I think there is a docker memory issue or similar bottlenecking my performance on the imdb notebook.
The learner.fit(lrs/2, 1, wds=wd, use_clr=(32,2), cycle_len=1)
step is taking a very long time.
I tried to benchmark against Colab, but the session keeps crashing from lack of RAM there.
Has anyone run this imbd notebook in Colab successfully?
Or if you have run it elsewhere, can you please let me know how fast the learner.fit was? It’s over 15 minutes, and I don’t think that’s typical?