Merge branch '15-integration-sota' into 'master'

Resolve "integration-sota"

Closes #15

See merge request !23

scripts/run_stage5_experiments.sh

 #!/bin/bash
 core_number=5
-core_number_sota=50
-walltime=1:00
+core_number_sota=5
+walltime=5:00
 walltime_sota=5:00
 seeds='1 2 3 4 5'
 
-for dataset in kin8nm kr-vs-kp spambase steel-plates diabetes diamonds boston california_housing
+for dataset in boston diabetes linnerud breast_cancer california_housing diamonds steel-plates kr-vs-kp kin8nm spambase musk gamma lfw_pairs
 do
-    oarsub -p "(gpu is null)" -l /core=$core_number,walltime=$walltime "conda activate test_env && python code/train.py --dataset_name=$dataset --seeds $seeds --extraction_strategy=none --extracted_forest_size_stop=1.0 --extracted_forest_size_samples=30 --experiment_id=1 --models_dir=models/$dataset/stage5 --subsets_used train+dev,train+dev"
-    oarsub -p "(gpu is null)" -l /core=$core_number,walltime=$walltime "conda activate test_env && python code/train.py --dataset_name=$dataset --seeds $seeds --extraction_strategy=random --extracted_forest_size_stop=1.0 --extracted_forest_size_samples=30 --experiment_id=2 --models_dir=models/$dataset/stage5 --subsets_used train+dev,train+dev"
-    oarsub -p "(gpu is null)" -l /core=$core_number,walltime=$walltime "conda activate test_env && python code/train.py --dataset_name=$dataset --seeds $seeds --extraction_strategy=omp --extracted_forest_size_stop=1.0 --extracted_forest_size_samples=30 --experiment_id=3 --models_dir=models/$dataset/stage5 --subsets_used train+dev,train+dev"
-    oarsub -p "(gpu is null)" -l /core=$core_number_sota,walltime=$walltime_sota "conda activate test_env && python code/train.py --dataset_name=$dataset --seeds $seeds --extraction_strategy=similarity --extracted_forest_size_stop=1.0 --extracted_forest_size_samples=30 --experiment_id=4 --models_dir=models/$dataset/stage5 --subsets_used train+dev,train+dev"
-    oarsub -p "(gpu is null)" -l /core=$core_number_sota,walltime=$walltime_sota "conda activate test_env && python code/train.py --dataset_name=$dataset --seeds $seeds --extraction_strategy=kmeans --extracted_forest_size_stop=1.0 --extracted_forest_size_samples=30 --experiment_id=5 --models_dir=models/$dataset/stage5 --subsets_used train+dev,train+dev"
-    oarsub -p "(gpu is null)" -l /core=$core_number_sota,walltime=$walltime_sota "conda activate test_env && python code/train.py --dataset_name=$dataset --seeds $seeds --extraction_strategy=ensemble --extracted_forest_size_stop=1.0 --extracted_forest_size_samples=30 --experiment_id=6 --models_dir=models/$dataset/stage5 --subsets_used train+dev,train+dev"
+    oarsub -p "(gpu is null)" -l /core=$core_number,walltime=$walltime "conda activate test_env && python code/train.py --dataset_name=$dataset --seeds $seeds --extraction_strategy=none --extracted_forest_size_stop=1.0 --extracted_forest_size_samples=30 --experiment_id=1 --models_dir=models/$dataset/stage5_new --subsets_used train+dev,train+dev"
+    oarsub -p "(gpu is null)" -l /core=$core_number,walltime=$walltime "conda activate test_env && python code/train.py --dataset_name=$dataset --seeds $seeds --extraction_strategy=random --extracted_forest_size_stop=1.0 --extracted_forest_size_samples=30 --experiment_id=2 --models_dir=models/$dataset/stage5_new --subsets_used train+dev,train+dev"
+    oarsub -p "(gpu is null)" -l /core=$core_number,walltime=$walltime "conda activate test_env && python code/train.py --dataset_name=$dataset --seeds $seeds --extraction_strategy=omp --extracted_forest_size_stop=1.0 --extracted_forest_size_samples=30 --experiment_id=3 --models_dir=models/$dataset/stage5_new --subsets_used train+dev,train+dev --normalize_D"
+    oarsub -p "(gpu is null)" -l /core=$core_number,walltime=$walltime "conda activate test_env && python code/train.py --dataset_name=$dataset --seeds $seeds --extraction_strategy=omp_distillation --extracted_forest_size_stop=1.0 --extracted_forest_size_samples=30 --experiment_id=4 --models_dir=models/$dataset/stage5_new --subsets_used train+dev,train+dev --normalize_D"
