Request for best performance model

[chocolate9624]

Hi, I really appreciate your wonderful work. But when I run the "segctcent_800w.sh" script, I get much lower Seq_Acc than they are in the paper.

I guess the performance gaps come from the training iteration steps. The scripts (segctcent_800w.sh) says 50 epochs. But in your paper, you says 150 epochs.
Am I right ?

It takes almost two months to finish the 50 epochs training steps. It seems impossible for me to finish the left 100 epochs. So could you please provide the best performance model on Synth90K dataset?

I just want to make a comparison and validate my work. Many thanks.

[Gaming722]

Hello, I want to know where I can download the dataset file "max90k_train_val.lmdb", "data/max90k_test.lmdb" to run the "segctcent_800w.sh" script.