Adapt to api updates. (PaddlePaddle#1629)

* Adapt to api updates.

* fix bugs.

example/auto_compression/detection/paddle_inference_eval.py

@@ -399,7 +399,8 @@ def eval(predictor, val_loader, metric, rerun_flag=False):
     input_names = predictor.get_input_names()
     output_names = predictor.get_output_names()
     boxes_tensor = predictor.get_output_handle(output_names[0])
-    boxes_num = predictor.get_output_handle(output_names[1])
+    if FLAGS.include_nms:
+        boxes_num = predictor.get_output_handle(output_names[1])
     for batch_id, data in enumerate(val_loader):
         data_all = {k: np.array(v) for k, v in data.items()}
         for i, _ in enumerate(input_names):

example/auto_compression/detection/run.py

@@ -139,9 +139,10 @@ def main():
     train_loader = create('EvalReader')(reader_cfg['TrainDataset'],
                                         reader_cfg['worker_num'],
                                         return_list=True)
-    global_config['input_list'] = get_feed_vars(
-        global_config['model_dir'], global_config['model_filename'],
-        global_config['params_filename'])
+    if global_config.get('input_list') is None:
+        global_config['input_list'] = get_feed_vars(
+            global_config['model_dir'], global_config['model_filename'],
+            global_config['params_filename'])
     train_loader = reader_wrapper(train_loader, global_config['input_list'])
 
     if 'Evaluation' in global_config.keys() and global_config[

paddleslim/analysis/_utils.py

@@ -139,7 +139,8 @@ def save_cls_model(model, input_shape, save_dir, data_type):
             batch_nums=1,
             weight_bits=8,
             activation_bits=8,
-            quantizable_op_type=["conv2d", "depthwise_conv2d"])
+            quantizable_op_type=["conv2d", "depthwise_conv2d"],
+            onnx_format=False)
 
         model_file = os.path.join(quantize_model_path, 'model.pdmodel')
         param_file = os.path.join(quantize_model_path, 'model.pdiparams')