)", "_____no_output_____" ] ], [ [ "# Load Data\nLoad the Boston house-prices dataset, fit a Linear Regression, and make prediction on the dataset (used to create the model).", "_____no_output_____" ] ], [ [ "data = load_boston()\n\nx = pd.DataFrame(data['data'], columns=data['feature_names'])\ny = pd.Series(data['target'])\n\nlr = LinearRegression()\nlr.fit(x, y)\n\ny_hat = lr.predict(x)", "_____no_output_____" ] ], [ [ "# Metrics with scikitlearn\n\nBelow follows a list of metrics made available by scikitlearn and its usage:", "_____no_output_____" ], [ "## Mean Squared Error\n\n$$MSE = \\frac{1}{N} \\sum_{n=1}^N (y_n - \\hat{y}_n)^2$$", "_____no_output_____" ] ], [ [ "mean_squared_error(y, y_hat)", "_____no_output_____" ] ], [ [ "## Root Mean Squared Error\n\n$$RMSE = \\sqrt{MSE}$$", "_____no_output_____" ] ], [ [ "np.sqrt(mean_squared_error(y, y_hat))", "_____no_output_____" ] ], [ [ "## Mean Absolute Error\n\n$$MAE = \\frac{1}{N} \\sum_{n=1}^N \\left| y_n - \\hat{y}_n \\right|$$", "_____no_output_____" ] ], [ [ "mean_absolute_error(y, y_hat)", "_____no_output_____" ] ], [ [ "## R² score\n\n$$\\bar{y} = \\frac{1}{N} \\sum_{n=1}^N y_n$$\n\n$$R² = 1 - \\frac{MSE(y, \\hat{y})}{MSE(y, \\bar{y})} \n= 1 - \\frac{\\frac{1}{N} \\sum_{n=1}^N (y_n - \\hat{y}_n)^2}{\\frac{1}{N} \\sum_{n=1}^N (y_n - \\bar{y})^2}\n= 1 - \\frac{\\sum_{n=1}^N (y_n - \\hat{y}_n)^2}{\\sum_{n=1}^N (y_n - \\bar{y})^2}$$", "_____no_output_____" ] ], [ [ "r2_score(y, y_hat)", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ] ]

[ "CC0-1.0" ]

[ "CC0-1.0" ]

[ "CC0-1.0" ]

[ [ [ "import pandas as pd ", "_____no_output_____" ], [ "train = pd.read_csv(\"http://bit.ly/kaggletrain\")", "_____no_output_____" ], [ "train.head() ", "_____no_output_____" ], [ "feature_cols = ['Pclass', 'Parch'] \nX = train.loc[:, feature_cols] ", "_____no_output_____" ], [ "X.shape", "_____no_output_____" ], [ "y = train.Survived", "_____no_output_____" ], [ "y.shape", "_____no_output_____" ], [ "from sklearn.linear_model import LogisticRegression\nlogreg = LogisticRegression() \nlogreg.fit(X, y)", "/home/jubayer/anaconda3/lib/python3.7/site-packages/sklearn/linear_model/logistic.py:432: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n FutureWarning)\n" ], [ "test = pd.read_csv(\"http://bit.ly/kaggletest\")", "_____no_output_____" ], [ "test.head() ", "_____no_output_____" ], [ "X_new = test.loc[:, feature_cols] ", "_____no_output_____" ], [ "X_new.shape", "_____no_output_____" ], [ "new_pred_class = logreg.predict(X_new)", "_____no_output_____" ], [ "test.PassengerId", "_____no_output_____" ], [ "new_pred_class", "_____no_output_____" ], [ "pd.DataFrame({'PassengerID' : test.PassengerId, 'Survived': new_pred_class}).to_csv(\"sub.csv\", index=False)", "_____no_output_____" ], [ "subdf = pd.read_csv(\"sub.csv\")", "_____no_output_____" ], [ "subdf.head() ", "_____no_output_____" ] ], [ [ "<h3>About the Author</h3>\nThis repo was created by <a href=\"https://www.linkedin.com/in/jubayer28/\" target=\"_blank\">Jubayer Hossain</a> <br>\n<a href=\"https://www.linkedin.com/in/jubayer28/\" target=\"_blank\">Jubayer Hossain</a> is a student of Microbiology at Jagannath University and the founder of <a href=\"https://github.com/hdro\" target=\"_blank\">Health Data Research Organization</a>. He is also a team member of a bioinformatics research group known as Bio-Bio-1. \n\n<a rel=\"license\" href=\"http://creativecommons.org/licenses/by-nc-sa/4.0/\"><img alt=\"Creative Commons License\" style=\"border-width:0\" src=\"https://i.creativecommons.org/l/by-nc-sa/4.0/88x31.png\" /></a><br />This work is licensed under a <a rel=\"license\" href=\"http://creativecommons.org/licenses/by-nc-sa/4.0/\">Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License</a>.m", "_____no_output_____" ] ] ]

[ "code", "markdown" ]

[ [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ], [ "markdown" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "from jqdatasdk import *\nauth('18620668927', 'minpeng123')", "_____no_output_____" ], [ "#记录由上市公司年报、中报、一季报、三季报统计出的分红转增情况\nq = query(finance.STK_XR_XD).filter(finance.STK_XR_XD.report_date >= '2019-01-01').limit(10)\nfinance.run_query(q)", "_____no_output_____" ], [ "#记录沪深两市股票交易的成交情况，包括市值、成交量，市盈率等情况\nq = query(finance.STK_EXCHANGE_TRADE_INFO).filter(finance.STK_EXCHANGE_TRADE_INFO.date >= '2019-07-09').limit(10)\nfinance.run_query(q)", "_____no_output_____" ], [ "#描述：记录上海交易所和深圳交易所的融资融券汇总数据\nfinance.run_query(query(finance.STK_MT_TOTAL).filter(finance.STK_MT_TOTAL.date == '2019-04-23').limit(10))", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Import the required libraries\nimport warnings\nwarnings.filterwarnings('ignore')\n\nimport pandas as pd\nimport numpy as np\nimport seaborn as sns\nimport matplotlib.pyplot as plt\nimport joblib\n%matplotlib inline\n\nfrom sklearn.linear_model import LogisticRegression", "_____no_output_____" ], [ "# Read the data and display\n\ndiabetesDF = pd.read_csv('diabetes.csv')\ndiabetesDF.head()", "_____no_output_____" ], [ "diabetesDF.drop(['Pregnancies', 'Glucose', 'Age'], axis=1, inplace=True)", "_____no_output_____" ], [ "diabetesDF.head()", "_____no_output_____" ], [ "# Total 768 patients record\n# Using 650 data for training\n# Using 100 data for testing\n# Using 18 data for validation\n\ndfTrain = diabetesDF[:650]\ndfTest = diabetesDF[650:750]\ndfCheck = diabetesDF[750:]", "_____no_output_____" ], [ "# Separating label and features and converting to numpy array to feed into our model\ntrainLabel = np.asarray(dfTrain['Outcome'])\ntrainData = np.asarray(dfTrain.drop('Outcome',1))\ntestLabel = np.asarray(dfTest['Outcome'])\ntestData = np.asarray(dfTest.drop('Outcome',1))", "_____no_output_____" ], [ "# Normalize the data \nmeans = np.mean(trainData, axis=0)\nstds = np.std(trainData, axis=0)\n\ntrainData = (trainData - means)/stds\ntestData = (testData - means)/stds", "_____no_output_____" ], [ "# models target t as sigmoid(w0 + w1*x1 + w2*x2 + ... + wd*xd)\ndiabetesCheck = LogisticRegression()\ndiabetesCheck.fit(trainData,trainLabel)\naccuracy = diabetesCheck.score(testData,testLabel)\nprint(\"accuracy = \",accuracy * 100,\"%\")", "accuracy = 67.0 %\n" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import pandas as pd\n\nfrom sklearn.cluster import MeanShift", "_____no_output_____" ], [ "df_candies = pd.read_csv('../data/raw/candy.csv')\n\nx = df_candies.drop('competitorname', axis=1)\n\nmeanshift = MeanShift().fit(x)\nprint(meanshift.labels_)\nprint('_'*64)\nprint(meanshift.cluster_centers_)", "[2 2 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 1 1 0 1 0 1 2 1 2 0 1 2 0 0 1 2 2 0 1 2\n 2 2 1 1 1 2 2 0 0 0 2 0 0 0 2 2 2 2 0 2 0 0 0 2 1 0 0 2 2 2 1 2 0 0 0 0 1\n 1 0 0 1 1 2 0 0 0 0 1]\n________________________________________________________________\n[[2.25000000e-01 5.75000000e-01 1.00000000e-01 2.50000000e-02\n 5.00000000e-02 2.50000000e-02 3.00000000e-01 1.00000000e-01\n 5.50000000e-01 4.57599993e-01 3.67824996e-01 4.10442122e+01]\n [4.68750000e-01 5.00000000e-01 1.25000000e-01 1.56250000e-01\n 9.37500000e-02 6.25000000e-02 1.25000000e-01 3.12500000e-01\n 5.31250000e-01 4.57281243e-01 4.67874998e-01 5.21138597e+01]\n [8.26086957e-01 1.73913043e-01 3.04347826e-01 3.04347826e-01\n 1.73913043e-01 1.73913043e-01 0.00000000e+00 5.21739130e-01\n 4.34782609e-01 5.81391293e-01 6.38086963e-01 6.47120799e+01]]\n" ] ] ]

[ "code" ]

[ [ "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Planar data classification with one hidden layer\n\nWelcome to your week 3 programming assignment! It's time to build your first neural network, which will have one hidden layer. Now, you'll notice a big difference between this model and the one you implemented previously using logistic regression.\n\nBy the end of this assignment, you'll be able to:\n\n- Implement a 2-class classification neural network with a single hidden layer\n- Use units with a non-linear activation function, such as tanh\n- Compute the cross entropy loss\n- Implement forward and backward propagation\n", "_____no_output_____" ], [ "## Table of Contents\n- [1 - Packages](#1)\n- [2 - Load the Dataset](#2)\n - [Exercise 1](#ex-1)\n- [3 - Simple Logistic Regression](#3)\n- [4 - Neural Network model](#4)\n - [4.1 - Defining the neural network structure](#4-1)\n - [Exercise 2 - layer_sizes](#ex-2)\n - [4.2 - Initialize the model's parameters](#4-2)\n - [Exercise 3 - initialize_parameters](#ex-3)\n - [4.3 - The Loop](#4-3)\n - [Exercise 4 - forward_propagation](#ex-4)\n - [4.4 - Compute the Cost](#4-4)\n - [Exercise 5 - compute_cost](#ex-5)\n - [4.5 - Implement Backpropagation](#4-5)\n - [Exercise 6 - backward_propagation](#ex-6)\n - [4.6 - Update Parameters](#4-6)\n - [Exercise 7 - update_parameters](#ex-7)\n - [4.7 - Integration](#4-7)\n - [Exercise 8 - nn_model](#ex-8)\n- [5 - Test the Model](#5)\n - [5.1 - Predict](#5-1)\n - [Exercise 9 - predict](#ex-9)\n - [5.2 - Test the Model on the Planar Dataset](#5-2)\n- [6 - Tuning hidden layer size (optional/ungraded exercise)](#6)\n- [7- Performance on other datasets](#7)", "_____no_output_____" ], [ "<a name='1'></a>\n# 1 - Packages\n\nFirst import all the packages that you will need during this assignment.\n\n- [numpy](https://www.numpy.org/) is the fundamental package for scientific computing with Python.\n- [sklearn](http://scikit-learn.org/stable/) provides simple and efficient tools for data mining and data analysis. \n- [matplotlib](http://matplotlib.org) is a library for plotting graphs in Python.\n- testCases provides some test examples to assess the correctness of your functions\n- planar_utils provide various useful functions used in this assignment", "_____no_output_____" ] ], [ [ "# Package imports\nimport numpy as np\nimport copy\nimport matplotlib.pyplot as plt\nfrom testCases_v2 import *\nfrom public_tests import *\nimport sklearn\nimport sklearn.datasets\nimport sklearn.linear_model\nfrom planar_utils import plot_decision_boundary, sigmoid, load_planar_dataset, load_extra_datasets\n\n%matplotlib inline\n\nnp.random.seed(2) # set a seed so that the results are consistent\n\n%load_ext autoreload\n%autoreload 2", "_____no_output_____" ] ], [ [ "<a name='2'></a>\n# 2 - Load the Dataset \n\nNow, load the dataset you'll be working on. The following code will load a \"flower\" 2-class dataset into variables X and Y.", "_____no_output_____" ] ], [ [ "X, Y = load_planar_dataset()", "_____no_output_____" ] ], [ [ "Visualize the dataset using matplotlib. The data looks like a \"flower\" with some red (label y=0) and some blue (y=1) points. Your goal is to build a model to fit this data. In other words, we want the classifier to define regions as either red or blue.", "_____no_output_____" ] ], [ [ "# Visualize the data:\nplt.scatter(X[0, :], X[1, :], c=Y, s=40, cmap=plt.cm.Spectral);", "_____no_output_____" ] ], [ [ "You have:\n - a numpy-array (matrix) X that contains your features (x1, x2)\n - a numpy-array (vector) Y that contains your labels (red:0, blue:1).\n\nFirst, get a better sense of what your data is like. \n\n<a name='ex-1'></a>\n### Exercise 1 \n\nHow many training examples do you have? In addition, what is the `shape` of the variables `X` and `Y`? \n\n**Hint**: How do you get the shape of a numpy array? [(help)](https://docs.scipy.org/doc/numpy/reference/generated/numpy.ndarray.shape.html)", "_____no_output_____" ] ], [ [ "# (≈ 3 lines of code)\n# shape_X = ...\n# shape_Y = ...\n# training set size\n# m = ...\n# YOUR CODE STARTS HERE\nshape_X = X.shape\nshape_Y = Y.shape\nm = Y.shape[1]\n# YOUR CODE ENDS HERE\n\nprint ('The shape of X is: ' + str(shape_X))\nprint ('The shape of Y is: ' + str(shape_Y))\nprint ('I have m = %d training examples!' % (m))\nprint(X.shape[0])", "The shape of X is: (2, 400)\nThe shape of Y is: (1, 400)\nI have m = 400 training examples!\n2\n" ] ], [ [ "**Expected Output**:\n \n<table style=\"width:20%\">\n <tr>\n <td> shape of X </td>\n <td> (2, 400) </td> \n </tr>\n <tr>\n <td>shape of Y</td>\n <td>(1, 400) </td> \n </tr>\n <tr>\n <td>m</td>\n <td> 400 </td> \n </tr>\n</table>", "_____no_output_____" ], [ "<a name='3'></a>\n## 3 - Simple Logistic Regression\n\nBefore building a full neural network, let's check how logistic regression performs on this problem. You can use sklearn's built-in functions for this. Run the code below to train a logistic regression classifier on the dataset.", "_____no_output_____" ] ], [ [ "# Train the logistic regression classifier\nclf = sklearn.linear_model.LogisticRegressionCV();\nclf.fit(X.T, Y.T);", "_____no_output_____" ] ], [ [ "You can now plot the decision boundary of these models! Run the code below.", "_____no_output_____" ] ], [ [ "# Plot the decision boundary for logistic regression\nplot_decision_boundary(lambda x: clf.predict(x), X, Y)\nplt.title(\"Logistic Regression\")\nprint(X.shape)\n# Print accuracy\nLR_predictions = clf.predict(X.T)\nprint ('Accuracy of logistic regression: %d ' % float((np.dot(Y,LR_predictions) + np.dot(1-Y,1-LR_predictions))/float(Y.size)*100) +\n '% ' + \"(percentage of correctly labelled datapoints)\")", "(2, 400)\nAccuracy of logistic regression: 47 % (percentage of correctly labelled datapoints)\n" ] ], [ [ "**Expected Output**:\n\n<table style=\"width:20%\">\n <tr>\n <td>Accuracy</td>\n <td> 47% </td> \n </tr>\n \n</table>\n", "_____no_output_____" ], [ "**Interpretation**: The dataset is not linearly separable, so logistic regression doesn't perform well. Hopefully a neural network will do better. Let's try this now! ", "_____no_output_____" ], [ "<a name='4'></a>\n## 4 - Neural Network model\n\nLogistic regression didn't work well on the flower dataset. Next, you're going to train a Neural Network with a single hidden layer and see how that handles the same problem.\n\n**The model**:\n<img src=\"images/classification_kiank.png\" style=\"width:600px;height:300px;\">\n\n**Mathematically**:\n\nFor one example $x^{(i)}$:\n$$z^{[1] (i)} = W^{[1]} x^{(i)} + b^{[1]}\\tag{1}$$ \n$$a^{[1] (i)} = \\tanh(z^{[1] (i)})\\tag{2}$$\n$$z^{[2] (i)} = W^{[2]} a^{[1] (i)} + b^{[2]}\\tag{3}$$\n$$\\hat{y}^{(i)} = a^{[2] (i)} = \\sigma(z^{ [2] (i)})\\tag{4}$$\n$$y^{(i)}_{prediction} = \\begin{cases} 1 & \\mbox{if } a^{[2](i)} > 0.5 \\\\ 0 & \\mbox{otherwise } \\end{cases}\\tag{5}$$\n\nGiven the predictions on all the examples, you can also compute the cost $J$ as follows: \n$$J = - \\frac{1}{m} \\sum\\limits_{i = 0}^{m} \\large\\left(\\small y^{(i)}\\log\\left(a^{[2] (i)}\\right) + (1-y^{(i)})\\log\\left(1- a^{[2] (i)}\\right) \\large \\right) \\small \\tag{6}$$\n\n**Reminder**: The general methodology to build a Neural Network is to:\n 1. Define the neural network structure ( # of input units, # of hidden units, etc). \n 2. Initialize the model's parameters\n 3. Loop:\n - Implement forward propagation\n - Compute loss\n - Implement backward propagation to get the gradients\n - Update parameters (gradient descent)\n\nIn practice, you'll often build helper functions to compute steps 1-3, then merge them into one function called `nn_model()`. Once you've built `nn_model()` and learned the right parameters, you can make predictions on new data.", "_____no_output_____" ], [ "<a name='4-1'></a>\n### 4.1 - Defining the neural network structure ####\n\n<a name='ex-2'></a>\n### Exercise 2 - layer_sizes \n\nDefine three variables:\n - n_x: the size of the input layer\n - n_h: the size of the hidden layer (set this to 4) \n - n_y: the size of the output layer\n\n**Hint**: Use shapes of X and Y to find n_x and n_y. Also, hard code the hidden layer size to be 4.", "_____no_output_____" ] ], [ [ "# GRADED FUNCTION: layer_sizes\n\ndef layer_sizes(X, Y):\n \"\"\"\n Arguments:\n X -- input dataset of shape (input size, number of examples)\n Y -- labels of shape (output size, number of examples)\n \n Returns:\n n_x -- the size of the input layer\n n_h -- the size of the hidden layer\n n_y -- the size of the output layer\n \"\"\"\n #(≈ 3 lines of code)\n # n_x = ... \n # n_h = ...\n # n_y = ... \n # YOUR CODE STARTS HERE\n n_x = X.shape[0]\n n_h = 4\n n_y = Y.shape[0]\n \n print(Y.shape)\n # YOUR CODE ENDS HERE\n return (n_x, n_h, n_y)", "_____no_output_____" ], [ "t_X, t_Y = layer_sizes_test_case()\n(n_x, n_h, n_y) = layer_sizes(t_X, t_Y)\nprint(\"The size of the input layer is: n_x = \" + str(n_x))\nprint(\"The size of the hidden layer is: n_h = \" + str(n_h))\nprint(\"The size of the output layer is: n_y = \" + str(n_y))\n\nlayer_sizes_test(layer_sizes)", "(2, 3)\nThe size of the input layer is: n_x = 5\nThe size of the hidden layer is: n_h = 4\nThe size of the output layer is: n_y = 2\n(2, 3)\n(2, 3)\n\u001b[92m All tests passed.\n" ] ], [ [ "***Expected output***\n```\nThe size of the input layer is: n_x = 5\nThe size of the hidden layer is: n_h = 4\nThe size of the output layer is: n_y = 2\n```", "_____no_output_____" ], [ "<a name='4-2'></a>\n### 4.2 - Initialize the model's parameters ####\n\n<a name='ex-3'></a>\n### Exercise 3 - initialize_parameters\n\nImplement the function `initialize_parameters()`.\n\n**Instructions**:\n- Make sure your parameters' sizes are right. Refer to the neural network figure above if needed.\n- You will initialize the weights matrices with random values. \n - Use: `np.random.randn(a,b) * 0.01` to randomly initialize a matrix of shape (a,b).\n- You will initialize the bias vectors as zeros. \n - Use: `np.zeros((a,b))` to initialize a matrix of shape (a,b) with zeros.", "_____no_output_____" ] ], [ [ "# GRADED FUNCTION: initialize_parameters\n\ndef initialize_parameters(n_x, n_h, n_y):\n \"\"\"\n Argument:\n n_x -- size of the input layer\n n_h -- size of the hidden layer\n n_y -- size of the output layer\n \n Returns:\n params -- python dictionary containing your parameters:\n W1 -- weight matrix of shape (n_h, n_x)\n b1 -- bias vector of shape (n_h, 1)\n W2 -- weight matrix of shape (n_y, n_h)\n b2 -- bias vector of shape (n_y, 1)\n \"\"\"\n \n np.random.seed(2) # we set up a seed so that your output matches ours although the initialization is random.\n \n #(≈ 4 lines of code)\n # W1 = ...\n # b1 = ...\n # W2 = ...\n # b2 = ...\n # YOUR CODE STARTS HERE\n W1 = np.random.randn(n_h, n_x) * 0.01\n b1 = np.zeros((n_h, 1))\n W2 = np.random.randn(n_y, n_h) * 0.01\n b2 = np.zeros((n_y, 1))\n \n # YOUR CODE ENDS HERE\n\n parameters = {\"W1\": W1,\n \"b1\": b1,\n \"W2\": W2,\n \"b2\": b2}\n \n return parameters", "_____no_output_____" ], [ "n_x, n_h, n_y = initialize_parameters_test_case()\nparameters = initialize_parameters(n_x, n_h, n_y)\n\nprint(\"W1 = \" + str(parameters[\"W1\"]))\nprint(\"b1 = \" + str(parameters[\"b1\"]))\nprint(\"W2 = \" + str(parameters[\"W2\"]))\nprint(\"b2 = \" + str(parameters[\"b2\"]))\n\ninitialize_parameters_test(initialize_parameters)", "W1 = [[-0.00416758 -0.00056267]\n [-0.02136196 0.01640271]\n [-0.01793436 -0.00841747]\n [ 0.00502881 -0.01245288]]\nb1 = [[0.]\n [0.]\n [0.]\n [0.]]\nW2 = [[-0.01057952 -0.00909008 0.00551454 0.02292208]]\nb2 = [[0.]]\n\u001b[92m All tests passed.\n" ] ], [ [ "**Expected output**\n```\nW1 = [[-0.00416758 -0.00056267]\n [-0.02136196 0.01640271]\n [-0.01793436 -0.00841747]\n [ 0.00502881 -0.01245288]]\nb1 = [[0.]\n [0.]\n [0.]\n [0.]]\nW2 = [[-0.01057952 -0.00909008 0.00551454 0.02292208]]\nb2 = [[0.]]\n```", "_____no_output_____" ], [ "<a name='4-3'></a>\n### 4.3 - The Loop \n\n<a name='ex-4'></a>\n### Exercise 4 - forward_propagation\n\nImplement `forward_propagation()` using the following equations:\n\n$$Z^{[1]} = W^{[1]} X + b^{[1]}\\tag{1}$$ \n$$A^{[1]} = \\tanh(Z^{[1]})\\tag{2}$$\n$$Z^{[2]} = W^{[2]} A^{[1]} + b^{[2]}\\tag{3}$$\n$$\\hat{Y} = A^{[2]} = \\sigma(Z^{[2]})\\tag{4}$$\n\n\n**Instructions**:\n\n- Check the mathematical representation of your classifier in the figure above.\n- Use the function `sigmoid()`. It's built into (imported) this notebook.\n- Use the function `np.tanh()`. It's part of the numpy library.\n- Implement using these steps:\n 1. Retrieve each parameter from the dictionary \"parameters\" (which is the output of `initialize_parameters()` by using `parameters[\"..\"]`.\n 2. Implement Forward Propagation. Compute $Z^{[1]}, A^{[1]}, Z^{[2]}$ and $A^{[2]}$ (the vector of all your predictions on all the examples in the training set).\n- Values needed in the backpropagation are stored in \"cache\". The cache will be given as an input to the backpropagation function.", "_____no_output_____" ] ], [ [ "# GRADED FUNCTION:forward_propagation\n\ndef forward_propagation(X, parameters):\n \"\"\"\n Argument:\n X -- input data of size (n_x, m)\n parameters -- python dictionary containing your parameters (output of initialization function)\n \n Returns:\n A2 -- The sigmoid output of the second activation\n cache -- a dictionary containing \"Z1\", \"A1\", \"Z2\" and \"A2\"\n \"\"\"\n # Retrieve each parameter from the dictionary \"parameters\"\n #(≈ 4 lines of code)\n # W1 = ...\n # b1 = ...\n # W2 = ...\n # b2 = ...\n # YOUR CODE STARTS HERE\n W1 = parameters['W1']\n b1 = parameters['b1']\n W2 = parameters['W2']\n b2 = parameters['b2']\n \n # YOUR CODE ENDS HERE\n \n # Implement Forward Propagation to calculate A2 (probabilities)\n # (≈ 4 lines of code)\n # Z1 = ...\n # A1 = ...\n # Z2 = ...\n # A2 = ...\n # YOUR CODE STARTS HERE\n Z1 = np.dot(W1,X) + b1\n A1 = np.tanh(Z1)\n Z2 = np.dot(W2,A1) + b2\n A2 = sigmoid(Z2)\n \n # YOUR CODE ENDS HERE\n \n assert(A2.shape == (1, X.shape[1]))\n \n cache = {\"Z1\": Z1,\n \"A1\": A1,\n \"Z2\": Z2,\n \"A2\": A2}\n \n return A2, cache", "_____no_output_____" ], [ "t_X, parameters = forward_propagation_test_case()\nA2, cache = forward_propagation(t_X, parameters)\nprint(\"A2 = \" + str(A2))\n\nforward_propagation_test(forward_propagation)", "A2 = [[0.21292656 0.21274673 0.21295976]]\n\u001b[92m All tests passed.\n" ] ], [ [ "***Expected output***\n```\nA2 = [[0.21292656 0.21274673 0.21295976]]\n```", "_____no_output_____" ], [ "<a name='4-4'></a>\n### 4.4 - Compute the Cost\n\nNow that you've computed $A^{[2]}$ (in the Python variable \"`A2`\"), which contains $a^{[2](i)}$ for all examples, you can compute the cost function as follows:\n\n$$J = - \\frac{1}{m} \\sum\\limits_{i = 1}^{m} \\large{(} \\small y^{(i)}\\log\\left(a^{[2] (i)}\\right) + (1-y^{(i)})\\log\\left(1- a^{[2] (i)}\\right) \\large{)} \\small\\tag{13}$$\n\n<a name='ex-5'></a>\n### Exercise 5 - compute_cost \n\nImplement `compute_cost()` to compute the value of the cost $J$.\n\n**Instructions**:\n- There are many ways to implement the cross-entropy loss. This is one way to implement one part of the equation without for loops:\n$- \\sum\\limits_{i=1}^{m} y^{(i)}\\log(a^{[2](i)})$:\n```python\nlogprobs = np.multiply(np.log(A2),Y)\ncost = - np.sum(logprobs) \n```\n\n- Use that to build the whole expression of the cost function.\n\n**Notes**: \n\n- You can use either `np.multiply()` and then `np.sum()` or directly `np.dot()`). \n- If you use `np.multiply` followed by `np.sum` the end result will be a type `float`, whereas if you use `np.dot`, the result will be a 2D numpy array. \n- You can use `np.squeeze()` to remove redundant dimensions (in the case of single float, this will be reduced to a zero-dimension array). \n- You can also cast the array as a type `float` using `float()`.", "_____no_output_____" ] ], [ [ "# GRADED FUNCTION: compute_cost\n\ndef compute_cost(A2, Y):\n \"\"\"\n Computes the cross-entropy cost given in equation (13)\n \n Arguments:\n A2 -- The sigmoid output of the second activation, of shape (1, number of examples)\n Y -- \"true\" labels vector of shape (1, number of examples)\n\n Returns:\n cost -- cross-entropy cost given equation (13)\n \n \"\"\"\n \n m = Y.shape[1] # number of examples\n\n # Compute the cross-entropy cost\n # (≈ 2 lines of code)\n # logprobs = ...\n # cost = ...\n # YOUR CODE STARTS HERE\n cost = (-1/m)*(np.dot(Y, np.log(A2).T) + np.dot(1-Y, np.log(1-A2).T))\n \n # YOUR CODE ENDS HERE\n \n cost = float(np.squeeze(cost)) # makes sure cost is the dimension we expect. \n # E.g., turns [[17]] into 17 \n \n return cost", "_____no_output_____" ], [ "A2, t_Y = compute_cost_test_case()\ncost = compute_cost(A2, t_Y)\nprint(\"cost = \" + str(compute_cost(A2, t_Y)))\n\ncompute_cost_test(compute_cost)", "cost = 0.6930587610394646\n\u001b[92m All tests passed.\n" ] ], [ [ "***Expected output***\n\n`cost = 0.6930587610394646`\n", "_____no_output_____" ], [ "<a name='4-5'></a>\n### 4.5 - Implement Backpropagation\n\nUsing the cache computed during forward propagation, you can now implement backward propagation.\n\n<a name='ex-6'></a>\n### Exercise 6 - backward_propagation\n\nImplement the function `backward_propagation()`.\n\n**Instructions**:\nBackpropagation is usually the hardest (most mathematical) part in deep learning. To help you, here again is the slide from the lecture on backpropagation. You'll want to use the six equations on the right of this slide, since you are building a vectorized implementation. \n\n<img src=\"images/grad_summary.png\" style=\"width:600px;height:300px;\">\n<caption><center><font color='purple'><b>Figure 1</b>: Backpropagation. Use the six equations on the right.</font></center></caption>\n\n<!--\n$\\frac{\\partial \\mathcal{J} }{ \\partial z_{2}^{(i)} } = \\frac{1}{m} (a^{[2](i)} - y^{(i)})$\n\n$\\frac{\\partial \\mathcal{J} }{ \\partial W_2 } = \\frac{\\partial \\mathcal{J} }{ \\partial z_{2}^{(i)} } a^{[1] (i) T} $\n\n$\\frac{\\partial \\mathcal{J} }{ \\partial b_2 } = \\sum_i{\\frac{\\partial \\mathcal{J} }{ \\partial z_{2}^{(i)}}}$\n\n$\\frac{\\partial \\mathcal{J} }{ \\partial z_{1}^{(i)} } = W_2^T \\frac{\\partial \\mathcal{J} }{ \\partial z_{2}^{(i)} } * ( 1 - a^{[1] (i) 2}) $\n\n$\\frac{\\partial \\mathcal{J} }{ \\partial W_1 } = \\frac{\\partial \\mathcal{J} }{ \\partial z_{1}^{(i)} } X^T $\n\n$\\frac{\\partial \\mathcal{J} _i }{ \\partial b_1 } = \\sum_i{\\frac{\\partial \\mathcal{J} }{ \\partial z_{1}^{(i)}}}$\n\n- Note that $*$ denotes elementwise multiplication.\n- The notation you will use is common in deep learning coding:\n - dW1 = $\\frac{\\partial \\mathcal{J} }{ \\partial W_1 }$\n - db1 = $\\frac{\\partial \\mathcal{J} }{ \\partial b_1 }$\n - dW2 = $\\frac{\\partial \\mathcal{J} }{ \\partial W_2 }$\n - db2 = $\\frac{\\partial \\mathcal{J} }{ \\partial b_2 }$\n \n!-->\n\n- Tips:\n - To compute dZ1 you'll need to compute $g^{[1]'}(Z^{[1]})$. Since $g^{[1]}(.)$ is the tanh activation function, if $a = g^{[1]}(z)$ then $g^{[1]'}(z) = 1-a^2$. So you can compute \n $g^{[1]'}(Z^{[1]})$ using `(1 - np.power(A1, 2))`.", "_____no_output_____" ] ], [ [ "# GRADED FUNCTION: backward_propagation\n\ndef backward_propagation(parameters, cache, X, Y):\n \"\"\"\n Implement the backward propagation using the instructions above.\n \n Arguments:\n parameters -- python dictionary containing our parameters \n cache -- a dictionary containing \"Z1\", \"A1\", \"Z2\" and \"A2\".\n X -- input data of shape (2, number of examples)\n Y -- \"true\" labels vector of shape (1, number of examples)\n \n Returns:\n grads -- python dictionary containing your gradients with respect to different parameters\n \"\"\"\n m = X.shape[1]\n \n # First, retrieve W1 and W2 from the dictionary \"parameters\".\n #(≈ 2 lines of code)\n # W1 = ...\n # W2 = ...\n # YOUR CODE STARTS HERE\n W1 = parameters['W1']\n W2 = parameters['W2']\n \n # YOUR CODE ENDS HERE\n \n # Retrieve also A1 and A2 from dictionary \"cache\".\n #(≈ 2 lines of code)\n # A1 = ...\n # A2 = ...\n # YOUR CODE STARTS HERE\n A1 = cache['A1']\n A2 = cache['A2']\n \n # YOUR CODE ENDS HERE\n \n # Backward propagation: calculate dW1, db1, dW2, db2. \n #(≈ 6 lines of code, corresponding to 6 equations on slide above)\n # dZ2 = ...\n # dW2 = ...\n # db2 = ...\n # dZ1 = ...\n # dW1 = ...\n # db1 = ...\n # YOUR CODE STARTS HERE\n dZ2 = A2 - Y\n dW2 = (1/m)*(np.dot(dZ2, A1.T))\n db2 = (1/m)*(np.sum(dZ2, axis = 1, keepdims=True))\n dZ1 = np.dot(W2.T, dZ2)*(1 - np.power(A1, 2))\n dW1 = (1/m)*(np.dot(dZ1, X.T))\n db1 = (1/m)*(np.sum(dZ1, axis = 1, keepdims=True))\n \n # YOUR CODE ENDS HERE\n \n grads = {\"dW1\": dW1,\n \"db1\": db1,\n \"dW2\": dW2,\n \"db2\": db2}\n \n return grads", "_____no_output_____" ], [ "parameters, cache, t_X, t_Y = backward_propagation_test_case()\n\ngrads = backward_propagation(parameters, cache, t_X, t_Y)\nprint (\"dW1 = \"+ str(grads[\"dW1\"]))\nprint (\"db1 = \"+ str(grads[\"db1\"]))\nprint (\"dW2 = \"+ str(grads[\"dW2\"]))\nprint (\"db2 = \"+ str(grads[\"db2\"]))\n\nbackward_propagation_test(backward_propagation)", "dW1 = [[ 0.00301023 -0.00747267]\n [ 0.00257968 -0.00641288]\n [-0.00156892 0.003893 ]\n [-0.00652037 0.01618243]]\ndb1 = [[ 0.00176201]\n [ 0.00150995]\n [-0.00091736]\n [-0.00381422]]\ndW2 = [[ 0.00078841 0.01765429 -0.00084166 -0.01022527]]\ndb2 = [[-0.16655712]]\n\u001b[92m All tests passed.\n" ] ], [ [ "***Expected output***\n```\ndW1 = [[ 0.00301023 -0.00747267]\n [ 0.00257968 -0.00641288]\n [-0.00156892 0.003893 ]\n [-0.00652037 0.01618243]]\ndb1 = [[ 0.00176201]\n [ 0.00150995]\n [-0.00091736]\n [-0.00381422]]\ndW2 = [[ 0.00078841 0.01765429 -0.00084166 -0.01022527]]\ndb2 = [[-0.16655712]]\n```", "_____no_output_____" ], [ "<a name='4-6'></a>\n### 4.6 - Update Parameters \n\n<a name='ex-7'></a>\n### Exercise 7 - update_parameters\n\nImplement the update rule. Use gradient descent. You have to use (dW1, db1, dW2, db2) in order to update (W1, b1, W2, b2).\n\n**General gradient descent rule**: $\\theta = \\theta - \\alpha \\frac{\\partial J }{ \\partial \\theta }$ where $\\alpha$ is the learning rate and $\\theta$ represents a parameter.\n\n<img src=\"images/sgd.gif\" style=\"width:400;height:400;\"> <img src=\"images/sgd_bad.gif\" style=\"width:400;height:400;\">\n<caption><center><font color='purple'><b>Figure 2</b>: The gradient descent algorithm with a good learning rate (converging) and a bad learning rate (diverging). Images courtesy of Adam Harley.</font></center></caption>\n\n**Hint**\n\n- Use `copy.deepcopy(...)` when copying lists or dictionaries that are passed as parameters to functions. It avoids input parameters being modified within the function. In some scenarios, this could be inefficient, but it is required for grading purposes.\n", "_____no_output_____" ] ], [ [ "# GRADED FUNCTION: update_parameters\n\ndef update_parameters(parameters, grads, learning_rate = 1.2):\n \"\"\"\n Updates parameters using the gradient descent update rule given above\n \n Arguments:\n parameters -- python dictionary containing your parameters \n grads -- python dictionary containing your gradients \n \n Returns:\n parameters -- python dictionary containing your updated parameters \n \"\"\"\n # Retrieve a copy of each parameter from the dictionary \"parameters\". Use copy.deepcopy(...) for W1 and W2\n #(≈ 4 lines of code)\n # W1 = ...\n # b1 = ...\n # W2 = ...\n # b2 = ...\n # YOUR CODE STARTS HERE\n W1 = parameters[\"W1\"]\n b1 = parameters[\"b1\"]\n W2 = parameters[\"W2\"]\n b2 = parameters[\"b2\"]\n \n # YOUR CODE ENDS HERE\n \n # Retrieve each gradient from the dictionary \"grads\"\n #(≈ 4 lines of code)\n # dW1 = ...\n # db1 = ...\n # dW2 = ...\n # db2 = ...\n # YOUR CODE STARTS HERE\n dW1 = grads[\"dW1\"]\n db1 = grads[\"db1\"]\n dW2 = grads[\"dW2\"]\n db2 = grads[\"db2\"]\n \n # YOUR CODE ENDS HERE\n \n # Update rule for each parameter\n #(≈ 4 lines of code)\n # W1 = ...\n # b1 = ...\n # W2 = ...\n # b2 = ...\n # YOUR CODE STARTS HERE\n W1 = W1 - learning_rate*dW1\n b1 = b1 - learning_rate*db1\n W2 = W2 - learning_rate*dW2\n b2 = b2 - learning_rate*db2\n \n # YOUR CODE ENDS HERE\n \n parameters = {\"W1\": W1,\n \"b1\": b1,\n \"W2\": W2,\n \"b2\": b2}\n \n return parameters", "_____no_output_____" ], [ "parameters, grads = update_parameters_test_case()\nparameters = update_parameters(parameters, grads)\n\nprint(\"W1 = \" + str(parameters[\"W1\"]))\nprint(\"b1 = \" + str(parameters[\"b1\"]))\nprint(\"W2 = \" + str(parameters[\"W2\"]))\nprint(\"b2 = \" + str(parameters[\"b2\"]))\n\nupdate_parameters_test(update_parameters)", "W1 = [[-0.00643025 0.01936718]\n [-0.02410458 0.03978052]\n [-0.01653973 -0.02096177]\n [ 0.01046864 -0.05990141]]\nb1 = [[-1.02420756e-06]\n [ 1.27373948e-05]\n [ 8.32996807e-07]\n [-3.20136836e-06]]\nW2 = [[-0.01041081 -0.04463285 0.01758031 0.04747113]]\nb2 = [[0.00010457]]\n\u001b[92m All tests passed.\n" ] ], [ [ "***Expected output***\n```\nW1 = [[-0.00643025 0.01936718]\n [-0.02410458 0.03978052]\n [-0.01653973 -0.02096177]\n [ 0.01046864 -0.05990141]]\nb1 = [[-1.02420756e-06]\n [ 1.27373948e-05]\n [ 8.32996807e-07]\n [-3.20136836e-06]]\nW2 = [[-0.01041081 -0.04463285 0.01758031 0.04747113]]\nb2 = [[0.00010457]]\n```", "_____no_output_____" ], [ "<a name='4-7'></a>\n### 4.7 - Integration\n\nIntegrate your functions in `nn_model()` \n\n<a name='ex-8'></a>\n### Exercise 8 - nn_model\n\nBuild your neural network model in `nn_model()`.\n\n**Instructions**: The neural network model has to use the previous functions in the right order.", "_____no_output_____" ] ], [ [ "# GRADED FUNCTION: nn_model\n\ndef nn_model(X, Y, n_h, num_iterations = 10000, print_cost=False):\n \"\"\"\n Arguments:\n X -- dataset of shape (2, number of examples)\n Y -- labels of shape (1, number of examples)\n n_h -- size of the hidden layer\n num_iterations -- Number of iterations in gradient descent loop\n print_cost -- if True, print the cost every 1000 iterations\n \n Returns:\n parameters -- parameters learnt by the model. They can then be used to predict.\n \"\"\"\n \n np.random.seed(3)\n n_x = layer_sizes(X, Y)[0]\n n_y = layer_sizes(X, Y)[2]\n \n # Initialize parameters\n #(≈ 1 line of code)\n # parameters = ...\n # YOUR CODE STARTS HERE\n parameters = initialize_parameters(n_x, n_h, n_y)\n \n # YOUR CODE ENDS HERE\n \n # Loop (gradient descent)\n\n for i in range(0, num_iterations):\n \n #(≈ 4 lines of code)\n # Forward propagation. Inputs: \"X, parameters\". Outputs: \"A2, cache\".\n # A2, cache = ...\n \n # Cost function. Inputs: \"A2, Y\". Outputs: \"cost\".\n # cost = ...\n \n # Backpropagation. Inputs: \"parameters, cache, X, Y\". Outputs: \"grads\".\n # grads = ...\n \n # Gradient descent parameter update. Inputs: \"parameters, grads\". Outputs: \"parameters\".\n # parameters = ...\n \n # YOUR CODE STARTS HERE\n A2, cache = forward_propagation(X, parameters)\n cost = compute_cost(A2, Y)\n grads = backward_propagation(parameters, cache, X, Y)\n parameters = update_parameters(parameters, grads)\n # YOUR CODE ENDS HERE\n \n # Print the cost every 1000 iterations\n if print_cost and i % 1000 == 0:\n print (\"Cost after iteration %i: %f\" %(i, cost))\n\n return parameters", "_____no_output_____" ], [ "t_X, t_Y = nn_model_test_case()\nparameters = nn_model(t_X, t_Y, 4, num_iterations=10000, print_cost=True)\n\nprint(\"W1 = \" + str(parameters[\"W1\"]))\nprint(\"b1 = \" + str(parameters[\"b1\"]))\nprint(\"W2 = \" + str(parameters[\"W2\"]))\nprint(\"b2 = \" + str(parameters[\"b2\"]))\n\nnn_model_test(nn_model)", "(1, 3)\n(1, 3)\nCost after iteration 0: 0.692739\nCost after iteration 1000: 0.000218\nCost after iteration 2000: 0.000107\nCost after iteration 3000: 0.000071\nCost after iteration 4000: 0.000053\nCost after iteration 5000: 0.000042\nCost after iteration 6000: 0.000035\nCost after iteration 7000: 0.000030\nCost after iteration 8000: 0.000026\nCost after iteration 9000: 0.000023\nW1 = [[-0.65848169 1.21866811]\n [-0.76204273 1.39377573]\n [ 0.5792005 -1.10397703]\n [ 0.76773391 -1.41477129]]\nb1 = [[ 0.287592 ]\n [ 0.3511264 ]\n [-0.2431246 ]\n [-0.35772805]]\nW2 = [[-2.45566237 -3.27042274 2.00784958 3.36773273]]\nb2 = [[0.20459656]]\n(1, 3)\n(1, 3)\n(1, 3)\n(1, 3)\n(1, 3)\n(1, 3)\n\u001b[92m All tests passed.\n" ] ], [ [ "***Expected output***\n```\nCost after iteration 0: 0.692739\nCost after iteration 1000: 0.000218\nCost after iteration 2000: 0.000107\n...\nCost after iteration 8000: 0.000026\nCost after iteration 9000: 0.000023\nW1 = [[-0.65848169 1.21866811]\n [-0.76204273 1.39377573]\n [ 0.5792005 -1.10397703]\n [ 0.76773391 -1.41477129]]\nb1 = [[ 0.287592 ]\n [ 0.3511264 ]\n [-0.2431246 ]\n [-0.35772805]]\nW2 = [[-2.45566237 -3.27042274 2.00784958 3.36773273]]\nb2 = [[0.20459656]]\n```", "_____no_output_____" ], [ "<a name='5'></a>\n## 5 - Test the Model\n\n<a name='5-1'></a>\n### 5.1 - Predict\n\n<a name='ex-9'></a>\n### Exercise 9 - predict\n\nPredict with your model by building `predict()`.\nUse forward propagation to predict results.\n\n**Reminder**: predictions = $y_{prediction} = \\mathbb 1 \\text{{activation > 0.5}} = \\begin{cases}\n 1 & \\text{if}\\ activation > 0.5 \\\\\n 0 & \\text{otherwise}\n \\end{cases}$ \n \nAs an example, if you would like to set the entries of a matrix X to 0 and 1 based on a threshold you would do: ```X_new = (X > threshold)```", "_____no_output_____" ] ], [ [ "# GRADED FUNCTION: predict\n\ndef predict(parameters, X):\n \"\"\"\n Using the learned parameters, predicts a class for each example in X\n \n Arguments:\n parameters -- python dictionary containing your parameters \n X -- input data of size (n_x, m)\n \n Returns\n predictions -- vector of predictions of our model (red: 0 / blue: 1)\n \"\"\"\n \n # Computes probabilities using forward propagation, and classifies to 0/1 using 0.5 as the threshold.\n #(≈ 2 lines of code)\n # A2, cache = ...\n # predictions = ...\n # YOUR CODE STARTS HERE\n A2, cache = forward_propagation(X, parameters)\n predictions = (A2 > 0.5)\n # YOUR CODE ENDS HERE\n \n return predictions", "_____no_output_____" ], [ "parameters, t_X = predict_test_case()\n\npredictions = predict(parameters, t_X)\nprint(\"Predictions: \" + str(predictions))\n\npredict_test(predict)", "Predictions: [[ True False True]]\n\u001b[92m All tests passed.\n" ] ], [ [ "***Expected output***\n```\nPredictions: [[ True False True]]\n```", "_____no_output_____" ], [ "<a name='5-2'></a>\n### 5.2 - Test the Model on the Planar Dataset\n\nIt's time to run the model and see how it performs on a planar dataset. Run the following code to test your model with a single hidden layer of $n_h$ hidden units!", "_____no_output_____" ] ], [ [ "# Build a model with a n_h-dimensional hidden layer\nparameters = nn_model(X, Y, n_h = 4, num_iterations = 10000, print_cost=True)\n\n# Plot the decision boundary\nplot_decision_boundary(lambda x: predict(parameters, x.T), X, Y)\nplt.title(\"Decision Boundary for hidden layer size \" + str(4))", "(1, 400)\n(1, 400)\nCost after iteration 0: 0.693048\nCost after iteration 1000: 0.288083\nCost after iteration 2000: 0.254385\nCost after iteration 3000: 0.233864\nCost after iteration 4000: 0.226792\nCost after iteration 5000: 0.222644\nCost after iteration 6000: 0.219731\nCost after iteration 7000: 0.217504\nCost after iteration 8000: 0.219430\nCost after iteration 9000: 0.218551\n" ], [ "# Print accuracy\npredictions = predict(parameters, X)\nprint ('Accuracy: %d' % float((np.dot(Y, predictions.T) + np.dot(1 - Y, 1 - predictions.T)) / float(Y.size) * 100) + '%')", "Accuracy: 90%\n" ] ], [ [ "**Expected Output**: \n\n<table style=\"width:30%\">\n <tr>\n <td><b>Accuracy</b></td>\n <td> 90% </td> \n </tr>\n</table>", "_____no_output_____" ], [ "Accuracy is really high compared to Logistic Regression. The model has learned the patterns of the flower's petals! Unlike logistic regression, neural networks are able to learn even highly non-linear decision boundaries. ", "_____no_output_____" ], [ "### Congrats on finishing this Programming Assignment! \n\nHere's a quick recap of all you just accomplished: \n\n- Built a complete 2-class classification neural network with a hidden layer\n- Made good use of a non-linear unit\n- Computed the cross entropy loss\n- Implemented forward and backward propagation\n- Seen the impact of varying the hidden layer size, including overfitting.\n\nYou've created a neural network that can learn patterns! Excellent work. Below, there are some optional exercises to try out some other hidden layer sizes, and other datasets. ", "_____no_output_____" ], [ "<a name='6'></a>\n## 6 - Tuning hidden layer size (optional/ungraded exercise)\n\nRun the following code(it may take 1-2 minutes). Then, observe different behaviors of the model for various hidden layer sizes.", "_____no_output_____" ] ], [ [ "# This may take about 2 minutes to run\n\nplt.figure(figsize=(16, 32))\nhidden_layer_sizes = [1, 2, 3, 4, 5, 20, 50]\nfor i, n_h in enumerate(hidden_layer_sizes):\n plt.subplot(5, 2, i+1)\n plt.title('Hidden Layer of size %d' % n_h)\n parameters = nn_model(X, Y, n_h, num_iterations = 5000)\n plot_decision_boundary(lambda x: predict(parameters, x.T), X, Y)\n predictions = predict(parameters, X)\n accuracy = float((np.dot(Y,predictions.T) + np.dot(1 - Y, 1 - predictions.T)) / float(Y.size)*100)\n print (\"Accuracy for {} hidden units: {} %\".format(n_h, accuracy))", "(1, 400)\n(1, 400)\nAccuracy for 1 hidden units: 67.5 %\n(1, 400)\n(1, 400)\nAccuracy for 2 hidden units: 67.25 %\n(1, 400)\n(1, 400)\nAccuracy for 3 hidden units: 90.75 %\n(1, 400)\n(1, 400)\nAccuracy for 4 hidden units: 90.5 %\n(1, 400)\n(1, 400)\nAccuracy for 5 hidden units: 91.25 %\n(1, 400)\n(1, 400)\nAccuracy for 20 hidden units: 90.0 %\n(1, 400)\n(1, 400)\nAccuracy for 50 hidden units: 90.25 %\n" ] ], [ [ "**Interpretation**:\n- The larger models (with more hidden units) are able to fit the training set better, until eventually the largest models overfit the data. \n- The best hidden layer size seems to be around n_h = 5. Indeed, a value around here seems to fits the data well without also incurring noticeable overfitting.\n- Later, you'll become familiar with regularization, which lets you use very large models (such as n_h = 50) without much overfitting. ", "_____no_output_____" ], [ "**Note**: Remember to submit the assignment by clicking the blue \"Submit Assignment\" button at the upper-right. \n\n**Some optional/ungraded questions that you can explore if you wish**: \n- What happens when you change the tanh activation for a sigmoid activation or a ReLU activation?\n- Play with the learning_rate. What happens?\n- What if we change the dataset? (See part 5 below!)", "_____no_output_____" ], [ "<a name='7'></a>\n## 7- Performance on other datasets", "_____no_output_____" ], [ "If you want, you can rerun the whole notebook (minus the dataset part) for each of the following datasets.", "_____no_output_____" ] ], [ [ "# Datasets\nnoisy_circles, noisy_moons, blobs, gaussian_quantiles, no_structure = load_extra_datasets()\n\ndatasets = {\"noisy_circles\": noisy_circles,\n \"noisy_moons\": noisy_moons,\n \"blobs\": blobs,\n \"gaussian_quantiles\": gaussian_quantiles}\n\n### START CODE HERE ### (choose your dataset)\ndataset = \"noisy_moons\"\n### END CODE HERE ###\n\nX, Y = datasets[dataset]\nX, Y = X.T, Y.reshape(1, Y.shape[0])\n\n# make blobs binary\nif dataset == \"blobs\":\n Y = Y%2\n\n# Visualize the data\nplt.scatter(X[0, :], X[1, :], c=Y, s=40, cmap=plt.cm.Spectral);", "_____no_output_____" ] ], [ [ "**References**:\n\n- http://scs.ryerson.ca/~aharley/neural-networks/\n- http://cs231n.github.io/neural-networks-case-study/", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown" ]

