v0.7.1

What's Changed

• Bump to v0.7.0 by @rhatdan in #1042
• Explain dryrun option better in container_build.sh by @ericcurtin in #1041
• Add openvino to all images by @rhatdan in #1045
• Print status message when emulating --pull=newer for docker by @edmcman in #1047
• Remove unused variable by @ericcurtin in #1044
• Default the number of threads to (nproc)/(2) by @ericcurtin in #982
• Attempt to install openvino using pip by @rhatdan in #1050
• feat: add --jinja to the list of arguments if MODEL_JINJA env var is true by @benoitf in #1053
• Never use entrypoint by @rhatdan in #1046
• fix ramalama rag build code by @rhatdan in #1049
• Combine Vulkan, Kompute and CPU inferencing into one image by @ericcurtin in #1022
• Hardcode threads to 2 in this test by @ericcurtin in #1056
• fixed chunk error by @bmahabirbu in #1059
• Don't display server port when using run --rag by @rhatdan in #1061
• Add support for /dev/accel being leaked into containers by @rhatdan in #1055

Full Changelog: v0.7.0...v0.7.1

v0.7.0

This is a big release, We now have working support for RAG inside of RamaLama.
Try out
ramalama rag XYZ.pdf ABC.doc quay.io/NAME/myrag
ramalama run --rag quay.io/NAME/myrag MYMODEL

What's Changed

• Default whisper-server.sh, llama-server.sh to /mnt/models/model.file by @rhatdan in #984
• Improve intel-gpu to work with whisper-server and llama-server by @rhatdan in #986
• whisper.cpp requires ffmpeg by @ericcurtin in #985
• Fix container_build.sh to build all images by @rhatdan in #989
• fix: use expected condition by @benoitf in #992
• [CANN]Fix the bug that openEuler repo does not have ffmpeg-free package, instand of using ffmpeg for openEuler by @leo-pony in #994
• Add docling support version 2 by @rhatdan in #979
• chore: use the reverse condition for models by @benoitf in #995
• FIX: Ollama install with brew for CI by @kush-gupt in #1002
• Add the ability to identify a wider set of Intel GPUs that have enough Execution Units to produce decent results by @cgruver in #996
• Add ramalama client by @ericcurtin in #997
• Fix errors found in RamaLama RAG by @rhatdan in #998
• Turn on verbose logging in llama-server if --debug is on by @ericcurtin in #1001
• Don't use relative paths for destination by @rhatdan in #1003
• Red Hat Konflux update ramalama by @red-hat-konflux in #1005
• Fix errors on python3.9 by @rhatdan in #1007
• Use this container if we detect ROCm accelerator by @ericcurtin in #1008
• Improve UX for ramalama-client by @ericcurtin in #1013
• update docs for Intel GPU support. Clean up code comments by @cgruver in #1011
• Generate quadlets with rag databases by @rhatdan in #1012
• Keep conversation history by @ericcurtin in #1014
• Fix ramalama serve --rag ABC --generate kube by @rhatdan in #1015
• Adds Rag chatbot to ramalama serve and preloads models for doc2rag and rag_framework by @bmahabirbu in #1010
• Rag condition should be and instead of or by @ericcurtin in #1016
• Show model name in API instead of model file path by @bachp in #1009
• Make install script more aesthetically pleasing by @ericcurtin in #1019
• Color each word individually by @ericcurtin in #1017
• Add feature to turn off colored text by @ericcurtin in #1021
• Fix up building of images by @rhatdan in #1023
• Change default ROCM image to rocm-fedora by @rhatdan in #1024
• Run build_rag.sh as root by @rhatdan in #1027
• added hacky method to use 'run' instead of 'serve' for rag by @bmahabirbu in #1026
• More fixes to build scripts by @rhatdan in #1028
• Updated rag to have much better querys at the cost of slight delay by @bmahabirbu in #1029
• More fixes to build scripts by @rhatdan in #1031
• Minor bugfix remove self. from self.prompt by @ericcurtin in #1032
• Added terminal name fixed eof bug and added another model to rag_framework load by @bmahabirbu in #1033
• Update registry.access.redhat.com/ubi9/ubi Docker tag to v9.5-1742918310 by @renovate in #1035
• Typo in the webui by @ericcurtin in #1039
• Fix errors on python3.9 by @marceloleitner in #1038
• More updates for builds by @rhatdan in #1036

