* Recommend using pre-built Docker image
* Update README.md
Co-authored-by: István Zoltán Szabó <istvan.szabo@elastic.co>
---------
Co-authored-by: István Zoltán Szabó <istvan.szabo@elastic.co>
PyTorch models traced in version 1.13 of PyTorch cannot be evaluated in
version 1.9 or earlier. With this upgrade Eland becomes incompatible with
pre 8.7 Elasticsearch and will refuse to upload a model to the cluster.
In this scenario either upgrade Elasticsearch or use an earlier version of Eland.
Users who follow the Eland README as a guide to importing
models can easily end up seeing inexplicably poor performance
due to unknowingly running the model with one allocation and
one thread per allocation.
This change spells out the effect of `--start` and links to
alternatives that allow better use of available hardware.
Co-authored-by: David Kyle <david.kyle@elastic.co>
Elasticsearch uses v1.11 of PyTorch. Models created with the latest PyTorch
release (v1.12) are not compatible with v1.11. This pins the PyTorch version
to 1.11 to prevent the incompatibility. The version of the Elasticsearch Python
client is now required to be >= Eland.
All users of Eland for importing NLP models should upgrade.
This adds some more definite types for our NLP tasks and tokenization configurations.
This is the first step in allowing users to more easily import their own transformer models via something other than hugging face.
The Machine Learning section now has two sub-sections — one for
traditional regression/classification and the other for NLP with
PyTorch. The examples show two ways to upload models from the Hugging
Face model hub.
* Significant refactor of task list in operations.py
Classes based on composite pattern replace tuples for
tasks.
* Addressing review comments for eland/operations.py
* Minor update to review fixes
* Minor fix for some better handling of non-aggregatable fields: https://github.com/elastic/eland/issues/71
* Test for non-aggrgatable value_counts
* Refactoring tasks/actions
* Removing debug and fixing doctest
