This release is meant to keep alignment with the Podman Desktop ai-lab extension release cadence.
What's Changed
- fix: object_detection recipes error in Podman AI Lab by @jeffmaury in #505
- Add BASEIMAGE as argument to Containerfile by @enriquebelarte in #506
- Simplify package installation by @enriquebelarte in #511
- ilab-wrapper: fix GPU_AMOUNT check by @markmc in #513
- Remove unused instructlab Makefile variables by @markmc in #512
- initial promptfoo commit by @MichaelClifford in #508
- Fix invalid instructions in README.md by @gcolman in #510
- chore(deps): update auto merged updates by @platform-engineering-bot in #515
- adding single concurrency value to e2e by @Gregory-Pereira in #523
- bump renovate by @cooktheryan in #521
- Configure model for chatbot and codegen recipes by @johnmcollier in #519
- Add initial tool for creating custom eval sets by @MichaelClifford in #518
- fix of arm err on previous run by @cooktheryan in #529
- fix passing of chat_format to llamacpp ms by @Gregory-Pereira in #531
- bump python 3.11 to 3.12 by @Gregory-Pereira in #517
- need to pin version specific version of fedora by @cooktheryan in #533
- Revert "bump python 3.11 to 3.12" by @Gregory-Pereira in #532
- fix of broken install by @cooktheryan in #535
- lamacpp-python-vulkan for arm64 by @Gregory-Pereira in #528
- fixes break that was introduced by @cooktheryan in #537
- Llamacpp python vulkan amd64 by @Gregory-Pereira in #541
- Improve Makefiles by @rhatdan in #534
- Fix Makefiles to allow alternative container engines by @rhatdan in #542
- try to fix the renovate bot from breaking milvus by @cooktheryan in #543
- defaulting no option but allow override chat format by @Gregory-Pereira in #539
- podman build does not support --squash and --layers at same time by @rhatdan in #545
- trying to pass runtime var to allow for model pull by @cooktheryan in #549
- rollback by @cooktheryan in #551
- attempting to pull model manually w ilab by @cooktheryan in #552
- init ilab before pull by @cooktheryan in #554
- Make more options optional by @rhatdan in #550
- ansible -vvv is not helpful at this time by @cooktheryan in #556
- Add pciutils to the OS image by @fabiendupont in #555
- Fix typo for pciutils in AMD bootc by @fabiendupont in #558
- Temporary fix for bootc CI builds by @enriquebelarte in #559
- Add tmux to the bootc image by @fabiendupont in #557
- temp fix revert to llama-2 format by default by @Gregory-Pereira in #565
- Add BASEIMAGE variable by @enriquebelarte in #563
- Remove DKMS from Intel drivers by @enriquebelarte in #564
- specify community package by @cooktheryan in #548
- Add support for BUILD_ARG_FILE by @rhatdan in #544
- typo in server by @cooktheryan in #573
- Make growfs conditional by @enriquebelarte in #572
- Update Auto merged updates by @platform-engineering-bot in #569
- Fixes and improvements for Intel Containerfile by @enriquebelarte in #571
- Update Intel Gaudi to 1.16.0-526 by @tiran in #570
- WIP: resolve build issues with bootc and amd by @cooktheryan in #578
- need to bring in oci artifacts to build by @cooktheryan in #579
- wrong var used new var introduced by @cooktheryan in #581
- Adds models to bootc-models image by @n1hility in #583
- Use multi-stage to build the Intel bootc image by @fabiendupont in #585
- move bootc builder (dtk) to common by @Gregory-Pereira in #586
- Add note about Gaudi updates by @tiran in #587
- Allow custom instructlab image url via env variables by @javipolo in #584
- Fix image builder by @javipolo in #590
- Eliminate growfs from training containers by @rhatdan in #589
- No longer using vllm and deepspeed images for AI bootc images by @rhatdan in #591
- Fix instruct-intel target by @rhatdan in #588
- Fix up pulling images by @rhatdan in #594
- bump std runners from 22.04 to 24.04 by @Gregory-Pereira in #595
- adding new top level variant bootc + models targets by @Gregory-Pereira in #600
- Bring your own podman 5 by @cooktheryan in #603
- get through testing and allow for users to inject their own key with … by @cooktheryan in #604
- netavark tool required by @cooktheryan in #605
- Top level bootc with images by @Gregory-Pereira in #596
- Change source from ilab wrapper by @enriquebelarte in #599
- Revert "Change source from ilab wrapper" by @n1hility in #606
- feat: adding support for hf_pretrained_model option by @axel7083 in #602
- self hosted runner capabilties by @cooktheryan in #609
- allowing pip to install requirements by @Gregory-Pereira in #608
- Otel trace workflow by @sallyom in #607
- training/Makefile: Fix top level vendor targets by @prarit in #598
- Workaround for ilab script source by @enriquebelarte in #612
- fixing aws region on tf stop by @Gregory-Pereira in #613
- Fix ilab by @enriquebelarte in #617
- moving tests from mirror to regular by @Gregory-Pereira in #614
- Update Auto merged updates by @platform-engineering-bot in #610
- Must lowercase the IMAGE name by @rhatdan in #615
- needs runner information by @cooktheryan in #618
- when to run the testing by @cooktheryan in #619
- fixes of image tagging by @cooktheryan in #620
- Make output a subdir of build, to allow chown to work by @rhatdan in #616
- need to set the arch for instructlab image by @cooktheryan in #621
- need to provide nvidia with oci artifacts by @cooktheryan in #622
- Add note about osbuild-selinux required to build disk images by @javipolo in #625
- Add changes to podman oci mount by @enriquebelarte in #624
- wait until the nvidia bootc tasks is complete before deletion by @cooktheryan in #626
- ami switch for nvidia by @cooktheryan in #627
- Install and enable cloud-init by default by @javipolo in #629
- Use --installed to get kernel version for packages to be installed by @javipolo in #628
- Temp workaround for amd-bootc by @enriquebelarte in #630
- Fix typo in Containerfile workaround by @enriquebelarte in #632
- Fix model targets by @javipolo in #631
- udpate otel-collector workflow config by @sallyom in #623
- Fix model targets and add additional improvements by @n1hility in #611
- Drop cloud targets which were unintentionally resurrected by @n1hility in #633
- Add check for installed kernel-core by @enriquebelarte in #635
- Add missing trailing backslash by @javipolo in #638
- Slim models to Granite LAB by @n1hility in #639
- Exclude kernel packages from dnf update by @javipolo in #637
- Remove comment from chained commands in RUN statement by @javipolo in #646
- Remove training images from ailab-images.md by @rhatdan in #647
- Primary training Makefile should not include common by @rhatdan in #642
- fix: add build_vulkan commands for amd64/arm64 on Makefile by @lstocchi in #634
- remove checkout with path so pathing is fixed by @Gregory-Pereira in #648
- Workaround hardlinked /etc changes by @cgwalters in #650
- adding logo to assets and readme by @Gregory-Pereira in #652
New Contributors
- @markmc made their first contribution in #513
- @gcolman made their first contribution in #510
- @johnmcollier made their first contribution in #519
- @fabiendupont made their first contribution in #555
- @tiran made their first contribution in #570
- @axel7083 made their first contribution in #602
- @prarit made their first contribution in #598
Full Changelog: v1.1.2...v1.1.3
Contributors
markmc, cgwalters, and 17 other contributors