diff --git a/CHANGELOG.md b/CHANGELOG.md index 2ebce514e..2d3753368 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -98,6 +98,7 @@ And (*as always*) many bugfixes and improvements to existing features! - **Nunchaku** support for *FLUX.1-Fill* and *FLUX.1-Depth* models - update requirements/packages - use model vae scale-factor for image width/heigt calculations + - **SDNQ** add modules_dtype_dict to quantize *Qwen Image* with mixed dtype - **Other** - **prompt enhance** add `allura-org/Gemma-3-Glitter-4B`, `Qwen/Qwen3-4B-Instruct-2507`, `Qwen/Qwen2.5-VL-3B-Instruct` model support - **prompt enhance** improve system prompt diff --git a/modules/sdnq/__init__.py b/modules/sdnq/__init__.py index d79c8aa0f..924283149 100644 --- a/modules/sdnq/__init__.py +++ b/modules/sdnq/__init__.py @@ -454,7 +454,7 @@ class SDNQConfig(QuantizationConfigMixin): modules_to_not_convert (`list`, *optional*, default to `None`): The list of modules to not quantize, useful for quantizing models that explicitly require to have some modules left in their original precision (e.g. Whisper encoder, Llava encoder, Mixtral gate layers). - modules_to_not_convert (`dict`, *optional*, default to `None`): + modules_dtype_dict (`dict`, *optional*, default to `None`): The dict of dtypes and list of modules, useful for quantizing some modules with a different dtype. """