why loss is 0???

[HHL-445]

I can access query.wikidata but why can't I count the loss. The model runs very fast, has anyone run it? dev reward=(0.0, 0.0) dev te reward=(0.0, 0.0) test reward=(0.0, 0.0) training loss=0.0 training reward=(0.0, 0.0) training te loss=0.0 training te reward=(0.0, 0.0)

[iCloud112]

I had the same problem before, but I don't know why it worked later without any changes.

[HHL-445]

I suspect that the not_update parameter of the retrieve_via_frontier function is wrong. He should be False. Otherwise, the subgraph cannot be retrieved. But I still run very slowly

[iCloud112]

Are you accessing the url "https://query.wikidata.org/sparql" with the help of vpn？I also don't know why. I am also trying to train a new model but I think there are some problems with the model I trained. The code provided by the author is a bit messy and has some errors.

[HHL-445]

Yes I use VPN.If you have any new progress, you can share it with me. Thanks

[iCloud112]

You too. Thank you!

[Bigchen8013]

I can access query.wikidata but why can't I count the loss. The model runs very fast, has anyone run it? dev reward=(0.0, 0.0) dev te reward=(0.0, 0.0) test reward=(0.0, 0.0) training loss=0.0 training reward=(0.0, 0.0) training te loss=0.0 training te reward=(0.0, 0.0)

Hi I have the same problem, have you solved it? If you have solved it, can you share with me？ Thank you！

[HHL-445]

@Bigchen8013 The value of not_update in the retrieve_KB function is args.do_train. The retrieve_via_frontier function is called in the retrieve_KB function, and the value of not_update is also passed. But the value of arg.do_train is 1 during training and the value that not_update needs to pass in retrieve_via_frontier is False. I can successfully calculate the loss after adding not_update = False to the front of retrieve_via_frontier. You can try