Upgrades the PyTorch, transformers and sentence transformer requirements.
Elasticsearch has upgraded to PyTorch to 2.3.1 in 8.16 and 8.15.2. For
compatibility reasons Eland will refuse to upload to an Elasticsearch cluster
that has is using an earlier version of PyTorch.
* Ensure the feature logger is using NaN for non matching query feature extractors (consistent with ES).
* Default score is None instead of 0.
* LTR model import API improvements.
* Fix feature logger tests.
* Fix export in eland.ml.ltr
* Apply suggestions from code review
Co-authored-by: Adam Demjen <demjened@gmail.com>
* Fix supported models for LTR
---------
Co-authored-by: Adam Demjen <demjened@gmail.com>
* Add XGBRanker and transformer
* Map XGBoostRegressorTransformer to XGBRanker
* Add unit tests
* Remove unused import
* Revert addition of type
* Update function comment
* Distinguish objective based on model class
* Support for supplying inference_config
* Fix linting errors
* Add unit test
* Add LTR type, throw exception on predict, refine test
* Add search step to LTR test
* Fix linter errors
* Update rescoring assertion in test + type defs
* Fix linting error
* Remove failing assertion
Fixes an error uploading the sentence-transformers/all-distilroberta-v1 model
which failed with "missing 2 required positional arguments: 'token_type_ids'
and 'position_ids'". The cause was that the tokenizer type was not recognised
due to a typo
This PR adds an ability to estimate per deployment and per allocation memory usage of NLP transformer models. It uses torch.profiler and performs logs the peak memory usage during the inference.
This information is then used in Elasticsearch to provision models with sufficient memory (elastic/elasticsearch#98874).
I updated the tree serialization format for the new scikit learn versions. I also updated the minimum requirement of scikit learn to 1.3 to ensure compatibility.
Fixes#555
The eland_import_hub_model script supports uploading a local file where
the --hub-model-id argument is a file path. If the --es-model-id option is
not used the model Id is generated from the hub model id and when that
is a file path the path must be converted to a valid elasticsearch model id.
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.
Closes#503
Note: I also had to fix the Sphinx version to 5.3.0 since, starting from 6.0, Sphinx suffers from a TypeError bug, which causes a CI failure.
For many model types, we don't need to require the task requested. We can infer the task type based on the model configuration and architecture.
This commit makes the `task-type` parameter optional for the model up load script and adds logic for auto-detecting the task type based on the 🤗 model.
This switches our sklearn.DecisionTreeClassifier serialization logic to account for multi-valued leaves in the tree.
The key difference between our inference and DecisionTreeClassifier, is that we run a softMax over the leaf where sklearn simply normalizes the results.
This means that our "probabilities" returned will be different than sklearn.
This improves the user consumed functions and classes for PyTorch NLP model upload to Elasticsearch.
Previously it was difficult to wrap your own module for uploading to Elasticsearch.
This commit splits some classes out, adds new ones, and adds tests showing how to wrap some simple modules.
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.
