Summary
- Replacing scipy usage in the adam initial point choosing with numpy (details)
- Removing l_rtm_nudge from flag set file used by run_bindiff_w_flags tests. This is technically a model_setting, so it shouldn't be here, and is already enabled by cgils_s12, so will remain tested in at least one case. (details)
| The file was modified | tuner/pytests/test_adam_baselines.py (diff) |
| The file was modified | tuner/adam_spsa_strategy.py (diff) |
| The file was modified | input/flag_sets/run_bindiff_w_flags_config_core_flags.json (diff) |