[ [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown", "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown", "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import numpy as np\nimport datetime\n\n\nimport matplotlib.pyplot as plt\nfrom PIL import Image\n\nfrom scipy.sparse import csr_matrix\nimport matplotlib.pyplot as plt\n\nfrom sklearn.cluster import KMeans\n\nfrom numpy.linalg import norm\nfrom sklearn.feature_extraction import image\nimport warnings\nwarnings.filterwarnings(\"ignore\")\n", "_____no_output_____" ], [ "def get_B_and_weight_vec(n_nodes,threshold,sigma=1):\n '''\n Generate graph structure from the image to be segmented.\n Inputs:\n n_nodes: number of nodes, i.e. number of pixels\n threshold: threshold to drop edges with small weights (weak similarities)\n sigma: parameter to scale edge weights\n Outputs:\n B: Incidence matrix\n Weight_vec: edge_wise weights\n '''\n N = n_nodes\n row = []\n col = []\n data = []\n weight_vec = []\n cnt = 0\n# \n for i in range(N):\n for j in [i+1,i+100]:\n if j>=2900:\n continue\n if np.exp(-norm(img[i]-img[j])/(2*sigma**2)) > threshold:\n row.append(cnt)\n col.append(i)\n data.append(1)\n\n row.append(cnt)\n col.append(j)\n data.append(-1)\n cnt += 1\n weight_vec.append(np.exp(-norm(img[i]-img[j])/(2*sigma**2)))\n\n B = csr_matrix((data, (row, col)), shape=(cnt, N))\n weight_vec = np.array(weight_vec)\n return B, weight_vec", "_____no_output_____" ], [ "def algorithm(B, weight_vec, seeds,K=15000,alpha=0.02, lambda_nLasso=None, check_s=False):\n E, N = B.shape\n# weight_vec = np.ones(E)\n\n Gamma_vec = np.array(1./(np.sum(abs(B), 0)))[0] # \\in [0, 1]\n Gamma = np.diag(Gamma_vec)\n\n Sigma = 0.5\n\n seednodesindicator= np.zeros(N)\n seednodesindicator[seeds] = 1 \n\n\n \n noseednodeindicator = np.ones(N)\n noseednodeindicator[seeds] = 0\n \n if lambda_nLasso == None:\n lambda_nLasso = 2 / math.sqrt(np.sum(weight_vec))\n \n if check_s:\n s = 0.0\n for item in range(len(weight_vec)):\n x = B[item].toarray()[0]\n i = np.where(x == -1)[0][0]\n j = np.where(x == 1)[0][0]\n if i < N1 <= j:\n s += weight_vec[item]\n elif i >= N1 > j:\n s += weight_vec[item]\n\n if lambda_nLasso * s >= alpha * N2 / 2:\n print ('eq(24)', lambda_nLasso * s, alpha * N2 / 2)\n \n fac_alpha = 1./(Gamma_vec*alpha+1) # \\in [0, 1]\n\n hatx = np.zeros(N)\n newx = np.zeros(N)\n prevx = np.zeros(N)\n haty = np.array([x/(E-1) for x in range(0, E)])\n history = []\n for iterk in range(K):\n # if 0 < np.max(abs(newx - prevx)) < 1e-4:\n # print(iterk)\n # break\n tildex = 2 * hatx - prevx\n newy = haty + Sigma * B.dot(tildex) # chould be negative\n haty = newy / np.maximum(abs(newy) / (lambda_nLasso * weight_vec), np.ones(E)) # could be negative\n\n newx = hatx - Gamma_vec * B.T.dot(haty) # could be negative\n newx[seeds] = (newx[seeds] + Gamma_vec[seeds]) / (1 + Gamma_vec[seeds])\n\n newx = seednodesindicator * newx + noseednodeindicator * (newx * fac_alpha)\n prevx = np.copy(hatx)\n hatx = newx # could be negative\n history.append(newx)\n \n history = np.array(history)\n\n return history\n ", "_____no_output_____" ], [ "#load the image\nimg=Image.open(\"stripes.png\")", "_____no_output_____" ] ], [ [ "# Preprocess the image", "_____no_output_____" ] ], [ [ "# resize the image\nbasewidth = 100\nwpercent = (basewidth / float(img.size[0]))\nhsize = int((float(img.size[1]) * float(wpercent)))\nimg = img.resize((basewidth, hsize), Image.ANTIALIAS)\nimg = np.array(img)[:,:,:3]\nprint(img.shape)\nplt.imshow(img)", "(29, 100, 3)\n" ], [ "img = img.reshape(-1,3)\nimg.shape", "_____no_output_____" ] ], [ [ "# Perform the segmentation task via Kmeans", "_____no_output_____" ] ], [ [ "kmeans = KMeans(n_clusters=2).fit(img)\nplt.imshow(kmeans.labels_.reshape(29,100))", "_____no_output_____" ] ], [ [ "# Perform the task via our algorithm", "_____no_output_____" ] ], [ [ "# generate graph from image\nimg = img.reshape(-1,3)/255\nn_nodes=img.shape[0]\nprint(\"number of nodes:\",n_nodes )\n\nB,weight=get_B_and_weight_vec(n_nodes,0.2,1)\n# plt.hist(weight,bins=30) #distribution of similarity measure", "number of nodes: 2900\n" ], [ "def run_seg(n_nodes,seeds,threshold, K=30, alpha=0.1, lambda_nLasso=0.1):\n B, weight_vec = get_B_and_weight_vec(n_nodes,threshold)\n \n start = datetime.datetime.now()\n history = algorithm(B, weight_vec, seeds=seeds, K=K, alpha=alpha, lambda_nLasso=lambda_nLasso)\n print('our method time: ', datetime.datetime.now() - start) \n return history", "_____no_output_____" ], [ "# generate seeds according to the labels assigned by kmeans\nseeds = np.random.choice(np.where(kmeans.labels_==0)[0],20)\n\n# run our algorithm and visulize the result before feed it to kmeans\nhistory = run_seg(n_nodes=n_nodes,seeds=seeds,threshold = 0.95, K=1000,alpha=0.01, lambda_nLasso=1)\nplt.imshow(history[-1].reshape(29,100))", "our method time: 0:00:00.431303\n" ], [ "# Feed the node signal from our algorithm to kmeans to complete clustering (2 clusters)\nhistory=np.nan_to_num(history)\nkmeans = KMeans(n_clusters=2).fit(history[-1].reshape(len(history[-1]), 1))\n\n#visulize the segmentation result\nsegmented = kmeans.labels_\nplt.imshow(segmented.reshape((29,100)))", "_____no_output_____" ] ], [ [ "# Perform the segmentation task via spectral clustering", "_____no_output_____" ] ], [ [ "from sklearn.cluster import SpectralClustering\n\ns=SpectralClustering(2).fit(img)\n\nplt.imshow(s.labels_.reshape(29,100))", "_____no_output_____" ], [ "# Python3 Program to print BFS traversal\n# from a given source vertex. BFS(int s)\n# traverses vertices reachable from s.\nfrom collections import defaultdict\n\n# This class represents a directed graph\n# using adjacency list representation\nclass Graph:\n\n\t# Constructor\n\tdef __init__(self):\n\n\t\t# default dictionary to store graph\n\t\tself.graph = defaultdict(list)\n\n\t# function to add an edge to graph\n\tdef addEdge(self,u,v):\n\t\tself.graph[u].append(v)\n\n\t# Function to print a BFS of graph\n\tdef BFS(self, s):\n\n\t\t# Mark all the vertices as not visited\n\t\tvisited = [False] * (max(self.graph) + 1)\n\n\t\t# Create a queue for BFS\n\t\tqueue = []\n\n\t\t# Mark the source node as\n\t\t# visited and enqueue it\n\t\tqueue.append(s)\n\t\tvisited[s] = True\n\n\t\twhile queue:\n\n\t\t\t# Dequeue a vertex from\n\t\t\t# queue and print it\n\t\t\ts = queue.pop(0)\n\t\t\tprint (s, end = \" \")\n\n\t\t\t# Get all adjacent vertices of the\n\t\t\t# dequeued vertex s. If a adjacent\n\t\t\t# has not been visited, then mark it\n\t\t\t# visited and enqueue it\n\t\t\tfor i in self.graph[s]:\n\t\t\t\tif visited[i] == False:\n\t\t\t\t\tqueue.append(i)\n\t\t\t\t\tvisited[i] = True\n\n# Driver code\n\n# Create a graph given in\n# the above diagram\ng = Graph()\ng.addEdge(0, 1)\ng.addEdge(0, 2)\ng.addEdge(1, 2)\ng.addEdge(2, 0)\ng.addEdge(2, 3)\ng.addEdge(3, 3)\n\nprint (\"Following is Breadth First Traversal\"\n\t\t\t\t\" (starting from vertex 2)\")\ng.BFS(2)\n\n# This code is contributed by Neelam Yadav\n\n", "Following is Breadth First Traversal (starting from vertex 2)\n2 0 3 1 " ] ] ]

[ "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import cv2\n\ncapture = cv2.VideoCapture('../data/drop.avi')\nframe_count = capture.get(cv2.CAP_PROP_FRAME_COUNT)\nprint('Frame count:', frame_count)\n\nprint('Position:', capture.get(cv2.CAP_PROP_POS_FRAMES))\n_, frame = capture.read()\ncv2.imshow('frame0', frame)\n\nprint('Position:', capture.get(cv2.CAP_PROP_POS_FRAMES))\n_, frame = capture.read()\ncv2.imshow('frame1', frame)\n\ncapture.set(cv2.CAP_PROP_POS_FRAMES, 100)\nprint('Position:', capture.get(cv2.CAP_PROP_POS_FRAMES))\n_, frame = capture.read()\ncv2.imshow('frame100', frame)\n \ncv2.waitKey()\ncv2.destroyAllWindows()", "Frame count: 182.0\nPosition: 0.0\nPosition: 1.0\nPosition: 100.0\n" ] ] ]

[ "code" ]

