Bart Broere
33cf029efe
Implement eland.DataFrame.to_json ( #661 )
...
Co-authored-by: Quentin Pradet <quentin.pradet@elastic.co>
2024-02-15 11:32:54 +04:00
Aurélien FOUCRET
9d492b03aa
Release 8.12.1
...
Co-authored-by: Quentin Pradet <quentin.pradet@elastic.co>
2024-02-01 10:50:18 +04:00
Quentin Pradet
fd2ceab846
Run Buildkite docs jobs in pull requests from forks ( #652 )
2024-01-31 20:55:19 +04:00
Quentin Pradet
02190e74e7
Switch to 2024 black style ( #657 )
2024-01-31 14:47:19 +04:00
Aurélien FOUCRET
2a6a4b1f06
Fix missing value support for XGBRanker. ( #654 )
...
* Fix missing value support for XGBRanker.
* lint
* Sort expected scores
* lint
2024-01-23 18:42:24 +01:00
Quentin Pradet
1190364abb
Release 8.12.0
2024-01-19 12:42:45 +04:00
David Kyle
64216d44fb
Add prefix_string config option to the import model hub script ( #642 )
2024-01-19 12:06:57 +04:00
Liam Thompson
0a6e3db157
[DOCS] Make online retail notebook runnable in Colab ( #641 )
...
* Make online retail notebook runnable in Colab
* Fix broken query
2024-01-18 15:55:20 +04:00
Aurélien FOUCRET
5169cc926a
Improve LTR ( #651 )
...
* 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>
2024-01-17 13:01:47 +04:00
Aurélien FOUCRET
d2291889f8
Fix typo ( #650 )
2024-01-12 09:34:09 -05:00
Aurélien FOUCRET
d3ed669a5e
LTR feature logger ( #648 )
2024-01-12 13:52:04 +01:00
Adam Demjen
926f0b9b5c
Add XGBRanker and transformer ( #649 )
...
* 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
2024-01-11 15:48:13 -05:00
Adam Demjen
840871f9d9
Accept LTR inference config when creating model ( #645 )
...
* 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
2024-01-08 09:19:03 -05:00
Aurélien FOUCRET
05c5859b8a
Adding a new movie dataset to the tests. ( #646 )
2024-01-04 16:14:56 +01:00
Aurélien FOUCRET
0f91224daf
Add 8.12 to CI and remove 8.10 ( #647 )
2024-01-04 10:06:19 -05:00
Bart Broere
927acc86ad
Small cosmetic fix to the docs ( #640 )
2023-11-30 08:34:59 +01:00
David Kyle
6ef418f465
Release 8.11.1
2023-11-22 11:55:53 +01:00
David Kyle
081250cdec
Fix failed import of ST RoBERTa models ( #637 )
...
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
2023-11-21 12:53:43 +00:00
Quentin Pradet
af26897313
Bumpy numpy and shap ( #636 )
2023-11-21 13:17:53 +01:00
David Kyle
add61a69ec
Update CI machine types to N2 ( #634 )
...
Use `n2-standard-2` for lint and doc builds
Use `n2-standard-4` for tests
2023-11-21 11:33:04 +00:00
David Kyle
b689759278
Skip model config tests ( #635 )
...
For #633
2023-11-21 11:07:55 +00:00
Liam Thompson
87d18bd850
Fix colab link ( #632 )
...
Co-authored-by: Quentin Pradet <quentin.pradet@elastic.co>
2023-11-16 10:24:06 +00:00
Quentin Pradet
dfc522eb31
Allow es-doc members to trigger CI ( #631 )
2023-11-13 11:55:39 +01:00
Liam Thompson
508de981ff
Make demo notebook runnable in Colab ( #630 )
...
* Make demo notebook runnable in Colab
* Index using IDs starting from 0
* Trivial change to trigger CI
2023-11-10 08:44:19 +01:00
Quentin Pradet
41db37246f
Release 8.11.0
2023-11-08 11:51:14 +01:00
Valeriy Khakhutskyy
6cecb454e3
[ML] Better memory estimation for NLP models ( #568 )
...
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 ).
2023-11-06 12:18:20 +01:00
Bart Broere
28e6d92430
Stream writes in to_csv()
...
