v1.1.0

Release v1.1.0 is once again targeted at bringing containers/ai-lab-recipes into alignment with containers/podman-desktop-extension-ai-lab.

Core differences:

• Introduction of grainte-7b-lab as the standard model for (mostly) all of our natural language processing applications
• General improvements to Summarizer and Audio-to-Text recipes
  • Greater variety of supported input file types
• Introduction of an experimental /training directory for building hardware accelerated bootc images to support training workloads
• Increased workflow build coverage
• Integration with @platform-engineering-bot to provide automated dependency bumps

What's Changed

• fix: Convert to Makefile-style and correct wrong usage by @PaulYuuu in #289
• Consistently name llamacpp-python by @rhatdan in #292
• Add support for bootc on audio-to-text recipe by @rhatdan in #287
• chore(deps): update auto merged updates by @platform-engineering-bot in #295
• recipes: update descriptions by @vrothberg in #298
• chore(deps): update all dependencies by @platform-engineering-bot in #296
• Fix document chunking in summarizer by @MichaelClifford in #293
• adding merlinite and granite models by @Gregory-Pereira in #300
• fixing testing framework with image ref being moved by @Gregory-Pereira in #301
• add model workflows by @sallyom in #304
• fix convert_models context by @sallyom in #306
• fix convert_models build with build path permissions (temporary) by @Gregory-Pereira in #307
• temporarily remove model-converter image build by @sallyom in #309
• make model-images only run with cron or manual temporarily by @sallyom in #312
• add mistral-7b-code model and change env by @sallyom in #313
• Show progress bar when downloading models by @rhatdan in #305
• add mistral-7b-code model and change env by @sallyom in #314
• rfe(bootc-image-builder): Support "--chown" option by @PaulYuuu in #299
• fix typo in model image names on push by @sallyom in #317
• after 10 failed attempts, finally updating the model-image workflow the right way by @sallyom in #318
• add PDF reading to summarizer by @MichaelClifford in #302
• Accelerated container image for AMD by @kwozyman in #316
• Fixup download-models by @rhatdan in #303
• Update model download example in models README.md by @MichaelClifford in #319
• remove extra registry in image push by @sallyom in #320
• quay defaults x86_64 arch to amd64 by @Gregory-Pereira in #321
• forgot endif by @Gregory-Pereira in #323
• Image names must be lower case by @rhatdan in #324
• joining the org by @Gregory-Pereira in #325
• Accelerated container image for NVidia by @kwozyman in #328
• Cleanup NVIDIA training. by @rhatdan in #331
• Update whisper model recipe to convert audio files by @hemajv in #326
• Better separation between CentOS and RHEL by @kwozyman in #332
• upping model download timeout and connection values by @Gregory-Pereira in #329
• Add model download utility by @MichaelClifford in #275
• set summarizer write to tempfile by @MichaelClifford in #330
• Update suggested model in chat and summary docs by @MichaelClifford in #333
• Intel Habanalabs bootc image by @kwozyman in #335
• make granite default for models/ and llamacpp/ by @MichaelClifford in #339
• Add pdf option to rag recipe by @MichaelClifford in #342
• Add clean for models, wastes lots of space by @rhatdan in #344
• Add -bootc identifier to training images by @rhatdan in #340
• Add building of instructlab container images by @rhatdan in #341
• Fix typo in instruclab Makefile by @rhatdan in #346
• add instructlab nvidia base image by @sallyom in #345
• Break apart push by @rhatdan in #347
• Get exact token count in summarizer by @MichaelClifford in #348
• try non protected build with path changes by @Gregory-Pereira in #308
• fix push image step double registry by @Gregory-Pereira in #355
• [Model Upgrade] upgrade mistral-7b-instruct-v0.1 to mistral-7b-instruct-v0.2 by @Gregory-Pereira in #352
• improper variable definition by @cooktheryan in #359
• Update the ailab images link in README by @hellohellenmao in #363
• Update all dependencies by @platform-engineering-bot in #362
• Fixes for configuring training by @rhatdan in #358
• only build from dispatch and cron, not paths by @Gregory-Pereira in #383
• add rouge score to summarizer by @MichaelClifford in #376
