CHANGELOG.md

v0.6.0 - 2026-05-19

What's Changed

New Features

• feat: adding typing to melleatools by @akihikokuroda in #959
• feat: add tool calling support to m serve by @markstur in #850
• feat: cli OpenAI-compatible API response_format support by @markstur in #884
• feat: allow async functions as tools by @ajbozarth in #1041
• feat: add as_generic_chat_history function to convert any Context to … by @akihikokuroda in #1007
• feat: add MCP tool integration by @ajbozarth in #1042
• feat(telemetry): close five OTel GenAI semantic convention emission gaps (#1035) by @planetf1 in #1036
• feat: add bash tool by @akihikokuroda in #1056
• refactor: get instructions from upstream guardian adapters by @psschwei in #1037
• feat(stdlib): add stream_with_chunking() with per-chunk validation (#901) by @planetf1 in #942
• feat(stdlib): add streaming event types, events() iterator, and OTEL bridge (#902) by @planetf1 in #1095

Bug Fixes

• fix: tool call arguments by @akihikokuroda in #896
• fix: Remove tests and examples for deprecated models by @frreiss in #1012
• fix: remove unused keys from Message template args by @ajbozarth in #1010
• fix: silence false-positive context warning in react framework by @ajbozarth in #1009
• fix: populate context_view in COMPONENT_PRE_EXECUTE payload by @araujof in #941
• fix: skip tool hooks for framework-internal tools by @araujof in #939
• fix: drop unused docs dependency group by @ajbozarth in #1044
• fix: log unexpected errors server-side, return generic message to client (#991) by @SAY-5 in #1057
• fix: make FunctionParameters and LogitBias OpenAI-compatible by @markstur in #1039
• fix(backends): capture vLLM reasoning field in mot._thinking by @planetf1 in #1063
• fix(test): loosen fragile equality assertions in astream incremental tests by @planetf1 in #1069
• fix: update docs and comment references to python 3.11 and remove rust mention by @jakelorocco in #1089
• docs(openai): document empty response from thinking-mode models by @planetf1 in #1062
• fix: revert feat: add bash tool by @akihikokuroda in #1098

Documentation

• docs: add canonical url headers by @AngeloDanducci in #961
• docs: add release and testing notes by @jakelorocco in #1031

Other Changes

• ci: more flexible pr template handling by @psschwei in #1082
• ci: better pr type checker experience by @psschwei in #1083
• chore: Add Ollama model name mappings for Granite 4.1 to intrinsics adapter resolution by @kndtran in #1085
• ci: migrate hold and pr-label actions to common org versions by @psschwei in #1081

New Contributors

• @SAY-5 made their first contribution in #1057
• @kndtran made their first contribution in #1085

Full Changelog: https://github.com/generative-computing/mellea/compare/v0.5.0...v0.6.0

