make acclerate avaliable.

Signed-off-by: wxywb <xy.wang@zilliz.com>

blip.py

@@ -30,7 +30,11 @@ from transformers.models.blip.modeling_blip import BlipOutput, blip_loss
 from towhee import register
 from towhee.operator.base import NNOperator, OperatorFlag
 from towhee.types.arg import arg, to_image_color
-#from towhee.dc2 import accelerate
+try:
+    from towhee import accelerate
+except:
+    def accelerate(func):
+        return func
 
 log = logging.getLogger('run_op')
 warnings.filterwarnings('ignore')
@@ -124,7 +128,7 @@ def create_model(cfg, modality, checkpoint_path, device):
         raise ValueError("modality[{}] not implemented.".format(modality))
     return blip
 
-#@accelerate
+@accelerate
 class BLIPModelVision(nn.Module):
     def __init__(self, model):
         super().__init__()
@@ -135,7 +139,7 @@ class BLIPModelVision(nn.Module):
         image_embeds = self.backbone.vision_proj(image_embeds[:,0,:])
         return image_embeds
     
-#@accelerate
+@accelerate
 class BLIPModelText(nn.Module):
     def __init__(self, model):
         super().__init__()
@@ -193,14 +197,12 @@ class Blip(NNOperator):
 
     def _inference_from_text(self, text):
         inputs = self.processor(text=text, padding=True, return_tensors='pt')
-        inputs = inputs.to(self.device)
         text_feature = self.model(input_ids = inputs.input_ids, attention_mask = inputs.attention_mask)[0]
         return text_feature
 
     @arg(1, to_image_color('RGB'))
     def _inference_from_image(self, img):
         inputs = self.processor(images=img, return_tensors='pt')
-        inputs = inputs.to(self.device)
         image_feature = self.model(inputs['pixel_values'])
         return image_feature