wrong evaluation function in lexical analysis model
Created by: BladeSun
For each batch, the evalution function saves F1 score based on the accumulated counts from begining. The final F1 score is the average of the F1 scores from each batch.
Created by: BladeSun
For each batch, the evalution function saves F1 score based on the accumulated counts from begining. The final F1 score is the average of the F1 scores from each batch.