Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Controllable Text Generation (Sentiment + Tense) on Yelp (test)
Loading...
98
Accuracy (Sentiment)
GPT2-FT
65.76
74.13
82.5
90.87
Aug 1, 2022
Accuracy (Sentiment)
Accuracy (Tense)
Geometric Mean Accuracy
Perplexity (PPL)
Self-BLEU (sBL)
Updated 4d ago
Evaluation Results
Method
Method
Links
Accuracy (Sentiment)
Accuracy (Tense)
Geometric Mean Accuracy
Perplexity (PPL)
Self-BLEU (sBL)
GPT2-FT
Backbone=GPT2-large, S...
2022.08
98
95
96.9
9
36.8
LATENTOPS
Backbone=GPT2-large, A...
2022.08
98
93
95.1
25.2
19.7
PPLM
Backbone=GPT2-large, A...
2022.08
81
59
67.7
15.7
28.7
FUDGE
Backbone=GPT2-large, A...
2022.08
67
63
56.5
11
35.9
Feedback
Search any
task
Search any
task
