better model

modules/ml-pipeline/src/pipeline/src/configs/prepare_data.yaml

@@ -1,6 +1,6 @@
 dataclient_type: minio
 data_location: s3://dev_bucket
-train_proportion: 0.7
+train_proportion: 0.8
 output_train_filepath: ./data/prepared_data/train.parquet
 output_test_filepath: ./data/prepared_data/test.parquet
 

modules/ml-pipeline/src/pipeline/src/dvc.lock

@@ -11,12 +11,12 @@ stages:
       configs/prepare_data.yaml:
         output_test_filepath: ./data/prepared_data/test.parquet
         output_train_filepath: ./data/prepared_data/train.parquet
-        train_proportion: 0.7
+        train_proportion: 0.8
     outs:
     - path: data/prepared_data/
       hash: md5
-      md5: fdb2377affab10953f85fdf280ee897b.dir
-      size: 36663
+      md5: f0d462fe6b1a856a827409a745539285.dir
+      size: 36169
       nfiles: 2
   build_model:
     cmd: python build_model.py
@@ -27,8 +27,8 @@ stages:
       size: 3576
     - path: data/prepared_data
       hash: md5
-      md5: fdb2377affab10953f85fdf280ee897b.dir
-      size: 36663
+      md5: f0d462fe6b1a856a827409a745539285.dir
+      size: 36169
       nfiles: 2
     params:
       configs/build_model.yaml:
@@ -41,7 +41,7 @@ stages:
     outs:
     - path: data/model/
       hash: md5
-      md5: 7afb942aee47aa6da23f7dec7ce666a8.dir
+      md5: fb7ae4137b445dc91e840b794d72e940.dir
       size: 1096
       nfiles: 1
   generate_predictions:
@@ -49,18 +49,18 @@ stages:
     deps:
     - path: data/model
       hash: md5
-      md5: 7afb942aee47aa6da23f7dec7ce666a8.dir
+      md5: fb7ae4137b445dc91e840b794d72e940.dir
       size: 1096
       nfiles: 1
     - path: data/prepared_data
       hash: md5
-      md5: fdb2377affab10953f85fdf280ee897b.dir
-      size: 36663
+      md5: f0d462fe6b1a856a827409a745539285.dir
+      size: 36169
       nfiles: 2
     - path: generate_predictions.py
       hash: md5
-      md5: 219c7b2aa920b14d4bb7a1ef7df0ea1b
-      size: 3420
+      md5: 424b9d89045eaf8a5a167ab2e0e363ae
+      size: 3400
     params:
       configs/generate_predictions.yaml:
         predictions_column_name: predictions
@@ -69,26 +69,26 @@ stages:
     outs:
     - path: data/predictions/
       hash: md5
-      md5: 25bb58c06ce3bc7ef20de87298db1567.dir
-      size: 2949
+      md5: 4d5854903b25bdae15d99c934ebcfb99.dir
+      size: 2531
       nfiles: 1
   generate_metrics:
     cmd: python generate_metrics.py
     deps:
     - path: data/predictions
       hash: md5
-      md5: 25bb58c06ce3bc7ef20de87298db1567.dir
-      size: 2949
+      md5: 4d5854903b25bdae15d99c934ebcfb99.dir
+      size: 2531
       nfiles: 1
     - path: data/prepared_data
       hash: md5
-      md5: fdb2377affab10953f85fdf280ee897b.dir
-      size: 36663
+      md5: f0d462fe6b1a856a827409a745539285.dir
+      size: 36169
       nfiles: 2
     - path: generate_metrics.py
       hash: md5
-      md5: 81f8eec20ffb542f27bde28dc028bace
-      size: 3741
+      md5: b456e207b152298428ba79c083d1b6ff
+      size: 3728
     params:
       configs/generate_metrics.yaml:
         metrics_output_filepath: ./metrics/metrics.json
@@ -96,5 +96,5 @@ stages:
     outs:
     - path: metrics/metrics.json
       hash: md5
-      md5: 6528bdbe67f968d8eb240533efab48d5
+      md5: 3c9306e992b07491ff7e642949d6bc47
       size: 182