Upload external_comparison.csv with huggingface_hub
Browse files- external_comparison.csv +3 -3
external_comparison.csv
CHANGED
|
@@ -1,7 +1,4 @@
|
|
| 1 |
name,source,algorithm,architecture,vocab_size,fertility_ar,fertility_az,fertility_overall,disparity,cpt_ar,cpt_az,exact_match_ar,exact_match_az
|
| 2 |
-
concat_bpe_8000,ours,bpe,concatenated,8000,1.677255929872786,1.5274486975717831,1.625089409755206,0.0893168595399542,3.3745329363977343,3.4826508889784487,0.9989120580235721,0.9964364500254539
|
| 3 |
-
concat_wordpiece_16000,ours,wordpiece,concatenated,16000,1.4835885687430344,1.3635334584201908,1.441782460775622,0.08092210526032824,3.81082194270718,3.9034221417934196,0.9990,0.9992
|
| 4 |
-
concat_bpe_32000,ours,bpe,concatenated,32000,1.3464,1.2327,1.3068,0.0844,4.1903,4.3094,0.9989,0.9964
|
| 5 |
CaMeLBERT-MSA,external_msa,WordPiece,shared,30000,1.816901627794181,3.173471216645639,2.2892921369888652,0.42747184273672184,3.1140517054288748,1.6388977943422216,0.2987307343608341,0.3885117936534872
|
| 6 |
Asafaya-BERT,external_msa,WordPiece,shared,32000,1.7951679730188856,2.794867411003224,2.143287620128837,0.35769118565288,3.1312619206100933,1.8618984630783484,0.19895738893925657,0.15238418462582726
|
| 7 |
Aranizer-SP-86k,external_msa,SentencePiece,shared,86000,1.5940707675919776,2.5232344233957935,1.9176281406140119,0.3682430959202508,3.532878084819776,2.06546468780593,0.9976881233000907,0.9955031393178347
|
|
@@ -11,3 +8,6 @@ DarijaBERT-az,external_darija,WordPiece,shared,110000,1.6053190877519825,1.51716
|
|
| 11 |
Moroccan-Darija-Tokenizer,external_darija,BPE,shared,30000,1.5701289164575563,2.9024430947615336,2.034073102047897,0.4590319723093283,3.626351312053874,1.7930892097539761,0.00018132366273798732,0.0
|
| 12 |
Translit-Darija,external_darija,BPE,shared,30000,1.7964830125579598,1.6576487239140443,1.7481375381752506,0.0772811586157074,3.1115965796575096,3.1602226731778864,0.0,0.0
|
| 13 |
Qwen2.5-Darija,external_darija,SentencePiece,shared,151643,2.340330300987948,2.245717380136819,2.307383781897024,0.04042716569160743,2.3801472507543755,2.325374802332435,1.0,1.0
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
name,source,algorithm,architecture,vocab_size,fertility_ar,fertility_az,fertility_overall,disparity,cpt_ar,cpt_az,exact_match_ar,exact_match_az
|
|
|
|
|
|
|
|
|
|
| 2 |
CaMeLBERT-MSA,external_msa,WordPiece,shared,30000,1.816901627794181,3.173471216645639,2.2892921369888652,0.42747184273672184,3.1140517054288748,1.6388977943422216,0.2987307343608341,0.3885117936534872
|
| 3 |
Asafaya-BERT,external_msa,WordPiece,shared,32000,1.7951679730188856,2.794867411003224,2.143287620128837,0.35769118565288,3.1312619206100933,1.8618984630783484,0.19895738893925657,0.15238418462582726
|
| 4 |
Aranizer-SP-86k,external_msa,SentencePiece,shared,86000,1.5940707675919776,2.5232344233957935,1.9176281406140119,0.3682430959202508,3.532878084819776,2.06546468780593,0.9976881233000907,0.9955031393178347
|
|
|
|
| 8 |
Moroccan-Darija-Tokenizer,external_darija,BPE,shared,30000,1.5701289164575563,2.9024430947615336,2.034073102047897,0.4590319723093283,3.626351312053874,1.7930892097539761,0.00018132366273798732,0.0
|
| 9 |
Translit-Darija,external_darija,BPE,shared,30000,1.7964830125579598,1.6576487239140443,1.7481375381752506,0.0772811586157074,3.1115965796575096,3.1602226731778864,0.0,0.0
|
| 10 |
Qwen2.5-Darija,external_darija,SentencePiece,shared,151643,2.340330300987948,2.245717380136819,2.307383781897024,0.04042716569160743,2.3801472507543755,2.325374802332435,1.0,1.0
|
| 11 |
+
concat_bpe_32000,ours,bpe,concatenated,32000,1.3463628334026025,1.2326871301490916,1.306777890060823,0.08443170030690308,4.190301018162676,4.309438424198928,0.9989120580235721,0.9964364500254539
|
| 12 |
+
concat_wordpiece_80000,ours,wordpiece,concatenated,80000,1.2210699643670833,1.1112555503947363,1.1828299158213738,0.08993294174528893,4.591976174635943,4.762024638941472,0.9989120580235721,0.9964364500254539
|
| 13 |
+
concat_wordpiece_110000,ours,wordpiece,concatenated,110000,1.1934305662132043,1.0820488921400155,1.1546447596001155,0.09332899393268149,4.690435896236248,4.884534963776444,0.9989120580235721,0.9964364500254539
|