What exactly does the "use_cache=False" do?
#9
by
ArthurParkerhouse
- opened
In the Colab under the "load model and tokenizer" section the Cache is disabled. Does this improve the performance of the model, or does changing it to "use_cache=True" use extra resources, or make it unstable? Just curious.
Hi! It does nothing in this case; just a leftover code copy/paste from training :) I have removed it from the inference notebook
pszemraj
changed discussion status to
closed