cherepanovic

Results 9 comments of cherepanovic

> Hi, is the output embedding structure like this: [ A, SEP, B, SEP, CLS]? If a sample has several sentences, what is to do? Should I take only two...

>with many design choices. yes of course as many other neural models. But in your standard settings, what is the exact idea behind StandardScaler and sigmoid output layer, where input...

>The reason to not use minmax scalaer for input is it is sensitive to outliers in the data. that makes sense too thank you for your response

@John-Almardeny I meant in general whether some papers exist about this topic (how to scale data for autoencoders in case of anomalies etc)

it does not work with newer release modAL 0.4.1 ^^ going to try 0.3.6 :)

> It's specified in the README. > > https://raw.githubusercontent.com/hermitdave/FrequencyWords/master/content/2016/kk/kk_50k.txt I didn't get you, but the list (link to a list) you shared is exactly the one I speak about.

have the same problem, even after have read the quick start guide :/

according to the main page stemming/lemmatization works ``` $ echo "Сәлем!" | apertium -d . kaz-morph ^Сәлем/сәлем/сәлем/сәлем/ сәлем+е/сәлем+е$ ^!/!$^./.$ $ echo "Оқу инемен құдық қазғандай." | apertium -d . kaz-tagger...