diff --git a/fastdeploy/model_executor/layers/lm_head.py b/fastdeploy/model_executor/layers/lm_head.py index 32f617160..163a5ddda 100644 --- a/fastdeploy/model_executor/layers/lm_head.py +++ b/fastdeploy/model_executor/layers/lm_head.py @@ -140,5 +140,4 @@ class ParallelLMHead(nn.Layer): logits = paddle.matmul(logits.astype(self.weight.dtype), self.weight) else: logits = self.linear(logits.astype(self.linear.weight.dtype)) - print(self.linear.weight.dtype) return logits