Table 5 USPTO-50K data statistics.

From: G2Retro as a two-step graph generative models for retrosynthesis prediction

Dataset

Statistics

# Training reactions

40,008

# Validation reactions

5001

# Test reactions

5007

Training reactions

Average size of products

26.0

 

Average size of larger reactants

21.9

 

Average size of smaller reactants

9.0

 

Average number of reactants

1.7

Validation reactions

Average size of products

25.9

 

Average size of larger reactants

21.8

 

Average size of smaller reactants

9.1

 

Average number of reactants

1.7

Test reactions

Average size of products

25.9

 

Average size of larger reactants

21.7

 

Average size of smaller reactants

9.2

 

Average number of reactants

1.7