From 6220cd17d3e8f8b961d694c1488a1f9aa9706262 Mon Sep 17 00:00:00 2001
From: Michael Duong <michaelduong22@gmail.com>
Date: Mon, 3 Nov 2025 14:43:26 +0000
Subject: [PATCH] faster inference

---
 .../src/pipeline/configs/build_model.yaml     |  8 +--
 modules/ml-pipeline/src/pipeline/dvc.lock     | 57 +++++++++----------
 2 files changed, 32 insertions(+), 33 deletions(-)

diff --git a/modules/ml-pipeline/src/pipeline/configs/build_model.yaml b/modules/ml-pipeline/src/pipeline/configs/build_model.yaml
index 50122ee..a03f430 100644
--- a/modules/ml-pipeline/src/pipeline/configs/build_model.yaml
+++ b/modules/ml-pipeline/src/pipeline/configs/build_model.yaml
@@ -17,15 +17,15 @@ default:
       time_limit: 1800
       presets: medium_quality
       excluded_model_types: ['RF', 'CAT', 'NN_TORCH', 'KNN', 'XT']
-      infer_limit: 0.001
+      infer_limit: 0.0005
       infer_limit_batch_size: 10000
-      fit_strategy: "sequential"
+      fit_strategy: "parallel"
       ag_args_ensemble: {'num_folds_parallel': 2}
-      num_gpus: auto
+      num_gpus: 0
       hyperparameters:
         {
         'NN_TORCH': [{}],
-        'GBM': [{'extra_trees': True, 'ag_args': {'name_suffix': 'XT'}}, {}, {'learning_rate': 0.03, 'num_leaves': 128, 'feature_fraction': 0.9, 'min_data_in_leaf': 3, 'ag_args': {'name_suffix': 'Large', 'priority': 0, 'hyperparameter_tune_kwargs': 'auto'}}],
+        'GBM': [{'extra_trees': True, 'ag_args': {'name_suffix': 'XT'}}, {}, {'learning_rate': 0.03, 'num_leaves': 128, 'feature_fraction': 0.9, 'min_data_in_leaf': 3, 'ag_args': {'name_suffix': 'Large', 'priority': 0,}}],
         # 'GBM': [{}],
         'CAT': [{}],
         'XGB': [{}],
diff --git a/modules/ml-pipeline/src/pipeline/dvc.lock b/modules/ml-pipeline/src/pipeline/dvc.lock
index cffd1b3..673285c 100644
--- a/modules/ml-pipeline/src/pipeline/dvc.lock
+++ b/modules/ml-pipeline/src/pipeline/dvc.lock
@@ -61,8 +61,8 @@ stages:
     outs:
     - path: data/prepared_data/
       hash: md5
-      md5: ba409a8c79863ddc407786b7aa7a053a.dir
-      size: 46113237
+      md5: bdf32829fa7d4084293e54f73084a25c.dir
+      size: 46126494
       nfiles: 3
   build_model:
     cmd: python 2_build_model.py
@@ -73,8 +73,8 @@ stages:
       size: 4820
     - path: data/prepared_data
       hash: md5
-      md5: ba409a8c79863ddc407786b7aa7a053a.dir
-      size: 46113237
+      md5: bdf32829fa7d4084293e54f73084a25c.dir
+      size: 46126494
       nfiles: 3
     params:
       configs/build_model.yaml:
@@ -99,12 +99,12 @@ stages:
               - NN_TORCH
               - KNN
               - XT
-              infer_limit: 0.001
+              infer_limit: 0.0005
               infer_limit_batch_size: 10000
-              fit_strategy: sequential
+              fit_strategy: parallel
               ag_args_ensemble:
                 num_folds_parallel: 2
-              num_gpus: auto
+              num_gpus: 0
               hyperparameters:
                 NN_TORCH:
                 - {}
@@ -120,7 +120,6 @@ stages:
                   ag_args:
                     name_suffix: Large
                     priority: 0
-                    hyperparameter_tune_kwargs: auto
                 CAT:
                 - {}
                 XGB:
@@ -175,18 +174,18 @@ stages:
     outs:
     - path: data/fit_predictions/
       hash: md5
-      md5: a9361ab31ff8fc08c3e5e3b96cec06d4.dir
-      size: 3474690
+      md5: bb3a39098652aee04227b8157cffe6dc.dir
+      size: 3475209
       nfiles: 1
     - path: data/model/
       hash: md5
-      md5: 19019e558886b1acd6d29442a47243d0.dir
-      size: 761937021
-      nfiles: 34
+      md5: 057ecca4ca9b6cabc60d81b36f5efee1.dir
+      size: 719633310
+      nfiles: 33
     - path: metrics/fit_metrics.json
       hash: md5
-      md5: 3af168aedf1f81a22024bb8c815f5d12
-      size: 221
+      md5: 5b63b6c51867279dea639df1c50b4416
+      size: 225
   generate_predictions:
     cmd: python 3_generate_predictions.py
     deps:
@@ -196,13 +195,13 @@ stages:
       size: 2464
     - path: data/model
       hash: md5
-      md5: 19019e558886b1acd6d29442a47243d0.dir
-      size: 761937021
-      nfiles: 34
+      md5: 057ecca4ca9b6cabc60d81b36f5efee1.dir
+      size: 719633310
+      nfiles: 33
     - path: data/prepared_data
       hash: md5
-      md5: ba409a8c79863ddc407786b7aa7a053a.dir
-      size: 46113237
+      md5: bdf32829fa7d4084293e54f73084a25c.dir
+      size: 46126494
       nfiles: 3
     params:
       configs/settings.yaml:
@@ -216,8 +215,8 @@ stages:
     outs:
     - path: data/predictions/
       hash: md5
-      md5: a9f32d70a4817df8092e52c5513a445f.dir
-      size: 484694
+      md5: 76dd44be410b11c62f701c645a9003a9.dir
+      size: 484698
       nfiles: 1
   generate_metrics:
     cmd: python 4_generate_metrics.py
@@ -228,13 +227,13 @@ stages:
       size: 3484
     - path: data/predictions
       hash: md5
-      md5: a9f32d70a4817df8092e52c5513a445f.dir
-      size: 484694
+      md5: 76dd44be410b11c62f701c645a9003a9.dir
+      size: 484698
       nfiles: 1
     - path: data/prepared_data
       hash: md5
-      md5: ba409a8c79863ddc407786b7aa7a053a.dir
-      size: 46113237
+      md5: bdf32829fa7d4084293e54f73084a25c.dir
+      size: 46126494
       nfiles: 3
     params:
       configs/settings.yaml:
@@ -244,7 +243,7 @@ stages:
     outs:
     - path: metrics/metrics.json
       hash: md5
-      md5: 736ef69da7edb94577139ae9ede5ac0d
+      md5: de4098ee21d5efbf8b769471ae2fa20f
       size: 224
   generate_scenerio_metrics:
     cmd: python 5_generate_scenarios.py
@@ -265,9 +264,9 @@ stages:
     outs:
     - path: metrics/scenario_metrics.md
       hash: md5
-      md5: adcc78833e7a0824ecb10ad78a646ea8
+      md5: d4f3c1eeee96720270e2467fc50b0b29
       size: 356
     - path: metrics/scenario_table.md
       hash: md5
-      md5: 35e704d0499e943c4110f66f1482d2ec
+      md5: 56e39ca1f444b0a97ce894cf9ff90d32
       size: 872