diff --git a/demo/semantic_role_labeling/api_train_v2.py b/demo/semantic_role_labeling/api_train_v2.py
new file mode 100644
index 0000000000000000000000000000000000000000..e946a792b5f51ab54355aac0b6e9aef51ae815fb
--- /dev/null
+++ b/demo/semantic_role_labeling/api_train_v2.py
@@ -0,0 +1,175 @@
+import sys
+import math
+import numpy as np
+import paddle.v2 as paddle
+import paddle.v2.dataset.conll05 as conll05
+
+
+def db_lstm():
+    word_dict, verb_dict, label_dict = conll05.get_dict()
+    word_dict_len = len(word_dict)
+    label_dict_len = len(label_dict)
+    pred_len = len(verb_dict)
+
+    mark_dict_len = 2
+    word_dim = 32
+    mark_dim = 5
+    hidden_dim = 512
+    depth = 8
+
+    #8 features
+    def d_type(size):
+        return paddle.data_type.integer_value_sequence(size)
+
+    word = paddle.layer.data(name='word_data', type=d_type(word_dict_len))
+    predicate = paddle.layer.data(name='verb_data', type=d_type(pred_len))
+
+    ctx_n2 = paddle.layer.data(name='ctx_n2_data', type=d_type(word_dict_len))
+    ctx_n1 = paddle.layer.data(name='ctx_n1_data', type=d_type(word_dict_len))
+    ctx_0 = paddle.layer.data(name='ctx_0_data', type=d_type(word_dict_len))
+    ctx_p1 = paddle.layer.data(name='ctx_p1_data', type=d_type(word_dict_len))
+    ctx_p2 = paddle.layer.data(name='ctx_p2_data', type=d_type(word_dict_len))
+    mark = paddle.layer.data(name='mark_data', type=d_type(mark_dict_len))
+
+    target = paddle.layer.data(name='target', type=d_type(label_dict_len))
+
+    default_std = 1 / math.sqrt(hidden_dim) / 3.0
+
+    emb_para = paddle.attr.Param(name='emb', initial_std=0., learning_rate=0.)
+    std_0 = paddle.attr.Param(initial_std=0.)
+    std_default = paddle.attr.Param(initial_std=default_std)
+
+    predicate_embedding = paddle.layer.embedding(
+        size=word_dim,
+        input=predicate,
+        param_attr=paddle.attr.Param(
+            name='vemb', initial_std=default_std))
+    mark_embedding = paddle.layer.embedding(
+        size=mark_dim, input=mark, param_attr=std_0)
+
+    word_input = [word, ctx_n2, ctx_n1, ctx_0, ctx_p1, ctx_p2]
+    emb_layers = [
+        paddle.layer.embedding(
+            size=word_dim, input=x, param_attr=emb_para) for x in word_input
+    ]
+    emb_layers.append(predicate_embedding)
+    emb_layers.append(mark_embedding)
+
+    hidden_0 = paddle.layer.mixed(
+        size=hidden_dim,
+        bias_attr=std_default,
+        input=[
+            paddle.layer.full_matrix_projection(
+                input=emb, param_attr=std_default) for emb in emb_layers
+        ])
+
+    mix_hidden_lr = 1e-3
+    lstm_para_attr = paddle.attr.Param(initial_std=0.0, learning_rate=1.0)
+    hidden_para_attr = paddle.attr.Param(
+        initial_std=default_std, learning_rate=mix_hidden_lr)
+
+    lstm_0 = paddle.layer.lstmemory(
+        input=hidden_0,
+        act=paddle.activation.Relu(),
+        gate_act=paddle.activation.Sigmoid(),
+        state_act=paddle.activation.Sigmoid(),
+        bias_attr=std_0,
+        param_attr=lstm_para_attr)
+
+    #stack L-LSTM and R-LSTM with direct edges
+    input_tmp = [hidden_0, lstm_0]
+
+    for i in range(1, depth):
+        mix_hidden = paddle.layer.mixed(
+            size=hidden_dim,
+            bias_attr=std_default,
+            input=[
+                paddle.layer.full_matrix_projection(
+                    input=input_tmp[0], param_attr=hidden_para_attr),
+                paddle.layer.full_matrix_projection(
+                    input=input_tmp[1], param_attr=lstm_para_attr)
+            ])
+
+        lstm = paddle.layer.lstmemory(
+            input=mix_hidden,
+            act=paddle.activation.Relu(),
+            gate_act=paddle.activation.Sigmoid(),
+            state_act=paddle.activation.Sigmoid(),
+            reverse=((i % 2) == 1),
+            bias_attr=std_0,
+            param_attr=lstm_para_attr)
+
+        input_tmp = [mix_hidden, lstm]
+
+    feature_out = paddle.layer.mixed(
+        size=label_dict_len,
+        bias_attr=std_default,
+        input=[
+            paddle.layer.full_matrix_projection(
+                input=input_tmp[0], param_attr=hidden_para_attr),
+            paddle.layer.full_matrix_projection(
+                input=input_tmp[1], param_attr=lstm_para_attr)
+        ], )
+
+    crf_cost = paddle.layer.crf(size=label_dict_len,
+                                input=feature_out,
+                                label=target,
+                                param_attr=paddle.attr.Param(
+                                    name='crfw',
+                                    initial_std=default_std,
+                                    learning_rate=mix_hidden_lr))
+
+    crf_dec = paddle.layer.crf_decoding(
+        name='crf_dec_l',
+        size=label_dict_len,
+        input=feature_out,
+        label=target,
+        param_attr=paddle.attr.Param(name='crfw'))
+
+    return crf_cost, crf_dec
+
+
+def load_parameter(file_name, h, w):
+    with open(file_name, 'rb') as f:
+        f.read(16)  # skip header.
+        return np.fromfile(f, dtype=np.float32).reshape(h, w)
+
+
+def main():
+    paddle.init(use_gpu=False, trainer_count=1)
+
+    # define network topology
+    crf_cost, crf_dec = db_lstm()
+
+    # create parameters
+    parameters = paddle.parameters.create([crf_cost, crf_dec])
+
+    # create optimizer
+    optimizer = paddle.optimizer.Momentum(
+        momentum=0,
+        learning_rate=2e-2,
+        regularization=paddle.optimizer.L2Regularization(rate=8e-4),
+        model_average=paddle.optimizer.ModelAverage(
+            average_window=0.5, max_average_window=10000), )
+
+    def event_handler(event):
+        if isinstance(event, paddle.event.EndIteration):
+            if event.batch_id % 100 == 0:
+                print "Pass %d, Batch %d, Cost %f, %s" % (
+                    event.pass_id, event.batch_id, event.cost, event.metrics)
+
+    trainer = paddle.trainer.SGD(cost=crf_cost,
+                                 parameters=parameters,
+                                 update_equation=optimizer)
+    parameters.set('emb', load_parameter(conll05.get_embedding(), 44068, 32))
+
+    trn_reader = paddle.reader.batched(
+        paddle.reader.shuffle(
+            conll05.test(), buf_size=8192), batch_size=10)
+
+    trainer.train(
+        reader=trn_reader, event_handler=event_handler, num_passes=10000)
+
+
+if __name__ == '__main__':
+    main()
diff --git a/python/paddle/v2/dataset/__init__.py b/python/paddle/v2/dataset/__init__.py
index 7c999752af22a58aca4ee9a504a909d97c53498d..a1b21bab3bac8b304abb4ae292b1c1e9f3e719de 100644
--- a/python/paddle/v2/dataset/__init__.py
+++ b/python/paddle/v2/dataset/__init__.py
@@ -17,5 +17,6 @@ import imikolov
 import imdb
 import cifar
 import movielens
+import conll05
 
