MichelBartels
MichelBartels
Sorry, I missed this discussion. I opened this issue because it was a big bottleneck when mining a lot of possible answers (about 50 per text). If I remember correctly,...
Hi, I don't have that much experience with DDP, but as far as I know most should be already implemented. So it should be mostly about passing down a `distributed`...
Hi @mumtazcem. Thanks for the input. We will definitely be considering this when working on the issue. However, this issue is currently in our backlog so please don't expect any...
I think saving the value of `use_no_answer_legacy_confidence` doesn't work for all cases because it's not set during training and only relevant during postprocessing. The issue it is addressing is that...
This node uses composition with another node which is generally not something we would like for v2. Instead, we would probably prefer if the node was split in two separate...
@mathislucka I agree with you and we have decided at our co.work last week that without "subpipelines" or a similar abstraction the v2 concept of pipelines is not really usable....
@elundaeva That sounds good to me. @ZanSara Just to confirm, would we then want some sort of abstraction for v1? Otherwise, it's not very usable until we release v2.
@Rocketknight1 Thanks for looking into this, I am afraid I currently don't have time to contribute a fix.
@BCreativeS , the issue still needs to be addressed.
This issue still needs to be addressed.