v0.5.0 - 2026-05-05

What's Changed

New Features

• feat(telemetry): latency histograms for LLM request duration and TTFB (#463) by @ajbozarth in #782
• feat: rename generative slots -> generative stubs by @jakelorocco in #801
• feat: (m-decompose) Module Prompt V3 by @csbobby in #770
• feat: simplify plugin tests; fix plugin resetting by @jakelorocco in #819
• feat: add examples and tooling tests to run_tests_with_ollama_and_vllm by @jakelorocco in #821
• feat: add return types to invoke_hook by @jakelorocco in #707
• feat: separate out remaining dependencies and improve tests by @jakelorocco in #789
• feat: add error counter metrics categorized by semantic type (#465) by @ajbozarth in #856
• refactor: improve fancylogger implementation by @AngeloDanducci in #792
• refactor: add otel tracing filter to logging by @AngeloDanducci in #859
• feat: streaming support in m serve OpenAI API server by @markstur in #823
• feat: first pass at carrying contextvars though async flows by @AngeloDanducci in #878
• refactor: add print statements to show code flow in mify example by @code4days in #870
• feat: add pricing registry and cost metrics (#464) by @ajbozarth in #882
• feat: add operational counters for sampling, requirements, and tools (#467) by @ajbozarth in #883
• feat: add --skip-resource-checks flag to bypass hardware capability g… by @ajbozarth in #889
• refactor!: partition ModelOutputThunk execution metadata into Generat… by @ajbozarth in #908
• feat: add additional logging handlers by @AngeloDanducci in #907
• feat(core): add PartialValidationResult with tri-state semantics by @planetf1 in #924
• feat(stdlib): add ChunkingStrategy ABC and built-in chunkers by @planetf1 in #923
• feat: add prompt cache token support to cost telemetry by @ajbozarth in #936
• feat: add stream_validate() hook to Requirement (#900) by @planetf1 in #925
• feat(examples): add extra_requirements param to IVR qiskit validation by @ajbozarth in #955
• feat: add embedded adapters (granite switch) to openai backend by @jakelorocco in #881
• refactor(telemetry): replace builtin_pricing.json with litellm pricing API by @ajbozarth in #956
• feat: simplify intrinsics (code and examples) by @jakelorocco in #946
• feat: granite4.1 by @avinash2692 in #964
• feat: allow name field in intrinsics io.yaml by @ink-pad in #980
• feat: handle message docs correctly by @jakelorocco in #975
• feat: update granite library examples to use Granite 4.1 3B adapters. by @nrfulton in #981

Bug Fixes

• fix: restore example collection during directory traversal (#794) by @planetf1 in #795
• fix: redirect /how-to/safety-guardrails to existing security page (#788) by @planetf1 in #803
• fix(cli): handle sync/async serve functions in m serve by @markstur in #784
• fix: evict Ollama models between test modules to prevent memory starvation by @planetf1 in #804
• fix: sofai graph coloring example — broken model and incorrect problem #806 by @planetf1 in #807
• fix: flush MPS cache in alora test GPU cleanup (#790) by @planetf1 in #800
• fix(test): widen hallucination detection tolerance (#809) by @planetf1 in #810
• fix: reload module for telemetry testing so all tests can run by @jakelorocco in #805
• fix: handle stale .vllm-venv in test runner by @planetf1 in #829
• fix: remove all mentions to RITS by @guicho271828 in #868
• fix: granite33 response_end span uses sentence length not full respon… by @planetf1 in #845
• fix: run zizmor checker for github actions to ensure security by @jakelorocco in #854
• fix: render Click \b verbatim blocks in CLI reference docs (#866) by @planetf1 in #867
• fix: fixes invalid workflow file by @markstur in #877
• fix: granite33 citation spans wrong for duplicate sentences (#851) by @planetf1 in #872
• fix: fixing test bugs with xfail by @avinash2692 in #886
• fix: handle nested JSON in parse_judge_output via raw_decode by @sjoerdvink99 in #875
• fix: disable OCR in RichDocument CI test to avoid modelscope.cn download by @ajbozarth in #888
• fix: update hallucination_detection fixture for upstream NA enum addition by @ajbozarth in #918
• fix: remove wall time checks from tracing_backend tests by @jakelorocco in #927
• fix: add missing nav and fix cli ref by @AngeloDanducci in #922
• fix: add vllm pytest marker back by @jakelorocco in #933
• fix: raise ValueError on duplicate subtask tags in reorder_subtasks by @sjoerdvink99 in #874
• fix: replace asyncio.sleep FAF guards with deterministic awaits by @ajbozarth in #919
• fix: removing ollama hardcoding in examples, guardian, and test by @avinash2692 in #912
• fix: pin uncertainty and context-attribution revisions and update uncertai… by @AngeloDanducci in #970
• fix: swap python decompose example model by @AngeloDanducci in #968
• fix: model options with intrinsics by @jakelorocco in #972
• fix: add guardian intrinsic document by @subhajitchaudhury in #966
• fix: key in json object returned by policy_guardrails intrinsic by @monindersingh in #979
• fix: default intrinsic adapter types by @jakelorocco in #994
• fix: issues introduced by intrinsic changes by @jakelorocco in #986
• fix: update model ids and documentation links for switch by @jakelorocco in #997
• fix: move test_huggingface.py to granite4.1; and small rag intrinsic … by @jakelorocco in #1008
• fix: prevent major releases by @jakelorocco in #1016

