update w4a8 (#290)

ModelTC · Jan 12, 2025 · 4344786 · 4344786
1 parent dbfd394
commit 4344786
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/llmc/compression/quantization/quant.py b/llmc/compression/quantization/quant.py
@@ -1250,7 +1250,7 @@ def __init__(self, bit, bit4, bit8, **kwargs):
                 self.bit_settings[bit]['zeros_qmin'] = self.bit_settings[bit]['qmin']
                 self.bit_settings[bit]['zeros_qmax'] = self.bit_settings[bit]['qmax']
 
-    def reshape_tensor(self, tensor, bit):
+    def reshape_tensor(self, tensor, bit=4):
         granularity = self.bit_settings[bit].get('granularity')
         if granularity == 'per_group':
             group_size = self.bit_settings[bit].get('group_size')