New Contributors

• @red-hat-konflux made their first contribution in #1005
• @bachp made their first contribution in #1009
• @marceloleitner made their first contribution in #1038

Full Changelog: v0.6.4...v0.7.0

v0.6.4

What's Changed

• Print error when converting from an OCI Image by @rhatdan in #932
• Make compatible with the macOS system python3 by @ericcurtin in #933
• Bugfixes noticed while installing on Raspberry Pi by @ericcurtin in #935
• Add note about updating nvidia.yaml file by @rhatdan in #938
• Fix docker handling of GPUs. by @rhatdan in #941
• macOS detection fix by @ericcurtin in #942
• Add chat template support by @engelmi in #917
• Consolidate gpu detection by @ericcurtin in #943
• Implement RamaLama shell by @ericcurtin in #915
• Add Linux x86-64 support for Ascend NPU accelerator in llama.cpp backend by @leo-pony in #950
• Handle CNAI annotation deprecation by @s3rj1k in #939
• Fix install.sh for OSTree system by @ericcurtin in #951
• Lets run container in all tests, to make sure it does not explode. by @rhatdan in #946
• Added --chat-template-file support to ramalama serve by @engelmi in #952
• Update registry.access.redhat.com/ubi9/ubi Docker tag to v9.5-1741850090 by @renovate in #956
• Add specified nvidia-oci runtime by @rhatdan in #953
• python3 validator by @ericcurtin in #959
• There must be at least one CDI device present to use CUDA by @ericcurtin in #954
• [NPU][Fix] only specify device num, but without ascend-docker-runtime installed, running ramalama/cann container image will failing by @leo-pony in #962
• Fix port rendering in README by @andreadecorte in #963
• Update docker.io/nvidia/cuda Docker tag to v12.8.1 by @renovate in #960
• Update llama.cpp to contain threads features by @ericcurtin in #967
• Fix ENTRYPOINTS of whisper-server and llama-server by @rhatdan in #965
• Add software to support using rag in RamaLama by @rhatdan in #968
• Update llama.cpp for some Gemma features by @ericcurtin in #973
• Only set this environment variable if we can resolve CDI by @ericcurtin in #971
• feat(cpu): add --threads option to specify number of cpu threads by @antheas in #966
• Asashi build is failing because of no python3-devel package by @rhatdan in #974
• GPG Check is failing on the Intel Repo by @cgruver in #976
• Add --runtime-arg option for run and serve by @edmcman in #949
• Fix handling of whisper-server and llama-server entrypoints by @rhatdan in #975
• Bump to v0.6.4 by @rhatdan in #978

New Contributors

• @s3rj1k made their first contribution in #939
• @antheas made their first contribution in #966
• @edmcman made their first contribution in #949

Full Changelog: v0.6.3...v0.6.4

v0.6.3

What's Changed

• Check if terminal is compatible with emojis before using them by @ericcurtin in #878
• Use vllm-openai upstream image by @ericcurtin in #880
• The package available via dnf is in a good place by @ericcurtin in #879
• Add Ollama to CI and system tests for its caching by @kush-gupt in #881
• Moved pruning protocol from model to factory by @engelmi in #882
• Remove emoji usage until linenoise.cpp and llama-run are compatible by @ericcurtin in #884
• Inject config to cli functions by @engelmi in #889
• Switch from tiny to smollm:135m by @ericcurtin in #891
• benchmark failing because of lack of flag by @ericcurtin in #888
• Update the README.md to point people at ramalama.ai web site by @rhatdan in #894
• fix: handling of date with python 3.8/3.9/3.10 by @benoitf in #897
• readme: fix artifactory link by @alaviss in #903
• Added support for mac cpu and clear warning message by @bmahabirbu in #902
• Use python variable instead of environment variable by @ericcurtin in #907
• Update llama.cpp by @ericcurtin in #908
• Build a non-kompute Vulkan container image by @ericcurtin in #910
• Reintroduce emoji prompts by @ericcurtin in #913
• Add new ramalama-*-core executables by @ericcurtin in #909
• Detect & get info on hugging face repos, fix sizing of symlinked directories by @kush-gupt in #901
• Add ramalama image built on Fedora using Fedora's rocm packages by @maxamillion in #596
• Add new model store by @engelmi in #905
• Add support for llama.cpp engine to use ascend NPU device by @leo-pony in #911
• Extend make validate check to do more by @ericcurtin in #916
• Modify GPU detection to match against env var value instead of prefix by @cgruver in #919
• Add Intel ARC 155H to list of supported hardware by @cgruver in #920
• Try to choose a free port on serve if default one is not available by @andreadecorte in #898
• Add passing of environment variables to ramalama commands by @rhatdan in #922
• Allow user to specify the images to use per hardware by @rhatdan in #921
• fix: CHAT_FORMAT variable should be expanded by @benoitf in #926
• Update registry.access.redhat.com/ubi9/ubi Docker tag to v9.5-1741600006 by @renovate in #928
• Bump to v0.6.3 by @rhatdan in #931

