questions about the paper on arxiv

[liangbright]

"4 Experiments and Discussion
For the sake of brevity, we only report mean BLEU, standard deviation, highest BLEU in
parantheses, and model size in the following tables. Log perplexity, tokens/sec and convergence
times can be found in the supplementary material tables"

Q1: where are the supplementary material tables?

---

"Table 3: BLEU scores on newstest2013, varying the encoder and decoder depth and type of residual
connections "

Q2: what units are used? LSTM?
Q3: when evaluating an encoder (e.g. Enc-2), a decoder must be chosen. So, a 2D table should be used
each column indicates an encoder and each row indicates an decoder. Basically, this table in the paper is confusing.