Model evaluation error

[binglinchengxiash]

When the model is evaluated, an error is reported using java-small dataset.

[urialon]

Hi @binglinchengxiash , thank you for your interest in our work!

It seems that the denominator is zero. This can be fixed by adding a line such as:

if self.nr_true_positives + self.nr_false_positives == 0:
   return 0

But I wonder why did it happen, it might hint that the test set is empty? Do you know whether there are any examples in the test/validation set?

[binglinchengxiash]

Hi @binglinchengxiash , thank you for your interest in our work!

It seems that the denominator is zero. This can be fixed by adding a line such as:

if self.nr_true_positives + self.nr_false_positives == 0:
   return 0

But I wonder why did it happen, it might hint that the test set is empty? Do you know whether there are any examples in the test/validation set?

I use astminer to generate data for feeding code2vec.And the dataset is java-small. I found all results are zeros.I am not sure the reason.

[urialon]

Hi @binglinchengxiash ,
Sorry for the delayed response.

Have you managed to solve this?
This may be a problem with astminer.

Best,
Uri

[binglinchengxiash]

Thank you, I have solved my problems.

[WQR53]

Hi @binglinchengxiash ,

I met the same problem that PRECISION, RECALL, and F1 were always zero when training.
"
2022-05-03 01:26:07,479 INFO After 12 epochs -- top10_acc: [0.32272727 0.51818182 0.61363636 0.66818182 0.69090909 0.73636364 0.75454545 0.76363636 0.76818182 0.79545455], precision: 0.0, recall: 0.0, F1: 0
2022-05-03 01:26:16,707 INFO Average loss at batch 100: 0.008121, throughput: 399 samples/sec
"
Do you know what causes this problem？What should I do to fix it?

I would be very grateful indeed for any help you could give me.

[urialon]

Hi @WQR53 ,
Feel free to create a new issue to describe your problem, your settings, and your command lines.

Uri

[WQR53]

Hi @urialon ,
I opened a new issue here . Thank you!