Accommodate UAPI

ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x1_0.yaml

@@ -30,6 +30,7 @@ AMP:
 Arch:
   name: MobileNetV3_small_x1_0
   class_num: 1000
+  pretrained: True
  
 # loss function config for traing/eval process
 Loss:

ppcls/configs/ImageNet/PPHGNet/PPHGNet_small.yaml

@@ -31,6 +31,7 @@ AMP:
 Arch:
   name: PPHGNet_small
   class_num: 1000
+  pretrained: True
  
 # loss function config for traing/eval process
 Loss:

ppcls/configs/ImageNet/PPHGNet/PPHGNet_tiny.yaml

@@ -31,6 +31,7 @@ AMP:
 Arch:
   name: PPHGNet_tiny
   class_num: 1000
+  pretrained: True
  
 # loss function config for traing/eval process
 Loss:

ppcls/configs/ImageNet/PPLCNet/PPLCNet_x1_0.yaml

@@ -29,6 +29,7 @@ AMP:
 Arch:
   name: PPLCNet_x1_0
   class_num: 1000
+  pretrained: True
  
 # loss function config for traing/eval process
 Loss:

ppcls/configs/ImageNet/ResNet/ResNet50.yaml

@@ -32,6 +32,7 @@ AMP:
 Arch:
   name: ResNet50
   class_num: 1000
+  pretrained: True
  
 # loss function config for traing/eval process
 Loss:

ppcls/configs/ImageNet/SwinTransformer/SwinTransformer_base_patch4_window7_224.yaml

@@ -32,6 +32,7 @@ AMP:
 Arch:
   name: SwinTransformer_base_patch4_window7_224
   class_num: 1000
+  pretrained: True
  
 # loss function config for traing/eval process
 Loss:

ppcls/configs/slim/ClasModels_general_quantization.yaml

+# for quantizaiton or prune model
+Slim:
+  ## for prune
+  quant:
+    name: pact
\ No newline at end of file

ppcls/engine/engine.py

@@ -15,6 +15,7 @@ from __future__ import division
 from __future__ import print_function
 
 import os
+import shutil
 import platform
 import paddle
 import paddle.distributed as dist
@@ -72,8 +73,7 @@ class Engine(object):
 
         # init logger
         self.output_dir = self.config['Global']['output_dir']
-        log_file = os.path.join(self.output_dir, self.config["Arch"]["name"],
-                                f"{mode}.log")
+        log_file = os.path.join(self.output_dir, f"{mode}.log")
         init_logger(log_file=log_file)
         print_config(config)
 
@@ -519,6 +519,10 @@ class Engine(object):
                                                           save_path + "_int8")
         else:
             paddle.jit.save(model, save_path)
+        if self.config["Global"].get("export_for_fd", False):
+            src_path = self.config["Global"]["infer_config_path"]
+            dst_path = os.path.join(self.config["Global"]["save_inference_dir"], 'inference.yml')
+            shutil.copy(src_path, dst_path)
         logger.info(
             f"Export succeeded! The inference model exported has been saved in \"{self.config['Global']['save_inference_dir']}\"."
         )

ppcls/utils/save_load.py

@@ -163,7 +163,11 @@ def save_model(net,
     """
     if paddle.distributed.get_rank() != 0:
         return
-    model_path = os.path.join(model_path, model_name)
+
+    if prefix == 'best_model':
+        uapi_best_model_path = os.path.join(model_path, 'best_model')
+        _mkdir_if_not_exist(uapi_best_model_path)
+
     _mkdir_if_not_exist(model_path)
     model_path = os.path.join(model_path, prefix)
 
@@ -182,6 +186,11 @@ def save_model(net,
             paddle.save(s_params, model_path + "_student.pdparams")
 
     paddle.save(params_state_dict, model_path + ".pdparams")
+
+    if prefix == 'best_model':
+        uapi_best_model_path = os.path.join(uapi_best_model_path, 'model')
+        paddle.save(params_state_dict, uapi_best_model_path + ".pdparams")
+
     if ema is not None:
         paddle.save(ema.state_dict(), model_path + ".ema.pdparams")
     paddle.save([opt.state_dict() for opt in optimizer], model_path + ".pdopt")