* Added modeld.py (WIP)
* No more VisionIpcBufExtra
* Started work on cython bindings for runmodel
* Got ONNXModel cython bindings mostly working, added ModelFrame bindings
* Got modeld main loop running without model eval
* Move everything into ModelState
* Doesn't crash!
* Moved ModelState into modeld.py
* Added driving_pyx
* Added cython bindings for message generation
* Moved CLContext definition to visionipc.pxd
* *facepalm*
* Move cl_pyx into commonmodel_pyx
* Split out ONNXModel into a subclass of RunModel
* Added snpemodel/thneedmodel bindings
* Removed modeld.cc
* Fixed scons for macOS
* Fixed sconscript
* Added flag for thneedmodel
* paths are now relative to openpilot root dir
* Set cl kernel paths in SConscript
* Set LD_PRELOAD=libthneed.so to fix ioctl interception
* Run from root dir
* A few more fixes
* A few more minor fixes
* Use C update_calibration for now to exactly match refs
* Add nav_instructions input
* Link driving_pyx.pyx with transformations
* Checked python FirstOrderFilter against C++ FirstOrderFilter
* Set process name to fix test_onroad
* Revert changes to onnxmodel.cc
* Fixed bad onnx_runner.py path in onnxmodel.cc
* Import all constants from driving.h
* logging -> cloudlog
* pylint import-error suppressions no longer needed?
* Loop in SConscript
* Added parens
* Bump modeld cpu usage in test_onroad
* Get rid of use_nav
* use config_realtime_process
* error message from ioctl sniffer was messing up pyenv
* cast distance_idx to int
* Removed cloudlog.infos in model.run
* Fixed rebase conflicts
* Clean up driving.pxd/pyx
* Fixed linter error
* Added all maneuvers to navInstruction message
* Added nav instruction logic to modeld
* New model: fcee01c1-96bb-414f-b00d-e4a994a00922/700
* Fixed bug in navd
* Added sharp/slight modifiers
* Updated refs
* Add support for aarch64 pc linux
* Add new libyuv path to release files
* Add libyuv's x86_64 lib dir
* Move left-over platform specific dirs from files_common
* Remove libyuv/lib directory (duplicate)
* Fix mpc Sconscripts
* Remove acados lib path from mpc sources
* Fix typo
* Add watch3 exec on aarch64
* first draft copy, needs updates
* gate behind exp mode
* can be one line
* this is no longer true
* toggle isn't just e2e long now, rename
* don't enable/disable
* kinda works!
* remove old giant qstring
* comment clean up
* similar breakpoint
* more clean up
* fixup onroad
* preserve planner behavior, and we don't need to look at confirmed param if we're onroad
* update translations
* vanish translations
* use scene for op long
* lowercase!
* updated copy
* need to check longitudinal here now too
* clean up!
* here is better
* used
* no nav emoji
* don't check param if not valid
* revert this
revert this
* ed
* revert experimental mode longitudinal checks (still need to revert toggle change)
* revert
* logic works
* revert test
* fix
* update translation
* fix (reset and update)
* first draft copy, needs updates
* gate behind exp mode
* can be one line
* this is no longer true
* toggle isn't just e2e long now, rename
* don't enable/disable
* kinda works!
* remove old giant qstring
* comment clean up
* similar breakpoint
* more clean up
* fixup onroad
* preserve planner behavior, and we don't need to look at confirmed param if we're onroad
* update translations
* vanish translations
* use scene for op long
* lowercase!
* updated copy
* need to check longitudinal here now too
* clean up!
* here is better
* used
* no nav emoji
* don't check param if not valid
* revert this
revert this
* ed
* Fix flicker in nav enabled state
* Move all relevant information for validity checks into navModel packet
* Ignore locationMonoTime in replay tests
* Check route valid in navmodeld
* sm update
* check that
* update refs
---------
Co-authored-by: Adeeb Shihadeh <adeebshihadeh@gmail.com>
* Started work on model runner refactor
* Fixed some compile errors
* everything compiles
* Fixed bug in SNPEModel
* updateInput -> setInputBuffer
* I understand nothing
* whoops lol
* use std::string instead of char*
* Move common logic into RunModel
* formatting fix
* 0.7
* magic
* faster magic
* more simple
* up
* empty
* more mid bits
* naive
* flatten
* dz
* that can stay
* this is fine
* what the
* what the
* giRevert "what the"
This reverts commit 1619ba68e6098dc581fe9b82e7ecb74562b619cc.
* Revert "what the"
This reverts commit 0037dd368290497a6d0009ca34adb2184b584d2e.
* 1x fine
* that was fine
* combined
* independent cum
* 0 is fine
* use metrics
* up cereal
* process and publish from modeld
* cleanup
* use s.output
* bg
* a greener approach
* dns
* serial
* update ref commit
* rebase
* ref
* cereal master
---------
Co-authored-by: Comma Device <device@comma.ai>
* vipc support. modeld and dmonitoringmodeld configs
* Update cereal
* Add vision meta
* Use vision meta in process replay
* Exclude modeld and dmonitoringmodeld from test_processes
* Refactor other pieces of code to use vision meta
* remove unnecessary imports
* Add MAX_FRAMES_PER_SEGMENT
* cameraState migration
* Move camera state migration to migration module
* model_replay refactor
* Fix lint issues
* Reduce number of vipc buffers to 2
* Update spinner message
* Update model replay ref
* Support the new log order
* Replace CameraFrameSizes with dict. Add encode_index field
* Migrate camera states based on encode indices
* Refactor process replay
* Refactor model replay
* Remove unnecessary print
* Update model replay ref
* Make camera state migration optional
* Remove unused import
* Add ignore_alive_pubs field
* Add assertion for carParams inside fingerprinting callback
* Remove WideCameraOnly param setup
* Refactor ModeldCameraSyncRcvCallback
* Update model replay ref commit
* Fix camera state migration
* Only migrate camera states when process has vision pubs
* Refactor model replay again
* Fix linter issues
* One more linter fix
* Update model replay ref commit
* Enable nav features in modeld
* Enable mapsd/navmodeld
* Updated model_replay_ref_commit and added mapsd/navmodeld to test_onroad
* fixed process name
* always publish from mapsd
