[to #42322933] fix forward input in token classification

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10103632

modelscope/pipelines/nlp/word_segmentation_pipeline.py

@@ -62,7 +62,7 @@ class WordSegmentationPipeline(Pipeline):
         text = inputs.pop(OutputKeys.TEXT)
         with torch.no_grad():
             return {
-                **self.model(inputs, **forward_params), OutputKeys.TEXT: text
+                **self.model(**inputs, **forward_params), OutputKeys.TEXT: text
             }
 
     def postprocess(self, inputs: Dict[str, Any],