Releases: triton-inference-server/model_analyzer
Releases · triton-inference-server/model_analyzer
Release 1.31.0 corresponding to NGC container 23.08
- Added Quick Start guides for Ensemble and BLS models
Release 1.30.0 corresponding to NGC container 23.07
- Implemented periodic checkpointing
- Added support for custom docker args
- Detect and handle invalid metrics url
- Profile will now automatically create the default detailed reports
Release 1.29.0 corresponding to NGC container 23.06
request-rate-range
can now be searched in brute mode- Capture PA errors in a log file
- Added detection for Triton Server launch failures
- Added
cpu_only
option for ensemble composing models - Added binary concurrency search to quick search mode
- Added binary parameter search to brute search mode
Release 1.28.0 corresponding to NGC container 23.05
Release 1.27.0 corresponding to NGC container 23.04
Release 1.21.0 corresponding to NGC container 22.10
Release 1.20.0 corresponding to NGC container 22.09
Release 1.14.0 corresponding to NGC container 22.03
- Added support to allow the user to specify a max batch size when automatically sweeping
Release 1.10.0 corresponding to NGC container 21.11
Release 1.5.0 corresponding to NGC container 21.06
- Initial release of Model Analyzer