How to train 10000000 chilly corpus?

I want to implement a chat robot。it has 10000000 chilly corpus,my linux has 32G memory 。but

hey @mars, i am not clear by your question can you please elaborate it and another thing i noticed, as far as the error i can see from your attachment, you need to follow the rasa nlu training data format :slight_smile:

my nlu.json has 10000000 intent,how can i train?

it depends on your machine processing capacity :sweat_smile:

There are other technical implementations? perfect match?fuzzy search?

nlu train 10000 intent , it is very slowly。。。:sweat_smile:

i am also curious for this query lets see what @akelad has to say for this :sweat_smile:

Um, you wait :smiley: and increase your memory, or delete some of your data. Is that really that meany distinct intents? because that seems excessive