fix bug of FusedMoE object has no attribute w13_weight (#94)

Signed-off-by: yuwenzho <[email protected]>
intel · Jan 5, 2025 · 370cc68 · 370cc68
1 parent 9ceb958
commit 370cc68
Showing 1 changed file with 3 additions and 0 deletions.
diff --git a/neural_compressor/torch/algorithms/fp8_quant/_quant_common/helper_modules.py b/neural_compressor/torch/algorithms/fp8_quant/_quant_common/helper_modules.py
@@ -14,6 +14,7 @@
 
 import torch
 import torch.nn as nn
+import types
 
 from .quant_config import QuantMode, get_hqt_config, ScaleFormat
 from .._core.quant_dequant import QuantDequant as qdq
@@ -294,6 +295,8 @@ def __init__(self, mod, mod_extra_config, *args, **kwargs):
         if self.quantization_mode in [QuantMode.QUANTIZE, QuantMode.LOAD]:
             delattr(mod, "w13_weight")
             delattr(mod, "w2_weight")
+            setattr(mod, "w13_weight", None)
+            setattr(mod, "w2_weight", None)
             setattr(self, "w13_weight", None)
             setattr(self, "w2_weight", None)
         self.forward = self.forward_orig