Ocr use new API (#4290)

* update new API for ocr
* fix the code style

PaddleCV/ocr_recognition/attention_model.py

@@ -24,6 +24,7 @@ sos = 0
 eos = 1
 beam_size = 1
 
+
 def conv_bn_pool(input,
                  group,
                  out_ch,
@@ -164,12 +165,13 @@ def gru_decoder_with_attention(target_embedding, encoder_vec, encoder_proj,
 
 
 def attention_train_net(args, data_shape, num_classes):
-
-    images = fluid.layers.data(name='pixel', shape=data_shape, dtype='float32')
-    label_in = fluid.layers.data(
-        name='label_in', shape=[1], dtype='int32', lod_level=1)
-    label_out = fluid.layers.data(
-        name='label_out', shape=[1], dtype='int32', lod_level=1)
+    if len(list(data_shape)) == 3:
+        data_shape = [None] + list(data_shape)
+    images = fluid.data(name='pixel', shape=data_shape, dtype='float32')
+    label_in = fluid.data(
+        name='label_in', shape=[None, 1], dtype='int32', lod_level=1)
+    label_out = fluid.data(
+        name='label_out', shape=[None, 1], dtype='int32', lod_level=1)
 
     gru_backward, encoded_vector, encoded_proj = encoder_net(images)
 
@@ -188,7 +190,8 @@ def attention_train_net(args, data_shape, num_classes):
     prediction = gru_decoder_with_attention(trg_embedding, encoded_vector,
                                             encoded_proj, decoder_boot,
                                             decoder_size, num_classes)
-    fluid.clip.set_gradient_clip(fluid.clip.GradientClipByGlobalNorm(args.gradient_clip))
+    fluid.clip.set_gradient_clip(
+        fluid.clip.GradientClipByGlobalNorm(args.gradient_clip))
     label_out = fluid.layers.cast(x=label_out, dtype='int64')
 
     _, maxid = fluid.layers.topk(input=prediction, k=1)
@@ -264,10 +267,10 @@ def attention_infer(images, num_classes, use_cudnn=True):
     ids_array = fluid.layers.create_array('int64')
     scores_array = fluid.layers.create_array('float32')
 
-    init_ids = fluid.layers.data(
-        name="init_ids", shape=[1], dtype="int64", lod_level=2)
-    init_scores = fluid.layers.data(
-        name="init_scores", shape=[1], dtype="float32", lod_level=2)
+    init_ids = fluid.data(
+        name="init_ids", shape=[None, 1], dtype="int64", lod_level=2)
+    init_scores = fluid.data(
+        name="init_scores", shape=[None, 1], dtype="float32", lod_level=2)
 
     fluid.layers.array_write(init_ids, array=ids_array, i=counter)
     fluid.layers.array_write(init_scores, array=scores_array, i=counter)
@@ -349,11 +352,13 @@ def attention_infer(images, num_classes, use_cudnn=True):
 
 
 def attention_eval(data_shape, num_classes, use_cudnn=True):
-    images = fluid.layers.data(name='pixel', shape=data_shape, dtype='float32')
-    label_in = fluid.layers.data(
-        name='label_in', shape=[1], dtype='int32', lod_level=1)
-    label_out = fluid.layers.data(
-        name='label_out', shape=[1], dtype='int32', lod_level=1)
+    if len(list(data_shape)) == 3:
+        data_shape = [None] + data_shape
+    images = fluid.data(name='pixel', shape=data_shape, dtype='float32')
+    label_in = fluid.data(
+        name='label_in', shape=[None, 1], dtype='int32', lod_level=1)
+    label_out = fluid.data(
+        name='label_out', shape=[None, 1], dtype='int32', lod_level=1)
     label_out = fluid.layers.cast(x=label_out, dtype='int64')
     label_in = fluid.layers.cast(x=label_in, dtype='int64')
 

