first 3 stammer training

.gitattributes

@@ -53,3 +53,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.webp filter=lfs diff=lfs merge=lfs -text
 *.tsv filter=lfs diff=lfs merge=lfs -text
 *.tmx filter=lfs diff=lfs merge=lfs -text
+*.jsonl filter=lfs diff=lfs merge=lfs -text

correct/correct_datafiles/correct_dev.tsv

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fec0f2575d5925b993aefdb9715ace66f3a4948258c8cfa1600fd9b264fd8963
-size 2320597
+oid sha256:53a4aa6c8acac14492a898359d250ef543091757e0a3b3e5dd5faab817a87a08
+size 2417746

correct/correct_datafiles/correct_test.tsv

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5e7595989509c671e3bc088129a080ab4292ba472ef21bc8c92d470ebc4f1196
-size 2339422
+oid sha256:f049a0f878985fa1efd50151a4089b45d3dd10a23e45f161fd084c2714354911
+size 2417751

correct/correct_datafiles/correct_train.tsv

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d42cbf75385b1918f7dd9fa479e65519ecbe9bfcf1996132c4236ea695cd9fdf
-size 18565347
+oid sha256:47357d479c1a62b85c131d5f1cbdb01b963cd2022ffe568725a61cf92a815280
+size 19408389

correct/create_correct_dataset.py

@@ -119,6 +119,8 @@ def main(args):
 
     print(f"Removing a random number of commas 10% - '{data.loc[0]['target']}' -> '{data.loc[0]['source']}'")
     data = data.sample(frac=1).reset_index(drop=True)
+    data.loc[:,'source'] = "correct: "+data['source']
+    
 
     # Train - test - dev
     train, test = train_test_split(data, test_size=0.2)

create_training_files.py

@@ -43,6 +43,14 @@ data['source'] = data['source'].str[4:]
 
 data = data.sample(frac=1).reset_index(drop=True)
 
+#Clean before splitting
+data['source'] = data['source'].str.replace('\t',' ')
+data['source'] = data['source'].str.replace('\n',' ')
+data['target'] = data['target'].str.replace('\t',' ')
+data['target'] = data['target'].str.replace('\n',' ')
+
+
+
 # Train - test - dev
 train, test = train_test_split(data, test_size=0.2)
 test, dev = train_test_split(test, test_size=0.5)
@@ -50,21 +58,32 @@ test, dev = train_test_split(test, test_size=0.5)
 # Write the datasets to disk
 test.to_csv('tsv_all_target/test.tsv', index=False, header=False, sep='\t')
 dev.to_csv('tsv_all_target/dev.tsv', index=False, header=False, sep='\t')
-
+print("Finished writint dev and test. Now we are adding some extra language detection and English translation")
 
 # Add the language detection to the training dataset
 filelist = glob.glob('langid/langid_datafiles/*train.tsv')
 
-print(len(train))
-
 for tsvfile in filelist:
     tmp = pd.read_csv(tsvfile, sep='\t')
     tmp.columns=['source','target']
 
     train = pd.concat([train,tmp])
 
-data = data.sample(frac=1).reset_index(drop=True)
-print(len(train))
+
+# Add extra English training set - Currently we do not add this to the test/dev-collection
+tmp = pd.read_csv('eng_nob/train.tsv', sep='\t')
+tmp.columns=['source','target']
+train = pd.concat([train,tmp])
+
+#Final shuffle
+train = train.sample(frac=1).reset_index(drop=True)
+
+#Clean before splitting
+train['source'] = train['source'].str.replace('\t',' ')
+train['source'] = train['source'].str.replace('\n',' ')
+train['target'] = train['target'].str.replace('\t',' ')
+train['target'] = train['target'].str.replace('\n',' ')
+
 
 train.to_csv('tsv_all_target/train.tsv', index=False, header=False, sep='\t')
 

eng_nob/create_english_train.py

@@ -0,0 +1,42 @@
+import pandas as pd
+import xmltodict
+from sklearn.model_selection import train_test_split
+import glob
+import sys
+import os
+
+filelist = glob.glob('*.jsonl')
+
+for jsonfile in filelist:
+    data = pd.DataFrame([['source','target']])
+    temp = pd.DataFrame()
+    
+    print(f"Processing {jsonfile}")
+    temp = pd.read_json(jsonfile, lines=True,encoding='utf8')
+    errors = 0 
+    for index, row in temp.iterrows():
+        try:
+            engnob = ['nob: '+str(row['en']),str(row['no'])]
+            data.loc[len(data)] = engnob
+            nobeng = ['eng: '+str(row['no']),str(row['en'])]
+            data.loc[len(data)] = nobeng
+            
+        except:
+            errors += 1
+            print("Unable to convert this line")
+            print(row)
+    try:
+        data['source'] = data['source'].str.replace('\t',' ')
+        data['target'] = data['target'].str.replace('\t',' ')
+    except:
+        errors += 1
+        print("Key error")
+
+    data = data.sample(frac=1).reset_index(drop=True)
+
+    filename = jsonfile.replace(".jsonl",".tsv")
+
+    # Write the datasets to disk
+    data.to_csv(filename, index=False, header=False, sep='\t')
+
+print("Finished")

eng_nob/dev.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7825999c9ad6a7a7e77515783b7cf6a1b43795194f6864058ca1f9cd7be0fea8
+size 3197577

eng_nob/dev.tsv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a3b9df30f2a0e69273055c8708d4ef042389d212a81ed5c03fda066a71f6057d
+size 4963224

eng_nob/test.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:442a5ec7b689293feea897cff4c56590d63c137f650b7a88416412e6e42713b2
+size 3206230

eng_nob/test.tsv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fb25a3d288398c2c8b4f5eab2d6ca0ff0fa6edfd57986f4ee75d0d182b69decd
+size 4983910

eng_nob/train.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:434f255a6b48b46b2f6124df3a10eb4cfccd74a2405aeb2ce36998d4644da597
+size 25675727

eng_nob/train.tsv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:78be2aeb8740785607219544dc2eb6182ebaa9fc6d985f734eff66e123b53eb8
+size 39894522

tsv_all_target/dev.tsv

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6033fef40a251fa84c0bd2371dc137c671ca3c97d255765b0a0903ecd69cd0fb
-size 29236881
+oid sha256:25c38a038bd776b625a3b70588a0262f02880c29800a0ffc883b3ee1b26aa30d
+size 29188226

tsv_all_target/dev_small.tsv

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2b8ecd9a7c55562eea70663bb7ed62639ed29cb88f5231136e690e72670f0cf8
-size 2011164
+oid sha256:c68ec7597310bcaf1f9fa7c6a2863d36775e75f6f9c434e3fdebc8d2e2aa2553
+size 2022316

tsv_all_target/test.tsv

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:433d5360a91971fc3ee2cc4627e0233603d37e30bf33d2aec6dcbfc58072c608
-size 29096305
+oid sha256:e6f6a9bf9fe10ec302d1b4d00e0a3c0914606e0111928f5697ee387a10dcbc67
+size 29178110

tsv_all_target/test_small.tsv

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6963609d621487dfab571e6bbda2beac49f5960f8f8cc46a9e39e88b0db8941a
-size 2046850
+oid sha256:dd36bdc1a55f3c44cadfbdde2c3b2992282f69613c3b46e63ea6b3264d9185a8
+size 2037392

tsv_all_target/train.tsv

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:301034a5c29ca96a946b999c9bb83a3e8f2e371f063ff62c0104e534094236c3
-size 244651793
+oid sha256:86e2efaec9f70719518ef21139247cbc3b3d4a3cb63a63498da7e13c44995c51
+size 284512806