Documentation

• docs: add redirects for former pages by @psschwei in #846
• docs: add CLI reference page and remove CLI from API docs (#704) by @planetf1 in #852
• docs: add AI attribution policy by @ajbozarth in #848
• docs: consolidate how-to section by @psschwei in #893
• docs: add generation_error hook to plugins page, remove stale plan doc by @ajbozarth in #887
• docs: fix 'convienance' -> 'convenience' (5 occurrences) by @MukundaKatta in #894
• docs: move glossary to reference section by @psschwei in #892
• docs: document two session creation patterns by @akihikokuroda in #906
• docs: add backend selection lookup table by @akihikokuroda in #905
• docs: restructure sidebar — split Observability from Evaluation, move LLM-as-a-Judge to How-To by @ajbozarth in #895
• docs: add metadata to code block by @akihikokuroda in #917
• docs: test based eval documentation by @seirasto in #916
• docs: fix link to CONTRIBUTING guide by @seirasto in #960
• docs: add expected output blocks and update quickstart examples by @AngeloDanducci in #957
• docs: add architecture diagram for intrinsics by @jakelorocco in #998

Other Changes

• chore: update governance by @psschwei in #799
• test: add unit tests for stdlib/requirements (#814) by @planetf1 in #820
• test: add tool_arg_validator edge case test, fix typo (#826) by @planetf1 in #831
• test: add unit tests for helpers (#815) by @planetf1 in #847
• test: add unit tests for granite formatters (#812) by @planetf1 in #818
• test: unit tests for backend pure logic (cache, catalog, bedrock) by @planetf1 in #832
• chore: add info for working with intrinsics to AGENTS.md by @psschwei in #768
• test: add unit and integration tests for stdlib components (#817) by @planetf1 in #830
• test: unit tests for CLI decompose and eval pure-logic helpers (#861) by @planetf1 in #863
• test: pure-logic unit tests for stdlib, core, backends, telemetry (#860) by @planetf1 in #862
• ci: add actionlint to validate workflow files on PRs by @planetf1 in #880
• chore: Update expected test outputs to reflect upstream config changes by @frreiss in #897
• chore: removing some comments by @avinash2692 in #978
• test: add tests for new intrinsic field name by @jakelorocco in #988
• release: bump minor version by @jakelorocco in #977
• ci: add action for holding PRs (preventing merge) by @psschwei in #1014

New Contributors

• @sjoerdvink99 made their first contribution in #875
• @MukundaKatta made their first contribution in #894
• @seirasto made their first contribution in #916
• @subhajitchaudhury made their first contribution in #966
• @monindersingh made their first contribution in #979

Full Changelog: https://github.com/generative-computing/mellea/compare/v0.4.2...v0.5.0

v0.4.2 - 2026-04-08

What's Changed

New Features

• feat: add tests for mellea optional dependencies by @jakelorocco in #724
• feat: further vram optimizations by @avinash2692 in #765
• feat: (m decomp) M Decompose Readme and Docstring Updates by @csbobby in #767
• feat: add top level async streaming by @jakelorocco in #655
• feat(serve): improve OpenAI API compatibility with usage, finish_reas… by @markstur in #771
• feat: removing vllm backend by @avinash2692 in #781