[ [ "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Solution Graded Exercise 1: Leaky-integrate-and-fire model", "_____no_output_____" ], [ "first name: Eve\n\nlast name: Rahbe\n\nsciper: 235549\n\ndate: 21.03.2018\n\n*Your teammate*\n\nfirst name of your teammate: Antoine\n\nlast name of your teammate: Alleon\n\nsciper of your teammate: 223333\n\n\nNote: You are allowed to discuss the concepts with your class mates. You are not allowed to share code. You have to understand every line of code you write in this notebook. We will ask you questions about your submission during a fraud detection session during the last week of the semester.\n\nIf you are asked for plots: The appearance of the plots (labelled axes, useful scaling etc.) is important!\n\nIf you are asked for discussions: Answer in a precise way and try to be concise. \n\n\n** Submission **\n\nRename this notebook to Ex2_FirstName_LastName_Sciper.ipynb and upload that single file on moodle before the deadline.\n\n** Link to the exercise **\n\nhttp://neuronaldynamics-exercises.readthedocs.io/en/stable/exercises/leaky-integrate-and-fire.html", "_____no_output_____" ], [ "# Exercise 2, getting started", "_____no_output_____" ] ], [ [ "%matplotlib inline\nimport brian2 as b2\nimport matplotlib.pyplot as plt\nimport numpy as np\nfrom neurodynex.leaky_integrate_and_fire import LIF\nfrom neurodynex.tools import input_factory, plot_tools\n\n\nLIF.getting_started()\nLIF.print_default_parameters()", "nr of spikes: 0\n" ] ], [ [ "# 2.1 Exercise: minimal current\n## 2.1.1. Question: minimal current (calculation) \n#### [2 points]", "_____no_output_____" ] ], [ [ "from neurodynex.leaky_integrate_and_fire import LIF\nprint(\"resting potential: {}\".format(LIF.V_REST))\ni_min = (LIF.FIRING_THRESHOLD-LIF.V_REST)/LIF.MEMBRANE_RESISTANCE\nprint(\"minimal current i_min: {}\".format(i_min))", "resting potential: -0.07\nminimal current i_min: 2e-09\n" ] ], [ [ "The minimal current is :\n\n$i_{min} = \\frac{\\theta-u_{rest}}{R} = \\frac{-50-(-70) [mV]}{10 [Mohm]} = 2 [nA]$\n\n$\\theta$ is the firing threshold\n\n$u_{rest}$ is the resting potential\n\n$R$ is the membrane resistance", "_____no_output_____" ], [ "## 2.1.2. Question: minimal current (simulation)\n#### [2 points]", "_____no_output_____" ] ], [ [ "# create a step current with amplitude= i_min\nstep_current = input_factory.get_step_current(\n t_start=5, t_end=100, unit_time=b2.ms,\n amplitude= i_min) # set i_min to your value\n\n# run the LIF model.\n# Note: As we do not specify any model parameters, the simulation runs with the default values\n(state_monitor,spike_monitor) = LIF.simulate_LIF_neuron(input_current=step_current, simulation_time = 100 * b2.ms)\n\n# plot I and vm\nplot_tools.plot_voltage_and_current_traces(\nstate_monitor, step_current, title=\"min input\", firing_threshold=LIF.FIRING_THRESHOLD)\nprint(\"nr of spikes: {}\".format(spike_monitor.count[0])) # should be 0", "nr of spikes: 0\n" ] ], [ [ "# 2.2. Exercise: f-I Curve \n## 2.2.1. Question: f-I Curve and refractoryness\n", "_____no_output_____" ], [ "1 - Sketch or plot the curve with some program. You don't have to include it here, it is just for your understanding and will not be graded.\n\n2 - What is the maximum rate at which this neuron can fire?\n#### [3 points]", "_____no_output_____" ] ], [ [ "# create a step current with amplitude i_max\ni_max = 125 * b2.namp\nstep_current = input_factory.get_step_current(\n t_start=5, t_end=100, unit_time=b2.ms,\n amplitude=i_max)\n\n# run the LIF model and set the absolute refractory period to 3ms\n(state_monitor,spike_monitor) = LIF.simulate_LIF_neuron(input_current=step_current, simulation_time = 500 * b2.ms, \n abs_refractory_period=3 * b2.ms)\n\n# number of spikes\nprint(\"nr of spikes: {}\".format(spike_monitor.count[0]))\n\n# firing frequency\nT = 95e-03/spike_monitor.count[0]\nprint(\"T : {}\".format(T))\nprint(\"firing frequency : {}\".format(1/T))", "nr of spikes: 32\nT : 0.00296875\nfiring frequency : 336.842105263\n" ] ], [ [ "The maximum rate at which this neuron can fire is $f = 336.84 [Hz]$. \n", "_____no_output_____" ], [ "3 - Inject currents of different amplitudes (from 0nA to 100nA) into a LIF neuron. \nFor each current, run the simulation for 500ms and determine the firing frequency in Hz. Then plot the f-I curve. \n#### [4 points]", "_____no_output_____" ] ], [ [ "import numpy as np\nimport matplotlib.pyplot as plt\n\nfiring_frequency = []\n\n# create a step current with amplitude i from 0 to 100 nA\nfor i in range(0,100,1) :\n step_current = input_factory.get_step_current(\n t_start=5, t_end=100, unit_time=b2.ms,\n amplitude=i * b2.namp) # stock amplitude i from 0 to 100 nA\n \n # run the LIF model\n (state_monitor,spike_monitor) = LIF.simulate_LIF_neuron(input_current=step_current, simulation_time = 500 * b2.ms, \n abs_refractory_period= 3 * b2.ms)\n\n if (spike_monitor.count[0] == 0) :\n firing_frequency.append(0)\n else :\n # firing frequency\n T = 95e-03/spike_monitor.count[0]\n firing_frequency.append(1/T)", "_____no_output_____" ], [ "plt.xlabel(\"step current amplitude [nA]\")\nplt.ylabel(\"firing frequency [Hz]\")\nplt.title(\"f-I curve for step current\")\nplt.plot(range(0,100,1), firing_frequency)", "_____no_output_____" ] ], [ [ "# 2.3. Exercise: “Experimentally” estimate the parameters of a LIF neuron\n## 2.3.1. Question: “Read” the LIF parameters out of the vm plot\n#### [6 points]", "_____no_output_____" ], [ "My estimates for the parameters :\n\\begin{itemize}\n\\item Resting potential : $u_{rest} = -66 [mV]$.\n\\item Reset potential : $u_{reset} = -63 [mV]$ is the membrane potential between spikes.\n\\item Firing threshold : $\\theta = -38 [mV]$ using $\\theta = u(t_{inf})$ with step current amplitude $i_{min}$.\n\\item Membrane resitance : $R = \\frac{\\theta-u_{rest}}{i_{min}} = 12.7 [Mohm]$ with $i_{min}=2.2[nA]$ at the firing threshold.\n\\item Membrane time scale : $\\tau = 11.5 [ms]$ time to reach $63\\%$ of $\\theta$ with step curent amplitude $i_{min}$.\n\\item Absolute refractory period : $ t = 5 [ms]$ time before new spike after reset of the potential.\n\\end{itemize} \n", "_____no_output_____" ] ], [ [ "# get a random parameter. provide a random seed to have a reproducible experiment\nrandom_parameters = LIF.get_random_param_set(random_seed=432)\n\n# define your test current\ntest_current = input_factory.get_step_current(\n t_start=5, t_end=100, unit_time=b2.ms, amplitude= 10 * b2.namp)\n\n# probe the neuron. pass the test current AND the random params to the function\nstate_monitor, spike_monitor = LIF.simulate_random_neuron(test_current, random_parameters)\n\n# plot\nplot_tools.plot_voltage_and_current_traces(state_monitor, test_current, title=\"experiment\")\n\n# print the parameters to the console and compare with your estimates\nLIF.print_obfuscated_parameters(random_parameters)", "Resting potential: -0.066\nReset voltage: -0.063\nFiring threshold: -0.038\nMembrane resistance: 13000000.0\nMembrane time-scale: 0.013\nAbsolute refractory period: 0.005\n" ] ], [ [ "# 2.4. Exercise: Sinusoidal input current and subthreshold response\n## 2.4.1. Question\n#### [5 points]", "_____no_output_____" ] ], [ [ "# note the higher resolution when discretizing the sine wave: we specify unit_time=0.1 * b2.ms\nsinusoidal_current = input_factory.get_sinusoidal_current(200, 1000, unit_time=0.1 * b2.ms,\n amplitude= 2.5 * b2.namp, frequency=250*b2.Hz,\n direct_current=0. * b2.namp)\n\n# run the LIF model. By setting the firing threshold to to a high value, we make sure to stay in the linear (non spiking) regime.\n(state_monitor, spike_monitor) = LIF.simulate_LIF_neuron(input_current=sinusoidal_current, simulation_time = 120 * b2.ms, \n firing_threshold=0*b2.mV)\n\n# plot the membrane voltage\nplot_tools.plot_voltage_and_current_traces(state_monitor, sinusoidal_current, title = \"Sinusoidal input current\")\nprint(\"nr of spikes: {}\".format(spike_monitor.count[0]))\n\n# Calculate the amplitude of the membrane voltage\n# get the difference betwwen min value of the voltage and the resting potential\nprint(\"Amplitude of the membrane voltage : {} V\" .format(abs(np.min(np.asarray(state_monitor.v))-(-0.07))))\n\nimport scipy.signal\n\n# Calculate the phase of the membrane voltage\n# interpolation of the signals\nxx = np.interp(np.linspace(1,1002,1002),np.linspace(1,1200,1200),np.transpose(np.asarray(state_monitor.v))[:,0])\n# correlation\ncorr = scipy.signal.correlate(xx,sinusoidal_current.values[:,0])\n\ndt = np.arange(-1001,1002)\n# find the max correlation\nrecovered_time_shift = dt[corr.argmax()]\n# convert timeshift in phase between 0 and 2pi\nperiod = 1/0.250\nrecovered_phase_shift = np.pi*(((0.5 + recovered_time_shift/period) %1.0)-0.5)\nprint(\"Phase shift : {}\" .format(recovered_phase_shift))", "nr of spikes: 0\nAmplitude of the membrane voltage : 0.001985117111761442 V\nPhase shift : -1.5707963267948966\n" ] ], [ [ "The results are : \n\n$ A = 2 [mV]$ (computationnally and visually) and $phase = -\\pi/2$ (computationnally and visually).", "_____no_output_____" ], [ "## 2.4.2. Question\n#### [5 points]", "_____no_output_____" ] ], [ [ "# For input frequencies between 10Hz and 1kHz plot the resulting amplitude of subthreshold oscillations of the \n# membrane potential vs. input frequency.\n\namplitude = []\nfor i in range(15) :\n\n sinusoidal_current = input_factory.get_sinusoidal_current(200, 1000, unit_time=0.1 * b2.ms,\n amplitude= 2.5 * b2.namp, frequency=10.**(1.+i/7.)*b2.Hz,\n direct_current=0. * b2.namp)\n\n # run the LIF model.\n (state_monitor, spike_monitor) = LIF.simulate_LIF_neuron(input_current=sinusoidal_current, simulation_time = 120 * b2.ms, \n firing_threshold=0*b2.mV)\n\n amplitude.append(abs(np.min(np.asarray(state_monitor.v))-(-0.07)))", "_____no_output_____" ], [ "plt.xlabel(\"sinusoidal current frequency [Hz]\")\nplt.ylabel(\"amplitude of the membrane potential [mV]\")\nplt.title(\"Amplitude vs Input frequency\")\nplt.plot([10.**(1.+i/7.) for i in range(15)], amplitude)", "_____no_output_____" ] ], [ [ "## 2.4.3. Question\n#### [5 points]", "_____no_output_____" ] ], [ [ "# For input frequencies between 10Hz and 1kHz\n# plot the resulting phase shift of subthreshold oscillations of the membrane potential vs. input frequency.\nphase = []\nfor f in [10,50,100,250,500,750,1000] :\n \n sinusoidal_current = input_factory.get_sinusoidal_current(200, 1000, unit_time=0.1 * b2.ms,\n amplitude= 2.5 * b2.namp, frequency = f*b2.Hz,\n direct_current=0. * b2.namp)\n\n # run the LIF model.\n (state_monitor, spike_monitor) = LIF.simulate_LIF_neuron(input_current=sinusoidal_current, simulation_time = 120 * b2.ms, \n firing_threshold=0*b2.mV)\n\n xx = np.interp(np.linspace(1,1002,1002),np.linspace(1,1200,1200),np.transpose(np.asarray(state_monitor.v))[:,0])\n corr = scipy.signal.correlate(xx,sinusoidal_current.values[:,0])\n\n dt = np.arange(-1001,1002)\n recovered_time_shift = dt[corr.argmax()]\n period = 1000/f\n recovered_phase_shift = np.pi*(((0.5 + recovered_time_shift/period) %1.0)-0.5)\n \n phase.append(recovered_phase_shift)", "_____no_output_____" ], [ "plt.xlabel(\"sinusoidal current frequency [Hz]\")\nplt.ylabel(\"phase shift between membrane potential and input current\")\nplt.title(\"Phase shift of membrane potential vs Input frequency\")\nplt.plot([10,50,100,250,500,750,1000], phase)", "_____no_output_____" ] ], [ [ "## 2.4.4. Question\n#### [3 points]", "_____no_output_____" ], [ "It is a \\textbf{Low-Pass} filter because it amplifies low frequencies and attenuates high frequencies.", "_____no_output_____" ], [ "# 2.5 Leaky integrate-and-fire neuron with noisy input\nThis exercise is not available online. All information is given here.\nSo far you have explored the leaky integrate-and-fire model with step and sinusoidal input currents. We will now investigate the same neuron model with noisy input.\nThe voltage equation now is:\n\\begin{eqnarray}\n\\tau \\frac{du}{dt} = -u(t) + u_{rest} + RI(t) + RI_{noise}(t)\n\\end{eqnarray}\nwhere the noise is simply an additional term.\n\nTo implement the noise term in the above equation we will consider it as 'white noise', $I_{noise}(t) = \\sigma \\xi(t)$. White noise $\\xi$ is a stochastic process with expectation value $<\\xi(t)>=0$ and autocorrelation $<\\xi(t)\\xi(t+\\Delta)>=\\delta(\\Delta)$ Note that, as we saw in the Exercise set of Week 1, the $\\delta$-function has units of $1/time$, so $\\xi$ has units of $1/\\sqrt{time}$.\n\nIt can be shown that the discrete time implementation of a noisy voltage trajectory is:\n\\begin{eqnarray}\ndu = (-u(t) + u_{rest} + RI(t))\\frac{dt}{\\tau} + \\frac{R}{\\tau}\\sigma \\sqrt{dt}\\ y,\n\\end{eqnarray}\nwhere $y \\sim \\mathcal{N}(0, 1)$\n\nWe can then write, again for implementational purposes:\n\\begin{eqnarray}\ndu = \\big[-u(t) + u_{rest} + R(I(t) + \\sigma \\frac{1}{\\sqrt{dt}} y) \\big]\\frac{dt}{\\tau}\n\\end{eqnarray}\n\n\nNote that for the physical units to be consistent $\\sigma$ in our formulation has units of $current * \\sqrt{time}$. \n\nDetails of the above are beyond the scope of this exercise. If you would like to get more insights we refer to the paragraph 8.1 of the book (http://neuronaldynamics.epfl.ch/online/Ch8.S1.html), to http://www.scholarpedia.org/article/Stochastic_dynamical_systems#Ornstein-Uhlenbeck_process and regarding the implementational scaling of the noise to http://brian2.readthedocs.io/en/stable/user/models.html#time-scaling-of-noise.", "_____no_output_____" ], [ "### 2.5.1 Noisy step input current\n\n#### [7 points]\n1 - Implement the noisy current $I_0 + I_{noise}$ as described above. In order to do this edit the function get_noisy_step_current provided below. This is simply a copy of the code of the function get_step_current that you used earlier, and you just need to add the noisy part of the current at the indicated line (indicated by \"???\").\n\nThen create a noisy step current with amplitude $I_0 = 1.5nA$ and $\\sigma = 1 nA* \\sqrt{\\text{your time unit}}$ (e.g.: time_unit = 1 ms), run the LIF model and plot the input current and the membrane potential, as you did in the previous exercises.", "_____no_output_____" ] ], [ [ "def get_noisy_step_current(t_start, t_end, unit_time, amplitude, sigma, append_zero=True):\n\n \"\"\"Creates a step current with added noise. If t_start == t_end, then a single\n entry in the values array is set to amplitude.\n\n Args:\n t_start (int): start of the step\n t_end (int): end of the step\n unit_time (Quantity, Time): unit of t_start and t_end. e.g. 0.1*brian2.ms\n amplitude (Quantity): amplitude of the step. e.g. 3.5*brian2.uamp\n sigma (float): amplitude (std) of the noise. e.g. 0.1*b2.uamp\n append_zero (bool, optional): if true, 0Amp is appended at t_end+1.\n Without that trailing 0, Brian reads out the last value in the array (=amplitude) for all indices > t_end.\n\n Returns:\n TimedArray: Brian2.TimedArray\n \"\"\"\n\n assert isinstance(t_start, int), \"t_start_ms must be of type int\"\n assert isinstance(t_end, int), \"t_end must be of type int\"\n assert b2.units.fundamentalunits.have_same_dimensions(amplitude, b2.amp), \\\n \"amplitude must have the dimension of current e.g. brian2.uamp\"\n\n tmp_size = 1 + t_end # +1 for t=0\n if append_zero:\n tmp_size += 1\n tmp = np.zeros((tmp_size, 1)) * b2.amp\n \n tmp[t_start] = amplitude\n for i in range(t_start+1, t_end) :\n tmp[i] = amplitude + sigma*(time_step**(-0.5))*np.random.randn()\n \n curr = b2.TimedArray(tmp, dt= unit_time)\n return curr\n\n# -------------------\namplitude = 1.5*b2.nA\ntime_unit = 1.*b2.ms\ntime_step = 1.*b2.ms\nsigma = 1*b2.nA*time_unit**(0.5)\n\n# Create a noisy step current\nnoisy_step_current = get_noisy_step_current(t_start=50, t_end=500, unit_time = time_step,\n amplitude= amplitude, sigma = sigma)\n\n# Run the LIF model\n(state_monitor,spike_monitor) = LIF.simulate_LIF_neuron(input_current=noisy_step_current, \\\n simulation_time = 500*b2.ms)\n\n# plot I and vm\nplot_tools.plot_voltage_and_current_traces(state_monitor, noisy_step_current, title=\"min input\", \\\n firing_threshold=LIF.FIRING_THRESHOLD)\nprint(\"nr of spikes: {}\".format(spike_monitor.count[0]))", "nr of spikes: 3\n" ] ], [ [ "2 - How does the neuron behave? Discuss your result. Your answer should be max 3 lines long.", "_____no_output_____" ], [ "The current behaves randomly increasing or decreasing at each time step (like in a Markov process). The membrane voltage reacts to this input current by following the increasing or decreasing pattern. When the current is large enough, the membrane potential reaches the firing threshold and spikes are generated. ", "_____no_output_____" ], [ "### 2.5.2 Subthreshold vs. superthreshold regime\n#### [7 + 5 = 12 points]\n1 - A time-dependent input current $I(t)$ is called subthreshold if it does not lead to spiking, i.e. if it leads to a membrane potential that stays - in the absence of noise - below the firing threshold. When noise is added, however, even subthreshold stimuli can induce spikes. Input stimuli that lead to spiking even in a noise-free neuron are called superthreshold. Sub- and superthreshold inputs, in the presence and absence of noise give rise to different spiking behaviour. These 4 different regimes (sub, super, noiseless, noisy) are what we will explore in this exercise.\n\nCreate a function that takes the amplitudes of a step current and the noise as arguments. It should simulate the LIF-model with this input, calculate the interspike intervals (ISI) and plot a histogram of the ISI (the interspike interval is the time interval between two consecutive spikes).\n\nIn order to do so edit the function test_effect_of_noise provided below. A few more details:\n* Use the function spike_tools.get_spike_train_stats (http://neuronaldynamics-exercises.readthedocs.io/en/latest/_modules/neurodynex/tools/spike_tools.html#get_spike_train_stats) to get the ISI. Have a look at its source code to understand how to use it and what it returns. You may need to use other parts of the documentation as well.\n* You will need to simulate the neuron model for long enough to get some statistics.\n* Optional and recommended: What would you expect the resulting histograms to look like?\n\n2 - Run your function and create the ISI histograms for the following four regimes:\n* No noise, subthreshold: $I_0 = 1.9nA$, $\\sigma = 0 nA* \\sqrt{\\text{your time unit}}$\n* Noise, subthreshold regime: $I_0 = 1.9nA$, $\\sigma = 1 nA* \\sqrt{\\text{your time unit}}$\n* No noise, superthreshold regime: $I_0 = 2.5nA$, $\\sigma = 0 nA* \\sqrt{\\text{your time unit}}$\n* Noise, superthreshold regime: $I_0 = 2.5nA$, $\\sigma = 1 nA* \\sqrt{\\text{your time unit}}$", "_____no_output_____" ] ], [ [ "from neurodynex.tools import spike_tools, plot_tools\n\n# time unit. e.g.\ntime_unit = 1.*b2.ms\ntime_step = time_unit\n\ndef test_effect_of_noise(amplitude, sigma, bins = np.linspace(0,1,50)):\n \n # Create a noisy step current \n noisy_step_current = get_noisy_step_current(t_start=50, t_end=5000, unit_time = time_step,\n amplitude= amplitude, sigma = sigma)\n \n # Run the LIF model\n (state_monitor,spike_monitor) = LIF.simulate_LIF_neuron(input_current=noisy_step_current, \\\n simulation_time = 5000 * b2.ms)\n\n plt.figure()\n plot_tools.plot_voltage_and_current_traces(state_monitor, noisy_step_current, title=\"\", \\\n firing_threshold=LIF.FIRING_THRESHOLD)\n plt.show()\n \n print(\"nr of spikes: {}\".format(spike_monitor.count[0]))\n \n # Use the function spike_tools.get_spike_train_stats\n spike_stats = spike_tools.get_spike_train_stats(spike_monitor) \n\n # Make the ISI histogram\n if len(spike_stats._all_ISI) != 0:\n plt.hist(np.asarray(spike_stats._all_ISI), bins)\n # choose an appropriate window size for the x-axis (ISI-axis)!\n plt.xlabel(\"ISI [s]\")\n plt.ylabel(\"Number of spikes\")\n plt.show()\n \n return spike_stats\n\n\n# 1. No noise, subthreshold\nstats1 = test_effect_of_noise(amplitude = 1.9 *b2.nA, sigma = 0*b2.nA*time_unit**(0.5))\n\n# 2. Noise, subthreshold regime\nstats2 = test_effect_of_noise(amplitude = 1.9*b2.nA, sigma = 1*b2.nA*time_unit**(0.5), bins = np.linspace(0, 0.1, 100))\n\n# 3. No noise, superthreshold regime\nstats3 = test_effect_of_noise(amplitude = 2.5*b2.nA, sigma = 0*b2.nA*time_unit**(0.5),bins = np.linspace(0, 0.1, 100))\n\n# 4. Noise, superthreshold regime\nstats4 = test_effect_of_noise(amplitude = 2.5*b2.nA, sigma = 1*b2.nA*time_unit**(0.5), bins = np.linspace(0, 0.1, 100))", "_____no_output_____" ] ], [ [ "2 - Discuss yout results (ISI histograms) for the four regimes regimes. For help and inspiration, as well as for verification of your results, have a look at the book chapter 8.3 (http://neuronaldynamics.epfl.ch/online/Ch8.S3.html). Your answer should be max 5 lines long.\n#### [5 points]", "_____no_output_____" ], [ "\\begin{itemize}\n\\item The first regime (no noise, subthreshold) does not generate any spikes. The histogram does not exist.\n\\item The second regime (noise, subthreshold) generates spikes, but less than the superthreshold regimes, as the voltage is generally under the firing threshold. The interspike interval is concentrated around 0.02 [s].\n\\item The third regime (no noise, superthreshold) generates regularly separated spikes. The histogram only has one column as the time between each spike is always the same (0.012 [s]).\n\\item The fourth regime (noise, superthreshold) generates as many spikes as the noiseless superthreshold regime. The interspike interval is concentrated around 0.012 [s] which is faster than the noisy subthreshold regime as the current is generally higher.\n\\end{itemize}", "_____no_output_____" ], [ "3 - For the ISI histograms you needed to simulate the neuron for a long time to gather enought statistics for the ISI. If you wanted to parallelize this procedure in order to reduce the computation time (e.g. you have multiple CPU cores on your machine), what would be a simple method to do that? Your answer should be max 3 lines long.\n\nHint: Temporal vs. ensemble average...\n#### [2 points]", "_____no_output_____" ], [ "You can simulate the neuron separately on 4 cores to obtain 4 times more ISI than if running on only 1 core. You can then aggregate the data of the 4 cores to compute the ISI histograms.", "_____no_output_____" ], [ "### 2.5.3 Noisy sinusoidal input current\nImplement the noisy sinusoidal input current $I(t) + I_{noise}$. As before, edit the function provided below; you only have to add the noisy part ot the current.\n\nThen create a noisy sinusoidal current with amplitude = $2.5nA$, frequency = $100Hz$, $\\sigma = 1 nA* \\sqrt{\\text{your time unit}}$ and direct_current = $1.5nA$, run the LIF model and plot the input current and the membrane potential, as you did in the previous exercises. What do you observe when compared to the noiseless case ($\\sigma = 0 nA*\\sqrt{\\text{your time unit}}$)?\n#### [5 points]", "_____no_output_____" ] ], [ [ "import math\n\ndef get_noisy_sinusoidal_current(t_start, t_end, unit_time,\n amplitude, frequency, direct_current, sigma, phase_offset=0.,\n append_zero=True):\n \"\"\"Creates a noisy sinusoidal current. If t_start == t_end, then ALL entries are 0.\n\n Args:\n t_start (int): start of the sine wave\n t_end (int): end of the sine wave\n unit_time (Quantity, Time): unit of t_start and t_end. e.g. 0.1*brian2.ms\n amplitude (Quantity, Current): maximum amplitude of the sinus e.g. 3.5*brian2.uamp\n frequency (Quantity, Hz): Frequency of the sine. e.g. 0.5*brian2.kHz\n direct_current(Quantity, Current): DC-component (=offset) of the current\n sigma (float): amplitude (std) of the noise. e.g. 0.1*b2.uamp\n phase_offset (float, Optional): phase at t_start. Default = 0.\n append_zero (bool, optional): if true, 0Amp is appended at t_end+1. Without that\n trailing 0, Brian reads out the last value in the array for all indices > t_end.\n\n\n Returns:\n TimedArray: Brian2.TimedArray\n \"\"\"\n assert isinstance(t_start, int), \"t_start_ms must be of type int\"\n assert isinstance(t_end, int), \"t_end must be of type int\"\n assert b2.units.fundamentalunits.have_same_dimensions(amplitude, b2.amp), \\\n \"amplitude must have the dimension of current. e.g. brian2.uamp\"\n assert b2.units.fundamentalunits.have_same_dimensions(direct_current, b2.amp), \\\n \"direct_current must have the dimension of current. e.g. brian2.uamp\"\n assert b2.units.fundamentalunits.have_same_dimensions(frequency, b2.Hz), \\\n \"frequency must have the dimension of 1/Time. e.g. brian2.Hz\"\n \n tmp_size = 1 + t_end # +1 for t=0\n if append_zero:\n tmp_size += 1\n tmp = np.zeros((tmp_size, 1)) * b2.amp\n if t_end > t_start: # if deltaT is zero, we return a zero current\n phi = range(0, (t_end - t_start) + 1)\n phi = phi * unit_time * frequency\n phi = phi * 2. * math.pi + phase_offset\n c = np.sin(phi)\n c = (direct_current + c * amplitude) # add direct current and scale by amplitude\n tmp[t_start: t_end + 1, 0] = c # add sinusoidal part of current\n \n for i in range(t_start, t_end) :\n # Add noisy part of current here\n # Pay attention to correct scaling with respect to the unit_time (time_step)\n tmp[i] += sigma*(time_step**(-0.5))*np.random.randn()\n \n curr = b2.TimedArray(tmp, dt= unit_time)\n return curr\n \n# ------------------\namplitude = 2.5 *b2.nA\nfrequency = 100 *b2.Hz \ntime_unit = 1.*b2.ms\ntime_step = 0.1*b2.ms # This is needed for higher temporal resolution \nsigma = 1*b2.nA*time_unit**(0.5)\ndirect_current = 1.5 *b2.nA\n\n# Create a noiseless sinusoidal current\nnoisy_sinusoidal_current = get_noisy_sinusoidal_current(200, 800, unit_time = time_step,\n amplitude= amplitude, frequency=frequency,\n direct_current=direct_current, sigma = 0 *b2.nA*time_unit**(0.5))\n\n# Run the LIF model\n(state_monitor,spike_monitor) = LIF.simulate_LIF_neuron(input_current=noisy_sinusoidal_current, \\\n simulation_time = 100 * b2.ms)\n\n# plot I and vm\nplot_tools.plot_voltage_and_current_traces(state_monitor, noisy_sinusoidal_current, title=\"\", \\\n firing_threshold=LIF.FIRING_THRESHOLD)\nprint(\"nr of spikes: {}\".format(spike_monitor.count[0]))", "nr of spikes: 0\n" ], [ "# Create a noisy sinusoidal current\nnoisy_sinusoidal_current = get_noisy_sinusoidal_current(200, 800, unit_time = time_step,\n amplitude= amplitude, frequency=frequency,\n direct_current=direct_current, sigma = sigma)\n\n# Run the LIF model\n(state_monitor,spike_monitor) = LIF.simulate_LIF_neuron(input_current=noisy_sinusoidal_current, \\\n simulation_time = 100 * b2.ms)\n\n# plot I and vm\nplot_tools.plot_voltage_and_current_traces(state_monitor, noisy_sinusoidal_current, title=\"\", \\\n firing_threshold=LIF.FIRING_THRESHOLD)\nprint(\"nr of spikes: {}\".format(spike_monitor.count[0]))", "nr of spikes: 2\n" ] ], [ [ "In the noiseles case, the voltage reaches the firing threshold but does not generate any spike. In the noisy case we observe some spikes when the firing threshold is exceeded around the maxima of the sinusoid.", "_____no_output_____" ], [ "### 2.5.4 Stochastic resonance (Bonus, not graded)\nContrary to what one may expect, some amount of noise under certain circumstances can improve the signal transmission properties of neurons. In the subthreshold regime, a neuron cannot transmit any information about the temporal structure of its input since it does not spike. With some noise, there is some probability to spike, with the probability depending on the time dependent input (inhomogeneous Poisson process). However to much noise covers the signal completely and thus, there is usually an optimal value for the amplitude of the noise. This phenomenon is called \"stochastic resonance\" and we will briefly touch upon it in this exercise. To get an idea of the effect we suggest reading section 9.4.2 in the book: http://neuronaldynamics.epfl.ch/online/Ch9.S4.html.\n\n1 - Simulate several (e.g. n_inits = 5) trials of a LIF neuron with noisy sinusoidal current. For each trial calculate the power spectrum of the resulting spike train (using the function spike_tools.get_averaged_single_neuron_power_spectrum). Finally calculate the average power spectrum and plot it. With appropriate noise amplitudes, you should see a pronounced peak at the driving frequency, while without noise we don't see anything in the power spectrum since no spike was elicited in the subthreshold regime we are in. \n\nIn order to do that use the provided parameters and edit the code provided below. Complete the function _run_sim() which creates the input current, runs a simulation and computes the power spectrum. Call it in a loop to execute several trials. Then average over the spectra to obtain a smooth spectrum to plot.", "_____no_output_____" ] ], [ [ "amplitude = 1.*b2.nA\nfrequency = 20*b2.Hz\ntime_unit = 1.*b2.ms\ntime_step = .1*b2.ms\ndirect_current = 1. * b2.nA\nsampling_frequency = .01/time_step\nnoise_amplitude = 2.\nn_inits = 5\n\n# run simulation and calculate power spectrum\ndef _run_sim(amplitude, noise_amplitude):\n noisy_sinusoidal_current = get_noisy_sinusoidal_current(50, 100000, unit_time = time_step,\n amplitude= ???, frequency= ???,\n direct_current= ???,\n sigma = noise_amplitude*b2.nA*np.sqrt(time_unit))\n # run the LIF model\n (state_monitor,spike_monitor) = LIF.simulate_LIF_neuron(input_current=noisy_sinusoidal_current, \\\n simulation_time = 10000 * b2.ms)\n # get power spectrum\n freq, mean_ps, all_ps_dict, mean_firing_rate, mean_firing_freqs_per_neuron = \\\n spike_tools.get_averaged_single_neuron_power_spectrum(spike_monitor, sampling_frequency,\n window_t_min = 1000*b2.ms, window_t_max = 9000*b2.ms,\n nr_neurons_average=1, subtract_mean=True)\n \n return freq, all_ps_dict, mean_firing_rate\n\n\n# initialize array\nspectra = []\n# run a few simulations, calculate the power spectrum and append it to the spectra array\nfor i in ???:\n freq, spectrum, mfr = ???\n spectra.append(spectrum[0])\n\n# average spectra over trials\nspectrum = ??? # hint: us np.mean with axis=0\n\n# plotting, frequencies vs the obtained spectrum:\nplt.figure()\nplt.plot(???,???)\nplt.xlabel(\"???\")\nplt.ylabel(\"???\")", "_____no_output_____" ] ], [ [ "2 - We now apply different noise levels to investigate the optimal noise level of stochastic resonance. \n\nThe quantity to optimize is the signal-to-noise ratio (SNR). Here, the SNR is defined as the intensity of the power spectrum at the driving frequency (the peak from above), divided by the value of the background noise (power spectrum averaged around the peak).\n\nIn order to do that edit the code provided below. You can re-use the function _run_sim() to obtain the power spectrum of on trial. The calculation of the SNR is already implemented and doesn't need to be changed.\n\nWhen you are done with completing the code, run the simulation with the proposed parameters (This could take several minutes...). The result should be a plot showing an optimal noise amplitude, i.e. a $\\sigma$ where the SNR is maximal.", "_____no_output_____" ] ], [ [ "def get_snr(amplitude, noise_amplitude, n_inits):\n \n spectra = []\n snr = 0.\n for i in range(0,n_inits):\n # run model with noisy sinusoidal \n freq_signal, spectrum, mfr = ???\n spectra.append(spectrum[0])\n\n # Average over trials to get power spectrum\n spectrum = ???\n\n if mfr != 0.*b2.Hz:\n peak = np.amax(spectrum)\n index_of_peak = np.argmax(spectrum)\n # snr: divide peak value by average of surrounding values\n snr = peak/np.mean(np.concatenate((spectrum[index_of_peak-100:index_of_peak-1],\\\n spectrum[index_of_peak+1:index_of_peak+100])))\n else:\n snr = 0.\n \n return snr\n\nnoise_amplitudes = np.arange(0.,5.,.5)\nsnr = np.zeros(len(noise_amplitudes))\nfor j in np.arange(0,len(noise_amplitudes)):\n snr[j] = get_snr(amplitude, noise_amplitudes[j], n_inits = 8)\n\n\nplt.figure()\nplt.plot(noise_amplitudes,snr)\nplt.xlabel(\"???\")\nplt.ylabel(\"???\")\nplt.show()", "_____no_output_____" ] ], [ [ "3 - For further reading on this topic, consult the book chapter 9.4.2 (http://neuronaldynamics.epfl.ch/online/Ch9.S4.html#Ch9.F10).", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown" ]

[ [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown", "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown", "markdown", "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Azure ML Training Pipeline for COVID-CXR\nThis notebook defines an Azure machine learning pipeline for a single training run and submits the pipeline as an experiment to be run on an Azure virtual machine.", "_____no_output_____" ] ], [ [ "# Import statements\nimport azureml.core\nfrom azureml.core import Experiment\nfrom azureml.core import Workspace, Datastore\nfrom azureml.data.data_reference import DataReference\nfrom azureml.pipeline.core import PipelineData\nfrom azureml.pipeline.core import Pipeline\nfrom azureml.pipeline.steps import PythonScriptStep, EstimatorStep\nfrom azureml.train.dnn import TensorFlow\nfrom azureml.train.estimator import Estimator\nfrom azureml.core.compute import ComputeTarget, AmlCompute\nfrom azureml.core.compute_target import ComputeTargetException\nfrom azureml.core.environment import Environment\nfrom azureml.core.runconfig import RunConfiguration\nimport shutil", "_____no_output_____" ] ], [ [ "### Register the workspace and configure its Python environment.", "_____no_output_____" ] ], [ [ "# Get reference to the workspace\nws = Workspace.from_config(\"./ws_config.json\")\n\n# Set workspace's environment\nenv = Environment.from_pip_requirements(name = \"covid-cxr_env\", file_path = \"./../requirements.txt\")\nenv.register(workspace=ws)\nrunconfig = RunConfiguration(conda_dependencies=env.python.conda_dependencies)\nprint(env.python.conda_dependencies.serialize_to_string())\n\n# Move AML ignore file to root folder\naml_ignore_path = shutil.copy('./.amlignore', './../.amlignore') ", "_____no_output_____" ] ], [ [ "### Create references to persistent and intermediate data\nCreate DataReference objects that point to our raw data on the blob. Configure a PipelineData object to point to preprocessed images stored on the blob.", "_____no_output_____" ] ], [ [ "# Get the blob datastore associated with this workspace\nblob_store = Datastore(ws, name='covid_cxr_ds')\n\n# Create data references to folders on the blob\nraw_data_dr = DataReference(\n datastore=blob_store,\n data_reference_name=\"raw_data\",\n path_on_datastore=\"data/\")\nmila_data_dr = DataReference(\n datastore=blob_store,\n data_reference_name=\"mila_data\",\n path_on_datastore=\"data/covid-chestxray-dataset/\")\nfig1_data_dr = DataReference(\n datastore=blob_store,\n data_reference_name=\"fig1_data\",\n path_on_datastore=\"data/Figure1-COVID-chestxray-dataset/\")\nrsna_data_dr = DataReference(\n datastore=blob_store,\n data_reference_name=\"rsna_data\",\n path_on_datastore=\"data/rsna/\")\ntraining_logs_dr = DataReference(\n datastore=blob_store,\n data_reference_name=\"training_logs_data\",\n path_on_datastore=\"logs/training/\")\nmodels_dr = DataReference(\n datastore=blob_store,\n data_reference_name=\"models_data\",\n path_on_datastore=\"models/\")\n\n# Set up references to pipeline data (intermediate pipeline storage).\nprocessed_pd = PipelineData(\n \"processed_data\",\n datastore=blob_store,\n output_name=\"processed_data\",\n output_mode=\"mount\")", "_____no_output_____" ] ], [ [ "### Compute Target\nSpecify and configure the compute target for this workspace. If a compute cluster by the name we specified does not exist, create a new compute cluster.", "_____no_output_____" ] ], [ [ "CT_NAME = \"nd12s-clust-hp\" # Name of our compute cluster\nVM_SIZE = \"STANDARD_ND12S\" # Specify the Azure VM for execution of our pipeline\n#CT_NAME = \"d2-cluster\" # Name of our compute cluster\n#VM_SIZE = \"STANDARD_D2\" # Specify the Azure VM for execution of our pipeline\n\n# Set up the compute target for this experiment\ntry:\n compute_target = AmlCompute(ws, CT_NAME)\n print(\"Found existing compute target.\")\nexcept ComputeTargetException:\n print(\"Creating new compute target\")\n provisioning_config = AmlCompute.provisioning_configuration(vm_size=VM_SIZE, min_nodes=1, max_nodes=4) \n compute_target = ComputeTarget.create(ws, CT_NAME, provisioning_config) # Create the compute cluster\n \n # Wait for cluster to be provisioned\n compute_target.wait_for_completion(show_output=True, min_node_count=None, timeout_in_minutes=20) \n \nprint(\"Azure Machine Learning Compute attached\")\nprint(\"Compute targets: \", ws.compute_targets)\ncompute_target = ws.compute_targets[CT_NAME]", "_____no_output_____" ] ], [ [ "### Define pipeline and submit experiment.\nDefine the steps of an Azure machine learning pipeline. Create an Azure Experiment that will run our pipeline. Submit the experiment to the execution environment.", "_____no_output_____" ] ], [ [ "# Define preprocessing step the ML pipeline\nstep1 = PythonScriptStep(name=\"preprocess_step\",\n script_name=\"azure/preprocess_step/preprocess_step.py\",\n arguments=[\"--miladatadir\", mila_data_dr, \"--fig1datadir\", fig1_data_dr, \n \"--rsnadatadir\", rsna_data_dr, \"--preprocesseddir\", processed_pd],\n inputs=[mila_data_dr, fig1_data_dr, rsna_data_dr],\n outputs=[processed_pd],\n compute_target=compute_target, \n source_directory=\"./../\",\n runconfig=runconfig,\n allow_reuse=True)\n\n# Define training step in the ML pipeline\nest = TensorFlow(source_directory='./../',\n script_params=None,\n compute_target=compute_target,\n entry_script='azure/train_step/train_step.py',\n pip_packages=['tensorboard', 'pandas', 'dill', 'numpy', 'imblearn', 'matplotlib', 'scikit-image', 'matplotlib',\n 'pydicom', 'opencv-python', 'tqdm', 'scikit-learn'],\n use_gpu=True,\n framework_version='2.0')\nstep2 = EstimatorStep(name=\"estimator_train_step\", \n estimator=est, \n estimator_entry_script_arguments=[\"--rawdatadir\", raw_data_dr, \"--preprocesseddir\", processed_pd, \n \"--traininglogsdir\", training_logs_dr, \"--modelsdir\", models_dr],\n runconfig_pipeline_params=None, \n inputs=[raw_data_dr, processed_pd, training_logs_dr, models_dr], \n outputs=[], \n compute_target=compute_target)\n\n# Construct the ML pipeline from the steps\nsteps = [step1, step2]\nsingle_train_pipeline = Pipeline(workspace=ws, steps=steps)\nsingle_train_pipeline.validate()\n\n# Define a new experiment and submit a new pipeline run to the compute target.\nexperiment = Experiment(workspace=ws, name='SingleTrainExperiment_v3')\nexperiment.submit(single_train_pipeline, regenerate_outputs=False)\nprint(\"Pipeline is submitted for execution\")\n\n# Move AML ignore file back to original folder\naml_ignore_path = shutil.move(aml_ignore_path, './.amlignore') ", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# SIT742: Modern Data Science \n**(Week 01: Programming Python)**\n\n---\n- Materials in this module include resources collected from various open-source online repositories.\n- You are free to use, change and distribute this package.\n- If you found any issue/bug for this document, please submit an issue at [tulip-lab/sit742](https://github.com/tulip-lab/sit742/issues)\n\n\n\nPrepared by **SIT742 Teaching Team**\n\n---\n\n\n# Session 1A - IPython notebook and basic data types\n\n\nIn this session, \nyou will learn how to run *Python* code under **IPython notebook**. You have two options for the environment:\n\n1. Install the [Anaconda](https://www.anaconda.com/distribution/), and run it locally; **OR**\n1. Use one cloud data science platform such as:\n - [Google Colab](https://colab.research.google.com): SIT742 lab session will use Google Colab. \n - [IBM Cloud](https://www.ibm.com/cloud)\n - [DataBricks](https://community.cloud.databricks.com)\n\n\n\nIn IPython notebook, you will be able to execute and modify your *Python* code more efficiently. \n\n- **If you are using Google Colab for SIT742 lab session practicals, you can ignore this Part 1 of this Session 1A, and start with Part 2.**\n\n\n\nIn addition, you will be given an introduction on *Python*'s basic data types, \ngetting familiar with **string**, **number**, data conversion, data comparison and \ndata input/output. \n\nHopefully, by using **Python** and the powerful **IPython Notebook** environment, \nyou will find writing programs both fun and easy. \n", "_____no_output_____" ], [ "## Content\n\n### Part 1 Create your own IPython notebook\n\n1.1 [Start a notebook server](#cell_start)\n\n1.2 [A tour of IPython notebook](#cell_tour)\n\n1.3 [IPython notebook infterface](#cell_interface)\n\n1.4 [Open and close notebooks](#cell_close)\n\n### Part 2 Basic data types\n\n2.1 [String](#cell_string)\n\n2.2 [Number](#cell_number)\n\n2.3 [Data conversion and comparison](#cell_conversion)\n\n2.4 [Input and output](#cell_input)\n", "_____no_output_____" ], [ "# Part 1. Create your own IPython notebook\n\n- **If you are using Google Colab for SIT742 lab session practicals, you can ignore this Part 1, and start with Part 2.**\n\n\nThis notebook will show you how to start an IPython notebook session. It guides you through the process of creating your own notebook. It provides you details on the notebook interface and show you how to nevigate with a notebook and manipulate its components. ", "_____no_output_____" ], [ "<a id = \"cell_start\"></a>", "_____no_output_____" ], [ " \n## 1. 1 Start a notebook server\n\nAs described in Part 1, you start the IPython notebnook server by keying in the command in a terminal window/command line window.\n\nHowever, before you do this, make sure you have created a folder **p01** under **H:/sit742**, download the file **SIT742P01A-Python.ipynb** notebook, and saved it under **H:/sit742/p01**.\n\nIf you are using [Google Colab](https://colab.research.google.com), you can upload this notebook to Google Colab and run it from there. If any difficulty, please ask your tutor, or check the CloudDeakin discussions.\n\n\nAfter you complete this, you can now switch working directory to **H:/sit742**, and start the IPython notebook server by the following commands:", "_____no_output_____" ] ], [ [ "ipython notebook %don't run this in notebook, run it on command line to activate the server", "_____no_output_____" ] ], [ [ "You can see the message in the terminal windows as follows:\n\n\n<img src=\"https://raw.githubusercontent.com/tuliplab/mds/master/Jupyter/image/start-workspace.jpg\">\n\nThis will open a new browser window(or a new tab in your browser window). In the browser, there is an **dashboard** page which shows you all the folders and files under **sit742** folder\n\n\n<img src=\"https://raw.githubusercontent.com/tuliplab/mds/master/Jupyter/image/start-index.jpg\">\n ", "_____no_output_____" ], [ "<a id = \"cell_tour\"></a>", "_____no_output_____" ], [ " \n ## 1.2 A tour of iPython notebook \n \n ### Create a new ipython notebook\n \n \n To create a new notebook, go to the menu bar and select **File -> New Notebook -> Python 3**\n\nBy default, the new notebook is named **Untitled**. To give your notebook a meaningful name, click on the notebook name and rename it. We would like to call our new notebook **hello.ipynb**. Therefore, key in the name **hello**. \n\n<img src=\"https://raw.githubusercontent.com/tuliplab/mds/master/Jupyter/image/emptyNotebook.jpg\">\n\n\n\n### Run script in code cells\n\nAfter a new notebook is created, there is an empty box in the notebook, called a **cell**. If you double click on the cell, you enter the **edit** mode of the notebook. Now we can enter the following code in the cell", "_____no_output_____" ] ], [ [ "text = \"Hello World\"\nprint(text)", "_____no_output_____" ] ], [ [ "After this, press **CTRL + ENTER**, and execute the cell. The result will be shown after the cell. \n\n\n\n<img src=\"https://raw.githubusercontent.com/tuliplab/mds/master/Jupyter/image/hello-world.jpg\">\n\n\n\n\nAfter a cell is executed , the notebook is switched to the **Commmand** mode. In this mode, you can manipulte the notebook and its commponent. Alternatively, you can use **ESC** key to switch from **Edit** mode to **Command** mode without executing code. \n\nTo modify the code you entered in the cell, **double click** the cell again and modify its content. For example, try to change the first line of previouse cell into the following code:", "_____no_output_____" ] ], [ [ "text = \"Good morning World!\"", "_____no_output_____" ] ], [ [ "Afterwards, press **CTRL + ENTER**, and the new output is displayed. \n\nAs you can see, you are switching bewteen two modes, **Command** and **Edit**, when editing a notebook. We will in later section look into these two operation modes of closely. Now practise switching between the two modes until you are comfortable with them. ", "_____no_output_____" ], [ "### Add new cells\n\nTo add a new cell to a notebook, you have to ensure the notebook is in **Command** mode. If not, refer to previous section to switch to **Command** mode. \n\n\nTo add cell below the currrent cell, go to menubar and click **Insert-> Insert Cell Below**. Alternatively, you can use shortcut i.e. pressing **b** (or **a** to create a cell above).\n\n\n<img src=\"https://raw.githubusercontent.com/tuliplab/mds/master/Jupyter/image/new-cell.jpg\">\n\n### Add markdown cells\n\nBy default, a code cell is created when adding a new cell. However, IPython notebook also use a **Markdown** cell for enter normal text. We use markdown cell to display the text in specific format and to provide structure for a notebook. \n\nTry to copy the text in the cell below and paste it into your new notebook. Then from toolbar(**Cell->Cell Type**), change cell type from **Code** to **Markdown**. \n\nPlease note in the following cell, there is a space between the leading **-, #, 0** and the text that follows. ", "_____no_output_____" ] ], [ [ "## Heading 2\nNormal text here!\n\n### Heading 3\nordered list here\n\n0. Fruits\n 0. Banana\n 0. Grapes\n0. Veggies\n 0. Tomato\n 0. Broccoli\n \nUnordered list here\n- Fruits\n - Banana\n - Grapes\n- Veggies\n - Tomato\n - Broccoli", "_____no_output_____" ] ], [ [ "Now execute the cell by press **CTRL+ ENTER**. You notebook should look like this:\n \n<img src=\"https://raw.githubusercontent.com/tuliplab/mds/master/Jupyter/image/new-markdown.jpg\">\n", "_____no_output_____" ], [ "Here is what the formated Markdown cell looks like:", "_____no_output_____" ], [ "### Exercise:\nClick this cell, and practise writing markdown language here....", "_____no_output_____" ], [ "<a id = \"cell_interface\"></a>", "_____no_output_____" ], [ " \n### 1.3 IPython notebook interface\n\nNow you have created your first notebook, let us have a close look at the user interface of IPython notebook.\n\n\n\n### Notebook component\nWhen you create a new notebook document, you will be presented with the notebook name, a menu bar, a toolbar and an empty code cell.\n\nWe can see the following components in a notebook:\n\n- **Title bar** is at the top of the page and contains the name of the notebook. Clicking on the notebook name brings up a dialog which allows you to rename it. Please renaming your notebook name from “Untitled0” to “hello”. This change the file name from **Untitled0.ipynb** to **hello.ipynb**.\n\n- **Menu bar** presents different options that can be used to manipulate the way the notebook functions.\n\n- **Toolbar** gives a quick way of performing the most-used operations within the notebook.\n\n- An empty computational cell is show in a new notebook where you can key in your code.\n\nThe notebook has two modes of operatiopn:\n\n- **Edit**: In this mode, a single cell comes into focus and you can enter text or execute code. You activate the **Edit mode** by **clicking on a cell** or **selecting a cell and then pressing Enter key**.\n\n- **Command**: In this mode, you can perform tasks that is related to the whole notebook structure. For example, you can move, copy, cut and paste cells. A series of keyboard shortcuts are also available to enable you to performa these tasks more effiencient. One easiest way of activating the command mode by pressing the **Esc** key to exit editing mode. \n\n\n\n\n\n### Get help and interrupting\n\n\nTo get help on the use of different cammands, shortcuts, you can go to the **Help** menu, which provides links to relevant documentation.\n\nIt is also easy to get help on any objects(including functions and methods). For example, to access help on the sum() function, enter the followsing line in a cell:", "_____no_output_____" ] ], [ [ "sum?", "_____no_output_____" ] ], [ [ "The other improtant thing to know is how to interrupt a compuation. This can be done through the menu **Kernel->Interrupt** or **Kernel->Restart**, depending on what works on the situation. We will have chance to try this in later session. ", "_____no_output_____" ], [ "\n### Notebook cell types\n\n\nThere are basically three types of cells in a IPython notebook: Code Cells, Markdown Cells, Raw Cells.\n\n\n**Code cells** : Code cell can be used to enter code and will be executed by Python interpreter. Although we will not use other language in this unit, it is good to know that Jupyter Notebooks also support JavaScript, HTML, and Bash commands.\n\n*** Markdown cells***: You have created markdown cell in the previouse section. Markdown Cells are the easiest way to write and format text. It is also give structure to the notebook. Markdown language is used in this type of cell. Follow this link https://daringfireball.net/projects/markdown/basics for the basics of the syntax. \n\nThis is a Markdown Cells example notebook sourced from : https://ipython.org/ipython-doc/3/notebook/notebook.html\nThis markdown cheat sheet can also be good reference to the main markdowns you might need to use in our pracs http://nestacms.com/docs/creating-content/markdown-cheat-sheet\n\n\n**Raw cells** : Raw cells, unlike all other Jupyter Notebook cells, have no input-output distinction. This means that raw Cells cannot be rendered into anything other than what they already are. They are mainly used to create examples.\n\n\nAs you have seen, you can use the toolbar to choose between different cell types. In addition, shortcut **M** and **Y** can be used to quickly change a cell to Code cell or Markdown cell under Command mode. \n\n\n### Operation modes of IPytho notebook\n\n**Edit mode**\n\n\n\nThe Edit mode is used to enter text in cells and to execute code. As you have seen, after typing some code in the notebook and pressing **CTRL+Enter**, the notebook executes the cell and diplays output. The other two shortcuts used to run code in a cell are **Shift +Enter** and **Alt + Enter**. \n\nThese three ways to run the the code in a cells are summarized as follows:\n\n\n- Pressing Shift + Enter: This runs the cell and select the next cell(A new cell is created if at the end of the notebook). This is the most usual way to execute a cell.\n\n- Pressing Ctrl + Enter: This runs the cell and keep the same cell selected. \n\n- Pressing Alt + Enter: This runs the cell and insert a new cell below it. \n\n\n**Command mode**\n\nIn Command mode, you can edit the notebook as a whole, but not type into individual cells.\n\nYou can use keyboard shortcut in this mode to perform the notebook and cell actions effeciently. For example, if you are in command mode and press **c**, you will copy the current cell. \n\n\n\n There are a large amount of shortcuts avaialbe in the command mode. However, you do not have to remember all of them, since most actions in the command mode are available in the menu. \n \n Here is a list of the most useful shortcuts. They are arrganged by the \n order we recommend you learn so that you can edit the cells effienctly.\n\n\n1. Basic navigation: \n\n - Enter: switch to Edit mode\n \n - Esc: switch to Command mode\n \n - Shift+enter: Eexecute a cell\n \n - Up, down: Move to the cell above or below\n\n2. Cell types: \n - y: switch to code cell)\n - m: switch to markdown cell) \n\n3. Cell creation: \n - a: insert new sell above\n - b: insert new cell below\n\n4. Cell deleting:\n - press D twice.\n\nNote that one of the most common (and frustrating) mistakes when using the\nnotebook is to type something in the wrong mode. Remember to use **Esc**\nto switch to the Command mode and **Enter** to switch to the Edit mode.\nAlso, remember that **clicking** on a cell automatically places it in the Edit\nmode, so it will be necessary to press **Esc** to go to the Command mode.\n\n### Exercise\nPlease go ahead and try these shortcuts. For example, try to insert new cell, modify and delete an existing cell. You can also switch cells between code type and markdown type, and practics different kinds of formatting in a markdown cell. \n \nFor a complete list of shortcut in **Command** mode, go to menu bar **Help->Keyboardshorcut**. Feel free to explore the other shortcuts. \n \n ", "_____no_output_____" ], [ "<a id = \"cell_close\"></a>", "_____no_output_____" ], [ " \n ## 1.4 open and close notebooks\n \n You can open multiple notebooks in a browser windows. Simply go to menubar and choose **File->open...**, and select one **.ipynb** file. The second notebook will be opened in a seperated tab. \n \nNow make sure you still have your **hello.ipynb** open. Also please download **ControlAdvData.ipynb** from cloudDeakin, and save under **H:/sit742/prac01**. Now go to the manu bar, click on **File->open ...**, locate the file **ControlAdvData.ipynb**, and open this file. \n \n When you finish your work, you will need to close your notebooks and shutdown the IPython notebook server. Instead of simply close all the tabs in the browser, you need to shutdown each notebook first. To do this, swich to the **Home** tab(**Dashboard page**) and **Running** section(see below). Click on **Shutdown** button to close each notebook. In case **Dashboard** page is not open, click on the **Jupyter** icon to reopen it. \n \n\n <img src=\"https://raw.githubusercontent.com/tuliplab/mds/master/Jupyter/image/close-index.jpg\">\n \n After each notebook is shutdown, it is time to showdown the IPython notebook server. To do this, go to the terminal window and press **CTRL + C**, and then enter **Y**. After the notebook server is shut down, the terminal window is ready for you to enter any new command. \n \n\n <img src=\"https://raw.githubusercontent.com/tuliplab/mds/master/Jupyter/image/close-terminal.jpg\">\n \n ", "_____no_output_____" ], [ "\n# Part 2 Basic Data Types\n", "_____no_output_____" ], [ "\n\nIn this part, you will get better understanding with Python's basic data type. We will \nlook at **string** and **number** data type in this section. Also covered are:\n\n- Data conversion\n- Data comparison\n- Receive input from users and display results effectively \n\nYou will be guided through completing a simple program which receives input from a user,\n process the information, and display results with specific format. ", "_____no_output_____" ], [ "<a id = \"cell_string\"></a>", "_____no_output_____" ], [ "## 2.1 String\n\nA string is a *sequence of characters*. We are using strings in almost every Python\nprograms. As we can seen in the **”Hello, World!”** example, strings can be specified\nusing single quotes **'**. The **print()** function can be used to display a string.", "_____no_output_____" ] ], [ [ "print('Hello, World!')", "_____no_output_____" ] ], [ [ "We can also use a variable to store the string value, and use the variable in the\n**print()** function.", "_____no_output_____" ] ], [ [ "# Assign a string to a variable \ntext = 'Hello, World!'\nprint(text)", "_____no_output_____" ] ], [ [ "A *variable* is basically a name that represents (or refers to) some value. We use **=**\nto assign a value to a variable before we use it. Variable names are given by a programer\nin a way that the program is easy to understanding. Variable names are *case sensitive*.\nIt can consist of letters, digits and underscores. However, it can not begin with a digit.\nFor example, **plan9** and **plan_9** are valid names, where **9plan** is not.", "_____no_output_____" ] ], [ [ "text = 'Hello, World!'", "_____no_output_____" ], [ "# with print() function, content is displayed without quotation mark\nprint(text)", "_____no_output_____" ] ], [ [ "With variables, we can also display its value without **print()** function. Note that\nyou can not display a variable without **print()** function in Python script(i.e. in a **.py** file). This method only works under interactive mode (i.e. in the notebook). ", "_____no_output_____" ] ], [ [ "# without print() function, quotation mark is displayed together with content\ntext ", "_____no_output_____" ] ], [ [ "Back to representation of string, there will be issues if you need to include a quotation\nmark in the text.", "_____no_output_____" ] ], [ [ "text = ’What’ s your name ’", "_____no_output_____" ] ], [ [ "Since strings in double quotes **\"** work exactly the same way as string in single quotes.\nBy mixing the two types, it is easy to include quaotation mark itself in the text.", "_____no_output_____" ] ], [ [ "text = \"What' s your name?\"\nprint(text)", "_____no_output_____" ] ], [ [ "Alternertively, you can use:", "_____no_output_____" ] ], [ [ "text = '\"What is the problem?\", he asked.'\nprint(text)", "_____no_output_____" ] ], [ [ "You can specify multi-line strings using triple quotes (**\"\"\"** or **'''**). In this way, single\nquotes and double quotes can be used freely in the text.\nHere is one example:", "_____no_output_____" ] ], [ [ "multiline = '''This is a test for multiline. This is the first line. \nThis is the second line. \nI asked, \"What's your name?\"'''\nprint(multiline)", "_____no_output_____" ] ], [ [ "Notice the difference when the variable is displayed without **print()** function in this case.", "_____no_output_____" ] ], [ [ "multiline = '''This is a test for multiline. This is the first line. \nThis is the second line. \nI asked, \"What's your name?\"'''\nmultiline", "_____no_output_____" ] ], [ [ "Another way of include the special characters, such as single quotes is with help of\nescape sequences **\\\\**. For example, you can specify the single quote using **\\\\' ** as follows.", "_____no_output_____" ] ], [ [ "string = 'What\\'s your name?'\nprint(string)", "_____no_output_____" ] ], [ [ "There are many more other escape sequences (See Section 2.4.1 in [Python3.0 official document](https://docs.python.org/3.1/reference/lexical_analysis.html)). But I am going to mention the most useful two examples here. \n\nFirst, use escape sequences to indicate the backslash itself e.g. **\\\\\\\\**", "_____no_output_____" ] ], [ [ "path = 'c:\\\\windows\\\\temp'\nprint(path)", "_____no_output_____" ] ], [ [ "Second, used escape sequences to specify a two-line string. Apart from using a triple-quoted\nstring as shown previously, you can use **\\n** to indicate the start of a new line.", "_____no_output_____" ] ], [ [ "multiline = 'This is a test for multiline. This is the first line.\\nThis is the second line.'\nprint(multiline)", "_____no_output_____" ] ], [ [ "To manipulate strings, the following two operators are most useful: \n* **+** is use to concatenate\ntwo strings or string variables; \n* ***** is used for concatenating several copies of the same\nstring.", "_____no_output_____" ] ], [ [ "print('Hello, ' + 'World' * 3)", "_____no_output_____" ] ], [ [ "Below is another example of string concatenation based on variables that store strings.", "_____no_output_____" ] ], [ [ "name = 'World'\ngreeting = 'Hello'\nprint(greeting + ', ' + name + '!')", "_____no_output_____" ] ], [ [ "Using variables, change part of the string text is very easy. ", "_____no_output_____" ] ], [ [ "name", "_____no_output_____" ], [ "greeting", "_____no_output_____" ], [ "# Change part of the text is easy\ngreeting = 'Good morning' \nprint(greeting + ', ' + name + '!')", "_____no_output_____" ] ], [ [ "<a id = \"cell_number\"></a>", "_____no_output_____" ], [ " ## 2.2 Number", "_____no_output_____" ], [ "There are two types of numbers that are used most frequently: integers and floats. As we\nexpect, the standard mathematic operation can be applied to these two types. Please\ntry the following expressions. Note that **\\*\\*** is exponent operator, which indicates\nexponentation exponential(power) caluclation.", "_____no_output_____" ] ], [ [ "2 + 3", "_____no_output_____" ], [ "3 * 5", "_____no_output_____" ], [ "#3 to the power of 4\n3 ** 4 ", "_____no_output_____" ] ], [ [ "Among the number operations, we need to look at division closely. In Python 3.0, classic division is performed using **/**. ", "_____no_output_____" ] ], [ [ "15 / 5", "_____no_output_____" ], [ "14 / 5", "_____no_output_____" ] ], [ [ "*//* is used to perform floor division. It truncates the fraction and rounds it to the next smallest whole number toward the left on the number line.", "_____no_output_____" ] ], [ [ "14 // 5", "_____no_output_____" ], [ "# Negatives move left on number line. The result is -3 instead of -2\n-14 // 5 ", "_____no_output_____" ] ], [ [ "Modulus operator **%** can be used to obtain remaider. Pay attention when negative number is involved.", "_____no_output_____" ] ], [ [ "14 % 5", "_____no_output_____" ], [ "# Hint: −14 // 5 equal to −3\n# (-3) * 5 + ? = -14\n\n-14 % 5 ", "_____no_output_____" ] ], [ [ "*Operator precedence* is a rule that affects how an expression is evaluated. As we learned in high school, the multiplication is done first than the addition. e.g. **2 + 3 * 4**. This means multiplication operator has higher precedence than the addition operator.\n\nFor your reference, a precedence table from the python reference manual is used to indicate the evaluation order in Python. For a complete precedence table, check the heading \"Python Operators Precedence\" in this [Python tutorial](http://www.tutorialspoint.com/python/python_basic_operators.htm)\n\n\nHowever, When things get confused, it is far better to use parentheses **()** to explicitly\nspecify the precedence. This makes the program more readable.\n\nHere are some examples on operator precedence:", "_____no_output_____" ] ], [ [ "2 + 3 * 4", "_____no_output_____" ], [ "(2 + 3) * 4", "_____no_output_____" ], [ "2 + 3 ** 2", "_____no_output_____" ], [ "(2 + 3) ** 2", "_____no_output_____" ], [ "-(4+3)+2", "_____no_output_____" ] ], [ [ "Similary as string, variables can be used to store a number so that it is easy to manipulate them.", "_____no_output_____" ] ], [ [ "x = 3\ny = 2\nx + 2", "_____no_output_____" ], [ "sum = x + y\nsum", "_____no_output_____" ], [ "x * y", "_____no_output_____" ] ], [ [ "One common expression is to run a math operation on a variable and then assign the result of the operation back to the variable. Therefore, there is a shortcut for such a expression. ", "_____no_output_____" ] ], [ [ "x = 2\nx = x * 3\nx", "_____no_output_____" ] ], [ [ "This is equivalant to:", "_____no_output_____" ] ], [ [ "x = 2\n# Note there is no space between '*' and '+'\nx *= 3\nx", "_____no_output_____" ] ], [ [ "<a id = \"cell_conversion\"></a>", "_____no_output_____" ], [ "## 2.3 Data conversion and comparison", "_____no_output_____" ], [ "So far, we have seen three types of data: interger, float, and string. With various data type, Python can define the operations possible on them and the storage method for each of them. In the later pracs, we will further introduce more data types, such as tuple, list and dictionary. \n\nTo obtain the data type of a variable or a value, we can use built-in function **type()**;\nwhereas functions, such as **str()**, **int()**, **float()**, are used to convert data one type to another. Check the following examples on the usage of these functions:", "_____no_output_____" ] ], [ [ "type('Hello, world!)')", "_____no_output_____" ], [ "input_Value = '45.6'\ntype(input_Value)", "_____no_output_____" ], [ "weight = float(input_Value)\nweight\ntype(weight)", "_____no_output_____" ] ], [ [ "Note the system will report error message when the conversion function is not compatible with the data.", "_____no_output_____" ] ], [ [ "input_Value = 'David'\nweight = float(input_Value)", "_____no_output_____" ] ], [ [ "Comparison between two values can help make decision in a program. The result of the comparison is either **True** or **False**. They are the two values of *Boolean* type.", "_____no_output_____" ] ], [ [ "5 > 10", "_____no_output_____" ], [ "type(5 > 10)", "_____no_output_____" ], [ "# Double equal sign is also used for comparison\n10.0 == 10", "_____no_output_____" ] ], [ [ "Check the following examples on comparison of two strings.", "_____no_output_____" ] ], [ [ "'cat' < 'dog'", "_____no_output_____" ], [ "# All uppercases are before low cases. \n'cat' < 'Dog'", "_____no_output_____" ], [ "'apple' < 'apricot'", "_____no_output_____" ] ], [ [ "There are three logical operators, *not*, *and* and *or*, which can be applied to the boolean values. ", "_____no_output_____" ] ], [ [ "# Both condition #1 and condition #2 are True?\n3 < 4 and 7 < 8", "_____no_output_____" ], [ "# Either condition 1 or condition 2 are True?\n3 < 4 or 7 > 8", "_____no_output_____" ], [ "# Both conditional #1 and conditional #2 are False?\nnot ((3 > 4) or (7 > 8))", "_____no_output_____" ] ], [ [ "<a id = \"cell_input\"></a>", "_____no_output_____" ], [ "## 2. 4. Input and output", "_____no_output_____" ], [ "All programing languages provide features to interact with user. Python provide *input()* function to get input. It waits for the user to type some input and press return. We can add some information for the user by putting a message inside the function's brackets. It must be a string or a string variable. The text that was typed can be saved in a variable. Here is one example:", "_____no_output_____" ] ], [ [ "nInput = input('Enter you number here:\\n')", "_____no_output_____" ] ], [ [ "However, be aware that the input received from the user are treated as a string, even\nthough a user entered a number. The following **print()** function invokes an error message.", "_____no_output_____" ] ], [ [ "print(nInput + 3)", "_____no_output_____" ] ], [ [ "The input need to be converted to an integer before the match operation can be performed as follows:", "_____no_output_____" ] ], [ [ "print(int(nInput) + 3)", "_____no_output_____" ] ], [ [ "After user's input are accepted, the messages need to be displayed to the user accordingly. String concatenation is one way to display messages which incorporate variable values. ", "_____no_output_____" ] ], [ [ "name = 'David'\nprint('Hello, ' + name)", "_____no_output_____" ] ], [ [ "Another way of achieving this is using **print()** funciton with *string formatting*. We need to use the *string formatting operator*, the percent(**%**) sign. ", "_____no_output_____" ] ], [ [ "name = 'David'\nprint('Hello, %s' % name)", "_____no_output_____" ] ], [ [ "Here is another example with two variables:", "_____no_output_____" ] ], [ [ "name = 'David'\nage = 23\nprint('%s is %d years old.' % (name, age))", "_____no_output_____" ] ], [ [ "Notice that the two variables, **name**, **age**, that specify the values are included at the end of the statement, and enclosed with a bracket. \n\nWith the quotation mark, **%s** and **%d** are used to specify formating for string and integer respectively. \nThe following table shows a selected set of symbols which can be used along with %. ", "_____no_output_____" ], [ "<table width=\"304\" border=\"1\">\n <tr>\n <th width=\"112\" scope=\"col\">Format symbol</th>\n <th width=\"176\" scope=\"col\">Conversion</th>\n </tr>\n <tr>\n <td>%s</td>\n <td>String</td>\n </tr>\n <tr>\n <td>%d</td>\n <td>Signed decimal integer</td>\n </tr>\n <tr>\n <td>%f</td>\n <td>Floating point real number</td>\n </tr>\n</table>", "_____no_output_____" ], [ "There are extra charaters that are used together with above symbols:", "_____no_output_____" ], [ "<table width=\"400\" border=\"1\">\n <tr>\n <th width=\"100\" scope=\"col\">Symbol</th>\n <th width=\"3000\" scope=\"col\">Functionality</th>\n </tr>\n <tr>\n <td>-</td>\n <td>Left justification</td>\n </tr>\n <tr>\n <td>+</td>\n <td>Display the sign</td>\n </tr>\n <tr>\n <td>m.n</td>\n <td>m is the minimum total width; n is the number of digits to display after the decimal point</td>\n </tr>\n</table>", "_____no_output_____" ], [ "Here are more examples that use above specifiers:", "_____no_output_____" ] ], [ [ "# With %f, the format is right justification by default. \n# As a result, white spaces are added to the left of the number\n# 10.4 means minimal width 10 with 4 decinal points\nprint('Output a float number: %10.4f' % (3.5))", "_____no_output_____" ], [ "# plus sign after % means to show positive sign\n# Zero after plus sign means using leading zero to fill width of 5\nprint('Output an integer: %+05d' % (23))", "_____no_output_____" ] ], [ [ "### 2.5 Notes on *Python 2*\n\nYou need to pay attention if you test examples in this prac under *Python* 2. \n\n1. In *Python 3, * **/** is float division, and **//** is integer division; while in Python 2, \n\tboth **/** and **//**\n\tperform *integer division*. \n\tHowever, if you stick to using **float(3)/2** for *float division*, \n\tand **3/2** for *integer division*, \n\tyou will have no problem in both version. \n \n2. Instead using function **input()**, \n\t**raw_input()** is used in Python 2. \n\tBoth functions have the same functionality,\n\ti.e. take what the user typed and passes it back as a string.\n \n3. Although both versions support **print()** function with same format, \n\tPython 2 also allows the print statement (e.g. **print \"Hello, World!\"**), \n\twhich is not valid in Python 3. \n\tHowever, if you stick to our examples and using **print()** function with parantheses, \n\tyour programs should works fine in both versions. ", "_____no_output_____" ] ] ]

