Open johnp2266 opened 6 years ago
labels are collected from training data automatically. https://github.com/jiegzhan/multi-class-text-classification-cnn/blob/8961f9277375c821023656235b8c41e13eb4b9a8/train.py#L39
Side question, why the accuracy so low? I tried training with the original files but instead of 1 epoch, I changed to 10. That's all I changed but the accuracy is only 89 when I predict.
If I want to change the labels, say I want to classify 7 labels, do I just change labels.json?