|
NMT's vocabulary
|
|
2
|
2955
|
February 7, 2018
|
|
Is there any efficient reference set(English) for bleu scoring?
|
|
2
|
1425
|
January 17, 2018
|
|
Word-level Distillation
|
|
1
|
1844
|
December 28, 2017
|
|
What is the use of Monolingual Corpora in SMT
|
|
2
|
1629
|
December 6, 2017
|
|
Training with out-of-domain data
|
|
0
|
1537
|
November 3, 2017
|
|
How is google's transformer translate between Chinese and English?
|
|
0
|
1680
|
November 1, 2017
|
|
Multiple translations open nmt
|
|
1
|
1371
|
October 31, 2017
|
|
Quality estimation / pred_score
|
|
3
|
1918
|
October 25, 2017
|
|
Paragraph vs sentence segmentation
|
|
2
|
2840
|
October 16, 2017
|
|
What is the best way to generate parallel corpus as per your past experience?
|
|
0
|
1329
|
September 26, 2017
|
|
In parallel corpus what should I keep, single sentence in each line or paragraph for MT?
|
|
2
|
1311
|
September 19, 2017
|
|
Metrics (Bleu, ppl, gold ppl, pred ....)
|
|
7
|
20351
|
September 14, 2017
|
|
About mask in attention layer and decoder
|
|
4
|
2429
|
September 12, 2017
|
|
How to create models that translate very quickly
|
|
7
|
3628
|
August 30, 2017
|
|
Exposure bias during training
|
|
3
|
2393
|
August 10, 2017
|
|
Improve BLEU by Coverage and Context Gate
|
|
6
|
2187
|
August 10, 2017
|
|
Use of Monolingual corpora in OPENNMT
|
|
7
|
3475
|
July 26, 2017
|
|
Using features for domain/client/subject adaptation
|
|
6
|
2281
|
July 12, 2017
|
|
Did someone test RNN with larger recurrence?
|
|
5
|
1420
|
June 23, 2017
|
|
Training chatbot with multiple inputs (to add context)
|
|
4
|
4617
|
June 22, 2017
|
|
Corpus compactization
|
|
2
|
1848
|
June 21, 2017
|
|
Incremental vocab
|
|
6
|
1880
|
June 21, 2017
|
|
Automatic training corpus filtering
|
|
15
|
4551
|
June 16, 2017
|
|
Implementing hashing to massively improve performance (by 95%)
|
|
0
|
1310
|
June 13, 2017
|
|
Speech-to-text using Convolutional LSTM layers
|
|
1
|
1457
|
June 8, 2017
|
|
Sentence length affect perplexity decrease
|
|
2
|
2513
|
June 2, 2017
|
|
Simple combination between SMT and NMT
|
|
0
|
1531
|
May 31, 2017
|
|
Alternative methods for <UNK> substitution
|
|
5
|
2700
|
May 29, 2017
|
|
Importance Sampling - training speed
|
|
1
|
1674
|
May 29, 2017
|
|
Weird Output after many attempts to train
|
|
0
|
1432
|
May 29, 2017
|