ajyl

Results 9 comments of ajyl

I suppose I could try removing the task-specific arguments in my regression test to see if multitask regression tests can be made, and that could narrow down the scope.

@dex314 What changes did you make?? Just curious. I also want to say thank you for saving me so much time.

Hmm perhaps it's because I'm using a different version of torch.

TL;DR you can probably do ``` pip3 uninstall geopy pip3 install geopy==1.13.0 ``` For repo authors, I think `setup.py` needs to be updated - I'm seeing that `setup.py` will install...

For what it's worth, I find [jupyter-ascending](https://github.com/untitled-ai/jupyter_ascending) really useful for this kind of thing -- I develop in `.py`, while using jupyter-ascending + vim plugins to automatically update an equivalent...

If I remember correctly, this specific issue for me was because I was using the wrong version of torch.

My issue was a mismatch in virtual environments -- I had `evaluate` installed system-wide and `sklearn` installed in my virtualenv. Hope this helps.

I think you want `num_samples` parameter in `explain_instance()` to be something much larger, ie the default value if I recall correctly is 5,000 or something.