New Contributors

• @alaviss made their first contribution in #903
• @leo-pony made their first contribution in #911
• @andreadecorte made their first contribution in #898

Full Changelog: v0.6.2...v0.6.3

v0.6.2

What's Changed

• Introduce basic renovate.json file by @gnaponie in #854
• Some tests around --network, --net options by @ericcurtin in #840
• Add demos script to show the power of RamaLama by @rhatdan in #855
• chore: add alias from llama-2 to llama2 by @benoitf in #859
• Define Environment variables to use by @rhatdan in #861
• Fix macOS GPU acceleration via podman by @ericcurtin in #863
• Change rune to run by @ericcurtin in #862
• Revert back to 12.6 version of cuda by @rhatdan in #864
• Make CI build all images by @ericcurtin in #831
• chore: do not format size for --json export in list command by @benoitf in #870
• Added model factory by @engelmi in #874
• feat: display emoji of the engine for the run in the prompt by @benoitf in #872
• Fix up handling of image selection on generate by @rhatdan in #856
• fix: use iso8601 for JSON modified field by @benoitf in #873
• Bump to 0.6.2 by @rhatdan in #875

New Contributors

• @gnaponie made their first contribution in #854

Full Changelog: v0.6.1...v0.6.2

v0.6.1

What's Changed

• chore: use absolute link for the RamaLama logo by @benoitf in #781
• Reuse Ollama cached image when available by @kush-gupt in #782
• Add env var RAMALAMA_GPU_DEVICE to allow for explicit declaration of the GPU device to use by @cgruver in #773
• Change RAMALAMA_GPU_DEVICE to RAMALAMA_DEVICE for AI accelerator device override by @cgruver in #786
• Add Security information to README.md by @rhatdan in #787
• Fix exiting on llama-serve when user hits ^c by @rhatdan in #785
• Check if file exists before sorting them into a list by @kush-gupt in #784
• Add ramalama run --keepalive option by @rhatdan in #789
• Stash output from container_manager by @rhatdan in #790
• Install llama.cpp for mac and nocontainer tests by @rhatdan in #792
• _engine is set to None or has a value by @ericcurtin in #793
• Only run dnf commands on platforms that have dnf by @ericcurtin in #794
• Add ramalama rag command by @rhatdan in #501
• Attempt to use build_llama_and_whisper.sh by @rhatdan in #795
• Change --network-mode to --network by @ericcurtin in #800
• Add some more gfx values to the default list by @ericcurtin in #806
• Update registry.access.redhat.com/ubi9/ubi Docker tag to v9.5-1739449058 by @renovate in #808
• Prepare containers to run with ai-lab-recipes by @rhatdan in #803
• If ngl is not specified by @ericcurtin in #802
• feat: add ramalama labels about the execution on top of container by @benoitf in #810
• Add run and serve arguments for --device and --privileged by @cgruver in #809
• chore: rewrite readarray function to make it portable by @benoitf in #815
• chore: replace RAMALAMA label by ai.ramalama by @benoitf in #814
• Upgrade from 6.3.1 to 6.3.2 by @ericcurtin in #816
• Removed error wrapping in urlopen by @engelmi in #818
• Encountered a bug where this function was returning -1 by @ericcurtin in #817
• Align runtime arguments with run, serve, bench, and perplexity by @cgruver in #820
• README: fix inspect command description by @kush-gupt in #826
• Pin dev dependencies to major version and improve formatting + linting by @engelmi in #824
• README: Fix typo by @bupd in #827
• Switch apt-get to apt by @ericcurtin in #832
• Update registry.access.redhat.com/ubi9/ubi Docker tag to v9.5-1739751568 by @renovate in #834
• Add entrypoint container images by @rhatdan in #819
• HuggingFace Cache Implementation by @kush-gupt in #833
• Make serve by default expose network by @ericcurtin in #830
• Fix up man page help verifacation by @rhatdan in #835
• Fix handling of --privileged flag by @rhatdan in #821
• chore: fix links of llama.cpp repository by @benoitf in #841
• Unify CLI options (verbosity, version) by @mkesper in #685
• Add system tests to pull from the Hugging Face cache by @kush-gupt in #846
• Just one add_argument call for --dryrun/--dry-run by @ericcurtin in #847
• Fix ramalama info to display NVIDIA and amd GPU information by @rhatdan in #848
• Remove LICENSE header from gpu_detector.py by @ericcurtin in #850
• Allowing modification of pull policy by @rhatdan in #843
• Include instructions for installing on Fedora 42+ by @stefwalter in #849
• Bump to 0.6.1 by @rhatdan in #851