[ "markdown", "raw", "markdown", "raw", "markdown", "raw", "markdown", "raw", "markdown", "raw", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown" ]

[ [ "markdown", "markdown", "markdown", "markdown", "markdown" ], [ "raw" ], [ "markdown", "markdown", "markdown" ], [ "raw" ], [ "markdown" ], [ "raw" ], [ "markdown", "markdown" ], [ "raw" ], [ "markdown", "markdown", "markdown", "markdown", "markdown" ], [ "raw" ], [ "markdown", "markdown", "markdown", "markdown", "markdown", "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown", "markdown", "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown", "markdown", "markdown" ], [ "code", "code" ], [ "markdown" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# General Equilibrium", "_____no_output_____" ], [ "This notebook illustrates **how to solve GE equilibrium models**. The example is a simple one-asset model without nominal rigidities.\n\nThe notebook shows how to:\n\n1. Solve for the **stationary equilibrium**.\n2. Solve for (non-linear) **transition paths** using a relaxtion algorithm.\n3. Solve for **transition paths** (linear vs. non-linear) and **impulse-responses** using the **sequence-space method** of **Auclert et. al. (2020)**.", "_____no_output_____" ] ], [ [ "LOAD = False # load stationary equilibrium\nDO_VARY_SIGMA_E = True # effect of uncertainty on stationary equilibrium\nDO_TP_RELAX = True # do transition path with relaxtion", "_____no_output_____" ] ], [ [ "# Setup", "_____no_output_____" ] ], [ [ "%load_ext autoreload\n%autoreload 2\n\nimport time\nimport numpy as np\nimport numba as nb\nfrom scipy import optimize\n\nimport matplotlib.pyplot as plt\nplt.style.use('seaborn-whitegrid')\nprop_cycle = plt.rcParams['axes.prop_cycle']\ncolors = prop_cycle.by_key()['color']\n\nfrom consav.misc import elapsed\n\nfrom GEModel import GEModelClass\nfrom GEModel import solve_backwards, simulate_forwards, simulate_forwards_transpose", "_____no_output_____" ] ], [ [ "## Choose number of threads in numba", "_____no_output_____" ] ], [ [ "import numba as nb\nnb.set_num_threads(8)", "_____no_output_____" ] ], [ [ "# Model", "_____no_output_____" ] ], [ [ "model = GEModelClass('baseline',load=LOAD)\nprint(model)", "Modelclass: GEModelClass\nName: baseline\n\nnamespaces: ['sol', 'sim', 'par']\nother_attrs: []\nsavefolder: saved\nnot_floats: ['Ne', 'Na', 'max_iter_solve', 'max_iter_simulate', 'path_T']\n\nsol:\n a = ndarray with shape = (7, 500) [dtype: float64]\n m = ndarray with shape = (7, 500) [dtype: float64]\n c = ndarray with shape = (7, 500) [dtype: float64]\n Va = ndarray with shape = (7, 500) [dtype: float64]\n i = ndarray with shape = (7, 500) [dtype: int32]\n w = ndarray with shape = (7, 500) [dtype: float64]\n path_a = ndarray with shape = (500, 7, 500) [dtype: float64]\n path_m = ndarray with shape = (500, 7, 500) [dtype: float64]\n path_c = ndarray with shape = (500, 7, 500) [dtype: float64]\n path_Va = ndarray with shape = (500, 7, 500) [dtype: float64]\n path_i = ndarray with shape = (500, 7, 500) [dtype: int32]\n path_w = ndarray with shape = (500, 7, 500) [dtype: float64]\n jac_K = ndarray with shape = (500, 500) [dtype: float64]\n jac_C = ndarray with shape = (500, 500) [dtype: float64]\n jac_curlyK_r = ndarray with shape = (500, 500) [dtype: float64]\n jac_curlyK_w = ndarray with shape = (500, 500) [dtype: float64]\n jac_C_r = ndarray with shape = (500, 500) [dtype: float64]\n jac_C_w = ndarray with shape = (500, 500) [dtype: float64]\n jac_r_K = ndarray with shape = (500, 500) [dtype: float64]\n jac_w_K = ndarray with shape = (500, 500) [dtype: float64]\n jac_r_Z = ndarray with shape = (500, 500) [dtype: float64]\n jac_w_Z = ndarray with shape = (500, 500) [dtype: float64]\n H_K = ndarray with shape = (500, 500) [dtype: float64]\n H_Z = ndarray with shape = (500, 500) [dtype: float64]\n G = ndarray with shape = (500, 500) [dtype: float64]\n memory, gb: 0.1\n\nsim:\n D = ndarray with shape = (7, 500) [dtype: float64]\n path_D = ndarray with shape = (500, 7, 500) [dtype: float64]\n path_K = ndarray with shape = (500,) [dtype: float64]\n path_C = ndarray with shape = (500,) [dtype: float64]\n path_Klag = ndarray with shape = (500,) [dtype: float64]\n memory, gb: 0.0\n\npar:\n r_ss = nan [float]\n w_ss = nan [float]\n K_ss = nan [float]\n Y_ss = nan [float]\n C_ss = nan [float]\n kd_ss = nan [float]\n ks_ss = nan [float]\n sigma = 1.0 [float]\n beta = 0.982 [float]\n Z = 1.0 [float]\n Z_sigma = 0.01 [float]\n Z_rho = 0.9 [float]\n alpha = 0.11 [float]\n delta = 0.025 [float]\n rho = 0.966 [float]\n sigma_e = 0.1 [float]\n Ne = 7 [int]\n a_max = 200.0 [float]\n Na = 500 [int]\n path_T = 500 [int]\n max_iter_solve = 5000 [int]\n max_iter_simulate = 5000 [int]\n solve_tol = 1e-10 [float]\n simulate_tol = 1e-10 [float]\n a_grid = ndarray with shape = (500,) [dtype: float64]\n e_grid = ndarray with shape = (7,) [dtype: float64]\n e_trans = ndarray with shape = (7, 7) [dtype: float64]\n e_ergodic = ndarray with shape = (7,) [dtype: float64]\n e_trans_cumsum = ndarray with shape = (7, 7) [dtype: float64]\n e_ergodic_cumsum = ndarray with shape = (7,) [dtype: float64]\n memory, gb: 0.0\n\n" ] ], [ [ "For easy access", "_____no_output_____" ] ], [ [ "par = model.par\nsim = model.sim\nsol = model.sol", "_____no_output_____" ] ], [ [ "**Productivity states:**", "_____no_output_____" ] ], [ [ "for e,pr_e in zip(par.e_grid,par.e_ergodic):\n print(f'Pr[e = {e:7.4f}] = {pr_e:.4f}')\n \nassert np.isclose(np.sum(par.e_grid*par.e_ergodic),1.0)", "Pr[e = 0.3599] = 0.0156\nPr[e = 0.4936] = 0.0938\nPr[e = 0.6769] = 0.2344\nPr[e = 0.9282] = 0.3125\nPr[e = 1.2729] = 0.2344\nPr[e = 1.7456] = 0.0937\nPr[e = 2.3939] = 0.0156\n" ] ], [ [ "# Find Stationary Equilibrium", "_____no_output_____" ], [ "**Step 1:** Find demand and supply of capital for a grid of interest rates.", "_____no_output_____" ] ], [ [ "if not LOAD:\n \n t0 = time.time()\n \n par = model.par\n\n # a. interest rate trial values\n Nr = 20\n r_vec = np.linspace(0.005,1.0/par.beta-1-0.002,Nr) # 1+r > beta not possible\n\n # b. allocate\n Ks = np.zeros(Nr) \n Kd = np.zeros(Nr)\n\n # c. loop\n r_min = r_vec[0]\n r_max = r_vec[Nr-1]\n for i_r in range(Nr):\n\n # i. firm side\n k = model.firm_demand(r_vec[i_r],par.Z)\n Kd[i_r] = k*1 # aggregate labor = 1.0\n \n # ii. household side\n success = model.solve_household_ss(r=r_vec[i_r])\n if success: \n success = model.simulate_household_ss()\n\n if success:\n\n # total demand\n Ks[i_r] = np.sum(model.sim.D*model.sol.a)\n\n # bounds on r\n diff = Ks[i_r]-Kd[i_r]\n if diff < 0: r_min = np.fmax(r_min,r_vec[i_r])\n if diff > 0: r_max = np.fmin(r_max,r_vec[i_r])\n\n else:\n\n Ks[i_r] = np.nan \n \n # d. save\n model.save()\n \n print(f'grid search done in {elapsed(t0)}')", "grid search done in 10.8 secs\n" ] ], [ [ "**Step 2:** Plot supply and demand.", "_____no_output_____" ] ], [ [ "if not LOAD:\n \n par = model.par\n\n fig = plt.figure(figsize=(6,4))\n ax = fig.add_subplot(1,1,1)\n\n ax.plot(r_vec,Ks,label='supply of capital')\n ax.plot(r_vec,Kd,label='demand for capital')\n\n ax.axvline(r_min,lw=0.5,ls='--',color='black')\n ax.axvline(r_max,lw=0.5,ls='--',color='black')\n\n ax.legend(frameon=True)\n ax.set_xlabel('interest rate, $r$')\n ax.set_ylabel('capital, $K_t$')\n \n fig.tight_layout()\n fig.savefig('figs/stationary_equilibrium.pdf')\n ", "_____no_output_____" ] ], [ [ "**Step 3:** Solve root-finding problem.", "_____no_output_____" ] ], [ [ "def obj(r,model):\n \n model.solve_household_ss(r=r)\n model.simulate_household_ss()\n return np.sum(model.sim.D*model.sol.a)-model.firm_demand(r,model.par.Z)\n\nif not LOAD:\n \n t0 = time.time()\n \n opt = optimize.root_scalar(obj,bracket=[r_min,r_max],method='bisect',args=(model,))\n model.par.r_ss = opt.root\n assert opt.converged\n \n print(f'search done in {elapsed(t0)}')\n ", "search done in 7.1 secs\n" ] ], [ [ "**Step 4:** Check market clearing conditions.", "_____no_output_____" ] ], [ [ "model.steady_state()", "household problem solved in 0.1 secs [652 iterations]\nhousehold problem simulated in 0.1 secs [747 iterations]\n\nr: 0.0127\nw: 1.0160\nY: 1.1415\nK/Y: 2.9186\n\ncapital market clearing: -0.00000000\ngoods market clearing: -0.00000755\n" ] ], [ [ "## Timings", "_____no_output_____" ] ], [ [ "%timeit model.solve_household_ss(r=par.r_ss)", "119 ms ± 8.83 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)\n" ], [ "%timeit model.simulate_household_ss()", "66.2 ms ± 1.42 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)\n" ] ], [ [ "## Income uncertainty and the equilibrium interest rate", "_____no_output_____" ], [ "The equlibrium interest rate decreases when income uncertainty is increased.", "_____no_output_____" ] ], [ [ "if DO_VARY_SIGMA_E:\n \n par = model.par\n\n # a. seetings\n sigma_e_vec = [0.20]\n\n # b. find equilibrium rates\n model_ = model.copy()\n for sigma_e in sigma_e_vec:\n\n # i. set new parameter\n model_.par.sigma_e = sigma_e\n model_.create_grids()\n \n # ii. solve\n print(f'sigma_e = {sigma_e:.4f}',end='')\n opt = optimize.root_scalar(\n obj,\n bracket=[0.00,model.par.r_ss],\n method='bisect',\n args=(model_,)\n )\n print(f' -> r_ss = {opt.root:.4f}')\n model_.par.r_ss = opt.root\n model_.steady_state() \n print('\\n')\n ", "sigma_e = 0.2000 -> r_ss = 0.0029\nhousehold problem solved in 0.1 secs [430 iterations]\nhousehold problem simulated in 0.0 secs [427 iterations]\n\nr: 0.0029\nw: 1.0546\nY: 1.1849\nK/Y: 3.9462\n\ncapital market clearing: -0.00000000\ngoods market clearing: -0.00000587\n\n\n" ] ], [ [ "## Test matrix formulation", "_____no_output_____" ], [ "**Step 1:** Construct $\\boldsymbol{Q}_{ss}$", "_____no_output_____" ] ], [ [ "# a. allocate Q\nQ = np.zeros((par.Ne*par.Na,par.Ne*par.Na))\n\n# b. fill\nfor i_e in range(par.Ne):\n \n # get view of current block\n q = Q[i_e*par.Na:(i_e+1)*par.Na,i_e*par.Na:(i_e+1)*par.Na]\n \n for i_a in range(par.Na):\n \n # i. optimal choice\n a_opt = sol.a[i_e,i_a]\n \n # ii. above -> all weight on last node\n if a_opt >= par.a_grid[-1]:\n \n q[i_a,-1] = 1.0 \n \n # iii. below -> all weight on first node\n elif a_opt <= par.a_grid[0]:\n \n q[i_a,0] = 1.0\n \n # iv. standard -> distribute weights on neighboring nodes\n else: \n \n i_a_low = np.searchsorted(par.a_grid,a_opt,side='right')-1 \n \n assert a_opt >= par.a_grid[i_a_low], f'{a_opt} < {par.a_grid[i_a_low]}'\n assert a_opt < par.a_grid[i_a_low+1], f'{a_opt} < {par.a_grid[i_a_low]}'\n \n q[i_a,i_a_low] = (par.a_grid[i_a_low+1]-a_opt)/(par.a_grid[i_a_low+1]-par.a_grid[i_a_low])\n q[i_a,i_a_low+1] = 1-q[i_a,i_a_low]\n ", "_____no_output_____" ] ], [ [ "**Step 2:** Construct $\\tilde{\\Pi}^e=\\Pi^e \\otimes \\boldsymbol{I}_{\\#_{a}\\times\\#_{a}}$", "_____no_output_____" ] ], [ [ "Pit = np.kron(par.e_trans,np.identity(par.Na))", "_____no_output_____" ] ], [ [ "**Step 3:** Test $\\overrightarrow{D}_{t+1}=\\tilde{\\Pi}^{e\\prime}\\boldsymbol{Q}_{ss}^{\\prime}\\overrightarrow{D}_{t}$", "_____no_output_____" ] ], [ [ "D = np.zeros(sim.D.shape)\nD[:,0] = par.e_ergodic \n\n# a. standard\nD_plus = np.zeros(D.shape)\nsimulate_forwards(D,sol.i,sol.w,par.e_trans.T.copy(),D_plus)\n\n# b. matrix product\nD_plus_alt = ((Pit.T@Q.T)@D.ravel()).reshape((par.Ne,par.Na))\n\n# c. test equality\nassert np.allclose(D_plus,D_plus_alt)", "_____no_output_____" ] ], [ [ "# Find transition path", "_____no_output_____" ], [ "**MIT-shock:** Transtion path for arbitrary exogenous path of $Z_t$ starting from the stationary equilibrium, i.e. $D_{-1} = D_{ss}$ and in particular $K_{-1} = K_{ss}$.", "_____no_output_____" ], [ "**Step 1:** Construct $\\{Z_t\\}_{t=0}^{T-1}$ where $Z_t = (1-\\rho_Z)Z_{ss} + \\rho_Z Z_t$ and $Z_0 = (1+\\sigma_Z) Z_{ss}$", "_____no_output_____" ] ], [ [ "path_Z = model.get_path_Z()", "_____no_output_____" ] ], [ [ "**Step 2:** Apply relaxation algorithm.", "_____no_output_____" ] ], [ [ "if DO_TP_RELAX:\n \n t0 = time.time()\n\n # a. allocate\n path_r = np.repeat(model.par.r_ss,par.path_T) # use steady state as initial guess\n path_r_ = np.zeros(par.path_T)\n path_w = np.zeros(par.path_T)\n\n # b. setting\n nu = 0.90 # relaxation parameter\n max_iter = 5000 # maximum number of iterations\n\n # c. iterate\n it = 0\n while True:\n\n # i. find wage \n for t in range(par.path_T):\n path_w[t] = model.implied_w(path_r[t],path_Z[t])\n\n # ii. solve and simulate\n model.solve_household_path(path_r,path_w)\n model.simulate_household_path(model.sim.D)\n\n # iii. implied prices\n for t in range(par.path_T):\n path_r_[t] = model.implied_r(sim.path_Klag[t],path_Z[t])\n\n # iv. difference\n max_abs_diff = np.max(np.abs(path_r-path_r_))\n if it%10 == 0: print(f'{it:4d}: {max_abs_diff:.8f}')\n if max_abs_diff < 1e-8: break\n\n # v. update\n path_r = nu*path_r + (1-nu)*path_r_\n\n # vi. increment\n it += 1\n if it > max_iter: raise Exception('too many iterations') \n\n print(f'\\n transtion path found in {elapsed(t0)}')", " 0: 0.00038764\n 10: 0.00013139\n 20: 0.00004581\n 30: 0.00001597\n 40: 0.00000557\n 50: 0.00000194\n 60: 0.00000068\n 70: 0.00000024\n 80: 0.00000008\n 90: 0.00000003\n 100: 0.00000001\n\n transtion path found in 23.7 secs\n" ] ], [ [ "**Plot transition-paths:**", "_____no_output_____" ] ], [ [ "if DO_TP_RELAX:\n \n fig = plt.figure(figsize=(10,6))\n\n ax = fig.add_subplot(2,2,1)\n ax.plot(np.arange(par.path_T),path_Z,'-o',ms=2)\n ax.set_title('technology, $Z_t$');\n\n ax = fig.add_subplot(2,2,2)\n ax.plot(np.arange(par.path_T),sim.path_K,'-o',ms=2)\n ax.set_title('capital, $k_t$');\n\n ax = fig.add_subplot(2,2,3)\n ax.plot(np.arange(par.path_T),path_r,'-o',ms=2)\n ax.set_title('interest rate, $r_t$');\n\n ax = fig.add_subplot(2,2,4)\n ax.plot(np.arange(par.path_T),path_w,'-o',ms=2)\n ax.set_title('wage, $w_t$')\n\n fig.tight_layout()\n fig.savefig('figs/transition_path.pdf')\n ", "_____no_output_____" ] ], [ [ "**Remember:**", "_____no_output_____" ] ], [ [ "if DO_TP_RELAX:\n \n path_Z_relax = path_Z\n path_K_relax = sim.path_K\n path_r_relax = path_r\n path_w_relax = path_w\n ", "_____no_output_____" ] ], [ [ "# Find impulse-responses using sequence-space method", "_____no_output_____" ], [ "**Paper:** Auclert, A., Bardóczy, B., Rognlie, M., and Straub, L. (2020). *Using the Sequence-Space Jacobian to Solve and Estimate Heterogeneous-Agent Models*. \n\n**Original code:** [shade-econ](https://github.com/shade-econ/sequence-jacobian/#sequence-space-jacobian)\n\n**This code:** Illustrates the sequence-space method. The original paper shows how to do it computationally efficient and for a general class of models.", "_____no_output_____" ], [ "**Step 1:** Compute the Jacobian for the household block around the stationary equilibrium", "_____no_output_____" ] ], [ [ "def jac(model,price,dprice=1e-4,do_print=True):\n \n t0_all = time.time()\n \n if do_print: print(f'price is {price}')\n \n par = model.par\n sol = model.sol\n sim = model.sim\n \n # a. step 1: solve backwards\n t0 = time.time()\n\n path_r = np.repeat(par.r_ss,par.path_T)\n path_w = np.repeat(par.w_ss,par.path_T)\n \n if price == 'r': path_r[-1] += dprice\n elif price == 'w': path_w[-1] += dprice\n \n model.solve_household_path(path_r,path_w,do_print=False)\n if do_print: print(f'solved backwards in {elapsed(t0)}')\n\n # b. step 2: derivatives\n t0 = time.time()\n \n diff_Ds = np.zeros((par.path_T,*sim.D.shape))\n diff_as = np.zeros(par.path_T)\n diff_cs = np.zeros(par.path_T)\n \n for s in range(par.path_T):\n \n t_ =(par.path_T-1)-s\n simulate_forwards(sim.D,sol.path_i[t_],sol.path_w[t_],par.e_trans.T,diff_Ds[s])\n \n diff_Ds[s] = (diff_Ds[s]-sim.D)/dprice\n diff_as[s] = (np.sum(sol.path_a[t_]*sim.D)-np.sum(sol.a*sim.D))/dprice \n diff_cs[s] = (np.sum(sol.path_c[t_]*sim.D)-np.sum(sol.c*sim.D))/dprice \n \n if do_print: print(f'derivatives calculated in {elapsed(t0)}')\n \n # c. step 3: expectation factors\n t0 = time.time()\n \n # demeaning improves numerical stability\n def demean(x):\n return x - x.sum()/x.size\n\n exp_as = np.zeros((par.path_T-1,*sol.a.shape))\n exp_as[0] = demean(sol.a)\n\n exp_cs = np.zeros((par.path_T-1,*sol.c.shape))\n exp_cs[0] = demean(sol.c)\n \n for t in range(1,par.path_T-1):\n \n simulate_forwards_transpose(exp_as[t-1],sol.i,sol.w,par.e_trans,exp_as[t])\n exp_as[t] = demean(exp_as[t])\n \n simulate_forwards_transpose(exp_cs[t-1],sol.i,sol.w,par.e_trans,exp_cs[t])\n exp_cs[t] = demean(exp_cs[t])\n \n if do_print: print(f'expecation factors calculated in {elapsed(t0)}')\n \n # d. step 4: F \n t0 = time.time()\n \n Fa = np.zeros((par.path_T,par.path_T))\n Fa[0,:] = diff_as\n \n Fc = np.zeros((par.path_T,par.path_T))\n Fc[0,:] = diff_cs \n \n Fa[1:, :] = exp_as.reshape((par.path_T-1, -1)) @ diff_Ds.reshape((par.path_T, -1)).T\n Fc[1:, :] = exp_cs.reshape((par.path_T-1, -1)) @ diff_Ds.reshape((par.path_T, -1)).T \n\n if do_print: print(f'f calculated in {elapsed(t0)}')\n \n t0 = time.time()\n \n # e. step 5: J\n Ja = Fa.copy()\n for t in range(1, Ja.shape[1]): Ja[1:, t] += Ja[:-1, t - 1]\n\n Jc = Fc.copy()\n for t in range(1, Jc.shape[1]): Jc[1:, t] += Jc[:-1, t - 1]\n\n if do_print: print(f'J calculated in {elapsed(t0)}')\n \n # f. save\n setattr(model.sol,f'jac_curlyK_{price}',Ja)\n setattr(model.sol,f'jac_C_{price}',Jc)\n \n if do_print: print(f'full Jacobian calculated in {elapsed(t0_all)}\\n')", "_____no_output_____" ], [ "jac(model,'r')\njac(model,'w')", "price is r\nsolved backwards in 0.2 secs\nderivatives calculated in 2.0 secs\nexpecation factors calculated in 0.8 secs\nf calculated in 0.1 secs\nJ calculated in 0.0 secs\nfull Jacobian calculated in 3.1 secs\n\nprice is w\nsolved backwards in 0.1 secs\nderivatives calculated in 0.2 secs\nexpecation factors calculated in 0.1 secs\nf calculated in 0.1 secs\nJ calculated in 0.0 secs\nfull Jacobian calculated in 0.5 secs\n\n" ] ], [ [ "**Inspect Jacobians:**", "_____no_output_____" ] ], [ [ "fig = plt.figure(figsize=(12,8))\nT_fig = 200\n\n# curlyK_r\nax = fig.add_subplot(2,2,1)\nfor s in [0,25,50,75,100]:\n ax.plot(np.arange(T_fig),sol.jac_curlyK_r[s,:T_fig],'-o',ms=2,label=f'$s={s}$')\n \nax.legend(frameon=True)\nax.set_title(r'$\\mathcal{J}^{\\mathcal{K},r}$')\nax.set_xlim([0,T_fig])\n\n# curlyK_w\nax = fig.add_subplot(2,2,2)\nfor s in [0,25,50,75,100]:\n ax.plot(np.arange(T_fig),sol.jac_curlyK_w[s,:T_fig],'-o',ms=2)\n \nax.set_title(r'$\\mathcal{J}^{\\mathcal{K},w}$')\nax.set_xlim([0,T_fig])\n\n# C_r\nax = fig.add_subplot(2,2,3)\nfor s in [0,25,50,75,100]:\n ax.plot(np.arange(T_fig),sol.jac_C_r[s,:T_fig],'-o',ms=2,label=f'$s={s}$')\n \nax.legend(frameon=True)\nax.set_title(r'$\\mathcal{J}^{C,r}$')\nax.set_xlim([0,T_fig])\n\n# curlyK_w\nax = fig.add_subplot(2,2,4)\nfor s in [0,25,50,75,100]:\n ax.plot(np.arange(T_fig),sol.jac_C_w[s,:T_fig],'-o',ms=2)\n \nax.set_title(r'$\\mathcal{J}^{C,w}$')\nax.set_xlim([0,T_fig])\n\nfig.tight_layout()\nfig.savefig('figs/jacobians.pdf')", "_____no_output_____" ] ], [ [ "**Step 2:** Compute the Jacobians for the firm block around the stationary equilibrium (analytical).", "_____no_output_____" ] ], [ [ "sol.jac_r_K[:] = 0\nsol.jac_w_K[:] = 0\nsol.jac_r_Z[:] = 0\nsol.jac_w_Z[:] = 0\n\nfor s in range(par.path_T):\n for t in range(par.path_T):\n\n if t == s+1:\n\n sol.jac_r_K[t,s] = par.alpha*(par.alpha-1)*par.Z*par.K_ss**(par.alpha-2)\n sol.jac_w_K[t,s] = (1-par.alpha)*par.alpha*par.Z*par.K_ss**(par.alpha-1)\n\n if t == s:\n sol.jac_r_Z[t,s] = par.alpha*par.Z*par.K_ss**(par.alpha-1)\n sol.jac_w_Z[t,s] = (1-par.alpha)*par.Z*par.K_ss**par.alpha\n", "_____no_output_____" ] ], [ [ "**Step 3:** Use the chain rule and solve for $G$.", "_____no_output_____" ] ], [ [ "H_K = sol.jac_curlyK_r @ sol.jac_r_K + sol.jac_curlyK_w @ sol.jac_w_K - np.eye(par.path_T)\nH_Z = sol.jac_curlyK_r @ sol.jac_r_Z + sol.jac_curlyK_w @ sol.jac_w_Z \nG_K_Z = -np.linalg.solve(H_K, H_Z) # H_K^(-1)H_Z", "_____no_output_____" ] ], [ [ "**Step 4:** Find effect on prices and other outcomes than $K$.", "_____no_output_____" ] ], [ [ "G_r_Z = sol.jac_r_Z + sol.jac_r_K@G_K_Z\nG_w_Z = sol.jac_w_Z + sol.jac_w_K@G_K_Z\nG_C_Z = sol.jac_C_r@G_r_Z + sol.jac_C_w@G_w_Z", "_____no_output_____" ] ], [ [ "**Step 5:** Plot impulse-responses.", "_____no_output_____" ], [ "**Example I:** News shock (i.e. in a single period) vs. persistent shock where $ dZ_t = \\rho dZ_{t-1} $ and $dZ_0$ is the initial shock.", "_____no_output_____" ] ], [ [ "fig = plt.figure(figsize=(12,4))\nT_fig = 50\n\n# left: news shock\nax = fig.add_subplot(1,2,1)\nfor s in [5,10,15,20,25]:\n dZ = (1+par.Z_sigma)*par.Z*(np.arange(par.path_T) == s)\n dK = G_K_Z@dZ \n ax.plot(np.arange(T_fig),dK[:T_fig],'-o',ms=2,label=f'$s={s}$')\n\nax.legend(frameon=True)\nax.set_title(r'1% TFP news shock in period $s$')\nax.set_ylabel('$K_t-K_{ss}$')\nax.set_xlim([0,T_fig])\n\n# right: persistent shock\nax = fig.add_subplot(1,2,2)\ndZ = model.get_path_Z()-par.Z\ndK = G_K_Z@dZ\nax.plot(np.arange(T_fig),dK[:T_fig],'-o',ms=2)\n \nax.set_title(r'1% TFP shock with persistence $\\rho=0.90$')\nax.set_ylabel('$K_t-K_{ss}$')\nax.set_xlim([0,T_fig])\n\nfig.tight_layout()\nfig.savefig('figs/news_vs_persistent_shock.pdf')", "_____no_output_____" ] ], [ [ "**Example II:** Further effects of persistent shock.", "_____no_output_____" ] ], [ [ "fig = plt.figure(figsize=(12,8))\nT_fig = 50\n\nax_K = fig.add_subplot(2,2,1)\nax_r = fig.add_subplot(2,2,2)\nax_w = fig.add_subplot(2,2,3)\nax_C = fig.add_subplot(2,2,4)\n\nax_K.set_title('$K_t-K_{ss}$ after 1% TFP shock')\nax_K.set_xlim([0,T_fig])\n \nax_r.set_title('$r_t-r_{ss}$ after 1% TFP shock')\nax_r.set_xlim([0,T_fig])\n\nax_w.set_title('$w_t-w_{ss}$ after 1% TFP shock')\nax_w.set_xlim([0,T_fig])\n\nax_C.set_title('$C_t-C_{ss}$ after 1% TFP shock')\nax_C.set_xlim([0,T_fig])\n\ndZ = model.get_path_Z()-par.Z\n \ndK = G_K_Z@dZ\nax_K.plot(np.arange(T_fig),dK[:T_fig],'-o',ms=2)\n\ndr = G_r_Z@dZ \nax_r.plot(np.arange(T_fig),dr[:T_fig],'-o',ms=2)\n\ndw = G_w_Z@dZ \nax_w.plot(np.arange(T_fig),dw[:T_fig],'-o',ms=2)\n\ndC = G_C_Z@dZ\nax_C.plot(np.arange(T_fig),dC[:T_fig],'-o',ms=2)\n \nfig.tight_layout()\nfig.savefig('figs/irfs.pdf')", "_____no_output_____" ] ], [ [ "## Non-linear transition path", "_____no_output_____" ], [ "Use the Jacobian to speed-up solving for the non-linear transition path using a quasi-Newton method.", "_____no_output_____" ], [ "**1. Solver**", "_____no_output_____" ] ], [ [ "def broyden_solver(f,x0,jac,tol=1e-8,max_iter=100,backtrack_fac=0.5,max_backtrack=30,do_print=False):\n \"\"\" numerical solver using the broyden method \"\"\"\n\n # a. initial\n x = x0.ravel()\n y = f(x)\n\n # b. iterate\n for it in range(max_iter):\n \n # i. current difference\n abs_diff = np.max(np.abs(y))\n if do_print: print(f' it = {it:3d} -> max. abs. error = {abs_diff:12.8f}')\n\n if abs_diff < tol: return x\n \n # ii. new x\n dx = np.linalg.solve(jac,-y)\n \n # iii. evalute with backtrack\n for _ in range(max_backtrack):\n\n try: # evaluate\n ynew = f(x+dx)\n except ValueError: # backtrack\n dx *= backtrack_fac\n else: # update jac and break from backtracking\n dy = ynew-y\n jac = jac + np.outer(((dy - jac @ dx) / np.linalg.norm(dx) ** 2), dx)\n y = ynew\n x += dx\n break\n\n else:\n\n raise ValueError('too many backtracks, maybe bad initial guess?')\n \n else:\n\n raise ValueError(f'no convergence after {max_iter} iterations') ", "_____no_output_____" ] ], [ [ "**2. Target function**", "_____no_output_____" ], [ "$$\\boldsymbol{H}(\\boldsymbol{K},\\boldsymbol{Z},D_{ss}) = \\mathcal{K}_{t}(\\{r(Z_{s},K_{s-1}),w(Z_{s},K_{s-1})\\}_{s\\geq0},D_{ss})-K_{t}=0$$", "_____no_output_____" ] ], [ [ "def target(path_K,path_Z,model,D0,full_output=False):\n\n par = model.par\n sim = model.sim\n \n path_r = np.zeros(path_K.size)\n path_w = np.zeros(path_K.size)\n\n # a. implied prices\n K0lag = np.sum(par.a_grid[np.newaxis,:]*D0)\n path_Klag = np.insert(path_K,0,K0lag)\n \n for t in range(par.path_T):\n path_r[t] = model.implied_r(path_Klag[t],path_Z[t])\n path_w[t] = model.implied_w(path_r[t],path_Z[t])\n \n # b. solve and simulate\n model.solve_household_path(path_r,path_w)\n model.simulate_household_path(D0)\n \n # c. market clearing\n if full_output:\n return path_r,path_w\n else:\n return sim.path_K-path_K\n", "_____no_output_____" ] ], [ [ "**3. Solve**", "_____no_output_____" ] ], [ [ "path_Z = model.get_path_Z()\nf = lambda x: target(x,path_Z,model,sim.D)", "_____no_output_____" ], [ "t0 = time.time()\npath_K = broyden_solver(f,x0=np.repeat(par.K_ss,par.path_T),jac=H_K,do_print=True)\npath_r,path_w = target(path_K,path_Z,model,sim.D,full_output=True)\nprint(f'\\nIRF found in {elapsed(t0)}')", " it = 0 -> max. abs. error = 0.05898269\n it = 1 -> max. abs. error = 0.00026075\n it = 2 -> max. abs. error = 0.00000163\n it = 3 -> max. abs. error = 0.00000000\n\nIRF found in 2.1 secs\n" ] ], [ [ "**4. Plot**", "_____no_output_____" ] ], [ [ "fig = plt.figure(figsize=(12,4))\n\nax = fig.add_subplot(1,2,1)\nax.set_title('capital, $K_t$')\ndK = G_K_Z@(path_Z-par.Z)\nax.plot(np.arange(T_fig),dK[:T_fig] + par.K_ss,'-o',ms=2,label=f'linear')\nax.plot(np.arange(T_fig),path_K[:T_fig],'-o',ms=2,label=f'non-linear')\nif DO_TP_RELAX:\n ax.plot(np.arange(T_fig),path_K_relax[:T_fig],'--o',ms=2,label=f'non-linear (relaxtion)')\n\nax.legend(frameon=True) \n \nax = fig.add_subplot(1,2,2)\nax.set_title('interest rate, $r_t$')\ndr = G_r_Z@(path_Z-par.Z)\nax.plot(np.arange(T_fig),dr[:T_fig] + par.r_ss,'-o',ms=2,label=f'linear')\nax.plot(np.arange(T_fig),path_r[:T_fig],'-o',ms=2,label=f'non-linear')\nif DO_TP_RELAX:\n ax.plot(np.arange(T_fig),path_r_relax[:T_fig],'--o',ms=2,label=f'non-linear (relaxtion)')\n \nfig.tight_layout()\nfig.savefig('figs/non_linear.pdf')", "_____no_output_____" ] ], [ [ "## Covariances", "_____no_output_____" ], [ "Assume that $Z_t$ is stochastic and follows\n\n$$ d\\tilde{Z}_t = \\rho d\\tilde{Z}_{t-1} + \\sigma\\epsilon_t,\\,\\,\\, \\epsilon_t \\sim \\mathcal{N}(0,1) $$\n\nThe covariances between all outcomes can be calculated as follows.", "_____no_output_____" ] ], [ [ "# a. choose parameter\nrho = 0.90\nsigma = 0.10\n\n# b. find change in outputs\ndZ = rho**(np.arange(par.path_T))\ndC = G_C_Z@dZ\ndK = G_K_Z@dZ\n\n# c. covariance of consumption\nprint('auto-covariance of consumption:\\n')\nfor k in range(5):\n if k == 0:\n autocov_C = sigma**2*np.sum(dC*dC)\n else:\n autocov_C = sigma**2*np.sum(dC[:-k]*dC[k:])\n print(f' k = {k}: {autocov_C:.4f}')\n \n# d. covariance of consumption and capital\ncov_C_K = sigma**2*np.sum(dC*dK)\nprint(f'\\ncovariance of consumption and capital: {cov_C_K:.4f}')", "auto-covariance of consumption:\n\n k = 0: 0.0445\n k = 1: 0.0431\n k = 2: 0.0415\n k = 3: 0.0399\n k = 4: 0.0382\n\ncovariance of consumption and capital: 0.2117\n" ] ], [ [ "# Extra: No idiosyncratic uncertainty", "_____no_output_____" ], [ "This section solve for the transition path in the case without idiosyncratic uncertainty.", "_____no_output_____" ], [ "**Analytical solution for steady state:**", "_____no_output_____" ] ], [ [ "r_ss_pf = (1/par.beta-1) # from euler-equation\nw_ss_pf = model.implied_w(r_ss_pf,par.Z)\nK_ss_pf = model.firm_demand(r_ss_pf,par.Z)\nY_ss_pf = model.firm_production(K_ss_pf,par.Z)\nC_ss_pf = Y_ss_pf-par.delta*K_ss_pf\n\nprint(f'r: {r_ss_pf:.6f}')\nprint(f'w: {w_ss_pf:.6f}')\nprint(f'Y: {Y_ss_pf:.6f}')\nprint(f'C: {C_ss_pf:.6f}')\nprint(f'K/Y: {K_ss_pf/Y_ss_pf:.6f}')", "r: 0.018330\nw: 0.998613\nY: 1.122037\nC: 1.050826\nK/Y: 2.538660\n" ] ], [ [ "**Function for finding consumption and capital paths given paths of interest rates and wages:**\n\nIt can be shown that\n\n$$ C_{0}=\\frac{(1+r_{0})a_{-1}+\\sum_{t=0}^{\\infty}\\frac{1}{\\mathcal{R}_{t}}w_{t}}{\\sum_{t=0}^{\\infty}\\beta^{t/\\rho}\\mathcal{R}_{t}^{\\frac{1-\\rho}{\\rho}}} $$\n\nwhere \n\n$$ \\mathcal{R}_{t} =\\begin{cases} 1 & \\text{if }t=0\\\\ (1+r_{t})\\mathcal{R}_{t-1} & \\text{else} \\end{cases} $$\n\nOtherwise the **Euler-equation** holds\n\n$$ C_t = (\\beta (1+r_{t}))^{\\frac{1}{\\sigma}}C_{t-1} $$", "_____no_output_____" ] ], [ [ "def path_CK_func(K0,path_r,path_w,r_ss,w_ss,model):\n \n par = model.par\n \n # a. initialize\n wealth = (1+path_r[0])*K0\n inv_MPC = 0\n \n # b. solve\n RT = 1\n max_iter = 5000\n t = 0\n while True and t < max_iter: \n \n # i. prices padded with steady state\n r = path_r[t] if t < par.path_T else r_ss\n w = path_w[t] if t < par.path_T else w_ss\n \n # ii. interest rate factor \n if t == 0:\n fac = 1\n else:\n fac *= (1+r)\n \n # iii. accumulate\n add_wealth = w/fac\n add_inv_MPC = par.beta**(t/par.sigma)*fac**((1-par.sigma)/par.sigma)\n if np.fmax(add_wealth,add_inv_MPC) < 1e-12:\n break\n else:\n wealth += add_wealth\n inv_MPC += add_inv_MPC\n \n # iv. increment\n t += 1\n \n # b. simulate\n path_C = np.empty(par.path_T)\n path_K = np.empty(par.path_T)\n \n for t in range(par.path_T):\n \n if t == 0:\n path_C[t] = wealth/inv_MPC\n K_lag = K0\n else:\n path_C[t] = (par.beta*(1+path_r[t]))**(1/par.sigma)*path_C[t-1]\n K_lag = path_K[t-1]\n \n path_K[t] = (1+path_r[t])*K_lag + path_w[t] - path_C[t]\n \n return path_K,path_C", "_____no_output_____" ] ], [ [ "**Test with steady state prices:**", "_____no_output_____" ] ], [ [ "path_r_pf = np.repeat(r_ss_pf,par.path_T)\npath_w_pf = np.repeat(w_ss_pf,par.path_T)\npath_K_pf,path_C_pf = path_CK_func(K_ss_pf,path_r_pf,path_w_pf,r_ss_pf,w_ss_pf,model)", "_____no_output_____" ], [ "print(f'C_ss: {C_ss_pf:.6f}')\nprint(f'C[0]: {path_C_pf[0]:.6f}')\nprint(f'C[-1]: {path_C_pf[-1]:.6f}')\nassert np.isclose(C_ss_pf,path_C_pf[0])", "C_ss: 1.050826\nC[0]: 1.050826\nC[-1]: 1.050826\n" ] ], [ [ "**Shock paths** where interest rate deviate in one period:", "_____no_output_____" ] ], [ [ "dr = 1e-4\nts = np.array([0,20,40])\npath_C_pf_shock = np.empty((ts.size,par.path_T))\npath_K_pf_shock = np.empty((ts.size,par.path_T))\n\nfor i,t in enumerate(ts):\n \n path_r_pf_shock = path_r_pf.copy()\n path_r_pf_shock[t] += dr\n \n K,C = path_CK_func(K_ss_pf,path_r_pf_shock,path_w_pf,r_ss_pf,w_ss_pf,model)\n \n path_K_pf_shock[i,:] = K\n path_C_pf_shock[i,:] = C", "_____no_output_____" ] ], [ [ "**Plot paths:**", "_____no_output_____" ] ], [ [ "fig = plt.figure(figsize=(12,4))\n\nax = fig.add_subplot(1,2,1)\nax.plot(np.arange(par.path_T),path_C_pf,'-o',ms=2,label=f'$r_t = r^{{\\\\ast}}$')\nfor i,t in enumerate(ts):\n ax.plot(np.arange(par.path_T),path_C_pf_shock[i],'-o',ms=2,label=f'shock to $r_{{{t}}}$')\n \nax.set_xlim([0,50])\nax.set_xlabel('periods')\nax.set_ylabel('consumtion, $C_t$');\n\nax = fig.add_subplot(1,2,2)\nax.plot(np.arange(par.path_T),path_K_pf,'-o',ms=2,label=f'$r_t = r^{{\\\\ast}}$')\nfor i,t in enumerate(ts):\n ax.plot(np.arange(par.path_T),path_K_pf_shock[i],'-o',ms=2,label=f'shock to $r_{{{t}}}$')\n \nax.legend(frameon=True)\nax.set_xlim([0,50])\nax.set_xlabel('$t$')\nax.set_ylabel('capital, $K_t$');\n\nfig.tight_layout()", "_____no_output_____" ] ], [ [ "**Find transition path with shooting algorithm:**", "_____no_output_____" ] ], [ [ "# a. allocate\ndT = 200\npath_C_pf = np.empty(par.path_T)\npath_K_pf = np.empty(par.path_T)\npath_r_pf = np.empty(par.path_T)\npath_w_pf = np.empty(par.path_T)\n\n# b. settings\nC_min = C_ss_pf\nC_max = C_ss_pf + K_ss_pf\n\nK_min = 1.5 # guess on lower consumption if below this\nK_max = 3 # guess on higher consumption if above this\n\ntol_pf = 1e-6\nmax_iter_pf = 5000\n\npath_K_pf[0] = K_ss_pf # capital is pre-determined\n\n# c. iterate\nt = 0\nit = 0\nwhile True:\n \n # i. update prices\n path_r_pf[t] = model.implied_r(path_K_pf[t],path_Z[t])\n path_w_pf[t] = model.implied_w(path_r_pf[t],path_Z[t])\n\n # ii. consumption \n if t == 0:\n C0 = (C_min+C_max)/2\n path_C_pf[t] = C0\n else:\n path_C_pf[t] = (1+path_r_pf[t])*par.beta*path_C_pf[t-1]\n \n # iii. check for steady state\n if path_K_pf[t] < K_min:\n t = 0\n C_max = C0\n continue\n elif path_K_pf[t] > K_max:\n t = 0\n C_min = C0\n continue\n elif t > 10 and np.sqrt((path_C_pf[t]-C_ss_pf)**2+(path_K_pf[t]-K_ss_pf)**2) < tol_pf:\n path_C_pf[t:] = path_C_pf[t]\n path_K_pf[t:] = path_K_pf[t]\n for k in range(par.path_T):\n path_r_pf[k] = model.implied_r(path_K_pf[k],path_Z[k])\n path_w_pf[k] = model.implied_w(path_r_pf[k],path_Z[k])\n break\n \n # iv. update capital\n path_K_pf[t+1] = (1+path_r_pf[t])*path_K_pf[t] + path_w_pf[t] - path_C_pf[t] \n \n # v. increment\n t += 1\n it += 1\n if it > max_iter_pf: break ", "_____no_output_____" ] ], [ [ "**Plot deviations from steady state:**", "_____no_output_____" ] ], [ [ "fig = plt.figure(figsize=(12,8))\n\nax = fig.add_subplot(2,2,1)\nax.plot(np.arange(par.path_T),path_Z,'-o',ms=2)\nax.set_xlim([0,200])\nax.set_title('technology, $Z_t$')\n\nax = fig.add_subplot(2,2,2)\nax.plot(np.arange(par.path_T),path_K-model.par.kd_ss,'-o',ms=2,label='$\\sigma_e = 0.5$')\nax.plot(np.arange(par.path_T),path_K_pf-K_ss_pf,'-o',ms=2,label='$\\sigma_e = 0$')\nax.legend(frameon=True)\nax.set_title('capital, $k_t$')\nax.set_xlim([0,200])\n\nax = fig.add_subplot(2,2,3)\nax.plot(np.arange(par.path_T),path_r-model.par.r_ss,'-o',ms=2,label='$\\sigma_e = 0.5$')\nax.plot(np.arange(par.path_T),path_r_pf-r_ss_pf,'-o',ms=2,label='$\\sigma_e = 0$')\nax.legend(frameon=True)\nax.set_title('interest rate, $r_t$')\nax.set_xlim([0,200])\n\nax = fig.add_subplot(2,2,4)\nax.plot(np.arange(par.path_T),path_w-model.par.w_ss,'-o',ms=2,label='$\\sigma_e = 0.5$')\nax.plot(np.arange(par.path_T),path_w_pf-w_ss_pf,'-o',ms=2,label='$\\sigma_e = 0$')\nax.legend(frameon=True)\nax.set_title('wage, $w_t$')\nax.set_xlim([0,200])\n\nfig.tight_layout()", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# LeNet Lab\n\nSource: Yan LeCun", "_____no_output_____" ], [ "## Load Data\n\nLoad the MNIST data, which comes pre-loaded with TensorFlow.\n\nYou do not need to modify this section.", "_____no_output_____" ] ], [ [ "from tensorflow.examples.tutorials.mnist import input_data\n\nmnist = input_data.read_data_sets(\"MNIST_data/\", reshape=False)\nX_train, y_train = mnist.train.images, mnist.train.labels\nX_validation, y_validation = mnist.validation.images, mnist.validation.labels\nX_test, y_test = mnist.test.images, mnist.test.labels\n\nassert(len(X_train) == len(y_train))\nassert(len(X_validation) == len(y_validation))\nassert(len(X_test) == len(y_test))\n\nprint()\nprint(\"Image Shape: {}\".format(X_train[0].shape))\nprint()\nprint(\"Training Set: {} samples\".format(len(X_train)))\nprint(\"Validation Set: {} samples\".format(len(X_validation)))\nprint(\"Test Set: {} samples\".format(len(X_test)))", "/Users/Kornet-Mac-1/anaconda3/lib/python3.6/site-packages/h5py/__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.\n from ._conv import register_converters as _register_converters\n" ] ], [ [ "The MNIST data that TensorFlow pre-loads comes as 28x28x1 images.\n\nHowever, the LeNet architecture only accepts 32x32xC images, where C is the number of color channels.\n\nIn order to reformat the MNIST data into a shape that LeNet will accept, we pad the data with two rows of zeros on the top and bottom, and two columns of zeros on the left and right (28+2+2 = 32).\n\nYou do not need to modify this section.", "_____no_output_____" ] ], [ [ "import numpy as np\n\n# Pad images with 0s\nX_train = np.pad(X_train, ((0,0),(2,2),(2,2),(0,0)), 'constant')\nX_validation = np.pad(X_validation, ((0,0),(2,2),(2,2),(0,0)), 'constant')\nX_test = np.pad(X_test, ((0,0),(2,2),(2,2),(0,0)), 'constant')\n \nprint(\"Updated Image Shape: {}\".format(X_train[0].shape))", "Updated Image Shape: (32, 32, 1)\n" ] ], [ [ "## Visualize Data\n\nView a sample from the dataset.\n\nYou do not need to modify this section.", "_____no_output_____" ] ], [ [ "import random\nimport numpy as np\nimport matplotlib.pyplot as plt\n%matplotlib inline\n\nindex = random.randint(0, len(X_train))\nimage = X_train[index].squeeze()\n\nplt.figure(figsize=(1,1))\nplt.imshow(image, cmap=\"gray\")\nprint(y_train[index])", "6\n" ] ], [ [ "## Preprocess Data\n\nShuffle the training data.\n\nYou do not need to modify this section.", "_____no_output_____" ] ], [ [ "from sklearn.utils import shuffle\n\nX_train, y_train = shuffle(X_train, y_train)", "_____no_output_____" ] ], [ [ "## Setup TensorFlow\nThe `EPOCH` and `BATCH_SIZE` values affect the training speed and model accuracy.\n\nYou do not need to modify this section.", "_____no_output_____" ] ], [ [ "import tensorflow as tf\n\nEPOCHS = 10\nBATCH_SIZE = 128", "_____no_output_____" ] ], [ [ "## TODO: Implement LeNet-5\nImplement the [LeNet-5](http://yann.lecun.com/exdb/lenet/) neural network architecture.\n\nThis is the only cell you need to edit.\n### Input\nThe LeNet architecture accepts a 32x32xC image as input, where C is the number of color channels. Since MNIST images are grayscale, C is 1 in this case.\n\n### Architecture\n**Layer 1: Convolutional.** The output shape should be 28x28x6.\n\n**Activation.** Your choice of activation function.\n\n**Pooling.** The output shape should be 14x14x6.\n\n**Layer 2: Convolutional.** The output shape should be 10x10x16.\n\n**Activation.** Your choice of activation function.\n\n**Pooling.** The output shape should be 5x5x16.\n\n**Flatten.** Flatten the output shape of the final pooling layer such that it's 1D instead of 3D. The easiest way to do is by using `tf.contrib.layers.flatten`, which is already imported for you.\n\n**Layer 3: Fully Connected.** This should have 120 outputs.\n\n**Activation.** Your choice of activation function.\n\n**Layer 4: Fully Connected.** This should have 84 outputs.\n\n**Activation.** Your choice of activation function.\n\n**Layer 5: Fully Connected (Logits).** This should have 10 outputs.\n\n### Output\nReturn the result of the 2nd fully connected layer.", "_____no_output_____" ] ], [ [ "from tensorflow.contrib.layers import flatten\n\ndef LeNet(x): \n # Arguments used for tf.truncated_normal, randomly defines variables for the weights and biases for each layer\n mu = 0\n sigma = 0.1\n \n # TODO: Layer 1: Convolutional. Input = 32x32x1. Output = 28x28x6.\n conv1_W = tf.Variable(tf.truncated_normal(shape=(5, 5, 1, 6), mean = mu, stddev = sigma))\n conv1_b = tf.Variable(tf.zeros(6))\n conv1 = tf.nn.conv2d(x, conv1_W, strides=[1, 1, 1, 1], padding='VALID') + conv1_b\n\n # TODO: Activation.\n conv1 = tf.nn.relu(conv1)\n\n # TODO: Pooling. Input = 28x28x6. Output = 14x14x6.\n conv1 = tf.nn.max_pool(conv1, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='VALID')\n\n # TODO: Layer 2: Convolutional. Output = 10x10x16.