This saves time when testing on CPU which is the only sensible thing
to do on GitHub CI for PRs. For releases or once the commit is merged
we could use an external runner with GPU or just wait.
Signed-off-by: Richard Palethorpe <io@richiejp.com>
* chore: default to gemma-3-12b-it-qat
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
* fix: simplify tests to run faster
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
---------
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
* chore: cleanup, identify goal from conversation when evaluting achievement
Signed-off-by: mudler <mudler@localai.io>
* change base cpu model
Signed-off-by: mudler <mudler@localai.io>
* this is not necessary anymore
Signed-off-by: mudler <mudler@localai.io>
* use 12b
Signed-off-by: mudler <mudler@localai.io>
* use openthinker, it's smaller
* chore(tests): set timeout
Signed-off-by: mudler <mudler@localai.io>
* Enable reasoning in some of the tests
Signed-off-by: mudler <mudler@localai.io>
* docker compose unification, small changes
Signed-off-by: mudler <mudler@localai.io>
* Simplify
Signed-off-by: mudler <mudler@localai.io>
* Back at arcee-agent as default
Signed-off-by: mudler <mudler@localai.io>
* Better error handling during planning
Signed-off-by: mudler <mudler@localai.io>
* Ci: do not run jobs for every branch
Signed-off-by: mudler <mudler@localai.io>
---------
Signed-off-by: mudler <mudler@localai.io>
* try to fixup tests, enable e2e
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
* Generate JSON character data with tools
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
* Rework generation of character
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
* Simplify text
Signed-off-by: mudler <mudler@localai.io>
* Relax some test constraints
Signed-off-by: mudler <mudler@localai.io>
* Fixups
* Properly fit schema generation
* Swap default model
* ci fixups
---------
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
Signed-off-by: mudler <mudler@localai.io>