PaddleCV/ocr_recognition/crnn_ctc_model.py

@@ -188,10 +188,11 @@ def ctc_train_net(args, data_shape, num_classes):
     MOMENTUM = args.momentum
     learning_rate_decay = None
     regularizer = fluid.regularizer.L2Decay(L2_RATE)
-
-    images = fluid.layers.data(name='pixel', shape=data_shape, dtype='float32')
-    label = fluid.layers.data(
-        name='label', shape=[1], dtype='int32', lod_level=1)
+    if len(list(data_shape)) == 3:
+        data_shape = [None] + list(data_shape)
+    images = fluid.data(name='pixel', shape=data_shape, dtype='float32')
+    label = fluid.data(
+        name='label', shape=[None, 1], dtype='int32', lod_level=1)
     fc_out = encoder_net(
         images,
         num_classes,
@@ -231,9 +232,11 @@ def ctc_infer(images, num_classes, use_cudnn=True):
 
 
 def ctc_eval(data_shape, num_classes, use_cudnn=True):
-    images = fluid.layers.data(name='pixel', shape=data_shape, dtype='float32')
-    label = fluid.layers.data(
-        name='label', shape=[1], dtype='int32', lod_level=1)
+    if len(list(data_shape)) == 3:
+        data_shape = [None] + list(data_shape)
+    images = fluid.data(name='pixel', shape=data_shape, dtype='float32')
+    label = fluid.data(
+        name='label', shape=[None, 1], dtype='int32', lod_level=1)
     fc_out = encoder_net(images, num_classes, is_test=True, use_cudnn=use_cudnn)
     decoded_out = fluid.layers.ctc_greedy_decoder(
         input=fc_out, blank=num_classes)

PaddleCV/ocr_recognition/data_reader.py

@@ -32,7 +32,8 @@ except NameError:
 SOS = 0
 EOS = 1
 NUM_CLASSES = 95
-DATA_SHAPE = [1, 48, 512]
+IMG_WIDTH = 384
+DATA_SHAPE = [1, 48, IMG_WIDTH]
 
 DATA_MD5 = "7256b1d5420d8c3e74815196e58cdad5"
 DATA_URL = "http://paddle-ocr-data.bj.bcebos.com/data.tar.gz"

PaddleCV/ocr_recognition/eval.py

@@ -64,11 +64,11 @@ def evaluate(args):
 
     # load init model
     model_dir = args.model_path
-    model_file_name = None
-    if not os.path.isdir(args.model_path):
-        model_dir = os.path.dirname(args.model_path)
-        model_file_name = os.path.basename(args.model_path)
-    fluid.io.load_params(exe, dirname=model_dir, filename=model_file_name)
+    if os.path.isdir(args.model_path):
+        raise Exception("{} should not be a directory".format(args.model_path))
+    fluid.load(program=fluid.default_main_program(),
+               model_path=model_dir,
+               executor=exe)
     print("Init model from: %s." % args.model_path)
 
     evaluator.reset(exe)

PaddleCV/ocr_recognition/infer.py

@@ -54,7 +54,9 @@ def inference(args):
     num_classes = data_reader.num_classes()
     data_shape = data_reader.data_shape()
     # define network
-    images = fluid.layers.data(name='pixel', shape=data_shape, dtype='float32')
+    if len(list(data_shape)) == 3:
+        data_shape = [None] + list(data_shape)
+    images = fluid.data(name='pixel', shape=data_shape, dtype='float32')
     ids = infer(images, num_classes, use_cudnn=True if args.use_gpu else False)
     # data reader
     infer_reader = data_reader.inference(
@@ -82,11 +84,10 @@ def inference(args):
 
     # load init model
     model_dir = args.model_path
-    model_file_name = None
-    if not os.path.isdir(args.model_path):
-        model_dir = os.path.dirname(args.model_path)
-        model_file_name = os.path.basename(args.model_path)
-    fluid.io.load_params(exe, dirname=model_dir, filename=model_file_name)
+    fluid.load(
+        program=fluid.default_main_program(),
+        model_path=model_dir,
+        executor=exe)
     print("Init model from: %s." % args.model_path)
 
     batch_times = []

PaddleCV/ocr_recognition/train.py

@@ -106,11 +106,7 @@ def train(args):
     # load init model
     if args.init_model is not None:
         model_dir = args.init_model
-        model_file_name = None
-        if not os.path.isdir(args.init_model):
-            model_dir = os.path.dirname(args.init_model)
-            model_file_name = os.path.basename(args.init_model)
-        fluid.io.load_params(exe, dirname=model_dir, filename=model_file_name)
+        fluid.load(fluid.default_main_program(), model_dir)
         print("Init model from: %s." % args.init_model)
 
     train_exe = exe
@@ -148,8 +144,8 @@ def train(args):
 
     def save_model(args, exe, iter_num):
         filename = "model_%05d" % iter_num
-        fluid.io.save_params(
-            exe, dirname=args.save_model_dir, filename=filename)
+        fluid.save(fluid.default_main_program(),
+                   os.path.join(args.save_model_dir, filename))
         print("Saved model to: %s/%s." % (args.save_model_dir, filename))
 
     iter_num = 0