What's Changed

• Update colpali engine to latest version by @wirthual in #563
• Add ColQwen2_5 to IMAGE_COL_MODELS by @yasyf in #568
• add deps by @michaelfeil in #570
• [Docs] Correct outdated mention of OpenAIEmbeddingInput by @HollowMan6 in #587
• Caching huggingface models in GH actions by @wirthual in #602
• Fix client generation pipeline by @wirthual in #603
• Fix broken link in README by @davidpaquin in #622
• adding hf-xet by @michaelfeil in #632
• fix linting in select_model.py by @wirthual in #633
• New torch version (2.8) New flash attnention, remove hf_transfer by @michaelfeil in #636

New Contributors

• @yasyf made their first contribution in #568
• @HollowMan6 made their first contribution in #587
• @davidpaquin made their first contribution in #622

Full Changelog: 0.0.76...0.0.77

• torch=2.6.0 update - 5-10% faster attention on hopper
  -> previously 2.4.1 -> does no longer work with torch.compile + bettertransformers. We recommend disabling torch.compile for this model class.
• flash-attn included in docker image for nvidia.

What's Changed

• bump client version by @wirthual in #522
• add new st version by @michaelfeil in #523
• Version check step by @wirthual in #524
• README: add example for using local model wtth docker container by @wirthual in #528
• add vision client template by @wirthual in #526
• bump to 2.6 torch by @michaelfeil in #556

Full Changelog: 0.0.75...0.0.76

What's Changed

• add better bettertransformers support by @michaelfeil in #509
• Update verbose parameter when initializing ModelWorker by @aRyBernAlTEglOTRO in #516 THANKS!
• 0.0.75 release by @michaelfeil in #520 -> DEFAULTS to v2 command in the CLI

New Contributors

• @aRyBernAlTEglOTRO made their first contribution in #516 THANKS !

Full Changelog: 0.0.74...0.0.75

What's Changed

• move cli to extra file, macOS tests add documentation by @michaelfeil in #501
• Modernbert support, adding latest huggingface from source by @michaelfeil in #508

Full Changelog: 0.0.73...0.0.74

What's Changed

• Support for matryoshka embeddings by @wirthual in #490
• Adding optimum option for PredictEngine by @wwymak in #492
• new-release by @michaelfeil in #494

New Contributors

• @wwymak made their first contribution in #492

Full Changelog: 0.0.72...0.0.73

What's Changed

• Adds neuron support by @michaelfeil in #486
• Fix python classification example by @molntamas in #487
• add max size by @michaelfeil in #489

New Contributors

• @molntamas made their first contribution in #487

Full Changelog: 0.0.71...0.0.72

What's Changed

• update transformers package by @wirthual in #471
• Fix: openvino by @michaelfeil in #475

Full Changelog: 0.0.70...0.0.71

What's Changed

• Queue refactor by @michaelfeil in #462
• Update README.md by @LetsssGo55 in #425
• Update functional.py by @YadlaMani in #419
• Revert "Update functional.py" by @michaelfeil in #466
• Optimum amd support by @michaelfeil in #464
• DROP PYDANTIC 1.x by @michaelfeil in #465
• update release pipeline by @michaelfeil in #469

Full Changelog: 0.0.69...0.0.70

What's Changed

• Readme and Doc update by @michaelfeil in #459
• add openvino, trt by @michaelfeil in #460
• add dtype-based loading by @michaelfeil in #461

Full Changelog: 0.0.68...0.0.69

What's Changed

• add benchmark for clip model by @wirthual in #438
• Vision fix for non-clip models by @michaelfeil in #439
• Add PR template (With changes) by @fadkeabhi in #441
• Fixed Broken License Link in README by @niranjan-kurhade in #446
• change embedding response object to list for openai compatibility by @wirthual in #444
• README: Add KubeAI under integrations by @samos123 in #453
• Model: Device-id and data-parallel inference in CLI and Torch by @michaelfeil in #452
• Colbert by @michaelfeil in #456

New Contributors

• @niranjan-kurhade made their first contribution in #446
• @samos123 made their first contribution in #453

Full Changelog: 0.0.67...0.0.68