Unitxt

Latest version: v1.21.0

Safety actively analyzes 723158 Python packages for vulnerabilities to keep your Python projects secure.

Scan your dependencies

Page 3 of 10

1.15.10

What's Changed
* Fix arenahard bluebench template by perlitz in https://github.com/IBM/unitxt/pull/1405
* Fixed formal types of infer() and also added runtime check by yoavkatz in https://github.com/IBM/unitxt/pull/1406
* not using "score" as metric main_score by lilacheden in https://github.com/IBM/unitxt/pull/1407
* Fix model strings for Llama 3 on Together AI by yifanmai in https://github.com/IBM/unitxt/pull/1411
* Adjust binary llmaj to new engines and add rits support by lilacheden in https://github.com/IBM/unitxt/pull/1408
* Granite Guardian RAG metrics by arielge in https://github.com/IBM/unitxt/pull/1393
* Solved many red lines in 'make docs-server' by dafnapension in https://github.com/IBM/unitxt/pull/1418
* Fix artifact dict assignment bug by elronbandel in https://github.com/IBM/unitxt/pull/1419
* Remove top level imports from guerdian metric (as it adds dependencis to unitxt) by elronbandel in https://github.com/IBM/unitxt/pull/1420
* Make types compatible with python 3.8 by elronbandel in https://github.com/IBM/unitxt/pull/1423
* Benjams/loaders fix separator by BenjSz in https://github.com/IBM/unitxt/pull/1424
* Update version to 1.15.10 by elronbandel in https://github.com/IBM/unitxt/pull/1426


**Full Changelog**: https://github.com/IBM/unitxt/compare/1.15.9...1.15.10

1.15.9

Main changes

* Artifacts in the catalog can now be [links to other artifacts](https://www.unitxt.ai/en/latest/docs/saving_and_loading_from_catalog.html#a-catalog-asset-linking-to-another-catalog-asset) and can also be marked [deprecated](https://www.unitxt.ai/en/latest/docs/saving_and_loading_from_catalog.html#deprecated-asset).

<img width="1159" alt="image" src="https://github.com/user-attachments/assets/b2240516-db5a-4fe0-934b-50e2a7c00b04">

What's Changed
* artifact link by dafnapension in https://github.com/IBM/unitxt/pull/1363
* Add processors also as operators by antonpibm in https://github.com/IBM/unitxt/pull/1397
* added 'add_link_to_catalog' for easily adding artifact_links with/without deprecation msg by dafnapension in https://github.com/IBM/unitxt/pull/1398
* Safety updates by bnayahu in https://github.com/IBM/unitxt/pull/1391
* Reduce error message clutter by yoavkatz in https://github.com/IBM/unitxt/pull/1401
* Update version to 1.15.9 by yoavkatz in https://github.com/IBM/unitxt/pull/1404


**Full Changelog**: https://github.com/IBM/unitxt/compare/1.15.8...1.15.9

1.15.8

Main changes

Added support for RITS Inference Engine

Inference Engines
* Add inference engines to the catalog by martinscooper in https://github.com/IBM/unitxt/pull/1394
* Add support for OpenAI custom base url and default headers + RITS Inference engine by martinscooper in https://github.com/IBM/unitxt/pull/1385

Assets
* Add vectara's hhem2.1 faithfulness model as a metric by lilacheden in https://github.com/IBM/unitxt/pull/1382

Bug Fixes
* fix template in Arena Hard card and example by OfirArviv in https://github.com/IBM/unitxt/pull/1390



**Full Changelog**: https://github.com/IBM/unitxt/compare/1.15.7...1.15.8

1.15.7

Assets
* add llama-3-405b-instruct wml classification engine by lilacheden in https://github.com/IBM/unitxt/pull/1383

Usability
* Support MerticsList - to store a list of metrics by lilacheden in https://github.com/IBM/unitxt/pull/1379


Bug fixes
* Made sure null augmentor works as expected by yoavkatz in https://github.com/IBM/unitxt/pull/1381
* Fixes and improvements to task based llm as judge by lilacheden in https://github.com/IBM/unitxt/pull/1366
* Fix package dir in settings by yoavkatz in https://github.com/IBM/unitxt/pull/1387

Documentation
* Typos in the rst files by dafnapension in https://github.com/IBM/unitxt/pull/1380
* Chat api blog post by elronbandel in https://github.com/IBM/unitxt/pull/1371

Inference Engine
* Tests and minor changes Changes to GenAI, WML and HF inference engines by pawelknes in https://github.com/IBM/unitxt/pull/1290


**Full Changelog**: https://github.com/IBM/unitxt/compare/1.15.6...1.15.7

1.15.6

Main changes

* Added support for generating output in ChatAPI format (user/assistant turns) and for inference engines to process ChatAPI input.
See details in [blog](https://www.unitxt.ai/en/latest/blog/inference_engines_blog.html).

