moved .sh files

src/experiment_scripts/10run_jrc_combinations.sh

@@ -5,12 +5,12 @@ runs='6 7 8 9' #0 1 2 3 4 5
 for run in $runs
 do
   dataset=$dataset_path$run.pickle
-  #python main_multimodal_cls.py $dataset -o $logfile -P -U -S -c -r -z --l2 --allprob # last combination for CIKM 3 Pr(views) concatenated  (done up to run5)
+  #python main_gFun.py $dataset -o $logfile -P -U -S -c -r -z --l2 --allprob # last combination for CIKM 3 Pr(views) concatenated  (done up to run5)
   python main_multimodal_cls.py $dataset -o $logfile -P -U -S -c -r -z --l2 --allprob # last combination for CIKM 3 views concatenated
-  #python main_multimodal_cls.py $dataset -o $logfile -P -U -S -c -r -a -z --l2 --allprob
-  #python main_multimodal_cls.py $dataset -o $logfile -P -U -c -r -a -z --l2 --allprob
-  #python main_multimodal_cls.py $dataset -o $logfile -P -S -c -r -z --l2 --allprob
-  #python main_multimodal_cls.py $dataset -o $logfile -P -U -c -r -z --l2 --allprob
-  #python main_multimodal_cls.py $dataset -o $logfile -c -P -U -r -z --l2
-  #python main_multimodal_cls.py $dataset -o $logfile -c -P -U -S -r -z --l2
+  #python main_gFun.py $dataset -o $logfile -P -U -S -c -r -a -z --l2 --allprob
+  #python main_gFun.py $dataset -o $logfile -P -U -c -r -a -z --l2 --allprob
+  #python main_gFun.py $dataset -o $logfile -P -S -c -r -z --l2 --allprob
+  #python main_gFun.py $dataset -o $logfile -P -U -c -r -z --l2 --allprob
+  #python main_gFun.py $dataset -o $logfile -c -P -U -r -z --l2
+  #python main_gFun.py $dataset -o $logfile -c -P -U -S -r -z --l2
 done

src/experiment_scripts/10run_rcv_combinations.sh

@@ -5,12 +5,12 @@ runs='0 1 2 3 4 5 6 7 8 9'
 for run in $runs
 do
   dataset=$dataset_path$run.pickle
-  #python main_multimodal_cls.py $dataset -o $logfile -P -U -S -c -r -z --l2 --allprob # last combination for CIKM 3 Pr(views) concatenated
+  #python main_gFun.py $dataset -o $logfile -P -U -S -c -r -z --l2 --allprob # last combination for CIKM 3 Pr(views) concatenated
   python main_multimodal_cls.py $dataset -o $logfile -P -U -S -c -r -z --l2 --allprob # last combination for CIKM 3 views concatenated
-  #python main_multimodal_cls.py $dataset -o $logfile -P -U -c -r -a -z --l2 --allprob
-  #python main_multimodal_cls.py $dataset -o $logfile -P -U -S -c -r -a -z --l2 --allprob
-  #python main_multimodal_cls.py $dataset -o $logfile -P -S -c -r -z --l2 --allprob
-  #python main_multimodal_cls.py $dataset -o $logfile -P -U -c -r -z --l2 --allprob
-  #python main_multimodal_cls.py $dataset -o $logfile -c -P -U -r -z --l2
-  #python main_multimodal_cls.py $dataset -o $logfile -c -P -U -S -r -z --l2
+  #python main_gFun.py $dataset -o $logfile -P -U -c -r -a -z --l2 --allprob
+  #python main_gFun.py $dataset -o $logfile -P -U -S -c -r -a -z --l2 --allprob
+  #python main_gFun.py $dataset -o $logfile -P -S -c -r -z --l2 --allprob
+  #python main_gFun.py $dataset -o $logfile -P -U -c -r -z --l2 --allprob
+  #python main_gFun.py $dataset -o $logfile -c -P -U -r -z --l2
+  #python main_gFun.py $dataset -o $logfile -c -P -U -S -r -z --l2
 done

src/experiment_scripts/main_deep_learning.py

@@ -1,6 +1,6 @@
 import argparse
 import torch.nn as nn
-from torch.optim.lr_scheduler import StepLR, MultiStepLR
+from torch.optim.lr_scheduler import StepLR
 from dataset_builder import MultilingualDataset
 from learning.transformers import load_muse_embeddings
 from models.lstm_class import RNNMultilingualClassifier

src/experiment_scripts/main_embeddings_cls.py

@@ -1,6 +1,5 @@
 import os
 from dataset_builder import MultilingualDataset
-from learning.learners import *
 from util.evaluation import *
 from optparse import OptionParser
 from util.file import exists

src/experiment_scripts/main_majorityvoting_cls.py

@@ -2,15 +2,12 @@ import os
 from dataset_builder import MultilingualDataset
 # from learning.learners import *
 # from learning.learners import FunnellingMultimodal
-from learning.transformers import Funnelling, PosteriorProbabilitiesEmbedder, MetaClassifier, \
-    TfidfVectorizerMultilingual, DocEmbedderList, WordClassEmbedder, MuseEmbedder, FeatureSet2Posteriors, Voting
+from learning.transformers import PosteriorProbabilitiesEmbedder, TfidfVectorizerMultilingual, WordClassEmbedder, MuseEmbedder, FeatureSet2Posteriors, Voting
 from util.evaluation import *
 from optparse import OptionParser
 from util.file import exists
 from util.results import PolylingualClassificationResults
 from sklearn.svm import SVC
-from util.util import get_learner, get_params
-from sklearn.linear_model import LogisticRegression, LogisticRegressionCV
 
 parser = OptionParser()
 

src/experiment_scripts/main_mbert_extractor.py

@@ -1,4 +1,4 @@
-from main_mbert import *
+from experiment_scripts.main_mbert import *
 import pickle
 
 

src/experiment_scripts/run_combinations_jrc.sh

@@ -7,10 +7,10 @@ logfile=./results/final_combinations_jrc.csv
 #		(no one seems to improve over standard funnelling [the improved version after A.1] with posteriors probabilities...)
 