\n conv2_W = tf.Variable(tf.truncated_normal(shape=(5, 5, 6, 16), mean = mu, stddev = sigma))\n conv2_b = tf.Variable(tf.zeros(16))\n conv2 = tf.nn.conv2d(conv1, conv2_W, strides=[1, 1, 1, 1], padding='VALID') + conv2_b\n \n # TODO: Activation.\n conv2 = tf.nn.relu(conv2)\n\n # TODO: Pooling. Input = 10x10x16. Output = 5x5x16.\n conv2 = tf.nn.max_pool(conv2, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='VALID')\n\n # TODO: Flatten. Input = 5x5x16. Output = 400.\n fc0 = flatten(conv2)\n \n # TODO: Layer 3: Fully Connected. Input = 400. Output = 120.\n fc1_W = tf.Variable(tf.truncated_normal(shape=(400, 120), mean = mu, stddev = sigma))\n fc1_b = tf.Variable(tf.zeros(120))\n fc1 = tf.matmul(fc0, fc1_W) + fc1_b\n \n # TODO: Activation.\n fc1 = tf.nn.relu(fc1)\n\n # TODO: Layer 4: Fully Connected. Input = 120. Output = 84.\n fc2_W = tf.Variable(tf.truncated_normal(shape=(120, 84), mean = mu, stddev = sigma))\n fc2_b = tf.Variable(tf.zeros(84))\n fc2 = tf.matmul(fc1, fc2_W) + fc2_b\n \n # TODO: Activation.\n fc2 = tf.nn.relu(fc2)\n\n # TODO: Layer 5: Fully Connected. Input = 84. Output = 10.\n fc3_W = tf.Variable(tf.truncated_normal(shape=(84, 10), mean = mu, stddev = sigma))\n fc3_b = tf.Variable(tf.zeros(10))\n logits = tf.matmul(fc2, fc3_W) + fc3_b\n \n return logits", "_____no_output_____" ] ], [ [ "## Features and Labels\nTrain LeNet to classify [MNIST](http://yann.lecun.com/exdb/mnist/) data.\n\n`x` is a placeholder for a batch of input images.\n`y` is a placeholder for a batch of output labels.\n\nYou do not need to modify this section.", "_____no_output_____" ] ], [ [ "x = tf.placeholder(tf.float32, (None, 32, 32, 1))\ny = tf.placeholder(tf.int32, (None))\none_hot_y = tf.one_hot(y, 10)", "_____no_output_____" ] ], [ [ "## Training Pipeline\nCreate a training pipeline that uses the model to classify MNIST data.\n\nYou do not need to modify this section.", "_____no_output_____" ] ], [ [ "rate = 0.001\n\nlogits = LeNet(x)\ncross_entropy = tf.nn.softmax_cross_entropy_with_logits(labels=one_hot_y, logits=logits)\nloss_operation = tf.reduce_mean(cross_entropy)\noptimizer = tf.train.AdamOptimizer(learning_rate = rate)\ntraining_operation = optimizer.minimize(loss_operation)", "_____no_output_____" ] ], [ [ "## Model Evaluation\nEvaluate how well the loss and accuracy of the model for a given dataset.\n\nYou do not need to modify this section.", "_____no_output_____" ] ], [ [ "correct_prediction = tf.equal(tf.argmax(logits, 1), tf.argmax(one_hot_y, 1))\naccuracy_operation = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))\nsaver = tf.train.Saver()\n\ndef evaluate(X_data, y_data):\n num_examples = len(X_data)\n total_accuracy = 0\n sess = tf.get_default_session()\n for offset in range(0, num_examples, BATCH_SIZE):\n batch_x, batch_y = X_data[offset:offset+BATCH_SIZE], y_data[offset:offset+BATCH_SIZE]\n accuracy = sess.run(accuracy_operation, feed_dict={x: batch_x, y: batch_y})\n total_accuracy += (accuracy * len(batch_x))\n return total_accuracy / num_examples", "_____no_output_____" ] ], [ [ "## Train the Model\nRun the training data through the training pipeline to train the model.\n\nBefore each epoch, shuffle the training set.\n\nAfter each epoch, measure the loss and accuracy of the validation set.\n\nSave the model after training.\n\nYou do not need to modify this section.", "_____no_output_____" ] ], [ [ "with tf.Session() as sess:\n sess.run(tf.global_variables_initializer())\n num_examples = len(X_train)\n \n print(\"Training...\")\n print()\n for i in range(EPOCHS):\n X_train, y_train = shuffle(X_train, y_train)\n for offset in range(0, num_examples, BATCH_SIZE):\n end = offset + BATCH_SIZE\n batch_x, batch_y = X_train[offset:end], y_train[offset:end]\n sess.run(training_operation, feed_dict={x: batch_x, y: batch_y})\n \n validation_accuracy = evaluate(X_validation, y_validation)\n print(\"EPOCH {} ...\".format(i+1))\n print(\"Validation Accuracy = {:.3f}\".format(validation_accuracy))\n print()\n \n saver.save(sess, './lenet')\n print(\"Model saved\")", "Training...\n\nEPOCH 1 ...\nValidation Accuracy = 0.972\n\nEPOCH 2 ...\nValidation Accuracy = 0.978\n\nEPOCH 3 ...\nValidation Accuracy = 0.984\n\nEPOCH 4 ...\nValidation Accuracy = 0.986\n\nEPOCH 5 ...\nValidation Accuracy = 0.989\n\nEPOCH 6 ...\nValidation Accuracy = 0.987\n\nEPOCH 7 ...\nValidation Accuracy = 0.990\n\nEPOCH 8 ...\nValidation Accuracy = 0.987\n\nEPOCH 9 ...\nValidation Accuracy = 0.989\n\nEPOCH 10 ...\nValidation Accuracy = 0.989\n\nModel saved\n" ] ], [ [ "## Evaluate the Model\nOnce you are completely satisfied with your model, evaluate the performance of the model on the test set.\n\nBe sure to only do this once!\n\nIf you were to measure the performance of your trained model on the test set, then improve your model, and then measure the performance of your model on the test set again, that would invalidate your test results. You wouldn't get a true measure of how well your model would perform against real data.\n\nYou do not need to modify this section.", "_____no_output_____" ] ], [ [ "with tf.Session() as sess:\n saver.restore(sess, tf.train.latest_checkpoint('.'))\n\n test_accuracy = evaluate(X_test, y_test)\n print(\"Test Accuracy = {:.3f}\".format(test_accuracy))", "INFO:tensorflow:Restoring parameters from ./lenet\nTest Accuracy = 0.990\n" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import numpy as np\nimport scipy as sp\nimport scipy.interpolate\nimport matplotlib.pyplot as plt\nimport pandas as pd\nimport scipy.stats\nimport scipy.optimize\n\nfrom scipy.optimize import curve_fit \n\nimport minkowskitools as mt", "_____no_output_____" ], [ "import importlib\nimportlib.reload(mt)", "_____no_output_____" ], [ "n=4000\nrand_points = np.random.uniform(size=(2, n-2))\nedge_points = np.array([[0.0, 1.0],[0.0, 1.0]])\npoints = np.concatenate((rand_points, edge_points), axis=1)", "_____no_output_____" ], [ "connections = mt.get_connections(points, pval=2, radius=0.05)", "_____no_output_____" ], [ "quick_data = []\n\nfor i in range(1000):\n n=1000\n rand_points = np.random.uniform(size=(2, n-2))\n edge_points = np.array([[0.0, 1.0],[0.0, 1.0]])\n points = np.concatenate((rand_points, edge_points), axis=1)\n connections = mt.get_connections(points, pval=2, radius=0.1)\n no_points = mt.perc_thresh_n(connections)\n quick_data.append(no_points)", "_____no_output_____" ], [ "plt.hist(quick_data, cumulative=True, bins=100)\nplt.gca().set(xlim=(0, 1000), xlabel='Number of Points', ylabel='Culmulative Density', title='Connectionn Threshold')\n# plt.savefig('img/pval2r05.pdf')\nplt.gca().set(xlim=(0, np.max(quick_data)))", "_____no_output_____" ], [ "plt.hist(quick_data, bins=100);", "_____no_output_____" ], [ "n=1000\nrand_points = np.random.uniform(size=(2, n-2))\nedge_points = np.array([[0.0, 1.0],[0.0, 1.0]])\npoints = np.concatenate((rand_points, edge_points), axis=1)\n\nmt.smallest_r(points, pval=2)", "_____no_output_____" ], [ "n=1000\ntrials = 100\nall_results = {}\nresults = []\nfor i in range(trials):\n \n rand_points = np.random.uniform(size=(2, n-2))\n edge_points = np.array([[0.0, 1.0],[0.0, 1.0]])\n points = np.concatenate((rand_points, edge_points), axis=1)\n\n results.append(mt.smallest_r(points, pval=2)[1])\n ", "_____no_output_____" ], [ "plt.hist(results, cumulative=True, bins=100);", "_____no_output_____" ], [ "mt.r1_area2D(2)*(.05**2)*n", "_____no_output_____" ], [ "ns = [1000]\nps = [2]\n\nmt.separate_perc_r(ns, ps, 'outputs/test_perc.txt', repeats=10)", "_____no_output_____" ], [ "import importlib\nimportlib.reload(mt)", "_____no_output_____" ], [ "data_dict = {}\nfor pval in [0.8, 1, 1.2]:\n data_dict[pval] = []\n n = 1000\n r = 0.1\n for i in range(1000):\n rand_points = np.random.uniform(size=(2, n-2))\n edge_points = np.array([[0.0, 1.0],[0.0, 1.0]])\n points = np.concatenate((rand_points, edge_points), axis=1)\n connections = mt.get_connections(points, pval=pval, radius=r)\n no_points = mt.perc_thresh_n(connections)\n data_dict[pval].append(no_points)", "_____no_output_____" ], [ "for pval in [0.8, 1, 1.2]:\n plt.hist(data_dict[pval], cumulative=True, bins=100, label=pval, alpha=.3);\n \nplt.legend()\nplt.gca().set(title='Number of Points for Connectedness', xlabel='Points', ylabel='Cumulative Frequency');\n# plt.savefig('img/PointsCumul.pdf')", "_____no_output_____" ], [ "data_dict_r = {}\nfor pval in [0.8, 1, 1.2]:\n data_dict_r[pval] = []\n n = 1000\n r = 0.1\n for i in range(1000):\n print(i, end=',')\n rand_points = np.random.uniform(size=(2, n-2))\n edge_points = np.array([[0.0, 1.0],[0.0, 1.0]])\n points = np.concatenate((rand_points, edge_points), axis=1)\n r_min = smallest_r(points, pval)\n data_dict_r[pval].append(r_min[1])", "0," ], [ "fig, [ax1, ax2] = plt.subplots(ncols=2, figsize=(14, 5))\n\nfor pval in [0.8, 1, 1.2]:\n ax1.hist(data_dict_r[pval], cumulative=True, bins=100, label=pval, alpha=.3);\n\nax1.legend()\nax1.set(xlabel='r', ylabel='Cumulative Frequency')\n# plt.savefig('img/RadCumul.pdf')\n# suptitle='Minimum r for Connectedness'", "_____no_output_____" ], [ "apprx_thresh = [0.065, 0.068, 0.08]\nps = [1.2, 1, 0.8]\n\nfor p, thresh, col in zip(ps, apprx_thresh, ['k', 'g', 'b']):\n rs = np.arange(0.05, 0.14, 0.01)\n ys = 1000*(mt.r1_area2D(p)*rs*rs)\n plt.scatter(thresh, 1000*(mt.r1_area2D(p)*thresh*thresh), c=col)\n plt.plot(rs, ys, c=col, alpha=0.6)\n plt.axvline(x=thresh, c=col, ls='--', label=p, alpha=0.6)", "_____no_output_____" ], [ "n=10\nrand_points = np.random.uniform(size=(2, n-2))\nedge_points = np.array([[0.0, 1.0],[0.0, 1.0]])\npoints = np.concatenate((rand_points, edge_points), axis=1)", "_____no_output_____" ], [ "fig, (ax1, ax2) = plt.subplots(ncols=2, figsize=(10, 5))\nfor pval, col in zip([0.8, 1, 1.2], ['k', 'g', 'b']):\n ax1.hist(data_dict_r[pval], bins=np.arange(0.05, 0.14, 0.0005), label=pval, alpha=.3, color=col, cumulative=1, histtype='step', lw=5)\n hist_out = ax2.hist(data_dict_r[pval], bins=50, color=col, alpha=0.3, label=pval)\n ys = hist_out[0]\n xs = (hist_out[1][1:]+hist_out[1][:-1])/2\n pt = thresh_calc(xs, ys, sig_fract=.8, n_av=5)[0]\n\n ax1.axvline(x=pt, ls='--', alpha=0.6, c=col)\n ax2.axvline(x=pt, ls='--', alpha=0.6, c=col)\n \nax1.axhline(y=500, alpha=0.2, c='r')\n \n# popt, pcov = curve_fit(skewed, xs, ys)\n# plt.plot(xs, skewed(xs, *popt))\nax1.set(xlim=(0.05, 0.12), xlabel='r', ylabel='Cumulative Frequency')\nax2.set(xlim=(0.05, 0.12), xlabel='r', ylabel='Frequency')\nax1.legend(loc='lower right')\nax2.legend()\nplt.savefig('img/r_perc.pdf')\n# plt.gca().set(title='Minimum r for Connectedness', xlabel='r', ylabel='Cumulative Frequency', xlim=(0.05, .1))", "_____no_output_____" ], [ "for pval in [0.8, 1, 1.2]:\n hist_out = np.histogram(data_dict_r[pval], bins=50);\n ys = hist_out[0]\n xs = (hist_out[1][1:]+hist_out[1][:-1])/2\n# popt, pcov = curve_fit(skewed, xs, ys)\n plt.scatter(np.log(xs), np.log(ys))", "/anaconda3/lib/python3.7/site-packages/ipykernel_launcher.py:6: RuntimeWarning: divide by zero encountered in log\n \n" ], [ "ys = hist_out[0]\nxs = (hist_out[1][1:]+hist_out[1][:-1])/2\npopt, pcov = curve_fit(skewed, xs, ys)\nplt.plot(xs, skewed(xs, *popt))", "_____no_output_____" ], [ "def skewed(x, a, b, c, d):\n# (100*(xs-.06), 4, 50)\n return d*sp.stats.skewnorm.pdf(a*x-b, c)\n \npopt, pcov = curve_fit(skewed, xs, ys)\n\nhist_out = plt.hist(data_dict_r[pval], bins=50, label=pval, alpha=.3)\nplt.plot(xs, skewed(xs, *popt))\n# plt.plot(xs, skewed(xs, 100, 6, 4, 50))\n# plt.plot(xs, ys, label='Fit')\nplt.legend()\npopt", "_____no_output_____" ], [ "def moving_average(a, n=3) :\n ret = np.cumsum(np.array(a))\n ret[n:] = ret[n:] - ret[:-n]\n return ret[n - 1:] / n\n\ndef thresh_calc(data, sig_fract=.8, n_av=5, bins=50):\n \n hist_data = np.histogram(data, bins=bins)\n \n xs, ys = (hist_data[1]+hist_data[1])/2, hist_data[0]\n smoothxs = (moving_average(xs, n=n_av))\n smoothys = (moving_average(ys, n=n_av))\n inds = np.where(smoothys > max(smoothys)*sig_fract)\n\n vals, err = np.polyfit(smoothxs[inds], smoothys[inds], 2, cov=True)\n \n stat_point = -.5*vals[1]/vals[0]\n fract_err = np.sqrt(err[0, 0]/(vals[0]**2) + err[1, 1]/(vals[1]**2))\n\n return stat_point, fract_err*stat_point", "_____no_output_____" ], [ "apprx_thresh = [450, 500, 600]\nps = [1.2, 1, 0.8]\n\nfor p, thresh, col in zip(ps, apprx_thresh, ['k', 'g', 'b']):\n xs = np.arange(1000)\n ys = xs*(mt.r1_area2D(p)*.1*.1)\n plt.scatter(thresh, thresh*(mt.r1_area2D(p)*.1*.1), c=col)\n plt.plot(xs, ys, c=col, alpha=0.6)\n plt.axvline(x=thresh, c=col, ls='--', label=p, alpha=0.6)", "_____no_output_____" ], [ "def separate_perc_n(p, r, n_max=None):\n\n if n_max==None:\n n_max=int(4/(mt.r1_area2D(p)*r*r))\n print(n_max)\n\n rand_points = np.random.uniform(size=(2, n_max-2))\n edge_points = np.array([[0.0, 1.0],[0.0, 1.0]])\n points = np.concatenate((rand_points, edge_points), axis=1)\n\n connections = mt.get_connections(points, radius=r, pval=p)\n\n return mt.perc_thresh_n(connections)\n\ndef ensemble_perc_n(fileName, ps, rs, repeats=1, verbose=True):\n\n for p, r in zip(ps, rs):\n \n if verbose:\n print(f'p:{p}, r:{r}')\n \n for i in range(repeats):\n if verbose:\n print(i, end=' ')\n \n thresh = separate_perc_n(p, r)\n file1 = open(\"{}\".format(fileName),\"a\") \n file1.writelines(f'{p} - {r} - {thresh}\\n')\n file1.close()\n\n if verbose:\n print()\n \n return fileName", "_____no_output_____" ], [ "ensemble_perc_n('new_test.txt', [.8, 1.2, 2], [0.2, 0.1, 0.05], repeats=10)", "p:0.8, r:0.2\n0 258\n1 258\n2 258\n3 258\n4 258\n5 258\n6 258\n7 258\n8 258\n9 258\n\np:1.2, r:0.1\n0 680\n1 680\n2 680\n3 680\n4 680\n5 680\n6 680\n7 680\n8 680\n9 680\n\np:2, r:0.05\n0 2037\n1 2037\n2 2037\n3 2037\n4 2037\n5 2037\n6 2037\n7 2037\n8 2037\n9 2037\n\n" ], [ "pd.read_csv('new_test.txt', header=None, delimiter=\" - \")", "/anaconda3/lib/python3.7/site-packages/ipykernel_launcher.py:1: ParserWarning: Falling back to the 'python' engine because the 'c' engine does not support regex separators (separators > 1 char and different from '\\s+' are interpreted as regex); you can avoid this warning by specifying engine='python'.\n \"\"\"Entry point for launching an IPython kernel.\n" ], [ "p=.8\nr=0.05\n\n4/(mt.r1_area2D(p)*r*r)", "_____no_output_____" ], [ "pn1 = pd.read_csv('outputs/perc_n.txt', names=['p', 'r', 'n'], delimiter=\" - \")\npn1.tail()", "/anaconda3/lib/python3.7/site-packages/ipykernel_launcher.py:1: ParserWarning: Falling back to the 'python' engine because the 'c' engine does not support regex separators (separators > 1 char and different from '\\s+' are interpreted as regex); you can avoid this warning by specifying engine='python'.\n \"\"\"Entry point for launching an IPython kernel.\n" ], [ "pn1['edges'] = pn1['n']*pn1['r']*pn1['r']*mt.kernel_area2D(pn1['p'])", "_____no_output_____" ], [ "plt.hist(pn1[pn1['edges'] < 2.95]['edges'], bins=50, cumulative=1);\n# plt.hist(pn1['edges'], bins=50, cumulative=1);\nplt.gca().set(xlabel='Average Number Edges from Node', ylabel='Cumulative Frequency', );", "_____no_output_____" ], [ "plt.hist(pn1[pn1['edges'] < 2.95]['edges'], bins=50, cumulative=0)\nplt.gca().set(xlabel='Average Number Edges from Node', ylabel='Frequency', )", "_____no_output_____" ], [ "for bins in [50, 75, 100]:\n plt.plot(np.arange(0.5, 0.95, 0.01), [thresh_calc(pn1[pn1['edges'] < 2.95]['edges'], sig_fract=elem, bins=bins)[0] for elem in np.arange(0.5, 0.95, 0.01)], label=f'{bins} bins')\nplt.legend()\nplt.gca().set(xlabel='Fraction for bars to be considered', ylabel='Percolation Threshold', );", "_____no_output_____" ], [ "# #input file\n# fin = open('outputs/perc_r5000clean.txt', \"rt\")\n# #output file to write the result to\n# fout = open(\"outputs/perc_r5000clean2.txt\", \"wt\")\n# #for each line in the input file\n# for line in fin:\n# \t#read replace the string and write to output file\n# \tfout.write(line.replace('-[[', '- [['))\n# #close input and output files\n# fin.close()\n# fout.close()", "_____no_output_____" ], [ "pr1 = pd.read_csv('outputs/perc_r5000clean2.txt', names=['p', 'n', 'r', 'path'], delimiter=\" - \")\npr1['edges'] = pr1['n']*pr1['r']*pr1['r']*mt.kernel_area2D(pr1['p'])", "/anaconda3/lib/python3.7/site-packages/ipykernel_launcher.py:1: ParserWarning: Falling back to the 'python' engine because the 'c' engine does not support regex separators (separators > 1 char and different from '\\s+' are interpreted as regex); you can avoid this warning by specifying engine='python'.\n \"\"\"Entry point for launching an IPython kernel.\n" ], [ "fig, ax = plt.subplots(figsize=(7, 7))\n# axins = ax.inset_axes([5, 8, 150, 250])\naxins = ax.inset_axes([0.5, 0.57, 0.5, 0.43])\n\nhist_data = axins.hist(pr1['edges'], bins=100, label='Raw Data')\n\naxins.legend(loc='upper right')\n\n\nn_av = 5\nsig_fract = .7\nplot_fract = 0.1\n\nxs, ys = (hist_data[1]+hist_data[1])/2, hist_data[0]\nsmoothxs = (moving_average(xs, n=n_av))\nsmoothys = (moving_average(ys, n=n_av))\ninds = np.where(smoothys > max(smoothys)*sig_fract)\nnotinds = np.where(smoothys <= max(smoothys)*sig_fract)\n\n[a, b, c], err = np.polyfit(smoothxs[inds], smoothys[inds], 2, cov=True)\n# plt.plot(xs, vals[0]*xs*xs + vals[1]*xs + vals[2])\n# plotx = xs[inds]\n\nax.scatter(smoothxs[inds], smoothys[inds], c='b', alpha=0.5, label='Points in Fit')\nax.scatter(smoothxs[notinds], smoothys[notinds], c='k', alpha=0.2, label='Smoothed Points')\n\nplotx = smoothxs[inds]\nlowerlim = max(smoothys)*plot_fract\nquadx = np.arange((-b+np.sqrt(b*b - 4*a*(c-lowerlim)))/(2*a), (-b-np.sqrt(b*b - 4*a*(c-lowerlim)))/(2*a), 0.001)\nquady = a*quadx*quadx + b*quadx + c\nplotinds = np.where(quady > 0)\n\nax.axhline(max(smoothys)*sig_fract, color='r', alpha=0.5, ls='--', label=f'Fraction={sig_fract}')\nax.axvline(thresh_calc(pr1['edges'])[0], color='g', alpha=0.5, ls='--', label=f'Threshold')\nax.plot(quadx, quady, c='b', alpha=0.6, label='Quadratic Fit')\n\nax.legend(loc='best', bbox_to_anchor=(0.5, 0., 0.5, 0.45))\nax.set(xlabel='Average number of edges per node', ylabel='Frequency', title='Determining the Percolation Threshold');\nplt.savefig('img/percthreshn5000.pdf')", "_____no_output_____" ], [ "ss = np.arange(0.2, 0.85, 0.01)\nplt.plot(ss, [thresh_calc(pr1['edges'], sig_fract=s)[0] for s in ss])", "_____no_output_____" ], [ "r5000 = pd.read_csv('outputs/perc_r5000clean2.txt', names=['p', 'n', 'r', 'path'], delimiter=\" - \")\nr5000['e'] = mt.kernel_area2D(r5000['p'])*r5000['r']*r5000['r']*r5000['n']", "/anaconda3/lib/python3.7/site-packages/ipykernel_launcher.py:1: ParserWarning: Falling back to the 'python' engine because the 'c' engine does not support regex separators (separators > 1 char and different from '\\s+' are interpreted as regex); you can avoid this warning by specifying engine='python'.\n \"\"\"Entry point for launching an IPython kernel.\n" ], [ "ps = [0.4, 0.6, 0.8, 1.0]\nthreshs = [thresh_calc(r5000[np.abs(r5000['p']-p) < 0.01]['e'], sig_fract=0.6)[0] for p in ps]", "_____no_output_____" ], [ "plt.plot(ps, threshs)", "_____no_output_____" ], [ "r5000[np.abs(r5000['p']-1) < .01]", "_____no_output_____" ], [ "thresh_calc(r5000[np.abs(r5000['p']-p) < 0.01]['e'])[0]", "_____no_output_____" ], [ "thresh_calc(r5000[np.abs(r5000['p']-0.6) < .1]['e'], sig_fract=.6)", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "# Praca domowa 3", "_____no_output_____" ], [ "## Ładowanie podstawowych pakietów", "_____no_output_____" ] ], [ [ "import pandas as pd\nimport numpy as np\nimport sklearn\nimport seaborn as sns\nimport matplotlib.pyplot as plt\nfrom sklearn.model_selection import train_test_split\nfrom sklearn.metrics import f1_score\nfrom sklearn.model_selection import StratifiedKFold # used in crossvalidation\nfrom sklearn.model_selection import KFold\n\nimport IPython\nfrom time import time", "_____no_output_____" ] ], [ [ "## Krótki wstęp", "_____no_output_____" ], [ "Celem zadania jest by bazujac na danych meteorologicznych z australi sprawdzić i wytrenowac 3 różne modele. Równie ważnym celem zadania jest przejrzenie oraz zmiana tzn. hiperparamterów z każdego nich. ", "_____no_output_____" ], [ "### Załadowanie danych", "_____no_output_____" ] ], [ [ "data = pd.read_csv(\"../../australia.csv\")", "_____no_output_____" ] ], [ [ "### Przyjrzenie się danym", "_____no_output_____" ] ], [ [ "data.info()", "<class 'pandas.core.frame.DataFrame'>\nRangeIndex: 56420 entries, 0 to 56419\nData columns (total 18 columns):\nMinTemp 56420 non-null float64\nMaxTemp 56420 non-null float64\nRainfall 56420 non-null float64\nEvaporation 56420 non-null float64\nSunshine 56420 non-null float64\nWindGustSpeed 56420 non-null float64\nWindSpeed9am 56420 non-null float64\nWindSpeed3pm 56420 non-null float64\nHumidity9am 56420 non-null float64\nHumidity3pm 56420 non-null float64\nPressure9am 56420 non-null float64\nPressure3pm 56420 non-null float64\nCloud9am 56420 non-null float64\nCloud3pm 56420 non-null float64\nTemp9am 56420 non-null float64\nTemp3pm 56420 non-null float64\nRainToday 56420 non-null int64\nRainTomorrow 56420 non-null int64\ndtypes: float64(16), int64(2)\nmemory usage: 7.7 MB\n" ] ], [ [ "Nie ma w danych żadnych braków, oraz są one przygotowane idealnie do uczenia maszynowego. Przyjżyjmy się jednak jak wygląda ramka. ", "_____no_output_____" ] ], [ [ "data.head()", "_____no_output_____" ] ], [ [ "## Random Forest", "_____no_output_____" ], [ "**Załadowanie potrzebnych bibliotek** ", "_____no_output_____" ] ], [ [ "from sklearn.ensemble import RandomForestClassifier", "_____no_output_____" ] ], [ [ "**Inicjalizowanie modelu**", "_____no_output_____" ] ], [ [ "rf_default = RandomForestClassifier()", "_____no_output_____" ] ], [ [ "**Hiperparametry**", "_____no_output_____" ] ], [ [ "params = rf_default.get_params()\nparams", "_____no_output_____" ] ], [ [ "**Zmiana kilku hiperparametrów**", "_____no_output_____" ] ], [ [ "params['n_estimators']=150\nparams['max_depth']=6\nparams['min_samples_leaf']=4\nparams['n_jobs']=4\nparams['random_state']=0", "_____no_output_____" ], [ "rf_modified = RandomForestClassifier()\nrf_modified.set_params(**params)", "_____no_output_____" ] ], [ [ "## Extreme Gradient Boosting", "_____no_output_____" ], [ "**Załadowanie potrzebnych bibliotek** ", "_____no_output_____" ] ], [ [ "from xgboost import XGBClassifier", "_____no_output_____" ] ], [ [ "**Inicjalizowanie modelu**", "_____no_output_____" ] ], [ [ "xgb_default = XGBClassifier()", "_____no_output_____" ] ], [ [ "**Hiperparametry**", "_____no_output_____" ] ], [ [ "params = xgb_default.get_params()\nparams", "_____no_output_____" ] ], [ [ "**Zmiana kilku hiperparametrów**", "_____no_output_____" ] ], [ [ "params['n_estimators']=150\nparams['max_depth']=6\nparams['n_jobs']=4\nparams['random_state']=0", "_____no_output_____" ], [ "xgb_modified = XGBClassifier()\nxgb_modified.set_params(**params)", "_____no_output_____" ] ], [ [ "## Support Vector Machines", "_____no_output_____" ], [ "**Załadowanie potrzebnych bibliotek** ", "_____no_output_____" ] ], [ [ "from sklearn.svm import SVC", "_____no_output_____" ] ], [ [ "**Inicjalizowanie modelu**", "_____no_output_____" ] ], [ [ "svc_default = SVC()", "_____no_output_____" ] ], [ [ "**Hiperparametry**", "_____no_output_____" ] ], [ [ "params = svc_default.get_params()\nparams", "_____no_output_____" ] ], [ [ "**Zmiana kilku hiperparametrów**", "_____no_output_____" ] ], [ [ "params['degree']=3\nparams['tol']=0.001\nparams['random_state']=0", "_____no_output_____" ], [ "svc_modified = SVC()\nsvc_modified.set_params(**params)", "_____no_output_____" ] ], [ [ "## Komentarz\nW tym momencie otrzymaliśmy 3 modele z zmienionymi hiperparametrami, oraz ich domyślne odpowiedniki. Zobaczmy teraz jak zmieniły się rezultaty osiągane przez te modele i chociaż nie był to cel tego zadania, zobaczmy czy może udało nam się poprawić jakiś model.", "_____no_output_____" ], [ "## Porównanie", "_____no_output_____" ], [ "**Załadowanie potrzebnych bibliotek** ", "_____no_output_____" ] ], [ [ "from sklearn.metrics import f1_score\nfrom sklearn.metrics import accuracy_score\nfrom sklearn.metrics import balanced_accuracy_score\nfrom sklearn.metrics import precision_score\nfrom sklearn.metrics import average_precision_score\nfrom sklearn.metrics import roc_auc_score", "_____no_output_____" ] ], [ [ "### Funkcje pomocnicze", "_____no_output_____" ] ], [ [ "def cv_classifier(classifier,kfolds = 10, X = data.drop(\"RainTomorrow\", axis = 1), y = data.RainTomorrow):\n start_time = time()\n \n scores ={}\n scores[\"f1\"]=[]\n scores[\"accuracy\"]=[]\n scores[\"balanced_accuracy\"]=[]\n scores[\"precision\"]=[]\n scores[\"average_precision\"]=[]\n scores[\"roc_auc\"]=[]\n \n # Hardcoded crossvalidation metod, could be \n cv= StratifiedKFold(n_splits=kfolds,shuffle=True,random_state=0)\n \n for i, (train, test) in enumerate(cv.split(X, y)):\n \n IPython.display.clear_output()\n print(f\"Model {i+1}/{kfolds}\")\n \n # Training model\n classifier.fit(X.iloc[train, ], y.iloc[train], )\n \n # Testing model\n prediction = classifier.predict(X.iloc[test,])\n \n # calculating and savings scores\n scores[\"f1\"].append( f1_score(y.iloc[test],prediction))\n scores[\"accuracy\"].append( accuracy_score(y.iloc[test],prediction))\n scores[\"balanced_accuracy\"].append( balanced_accuracy_score(y.iloc[test],prediction))\n scores[\"precision\"].append( precision_score(y.iloc[test],prediction))\n scores[\"average_precision\"].append( average_precision_score(y.iloc[test],prediction))\n scores[\"roc_auc\"].append( roc_auc_score(y.iloc[test],prediction))\n \n IPython.display.clear_output()\n print(f\"Crossvalidation on {kfolds} folds done in {round((time()-start_time),2)}s\")\n \n return scores", "_____no_output_____" ], [ "def get_mean_scores(scores_dict):\n means={}\n for score_name in scores_dict:\n means[score_name] = np.mean(scores_dict[score_name])\n return means", "_____no_output_____" ], [ "def print_mean_scores(mean_scores_dict,precision=4):\n for score_name in mean_scores_dict:\n print(f\"Mean {score_name} score is {round(mean_scores_dict[score_name]*100,precision)}%\")", "_____no_output_____" ] ], [ [ "### Wyniki\n\nPoniżej zamieszczam wyniki predykcji pokazanych wcześniej modeli. Dla kontrastu nauczyłem zmodyfikowane wersję klasyfikatorów jak i również te domyślne. Ze smutkiem muszę stwierdzić, że nie jestem najlepszy w strzelaniu, ponieważ parametry, które dobrałem znacznie pogarszają skutecznść każdego z modeli. Niemniej jednak by to stwierdzić musiałem sie posłóżyć pewnymi miarami. Są to:\n* F1 \n* Accuracy\n* Balanced Accuracy\n* Precision\n* Average Precision\n* ROC AUC\n\nWszystkie modele zostały poddane 10 krotnej kroswalidacji, więc przedstawione wyniki są średnią. Kroswalidacja pozwala dokładniej ocenić skutecznosć modelu oraz wyciągajac z nich takie informacje jak odchylenie standardowe wyników, co daje nam możliowść dyskusji na temat działania modelu w skrajnych przypadkach. ", "_____no_output_____" ], [ "### Random Forest", "_____no_output_____" ], [ "### Kroswalidacja modeli", "_____no_output_____" ] ], [ [ "scores_rf_default = cv_classifier(rf_default)", "Crossvalidation on 10 folds done in 169.77s\n" ], [ "scores_rf_modified = cv_classifier(rf_modified)", "Crossvalidation on 10 folds done in 35.2s\n" ], [ "mean_scores_rf_default = get_mean_scores(scores_rf_default)\nmean_scores_rf_modified = get_mean_scores(scores_rf_modified)", "_____no_output_____" ] ], [ [ "**Random forest default**", "_____no_output_____" ] ], [ [ "print_mean_scores(mean_scores_rf_default,precision=2)", "Mean f1 score is 62.85%\nMean accuracy score is 86.09%\nMean balanced_accuracy score is 74.38%\nMean precision score is 76.32%\nMean average_precision score is 51.05%\nMean roc_auc score is 74.38%\n" ] ], [ [ "**Random forest modified**", "_____no_output_____" ] ], [ [ "print_mean_scores(mean_scores_rf_modified,precision=2)", "Mean f1 score is 56.74%\nMean accuracy score is 84.97%\nMean balanced_accuracy score is 70.54%\nMean precision score is 77.55%\nMean average_precision score is 46.88%\nMean roc_auc score is 70.54%\n" ] ], [ [ "## Extreme Gradient Boosting", "_____no_output_____" ], [ "### Kroswalidacja modeli", "_____no_output_____" ] ], [ [ "scores_xgb_default = cv_classifier(xgb_default)", "Crossvalidation on 10 folds done in 41.98s\n" ], [ "scores_xgb_modified = cv_classifier(xgb_modified)", "Crossvalidation on 10 folds done in 67.69s\n" ], [ "mean_scores_xgb_default = get_mean_scores(scores_xgb_default)\nmean_scores_xgb_modified = get_mean_scores(scores_xgb_modified)", "_____no_output_____" ] ], [ [ "**XGBoost default**", "_____no_output_____" ] ], [ [ "print_mean_scores(mean_scores_xgb_default,precision=2)", "Mean f1 score is 63.79%\nMean accuracy score is 85.92%\nMean balanced_accuracy score is 75.3%\nMean precision score is 73.56%\nMean average_precision score is 51.06%\nMean roc_auc score is 75.3%\n" ] ], [ [ "**XGBoost modified**", "_____no_output_____" ] ], [ [ "print_mean_scores(mean_scores_xgb_modified,precision=2)", "Mean f1 score is 63.93%\nMean accuracy score is 85.89%\nMean balanced_accuracy score is 75.44%\nMean precision score is 73.18%\nMean average_precision score is 51.07%\nMean roc_auc score is 75.44%\n" ] ], [ [ "## Support Vector Machines", "_____no_output_____" ], [ "### Kroswalidacja modeli", "_____no_output_____" ], [ "**warning this takes a while**", "_____no_output_____" ] ], [ [ "scores_svc_default = cv_classifier(svc_default)", "Crossvalidation on 10 folds done in 923.06s\n" ], [ "scores_svc_modified = cv_classifier(svc_modified)", "Crossvalidation on 10 folds done in 864.31s\n" ], [ "mean_scores_svc_default = get_mean_scores(scores_svc_default)\nmean_scores_svc_modified = get_mean_scores(scores_svc_modified)", "_____no_output_____" ] ], [ [ "**SVM default**", "_____no_output_____" ] ], [ [ "print_mean_scores(mean_scores_svc_default,precision=2)", "Mean f1 score is 51.52%\nMean accuracy score is 84.38%\nMean balanced_accuracy score is 67.63%\nMean precision score is 81.47%\nMean average_precision score is 44.44%\nMean roc_auc score is 67.63%\n" ] ], [ [ "**SVM modified**", "_____no_output_____" ] ], [ [ "print_mean_scores(mean_scores_svc_modified,precision=2)", "Mean f1 score is 51.52%\nMean accuracy score is 84.38%\nMean balanced_accuracy score is 67.63%\nMean precision score is 81.47%\nMean average_precision score is 44.44%\nMean roc_auc score is 67.63%\n" ] ], [ [ "## Podsumowanie", "_____no_output_____" ], [ "Wyniki random forest oraz xgboost były dośyć zbliżone i szczerze mówiąc dosyć słabe. Jeszcze gorzej wypadł SVM, co pewnie wielu nie zdziwi. Ma okropnie długi czas uczenia, ponad minuta na model. Wypada dużo gorzej niż pozostałe algorytmy, gdzie 10 modeli xgboost zostało wyszkolone w 41s. Natomiast wyniki random forest oraz xgboost są dosyć zbliżone. Gdybym jednak miał wybrać jeden z tych trzech modeli, by dalej go dostrajać na pewno zdecydowałbym się na xgboosta. Między innymi dlatego, że czas uczenia i testowania byłby dużo krótszy niż w przypadku random forest, oraz prawdopodobnie z odpowiednimi parametrami xgboost będzie sobie radził lepiej niż random forest. \n\nNatomiast wybór najlepszej miary nie jest już taki prosty, a nawet śmiem twierdzić, że nie znalazłem takiej która zasługiwałaby na takie miano. Większość miar w niebanalny sposób reprezentuje jakąś cechę modelu. Natomiast jeżeli musiałbym się ograniczyć do jednej prawdopodobnie wybrałbym ROC_AUC. Z tego powodu, że przez użycie True Positive Rate i False Positive Rate jest ona całkiem logiczna (w przeciwieństwie do wielu), a zarazem wiąć pozwala dobrze tłumaczyć skutecznośc modeli.", "_____no_output_____" ], [ "# Część bonusowa - Regresja", "_____no_output_____" ], [ "### Przygotowanie danych", "_____no_output_____" ] ], [ [ "data2 = pd.read_csv('allegro-api-transactions.csv')\ndata2 = data2.drop(['lp','date'], axis = 1)\ndata2.head()", "_____no_output_____" ] ], [ [ "Dane są prawie gotowe do procesu czenia, trzeba jedynie poprawić `it_location` w którym mogą pojawić się powtórki w stylu *Warszawa* i *warszawa*, a następnie zakodować zmienne kategoryczne", "_____no_output_____" ] ], [ [ "data2.it_location = data2.it_location.str.lower()\ndata2.head()", "_____no_output_____" ], [ "encoding_columns = ['categories','seller','it_location','main_category']", "_____no_output_____" ] ], [ [ "## Kodowanie zmiennych kategorycznych ", "_____no_output_____" ] ], [ [ "import category_encoders\nfrom sklearn.preprocessing import OneHotEncoder", "_____no_output_____" ] ], [ [ "### Podział danych \nNie wykonam standardowego podziału na dane test i train, ponieważ w dalszej części dokumentu do oceny skutecznosci użytych kodowań posłużę się kroswalidacją. Pragnę zaznaczyć, ze prawdopodbnie najlepszą metodą tutaj byłoby rozbicie kategorii `categories` na 26 kolumn, zero-jedynkowych, jednak znacznie by to powiększyło rozmiar danych. Z dokładnie tego samego powodu nie wykonam one hot encodingu, tylko posłużę się kodowaniami, które nie powiększą rozmiatu danych.", "_____no_output_____" ] ], [ [ "X = data2.drop('price', axis = 1)\ny = data2.price", "_____no_output_____" ] ], [ [ "## Target encoding", "_____no_output_____" ] ], [ [ "te = category_encoders.target_encoder.TargetEncoder(data2, cols = encoding_columns)\ntarget_encoded = te.fit_transform(X,y)\ntarget_encoded", "_____no_output_____" ] ], [ [ "## James-Stein Encoding", "_____no_output_____" ] ], [ [ "js = category_encoders.james_stein.JamesSteinEncoder(cols = encoding_columns)\nencoded_js = js.fit_transform(X,y)\nencoded_js", "_____no_output_____" ] ], [ [ "## Cat Boost Encoding", "_____no_output_____" ] ], [ [ "cb = category_encoders.cat_boost.CatBoostEncoder(cols = encoding_columns)\nencoded_cb = cb.fit_transform(X,y)\nencoded_cb", "_____no_output_____" ] ], [ [ "## Testowanie", "_____no_output_____" ] ], [ [ "from sklearn.metrics import r2_score, mean_squared_error\nfrom sklearn import linear_model", "_____no_output_____" ], [ "def cv_encoding(model,kfolds = 10, X = data.drop(\"RainTomorrow\", axis = 1), y = data.RainTomorrow):\n start_time = time()\n \n scores ={}\n scores[\"r2_score\"] = []\n scores['RMSE'] = []\n \n # Standard k-fold\n cv = KFold(n_splits=kfolds,shuffle=False,random_state=0)\n \n for i, (train, test) in enumerate(cv.split(X, y)):\n \n IPython.display.clear_output()\n print(f\"Model {i+1}/{kfolds}\")\n \n # Training model\n model.fit(X.iloc[train, ], y.iloc[train], )\n \n # Testing model\n prediction = model.predict(X.iloc[test,])\n \n # calculating and savings score\n scores['r2_score'].append( r2_score(y.iloc[test],prediction))\n scores['RMSE'].append( mean_squared_error(y.iloc[test],prediction))\n \n \n IPython.display.clear_output()\n print(f\"Crossvalidation on {kfolds} folds done in {round((time()-start_time),2)}s\")\n \n return scores", "_____no_output_____" ] ], [ [ "## Mierzenie skutecznosci kodowań\nZdecydowałem isę skorzystać z modelu regresji liniowej `Lasso` początkowo chciałem skorzystać z `Elastic Net`, ale jak się okazało zmienne nie sa ze sobą zbytnio powiązane, a to miał być główny powód do jego użycia.", "_____no_output_____" ] ], [ [ "corr=data2.corr()\nfig, ax=plt.subplots(figsize=(9,6)) \nax=sns.heatmap(corr, xticklabels=corr.columns, yticklabels=corr.columns, annot=True, cmap=\"PiYG\", center=0, vmin=-1, vmax=1)\nax.set_title('Korelacje zmiennych')\nplt.show();", "_____no_output_____" ] ], [ [ "### Wybór modelu liniowego\nOkreślam go w tym miejscu, ponieważ w dalszych częściach dokumentu będę z niego wielokrotnie korzystał przy kroswalidacji", "_____no_output_____" ] ], [ [ "lasso = linear_model.Lasso()", "_____no_output_____" ] ], [ [ "## Wyniki target encodingu", "_____no_output_____" ] ], [ [ "target_encoding_scores = cv_encoding(model = lasso,kfolds=20, X = target_encoded, y = y)", "Crossvalidation on 20 folds done in 8.57s\n" ], [ "target_encoding_scores_mean = get_mean_scores(target_encoding_scores)\ntarget_encoding_scores_mean", "_____no_output_____" ] ], [ [ "## Wyniki James-Stein Encodingu", "_____no_output_____" ] ], [ [ "js_encoding_scores = cv_encoding(lasso, 20, encoded_js, y)", "Crossvalidation on 20 folds done in 24.17s\n" ], [ "js_encoding_scores_mean = get_mean_scores(js_encoding_scores)\njs_encoding_scores_mean", "_____no_output_____" ] ], [ [ "## Wyniki Cat Boost Encodingu", "_____no_output_____" ] ], [ [ "cb_encoding_scores = cv_encoding(lasso, 20 ,encoded_cb, y)", "Crossvalidation on 20 folds done in 11.04s\n" ], [ "cb_encoding_scores_mean = get_mean_scores(cb_encoding_scores)\ncb_encoding_scores_mean", "_____no_output_____" ] ], [ [ "## Porównanie", "_____no_output_____" ], [ "## Wyniki metryki r2", "_____no_output_____" ] ], [ [ "r2_data = [target_encoding_scores[\"r2_score\"], js_encoding_scores[\"r2_score\"], cb_encoding_scores[\"r2_score\"]]\nlabels = [\"Target\", \" James-Stein\", \"Cat Boost\"]\nfig, ax = plt.subplots(figsize = (12,9))\nax.set_title('Wyniki r2')\nax.boxplot(r2_data, labels = labels)\nplt.show()", "_____no_output_____" ] ], [ [ "**Komentarz** \n\nWidać, że użycie kodowania Jamesa-Steina pozwoliło modelowi dużo lepiej się dopasować do danych, jednak stwarza to ewentaulny problem nadmiernego dopasowania się do danych. Warto by było sprawdzić czy przez ten sposób kodowania nie dochodzi do dużo silniejszego overfittingu. ", "_____no_output_____" ], [ "## Wynikii metryki RMSE", "_____no_output_____" ] ], [ [ "rmse_data = [target_encoding_scores[\"RMSE\"], js_encoding_scores[\"RMSE\"], cb_encoding_scores[\"RMSE\"]]\nlabels = [\"Target\", \" James-Stein\", \"Cat Boost\"]\nfig, ax = plt.subplots(figsize = (12,9))\nax.set_title('Wyniki RMSE w skali logarytmicznej')\nax.set_yscale('log')\nax.boxplot(rmse_data, labels = labels)\nplt.show()", "_____no_output_____" ] ], [ [ "**Komentarz**\n\nNajlepiej poradził sobie James-Stein Encoding, nic dziwnego ponieważ r2 wskazał nam już lepsze dopasowanie modelu. ", "_____no_output_____" ], [ "## Podsumowanie", "_____no_output_____" ], [ "Kodowanie Jamesa-Steina osiąga dużo lepsze wyniki niż pozostałe dwa przez mnie wybrane. O ile nie dochodzi w tym przypadku do overfittingu, to w tej grupie z pewnością wybrałbym właśnie to kodowanie. Warto się jednak zastanowić nad kodowaniem one-hot, które w tym przypadku wydaje się bardzo naturale, jednakże wiąże się z kilkukrotnym powiększeniem danych.", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown" ]