* Improved catalog browsing experience with cleaner formatting of catalog assets, and clickable hyper links between catalog assets and between catalog assets and code. See for [example](https://www.unitxt.ai/en/latest/catalog/catalog.cards.sst2.html).

New Features

Inference Engines that support ChatApi interface
* Add target_prefix erasing post processor by default by elronbandel in https://github.com/IBM/unitxt/pull/1361
* Add multi api inference engine by elronbandel in https://github.com/IBM/unitxt/pull/1343
* Add chat api format with standard open ai chat format by elronbandel in https://github.com/IBM/unitxt/pull/1314
* Add option selecting huggingface inference engine by elronbandel in https://github.com/IBM/unitxt/pull/1357

Improved multi model support
* Add seed bench dataset and support for videos by elronbandel in https://github.com/IBM/unitxt/pull/1309
* Add LMMSEvalInferenceEngine by elronbandel in https://github.com/IBM/unitxt/pull/1301
* Vision robustness blog by elronbandel in https://github.com/IBM/unitxt/pull/1318

New Asserts
* added QTSUMM taskcard for query-focused table summarization task by csrajmohan in https://github.com/IBM/unitxt/pull/1304
* Add OptionSelectingByLogProbsInferenceEngine by martinscooper in https://github.com/IBM/unitxt/pull/1317
* Replace 20 newsgroup with a shorter version in bluebench by perlitz in https://github.com/IBM/unitxt/pull/1347
* Bluebench Update by perlitz in https://github.com/IBM/unitxt/pull/1342
* Update Blue Bench description by elronbandel in https://github.com/IBM/unitxt/pull/1354
* Batched multi class classification by yoavkatz in https://github.com/IBM/unitxt/pull/1340
* move rag binary llmaj under rag metrics by lilacheden in https://github.com/IBM/unitxt/pull/1338
* adding generic inference binary+idk judges by Roni-Friedman in https://github.com/IBM/unitxt/pull/1316
* Add table augmentors by elronbandel in https://github.com/IBM/unitxt/pull/1328
* Align augmenters with task and types mechanisms by elronbandel in https://github.com/IBM/unitxt/pull/1356
* add serializers to catalog + new table operators by ShirApp in https://github.com/IBM/unitxt/pull/1365

Performance
* Add loaders cache by elronbandel in https://github.com/IBM/unitxt/pull/1333

Usuability
* Allow turning single stream to dataset by elronbandel in https://github.com/IBM/unitxt/pull/1335

Documentation
* Add ability to load_dataset without a template for simpler usage for beginners by elronbandel in https://github.com/IBM/unitxt/pull/1350
* add score name prefix for judge_raw_output/input in llmaj metric by OfirArviv in https://github.com/IBM/unitxt/pull/1323
* Add link to source in catalog assets by elronbandel in https://github.com/IBM/unitxt/pull/1362
* Fix docs compilation and links from docs to github by elronbandel in https://github.com/IBM/unitxt/pull/1359
* Fix website docs-code links by elronbandel in https://github.com/IBM/unitxt/pull/1360
* Update error checking and documentation of processors by yoavkatz in https://github.com/IBM/unitxt/pull/1325
* Unified catalog terminology by yoavkatz in https://github.com/IBM/unitxt/pull/1355
* Improved documentation formatting by dafnapension in https://github.com/IBM/unitxt/pull/1334
* Fix catalog links by elronbandel in https://github.com/IBM/unitxt/pull/1348
* Print catalog entries as yamls by dafnapension in https://github.com/IBM/unitxt/pull/1351

