diff --git a/python/fate/components/components/homo_lr.py b/python/fate/components/components/homo_lr.py
index 7dc567cc05..618b74c1d0 100644
--- a/python/fate/components/components/homo_lr.py
+++ b/python/fate/components/components/homo_lr.py
@@ -13,16 +13,13 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 import logging
-import os
-
-import pandas as pd
 from fate.arch import Context
-from fate.arch.dataframe import PandasReader
 from fate.ml.glm.homo_lr.client import HomoLRClient
 from fate.ml.glm.homo_lr.server import HomoLRServer
-from fate.components.components.utils.predict_format import LABEL
 from fate.components.core import ARBITER, GUEST, HOST, Role, cpn, params
 from fate.components.components.utils import consts
+from fate.ml.utils.model_io import ModelIO
+
 
 logger = logging.getLogger(__name__)
 
@@ -59,25 +56,40 @@ def train(
                                 desc="Model param init setting."),
     threshold: cpn.parameter(type=params.confloat(ge=0.0, le=1.0), default=0.5,
                                 desc="predict threshold for binary data"),
+    ovr: cpn.parameter(type=bool, default=False,
+                                desc="predict threshold for binary data"),
+    label_num: cpn.parameter(type=params.conint(ge=2), default=None),
     train_output_data: cpn.dataframe_output(roles=[GUEST, HOST]),
     train_input_model: cpn.json_model_input(roles=[GUEST, HOST], optional=True),
-    output_model: cpn.json_model_output(roles=[GUEST, HOST])
+    train_output_model: cpn.json_model_output(roles=[GUEST, HOST])
 ):
 
     sub_ctx = ctx.sub_ctx(consts.TRAIN)
 
     if role.is_guest or role.is_host:  # is client
         
+        logger.info('homo lr component: client start training')
         logger.info('optim param {} init param {}'.format(optimizer.dict(), init_param.dict()))
 
-        client = HomoLRClient(epochs=epochs, batch_size=batch_size, optimizer_param=optimizer, init_param=init_param, 
-                              learning_rate_scheduler=0.01, threshold=threshold)
+        client = HomoLRClient(epochs=epochs, batch_size=batch_size, optimizer_param=optimizer.dict(), init_param=init_param.dict(), 
+                              learning_rate_scheduler=0.01, threshold=threshold, ovr=ovr, label_num=label_num)
         train_df = train_data.read()
         validate_df = validate_data.read() if validate_data else None
         client.fit(sub_ctx, train_df, validate_df)
-        output_model.write({"aaa": 1}, metadata={"bbb": 2})
+        model_dict = client.get_model().dict()
+        
+        train_rs = client.predict(sub_ctx, train_df)
+        if validate_df:
+            validate_rs = client.predict(sub_ctx, validate_df)
+            ret_df = train_rs.vstack(validate_rs)
+        else:
+            ret_df = train_rs
+
+        train_output_data.write(ret_df)
+        train_output_model.write(model_dict, metadata=model_dict['meta'])
+
     elif role.is_arbiter:  # is server
-        logger.info('hello')
+        logger.info('homo lr component: server start training')
         server = HomoLRServer()
         server.fit(sub_ctx)
 
@@ -97,7 +109,13 @@ def predict(
 ):
 
     if role.is_guest or role.is_host:  # is client
-        pass
+
+        client = HomoLRClient(batch_size=batch_size, threshold=threshold)
+        model_input = predict_input_model.read()
+        model_data = ModelIO.from_dict(model_input)
+        logger.info('model input is {}'.format(model_input))
+        pred_rs = client.predict(ctx, test_data.read())
+        
 
     elif role.is_arbiter:  # is server
         logger.info("arbiter skip predict")