Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Sentence embedding computation speed on STS benchmark 2017 (test)
Loading...
6,469
Throughput (CPU, sentences/sec)
Avg. GloVe embeddings
-213
1,521.75
3,256.5
4,991.25
Aug 27, 2019
Throughput (CPU, sentences/sec)
Throughput (GPU, sentences/sec)
Updated 4d ago
Evaluation Results
Method
Method
Links
Throughput (CPU, sentences/sec)
Throughput (GPU, sentences/sec)
Avg. GloVe embeddings
2019.08
6,469
-
InferSent
Architecture=BiLSTM
2019.08
137
1,876
SBERT-base
Smart Batching=enabled
2019.08
83
2,042
Universal Sentence Encoder
Framework=TensorFlow Hub
2019.08
67
1,318
SBERT-base
Smart Batching=disabled
2019.08
44
1,378
Feedback
Search any
task
Search any
task