Skip to content

Commit ef2201d

Browse files
authored
fix vit quantcfg init error. (#709)
1 parent 5c36aa7 commit ef2201d

File tree

1 file changed

+1
-1
lines changed

1 file changed

+1
-1
lines changed

lightllm/models/vit/model.py

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -107,7 +107,7 @@ def _init_weights(self):
107107
return
108108

109109
def _init_quant(self):
110-
self.quant_cfg = Quantcfg(self.config["num_hidden_layers"], self.quant_type, self.quant_cfg_path)
110+
self.quant_cfg = Quantcfg(self.config, self.quant_type, self.quant_cfg_path)
111111
logger.info(f"Initial quantization. " f"The default quantization method is {self.quant_cfg.quant_type}")
112112

113113
def _init_infer_layer(self):

0 commit comments

Comments
 (0)