What's Changed Support PEP 604 unions (int | str) as output types by @sohumt123 in #1932 fix: reject negative counts in exactly/at_least/at_most/between by @chuenchen309 in #1922 fix: raise clearly instead of asserting on a rejected XGrammar token by @chuenchen309 in #1923 fix(models): handle nullable object type (list form) in additionalProperties pass by @ErenAta16 in #1918 fix: reject leading-z…
What's Changed fix(types): map JSON Schema const to a Literal by @sarathfrancis90 in #1894 fix: reject invalid day '00' in the date type regex by @hoangsonww in #1897 fix(exceptions): don't require a provider's optional SDK to normalize errors by @Sanjays2402 in #1898 fix(vllm): send structured_outputs instead of removed guided_* keys by @pablopupo in #1888 Escape regex metacharacters in numeric l…
What's Changed Add semver and MAC address custom types by @Yeuvoir in #1863 Update docs with OrcaRouter guide by @xilema2 in #1865 fix(types): handle bool instances in python_types_to_terms for Literal[True/False] by @devteamaegis in #1869 Add ipv6 custom type by @u7k4rs6 in #1867 Fix unformatted f-string in List type error message by @tan7vir in #1870 Docs: remove stale content by @plpxsk in #187…
What's Changed Fix error message for model type determination by @Reslan-Tinawi in #1862 Implement uniform exceptions classes across model providers by @plpxsk in #1823 New Contributors @Reslan-Tinawi made their first contribution in #1862 @plpxsk made their first contribution in #1823 Full Changelog: 1.2.13...1.3.0 Minor Breaking Change: API-based models (OpenAI, Anthropic...) are now raising sta…
What's Changed Fix XDG_CACHE_HOME double .cache in path construction by @jnMetaCode in #1828 fix(llamacpp): correct EOS attention mask and vocab truncation bugs by @giulio-leone in #1820 fix(dsl): add JSON double-quotes to string literals inside container types by @giulio-leone in #1821 fix: accumulate duplicate token IDs in vocabulary construction by @RoloPineda in #1831 fix: inline SPIECE_UNDERL…
What's Changed Add a link to the audit form in the README and the doc website by @RobinPicard in #1822 Full Changelog: 1.2.11...1.2.12
What's Changed Add documentation preview when opening a PR by @RobinPicard in #1818 Full Changelog: 1.2.10...1.2.11
What's Changed Correct mlx-lm example unpacking operator by @Anri-Lombard in #1786 Remove todos by @Anri-Lombard in #1788 fix: correct chat format usage by @Ki-Seki in #1790 Add AsyncOllama to all exports by @Anri-Lombard in #1791 fix: Best-effort use of chat completion by @Ki-Seki in #1789 Fix bug in JsonSchema resulting in the schema validation not running by @RobinPicard in #1795 Update the is_…
What's Changed fix: Refactor sampling parameters in VLLMOffline to use StructuredOutputsParams by @Ki-Seki in #1779 docs: clarify use of SamplingParams in model response examples by @Ki-Seki in #1777 Update the uv.lock by @RobinPicard in #1783 New Contributors @Ki-Seki made their first contribution in #1779 Full Changelog: 1.2.8...1.2.9
What's Changed Use uv sync in the CI instead of pip install by @RobinPicard in #1767 Fix error in the batch method of the MLXLM model by @RobinPicard in #1772 Full Changelog: 1.2.7...1.2.8
What's Changed Add the device_dtype init parameter to the transformers model by @RobinPicard in #1762 Full Changelog: 1.2.6...1.2.7
What's Changed Fix correct handling of chat multimodal inputs in TransformersMM class by @laitifranz in #1728 Add batch generation to the MLXLM model by @RobinPicard in #1759 Feature/mistral ai integration by @yasteven in #1755 Add json type conversion by @RobinPicard in #1761 Fix installation instructions by @RobinPicard in #1764 Add design partner CTA in docs by @rlouf in #1766 New Contributors…