medium model with scenario and upgraded autogluon

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

@@ -85,21 +85,12 @@ stages:
         default.prepare_data.output_dataclient_type: local
         default.prepare_data.output_test_filepath: ./data/prepared_data/test.parquet
         default.prepare_data.output_train_filepath: ./data/prepared_data/train.parquet
-<<<<<<< HEAD
         default.prepare_data.train_proportion: 0.9
     outs:
     - path: data/prepared_data/
       hash: md5
       md5: efa416abea618ae6220a0c3d597603cf.dir
       size: 44750997
-=======
-        default.prepare_data.train_proportion: 0.98
-    outs:
-    - path: data/prepared_data/
-      hash: md5
-      md5: 544427230544c2cc526334e246db4845.dir
-      size: 26132493
->>>>>>> d5f40a8eb294924e0525904d6ee1864999d77c23
       nfiles: 2
   build_model:
     cmd: python 2_build_model.py
@@ -110,13 +101,8 @@ stages:
       size: 4820
     - path: data/prepared_data
       hash: md5
-<<<<<<< HEAD
       md5: efa416abea618ae6220a0c3d597603cf.dir
       size: 44750997
-=======
-      md5: 544427230544c2cc526334e246db4845.dir
-      size: 26132493
->>>>>>> d5f40a8eb294924e0525904d6ee1864999d77c23
       nfiles: 2
     params:
       configs/build_model.yaml:
@@ -148,7 +134,6 @@ stages:
     outs:
     - path: data/fit_predictions/
       hash: md5
-<<<<<<< HEAD
       md5: de46250d454c4d713ab580b10ff3fd31.dir
       size: 3349318
       nfiles: 1
@@ -161,20 +146,6 @@ stages:
       hash: md5
       md5: 8a952a5e884c268e6059357a627b9251
       size: 224
-=======
-      md5: 8f9e2059782dd55d3ecdad54b4551f6a.dir
-      size: 3630849
-      nfiles: 1
-    - path: data/model/
-      hash: md5
-      md5: e031eb3c3fdb63917aabfea745b56ac6.dir
-      size: 618445494
-      nfiles: 31
-    - path: metrics/fit_metrics.json
-      hash: md5
-      md5: e68009f5b66230b3ee4cd2ffc9a2d697
-      size: 222
->>>>>>> d5f40a8eb294924e0525904d6ee1864999d77c23
   generate_predictions:
     cmd: python 3_generate_predictions.py
     deps:
@@ -184,7 +155,6 @@ stages:
       size: 2464
     - path: data/model
       hash: md5
-<<<<<<< HEAD
       md5: 18bd7a93ece75a65d3a950b7dfdab4fb.dir
       size: 735951861
       nfiles: 35
@@ -192,15 +162,6 @@ stages:
       hash: md5
       md5: efa416abea618ae6220a0c3d597603cf.dir
       size: 44750997
-=======
-      md5: e031eb3c3fdb63917aabfea745b56ac6.dir
-      size: 618445494
-      nfiles: 31
-    - path: data/prepared_data
-      hash: md5
-      md5: 544427230544c2cc526334e246db4845.dir
-      size: 26132493
->>>>>>> d5f40a8eb294924e0525904d6ee1864999d77c23
       nfiles: 2
     params:
       configs/settings.yaml:
@@ -212,13 +173,8 @@ stages:
     outs:
     - path: data/predictions/
       hash: md5
-<<<<<<< HEAD
       md5: 07ef721a0dc94a52e3ba7a70ac45b8ff.dir
       size: 463563
-=======
-      md5: 1c14c9ac9711f5d33a60890e3ca72454.dir
-      size: 90361
->>>>>>> d5f40a8eb294924e0525904d6ee1864999d77c23
       nfiles: 1
   generate_metrics:
     cmd: python 4_generate_metrics.py
@@ -229,7 +185,6 @@ stages:
       size: 3484
     - path: data/predictions
       hash: md5
-<<<<<<< HEAD
       md5: 07ef721a0dc94a52e3ba7a70ac45b8ff.dir
       size: 463563
       nfiles: 1
@@ -237,15 +192,6 @@ stages:
       hash: md5
       md5: efa416abea618ae6220a0c3d597603cf.dir
       size: 44750997
-=======
-      md5: 1c14c9ac9711f5d33a60890e3ca72454.dir
-      size: 90361
-      nfiles: 1
-    - path: data/prepared_data
-      hash: md5
-      md5: 544427230544c2cc526334e246db4845.dir
-      size: 26132493
->>>>>>> d5f40a8eb294924e0525904d6ee1864999d77c23
       nfiles: 2
     params:
       configs/settings.yaml:
@@ -255,55 +201,25 @@ stages:
     outs:
     - path: metrics/metrics.json
       hash: md5
-<<<<<<< HEAD
       md5: 9f863f47799d42c101eba3b03a179455
       size: 224
   generate_scenerio_metrics:
     cmd: python 5_generate_scenarios.py
-=======
-      md5: 98e59ea9569522a8665c4e6c1bea7473
-      size: 222
-  startup_cleanup:
-    cmd: python 0_startup_cleanup.py
->>>>>>> d5f40a8eb294924e0525904d6ee1864999d77c23
     deps:
     - path: 5_generate_scenarios.py
       hash: md5
       md5: 30f80ffeb6ee50c5f7b82943a4dc7702
       size: 4014
     params:
-<<<<<<< HEAD
-=======
-      configs/settings.yaml:
-        default.startup_cleanup.artefacts: ./data
-        default.startup_cleanup.metrics: ./metrics
-  generate_scenerio_metrics:
-    cmd: python 5_generate_scenarios.py
-    deps:
-    - path: 5_generate_scenarios.py
-      hash: md5
-      md5: 30f80ffeb6ee50c5f7b82943a4dc7702
-      size: 4014
-    params:
->>>>>>> d5f40a8eb294924e0525904d6ee1864999d77c23
       configs/scenarios.yaml:
         default.scenarios:
           input_dataclient_type: aws-s3
           output_dataclient_type: local
           scenario_data_filepaths:
-<<<<<<< HEAD
           - s3://retrofit-data-dev/scenario_data/24-03-2024-20-23-25/recommendations_scoring_data.parquet
-=======
-          - s3://retrofit-data-dev/scenario_data/recommendations_scoring_data.parquet
->>>>>>> d5f40a8eb294924e0525904d6ee1864999d77c23
           output_filepath: ./metrics/scenario_table.md
     outs:
     - path: metrics/scenario_table.md
       hash: md5
-<<<<<<< HEAD
       md5: 54856c66fca8b2ebd1fa4dea2d25734a
       size: 2133
-=======
-      md5: 3ee1966a06c1e5b9c37797597be94797
-      size: 1648
->>>>>>> d5f40a8eb294924e0525904d6ee1864999d77c23