CI/CD
* a more elaborated message from performace-test-summary, and doc-string of card_profiler by dafnapension in https://github.com/IBM/unitxt/pull/1307
* Make package requirements compatible with requirements.txt like format by elronbandel in https://github.com/IBM/unitxt/pull/1310
* Make inference engine tests run only when inference.py has changed by elronbandel in https://github.com/IBM/unitxt/pull/1311
* Seperate examples tests by elronbandel in https://github.com/IBM/unitxt/pull/1322
* Fix pyproject.toml to be standalone and comply with modern standards by elronbandel in https://github.com/IBM/unitxt/pull/1324
* Fix GitHub Actions concurrence execution by elronbandel in https://github.com/IBM/unitxt/pull/1349
* Make tests faster and clearer by dafnapension in https://github.com/IBM/unitxt/pull/1345

New Contributors
* martinscooper made their first contribution in https://github.com/IBM/unitxt/pull/1317

**Full Changelog**: https://github.com/IBM/unitxt/compare/1.14.1...1.15.6

1.14.1

Important Change: Unitxt is Faster!

To improve Unitxt’s performance, we've made several optimizations:

1. **Operator Acceleration**: Many operators have been sped up by removing unnecessary deep copying in their code, enhancing runtime efficiency.

2. **Caching Hugging Face Datasets**: We added the option to cache Hugging Face datasets in loaders, which can prevent redundant loading operations. To enable this, you can either:
- Set it globally in code:
python
import unitxt

unitxt.settings.disable_hf_datasets_cache = False

- Use the settings context:
python
with settings.context(disable_hf_datasets_cache=False):
your code

- Or set the environment variable:
bash
export UNITXT_DISABLE_HF_DATASETS_CACHE=False


3. **Eager Execution Mode**: Running Unitxt without streaming, which can be faster in certain scenarios. Enable eager execution using the environment variable or directly in code:
python
unitxt.settings.use_eager_execution = True
or
with settings.context(use_eager_execution=True):
your code


4. **Partial Stream Loading**: This feature lets you load only the necessary data instances, avoiding full dataset loads when not required. Here's an example:
python
from unitxt import load_dataset

dataset = load_dataset(
card="cards.doc_vqa.lmms_eval",
template="templates.qa.with_context.title",
format="formats.models.llava_interleave",
loader_limit=300,
streaming=True,
)
print(next(iter(dataset["test"][0]))) Loads only the first instance


**Complete Example**: Combining the optimizations above can lead to near 1000x faster dataset loading:
python
from unitxt import load_dataset, settings

with settings.context(
disable_hf_datasets_cache=False,
use_eager_execution=True,
):
dataset = load_dataset(
card="cards.doc_vqa.lmms_eval",
template="templates.qa.with_context.title",
format="formats.models.llava_interleave",
loader_limit=300,
streaming=True,
)
print(next(iter(dataset["test"][0]))) Loads only the first instance


5. **Execution Speed Tracking**: A GitHub action has been added to monitor Unitxt’s execution speed in new pull requests, helping ensure that optimizations are maintained.

---

Summary

This release is focused on accelerating performance in Unitxt by introducing several key optimizations. Operator efficiency has been enhanced by removing deep copies, making operations faster. Users can now enable dataset caching for Hugging Face datasets to prevent redundant loading, configured directly in code or through environment variables. An optional eager execution mode has been added, bypassing streaming to increase speed in certain scenarios. Additionally, partial stream loading allows selective instance loading, reducing memory usage and improving response times. To maintain these improvements, a new GitHub action now monitors Unitxt’s execution speed in pull requests, ensuring consistent performance across updates.

