# NLP models on Elastic 8 Cloud

**URL:** <https://discuss.elastic.co/t/nlp-models-on-elastic-8-cloud/303199>\
**Category:** Elasticsearch\
**Tags:** elastic-stack-machine-learning\
**Created:** [April 25, 2022, 8:39pm UTC](https://discuss.elastic.co/t/nlp-models-on-elastic-8-cloud/303199 "2022-04-25T20:39:44Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![Wonder\_Garance](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/wonder_garance/32/91180_2.png) [@Wonder\_Garance](https://discuss.elastic.co/u/Wonder_Garance)\
**Post date:** [April 25, 2022, 8:39pm UTC](https://discuss.elastic.co/t/nlp-models-on-elastic-8-cloud/303199/1 "2022-04-25T20:39:44Z")

</div>

Hello, I started a free 14-days trial of Elastic 8 Cloud, to test NLP models.

On this page [Third party NLP models | Machine Learning in the Elastic Stack [8.1] | Elastic](https://www.elastic.co/guide/en/machine-learning/current/ml-nlp-model-ref.html) there are a lot of models, I just tried [sentence-transformers/LaBSE · Hugging Face](https://huggingface.co/sentence-transformers/LaBSE)  
It needs 3.7 Gb of memory to start the model, but by default only 1 Gb of memory is available for the machine learning.

Besides of the default model lang\_ident\_model\_1, which one in the third party NLP models may I test on the free cloud version? Thanks.

---

<div class="post-metadata">

**Author:** ![stephenb](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/stephenb/32/40856_2.png) [@stephenb](https://discuss.elastic.co/u/stephenb)\
**Post date:** [April 25, 2022, 11:11pm UTC](https://discuss.elastic.co/t/nlp-models-on-elastic-8-cloud/303199/2 "2022-04-25T23:11:54Z")

</div>

Unfortunately at this time the Free Trial only supports 1GB of ML Memory... if you want to go bigger you will need to put in a Credit Card ... I pinged some of the ML folks if they can suggest a model that will fit in the 1GB.

---

<div class="post-metadata">

**Author:** ![Wonder\_Garance](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/wonder_garance/32/91180_2.png) [@Wonder\_Garance](https://discuss.elastic.co/u/Wonder_Garance)\
**Post date:** [April 27, 2022, 9:54am UTC](https://discuss.elastic.co/t/nlp-models-on-elastic-8-cloud/303199/3 "2022-04-27T09:54:49Z")

</div>

Thank you for your response, I wait for the feedback.

---

<div class="post-metadata">

**Author:** ![stephenb](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/stephenb/32/40856_2.png) [@stephenb](https://discuss.elastic.co/u/stephenb)\
**Post date:** [April 27, 2022, 9:31pm UTC](https://discuss.elastic.co/t/nlp-models-on-elastic-8-cloud/303199/4 "2022-04-27T21:31:46Z")

</div>

I think you can try to expand the ml node to 2GB

Here are some smaller modes that may fit into 2GB

# Fill Mask Models

The smallest supported Fill Mask model will not fit no a 2GB node after 8.3

- [distilbert/distilroberta-base · Hugging Face](https://huggingface.co/distilroberta-base)
  - Size of binary PyTorch file: 316MB
  - Size of TorchScript file: 328MB
  - Quantized Size of TorchScript file: exactly the same as un-quantized

# Named Entity Recognition Models

- [elastic/distilbert-base-uncased-finetuned-conll03-english · Hugging Face](https://huggingface.co/elastic/distilbert-base-uncased-finetuned-conll03-english)
  - Size of binary PyTorch file: 253MB
  - Size of TorchScript file: 265MB
  - Quantized size of TorchScript file: exactly the same as un-quantized

# Text Classification Models

- [distilbert/distilbert-base-uncased-finetuned-sst-2-english · Hugging Face](https://huggingface.co/distilbert-base-uncased-finetuned-sst-2-english)
  - Size of binary PyTorch file: 255MB
  - Size of TorchScript file: 267MB
  - Quantized Size of TorchScript file: exactly the same as un-quantized

- [bhadresh-savani/distilbert-base-uncased-emotion at main](https://huggingface.co/bhadresh-savani/distilbert-base-uncased-emotion/tree/main)
  - Size of binary PyTorch file: 255MB
  - Size of TorchScript file: 267MB
  - Quantized Size of TorchScript file: exactly the same as un-quantized

# Text Embedding Models

- [sentence-transformers/all-MiniLM-L12-v2 · Hugging Face](https://huggingface.co/sentence-transformers/all-MiniLM-L12-v2)
  - Size of binary PyTorch file: 127MB
  - Size of TorchScript file: 133.4MB
  - Quantized size of TorchScript file: exactly the same as un-quantized

- [sentence-transformers/msmarco-distilbert-base-tas-b · Hugging Face](https://huggingface.co/sentence-transformers/msmarco-distilbert-base-tas-b)
  - Size of binary PyTorch file: 253MB
  - Size of TorchScript file: 265.4MB

# Zero Shot Classification Models

- [typeform/distilbert-base-uncased-mnli · Hugging Face](https://huggingface.co/typeform/distilbert-base-uncased-mnli)
  - Size of binary PyTorch file: 255MB
  - Size of TorchScript file: 267MB
  - Quantized size of TorchScript file: exactly the same as un-quantized

- [typeform/mobilebert-uncased-mnli · Hugging Face](https://huggingface.co/typeform/mobilebert-uncased-mnli)
  - Size of binary PyTorch file: 94.2MB
  - Size of TorchScript file: 100MB
  - Quantized size of TorchScript file: exactly the same as un-quantized

---

<div class="post-metadata">

**Author:** ![system](https://us1.discourse-cdn.com/elastic/original/3X/1/a/1ac57faf039f6b580b3f104ef42a2a89e41014de.png) [@system](https://discuss.elastic.co/u/system)\
**Post date:** [May 25, 2022, 9:32pm UTC](https://discuss.elastic.co/t/nlp-models-on-elastic-8-cloud/303199/5 "2022-05-25T21:32:41Z")

</div>

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.