 # aggregation=concatenation
-#python main_multimodal_cls.py $dataset -o $logfile -P -U -r -z --l2
-#python main_multimodal_cls.py $dataset -o $logfile -P -S -r -z --l2
-#python main_multimodal_cls.py $dataset -o $logfile -U -S -r -z --l2
-#python main_multimodal_cls.py $dataset -o $logfile -P -U -S -r -z --l2
+#python main_gFun.py $dataset -o $logfile -P -U -r -z --l2
+#python main_gFun.py $dataset -o $logfile -P -S -r -z --l2
+#python main_gFun.py $dataset -o $logfile -U -S -r -z --l2
+#python main_gFun.py $dataset -o $logfile -P -U -S -r -z --l2
 #
 
 ##FeatureSetToPosteriors (aggregation mean)
@@ -20,10 +20,10 @@ python main_multimodal_cls.py $dataset -o $logfile -U -S -r -a -z --l2 --allprob
 python main_multimodal_cls.py $dataset -o $logfile -P -U -S -r -a -z --l2 --allprob
 
 ##FeatureSetToPosteriors
-#python main_multimodal_cls.py $dataset -o $logfile -P -U -r -z --l2 --allprob
-#python main_multimodal_cls.py $dataset -o $logfile -P -S -r -z --l2 --allprob
-#python main_multimodal_cls.py $dataset -o $logfile -U -S -r -z --l2 --allprob
-#python main_multimodal_cls.py $dataset -o $logfile -P -U -S -r -z --l2 --allprob
+#python main_gFun.py $dataset -o $logfile -P -U -r -z --l2 --allprob
+#python main_gFun.py $dataset -o $logfile -P -S -r -z --l2 --allprob
+#python main_gFun.py $dataset -o $logfile -U -S -r -z --l2 --allprob
+#python main_gFun.py $dataset -o $logfile -P -U -S -r -z --l2 --allprob
 
 #MajorityVoting
 #python main_majorityvoting_cls.py $dataset -o $logfile -P -U -r

src/experiment_scripts/run_combinations_rcv.sh

@@ -7,10 +7,10 @@ logfile=./results/final_combinations_rcv.csv
 #		(no one seems to improve over standard funnelling [the improved version after A.1] with posteriors probabilities...)
 
 # aggregation=concatenation
-#python main_multimodal_cls.py $dataset -o $logfile -P -U -r -z --l2
-#python main_multimodal_cls.py $dataset -o $logfile -P -S -r -z --l2
-#python main_multimodal_cls.py $dataset -o $logfile -U -S -r -z --l2
-#python main_multimodal_cls.py $dataset -o $logfile -P -U -S -r -z --l2
+#python main_gFun.py $dataset -o $logfile -P -U -r -z --l2
+#python main_gFun.py $dataset -o $logfile -P -S -r -z --l2
+#python main_gFun.py $dataset -o $logfile -U -S -r -z --l2
+#python main_gFun.py $dataset -o $logfile -P -U -S -r -z --l2
 #
 ##FeatureSetToPosteriors (aggregation mean)
 python main_multimodal_cls.py $dataset -o $logfile -P -U -r -a -z --l2 --allprob
@@ -19,10 +19,10 @@ python main_multimodal_cls.py $dataset -o $logfile -U -S -r -a -z --l2 --allprob
 python main_multimodal_cls.py $dataset -o $logfile -P -U -S -r -a -z --l2 --allprob
 
 ##FeatureSetToPosteriors
-#python main_multimodal_cls.py $dataset -o $logfile -P -U -r -z --l2 --allprob
-#python main_multimodal_cls.py $dataset -o $logfile -P -S -r -z --l2 --allprob
-#python main_multimodal_cls.py $dataset -o $logfile -U -S -r -z --l2 --allprob
-#python main_multimodal_cls.py $dataset -o $logfile -P -U -S -r -z --l2 --allprob
+#python main_gFun.py $dataset -o $logfile -P -U -r -z --l2 --allprob
+#python main_gFun.py $dataset -o $logfile -P -S -r -z --l2 --allprob
+#python main_gFun.py $dataset -o $logfile -U -S -r -z --l2 --allprob
+#python main_gFun.py $dataset -o $logfile -P -U -S -r -z --l2 --allprob
 
 #MajorityVoting
 #python main_majorityvoting_cls.py $dataset -o $logfile -P -U -r

src/util/common.py

@@ -235,7 +235,7 @@ class MultilingualIndex:
     def bert_embeddings(self, bert_path, max_len=512, batch_size=64, stored_embeddings=False):
         show_gpu('GPU memory before initializing mBert model:')
         # TODO: load dumped embeddings?
-        from main_mbert_extractor import do_tokenization, ExtractorDataset, DataLoader
+        from experiment_scripts.main_mbert_extractor import do_tokenization, ExtractorDataset, DataLoader
         from transformers import BertConfig, BertForSequenceClassification
 
         print('[mBERT] generating mBERT doc embeddings')