Support TritonServe

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

test_onnx.py

@@ -11,8 +11,8 @@ import logging
 import platform
 import psutil
 
-models = TimmImage.supported_model_names()[:2]
-# models = ['resnet50']
+# models = TimmImage.supported_model_names()[:2]
+models = ['resnet50']
 
 atol = 1e-3
 log_path = 'timm_onnx.log'
@@ -57,7 +57,7 @@ for name in models:
     status = [name] + ['fail'] * 5
 
     try:
-        out1 = op.model.forward_features(data).detach().numpy()
+        out1 = op.accelerate_model(data).detach().numpy()
         logger.info('OP LOADED.')
         status[1] = 'success'
     except Exception as e:

timm_image.py

@@ -23,6 +23,7 @@ from towhee.operator.base import NNOperator, OperatorFlag
 from towhee.types.arg import arg, to_image_color
 from towhee import register
 from towhee.types import Image
+from towhee.dc2 import accelerate
 
 import torch
 from torch import nn
@@ -40,6 +41,17 @@ warnings.filterwarnings('ignore')
 log = logging.getLogger('timm_op')
 
 
+@accelerate
+class Model:
+    def __init__(self, model_name, device, num_classes):
+        self.model = create_model(model_name, pretrained=True, num_classes=num_classes)
+        self.model.eval()
+        self.model.to(device)
+
+    def __call__(self, x: torch.Tensor):
+        return self.model.forward_features(x)
+
+
 @register(output_schema=['vec'])
 class TimmImage(NNOperator):
     """
@@ -65,10 +77,12 @@ class TimmImage(NNOperator):
         self.device = device
         self.model_name = model_name
         if self.model_name:
-            self.model = create_model(self.model_name, pretrained=True, num_classes=num_classes)
-            self.model.eval()
-            self.model.to(self.device)
-
+            self.accelerate_model = Model(
+                model_name=model_name,
+                device=self.device,
+                num_classes=num_classes
+            )
+            self.model = self.accelerate_model.model
             self.config = resolve_data_config({}, model=self.model)
             self.tfms = create_transform(**self.config)
             self.skip_tfms = skip_preprocess
@@ -88,7 +102,7 @@ class TimmImage(NNOperator):
             img_list.append(img)
         inputs = torch.stack(img_list)
         inputs = inputs.to(self.device)
-        features = self.model.forward_features(inputs)
+        features = self.accelerate_model(inputs)
         if features.dim() == 4:
             global_pool = nn.AdaptiveAvgPool2d(1).to(self.device)
             features = global_pool(features)
@@ -111,12 +125,16 @@ class TimmImage(NNOperator):
             os.makedirs(path, exist_ok=True)
             name = self.model_name.replace('/', '-')
             path = os.path.join(path, name)
+            if format in ['pytorch', 'torchscript']:
+                path = path + '.pt'
+            elif format == 'onnx':
+                path = path + '.onnx'
+            else:
+                raise AttributeError(f'Invalid format {format}.')
         dummy_input = torch.rand((1,) + self.config['input_size'])
         if format == 'pytorch':
-            path = path + '.pt'
             torch.save(self.model, path)
         elif format == 'torchscript':
-            path = path + '.pt'
             try:
                 try:
                     jit_model = torch.jit.script(self.model)
@@ -127,7 +145,6 @@ class TimmImage(NNOperator):
                 log.error(f'Fail to save as torchscript: {e}.')
                 raise RuntimeError(f'Fail to save as torchscript: {e}.')
         elif format == 'onnx':
-            path = path + '.onnx'
             self.model.forward = self.model.forward_features
             try:
                 torch.onnx.export(self.model,
@@ -227,11 +244,9 @@ class TimmImage(NNOperator):
             log.error(f'Invalid format "{format}". Currently supported formats: "pytorch".')
         return model_list
 
-    def input_schema(self):
-        return [(Image, (-1, -1, 3))]
-
-    def output_schema(self):
-        image = Image(numpy.random.randn(480, 480, 3), "RGB")
-        ret = self(image)
-        data_type = type(ret.reshape(-1)[0])
-        return [(data_type, ret.shape)]
+    @property
+    def supported_formats(self):
+        if self.model_name in self.supported_model_names(format='onnx'):
+            return ['onnx']
+        else:
+            return []