Co-authored-by: P. Sai Vinay <pvinay1998@gmail.com>
2023-11-06 11:39:31 +01:00
Quentin Pradet
adf0535608
Fix docs build
...
Some dependencies like numpy are pinned to versions that do not support
Python 3.12. Python 3.10 is the latest version supported by Eland.
2023-11-06 13:25:30 +04:00
Bart Broere
5e5f36bdf8
Deal with the mad aggregation being removed in Pandas 2 ( #602 )
2023-11-06 06:12:16 +01:00
David Kyle
5b3a83e7f2
[NLP] Support E5 small multi-lingual ( #625 )
...
Although E5 small is a BERT based model it takes 2 parameters to forward
not 4. Use the tokenizer type to decide the number of parameters
2023-10-31 17:49:43 +00:00
David Kyle
ab6e44f430
[NLP] Tests for NLP model configurations ( #623 )
...
Add tests for generated Elasticsearch model configurations
2023-10-19 12:39:57 +01:00
Quentin Pradet
0c0a8ab19f
Bump tested stack versions ( #621 )
2023-10-11 19:48:47 +02:00
Bart Broere
36b941e336
Use _append instead of append since it's still available after 2.0 of pandas ( #603 )
2023-10-11 15:41:05 +01:00
Quentin Pradet
6a4fd511cc
Release 8.10.1 ( #620 )
2023-10-11 12:56:24 +02:00
Quentin Pradet
c6ce4b2c46
Fix direct usage of TransformerModel ( #619 )
2023-10-11 11:56:14 +02:00
Bart Broere
48e290a927
Prepare for deprecation of is_datetime_or_timedelta_dtype in Pandas 2.0 ( #592 )
2023-10-10 19:37:13 +01:00
Quentin Pradet
bb0c111a68
Release Eland 8.10.0
2023-10-09 11:49:12 +02:00
Quentin Pradet
9273636026
Reduce Docker image size and support arm64 ( #615 )
...
Co-authored-by: David Olaru <dolaru@elastic.co>
* Reduce Docker image size from 4.8GB to 2.2GB
* Use torch+cpu variant if target platform is linux/amd64
Avoids downloading large & unnecessary NVIDIA deps defined in the package on PyPI
* Build linux/arm64 image using buildx and QEMU
2023-10-05 18:43:52 +04:00
Quentin Pradet
b8a7b60c03
Stop mentioning Python 3.7 and Pandas 1.13 are supported ( #612 )
2023-10-04 10:56:51 +02:00
Quentin Pradet
3be610b6fc
Recommend using pre-built Docker image ( #614 )
...
* 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>
2023-10-03 19:40:24 +02:00
Quentin Pradet
352e31ed14
Add Buildkite pipeline to push Docker image ( #613 )
...
* Add Buildkite pipeline to push Docker image
* Fix lint
* Fix Read the Docs build
* Replace distutils with packaging
2023-10-03 14:39:54 +02:00
Quentin Pradet
9d7c042bdb
Bump transformers to fix private model support ( #611 )
2023-09-26 14:54:23 +02:00
Enrico Zimuel
235c490e0c
Updated bullseye docker image ( #610 )
2023-09-26 09:53:24 +02:00
Bart Broere
3908f43905
Remove deprecated check_less_precise ( #596 )
2023-09-26 07:34:52 +02:00
Quentin Pradet
566bb9e990
Allow importing private HuggingFace models ( #608 )
2023-09-25 15:10:58 +02:00
Quentin Pradet
5ec760635b
Recommend installing Eland in a virtual environment ( #606 )
2023-09-22 13:14:05 +02:00
Jonathan Buttner
a8b76c390f
Setting chunk size to 1mb ( #605 )
2023-09-20 11:40:11 -04:00
Bart Broere
12200039f5
Fix iteritems deprecation ( #593 )
2023-09-19 12:00:32 +02:00
David Kyle
301cda8d69
Error measuring embedding size for some DPR models ( #573 )
...
Fixes an error unpacking a tuple that contains a single element.
2023-09-19 10:44:15 +01:00
Bart Broere
5c5ef63a69
Use the workaround if we can't determine the server's version ( #581 )
2023-09-15 15:29:36 +04:00
