judge model type only when compress_config is None (#1461)

* fix ptq hpo in ac

* judge model type only when compress_config is None

* update

* update demo

* revert compressor

README.md

@@ -182,6 +182,9 @@ git clone https://github.com/PaddlePaddle/PaddleSlim.git & cd PaddleSlim
 python setup.py install
 ```
 
+### 验证安装
+
+安装完成后您可以使用 python 或 python3 进入 python 解释器，输入import paddleslim, 没有报错则说明安装成功。
 
 ### 快速开始
 

example/auto_compression/README.md

@@ -220,21 +220,21 @@ ac.compress()
   - 测试FP32模型的速度
 
     ```
-    python ./image_classification/infer.py
+    python ./image_classification/paddle_inference_eval.py --model_path='./MobileNetV1_infer' --use_gpu=True --use_trt=True
     ### using tensorrt FP32    batch size: 1 time(ms): 0.6140608787536621
     ```
 
   - 测试FP16模型的速度
 
     ```
-    python ./image_classification/infer.py --use_fp16=True
+    python ./image_classification/paddle_inference_eval.py --model_path='./MobileNetV1_infer' --use_gpu=True --use_trt=True --use_fp16=True
     ### using tensorrt FP16    batch size: 1 time(ms): 0.5795984268188477
     ```
 
   - 测试INT8模型的速度
 
     ```
-    python ./image_classification/infer.py --model_dir=./MobileNetV1_quant/ --use_int8=True
+    python ./image_classification/paddle_inference_eval.py --model_path='./MobileNetV1_quant/' --use_gpu=True --use_trt=True --use_int8=True
     ### using tensorrt INT8 batch size: 1 time(ms): 0.5213963985443115
     ```
 

example/auto_compression/image_classification/configs/eval.yaml

-model_dir: './MobileNetV1_infer'
-model_filename: 'inference.pdmodel'
-params_filename: "inference.pdiparams"
-batch_size: 128
-data_dir: './ILSVRC2012_data_demo/ILSVRC2012/'
-img_size: 224
-resize_size: 256
+Global:
+  model_dir: './MobileNetV1_infer'
+  model_filename: 'inference.pdmodel'
+  params_filename: "inference.pdiparams"
+  batch_size: 128
+  data_dir: './ILSVRC2012_data_demo/ILSVRC2012/'
+  img_size: 224
+  resize_size: 256

example/auto_compression/image_classification/paddle_inference_eval.py

@@ -172,9 +172,9 @@ class Predictor(object):
                 crop_size=args.img_size,
                 resize_size=args.resize_size)
         else:
-            image = np.ones(
-                (1, 3, args.img_size, args.img_size)).astype(np.float32)
-            label = None
+            image = np.ones((args.batch_size, 3, args.img_size,
+                             args.img_size)).astype(np.float32)
+            label = [[None]] * args.batch_size
             val_loader = [[image, label]]
         results = []
         input_names = self.paddle_predictor.get_input_names()