All Changes
* Enhancements to inference engines by lilacheden in https://github.com/IBM/unitxt/pull/1243
* add post processor to convert log probs dictionary to probabilities of a specific class by lilacheden in https://github.com/IBM/unitxt/pull/1247
* CI for metrics other than main + Bugfix in RetrievalAtK by lilacheden in https://github.com/IBM/unitxt/pull/1246
* Add huggingface cache disabling option to unitxt settings by elronbandel in https://github.com/IBM/unitxt/pull/1250
* Make F1Strings faster by elronbandel in https://github.com/IBM/unitxt/pull/1248
* Fix duplicate column deletion bug in pandas serializer by elronbandel in https://github.com/IBM/unitxt/pull/1249
* revived no_deep just to compare performance by dafnapension in https://github.com/IBM/unitxt/pull/1254
* fixed scigen post-processor by csrajmohan in https://github.com/IBM/unitxt/pull/1253
* Add prediction length metric by perlitz in https://github.com/IBM/unitxt/pull/1252
* Fix faithfulness confidence intervals by matanor in https://github.com/IBM/unitxt/pull/1257
* Allow role names to be captialized in SerializeOpenAiFormatDialog by yoavkatz in https://github.com/IBM/unitxt/pull/1259
* Accelerate image example 1000X by elronbandel in https://github.com/IBM/unitxt/pull/1258
* Fix the empty few-shot target issue when using produce() by marukaz in https://github.com/IBM/unitxt/pull/1266
* fix postprocessors in turl_col_type taskcard by csrajmohan in https://github.com/IBM/unitxt/pull/1261
* Fix answer correctness confidence intervals by matanor in https://github.com/IBM/unitxt/pull/1256
* add BlueBench as a benchmark to the catalog by shachardon in https://github.com/IBM/unitxt/pull/1262
* Fix MultipleSourceLoader documentation by marukaz in https://github.com/IBM/unitxt/pull/1270
* Ignore unitxt-venv by marukaz in https://github.com/IBM/unitxt/pull/1269
* Add mmmu by elronbandel in https://github.com/IBM/unitxt/pull/1271
* A fix for a bug in metric pipeline by elronbandel in https://github.com/IBM/unitxt/pull/1268
* Added Tablebench taskcard by csrajmohan in https://github.com/IBM/unitxt/pull/1273
* Fix missing deep copy in MapInstanceValues by yoavkatz in https://github.com/IBM/unitxt/pull/1267
* Add stream name to generation of dataset by elronbandel in https://github.com/IBM/unitxt/pull/1276
* Fix demos pool inference by elronbandel in https://github.com/IBM/unitxt/pull/1278
* Fix quality github action by elronbandel in https://github.com/IBM/unitxt/pull/1281
* add operators for robustness check on tables by csrajmohan in https://github.com/IBM/unitxt/pull/1279
* Instruction in SystemFormet demo support. by piotrhelm in https://github.com/IBM/unitxt/pull/1274
* change the max_test_instances of bluebench.recipe.attaq_500 to 100 by shachardon in https://github.com/IBM/unitxt/pull/1285
* Add documentation for types and serializers by elronbandel in https://github.com/IBM/unitxt/pull/1286
* Add example for image processing with different templates by elronbandel in https://github.com/IBM/unitxt/pull/1280
* Integrate metrics team LLMaJ with current unitxt implemantation by lilacheden in https://github.com/IBM/unitxt/pull/1205
* performance profiler with visualization by dafnapension in https://github.com/IBM/unitxt/pull/1255
* Remove split arg to support old hf datasets versions by elronbandel in https://github.com/IBM/unitxt/pull/1288
* add post-processors for tablebench taskcard by csrajmohan in https://github.com/IBM/unitxt/pull/1289
* recursive copy seems safer here by dafnapension in https://github.com/IBM/unitxt/pull/1295
* Fix performance tracking action by elronbandel in https://github.com/IBM/unitxt/pull/1296
* try num of instances in nested global scores by dafnapension in https://github.com/IBM/unitxt/pull/1282
* Update version to 1.14.0 by elronbandel in https://github.com/IBM/unitxt/pull/1298
* expand performance table by dafnapension in https://github.com/IBM/unitxt/pull/1299
* Fix doc_vqa lmms_eval by elronbandel in https://github.com/IBM/unitxt/pull/1300
* prepare for int-ish group names and type names and add the exposing card by dafnapension in https://github.com/IBM/unitxt/pull/1303
* remove groups breakdowns from global score of grouped instance metrics by dafnapension in https://github.com/IBM/unitxt/pull/1306
* Update the safety metric batch size to 10 by perlitz in https://github.com/IBM/unitxt/pull/1305

New Contributors
* piotrhelm made their first contribution in https://github.com/IBM/unitxt/pull/1274

**Full Changelog**: https://github.com/IBM/unitxt/compare/1.13.1...1.14.1

Page 3 of 10

© 2025 Safety CLI Cybersecurity Inc. All Rights Reserved.