New Contributors

• @benoitf made their first contribution in #781
• @bupd made their first contribution in #827
• @mkesper made their first contribution in #685
• @stefwalter made their first contribution in #849

Full Changelog: v0.6.0...v0.6.1

v0.6.0

What's Changed

• fix error on macOS for M1 pro by @volker48 in #687
• This should be a global variable by @ericcurtin in #703
• Update registry.access.redhat.com/ubi9/ubi Docker tag to v9.5-1736404036 by @renovate in #702
• Update install.sh to include "gpu_detector.py" by @graystevens in #704
• add --ngl to specify the number of gpu layers, and --keep-groups so podman has access to gpu by @khumarahn in #659
• We are displaying display driver info, scope creep by @ericcurtin in #710
• Use CODEOWNERS file for autoassign by @dougsland in #706
• common: general improvements by @dougsland in #713
• Fix macOS emoji compatibility with Alacritty by @ericcurtin in #716
• Makelint by @dougsland in #715
• Adding slp, engelmi, also by @ericcurtin in #711
• Report error when huggingface-cli is not available by @rhatdan in #719
• Add --network-mode option by @rhjostone in #674
• README: add convert to commands list by @kush-gupt in #723
• Revert "Add --network-mode option" by @ericcurtin in #731
• Check for apple,arm-platform in /proc by @ericcurtin in #730
• Packit: downstream jobs for EPEL 9,10 by @lsm5 in #728
• Add logic to build intel-gpu image to build_llama_and_whisper.sh by @cgruver in #724
• Add --network-mode option by @rhatdan in #734
• Honor RAMALAMA_IMAGE if set by @rhatdan in #733
• ramalama container: Make it possible to build basic container on all RHEL architectures by @jcajka in #722
• Add docs for using podman farm to build multi-arch images by @cgruver in #735
• Update registry.access.redhat.com/ubi9/ubi Docker tag to v9.5-1738643550 by @renovate in #729
• modify container_build.sh to add capability to use podman farm for multi-arch images by @cgruver in #736
• There's a comma in the list of files in install.sh by @ericcurtin in #739
• Make the default of ngl be -1 by @ericcurtin in #707
• github actions: ramalama install by @dougsland in #738
• [skip-ci] Update actions/checkout action to v4 by @renovate in #740
• On macOS this was returning an incorrect path by @ericcurtin in #741
• Begin process of packaging PRAGmatic by @rhatdan in #597
• Allow users to build RAG versus Docling images by @rhatdan in #744
• Update vLLM containers by @ericcurtin in #746
• Update README.md by @bmbouter in #748
• Update progress bar only once every 100ms by @ericcurtin in #717
• Remove reference to non-existent docs in CONTRIBUTING.md by @cgruver in #761
• Check if krunkit process is running with --all-providers by @ericcurtin in #763
• update_progress only takes one parameter by @ericcurtin in #764
• Detect Intel ARC GPU in Meteor Lake chipset by @cgruver in #749
• Drop all capablities and run with no-new-privileges by @rhatdan in #765
• Progress bar fixes by @ericcurtin in #767
• typo: Add quotes to intel-gpu argument in build llama and whisper script by @hanthor in #766
• chore(deps): update registry.access.redhat.com/ubi9/ubi docker tag to v9.5-1738814488 by @renovate in #771
• There would be one case where this wouldn't work by @ericcurtin in #768
• docs: update ramalama.1.md by @eltociear in #775
• Add community documents by @rhatdan in #777
• Parse https://ollama.com/library/ syntax by @ericcurtin in #648
• Use containers CODE-OF-CONDUCT.md by @rhatdan in #778
• Add model inspect cli by @engelmi in #776
• Cleanup READMEs and man pages. by @rhatdan in #780
• Bump to v0.6.0 by @rhatdan in #779

