Pasquale Minervini

@dirkweissenborn I would really like to sysematically test our models with the following two functions: https://github.com/uclmr/jack/blob/master/jack/debug/base.py One checks whether all layers are updated during training, and the other verifies that...

@dirkweissenborn no I mean - I just need the training operator (`train_op`) and a `feed_dict`: where could I get those? Or maybe I can just train in the normal way,...

This is also useful for writing more fine-grained tests that check things during training I quickly turned your code here into a test: https://github.com/uclmr/jack/blob/master/tests/jack/readers/test_fastqa_loop.py

Can this be a solution? https://github.com/explosion/spaCy/issues/1725 Where's the bottleneck exactly? What's the slow code/line ?

hey Paul! It was working on another platform, but haven't tried it on osx - I'm on it!

> shows a simple 2D interactive animation that shows the adversarial learning process (and also the default learning process), using a simple model and problem. If we use a model...

Yep it's currently a problem - our code is still based on TF 1.x I can start porting it to TF 2.x in the second half of February

> For reference this is what we do in tests: https://github.com/huggingface/peft/blob/542f2470e7679768f594692fc7f59e83f48585a8/tests/test_peft_model.py#L177 If you folks want things to stay this way, I can close the issue!

@dfdazac I looked into that, and your concern seems legit is there any news on that side ? maybe a better codebase ?

Setting `max_length` fixes this Shall we set `max_length` to `getattr(tokenizer, "model_max_length", getattr(tokenizer, "max_len", None))` automatically if it's `None` or `"auto"`?