Optimize triton

Signed-off-by: Jael Gu <mengjia.gu@zilliz.com>

auto_transformers.py

@@ -59,10 +59,17 @@ def create_model(model_name, checkpoint_path, device):
 # @accelerate
 class Model:
     def __init__(self, model_name, checkpoint_path, device):
+        self.device = device
         self.model = create_model(model_name, checkpoint_path, device)
 
     def __call__(self, *args, **kwargs):
-        outs = self.model(*args, **kwargs, return_dict=True)
+        new_args = []
+        for x in args:
+            new_args.append(x.to(self.device))
+        new_kwargs = {}
+        for k, v in kwargs.items():
+            new_kwargs[k] = v.to(self.device)
+        outs = self.model(*new_args, **new_kwargs, return_dict=True)
         return outs['last_hidden_state']
 
 
@@ -84,7 +91,6 @@ class AutoTransformers(NNOperator):
                  checkpoint_path: str = None,
                  tokenizer: object = None,
                  device: str = None,
-                 norm: bool = False
                  ):
         super().__init__()
         if device:
@@ -95,7 +101,6 @@ class AutoTransformers(NNOperator):
             self.model_name = 'sentence-transformers/' + model_name
         else:
             self.model_name = model_name
-        self.norm = norm
         self.checkpoint_path = checkpoint_path
 
         if self.model_name:
@@ -118,7 +123,7 @@ class AutoTransformers(NNOperator):
         else:
             txt = data
         try:
-            inputs = self.tokenizer(txt, padding=True, truncation=True, return_tensors='pt').to(self.device)
+            inputs = self.tokenizer(txt, padding=True, truncation=True, return_tensors='pt')
         except Exception as e:
             log.error(f'Fail to tokenize inputs: {e}')
             raise e
@@ -128,8 +133,6 @@ class AutoTransformers(NNOperator):
             log.error(f'Invalid input for the model: {self.model_name}')
             raise e
         outs = self.post_proc(outs, inputs)
-        if self.norm:
-            outs = torch.nn.functional.normalize(outs, )
         features = outs.cpu().detach().numpy()
         if isinstance(data, str):
             features = features.squeeze(0)
@@ -169,8 +172,8 @@ class AutoTransformers(NNOperator):
         return onnx_config
 
     def post_proc(self, token_embeddings, inputs):
-        token_embeddings = token_embeddings.to(self.device)
-        attention_mask = inputs['attention_mask'].to(self.device)
+        token_embeddings = token_embeddings
+        attention_mask = inputs['attention_mask']
         input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float()
         sentence_embs = torch.sum(
             token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9)
@@ -191,7 +194,7 @@ class AutoTransformers(NNOperator):
                 raise AttributeError('Unsupported model_type.')
 
         dummy_input = 'test sentence'
-        inputs = self.tokenizer(dummy_input, padding=True, truncation=True, return_tensors='pt').to(self.device)
+        inputs = self.tokenizer(dummy_input, padding=True, truncation=True, return_tensors='pt')
         if model_type == 'pytorch':
             torch.save(self._model, output_file)
         elif model_type == 'torchscript':
@@ -214,7 +217,7 @@ class AutoTransformers(NNOperator):
             if hasattr(self._model.config, 'use_cache'):
                 self._model.config.use_cache = False
             torch.onnx.export(
-                self._model,
+                self._model.to('cpu'),
                 tuple(inputs.values()),
                 output_file,
                 input_names=list(self.onnx_config['inputs'].keys()),