[ [ "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown", "markdown", "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "This notebook was prepared by [Donne Martin](https://github.com/donnemartin). Source and license info is on [GitHub](https://github.com/donnemartin/interactive-coding-challenges).", "_____no_output_____" ], [ "# Challenge Notebook", "_____no_output_____" ], [ "## Problem: Implement depth-first traversals (in-order, pre-order, post-order) on a binary tree.\n\n* [Constraints](#Constraints)\n* [Test Cases](#Test-Cases)\n* [Algorithm](#Algorithm)\n* [Code](#Code)\n* [Unit Test](#Unit-Test)", "_____no_output_____" ], [ "## Constraints\n\n* Can we assume we already have a Node class with an insert method?\n * Yes\n* What should we do with each node when we process it?\n * Call an input method `visit_func` on the node\n* Can we assume this fits in memory?\n * Yes", "_____no_output_____" ], [ "## Test Cases\n\n### In-Order Traversal\n\n* 5, 2, 8, 1, 3 -> 1, 2, 3, 5, 8\n* 1, 2, 3, 4, 5 -> 1, 2, 3, 4, 5\n\n### Pre-Order Traversal\n\n* 5, 2, 8, 1, 3 -> 5, 2, 1, 3, 8\n* 1, 2, 3, 4, 5 -> 1, 2, 3, 4, 5\n\n### Post-Order Traversal\n\n* 5, 2, 8, 1, 3 -> 1, 3, 2, 8, 5\n* 1, 2, 3, 4, 5 -> 5, 4, 3, 2, 1", "_____no_output_____" ], [ "## Algorithm\n\nRefer to the [Solution Notebook](http://nbviewer.ipython.org/github/donnemartin/interactive-coding-challenges/blob/master/graphs_trees/tree_dfs/dfs_solution.ipynb). If you are stuck and need a hint, the solution notebook's algorithm discussion might be a good place to start.", "_____no_output_____" ], [ "## Code", "_____no_output_____" ] ], [ [ "# %load ../bst/bst.py\nclass Node(object):\n\n def __init__(self, data):\n self.data = data\n self.left = None\n self.right = None\n self.parent = None\n\n def __repr__(self):\n return str(self.data)\n\n\nclass Bst(object):\n\n def __init__(self, root=None):\n self.root = root\n\n def insert(self, data):\n if data is None:\n raise TypeError('data cannot be None')\n if self.root is None:\n self.root = Node(data)\n return self.root\n else:\n return self._insert(self.root, data)\n\n def _insert(self, node, data):\n if node is None:\n return Node(data)\n if data <= node.data:\n if node.left is None:\n node.left = self._insert(node.left, data)\n node.left.parent = node\n return node.left\n else:\n return self._insert(node.left, data)\n else:\n if node.right is None:\n node.right = self._insert(node.right, data)\n node.right.parent = node\n return node.right\n else:\n return self._insert(node.right, data)", "_____no_output_____" ], [ "class BstDfs(Bst):\n\n def in_order_traversal(self, node, visit_func):\n if node is None:\n return\n \n self.in_order_traversal(node.left, visit_func)\n \n visit_func(node)\n \n self.in_order_traversal(node.right, visit_func)\n\n def pre_order_traversal(self, node, visit_func):\n if node is None:\n return\n \n visit_func(node)\n\n self.pre_order_traversal(node.left, visit_func)\n \n self.pre_order_traversal(node.right, visit_func)\n\n def post_order_traversal(self,node, visit_func):\n if node is None:\n return\n \n self.post_order_traversal(node.left, visit_func)\n \n self.post_order_traversal(node.right, visit_func)\n\n visit_func(node)\n ", "_____no_output_____" ] ], [ [ "## Unit Test", "_____no_output_____" ] ], [ [ "%run ../utils/results.py", "_____no_output_____" ], [ "# %load test_dfs.py\nfrom nose.tools import assert_equal\n\n\nclass TestDfs(object):\n\n def __init__(self):\n self.results = Results()\n\n def test_dfs(self):\n bst = BstDfs(Node(5))\n bst.insert(2)\n bst.insert(8)\n bst.insert(1)\n bst.insert(3)\n\n bst.in_order_traversal(bst.root, self.results.add_result)\n assert_equal(str(self.results), \"[1, 2, 3, 5, 8]\")\n self.results.clear_results()\n\n bst.pre_order_traversal(bst.root, self.results.add_result)\n assert_equal(str(self.results), \"[5, 2, 1, 3, 8]\")\n self.results.clear_results()\n\n bst.post_order_traversal(bst.root, self.results.add_result)\n assert_equal(str(self.results), \"[1, 3, 2, 8, 5]\")\n self.results.clear_results()\n\n bst = BstDfs(Node(1))\n bst.insert(2)\n bst.insert(3)\n bst.insert(4)\n bst.insert(5)\n\n bst.in_order_traversal(bst.root, self.results.add_result)\n assert_equal(str(self.results), \"[1, 2, 3, 4, 5]\")\n self.results.clear_results()\n\n bst.pre_order_traversal(bst.root, self.results.add_result)\n assert_equal(str(self.results), \"[1, 2, 3, 4, 5]\")\n self.results.clear_results()\n\n bst.post_order_traversal(bst.root, self.results.add_result)\n assert_equal(str(self.results), \"[5, 4, 3, 2, 1]\")\n\n print('Success: test_dfs')\n\n\ndef main():\n test = TestDfs()\n test.test_dfs()\n\n\nif __name__ == '__main__':\n main()", "Success: test_dfs\n" ] ], [ [ "## Solution Notebook\n\nReview the [Solution Notebook](http://nbviewer.ipython.org/github/donnemartin/interactive-coding-challenges/blob/master/graphs_trees/tree_dfs/dfs_solution.ipynb) for a discussion on algorithms and code solutions.", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown" ]