New Contributors

• @volker48 made their first contribution in #687
• @graystevens made their first contribution in #704
• @khumarahn made their first contribution in #659
• @rhjostone made their first contribution in #674
• @jcajka made their first contribution in #722
• @bmbouter made their first contribution in #748
• @hanthor made their first contribution in #766
• @eltociear made their first contribution in #775

Full Changelog: v0.5.5...v0.6.0

v0.5.5

What's Changed

• Add perplexity subcommand to RamaLama CLI by @ericcurtin in #637
• throwing an exception with there is a failure in http_client.init by @jhjaggars in #647
• Add container image to support Intel ARC GPU by @cgruver in #644
• Guide users to install huggingface-cli to login to huggingface by @pbabinca in #645
• Update intel-gpu Containerfile to reduce the size of the builder image by @cgruver in #657
• Look for configs also in /usr/local/share/ramalama by @jistr in #672
• remove ro as an option when mounting images by @kush-gupt in #676
• Add generated man pages for section 7 into gitignore by @jistr in #673
• Revert "Added --jinja to llama-run command" by @ericcurtin in #683
• Pull the source model if it isn't already in local storage for the convert and push functions by @kush-gupt in #680
• bump llama.cpp to latest release hash aa6fb13 by @maxamillion in #692
• Introduce a mode so one call install from git by @ericcurtin in #690
• Add ramalama gpu_detector by @dougsland in #670
• Bump to v0.5.5 by @rhatdan in #701

New Contributors

• @cgruver made their first contribution in #644
• @pbabinca made their first contribution in #645
• @jistr made their first contribution in #672
• @kush-gupt made their first contribution in #676
• @maxamillion made their first contribution in #692
• @dougsland made their first contribution in #670

Full Changelog: v0.5.4...v0.5.5

v0.5.4

What's Changed

• Attempt to install podman by @ericcurtin in #621
• Introduce ramalama bench by @ericcurtin in #620
• Add man page for cuda support by @rhatdan in #623
• Less verbose output by @ericcurtin in #624
• Avoid dnf install on OSTree system by @ericcurtin in #622
• Fix list in README - Credits section by @kubealex in #627
• added mac cpu only support by @bmahabirbu in #628
• Added --jinja to llama-run command by @engelmi in #625
• Update llama.cpp version by @ericcurtin in #630
• Add shortname for deepseek by @rhatdan in #631
• fixed rocm detection by adding gfx targets in containerfile by @bmahabirbu in #632
• Point macOS users to script install by @kubealex in #635
• Update docker.io/nvidia/cuda Docker tag to v12.8.0 by @renovate in #633
• feat: add argument to define amd gpu targets by @jobcespedes in #634
• Bump to v0.5.4 by @rhatdan in #641

New Contributors

• @kubealex made their first contribution in #627
• @engelmi made their first contribution in #625
• @jobcespedes made their first contribution in #634

Full Changelog: v0.5.3...v0.5.4

v0.5.3

What's Changed

• We no longer have python dependancies by @ericcurtin in #588
• container_build.sh works on MAC by @rhatdan in #590
• Added vllm cuda support by @bmahabirbu in #582
• Remove omlmd from OCI calls by @rhatdan in #591
• Build with curl support by @pepijndevos in #595
• Add model transport info to ramalama run/serve manpage by @rhatdan in #593
• Various README.md updates by @ericcurtin in #600
• code crashes for rocm added proper type cast for env var by @bmahabirbu in #602
• ROCm build broken by @ericcurtin in #605
• Cleaner output if a machine executes this command by @ericcurtin in #604
• Update to version that has command history by @ericcurtin in #603
• Remove these lines they are unused by @ericcurtin in #606
• Had to make this change for my laptop to suppor nvidia by @rhatdan in #609
• Start making vllm work with RamaLama by @rhatdan in #610
• Treat hf.co/ prefix the same as hf:// by @ericcurtin in #612
• We need the rocm libraries in here by @ericcurtin in #613
• A couple of cleanups in build_llama_and_whisper.sh by @rhatdan in #615
• Bump to v0.5.3 by @rhatdan in #614

New Contributors

• @pepijndevos made their first contribution in #595

Full Changelog: v0.5.2...v0.5.3