-__all__ = ['mnist', 'imikolov', 'imdb', 'cifar', 'movielens']
+__all__ = ['mnist', 'imikolov', 'imdb', 'cifar', 'movielens', 'conll05']
diff --git a/python/paddle/v2/dataset/conll05.py b/python/paddle/v2/dataset/conll05.py
index 7874161a05968921a8e6789a7ff6a229d6cc6c01..e96a701c1a944e2d6d84f897157cb357c5aa0824 100644
--- a/python/paddle/v2/dataset/conll05.py
+++ b/python/paddle/v2/dataset/conll05.py
@@ -12,10 +12,10 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import paddle.v2.dataset.common
 import tarfile
 import gzip
 import itertools
+from common import download
 
 __all__ = ['test, get_dict', 'get_embedding']
 """
@@ -160,7 +160,6 @@ def reader_creator(corpus_reader,
                 ctx_p2 = 'eos'
 
             word_idx = [word_dict.get(w, UNK_IDX) for w in sentence]
-            pred_idx = [predicate_dict.get(predicate)] * sen_len
 
             ctx_n2_idx = [word_dict.get(ctx_n2, UNK_IDX)] * sen_len
             ctx_n1_idx = [word_dict.get(ctx_n1, UNK_IDX)] * sen_len
@@ -168,38 +167,30 @@ def reader_creator(corpus_reader,
             ctx_p1_idx = [word_dict.get(ctx_p1, UNK_IDX)] * sen_len
             ctx_p2_idx = [word_dict.get(ctx_p2, UNK_IDX)] * sen_len
 
+            pred_idx = [predicate_dict.get(predicate)] * sen_len
             label_idx = [label_dict.get(w) for w in labels]
 
-            yield word_idx, pred_idx, ctx_n2_idx, ctx_n1_idx, \
-              ctx_0_idx, ctx_p1_idx, ctx_p2_idx, mark, label_idx
+            yield word_idx, ctx_n2_idx, ctx_n1_idx, \
+              ctx_0_idx, ctx_p1_idx, ctx_p2_idx, pred_idx, mark, label_idx
 
-    return reader()
+    return reader
 
 
 def get_dict():
-    word_dict = load_dict(
-        common.download(WORDDICT_URL, 'conll05st', WORDDICT_MD5))
-    verb_dict = load_dict(
-        common.download(VERBDICT_URL, 'conll05st', VERBDICT_MD5))
-    label_dict = load_dict(
-        common.download(TRGDICT_URL, 'conll05st', TRGDICT_MD5))
+    word_dict = load_dict(download(WORDDICT_URL, 'conll05st', WORDDICT_MD5))
+    verb_dict = load_dict(download(VERBDICT_URL, 'conll05st', VERBDICT_MD5))
+    label_dict = load_dict(download(TRGDICT_URL, 'conll05st', TRGDICT_MD5))
     return word_dict, verb_dict, label_dict
 
 
 def get_embedding():
-    return common.download(EMB_URL, 'conll05st', EMB_MD5)
+    return download(EMB_URL, 'conll05st', EMB_MD5)
 
 
 def test():
     word_dict, verb_dict, label_dict = get_dict()
     reader = corpus_reader(
-        common.download(DATA_URL, 'conll05st', DATA_MD5),
+        download(DATA_URL, 'conll05st', DATA_MD5),
         words_name='conll05st-release/test.wsj/words/test.wsj.words.gz',
         props_name='conll05st-release/test.wsj/props/test.wsj.props.gz')
     return reader_creator(reader, word_dict, verb_dict, label_dict)
-
-
-if __name__ == '__main__':
-    print get_embedding()
-    for f in test():
-        print f