diff --git a/README.md b/README.md
index 0403975..2265025 100644
--- a/README.md
+++ b/README.md
@@ -12,7 +12,7 @@
 
 生成句向量不需要做fine tune，使用预先训练好的模型即可，可参考`extract_feature.py`的`main`方法，注意参数必须是一个list。
 
-第一次生成句向量时需要加载graph，速度比较慢，后续速度会很快
+首次生成句向量时需要加载graph，并在output_dir路径下生成一个新的graph文件，因此速度比较慢，再次调用速度会很快
 ```
 from bert.extrac_feature import BertVector
 bv = BertVector()
diff --git a/args.py b/args.py
index 3cc7441..f7db308 100644
--- a/args.py
+++ b/args.py
@@ -1,31 +1,26 @@
 import os
-from enum import Enum
+import tensorflow as tf
+
+tf.logging.set_verbosity(tf.logging.INFO)
 
 file_path = os.path.dirname(__file__)
 
 model_dir = os.path.join(file_path, 'chinese_L-12_H-768_A-12/')
 config_name = os.path.join(model_dir, 'bert_config.json')
 ckpt_name = os.path.join(model_dir, 'bert_model.ckpt')
-
 output_dir = os.path.join(model_dir, '../tmp/result/')
-
 vocab_file = os.path.join(model_dir, 'vocab.txt')
 data_dir = os.path.join(model_dir, '../data/')
 
-max_seq_len = 32
-
-layer_indexes = [-2, -3, -4]
-
+num_train_epochs = 10
 batch_size = 128
-
-gpu_memory_fraction = 0.8
-
 learning_rate = 0.00005
 
-num_train_epochs = 10
+# gpu使用率
+gpu_memory_fraction = 0.8
 
-use_gpu = False
-if use_gpu:
-    device_id = '0'
-else:
-    device_id = '-1'
+# 默认取倒数第二层的输出值作为句向量
+layer_indexes = [-2]
+
+# 序列的最大程度，单文本建议把该值调小
+max_seq_len = 32
diff --git a/extract_feature.py b/extract_feature.py
index 5323167..9540f1e 100644
--- a/extract_feature.py
+++ b/extract_feature.py
@@ -1,12 +1,10 @@
-from graph import import_tf
 import modeling
 import tokenization
 from graph import optimize_graph
 import args
 from queue import Queue
 from threading import Thread
-
-tf = import_tf(0, True)
+import tensorflow as tf
 
 
 class InputExample(object):
diff --git a/graph.py b/graph.py
index cc86a4b..ff5ec4c 100644
--- a/graph.py
+++ b/graph.py
@@ -1,19 +1,10 @@
-import os
 import tempfile
 import json
 import logging
 from termcolor import colored
 import modeling
 import args
-import contextlib
-
-
-def import_tf(device_id=-1, verbose=False):
-    os.environ['CUDA_VISIBLE_DEVICES'] = '-1' if device_id < 0 else str(device_id)
-    os.environ['TF_CPP_MIN_LOG_LEVEL'] = '0' if verbose else '3'
-    import tensorflow as tf
-    tf.logging.set_verbosity(tf.logging.DEBUG if verbose else tf.logging.ERROR)
-    return tf
+import tensorflow as tf
 
 
 def set_logger(context, verbose=False):
@@ -35,7 +26,6 @@ def optimize_graph(logger=None, verbose=False):
         logger = set_logger(colored('BERT_VEC', 'yellow'), verbose)
     try:
         # we don't need GPU for optimizing the graph
-        tf = import_tf(device_id=0, verbose=verbose)
         from tensorflow.python.tools.optimize_for_inference_lib import optimize_for_inference
 
         # allow_soft_placement:自动选择运行设备
@@ -75,9 +65,7 @@ def optimize_graph(logger=None, verbose=False):
 
             tf.train.init_from_checkpoint(init_checkpoint, assignment_map)
 
-            minus_mask = lambda x, m: x - tf.expand_dims(1.0 - m, axis=-1) * 1e30
             mul_mask = lambda x, m: x * tf.expand_dims(m, axis=-1)
-            masked_reduce_max = lambda x, m: tf.reduce_max(minus_mask(x, m), axis=1)
             masked_reduce_mean = lambda x, m: tf.reduce_sum(mul_mask(x, m), axis=1) / (
                     tf.reduce_sum(m, axis=1, keepdims=True) + 1e-10)
 
@@ -113,7 +101,7 @@ def optimize_graph(logger=None, verbose=False):
                 [n.name[:-2] for n in output_tensors],
                 [dtype.as_datatype_enum for dtype in dtypes],
                 False)
-        tmp_file = tempfile.NamedTemporaryFile('w', delete=False).name
+        tmp_file = tempfile.NamedTemporaryFile('w', delete=False, dir=args.output_dir).name
         logger.info('write graph to a tmp file: %s' % tmp_file)
         with tf.gfile.GFile(tmp_file, 'wb') as f:
             f.write(tmp_g.SerializeToString())
diff --git a/similarity.py b/similarity.py
index 0472070..6a77293 100644
--- a/similarity.py
+++ b/similarity.py
@@ -181,7 +181,7 @@ class BertSim:
     def model_fn_builder(self, bert_config, num_labels, init_checkpoint, learning_rate,
                          num_train_steps, num_warmup_steps,
                          use_one_hot_embeddings):
-        """Returns `model_fn` closure for TPUEstimator."""
+        """Returns `model_fn` closurimport_tfe for TPUEstimator."""
 
         def model_fn(features, labels, mode, params):  # pylint: disable=unused-argument
             from tensorflow.python.estimator.model_fn import EstimatorSpec