PTQ-Level Efficiency, QAT-Level Accuracy | 训练后量化的效率,量化感知训练的精度
fix: preserve layer config for secondary diffusion transformers (#2337)
Wang, Changc78c11a66abb
diff --git a/auto_round/compressors/diffusion_mixin.py b/auto_round/compressors/diffusion_mixin.pyindex 844fcd572..1b285138c 100644--- a/auto_round/compressors/diffusion_mixin.py+++ b/auto_round/compressors/diffusion_mixin.py@@ -365,7 +365,6 @@ class DiffusionMixin:# Get block names for new transformerall_blocks = get_block_names(self.model_context.model)self.quant_block_list = find_matching_blocks(self.model_context.model, all_blocks, None)- self.layer_config = {}# Get new block names for cachingif bool(self.quant_block_list):