Bug Fixes

• fix: modifications to granite formatter tests by @jakelorocco in #703
• fix: exclude tooling from mypy check by @planetf1 in #748
• fix: setting ollama host in conftest by @avinash2692 in #751
• fix: Add qualitative and slow markers so the example is skipped by @markstur in #764
• fix(tools): correct args validation in langchain tool wrapper by @markstur in #761
• fix: remove references to old pytest markers by @jakelorocco in #776
• fix: add error handling to OpenAI-compatible serve endpoint by @markstur in #774
• fix: assertion for test_find_context_attributions and range for hallucination detection by @jakelorocco in #779
• fix: add xfail to citation test; functionality is tested elsewhere by @jakelorocco in #787

Documentation

• docs: remove discord link in main readme by @AngeloDanducci in #720
• docs: note virtual environment requirement for pre-commit hooks by @ajbozarth in #745
• docs: condense README to elevator pitch (#478) by @planetf1 in #688
• docs: update qiskit_code_validation example defaults by @ajbozarth in #743
• docs: remove pre-IVR validation and update readme with v2 benchmark results by @ajbozarth in #769

Other Changes

• docs: add multi-turn strategy option to Qiskit code validation example by @vabarbosa in #717
• chore: use github tooling to build release notes by @psschwei in #710
• docs: add release.md by @psschwei in #723
• fix: proper permissions on pr labeling job by @psschwei in #741
• ci: memory management in tests by @avinash2692 in #721
• chore: enforce commit formatting on PR titles by @psschwei in #750
• chore: Update HF repo names by @frreiss in #753
• ci: drop mergify, add release entry to pr-labels action by @psschwei in #752
• ci: fix to make pr label job required check by @psschwei in #756
• test: agent skills infrastructure and marker taxonomy audit (#727, #728) by @planetf1 in #742
• chore: add governance doc by @psschwei in #786
• chore: updating governance doc to use maintainers by @psschwei in #791

New Contributors

• @markstur made their first contribution in #764

Full Changelog: https://github.com/generative-computing/mellea/compare/v0.4.1...v0.4.2

v0.4.1 - 2026-03-23

Feature

• Move ruff hooks locally; add output for ci/cd autofixes; update (#709) (f0e778e)
• m-decomp: Upgraded pipeline and added README, examples, and fixed module issues (#676) (cf63d92)

Fix

• Add missing dependencies (#715) (4bb16c8)
• Add special handling for mellea global event loop when forked (#624) (a620440)
• Update github action versions to Node24 compatible (#713) (4c0bb1b)
• Increase test timeout and remove unnecessary hook debugging (#706) (871a4bf)

Documentation

• Add documentation listing required models for non qualitative tests to run (#674) (7501093)

v0.4.0 - 2026-03-18

Feature

• Guardianlib intrinsics (#8) (#678) (224d14f)
• Add find_context_attributions() intrinsic function (#679) (7eaf9b7)
• Add codeowners for the granite-common part of mellea intrinsics (#669) (a4ec484)
• UQ & requirement_check as core Intrinsic (#551) (3e47d15)
• Add OTLP logging export (#635) (c4cb59f)
• telemetry: Add configurable metrics exporters (OTLP and Prometheus) (#610) (5ec3c7a)
• Hook system and plugin support for Mellea (#582) (cbd63bd)
• Add token usage metrics with OpenTelemetry integration (#563) (0e71558)
• Move functionality of granite-common to mellea (#571) (6901c93)
• Add OpenTelemetry metrics support (#553) (78c5aab)

