0.1.35

What's Changed

• small cleanup to pin code by @MarcusDunn in #123 Potentially breaking
• updated llama.cpp by @github-actions in #124
• updated llama.cpp by @github-actions in #125
• updated llama.cpp by @github-actions in #126
• Bump docker/setup-buildx-action from 3.0.0 to 3.1.0 by @dependabot in #129
• updated llama.cpp by @github-actions in #128
• updated llama.cpp by @github-actions in #131

Full Changelog: 0.1.34...0.1.35

0.1.34

What's Changed

• Add CPU Feature Support by @Hirtol in #121
• override model values by @MarcusDunn in #120
• prep 0.1.34 by @MarcusDunn in #122

New Contributors

• @Hirtol made their first contribution in #121

Full Changelog: 0.1.33...0.1.34

0.1.33

What's Changed

• updated llama.cpp by @github-actions in #115
• updated llama.cpp by @github-actions in #117
• Expose the complete API for dealing with KV cache and states by @zh217 in #116
• add with_main_gpu to LlamaModelParams by @danbev in #118
• updated llama cpp and removed cast to mut by @MarcusDunn in #119

New Contributors

• @danbev made their first contribution in #118

Full Changelog: 0.1.32...0.1.33

0.1.32

What's Changed

• updated llama.cpp by @github-actions in #105
• Bump cc from 1.0.83 to 1.0.88 by @dependabot in #106
• added more sampling options by @MarcusDunn in #110
• updated llama.cpp by @github-actions in #111
• Expose functions llama_load_session_file and llama_save_session_file by @zh217 in #112
• Improved docs for new sampling options @MarcusDunn
• Fix clippy errors @MarcusDunn

New Contributors

• @zh217 made their first contribution in #112

Full Changelog: 0.1.31...0.1.32

0.1.31

What's Changed

• added docs for cublas support by @MarcusDunn in #103
• added with_use_mlock and llama_supports_mlock by @MarcusDunn in #104
• moved simple to its own binary for easier use + faster compile times by @MarcusDunn in #101

Full Changelog: 0.1.30...0.1.31

0.1.30

fixed not including ggml-metal.h in release

Full Changelog: 0.1.29...0.1.30

0.1.29

h/t @SilasMarvin for bringing metal support across the finish line!

What's Changed

• updated llama.cpp by @github-actions in #95
• updated llama.cpp by @github-actions in #97
• updated llama.cpp by @github-actions in #98
• updated llama.cpp by @github-actions in #100
• Working build.rs for apple metal by @SilasMarvin in #96
• attempt to add metal on mac by @MarcusDunn in #65
• Prep 0 1 29 by @MarcusDunn in #102

Full Changelog: 0.1.28...0.1.29

0.1.28

Known Breaking

• init_numa has been modified to accept an enum instead of a boolean.

What's Changed

• Bump anyhow from 1.0.79 to 1.0.80 by @dependabot in #89
• Bump clap from 4.5.0 to 4.5.1 by @dependabot in #90
• prep 0.1.28 by @MarcusDunn in #94
• Process user defined tokens by @SilasMarvin in #93
• updated llama.cpp (includes breaking backend init changes) by @github-actions in #92

New Contributors

• @SilasMarvin made their first contribution in #93

Full Changelog: 0.1.27...0.1.28

0.1.27

What's Changed

• updated llama.cpp by @github-actions in #83
• updated llama.cpp by @github-actions in #84
• updated llama.cpp by @github-actions in #85

Full Changelog: 0.1.26...0.1.27

0.1.26

What's Changed

• updated llama.cpp by @github-actions in #63
• exposed n_threads by @MarcusDunn in #66
• updated llama.cpp by @github-actions in #68
• Bump bindgen from 0.69.2 to 0.69.4 by @dependabot in #69
• update-toml fix by @sepehr455 in #73
• updated llama.cpp by @github-actions in #77
• Bump clap from 4.4.18 to 4.5.0 by @dependabot in #80
• Bump thiserror from 1.0.56 to 1.0.57 by @dependabot in #81
• updated llama.cpp by @github-actions in #79
• Add Windows MSVC support by @Systemcluster in #78

New Contributors

• @Systemcluster made their first contribution in #78

Full Changelog: 0.1.25...0.1.26