+    oarsub -p "(gpu is null)" -l /core=$core_number_sota,walltime=$walltime_sota "conda activate test_env && python code/train.py --dataset_name=$dataset --seeds $seeds --extraction_strategy=kmeans --extracted_forest_size_stop=1.0 --extracted_forest_size_samples=30 --experiment_id=5 --models_dir=models/$dataset/stage5_new --subsets_used train+dev,train+dev"
+    oarsub -p "(gpu is null)" -l /core=$core_number_sota,walltime=$walltime_sota "conda activate test_env && python code/train.py --dataset_name=$dataset --seeds $seeds --extraction_strategy=similarity_similarities --extracted_forest_size_stop=1.0 --extracted_forest_size_samples=30 --experiment_id=6 --models_dir=models/$dataset/stage5_new --subsets_used train+dev,train+dev"
+    oarsub -p "(gpu is null)" -l /core=$core_number_sota,walltime=$walltime_sota "conda activate test_env && python code/train.py --dataset_name=$dataset --seeds $seeds --extraction_strategy=similarity_predictions --extracted_forest_size_stop=1.0 --extracted_forest_size_samples=30 --experiment_id=7 --models_dir=models/$dataset/stage5_new --subsets_used train+dev,train+dev"
+    oarsub -p "(gpu is null)" -l /core=$core_number_sota,walltime=$walltime_sota "conda activate test_env && python code/train.py --dataset_name=$dataset --seeds $seeds --extraction_strategy=ensemble --extracted_forest_size_stop=1.0 --extracted_forest_size_samples=30 --experiment_id=8 --models_dir=models/$dataset/stage5_new --subsets_used train+dev,train+dev"
+
+    oarsub -p "(gpu is null)" -l /core=$core_number,walltime=$walltime "conda activate test_env && python code/train.py --dataset_name=$dataset --seeds $seeds --extraction_strategy=none --extracted_forest_size_stop=1.0 --extracted_forest_size_samples=30 --experiment_id=9 --models_dir=models/$dataset/stage5_new --subsets_used train,dev"
+    oarsub -p "(gpu is null)" -l /core=$core_number,walltime=$walltime "conda activate test_env && python code/train.py --dataset_name=$dataset --seeds $seeds --extraction_strategy=random --extracted_forest_size_stop=1.0 --extracted_forest_size_samples=30 --experiment_id=10 --models_dir=models/$dataset/stage5_new --subsets_used train,dev"
+    oarsub -p "(gpu is null)" -l /core=$core_number,walltime=$walltime "conda activate test_env && python code/train.py --dataset_name=$dataset --seeds $seeds --extraction_strategy=omp --extracted_forest_size_stop=1.0 --extracted_forest_size_samples=30 --experiment_id=11 --models_dir=models/$dataset/stage5_new --subsets_used train,dev --normalize_D"
+    oarsub -p "(gpu is null)" -l /core=$core_number,walltime=$walltime "conda activate test_env && python code/train.py --dataset_name=$dataset --seeds $seeds --extraction_strategy=omp_distillation --extracted_forest_size_stop=1.0 --extracted_forest_size_samples=30 --experiment_id=12 --models_dir=models/$dataset/stage5_new --subsets_used train,dev --normalize_D"
+    oarsub -p "(gpu is null)" -l /core=$core_number_sota,walltime=$walltime_sota "conda activate test_env && python code/train.py --dataset_name=$dataset --seeds $seeds --extraction_strategy=kmeans --extracted_forest_size_stop=1.0 --extracted_forest_size_samples=30 --experiment_id=13 --models_dir=models/$dataset/stage5_new --subsets_used train,dev"
+    oarsub -p "(gpu is null)" -l /core=$core_number_sota,walltime=$walltime_sota "conda activate test_env && python code/train.py --dataset_name=$dataset --seeds $seeds --extraction_strategy=similarity_similarities --extracted_forest_size_stop=1.0 --extracted_forest_size_samples=30 --experiment_id=14 --models_dir=models/$dataset/stage5_new --subsets_used train,dev"
+    oarsub -p "(gpu is null)" -l /core=$core_number_sota,walltime=$walltime_sota "conda activate test_env && python code/train.py --dataset_name=$dataset --seeds $seeds --extraction_strategy=similarity_predictions --extracted_forest_size_stop=1.0 --extracted_forest_size_samples=30 --experiment_id=15 --models_dir=models/$dataset/stage5_new --subsets_used train,dev"
+    oarsub -p "(gpu is null)" -l /core=$core_number_sota,walltime=$walltime_sota "conda activate test_env && python code/train.py --dataset_name=$dataset --seeds $seeds --extraction_strategy=ensemble --extracted_forest_size_stop=1.0 --extracted_forest_size_samples=30 --experiment_id=16 --models_dir=models/$dataset/stage5_new --subsets_used train,dev"
 done