Fix

• Always populate mot.usage in HuggingFace backend (#694) (#697) (4d3fc1b)
• Add opencv-python-headless to docling extras (#682) (#685) (80000af)
• Skip pytest collection of qiskit validation_helpers module (#683) (#686) (ab56c85)
• Remove answer_relevance* intrinsics; fix other intrinsics issues (#690) (1734900)
• Use tuple instead of generator for DropDuplicates dictionary key (#652) (f7ad489)
• Document.parts() returns [] instead of raising NotImplementedError (#637) (3888476)
• Add missing type annotations to public API functions (#619) (97b2ceb)
• Update MultiTurnStrategy to include validation failure reasons in repair messages (#633) (ebdd092)
• Restore VSCode test discovery and make GPU isolation opt-in (#605) (21746b1)
• Hf metrics tests run out of memory (#623) (5411760)
• Guarding optional imports for hooks (#627) (9588284)
• Python decompose model change and pipeline fix (#569) (15d8fff)
• Explicit PYTHONPATH for isolated test subprocesses (#593) (#594) (7bfd18d)
• Use device_map for HF model loading (#581) (#587) (8a385d5)
• Ensure enough tokens for structured output in vLLM test (#591) (#595) (ac6a4cf)
• Prevent example collection crash for readme_generator (#596) (0e56243)
• Include fixes issue in pr template (#602) (a3f3f71)
• Do not post_process before finally in ModelOutputThunk.astream (#580) (af25037)
• Correct type annotations and improve CI cache invalidation (#579) (dfc8942)
• Issues with tests (alora example, rag intrinsics, mistral tool use, vllm auto-skip) (#570) (4cc75c8)

Documentation

• Refactor telemetry docs into dedicated tracing, metrics, and logging pages (#662) (56e7ff9)
• Add missing example categories to examples catalogue (#645) (#672) (a86fe40)
• Fix MelleaPlugin/MelleaBasePayload missing from API coverage (#… (#670) (17d48d7)
• Removed outdated tutorial.md (#555) (a0e2a46)
• Pre-release verification (resync with latest docs, fix discrepancies) (#665) (e1f34cd)
• Fix RST double-backtick notation breaking API cross-reference links (#658) (98c0e22)
• Add plugins page to nav, apply standards, trim design doc (#663) (3c0cfa4)
• Fix missing docstring sections in plugins and telemetry (#654) (#664) (8a84987)
• Improve docstrings for API reference (#612) (#614) (f7294d0)
• Add Qiskit code validation IVR example (#576) (ea8d21e)
• Implement publishing pipeline (#617) (#646) (0c5d9c9)
• Complete developer documentation rewrite (#480) (#601) (ed01c87)
• Docs/api pipeline improvements (#611) (3d6755d)

v0.3.2 - 2026-02-26

Feature

• Add tool decorator (#387) (bfbbe46)

Fix

• Issues found in comprehensive tests: cache capacity, watsonx (#560) (ff00e89)
• Nonhybrid granite model id (#546) (dc94364)
• Huggingface memory leak (#544) (2f74853)
• Self._tokenizer is unset (#549) (5ac4b2f)
• Avoid instantiating an additional tokenizer (#548) (05f0a91)
• Allow mypy to install type stubs (#487) (2bb34d6)
• mellea decomp: Solve ConstraintExtractor parsing fails and improve robustness (#445) (ca3a7f2)

Documentation

• api: Generate API docs from latest PyPI release (#361) (0cf5d37)

v0.3.1 - 2026-02-11

Feature

• Migrate from Granite 3 to Granite 4 hybrid models (#357) (8f9e18c)
• Add MelleaTool.from_smolagents() for smolagents integration (#430) (0471006)
• Add tool calling argument validation (#364) (840a02d)
• Instrument telemetry (#355) (b2e5a52)
• Add query clarification RAG intrinsic support (#391) (d38698a)
• Add mellea react agent (#402) (7884b8d)
• Optimize example test discovery and execution speed (#372) (e9aefaf)
• New MelleaTool class and adoption across mellea (#380) (ffb8b6c)
• Add code coverage tracking with pytest-cov (#353) (b45a4b6)
• Add pytest markers for test categorization (#322) (#326) (0d8d020)

