Graph-to-seq-comment-generation/log.txt at master · DesmonDay/Graph-to-seq-comment-generation · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26

my model: 87.3 running
1.word-level bert + gcn + copy --- done
2.double bert + copy --- done

baselines:  87.185
1.seq2seq (original content) --- done
2.seq2seq (title) --- done
3.seq2seq (title+content) 87.22 done
3.bow2seq (content bag of words)    --- done
4.bow2seq (keywords)    --- done
5.h_attention --87.150 (mask running)

sport
1.word-level bert + gcn + copy 87.9 done
2.double bert + copy 87.185 done
3.word-level bert + gnn + copy 87.9, 87.3
4.word-level bert + copy 87.185, 87.3

baselines:
1. seq2seq (content) 87.3 done
2. seq2seq (title) 87.3 done
3. seq2seq (title+content)  87.54 running
3. bow2seq 87.54 done
4. keyword 87.9 done
5. h_attention --87.185 (mask running)