11–20 of 22 posts
So... We are now automating laziness? There goes my last stronghold...
Not quiet! Choosing between using a fast mostly accurate model and a slow very accurate model. Time saved vs accuracy gained.
No usually "that person", but: quite
Only pointing it out because of the irony: Choosing between using a fast mostly accurate model and a slow very accurate model.
So... We are now automating laziness? There goes my last stronghold...
Skimmming is definitely not a bad thing.
cool paper, i enjoyed following the post (and really appreciated the effort put into the bokeh viz!) i wonder if this would have been improved by being clearer about the motivation. the authors frame it as though the ~60ms penalty for using the LSTM for prediction is a huge burden, and i can imagine situations where it is. however, it seems like if this is the case, we need some real life/"scaled out" examples of how…
Earlier quoted context omitted.
Not quiet! Choosing between using a fast mostly accurate model and a slow very accurate model. Time saved vs accuracy gained.
Not quiet! No usually "that person", but: quite Only pointing it out because of the irony: Choosing between using a fast mostly accurate model and a slow very accurate model.
You did that on purpose ;)
Earlier quoted context omitted.
Skimmming is definitely not a bad thing.
Not so sure. I do skim sometimes, but it always leaves me with are dirty feeling of cheating. Also one needs to go back and check whether something important was missed and this undoes most of the benefit of skimming. The proper place or skimming would be to skim before reading just to get an overview of what's to come, or to skim while evaluating if the thing is worth reading in the first place.
cool paper, i enjoyed following the post (and really appreciated the effort put into the bokeh viz!) i wonder if this would have been improved by being clearer about the motivation. the authors frame it as though the ~60ms penalty for using the LSTM for prediction is a huge burden, and i can imagine situations where it is. however, it seems like if this is the case, we need some real life/"scaled out" examples of how…
A 'batch' is how much of the data you put in memory at once while training the NN. To train even a small language model, you'll go through 1000s of batches, so the time difference is way bigger than it sounds. I agree a more practical example would have been nice -- maybe it'll come out in the paper.
Earlier quoted context omitted.
No usually You did that on purpose ;)
Wow, I deserve that!