Our new X account is live! Follow @wizwand_team for updates

Natural Language Inference on SNLI (train)

99.7Accuracy

Lexicalized classifier

Updated 4d ago

Evaluation Results

Method
Lexicalized classifier 2015.09	99.7	-
Lexicalized classifier 2016.03	99.7	-
Lexicalized Classifier 2016.06	99.7	-
Classifier 2015.12	99.7	-
Classifier with handcrafted features 2016.07	99.7	-
Classifier with handcrafted features 2016.07	99.7	-
+ Unigram and bigram features 2017.09	99.7	-
Bowman et al. 2018.02	99.7	-
Handcrafted features 2016.09	99.7	-
1024D pretrained GRU encoders 2016.03	98.8	-
1024D pretrained GRU encoders 2016.06	98.8	-
1024D GRU encoders 2017.09	98.8	-
Vendrov et al. 2018.02	98.8	-
1024D pretrained GRU encoders 2016.09	98.8	-
Human Performance (Estimated) 2018.02	97.2	-
DR-BiLSTM 2018.02	94.8	-
DR-BiLSTM 2018.02	94.8	-
DR-BiLSTM 2018.02	94.1	-
DR-BiLSTM 2018.02	94.1	-
Chen et al. 2018.02	93.5	-
HIM (600D ESIM + 300D Syntactic tree-LSTM) 2016.09	93.5	-
Wang et al. 2018.02	93.2	-
600D Gumbel TreeLSTM encoders 2018.01	93.1	-
300D ReSAN 2018.01	92.6	-
Chen et al. 2018.02	92.6	-
600D ESIM 2016.09	92.6	-
Fine-tuning 2021.08	92.6	-
Gong et al. 2018.02	92.3	-
300D mLSTM 2016.06	92	-
mLSTM 2015.12	92	-
mLSTM word-by-word attention 2016.07	92	-
mLSTM word-by-word attention 2016.07	92	-
Wang and Jiang 2018.02	92	-
300D mLSTM 2016.09	92	-
Bi-GRU 2018.01	91.9	-
Hierarchical CNN 2018.01	91.3	-
mLSTM with bi-LSTM sentence modeling 2015.12	91.3	-
Gong et al. 2018.02	91.2	-
DISAN 2018.01	91.1	-
Directional self-attention network (DiSAN) 2017.09	91.08	-
600D Residual stacked encoders 2018.01	91	-
mLSTM 2015.12	91	-
Wang et al. 2018.02	90.9	-
Sha et al. 2018.02	90.7	-
300D re-read LSTM 2016.09	90.7	-
600D Deep Gated Attn. 2018.01	90.5	-
Decomposable Attention Model (intra-sentence attention) 2016.06	90.5	-
Decomposable attention model 2016.07	90.5	-
Decomposable Attention Model 2016.07	90.5	-
Parikh et al. 2018.02	90.5	-
Intra-sentence attention + 200D decomposable attention model 2016.09	90.5	-
Bi-LSTM 2018.01	90.4	-
Bi-LSTM with s2t self-attention 2017.09	90.39	-
LM-BFF 2021.08	90.3	-
DISAN without directions 2017.09	90.18	-
Multi-head 2018.01	89.6	-
Multi-head with s2t self-attention 2017.09	89.58	-
Decomposable Attention Model (vanilla) 2016.06	89.5	-
LSTMN with deep attention fusion 2016.07	89.5	-
200D decomposable attention model 2016.09	89.5	-
DART 2021.08	89.5	-
Multi-window CNN 2018.01	89.3	-
300D SPINN-PI encoders 2018.01	89.2	-
300D SPINN-PI (parsed input) encoders 2016.03	89.2	-
300D SPINN-PI encoders 2016.06	89.2	-
SPINN-PI encoders 2016.07	89.2	-
SPINN-PI encoders 2016.07	89.2	-
300D SPINN-PI encoders 2017.09	89.2	-
Bowman et al. 2018.02	89.2	-
300D SPINN-PI encoders 2016.09	89.2	-
mLSTM with word embedding 2015.12	88.6	-
300D btree-LSTM encoders 2016.09	88.6	-
450D LSTMN with deep attention fusion 2016.06	88.5	-
Full tree matching NTI-SLSTM-LSTM global attention 2016.07	88.5	-
LSTMN with deep attention fusion 2016.07	88.5	-
Full tree matching NTI-SLSTM-LSTM global attention 2016.07	88.5	-
Liu et al. 2018.02	88.5	-
Yu and Munkhdalai 2018.02	88.5	-
450D LSTMN with deep attention fusion 2016.09	88.5	-
300D NTI-SLSTM-LSTM 2016.09	88.5	-
NTI-SLSTM-LSTM node-by-node tree attention 2016.07	88.1	-
NTI-SLSTM-LSTM node-by-node global attention 2016.07	87.6	-
Tree matching NTI-SLSTM-LSTM global attention 2016.07	87.6	-
Tree matching NTI-SLSTM-LSTM tree attention 2016.07	87.3	-
300D SPINN (unparsed input) encoders 2016.03	87.2	-
MMA-NSE 2016.07	87.1	-
MMA-NSE attention 2016.07	86.9	-
four stacked TC-LSTMs 2016.05	86.7	-
Word-by-word attention 2015.09	86.6	-
Attention 2015.09	86.5	-
600D Bi-LSTM encoders 2018.01	86.4	-
600D Bi-LSTM encoders 2017.09	86.4	-
Word Embedding with s2t self-attention 2017.09	86.22	-
300D NSE encoders 2018.01	86.2	-
NSE 2016.07	86.2	-
300D NSE encoders 2017.09	86.2	-
Yu and Munkhdalai 2018.02	86.2	-
300D NSE encoders 2016.09	86.2	-
NTI-SLSTM node-by-node tree attention 2016.07	86	-
Word-by-word attention (our implementation) 2015.12	85.5	-

Showing 100 of 169 rows