[ [ "markdown", "markdown", "markdown", "markdown", "markdown", "markdown", "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "# Generative Adversarial Networks\n\n\nThroughout most of this book, we've talked about how to make predictions.\nIn some form or another, we used deep neural networks learned mappings from data points to labels.\nThis kind of learning is called discriminative learning,\nas in, we'd like to be able to discriminate between photos cats and photos of dogs. \nClassifiers and regressors are both examples of discriminative learning. \nAnd neural networks trained by backpropagation \nhave upended everything we thought we knew about discriminative learning \non large complicated datasets. \nClassification accuracies on high-res images has gone from useless \nto human-level (with some caveats) in just 5-6 years. \nWe'll spare you another spiel about all the other discriminative tasks \nwhere deep neural networks do astoundingly well.\n\nBut there's more to machine learning than just solving discriminative tasks.\nFor example, given a large dataset, without any labels,\nwe might want to learn a model that concisely captures the characteristics of this data.\nGiven such a model, we could sample synthetic data points that resemble the distribution of the training data.\nFor example, given a large corpus of photographs of faces,\nwe might want to be able to generate a *new* photorealistic image \nthat looks like it might plausibly have come from the same dataset. \nThis kind of learning is called *generative modeling*. \n\nUntil recently, we had no method that could synthesize novel photorealistic images. \nBut the success of deep neural networks for discriminative learning opened up new possiblities.\nOne big trend over the last three years has been the application of discriminative deep nets\nto overcome challenges in problems that we don't generally think of as supervised learning problems.\nThe recurrent neural network language models are one example of using a discriminative network (trained to predict the next character)\nthat once trained can act as a generative model. \n\n\nIn 2014, a young researcher named Ian Goodfellow introduced [Generative Adversarial Networks (GANs)](https://arxiv.org/abs/1406.2661) a clever new way to leverage the power of discriminative models to get good generative models. \nGANs made quite a splash so it's quite likely you've seen the images before. \nFor instance, using a GAN you can create fake images of bedrooms, as done by [Radford et al. in 2015](https://arxiv.org/pdf/1511.06434.pdf) and depicted below. \n\n\n\nAt their heart, GANs rely on the idea that a data generator is good\nif we cannot tell fake data apart from real data. \nIn statistics, this is called a two-sample test - a test to answer the question whether datasets $X = \\{x_1, \\ldots x_n\\}$ and $X' = \\{x_1', \\ldots x_n'\\}$ were drawn from the same distribution. \nThe main difference between most statistics papers and GANs is that the latter use this idea in a constructive way.\nIn other words, rather than just training a model to say 'hey, these two datasets don't look like they came from the same distribution', they use the two-sample test to provide training signal to a generative model.\nThis allows us to improve the data generator until it generates something that resembles the real data. \nAt the very least, it needs to fool the classifier. And if our classifier is a state of the art deep neural network.\n\nAs you can see, there are two pieces to GANs - first off, we need a device (say, a deep network but it really could be anything, such as a game rendering engine) that might potentially be able to generate data that looks just like the real thing. \nIf we are dealing with images, this needs to generate images. \nIf we're dealing with speech, it needs to generate audio sequences, and so on. \nWe call this the *generator network*. The second component is the *discriminator network*. \nIt attempts to distinguish fake and real data from each other. \nBoth networks are in competition with each other. \nThe generator network attempts to fool the discriminator network. At that point, the discriminator network adapts to the new fake data. This information, in turn is used to improve the generator network, and so on. \n\n**Generator**\n* Draw some parameter $z$ from a source of randomness, e.g. a normal distribution $z \\sim \\mathcal{N}(0,1)$.\n* Apply a function $f$ such that we get $x' = G(u,w)$\n* Compute the gradient with respect to $w$ to minimize $\\log p(y = \\mathrm{fake}|x')$ \n\n**Discriminator**\n* Improve the accuracy of a binary classifier $f$, i.e. maximize $\\log p(y=\\mathrm{fake}|x')$ and $\\log p(y=\\mathrm{true}|x)$ for fake and real data respectively.\n\n\n\n\nIn short, there are two optimization problems running simultaneously, and the optimization terminates if a stalemate has been reached. There are lots of further tricks and details on how to modify this basic setting. For instance, we could try solving this problem in the presence of side information. This leads to cGAN, i.e. conditional Generative Adversarial Networks. We can change the way how we detect whether real and fake data look the same. This leads to wGAN (Wasserstein GAN), kernel-inspired GANs and lots of other settings, or we could change how closely we look at the objects. E.g. fake images might look real at the texture level but not so at the larger level, or vice versa. \n\nMany of the applications are in the context of images. Since this takes too much time to solve in a Jupyter notebook on a laptop, we're going to content ourselves with fitting a much simpler distribution. We will illustrate what happens if we use GANs to build the world's most inefficient estimator of parameters for a Gaussian. Let's get started.\n", "_____no_output_____" ] ], [ [ "from __future__ import print_function\nimport matplotlib as mpl\nfrom matplotlib import pyplot as plt\nimport mxnet as mx\nfrom mxnet import gluon, autograd, nd\nfrom mxnet.gluon import nn\nimport numpy as np\n\nctx = mx.cpu()", "_____no_output_____" ] ], [ [ "## Generate some 'real' data\n\nSince this is going to be the world's lamest example, we simply generate data drawn from a Gaussian. And let's also set a context where we'll do most of the computation.", "_____no_output_____" ] ], [ [ "X = nd.random_normal(shape=(1000, 2))\nA = nd.array([[1, 2], [-0.1, 0.5]])\nb = nd.array([1, 2])\nX = nd.dot(X, A) + b\nY = nd.ones(shape=(1000, 1))\n\n# and stick them into an iterator\nbatch_size = 4\ntrain_data = mx.io.NDArrayIter(X, Y, batch_size, shuffle=True)", "_____no_output_____" ] ], [ [ "Let's see what we got. This should be a Gaussian shifted in some rather arbitrary way with mean $b$ and covariance matrix $A^\\top A$.", "_____no_output_____" ] ], [ [ "plt.scatter(X[:,0].asnumpy(), X[:,1].asnumpy())\nplt.show()\nprint(\"The covariance matrix is\")\nprint(nd.dot(A.T, A))", "_____no_output_____" ] ], [ [ "## Defining the networks\n\nNext we need to define how to fake data. Our generator network will be the simplest network possible - a single layer linear model. This is since we'll be driving that linear network with a Gaussian data generator. Hence, it literally only needs to learn the parameters to fake things perfectly. For the discriminator we will be a bit more discriminating: we will use an MLP with 3 layers to make things a bit more interesting. \n\nThe cool thing here is that we have *two* different networks, each of them with their own gradients, optimizers, losses, etc. that we can optimize as we please. ", "_____no_output_____" ] ], [ [ "# build the generator\nnetG = nn.Sequential()\nwith netG.name_scope():\n netG.add(nn.Dense(2))\n\n# build the discriminator (with 5 and 3 hidden units respectively)\nnetD = nn.Sequential()\nwith netD.name_scope():\n netD.add(nn.Dense(5, activation='tanh'))\n netD.add(nn.Dense(3, activation='tanh'))\n netD.add(nn.Dense(2))\n\n# loss\nloss = gluon.loss.SoftmaxCrossEntropyLoss()\n\n# initialize the generator and the discriminator\nnetG.initialize(mx.init.Normal(0.02), ctx=ctx)\nnetD.initialize(mx.init.Normal(0.02), ctx=ctx)\n\n# trainer for the generator and the discriminator\ntrainerG = gluon.Trainer(netG.collect_params(), 'adam', {'learning_rate': 0.01})\ntrainerD = gluon.Trainer(netD.collect_params(), 'adam', {'learning_rate': 0.05})", "_____no_output_____" ] ], [ [ "## Setting up the training loop\n\nWe are going to iterate over the data a few times. To make life simpler we need a few variables", "_____no_output_____" ] ], [ [ "real_label = mx.nd.ones((batch_size,), ctx=ctx)\nfake_label = mx.nd.zeros((batch_size,), ctx=ctx)\nmetric = mx.metric.Accuracy()\n\n# set up logging\nfrom datetime import datetime\nimport os\nimport time", "_____no_output_____" ] ], [ [ "## Training loop\n\n", "_____no_output_____" ] ], [ [ "stamp = datetime.now().strftime('%Y_%m_%d-%H_%M')\nfor epoch in range(10):\n tic = time.time()\n train_data.reset()\n for i, batch in enumerate(train_data):\n ############################\n # (1) Update D network: maximize log(D(x)) + log(1 - D(G(z)))\n ###########################\n # train with real_t\n data = batch.data[0].as_in_context(ctx)\n noise = nd.random_normal(shape=(batch_size, 2), ctx=ctx)\n\n with autograd.record():\n real_output = netD(data)\n errD_real = loss(real_output, real_label)\n \n fake = netG(noise)\n fake_output = netD(fake.detach())\n errD_fake = loss(fake_output, fake_label)\n errD = errD_real + errD_fake\n errD.backward()\n\n trainerD.step(batch_size)\n metric.update([real_label,], [real_output,])\n metric.update([fake_label,], [fake_output,])\n\n ############################\n # (2) Update G network: maximize log(D(G(z)))\n ###########################\n with autograd.record():\n output = netD(fake)\n errG = loss(output, real_label)\n errG.backward()\n\n trainerG.step(batch_size)\n\n name, acc = metric.get()\n metric.reset()\n print('\\nbinary training acc at epoch %d: %s=%f' % (epoch, name, acc))\n print('time: %f' % (time.time() - tic))\n noise = nd.random_normal(shape=(100, 2), ctx=ctx)\n fake = netG(noise)\n plt.scatter(X[:,0].asnumpy(), X[:,1].asnumpy())\n plt.scatter(fake[:,0].asnumpy(), fake[:,1].asnumpy())\n plt.show()", "\nbinary training acc at epoch 0: accuracy=0.764500\ntime: 5.838877\n" ] ], [ [ "## Checking the outcome\n\nLet's now generate some fake data and check whether it looks real.", "_____no_output_____" ] ], [ [ "noise = mx.nd.random_normal(shape=(100, 2), ctx=ctx)\nfake = netG(noise)\n\nplt.scatter(X[:,0].asnumpy(), X[:,1].asnumpy())\nplt.scatter(fake[:,0].asnumpy(), fake[:,1].asnumpy())\nplt.show()", "_____no_output_____" ] ], [ [ "## Conclusion \n\nA word of caution here - to get this to converge properly, we needed to adjust the learning rates *very carefully*. And for Gaussians, the result is rather mediocre - a simple mean and covariance estimator would have worked *much better*. However, whenever we don't have a really good idea of what the distribution should be, this is a very good way of faking it to the best of our abilities. Note that a lot depends on the power of the discriminating network. If it is weak, the fake can be very different from the truth. E.g. in our case it had trouble picking up anything along the axis of reduced variance. \nIn summary, this isn't exactly easy to set and forget. One nice resource for dirty practioner's knowledge is [Soumith Chintala's handy list of tricks](https://github.com/soumith/ganhacks) for how to babysit GANs. ", "_____no_output_____" ], [ "For whinges or inquiries, [open an issue on GitHub.](https://github.com/zackchase/mxnet-the-straight-dope)", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown" ]

[ [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "*This notebook is part of course materials for CS 345: Machine Learning Foundations and Practice at Colorado State University.\nOriginal versions were created by Asa Ben-Hur.\nThe content is availabe [on GitHub](https://github.com/asabenhur/CS345).*\n\n*The text is released under the [CC BY-SA license](https://creativecommons.org/licenses/by-sa/4.0/), and code is released under the [MIT license](https://opensource.org/licenses/MIT).*\n\n<img style=\"padding: 10px; float:right;\" alt=\"CC-BY-SA icon.svg in public domain\" src=\"https://upload.wikimedia.org/wikipedia/commons/d/d0/CC-BY-SA_icon.svg\" width=\"125\">\n", "_____no_output_____" ], [ "<a href=\"https://colab.research.google.com/github//asabenhur/CS345/blob/master/notebooks/module05_01_cross_validation.ipynb\">\n <img align=\"left\" src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/>\n</a>", "_____no_output_____" ] ], [ [ "import numpy as np\nimport matplotlib.pyplot as plt\n%matplotlib inline\n%autosave 0", "_____no_output_____" ] ], [ [ "# Evaluating classifiers: cross validation \n\n### Learning curves\n\nIntuitively, the more data we have available, the more accurate our classifiers become. To demonstrate this, let's read in some data and evaluate a k-nearest neighbor classifier on a fixed test set with increasing number of training examples. The resulting curve of accuracy as a function of number of examples is called a **learning curve**.\n\n", "_____no_output_____" ] ], [ [ "from sklearn.datasets import load_digits\nfrom sklearn.model_selection import train_test_split\nfrom sklearn.neighbors import KNeighborsClassifier\n\nX, y = load_digits(return_X_y=True)\n\ntraining_sizes = [20, 40, 100, 200, 400, 600, 800, 1000, 1200]\n\n# note the use of the stratify keyword: it makes it so that each \n# class is equally represented in both train and test set\nX_full_train, X_test, y_full_train, y_test = train_test_split(\n X, y, test_size = len(y)-max(training_sizes), \n stratify=y, random_state=1)\n\naccuracy = []\nfor training_size in training_sizes :\n X_train,_ , y_train,_ = train_test_split(\n X_full_train, y_full_train, test_size = \n len(y_full_train)-training_size+10, stratify=y_full_train)\n knn = KNeighborsClassifier(n_neighbors=1)\n knn.fit(X_train, y_train)\n y_pred = knn.predict(X_test)\n accuracy.append(np.sum((y_pred==y_test))/len(y_test))\n", "_____no_output_____" ], [ "plt.figure(figsize=(6,4))\nplt.plot(training_sizes, accuracy, 'ob')\nplt.xlabel('training set size')\nplt.ylabel('accuracy')\nplt.ylim((0.5,1));", "_____no_output_____" ] ], [ [ "It's also instructive to look at the numbers themselves:", "_____no_output_____" ] ], [ [ "print (\"# training examples\\t accuracy\")\nfor i in range(len(accuracy)) :\n print (\"\\t{:d}\\t\\t {:f}\".format(training_sizes[i], accuracy[i]))\n", "# training examples\t accuracy\n\t20\t\t 0.636516\n\t40\t\t 0.889447\n\t100\t\t 0.914573\n\t200\t\t 0.953099\n\t400\t\t 0.966499\n\t600\t\t 0.979899\n\t800\t\t 0.983250\n\t1000\t\t 0.983250\n\t1200\t\t 0.983250\n" ] ], [ [ "### Exercise\n\n* What can you conclude from this plot? \n* Why would you want to compute a learning curve on your data?\n", "_____no_output_____" ], [ "### Making better use of our data with cross validation\n\nThe discussion above demonstrates that it is best to have as large of a training set as possible. We also need to have a large enough test set, so that the accuracy estimates are accurate. How do we balance these two contradictory requirements? Cross-validation provides us a more effective way to make use of our data. Here it is:\n\n**Cross validation**\n\n* Randomly partition the data into $k$ subsets (\"folds\").\n* Set one fold aside for evaluation and train a model on the remaining $k$ folds and evaluate it on the held-out fold.\n* Repeat until each fold has been used for evaluation\n* Compute accuracy by averaging over the accuracy estimates generated for each fold.\n\nHere is an illustration of 8-fold cross validation:\n\n<img style=\"padding: 10px; float:left;\" alt=\"cross-validation by MBanuelos22 CC BY-SA 4.0\" src=\"https://upload.wikimedia.org/wikipedia/commons/c/c7/LOOCV.gif\">\n\n width=\"600\">\n\nAs you can see, this procedure is more expensive than dividing your data into train and test set. When dealing with relatively small datasets, which is when you want to use this procedure, this won't be an issue.\n\nTypically cross-validation is used with the number of folds being in the range of 5-10. An extreme case is when the number of folds equals the number of training examples. This special case is called *leave-one-out cross-validation*.", "_____no_output_____" ] ], [ [ "from sklearn.linear_model import LogisticRegression\nfrom sklearn.model_selection import cross_validate\nfrom sklearn.model_selection import cross_val_score\n\nfrom sklearn import metrics", "_____no_output_____" ] ], [ [ "Let's use the scikit-learn breast cancer dataset to demonstrate the use of cross-validation.", "_____no_output_____" ] ], [ [ "from sklearn.datasets import load_breast_cancer\ndata = load_breast_cancer()", "_____no_output_____" ] ], [ [ "A scikit-learn data object is container object with whose interesting attributes are: \n * ‘data’, the data to learn, \n * ‘target’, the classification labels, \n * ‘target_names’, the meaning of the labels,\n * ‘feature_names’, the meaning of the features, and \n * ‘DESCR’, the full description of the dataset.\n\n", "_____no_output_____" ] ], [ [ "X = data.data\ny = data.target\nprint('number of examples ', len(y))\nprint('number of features ', len(X[0]))\n\nprint(data.target_names)\nprint(data.feature_names)", "number of examples 569\nnumber of features 30\n['malignant' 'benign']\n['mean radius' 'mean texture' 'mean perimeter' 'mean area'\n 'mean smoothness' 'mean compactness' 'mean concavity'\n 'mean concave points' 'mean symmetry' 'mean fractal dimension'\n 'radius error' 'texture error' 'perimeter error' 'area error'\n 'smoothness error' 'compactness error' 'concavity error'\n 'concave points error' 'symmetry error' 'fractal dimension error'\n 'worst radius' 'worst texture' 'worst perimeter' 'worst area'\n 'worst smoothness' 'worst compactness' 'worst concavity'\n 'worst concave points' 'worst symmetry' 'worst fractal dimension']\n" ], [ "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.4, \n random_state=0)\nclassifier = KNeighborsClassifier(n_neighbors=3)\n#classifier = LogisticRegression()\n\n_ = classifier.fit(X_train, y_train)\ny_pred = classifier.predict(X_test)\n", "_____no_output_____" ] ], [ [ "Let's compute the accuracy of our predictions:", "_____no_output_____" ] ], [ [ "np.mean(y_pred==y_test)", "_____no_output_____" ] ], [ [ "We can do the same using scikit-learn:", "_____no_output_____" ] ], [ [ "metrics.accuracy_score(y_test, y_pred)", "_____no_output_____" ] ], [ [ "Now let's compute accuracy using [cross_validate](https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.cross_validate.html) instead:", "_____no_output_____" ] ], [ [ "accuracy = cross_val_score(classifier, X, y, cv=5, \n scoring='accuracy')\nprint(accuracy)", "[0.87719298 0.92105263 0.94736842 0.93859649 0.91150442]\n" ] ], [ [ "This yields an array containing the accuracy values for each fold.\nWhen reporting your results, you will typically show the mean:", "_____no_output_____" ] ], [ [ "np.mean(accuracy)", "_____no_output_____" ] ], [ [ "The arguments of `cross_val_score`:\n* A classifier (anything that satisfies the scikit-learn classifier API)\n* data (features/labels)\n* `cv` : an integer that specifies the number of folds (can be used in more sophisticated ways as we will see below).\n* `scoring`: this determines which accuracy measure is evaluated for each fold. Here's a link to the [list of available measures](https://scikit-learn.org/stable/modules/model_evaluation.html#scoring-parameter) in scikit-learn.\n\nYou can obtain accuracy for other metrics. *Balanced accuracy* for example, is appropriate when the data is unbalanced (e.g. when one class contains a much larger number of examples than other classes in the data).", "_____no_output_____" ] ], [ [ "accuracy = cross_val_score(classifier, X, y, cv=5, \n scoring='balanced_accuracy')\nnp.mean(accuracy)", "_____no_output_____" ] ], [ [ "`cross_val_score` is somewhat limited, in that it simply returns a list of accuracy scores. In practice, we often want to have more information about what happened during training, and also to compute multiple accuracy measures.\n`cross_validate` will provide you with that information:", "_____no_output_____" ] ], [ [ "results = cross_validate(classifier, X, y, cv=5, \n scoring='accuracy', return_estimator=True)\nprint(results)", "{'fit_time': array([0.0010581 , 0.00099397, 0.00085902, 0.00093985, 0.00105977]), 'score_time': array([0.00650811, 0.00734997, 0.01043916, 0.00643301, 0.00529218]), 'estimator': (KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='minkowski',\n metric_params=None, n_jobs=None, n_neighbors=3, p=2,\n weights='uniform'), KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='minkowski',\n metric_params=None, n_jobs=None, n_neighbors=3, p=2,\n weights='uniform'), KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='minkowski',\n metric_params=None, n_jobs=None, n_neighbors=3, p=2,\n weights='uniform'), KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='minkowski',\n metric_params=None, n_jobs=None, n_neighbors=3, p=2,\n weights='uniform'), KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='minkowski',\n metric_params=None, n_jobs=None, n_neighbors=3, p=2,\n weights='uniform')), 'test_score': array([0.87719298, 0.92105263, 0.94736842, 0.93859649, 0.91150442])}\n" ] ], [ [ "The object returned by `cross_validate` is a Python dictionary as the output suggests. To extract a specific piece of data from this object, simply access the dictionary with the appropriate key:", "_____no_output_____" ] ], [ [ "results['test_score']", "_____no_output_____" ] ], [ [ "If you would like to know the predictions made for each training example during cross-validation use [cross_val_predict](https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.cross_val_predict.html) instead:", "_____no_output_____" ] ], [ [ "from sklearn.model_selection import cross_val_predict\n\ny_pred = cross_val_predict(classifier, X, y, cv=5)\nmetrics.accuracy_score(y, y_pred)", "_____no_output_____" ] ], [ [ "The above way of performing cross-validation doesn't always give us enough control on the process: we usually want our machine learning experiments be reproducible, and to be able to use the same cross-validation splits with multiple algorithms. The scikit-learn `KFold` and `StratifiedKFold` cross-validation generators are the way to achieve that. \n\n`KFold` simply chooses a random subset of examples for each fold. This strategy can lead to cross-validation folds in which the classes are not well-represented as the following toy example demonstrates:", "_____no_output_____" ] ], [ [ "from sklearn.model_selection import StratifiedKFold, KFold\n\nX_toy = np.array([[1, 2], [3, 4], [5, 6], [7, 8], [9,10], [11, 12]])\ny_toy = np.array([0, 0, 1, 1, 1, 1])\ncv = KFold(n_splits=2, random_state=3, shuffle=True)\nfor train_idx, test_idx in cv.split(X_toy, y_toy):\n print(\"train:\", train_idx, \"test:\", test_idx)\n X_train, X_test = X_toy[train_idx], X_toy[test_idx]\n y_train, y_test = y_toy[train_idx], y_toy[test_idx]\n print(y_train)", "train: [0 1 2] test: [3 4 5]\n[0 0 1]\ntrain: [3 4 5] test: [0 1 2]\n[1 1 1]\n" ] ], [ [ "`StratifiedKFold` addresses this issue by making sure that each class is represented in each fold in proportion to its overall fraction in the data. This is particularly important when one or more of the classes have few examples.\n\n`StratifiedKFold` and `KFold` generate folds that can be used in conjunction with the cross-validation methods we saw above.\nAs an example, we will demonstrate the use of `StratifiedKFold` with `cross_val_score` on the breast cancer datast:", "_____no_output_____" ] ], [ [ "cv = StratifiedKFold(n_splits=5, random_state=1, shuffle=True)\naccuracy = cross_val_score(classifier, X, y, cv=cv, \n scoring='accuracy')\nnp.mean(accuracy)", "_____no_output_____" ] ], [ [ "For classification problems, `StratifiedKFold` is the preferred strategy. However, for regression problems `KFold` is the way to go.\n\n#### Question\n\nWhy is `KFold` used in regression probelms rather than `StratifiedKFold`?", "_____no_output_____" ], [ "To clarify the distinction between the different methods of generating cross-validation folds and their different parameters let's look at the following figures:", "_____no_output_____" ] ], [ [ "# the code for the figure is adapted from\n# https://scikit-learn.org/stable/auto_examples/model_selection/plot_cv_indices.html\n\nnp.random.seed(42)\ncmap_data = plt.cm.Paired\ncmap_cv = plt.cm.coolwarm\nn_folds = 4\n\n# Generate the data\nX = np.random.randn(100, 10)\n\n# generate labels - classes 0,1,2 and 10,30,60 examples, respectively\ny = np.array([0] * 10 + [1] * 30 + [2] * 60)\n\ndef plot_cv_indices(cv, X, y, ax, n_folds):\n \"\"\"plot the indices of a cross-validation object.\"\"\"\n\n # Generate the training/testing visualizations for each CV split\n for ii, (tr, tt) in enumerate(cv.split(X=X, y=y)):\n # Fill in indices with the training/test groups\n indices = np.zeros(len(X))\n indices[tt] = 1\n # Visualize the results\n ax.scatter(range(len(indices)), [ii + .5] * len(indices),\n c=indices, marker='_', lw=15, cmap=cmap_cv,\n vmin=-.2, vmax=1.2)\n\n # Plot the data classes and groups at the end\n ax.scatter(range(len(X)), [ii + 1.5] * len(X), c=y, marker='_', lw=15, cmap=cmap_data)\n\n # Formatting\n yticklabels = list(range(n_folds)) + ['class']\n ax.set(yticks=np.arange(n_folds+2) + .5, yticklabels=yticklabels, \n xlabel='index', ylabel=\"CV fold\",\n ylim=[n_folds+1.2, -.2], xlim=[0, 100])\n ax.set_title('{}'.format(type(cv).__name__), fontsize=15)\n return ax\n", "_____no_output_____" ] ], [ [ "Let's visualize the results of using `KFold` for fold generation:", "_____no_output_____" ] ], [ [ "fig, ax = plt.subplots()\ncv = KFold(n_folds)\nplot_cv_indices(cv, X, y, ax, n_folds);", "_____no_output_____" ] ], [ [ "As you can see, this naive way of using `KFold` can lead to highly undesirable splits into cross-validation folds.\nUsing `StratifiedKFold` addresses this to some extent:", "_____no_output_____" ] ], [ [ "fig, ax = plt.subplots()\ncv = StratifiedKFold(n_folds)\nplot_cv_indices(cv, X, y, ax, n_folds);", "_____no_output_____" ] ], [ [ "Using `StratifiedKFold` with shuffling of the examples is the preferred way of splitting the data into folds:", "_____no_output_____" ] ], [ [ "fig, ax = plt.subplots()\ncv = StratifiedKFold(n_folds, shuffle=True)\nplot_cv_indices(cv, X, y, ax, n_folds);", "_____no_output_____" ] ], [ [ "### Question\n\nConsider the task of digitizing handwritten text (aka optical character recognition, or OCR). For each letter in the alphabet you have multiple labeled examples generated by the same writer. How would this setup affect the way you divide your examples into training and test sets, or when performing cross-validation?\n", "_____no_output_____" ], [ "### Summary and Discussion\n\nIn this notebook we discussed cross-validation as a more effective way to make use of limited amounts of data compared to the strategy of splitting data into train and test sets. For very large datasets where training is time consuming you might still opt for evaluation on a single test set.\n", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown" ]

[ [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ] ]

[ "CC-BY-4.0" ]

[ "CC-BY-4.0" ]

[ "CC-BY-4.0" ]

[ [ [ "# Lecture 3.3: Anomaly Detection\n\n[**Lecture Slides**](https://docs.google.com/presentation/d/1_0Z5Pc5yHA8MyEBE8Fedq44a-DcNPoQM1WhJN93p-TI/edit?usp=sharing)\n\nThis lecture, we are going to use gaussian distributions to detect anomalies in our emoji faces dataset\n\n**Learning goals:**\n\n- Introduce an anomaly detection problem\n- Implement Gaussian distribution anomaly detection for images\n- Debug the optimisation of a learning algorithm\n- Discuss the imperfection of learning algorithms\n- Acknowledge other outlier detection methods", "_____no_output_____" ], [ "\n## 1. Introduction\n\nWe have an `emoji_faces` dataset of all our favourite emojis. However, Skynet hates their friendly expressiveness, and wants to destroy emojis forever! 🙀 It sent _terminator robots_ from the future to invade our dataset. We must act fast, and detect them amongst the emojis to prevent the catastrophy. \n\nOur challenge here, is that we don't watch many movies, so we don't have a clear idea of what those _terminators_ look like. 🤖 All we know, is that they look very different compared to emojis, and that only a handful managed to infiltrate our dataset.\n\nThis is a typical scenario of _anomaly detection_. We would like to identify rare examples that differ from our \"normal\" data points. We choose to use a Gaussian Distribution to model this \"normality\" and detect the killer robots.", "_____no_output_____" ], [ "\n## 2. Data Munging\n\nFirst let's load the images using [pillow](https://pillow.readthedocs.io/en/stable/), like in lecture 2.5:", "_____no_output_____" ] ], [ [ "from PIL import Image\nimport glob\n\npaths = glob.glob('emoji_faces/*.png')\n\nimages = [Image.open(path) for path in paths]\nlen(images)", "_____no_output_____" ] ], [ [ "We have 134 emoji faces, including a few terminator robots. We'll again be using the [sklearn](https://scikit-learn.org/) library to create our model. The interface is usually the same, and for gaussian anomaly detection, sklearn again expect a NumPy matrix where the rows are our images and the columns are the pixels. So we can apply the same transformations as notebook 3.2:", "_____no_output_____" ] ], [ [ "import numpy as np\n\narrays = [np.asarray(im) for im in images]\n# 64 * 64 = 4096\nvectors = [arr.reshape((4096,)) for arr in arrays]\ndata = np.stack(vectors)", "_____no_output_____" ] ], [ [ "## 3. Training \n\nNext, we will create an [`EllipticEnvelope`](https://scikit-learn.org/stable/modules/generated/sklearn.covariance.EllipticEnvelope.html) object. This will fit a multi-variate gaussian distribution to our data. It then allows us to pick a threshold to define an _ellipsoid_ decision boundary , and detect outliers. \n\nRemember that we are using a _learning_ algorithm, which must therefore be _trained_ before it can be used. This is why we'll use the `.fit()` method first, before calling `.predict()`:", "_____no_output_____" ] ], [ [ "from sklearn.covariance import EllipticEnvelope\n\ncov = EllipticEnvelope(random_state=0).fit(data)", "_____no_output_____" ] ], [ [ "😰 What's happening? Why is it stuck? Have the killer robots already taken over? \n\nNo need to panic, this kind of hiccup is very common when dealing with machine learning algorithms. We can kill the process (before it fries our laptop fan) by clicking the `stop` button ⬛️ in the notebook toolbar.\n\nMost learning algorithms are based around an _optimisation_ procedure. This step is often iterative and stochastic, i.e it tries its statistical best to maximise the learning in incremental steps. \n\nThis process isn't fail proof:\n* it can dramatically stop because of out of memory errors, or overflow errors 💥\n* it can get stuck, e.g when the optimisation is too slow 🐌\n* it can fail silently, and return wrong results 💩\n\nℹ️ We will encounter many of these failures throughout our ML experiments, so knowing how to overcome them is a part of the data scientist skillset. \n\nLet's go back to our killer robot detection: the model fitting got _stuck_ , which suggests that something about our data was too much to handle. We find the following \"notes\" in the [official documentation](https://scikit-learn.org/stable/modules/generated/sklearn.covariance.EllipticEnvelope.html#sklearn.covariance.EllipticEnvelope):\n\n> Outlier detection from covariance estimation may break or not perform well in high-dimensional settings.\n\nWe recall that our images are $64 \\times 64$ pixels, so $4096$ dimensions.... that's a lot. It seems a good candidate to explain why our multivariate gaussian distribution failed to fit our dataset. If only there was a way to reduce the dimensions of our data... 😏\n\nLet's apply PCA to reduce the number of dimensions of our dataset. Our emoji faces dataset is smaller than the full emoji dataset, so 40 dimensions should suffice to explain its variance: \n", "_____no_output_____" ] ], [ [ "from sklearn.decomposition import PCA\n\npca = PCA(n_components=40)\npca.fit(data)\ncomponents = pca.transform(data)\ncomponents.shape", "_____no_output_____" ] ], [ [ "💪 Visualise the eigenvector images of our PCA model. You can use the code from lecture 3.2!\n\n🧠 Can you explain what those eigenvector images represent? Why are they different than from the full emoji dataset?\n\nFantastic, we've managed to reduce the number of dimensions by 99%! Hopefully that should be enough to make our gaussian distribution fitting happy. Let's try again with the _principal components_ instead of the original data:", "_____no_output_____" ] ], [ [ "cov = EllipticEnvelope(random_state=0).fit(components)", "_____no_output_____" ] ], [ [ "😅 that was fast!\n\n## 4. Prediction\n\nWe can now use our fitted gaussian distribution to detect the outliers in our `data`. For this, we use the `.predict()` method:", "_____no_output_____" ] ], [ [ "y = cov.predict(components)\ny", "_____no_output_____" ] ], [ [ "`y` is our vector of predictions, where $1$ is a normal data point, and $-1$ is an anomaly. We can therefore iterate through our original `arrays` to find outliers:", "_____no_output_____" ] ], [ [ "outliers = []\n\nfor i in range(0, len(arrays)):\n if y[i] == -1:\n outliers.append(arrays[i])\nlen(outliers)", "_____no_output_____" ], [ "import matplotlib.pyplot as plt\n\nfig, axs = plt.subplots(dpi=150, nrows=2, ncols=7)\nfor outlier, ax in zip(outliers, axs.flatten()):\n ax.imshow(outlier, cmap='gray', vmin=0, vmax=255)\n ax.get_xaxis().set_visible(False)\n ax.get_yaxis().set_visible(False)", "_____no_output_____" ] ], [ [ "THERE'S OUR TERMINATORS! 🤖 We can count 5 of them in total. Notice how some real emoji faces were also detected as outliers. This is perhaps a sign that we should change our _threshold_ , to make the ellipsoid decision boundary smaller. \n\nIn fact, we didn't even specify a threshold before, we just used the default value of `contamination=0.1` in the [`EllipticEnvelope`](https://scikit-learn.org/stable/modules/generated/sklearn.covariance.EllipticEnvelope.html) class. This represents our estimation of the proportion of data points which are outliers. Since it looks like we detected double the amount of actual anomalies, let's try again with `contamination=0.05`:", "_____no_output_____" ] ], [ [ "cov = EllipticEnvelope(random_state=0, contamination=0.05).fit(components)\ny = cov.predict(components)\n\noutliers = []\n\nfor i in range(0, len(arrays)):\n if y[i] == -1:\n outliers.append(arrays[i])\n \nfig, axs = plt.subplots(dpi=150, nrows=1, ncols=7)\nfor outlier, ax in zip(outliers, axs.flatten()):\n ax.imshow(outlier, cmap='gray', vmin=0, vmax=255)\n ax.get_xaxis().set_visible(False)\n ax.get_yaxis().set_visible(False)", "_____no_output_____" ] ], [ [ "Better! `contamination=0.05` was a better choice of threshold, and we assessed this through _manual inspection_. This means we went through the results and used our human jugement to change the value of this _hyperparameter_.\n\nℹ️ Notice how our outlier detection is not _perfect_. Some emojis were also erroneously detected as anomalous killer robots. This can seem like a problem, or a sign that our model was malfunctioning. But, quite the contrary, _imperfection_ is a core aspect of all _learning_ algorithms. Instead of seeing the glass half-empty and looking at the outlier detector's mistakes, we should reflect on the task itself. It would have been almost impossible to detect those killer robot images using rule-based algorithms, and our model _accuracy_ was good _enough_ to save the emojis from Skynet. As data scientists, our goal is to make models which are accurate _enough_ to be useful, not to aim for perfect scores. We will revisit these topics later in the course when discussing Machine Learning Engineering 🛠", "_____no_output_____" ], [ "## 5. Analysis", "_____no_output_____" ], [ "We have detected the robot intruders and saved the emojis from a jealous AI from the future, all is good! We still want to better understand how anomaly detection defeated Skynet. For this, we would like to leverage our shiny new data visualization skills. Representing our dataset in space would allow us to identify its structures and hopefully understand how our gaussian distribution model identified terminators as \"abnormal\".\n\nOur data is high dimensional, so we can use our trusted PCA once again to project it down to 2 dimensions. We understand that this will lose a lot of the variance of our data, but the results were still somewhat interpretable with the full emoji dataset, so let's go!", "_____no_output_____" ] ], [ [ "# Dimesionality reduction to 2\npca_model = PCA(n_components=2)\npca_model.fit(data) # fit the model\nT = pca_model.transform(data) # transform the 'normalized model'\n\nplt.scatter(T[:, 0], T[:, 1],\n # use the predictions as color\n c=y, \n marker='o',\n alpha=0.4\n )\nplt.title('Anomaly detection of the emoji faces dataset with PCA dimensionality reduction');", "_____no_output_____" ] ], [ [ "We can notice that most of the outliers are clearly _separable_ from the bulk of the dataset, even with only 2 principal components. One outlier is very much within the main cluster however. This could be explained by the dimensionality reduction, i.e that this point is separated from the cluster in other dimensions, or by the fact our threshold might be too permissive.\n\nWe can check this by displaying the images directly on the scatter plot:", "_____no_output_____" ] ], [ [ "from matplotlib import offsetbox\n\ndef plot_components(data, model, images=None, ax=None,\n thumb_frac=0.05, cmap='gray'):\n ax = ax or plt.gca()\n \n proj = model.fit_transform(data)\n ax.plot(proj[:, 0], proj[:, 1], '.k')\n \n if images is not None:\n min_dist_2 = (thumb_frac * max(proj.max(0) - proj.min(0))) ** 2\n shown_images = np.array([2 * proj.max(0)])\n for i in range(data.shape[0]):\n dist = np.sum((proj[i] - shown_images) ** 2, 1)\n if np.min(dist) < min_dist_2:\n # don't show points that are too close\n continue\n shown_images = np.vstack([shown_images, proj[i]])\n imagebox = offsetbox.AnnotationBbox(\n offsetbox.OffsetImage(images[i], cmap=cmap),\n proj[i])\n ax.add_artist(imagebox)", "_____no_output_____" ], [ "small_images = [im[::2, ::2] for im in arrays]\nfig, ax = plt.subplots(figsize=(10, 10))\nplot_components(data,\n model=PCA(n_components=2),\n images=small_images, thumb_frac=0.02)\nplt.title('Anomaly detection of the emoji faces dataset with PCA dimensionality reduction');", "_____no_output_____" ] ], [ [ "We could probably have reduced the value of `contamination` further, since we can see how the killer robots are clearly \"abnormal\" with this visualisation. We also have a \"feel\" of how our gaussian distribution model could successfully detect them as outliers. Although remember that all of modeling magic happens in 40 dimensional space!\n\n🧠🧠 Can you explain why it is not very useful to display the ellipsoid decision boundary of our anomaly detection model on this graph?", "_____no_output_____" ], [ "## 6. More Anomaly Detection\n\nAnomaly detection is an active field in ML research, which combines supervised, unsupervised, non-linear, Bayesian, ... a whole bunch of methods! Each solution will have its pros and cons, and developing a production level outlier detection system will require empirically evaluating and comparing them. For a breakdown of the methods available in sklearn, check out this excellent [blogpost](https://sdsawtelle.github.io/blog/output/week9-anomaly-andrew-ng-machine-learning-with-python.html), or the [official documentation](https://scikit-learn.org/stable/modules/outlier_detection.html). For an in-depth view of modern anomaly detection, watch this [video](https://youtu.be/LRqX5uO5StA). And for everything else, feel free to experiment with this dataset or any other. Good luck on finding all the killer robots!", "_____no_output_____" ], [ "## 7. Summary\n\nToday, we defined **anomaly detection**, and listed some of its common applications including fraud detection and data cleaning. We then described how to use **fitted Gaussian distributions** to identify outliers. This lead us to a discussion about the choice of **thresholds** and **hyperparameters**, where we went over a few different realistic scenarios. We then used a Gaussian distribution to remove terminator images from an emoji faces dataset. We learned how learning algorithms **fail** and that data scientists must know how to **debug** them. Finally, we used **PCA** to visualize our killer robot detection. \n\n\n# Resources\n\n## Core Resources\n\n- [Anomaly detection algorithm](https://www.coursera.org/lecture/machine-learning/algorithm-C8IJp) \nAndrew Ng's limpid breakdown of anomaly detection\n\n## Additional Resources\n\n- [A review of ML techniques for anomaly detection](https://youtu.be/LRqX5uO5StA) \nMore in depth review of modern techniques for anomaly detection\n- [Anomaly Detection in sklearn](https://sdsawtelle.github.io/blog/output/week9-anomaly-andrew-ng-machine-learning-with-python.html) \nVisual blogpost experimenting with the various outlier detection algorithms available in sklearn\n- [sklearn official documentation - outlier detection](https://scikit-learn.org/stable/modules/outlier_detection.html) ", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown" ]

