search

hy-struggle / PRGC

PRGC: Potential Relation and Global Correspondence Based Joint Relational Triple Extraction
107 stars 16 forks source link

Whether need to run 100 epochs? #4

Closed CheaSim closed 2 years ago

CheaSim commented 2 years ago

image I have run 20 epochs, and the f1 score seems not to be higher. Also, I can not reproduce the same result in NYT as in the paper? So we must run 100 epochs?

hy-struggle commented 2 years ago

For fair comparison, we chose the same epochs as Tplinker. In NYT*, we trained 100 epochs and chose the last performance on the validation set, and set rel_threshold=0.2, mat_threshold=0.8 on the test set. You can reproduce the same result when you do the same.

CheaSim commented 2 years ago

ok, thanks for your reply.

32209056 commented 11 months ago

File "D:/sulisha/HGT-DGL-master/PRGC-main/PRGC-main/train.py", line 224, in train_and_evaluate(model, params, ex_params, args.restore_file) File "D:/sulisha/HGT-DGL-master/PRGC-main/PRGC-main/train.py", line 153, in train_and_evaluate valmetrics, , _ = evaluate(model, val_loader, params, ex_params, mark='Val') File "D:\sulisha\HGT-DGL-master\PRGC-main\PRGC-main\evaluate.py", line 92, in evaluate ex_params=ex_params) ValueError: not enough values to unpack (expected 4, got 3) 请问一下怎么解决