Fix bug of seg demo (#1449)

example/auto_compression/semantic_segmentation/configs/deeplabv3/deeplabv3_qat.yaml

 Global:
-  reader_config: configs/dataset/cityscapes_1024x512_scale1.0.yml
+  reader_config: configs/deeplabv3/deeplabv3_reader.yml
   model_dir: ./RES-paddle2-Deeplabv3-ResNet50
   model_filename: model
   params_filename: params
@@ -11,15 +11,16 @@ Distillation:
   - conv2d_123.tmp_1
 
 Quantization:
+  onnx_format: True
   quantize_op_types:
   - conv2d
   - depthwise_conv2d
 
 TrainConfig:
-  epochs: 10
+  epochs: 1
   eval_iter: 360
   learning_rate: 0.0001
   optimizer_builder:
     optimizer: 
       type: SGD
-    weight_decay: 0.0005
+    weight_decay: 0.0005
\ No newline at end of file

example/auto_compression/semantic_segmentation/configs/deeplabv3/deeplabv3_reader.yml

+batch_size: 4
+train_dataset:
+  type: Cityscapes
+  dataset_root: data/cityscapes
+  transforms:
+    - type: ResizeStepScaling
+      min_scale_factor: 0.5
+      max_scale_factor: 2.0
+      scale_step_size: 0.25
+    - type: RandomPaddingCrop
+      crop_size: [1024, 512]
+    - type: RandomHorizontalFlip
+    - type: RandomDistort
+      brightness_range: 0.5
+      contrast_range: 0.5
+      saturation_range: 0.5
+    - type: Normalize
+  mode: train
+val_dataset:
+  type: Cityscapes
+  dataset_root: data/cityscapes
+  transforms:
+    - type: Normalize
+  mode: val

example/auto_compression/semantic_segmentation/run.py

@@ -137,6 +137,7 @@ def main(args):
     # step1: load dataset config and create dataloader
     data_cfg = PaddleSegDataConfig(config['reader_config'])
     train_dataset = data_cfg.train_dataset
+    global eval_dataset
     eval_dataset = data_cfg.val_dataset
     batch_sampler = paddle.io.DistributedBatchSampler(
         train_dataset,
@@ -163,7 +164,7 @@ def main(args):
         save_dir=args.save_dir,
         config=all_config,
         train_dataloader=train_dataloader,
-        eval_callback=eval_function if nranks > 1 and rank_id != 0 else None,
+        eval_callback=eval_function if rank_id == 0 else None,
         deploy_hardware=config.get('deploy_hardware') or None,
         input_shapes=config.get('input_shapes', None))