[ [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown", "markdown", "markdown" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "from fastai.vision.all import *\nimport pandas as pd\nimport cam\nimport util", "_____no_output_____" ], [ "dls, labels = util.chexpert_data_loader()", "_____no_output_____" ], [ "dls.show_batch(max_n=9, figsize=(20,9))", "_____no_output_____" ], [ "# First train on conditional probabilities\nchexpert_learner_conditional = util.ChexpertLearner(dls, densenet121, n_out=len(labels), y_range=(0, 1),\n loss_func=util.BCEFlatHLCP(hierarchy_map=util.hierarchy_map),\n metrics=[RocAucMulti(average=None),\n RocAucMulti(average='weighted')])\nchexpert_learner_conditional.learn_model(use_saved=False , epochs=10, freeze_epochs=1)", "_____no_output_____" ], [ "# Next train unconditionally for only transfer learning\nprint('-------Running unconditional-------')\nchexpert_learner_unconditional = util.ChexpertLearner(dls, densenet121, n_out=len(labels), y_range=(0, 1),\n loss_func=BCELossFlat(),\n metrics=[RocAucMulti(average=None),\n RocAucMulti(average='weighted')])\nchexpert_learner_unconditional.learn_model(use_saved=True, train_saved=True, epochs=10, freeze_epochs=10)", "-------Running unconditional-------\n" ], [ "chexpert_learner = chexpert_learner_unconditional\ncam.plot_cam(chexpert_learner.learn)", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Import Necessary Libraries", "_____no_output_____" ] ], [ [ "import numpy as np\nimport pandas as pd\nfrom sklearn.model_selection import train_test_split\nfrom sklearn import preprocessing\nfrom sklearn.ensemble import RandomForestClassifier\nfrom sklearn import svm\nfrom sklearn.metrics import precision_score, recall_score\n# display images \nfrom IPython.display import Image\n# linear algebra\nimport numpy as np \n# data processing\nimport pandas as pd \n# data visualization\nimport seaborn as sns \n%matplotlib inline\nfrom matplotlib import pyplot as plt\nfrom matplotlib import style\n# Algorithms\nfrom sklearn import linear_model\nfrom sklearn.linear_model import LogisticRegression\nfrom sklearn.ensemble import RandomForestClassifier\nfrom sklearn.linear_model import Perceptron\nfrom sklearn.linear_model import SGDClassifier\nfrom sklearn.tree import DecisionTreeClassifier\nfrom sklearn.neighbors import KNeighborsClassifier\nfrom sklearn.svm import SVC, LinearSVC\nfrom sklearn.naive_bayes import GaussianNB", "_____no_output_____" ] ], [ [ "# Titanic \nTitanic was a British passenger liner that sank in the North Atlantic Ocean in the early morning hours of 15 April 1912, after it collided with an iceberg during its maiden voyage from Southampton to New York City. There were an estimated 2,224 passengers and crew aboard the ship, and more than 1,500 died, making it one of the deadliest commercial peacetime maritime disasters in modern history. The RMS Titanic was the largest ship afloat at the time it entered service and was the second of three Olympic-class ocean liners operated by the White Star Line. The Titanic was built by the Harland and Wolff shipyard in Belfast. Thomas Andrews, her architect, died in the disaster.", "_____no_output_____" ] ], [ [ "# Image of Titanic ship\nImage(filename='C:/Users/Nemgeree Armanonah/Documents/GitHub/Titanic/images/ship.jpeg')", "_____no_output_____" ] ], [ [ "# Getting the Data", "_____no_output_____" ] ], [ [ "#reading train.csv\ndata = pd.read_csv('./titanic datasets/train.csv')\ndata", "_____no_output_____" ] ], [ [ "## Exploring Data", "_____no_output_____" ] ], [ [ "data.info()", "<class 'pandas.core.frame.DataFrame'>\nRangeIndex: 891 entries, 0 to 890\nData columns (total 12 columns):\nPassengerId 891 non-null int64\nSurvived 891 non-null int64\nPclass 891 non-null int64\nName 891 non-null object\nSex 891 non-null object\nAge 714 non-null float64\nSibSp 891 non-null int64\nParch 891 non-null int64\nTicket 891 non-null object\nFare 891 non-null float64\nCabin 204 non-null object\nEmbarked 889 non-null object\ndtypes: float64(2), int64(5), object(5)\nmemory usage: 83.7+ KB\n" ] ], [ [ "### Describe Statistics \nDescribe method is used to view some basic statistical details like PassengerId,Servived,Age etc.", "_____no_output_____" ] ], [ [ "data.describe()", "_____no_output_____" ] ], [ [ "### View All Features ", "_____no_output_____" ] ], [ [ "data.columns.values", "_____no_output_____" ] ], [ [ "### What features could contribute to a high survival rate ?", "_____no_output_____" ], [ "To Us it would make sense if everything except ‘PassengerId’, ‘Ticket’ and ‘Name’ would be correlated with a high survival rate.", "_____no_output_____" ] ], [ [ "# defining variables\nsurvived = 'survived'\nnot_survived = 'not survived'\n# data to be plotted \nfig, axes = plt.subplots(nrows=1, ncols=2,figsize=(10, 4))\nwomen = data[data['Sex']=='female']\nmen = data[data['Sex']=='male']\n# plot the data\nax = sns.distplot(women[women['Survived']==1].Age.dropna(), bins=18, label = survived, ax = axes[0], kde =False)\nax = sns.distplot(women[women['Survived']==0].Age.dropna(), bins=40, label = not_survived, ax = axes[0], kde =False)\nax.legend()\nax.set_title('Female')\nax = sns.distplot(men[men['Survived']==1].Age.dropna(), bins=18, label = survived, ax = axes[1], kde = False)\nax = sns.distplot(men[men['Survived']==0].Age.dropna(), bins=40, label = not_survived, ax = axes[1], kde = False)\nax.legend()\n_ = ax.set_title('Male')", "_____no_output_____" ], [ "# count the null values \nnull_values = data.isnull().sum()\nnull_values", "_____no_output_____" ], [ "plt.plot(null_values)\nplt.grid()\nplt.show()", "_____no_output_____" ] ], [ [ "## Data Processing", "_____no_output_____" ] ], [ [ "\ndef handle_non_numerical_data(df):\n \n columns = df.columns.values\n\n for column in columns:\n text_digit_vals = {}\n def convert_to_int(val):10\n return text_digit_vals[val]\n\n #print(column,df[column].dtype)\n if df[column].dtype != np.int64 and df[column].dtype != np.float64:\n \n column_contents = df[column].values.tolist()\n #finding just the uniques\n unique_elements = set(column_contents)\n # great, found them. \n x = 0\n for unique in unique_elements:\n if unique not in text_digit_vals:\n text_digit_vals[unique] = x\n x+=1\n df[column] = list(map(convert_to_int,df[column]))\n\n return df", "_____no_output_____" ], [ "y_target = data['Survived']\n# Y_target.reshape(len(Y_target),1)\nx_train = data[['Pclass', 'Age', 'Sex', 'SibSp', 'Parch', 'Fare','Embarked', 'Ticket']]\n\nx_train = handle_non_numerical_data(x_train)\nx_train.head()", "_____no_output_____" ], [ "fare = pd.DataFrame(x_train['Fare'])\n# Normalizing\nmin_max_scaler = preprocessing.MinMaxScaler()\nnewfare = min_max_scaler.fit_transform(fare)\nx_train['Fare'] = newfare\nx_train", "c:\\users\\nemgeree armanonah\\appdata\\local\\programs\\python\\python36\\lib\\site-packages\\ipykernel_launcher.py:5: SettingWithCopyWarning: \nA value is trying to be set on a copy of a slice from a DataFrame.\nTry using .loc[row_indexer,col_indexer] = value instead\n\nSee the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n \"\"\"\n" ], [ "null_values = x_train.isnull().sum()\nnull_values", "_____no_output_____" ], [ "plt.plot(null_values)\nplt.show()", "_____no_output_____" ], [ "# Fill the NAN values with the median values in the datasets\nx_train['Age'] = x_train['Age'].fillna(x_train['Age'].median())\nprint(\"Number of NULL values\" , x_train['Age'].isnull().sum())\nx_train.head()", "Number of NULL values 0\n" ], [ "x_train['Sex'] = x_train['Sex'].replace('male', 0)\nx_train['Sex'] = x_train['Sex'].replace('female', 1)\n# print(type(x_train))\ncorr = x_train.corr()\ncorr.style.background_gradient()", "c:\\users\\nemgeree armanonah\\appdata\\local\\programs\\python\\python36\\lib\\site-packages\\ipykernel_launcher.py:1: SettingWithCopyWarning: \nA value is trying to be set on a copy of a slice from a DataFrame.\nTry using .loc[row_indexer,col_indexer] = value instead\n\nSee the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n \"\"\"Entry point for launching an IPython kernel.\nc:\\users\\nemgeree armanonah\\appdata\\local\\programs\\python\\python36\\lib\\site-packages\\ipykernel_launcher.py:2: SettingWithCopyWarning: \nA value is trying to be set on a copy of a slice from a DataFrame.\nTry using .loc[row_indexer,col_indexer] = value instead\n\nSee the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n \n" ], [ "def plot_corr(df,size=10):\n corr = df.corr()\n fig, ax = plt.subplots(figsize=(size, size))\n ax.matshow(corr)\n plt.xticks(range(len(corr.columns)), corr.columns);\n plt.yticks(range(len(corr.columns)), corr.columns);\n# plot_corr(x_train)\nx_train.corr()\ncorr.style.background_gradient()", "_____no_output_____" ], [ "# Dividing the data into train and test data set\nX_train, X_test, Y_train, Y_test = train_test_split(x_train, y_target, test_size = 0.4, random_state = 40)", "_____no_output_____" ], [ "clf = RandomForestClassifier()\nclf.fit(X_train, Y_train)", "c:\\users\\nemgeree armanonah\\appdata\\local\\programs\\python\\python36\\lib\\site-packages\\sklearn\\ensemble\\forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n" ], [ "print(clf.predict(X_test))\nprint(\"Accuracy: \",clf.score(X_test, Y_test))", "_____no_output_____" ], [ "## Testing the model.\ntest_data = pd.read_csv('./titanic datasets/test.csv')\ntest_data.head(3)\n# test_data.isnull().sum()", "_____no_output_____" ], [ "### Preprocessing on the test data\ntest_data = test_data[['Pclass', 'Age', 'Sex', 'SibSp', 'Parch', 'Fare', 'Ticket', 'Embarked']]\ntest_data = handle_non_numerical_data(test_data)\n\nfare = pd.DataFrame(test_data['Fare'])\nmin_max_scaler = preprocessing.MinMaxScaler()\nnewfare = min_max_scaler.fit_transform(fare)\ntest_data['Fare'] = newfare\ntest_data['Fare'] = test_data['Fare'].fillna(test_data['Fare'].median())\ntest_data['Age'] = test_data['Age'].fillna(test_data['Age'].median())\ntest_data['Sex'] = test_data['Sex'].replace('male', 0)\ntest_data['Sex'] = test_data['Sex'].replace('female', 1)\nprint(test_data.head())\n", "_____no_output_____" ], [ "print(clf.predict(test_data))", "_____no_output_____" ], [ "from sklearn.model_selection import cross_val_predict\npredictions = cross_val_predict(clf, X_train, Y_train, cv=3)\nprint(\"Precision:\", precision_score(Y_train, predictions))\nprint(\"Recall:\",recall_score(Y_train, predictions))", "_____no_output_____" ], [ "from sklearn.metrics import precision_recall_curve\n\n# getting the probabilities of our predictions\ny_scores = clf.predict_proba(X_train)\ny_scores = y_scores[:,1]\n\nprecision, recall, threshold = precision_recall_curve(Y_train, y_scores)\ndef plot_precision_and_recall(precision, recall, threshold):\n plt.plot(threshold, precision[:-1], \"r-\", label=\"precision\", linewidth=5)\n plt.plot(threshold, recall[:-1], \"b\", label=\"recall\", linewidth=5)\n plt.xlabel(\"threshold\", fontsize=19)\n plt.legend(loc=\"upper right\", fontsize=19)\n plt.ylim([0, 1])\n\nplt.figure(figsize=(14, 7))\nplot_precision_and_recall(precision, recall, threshold)\nplt.axis([0.3,0.8,0.8,1])\nplt.show()", "_____no_output_____" ], [ "def plot_precision_vs_recall(precision, recall):\n plt.plot(recall, precision, \"g--\", linewidth=2.5)\n plt.ylabel(\"recall\", fontsize=19)\n plt.xlabel(\"precision\", fontsize=19)\n plt.axis([0, 1.5, 0, 1.5])\n\nplt.figure(figsize=(14, 7))\nplot_precision_vs_recall(precision, recall)\nplt.show()", "_____no_output_____" ], [ "from sklearn.model_selection import cross_val_predict\nfrom sklearn.metrics import confusion_matrix\npredictions = cross_val_predict(clf, X_train, Y_train, cv=3)\nconfusion_matrix(Y_train, predictions)", "_____no_output_____" ] ], [ [ "True positive: 293 (We predicted a positive result and it was positive)\nTrue negative: 143 (We predicted a negative result and it was negative)\nFalse positive: 34 (We predicted a positive result and it was negative)\nFalse negative: 64 (We predicted a negative result and it was positive)", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown" ]

[ [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ], [ "markdown" ] ]

[ "BSD-3-Clause" ]

[ "BSD-3-Clause" ]

[ "BSD-3-Clause" ]

[ [ [ "## Gaussian Transformation with Scikit-learn\n\nScikit-learn has recently released transformers to do Gaussian mappings as they call the variable transformations. The PowerTransformer allows to do Box-Cox and Yeo-Johnson transformation. With the FunctionTransformer, we can specify any function we want.\n\nThe transformers per se, do not allow to select columns, but we can do so using a third transformer, the ColumnTransformer\n\nAnother thing to keep in mind is that Scikit-learn transformers return NumPy arrays, and not dataframes, so we need to be mindful of the order of the columns not to mess up with our features.\n\n## Important\n\nBox-Cox and Yeo-Johnson transformations need to learn their parameters from the data. Therefore, as always, before attempting any transformation it is important to divide the dataset into train and test set.\n\nIn this demo, I will not do so for simplicity, but when using this transformation in your pipelines, please make sure you do so.\n\n\n## In this demo\n\nWe will see how to implement variable transformations using Scikit-learn and the House Prices dataset.", "_____no_output_____" ] ], [ [ "import pandas as pd\nimport numpy as np\n\nimport matplotlib.pyplot as plt\n\nimport scipy.stats as stats\n\nfrom sklearn.preprocessing import FunctionTransformer, PowerTransformer", "_____no_output_____" ], [ "# load the data\n\ndata = pd.read_csv('../houseprice.csv')\n\ndata.head()", "_____no_output_____" ] ], [ [ "Let's select the numerical and positive variables in the dataset for this demonstration. As most of the transformations require the variables to be positive.", "_____no_output_____" ] ], [ [ "cols = []\n\nfor col in data.columns:\n\n if data[col].dtypes != 'O' and col != 'Id': # if the variable is numerical\n \n if np.sum(np.where(data[col] <= 0, 1, 0)) == 0: # if the variable is positive\n \n cols.append(col) # append variable to the list\n\ncols", "_____no_output_____" ], [ "# let's explore the distribution of the numerical variables\n\ndata[cols].hist(figsize=(20,20))\nplt.show()", "_____no_output_____" ] ], [ [ "## Plots to assess normality\n\nTo visualise the distribution of the variables, we plot a histogram and a Q-Q plot. In the Q-Q pLots, if the variable is normally distributed, the values of the variable should fall in a 45 degree line when plotted against the theoretical quantiles. We discussed this extensively in Section 3 of this course.", "_____no_output_____" ] ], [ [ "# plot the histograms to have a quick look at the variable distribution\n# histogram and Q-Q plots\n\ndef diagnostic_plots(df, variable):\n \n # function to plot a histogram and a Q-Q plot\n # side by side, for a certain variable\n \n plt.figure(figsize=(15,6))\n plt.subplot(1, 2, 1)\n df[variable].hist(bins=30)\n\n plt.subplot(1, 2, 2)\n stats.probplot(df[variable], dist=\"norm\", plot=plt)\n\n plt.show()", "_____no_output_____" ] ], [ [ "### Logarithmic transformation", "_____no_output_____" ] ], [ [ "# create a log transformer\n\ntransformer = FunctionTransformer(np.log, validate=True)", "_____no_output_____" ], [ "# transform all the numerical and positive variables\n\ndata_t = transformer.transform(data[cols].fillna(1))", "_____no_output_____" ], [ "# Scikit-learn returns NumPy arrays, so capture in dataframe\n# note that Scikit-learn will return an array with\n# only the columns indicated in cols\n\ndata_t = pd.DataFrame(data_t, columns = cols)", "_____no_output_____" ], [ "# original distribution\n\ndiagnostic_plots(data, 'GrLivArea')", "_____no_output_____" ], [ "# transformed distribution\n\ndiagnostic_plots(data_t, 'GrLivArea')", "_____no_output_____" ], [ "# original distribution\n\ndiagnostic_plots(data, 'MSSubClass')", "_____no_output_____" ], [ "# transformed distribution\n\ndiagnostic_plots(data_t, 'MSSubClass')", "_____no_output_____" ] ], [ [ "### Reciprocal transformation", "_____no_output_____" ] ], [ [ "# create the transformer\ntransformer = FunctionTransformer(lambda x: 1/x, validate=True)\n\n# also\n# transformer = FunctionTransformer(np.reciprocal, validate=True)\n\n# transform the positive variables\ndata_t = transformer.transform(data[cols].fillna(1))\n\n# re-capture in a dataframe\ndata_t = pd.DataFrame(data_t, columns = cols)", "_____no_output_____" ], [ "# transformed variable\n\ndiagnostic_plots(data_t, 'GrLivArea')", "_____no_output_____" ], [ "# transformed variable\n\ndiagnostic_plots(data_t, 'MSSubClass')", "_____no_output_____" ] ], [ [ "### Square root transformation", "_____no_output_____" ] ], [ [ "transformer = FunctionTransformer(lambda x: x**(1/2), validate=True)\n\n# also\n# transformer = FunctionTransformer(np.sqrt, validate=True)\n\ndata_t = transformer.transform(data[cols].fillna(1))\n\ndata_t = pd.DataFrame(data_t, columns = cols)", "_____no_output_____" ], [ "diagnostic_plots(data_t, 'GrLivArea')", "_____no_output_____" ], [ "diagnostic_plots(data_t, 'MSSubClass')", "_____no_output_____" ] ], [ [ "### Exponential", "_____no_output_____" ] ], [ [ "transformer = FunctionTransformer(lambda x: x**(1/1.2), validate=True)\n\ndata_t = transformer.transform(data[cols].fillna(1))\n\ndata_t = pd.DataFrame(data_t, columns = cols)", "_____no_output_____" ], [ "diagnostic_plots(data_t, 'GrLivArea')", "_____no_output_____" ], [ "diagnostic_plots(data_t, 'MSSubClass')", "_____no_output_____" ] ], [ [ "### Box-Cox transformation", "_____no_output_____" ] ], [ [ "# create the transformer\ntransformer = PowerTransformer(method='box-cox', standardize=False)\n\n# find the optimal lambda using the train set\ntransformer.fit(data[cols].fillna(1))\n\n# transform the data\ndata_t = transformer.transform(data[cols].fillna(1))\n\n# capture data in a dataframe\ndata_t = pd.DataFrame(data_t, columns = cols)", "C:\\Users\\Sole\\Documents\\Repositories\\envs\\fe_test\\lib\\site-packages\\numpy\\core\\_methods.py:195: RuntimeWarning: overflow encountered in multiply\n x = um.multiply(x, x, out=x)\nC:\\Users\\Sole\\Documents\\Repositories\\envs\\fe_test\\lib\\site-packages\\scipy\\stats\\morestats.py:910: RuntimeWarning: divide by zero encountered in log\n return (lmb - 1) * np.sum(logdata, axis=0) - N/2 * np.log(variance)\n" ], [ "diagnostic_plots(data_t, 'GrLivArea')", "_____no_output_____" ], [ "diagnostic_plots(data_t, 'MSSubClass')", "_____no_output_____" ] ], [ [ "### Yeo-Johnson\n\nYeo-Johnson is an adaptation of Box-Cox that can also be used in negative value variables. So let's expand the list of variables for the demo, to include those that contain zero and negative values as well. ", "_____no_output_____" ] ], [ [ "cols = [\n 'MSSubClass', 'LotFrontage', 'LotArea', 'OverallQual',\n 'OverallCond', 'MasVnrArea', 'BsmtFinSF1',\n 'BsmtFinSF2', 'BsmtUnfSF', 'TotalBsmtSF', '1stFlrSF', '2ndFlrSF',\n 'LowQualFinSF', 'GrLivArea', 'BsmtFullBath', 'BsmtHalfBath', 'FullBath',\n 'HalfBath', 'BedroomAbvGr', 'KitchenAbvGr', 'TotRmsAbvGrd',\n 'Fireplaces', 'GarageYrBlt', 'GarageCars', 'GarageArea', 'WoodDeckSF',\n 'OpenPorchSF', 'EnclosedPorch', '3SsnPorch', 'ScreenPorch', 'PoolArea',\n 'MiscVal', 'SalePrice'\n]", "_____no_output_____" ], [ "# call the transformer\ntransformer = PowerTransformer(method='yeo-johnson', standardize=False)\n\n# learn the lambda from the train set\ntransformer.fit(data[cols].fillna(1))\n\n# transform the data\ndata_t = transformer.transform(data[cols].fillna(1))\n\n# capture data in a dataframe\ndata_t = pd.DataFrame(data_t, columns = cols)", "_____no_output_____" ], [ "diagnostic_plots(data_t, 'GrLivArea')", "_____no_output_____" ], [ "diagnostic_plots(data_t, 'MSSubClass')", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "from IPython.core.display import display, HTML\ndisplay(HTML(\"<style>.container { width:95% !important; }</style>\"))\n\nfrom jupyterthemes import jtplot\njtplot.style()\n\nfrom plotly.offline import download_plotlyjs, init_notebook_mode, plot, iplot\ninit_notebook_mode(connected=True)\n\nimport os\nimport json\nimport numpy as np\nimport pandas as pd\nfrom tqdm import tqdm_notebook\nfrom datetime import datetime\nfrom matplotlib import pyplot as plt\n%matplotlib inline\n\ntarget_stocks = ['BANPU','IRPC','PTT','BBL','KBANK','SCB','AOT','THAI','CPF','MINT',\n 'TU','SCC','CPN','CK','CPALL','HMPRO','BDMS','BH','ADVANC','JAS','TRUE']", "_____no_output_____" ], [ "with open('../data/kaohoon.json') as json_data:\n data = json.load(json_data)\nlen(data)", "_____no_output_____" ], [ "data[20]", "_____no_output_____" ], [ "entry = []\nfor i, row in tqdm_notebook(enumerate(data)):\n if len(row['Stock Include']) > 1: continue\n for stock in row['Stock Include']:\n if stock in target_stocks:\n entry.append([\n row['Date'],\n stock,\n# row['Stock Include'],\n row['Content']\n ])", "_____no_output_____" ], [ "df = pd.DataFrame.from_records(entry)\ndf[0] = pd.to_datetime(df[0], format='%Y-%m-%d')\ndf.columns = ['Date','Ticker','Text']", "_____no_output_____" ], [ "df.to_csv('../data/kaohoon.csv', index=False)", "_____no_output_____" ], [ "df.head()", "_____no_output_____" ] ], [ [ "## Money Channel", "_____no_output_____" ] ], [ [ "with open('../data/moneychanel.json') as json_data:\n data = json.load(json_data)\nlen(data)", "_____no_output_____" ], [ "data[13]", "_____no_output_____" ], [ "entry = []\nfor i, row in tqdm_notebook(enumerate(data)):\n if len(row['Stock Include']) > 1: continue\n for stock in row['Stock Include']:\n if stock in target_stocks:\n entry.append([\n row['Date'],\n stock,\n row['Content']\n ])", "_____no_output_____" ], [ "df = pd.DataFrame.from_records(entry)\ndf[0] = pd.to_datetime(df[0], format='%Y-%m-%d')\ndf.columns = ['Date','Ticker','Text']\ndf.head()", "_____no_output_____" ], [ "df.to_csv('../data/moneychanel.csv', index=False)", "_____no_output_____" ] ], [ [ "## Pantip", "_____no_output_____" ] ], [ [ "with open('../data/pantip.json') as json_data:\n data = json.load(json_data)\nlen(data)", "_____no_output_____" ], [ "data[3]", "_____no_output_____" ], [ "data[3]['date']\ndata[3]['stock']\ntext = data[3]['head']+' '+data[3]['content']\ntext", "_____no_output_____" ], [ "for x in data[3]['comments']:\n text += x['message']\ntext", "_____no_output_____" ], [ "entry = []\nfor i, row in tqdm_notebook(enumerate(data)):\n if len(row['stock']) > 1: continue\n for stock in row['stock']:\n if stock in target_stocks:\n \n text = row['head']+' '+row['content']\n for comment in row['comments']:\n text += comment['message']\n \n entry.append([\n row['date'],\n stock,\n text\n ])", "_____no_output_____" ], [ "df = pd.DataFrame.from_records(entry)\ndf[0] = pd.to_datetime(df[0], format='%Y-%m-%d')\ndf.columns = ['Date','Ticker','Text']\ndf.head()", "_____no_output_____" ], [ "df.to_csv('../data/pantip.csv', index=False)", "_____no_output_____" ] ], [ [ "## Twitter", "_____no_output_____" ] ], [ [ "with open('../data/twitter.json') as json_data:\n data = json.load(json_data)\nlen(data)", "_____no_output_____" ], [ "data[0]", "_____no_output_____" ], [ "entry = []\nfor i, row in tqdm_notebook(enumerate(data)):\n if len(row['Stock Include']) > 1: continue\n for stock in row['Stock Include']:\n if stock in target_stocks:\n entry.append([\n row['date'],\n stock,\n row['text']\n ])", "_____no_output_____" ], [ "df = pd.DataFrame.from_records(entry)\ndf[0] = pd.to_datetime(df[0], format='%Y-%m-%d')\ndf.columns = ['Date','Ticker','Text']\ndf.head()", "_____no_output_____" ], [ "df.to_csv('../data/twitter.csv', index=False)", "_____no_output_____" ] ] ]

[ "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# In-Class Coding Lab: Iterations\n\nThe goals of this lab are to help you to understand:\n\n- How loops work.\n- The difference between definite and indefinite loops, and when to use each.\n- How to build an indefinite loop with complex exit conditions.\n- How to create a program from a complex idea.\n", "_____no_output_____" ], [ "# Understanding Iterations\n\nIterations permit us to repeat code until a Boolean expression is `False`. Iterations or **loops** allow us to write succint, compact code. Here's an example, which counts to 3 before [Blitzing the Quarterback in backyard American Football](https://www.quora.com/What-is-the-significance-of-counting-one-Mississippi-two-Mississippi-and-so-on):", "_____no_output_____" ] ], [ [ "i = 1\nwhile i <= 3:\n print(i,\"Mississippi...\")\n i=i+1\nprint(\"Blitz!\")", "1 Mississippi...\n2 Mississippi...\n3 Mississippi...\nBlitz!\n" ] ], [ [ "## Breaking it down...\n\nThe `while` statement on line 2 starts the loop. The code indented beneath it (lines 3-4) will repeat, in a linear fashion until the Boolean expression on line 2 `i <= 3` is `False`, at which time the program continues with line 5.\n\n### Some Terminology\n\nWe call `i <=3` the loop's **exit condition**. The variable `i` inside the exit condition is the only thing that we can change to make the exit condition `False`, therefore it is the **loop control variable**. On line 4 we change the loop control variable by adding one to it, this is called an **increment**.\n\nFurthermore, we know how many times this loop will execute before it actually runs: 3. Even if we allowed the user to enter a number, and looped that many times, we would still know. We call this a **definite loop**. Whenever we iterate over a fixed number of values, regardless of whether those values are determined at run-time or not, we're using a definite loop.\n\nIf the loop control variable never forces the exit condition to be `False`, we have an **infinite loop**. As the name implies, an Infinite loop never ends and typically causes our computer to crash or lock up. ", "_____no_output_____" ] ], [ [ "## WARNING!!! INFINITE LOOP AHEAD\n## IF YOU RUN THIS CODE YOU WILL NEED TO KILL YOUR BROWSER AND SHUT DOWN JUPYTER NOTEBOOK\n\ni = 1\nwhile i <= 3:\n print(i,\"Mississippi...\")\n# i=i+1\nprint(\"Blitz!\")", "_____no_output_____" ] ], [ [ "### For loops\n\nTo prevent an infinite loop when the loop is definite, we use the `for` statement. Here's the same program using `for`:", "_____no_output_____" ] ], [ [ "for i in range(1,4):\n print(i,\"Mississippi...\")\nprint(\"Blitz!\")", "1 Mississippi...\n2 Mississippi...\n3 Mississippi...\nBlitz!\n" ] ], [ [ "One confusing aspect of this loop is `range(1,4)` why does this loop from 1 to 3? Why not 1 to 4? Well it has to do with the fact that computers start counting at zero. The easier way to understand it is if you subtract the two numbers you get the number of times it will loop. So for example, 4-1 == 3.\n\n### Now Try It\n\nIn the space below, Re-Write the above program to count from 10 to 15. Note: How many times will that loop?", "_____no_output_____" ] ], [ [ "# TODO Write code here\n", "10 Mississippi...\n11 Mississippi...\n12 Mississippi...\n13 Mississippi...\n14 Mississippi...\n15 Mississippi...\nBlitz!\n" ] ], [ [ "## Indefinite loops\n\nWith **indefinite loops** we do not know how many times the program will execute. This is typically based on user action, and therefore our loop is subject to the whims of whoever interacts with it. Most applications like spreadsheets, photo editors, and games use indefinite loops. They'll run on your computer, seemingly forever, until you choose to quit the application. \n\nThe classic indefinite loop pattern involves getting input from the user inside the loop. We then inspect the input and based on that input we might exit the loop. Here's an example:", "_____no_output_____" ] ], [ [ "name = \"\"\nwhile name != 'mike':\n name = input(\"Say my name! : \")\n print(\"Nope, my name is not %s! \" %(name))", "Say my name! : rizzen\nNope, my name is not rizzen! \nSay my name! : mike\nNope, my name is not mike! \n" ] ], [ [ "The classic problem with indefinite loops is that its really difficult to get the application's logic to line up with the exit condition. For example we need to set `name = \"\"` in line 1 so that line 2 start out as `True`. Also we have this wonky logic where when we say `'mike'` it still prints `Nope, my name is not mike!` before exiting.\n\n### Break statement\n\nThe solution to this problem is to use the break statement. **break** tells Python to exit the loop immediately. We then re-structure all of our indefinite loops to look like this:\n\n```\nwhile True:\n if exit-condition:\n break\n```\n\nHere's our program we-written with the break statement. This is the recommended way to write indefinite loops in this course.", "_____no_output_____" ] ], [ [ "while True:\n name = input(\"Say my name!: \")\n if name == 'mike':\n break\n print(\"Nope, my name is not %s!\" %(name))", "Say my name!: bill\nNope, my name is not bill!\nSay my name!: dave\nNope, my name is not dave!\nSay my name!: mike\n" ] ], [ [ "### Multiple exit conditions\n\nThis indefinite loop pattern makes it easy to add additional exit conditions. For example, here's the program again, but it now stops when you say my name or type in 3 wrong names. Make sure to run this program a couple of times. First enter mike to exit the program, next enter the wrong name 3 times.", "_____no_output_____" ] ], [ [ "times = 0\nwhile True:\n name = input(\"Say my name!: \")\n times = times + 1\n if name == 'mike':\n print(\"You got it!\")\n break\n if times == 3:\n print(\"Game over. Too many tries!\")\n break\n print(\"Nope, my name is not %s!\" %(name))", "Say my name!: mike\nYou got it!\n" ] ], [ [ "# Number sums\n\nLet's conclude the lab with you writing your own program which\n\nuses an indefinite loop. We'll provide the to-do list, you write the code. This program should ask for floating point numbers as input and stops looping when **the total of the numbers entered is over 100**, or **more than 5 numbers have been entered**. Those are your two exit conditions. After the loop stops print out the total of the numbers entered and the count of numbers entered. ", "_____no_output_____" ] ], [ [ "## TO-DO List\n\n#1 count = 0\n#2 total = 0\n#3 loop Indefinitely\n#4. input a number\n#5 increment count\n#6 add number to total\n#7 if count equals 5 stop looping\n#8 if total greater than 100 stop looping\n#9 print total and count", "_____no_output_____" ], [ "# Write Code here:\ncount = 0 \ntotal = 0 \nwhile true:\n number = int(input(\"enter a number:\"))\n count = count + 1\n total = total + number \n if count == 5:\n break\n if total >100 \n break\n print(\"total:\", total,\"count:\", count)", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "# default_exp data_processing.numeric", "_____no_output_____" ] ], [ [ "# data_processing.numeric\n\n> Numeric related data processing\n\n- toc: True", "_____no_output_____" ] ], [ [ "#export\nimport pandas as pd", "_____no_output_____" ], [ "# export\ndef moving_average(data_frame: pd.DataFrame = None,\n window: int=7,\n group_col: str = None,\n value_col: str = None,\n shift=0)->pd.DataFrame:\n df = data_frame.copy()\n ma_col = '{value_col}_ma{window}'.format(value_col=value_col, window=window)\n if group_col is None:\n df[ma_col] = df[value_col].rolling(window=window).mean().shift(periods=shift)\n else:\n df[ma_col] = df.groupby(group_col)[value_col].apply(lambda x: x.rolling(window=window).mean().shift(periods=shift))\n return df[ma_col]", "_____no_output_____" ] ] ]

[ "code", "markdown", "code" ]

[ [ "code" ], [ "markdown" ], [ "code", "code" ] ]