Fix

• Lint/format issues (#536) (781bb6b)
• Tools in examples (#535) (a49bdf8)
• Quick fix to get the role / content from specifically parsed messages (#533) (2f54cc8)
• Migrate from IBM alora to PEFT 0.18.1 native aLoRA (#422) (c6a3e64)
• Flag more tests that require ollama (#420) (b06851f)
• Guarantee proper ordering of decompose subtask dependencies (#407) (f0b1346)
• Astream output (#358) (9cafe05)
• Update ci for merge-queue (#417) (5cf8eee)
• Some examples needed update (#408) (3d5ab56)
• Formatting model_ids for better readability (#386) (318a962)
• Update agents.md to strongly encourage using uv (#388) (8b2e2cf)
• Restrict transformers version to 4.x (#379) (67f8bc0)
• Friendly error messages for optional backend dependencies (#343) (4f7091f)
• Add missing await keywords in async tests (#346) (a7442a6)
• Use repr for helpful debug display in Message/ToolMessage (#339) (f15fadb)
• Add skip to timeout test for python < 3.11 (#333) (2cc3352)
• Don't overwrite user-configured logging levels (#298) (119ea86)

Documentation

• Bedrock example. (#410) (3204b3a)
• Add decompose to tutorial with example (#366) (ef0a964)
• Create contributing doc (#369) (1cacbf9)
• Add security policy (#363) (afbda1d)
• Add code of conduct (#365) (94b21d9)
• Add discord badge to readme (#362) (168ccca)

Performance

• Use module-scoped fixture for RAG tests (#337) (d72ecfd)

v0.3.0 - 2026-01-21

Feature

• SOFAI Sampling Strategy (#311) (cbf3913)
• Reorg of codebase (#310) (cbc456b)
• Add typed components; add typing to model output thunks and sampling results (#300) (2eb689d)

Fix

• Tool calling code sample in tutorial (#313) (a42a487)
• Adds granite-common[transformers] to Mellea's huggingface depedency group. (#330) (87a8166)
• Rename file from test_* (#332) (6512b32)
• Readd init file for mellea/stdlib (#328) (cb156a7)
• ImageBlocks are CBlocks (#323) (8a4c910)
• Additional tests optimization when running on github actions. (#293) (c5398e4)
• Import times by not exporting RichDocument at module level (#321) (565b27f)
• Add logging for start_session details (#299) (6e68f57)
• Typos in READMEs and documentation (#303) (9f6a086)
• Add explicit exports to init.py (#317) (6e7b09b)
• Mify protocol issues (#304) (7013b04)
• Linting error (#302) (c6e3b08)
• Add double quotes around brackets used in pip install (#301) (2d017f1)

Documentation

• Add AGENTS.md to guide AI coding assistants (#320) (a89256a)
• Improve contributor instructions in README. (#314) (2be67c8)

v0.2.4 - 2026-01-08

Fix

• Fix gc in instructions and add exception to generate walk (#295) (5fc7df0)
• Marks span tests as qualitative & removes chat error message. (#294) (5ce6360)

v0.2.3 - 2026-01-07

Feature

• Allow forcing a release through test failures (#292) (14b55a3)
• Lazy Spans and KV Blocks (#249) (b9e4a33)
• Switch to new RAG intrinsics repo (#289) (94c35ad)

Fix

• OpenAI base_url default and reasoning effort model option. (#271) (9733df8)
• Unpin granite_commons version from 0.3.5 (#287) (0b402bd)

v0.2.2 - 2025-12-18

Feature

• Add langchain / message interop example (#257) (9b1f299)
• Add better error messages for incorrect genslot args (#248) (9d875d6)

Fix

• Uv-lock package changes (#261) (cb0623f)
• Lock granite-common version to avoid arg changes (#260) (03716c1)
• Docstrings to have code blocks (#256) (94a7b40)

v0.2.1 - 2025-12-10

Feature

• Test-based Evaluation with LLM-as-a-judge (#225) (0f1f0f8)
• Add a code_interpreter tool (#232) (b03c964)