• Expand documentation for training images by @kwozyman in #369
• Rename training repos to add bootc- indicator by @rhatdan in #368
• minor update for bootc-image-builder target by @PaulYuuu in #367
• Downloading models makefile moved by @lmilbaum in #385
• download models from models dir by @Gregory-Pereira in #389
• refactor instructlab training images to use make targets by @Gregory-Pereira in #387
• swap granite for mistral cleanup + model and recipe housekeeping by @Gregory-Pereira in #350
• update ailab-images with intel-bootc and amd-bootc by @Gregory-Pereira in #392
• version pinning ubuntu runner by @Gregory-Pereira in #388
• Update Auto merged updates by @platform-engineering-bot in #361
• add slack notifications to instructlab wofklows by @Gregory-Pereira in #390
• With latest rocm-smi no need for EPEL in Containerfile by @rhatdan in #386
• model container builder by @kwozyman in #398
• Add missing habanalabs modules by @enriquebelarte in #396
• Add nvidia-instructlab-bootc and supporting materials by @n1hility in #360
• adding instructlab workflow and renaming bootc workflow by @Gregory-Pereira in #400
• Fix name of intel-bootc by @rhatdan in #401
• reworking workflow names for visibility from actions tab by @Gregory-Pereira in #394
• updating instructlab workflow name by @Gregory-Pereira in #403
• Add hl-smi binary from firmware-tools by @enriquebelarte in #402
• Minor text edits to top dir files by @MichaelClifford in #404
• Improve summary chunking by @MichaelClifford in #397
• fix bootc workflow with intel image rename by @Gregory-Pereira in #407
• Remove old fine tuning example by @MichaelClifford in #406
• Pull instructlab images out of local storage by @rhatdan in #409
• create basic instructlab testing framework by @Gregory-Pereira in #411
• Use --compress=false to speed up pushing of instruct-lab images by @rhatdan in #412
• add vllm server container, ability to start the server, ilab generate by @cdoern in #408
• Add containerfile for deepspeed training container by @jaideepr97 in #414
• add TMPDIR documentation by @kwozyman in #417
• Remove beta, rhel9.4 is released by @rhatdan in #418
• Add cloud-init build feature by @rhatdan in #415
• Repair history by @n1hility in #420
• Add growfs service to chatbot/bootc Containerfiles by @sallyom in #410
• Fix Nvidia driver build for RHEL 9.4 by @kwozyman in #421
• Bootc testing framework run playbook and install bootc binary by @Gregory-Pereira in #413
• Timeout of 300 seconds too low by @ericcurtin in #424
• Use ARCH variable, so we fail early on aarch64 by @ericcurtin in #425
• Small text updates to converter readme by @MichaelClifford in #423
• upping worker and running bootc install by @Gregory-Pereira in #422
• fix(growfs): Correct the source path to be copied by @PaulYuuu in #432
• Add EXTRA_RPM_PACKAGES to image Containerfile by @javipolo in #433
• Add object detection model server readme by @MichaelClifford in #426
• Use OCI Transport for embeded images to save time by @rhatdan in #427
• Create build dir if not previously created by @rhatdan in #438
• restart test environment instance after bootc install by @lmilbaum in #436
• Add growfs systemd service to bootc images by @rhatdan in #437
• intel training container is now pulling in vllm container by @rhatdan in #439
• Include new hi-fidelity implementation and container images by @n1hility in #440
• Growfs is on the wrong line by @rhatdan in #441
• Add installation support for deepspeed by @rhatdan in #442
• Make sure all scripts are included, fix template issues by @n1hility in #445
• Fix type-o that breaks pull on vllm by @n1hility in #444
• Fix deep speed by @n1hility in #443
• Add prune target to cleanup wasted containers/storage by @rhatdan in #446

New Contributors

• @PaulYuuu made their first contribution in #289
• @kwozyman made their first contribution in #316
• @cooktheryan made their first contribution in #359
• @hellohellenmao made their first contribution in #363
• @enriquebelarte made their first contribution in #366
• @n1hility made their first contribution in #360
• @cdoern made their first contribution in #408
• @jaideepr97 made their first contribution in #414
• @ericcurtin made their first contribution in #424
• @javipolo made their first contribution in #433

Full Changelog: v1.0.0...v1.1.0