New
v1.7
What's Changed
- Brew setup @HenryNdubuaku
- Cactus auth @HenryNdubuaku
- Hybrid inference by the cactus team
- Karen/vlm fix by @kar-m in https://github.com/cactus-compute/cactus/pull/311
- fixed moonshine state resetting and gemma3 4b layernorm loading by @kar-m in https://github.com/cactus-compute/cactus/pull/317
- fix: LFM2 multiple tool calls by @mhayes853 in https://github.com/cactus-compute/cactus/pull/316
- fix hf publish by @jakmro in https://github.com/cactus-compute/cactus/pull/323
- update models list by @jakmro in https://github.com/cactus-compute/cactus/pull/324
- Fixing pip command errors by @rshemet in https://github.com/cactus-compute/cactus/pull/322
- Add instructions for installing Ruby version for xcodeproj gem by @jakmro in https://github.com/cactus-compute/cactus/pull/327
- tests: remove duplicate vlm_multiturn test in runner by @AI-I224 in https://github.com/cactus-compute/cactus/pull/332
- fix: replace NSLog with CACTUS_LOG for iOS NPU debuggability by @KayaanT in https://github.com/cactus-compute/cactus/pull/328
- Kernel_attention optimization by @Ayan9074 in https://github.com/cactus-compute/cactus/pull/319
- M4airbenchmarks by @Ayan9074 in https://github.com/cactus-compute/cactus/pull/336
- docs: update cactus test command description for transcribe models (#297) by @AI-I224 in https://github.com/cactus-compute/cactus/pull/339
- Accelerate FP16 matmul via cblas_sgemm for Apple AMX by @KayaanT in https://github.com/cactus-compute/cactus/pull/340
- Fix hybrid attention sliding window for Gemma (#320) by @jrajala6 in https://github.com/cactus-compute/cactus/pull/338
- bench: update README benchmark with M2 MacBook Air results by @vyomshah05 in https://github.com/cactus-compute/cactus/pull/335
- docs: add iPad Pro (12.9") (6th Gen) benchmarks (#296) by @AI-I224 in https://github.com/cactus-compute/cactus/pull/333
- removed unused graph i/o methods by @ncylich in https://github.com/cactus-compute/cactus/pull/345
- feat: cpp-native telemetry by @justinl66 in https://github.com/cactus-compute/cactus/pull/326
- Update CPP Telemetry to point to main DB by @justinl66 in https://github.com/cactus-compute/cactus/pull/350
- update python bindings for stream transcribe by @jakmro in https://github.com/cactus-compute/cactus/pull/351
- Update CPP Telemetry by @justinl66 in https://github.com/cactus-compute/cactus/pull/352
- added only flag by @nshejwalkar in https://github.com/cactus-compute/cactus/pull/347
- Added warmups and increased iterations for performance testing by @nshejwalkar in https://github.com/cactus-compute/cactus/pull/355
- CMF Phone 2 Pro benchmarks by @jakmro in https://github.com/cactus-compute/cactus/pull/356
- Vad by @jakmro in https://github.com/cactus-compute/cactus/pull/353
- Cli reconvert by @jakmro in https://github.com/cactus-compute/cactus/pull/357
- Asr cloud merging by @kar-m in https://github.com/cactus-compute/cactus/pull/348
- Add optional cloud key prompt for transcribe by @rshemet in https://github.com/cactus-compute/cactus/pull/359
- HF support multiple precision options by @jakmro in https://github.com/cactus-compute/cactus/pull/361
- Add precision parameter to download_from_hf by @jakmro in https://github.com/cactus-compute/cactus/pull/362
- revert silero download logic by @jakmro in https://github.com/cactus-compute/cactus/pull/365
- Cactus clean now clears cache, Session metrics initialized properly for telemetry by @justinl66 in https://github.com/cactus-compute/cactus/pull/363
- Curl prepack by @kar-m in https://github.com/cactus-compute/cactus/pull/358
- Fix/f16 reduction accum by @vyomshah05 in https://github.com/cactus-compute/cactus/pull/344
- Update telemetry by @justinl66 in https://github.com/cactus-compute/cactus/pull/366
- Accelerate FP16 attention via cblas_sgemm for Apple AMX by @KayaanT in https://github.com/cactus-compute/cactus/pull/346
New Contributors
- @AI-I224 made their first contribution in https://github.com/cactus-compute/cactus/pull/332
- @jrajala6 made their first contribution in https://github.com/cactus-compute/cactus/pull/338
- @vyomshah05 made their first contribution in https://github.com/cactus-compute/cactus/pull/335
- @nshejwalkar made their first contribution in https://github.com/cactus-compute/cactus/pull/347
Full Changelog: https://github.com/cactus-compute/cactus/compare/v1.6.0...v1.7
@mhayes853 API has breaking changes