pull all the requestes

.dvc/config

@@ -1,4 +0,0 @@
-[core]
-    remote = remote
-['remote "remote"']
-    url = ../TEMP

data/raw.dvc

@@ -1,6 +1,6 @@
 outs:
-- md5: 63310a287c47722acf1521efd0e0d061.dir
-  size: 150825054
-  nfiles: 2
+- md5: ff6a3a61f72477d43b577210859f24e0.dir
+  size: 102634230
+  nfiles: 1
   hash: md5
   path: raw

dvc.lock

@@ -5,8 +5,8 @@ stages:
     deps:
     - path: .\data\raw\creditcard.csv
       hash: md5
-      md5: e8ffcd27762aaf279e268eb849a060f9
-      size: 150825054
+      md5: 61b475b2696ec29b4f9acac60c29ecae
+      size: 102634230
     - path: .\src\data\make_dataset.py
       hash: md5
       md5: 5d87ea0853101b748b17af0be510c2ac
@@ -18,16 +18,16 @@ stages:
     outs:
     - path: .\data\processed\
       hash: md5
-      md5: b21101ae3f5090f5af39e6f9cbdb21c3.dir
-      size: 151098829
+      md5: 299f94739fdbe9c803f4d5162a48387d.dir
+      size: 103630255
       nfiles: 2
   train_model:
     cmd: python .\src\models\train_model.py .\data\processed\
     deps:
     - path: .\data\processed\
       hash: md5
-      md5: b21101ae3f5090f5af39e6f9cbdb21c3.dir
-      size: 151098829
+      md5: 299f94739fdbe9c803f4d5162a48387d.dir
+      size: 103630255
       nfiles: 2
     - path: .\src\models\train_model.py
       hash: md5
@@ -36,28 +36,28 @@ stages:
     outs:
     - path: .\models\
       hash: md5
-      md5: 6191af90def2b514e1128c7d91ae5877.dir
-      size: 397817
+      md5: 5ec2d5bc2c9c3c5f4a13a3feadb06198.dir
+      size: 399609
       nfiles: 1
   visualize:
     cmd: python .\src\visualization\visualize.py .\models\model.joblib .\data\processed\
     deps:
     - path: .\data\processed\
       hash: md5
-      md5: 45207c27b5aec2846e4532f5bee1631f.dir
-      size: 151098829
+      md5: 299f94739fdbe9c803f4d5162a48387d.dir
+      size: 103630255
       nfiles: 2
     - path: .\models\model.joblib
       hash: md5
-      md5: e01830bd03da0f4f328c34580d40d5b3
-      size: 3865
+      md5: e543edd128da428d8eab5bbe5377072f
+      size: 399609
     - path: .\src\visualization\visualize.py
       hash: md5
-      md5: db86d2e43c664f095353e9ba6afe3393
-      size: 3687
+      md5: a91ca79bfaea995c333fa50841f86fde
+      size: 3683
     outs:
     - path: dvclive
       hash: md5
-      md5: e887b02afa678d483c083f4d3632dc94.dir
-      size: 18530385
+      md5: 5141dcc71cf1097da4b1b730fd527f1a.dir
+      size: 20018970
       nfiles: 8

params.yaml

@@ -1,7 +1,7 @@
 make_dataset:
-  test_split: 0.2
+  test_split: 0.15
   seed: 2023
 train_model:
-  seed: 21
-  n_estimators: 50
-  max_depth: 8
+  seed: 42
+  n_estimators: 10
+  max_depth: 10

requirements.txt

@@ -3,3 +3,4 @@ fastapi
 joblib
 uvicorn
 scikit-learn # This is needed by joblib to read and load the model
+dvclive

src/models/train_model.py

@@ -3,7 +3,7 @@
 import sys
 import yaml
 import joblib
-
+from dvclive import live
 import pandas as pd
 from sklearn.ensemble import RandomForestClassifier