Fix evaluator in parallel mode (#923)

fluid/ocr_recognition/ctc_train.py

@@ -71,6 +71,7 @@ def train(args, data_reader=ctc_reader):
         print "Init model from: %s." % args.init_model
 
     train_exe = exe
+    error_evaluator.reset(exe)
     if args.parallel:
         train_exe = fluid.ParallelExecutor(
             use_cuda=True, loss_name=sum_cost.name)
@@ -81,7 +82,7 @@ def train(args, data_reader=ctc_reader):
         var_names = [var.name for var in fetch_vars]
         if args.parallel:
             results = train_exe.run(var_names,
-                                    feed_dict=get_feeder_data(data, place))
+                                    feed=get_feeder_data(data, place))
             results = [np.array(result).sum() for result in results]
         else:
             results = exe.run(feed=get_feeder_data(data, place),
@@ -103,7 +104,6 @@ def train(args, data_reader=ctc_reader):
             exe, dirname=args.save_model_dir, filename=filename)
         print "Saved model to: %s/%s." % (args.save_model_dir, filename)
 
-    error_evaluator.reset(exe)
     for pass_id in range(args.pass_num):
         batch_id = 1
         total_loss = 0.0