Skip to content

Docker images for production NLP usage including deep learning

License

Notifications You must be signed in to change notification settings

poteha/docker-nlp

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 

Repository files navigation

How to use

In your docker file write FROM iwitaly/nlp:cpu for CPU version or FROM iwitaly/nlp:gpu for GPU version.

Find the image at Docker hub https://hub.docker.com/r/iwitaly/nlp/

Whats inside

Modern NLP frameworks including deep learning:

  • torch==0.4.1
  • flair==0.3.2
  • spacy==2.0.11
  • dateparser==0.7.0
  • pymorphy2==0.8
  • yargy==0.11.0
  • natasha==0.10.0
  • nltk==3.2.1
  • yake==0.3.7

How to build

CPU

docker build -t nlp-cpu -f ./Dockerfile.cpu .

GPU

docker build -t nlp-gpu -f ./Dockerfile.gpu .

Please keep in mind that this version works with CUDA 10. Choose another base image if your GPU does not support it.

You can find NVIDIA images here https://hub.docker.com/r/nvidia/cuda/

How to use

docker run -it --runtime=nvidia iwitaly/nlp:gpu nvidia-smi

You can also pass CUDA_VISIBLE_DEVICES environment variable.

About

Docker images for production NLP usage including deep learning

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published