# Training Rasa NLU model on AWS EC2 p2.xlarge Instance **URL:** \ **Category:** Rasa Open Source\ **Created:** [October 15, 2019, 3:28am UTC](https://forum.rasa.com/t/training-rasa-nlu-model-on-aws-ec2-p2-xlarge-instance/19793 "2019-10-15T03:28:56Z")\ **Posts on this page:** 11\ **Page:** 1
**Author:** ![BHARATHPABBA](https://dub1.discourse-cdn.com/flex013/user_avatar/forum.rasa.com/bharathpabba/32/2677_2.png) [@BHARATHPABBA](https://forum.rasa.com/u/BHARATHPABBA)\ **Post date:** [October 15, 2019, 3:28am UTC](https://forum.rasa.com/t/training-rasa-nlu-model-on-aws-ec2-p2-xlarge-instance/19793/1 "2019-10-15T03:28:56Z")
Hi, It takes 2 days to train the NLU model on my local computer. Can I use p2.xlarge instances to reduce the training time? Does Rasa NLU supports p2.xlarge ? are there any limitations? ---
**Author:** ![amn41](https://dub1.discourse-cdn.com/flex013/user_avatar/forum.rasa.com/amn41/32/20_2.png) [@amn41](https://forum.rasa.com/u/amn41)\ **Post date:** [October 15, 2019, 7:03am UTC](https://forum.rasa.com/t/training-rasa-nlu-model-on-aws-ec2-p2-xlarge-instance/19793/2 "2019-10-15T07:03:04Z")
my first question is how much data do you have that it takes two days to train? My guess is that you synthetically generated this data with a script or a tool like chatito. That’s a bad idea, much cleaner to use a lookup table if you have a large number of predefined values. ---
**Author:** ![BHARATHPABBA](https://dub1.discourse-cdn.com/flex013/user_avatar/forum.rasa.com/bharathpabba/32/2677_2.png) [@BHARATHPABBA](https://forum.rasa.com/u/BHARATHPABBA)\ **Post date:** [October 15, 2019, 7:13am UTC](https://forum.rasa.com/t/training-rasa-nlu-model-on-aws-ec2-p2-xlarge-instance/19793/3 "2019-10-15T07:13:50Z")
Hi @amn41 Thanks for your reply. The Rasa NLU dataset i use contains 411 intents with around 70k utterances. This data is manually generated. ---
**Author:** ![amn41](https://dub1.discourse-cdn.com/flex013/user_avatar/forum.rasa.com/amn41/32/20_2.png) [@amn41](https://forum.rasa.com/u/amn41)\ **Post date:** [October 15, 2019, 7:27am UTC](https://forum.rasa.com/t/training-rasa-nlu-model-on-aws-ec2-p2-xlarge-instance/19793/4 "2019-10-15T07:27:05Z")
wow, ok, that’s a lot of annotated data. I would recommend the supervised embeddings pipeline which uses tensorflow and should scale better to this size. I’m not sure how many epochs you have set, but with that much data I suspect you can reduce it to a v small number without losing performance ---
**Author:** ![BHARATHPABBA](https://dub1.discourse-cdn.com/flex013/user_avatar/forum.rasa.com/bharathpabba/32/2677_2.png) [@BHARATHPABBA](https://forum.rasa.com/u/BHARATHPABBA)\ **Post date:** [October 15, 2019, 7:31am UTC](https://forum.rasa.com/t/training-rasa-nlu-model-on-aws-ec2-p2-xlarge-instance/19793/5 "2019-10-15T07:31:40Z")
yes i use supervised embeddings. currently using default epochs. i will try to change it and see again. thanks a lot for replying 🙂 ---
**Author:** ![BHARATHPABBA](https://dub1.discourse-cdn.com/flex013/user_avatar/forum.rasa.com/bharathpabba/32/2677_2.png) [@BHARATHPABBA](https://forum.rasa.com/u/BHARATHPABBA)\ **Post date:** [October 16, 2019, 1:31am UTC](https://forum.rasa.com/t/training-rasa-nlu-model-on-aws-ec2-p2-xlarge-instance/19793/8 "2019-10-16T01:31:41Z")
hi @amn41 can you please confirm if i train the nlu model on gpu instances may reduce the training time? ---
**Author:** ![prasgaut](https://avatars.discourse-cdn.com/v4/letter/p/3ab097/32.png) [@prasgaut](https://forum.rasa.com/u/prasgaut)\ **Post date:** [February 27, 2020, 9:20am UTC](https://forum.rasa.com/t/training-rasa-nlu-model-on-aws-ec2-p2-xlarge-instance/19793/9 "2020-02-27T09:20:38Z")
Hi Bharat, I am also facing the same issue. Can you please share any finding you have got to reduce the training time. Thanks Prashant ---
**Author:** ![dgslv](https://dub1.discourse-cdn.com/flex013/user_avatar/forum.rasa.com/dgslv/32/9849_2.png) [@dgslv](https://forum.rasa.com/u/dgslv)\ **Post date:** [October 30, 2020, 10:45am UTC](https://forum.rasa.com/t/training-rasa-nlu-model-on-aws-ec2-p2-xlarge-instance/19793/10 "2020-10-30T10:45:59Z")
Hi guys. I’m also facing the same issue. Would love to hear from you some tips on reducing the training time. ---
**Author:** ![amn41](https://dub1.discourse-cdn.com/flex013/user_avatar/forum.rasa.com/amn41/32/20_2.png) [@amn41](https://forum.rasa.com/u/amn41)\ **Post date:** [November 2, 2020, 10:38am UTC](https://forum.rasa.com/t/training-rasa-nlu-model-on-aws-ec2-p2-xlarge-instance/19793/11 "2020-11-02T10:38:54Z")
hi @dgslv - do you also have such a large amount of training data? I have to warn you that synthetically generating a lot of training data with a script is a bad idea. ---
**Author:** ![dgslv](https://dub1.discourse-cdn.com/flex013/user_avatar/forum.rasa.com/dgslv/32/9849_2.png) [@dgslv](https://forum.rasa.com/u/dgslv)\ **Post date:** [November 3, 2020, 10:04am UTC](https://forum.rasa.com/t/training-rasa-nlu-model-on-aws-ec2-p2-xlarge-instance/19793/12 "2020-11-03T10:04:49Z")
Hi @amn41! I have 41k manually annotated phrases on my training data. they are not synthetically generated. ---
**Author:** ![amn41](https://dub1.discourse-cdn.com/flex013/user_avatar/forum.rasa.com/amn41/32/20_2.png) [@amn41](https://forum.rasa.com/u/amn41)\ **Post date:** [November 18, 2020, 12:01pm UTC](https://forum.rasa.com/t/training-rasa-nlu-model-on-aws-ec2-p2-xlarge-instance/19793/13 "2020-11-18T12:01:42Z")
wow that’s awesome! what annotation tool do you use?