Fix

• Add simple lock to hf generation to prevent using incorrect weights (#237) (6b2a527)
• Collection of small fixes (#238) (2120112)
• Fix unused litellm import (#246) (633bfd7)
• Minor updates to answer relevance (#245) (bde9b4d)
• Pre-commit file selection (#243) (e70d307)

Documentation

• Fixed copyright in LICENSE (#210) (3087051)

v0.2.0 - 2025-11-19

Feature

• Change backend functions to use async; add generate_from_raw (16b8aea)
• Updates for intrinsics support (#227) (52953a5)
• Add requirements and preconditions to gen slots (#226) (f73d8e2)
• MelleaSession.register for functional interface and MelleaSession.powerup for dynamic mixin (register all methods in a class) (#224) (662cfcc)
• Add secure Python code execution with llm-sandbox support (#217) (9d12458)
• Adds think budget-forcing (#107) (a2e29e6)
• Making generate_from_raw public (#219) (7eae224)
• Conda/Mamba-based installation script (#138) (6aea9dc)
• Adds a vllm backend (#122) (21908e5)
• Add the ability to run examples with pytest (#198) (e30afe6)
• Ollama generate_from_raw uses existing event loop (#204) (36a069f)

Fix

• Vllm format issues (abbde23)
• Some minor fixes (#223) (7fa0891)
• Watsonx self._project_id not getting set (#220) (10f6ffa)
• Decomp subtask regex (#218) (5ac34be)

Documentation

• Adding pii m serve example (#215) (54f13f4)

v0.1.3 - 2025-10-22

Feature

• Decompose cli tool enhancements & new prompt_modules (#170) (b8fc8e1)
• Add async functions (#169) (689e1a9)
• Add Granite Guardian 3.3 8B with updated examples function call validation and repair with reason. (#167) (517e9c5)
• Majority voting sampling strategy (#142) (36eaca4)

Fix

• Fix vllm install script (#185) (abcf622)
• Watsonx and litellm parameter filtering (#187) (793844c)
• Pin trl to version 0.19.1 to avoid deprecation (#202) (9948907)
• Rename format argument in internal methods for better mypiability (#172) (7a6f780)
• Async overhaul; create global event loop; add client cache (#186) (1e236dd)
• Update readme and other places with granite model and tweaks (#184) (519a35a)

v0.1.2 - 2025-10-03

Feature

• Making Granite 4 the default model (#178) (545c1b3)

Fix

• Default sampling strats to None for query, transform, chat (#179) (c8d4601)
• Docstrings (#177) (6126bd9)
• Always call sample when a strategy is provided (#176) (8fece40)

v0.1.1 - 2025-10-01

Fix

• Bump patch version to allow publishing (#175) (cf7a24b)

v0.1.0 - 2025-10-01

Feature

• Add fix to watsonx and note to litellm (#173) (307dbe1)
• New context, new sampling,. (#166) (4ae6d7c)
• Add async and streaming support (#137) (4ee56a9)
• Best-of-N Sampling with Process Reward Models (#118) (b18e03d)

v0.0.6 - 2025-09-18

Feature

• Test update pypi.yml for cd pipeline test (#155) (91003e5)

v0.0.5 - 2025-09-17

Feature

• Enable VLMs (#126) (629cd9b)
• LiteLLM backend (#60) (61d7f0e)
• New logo by Ja Young Lee (#120) (c8837c6)

Fix

• Adding pillow as dependency (#147) (160c6ef)
• Huggingface backend does not properly pad inputs (#145) (a079c77)
• Return to old logo (#132) (f08d2ec)
• Alora version and image printing in messages (#130) (2b3ff55)
• Remove ModelOption.THINKING from automatic mapping because it's explicitly handled in line #417 (which was causing parameter conflicts) (#124) (b5c2a39)

Documentation

• Improved documentation on model_options (#134) (ad10f3b)
• Explain that the tool must be called (#140) (a24a8fb)
• Fix typo on README (#116) (dc610ae)
• Fix README typos and broken links (4d90c81)