xiaolan77

@SmileYuhao I've got the same doubt. Have you tried other implementation (e.g., using previous prediction in loop_fn)?

@SmileYuhao I've compared both implementation on synthetic dataset. They are having similar performance. The new one, using previous predict, performed slightly better. In case you are curious.

In case you are struggling in reproducing the results. I'm using the latest code and reached F1 0.93466 on MSRA at epoch 9. The training has not stopped yet.