diff --git a/common/config_models.py b/common/config_models.py index 6aac505..b1266eb 100644 --- a/common/config_models.py +++ b/common/config_models.py @@ -213,10 +213,12 @@ class ModelConfig(BaseConfigModel): ), ) rope_alpha: Optional[Union[float, Literal["auto"]]] = Field( - 1.0, + None, description=( - "Rope alpha (default: 1.0).\n" - 'Same as alpha_value. Set to "auto" to auto-calculate.' + "Rope alpha (default: None).\n" + 'Same as alpha_value. Set to "auto" to auto-calculate.\n' + "Leaving this value blank will either pull from the model " + "or auto-calculate." ), ) cache_mode: Optional[CACHE_SIZES] = Field( @@ -318,7 +320,9 @@ class DraftModelConfig(BaseConfigModel): None, description=( "Rope alpha for draft models (default: None).\n" - 'Same as alpha_value. Set to "auto" to auto-calculate.' + 'Same as alpha_value. Set to "auto" to auto-calculate.\n' + "Leaving this value blank will either pull from the model " + "or auto-calculate." ), ) draft_cache_mode: Optional[CACHE_SIZES] = Field( diff --git a/config_sample.yml b/config_sample.yml index 0da6a36..507d7d5 100644 --- a/config_sample.yml +++ b/config_sample.yml @@ -97,9 +97,10 @@ model: # Leave blank to pull the value from the model. rope_scale: 1.0 - # Rope alpha (default: 1.0). + # Rope alpha (default: None). # Same as alpha_value. Set to "auto" to auto-calculate. - rope_alpha: 1.0 + # Leaving this value blank will either pull from the model or auto-calculate. + rope_alpha: # Enable different cache modes for VRAM savings (default: FP16). # Possible values: 'FP16', 'Q8', 'Q6', 'Q4'. @@ -154,6 +155,7 @@ draft_model: # Rope alpha for draft models (default: None). # Same as alpha_value. Set to "auto" to auto-calculate. + # Leaving this value blank will either pull from the model or auto-calculate. draft_rope_alpha: # Cache mode for draft models to save VRAM (default: FP16).