liuyuemaicha / deep-reinforcement-learning-for-dialogue-generation-in-tensorflow Goto Github PK

AttributeError: module 'tensorflow.python.ops.rnn_cell' has no attribute '_linear'

感谢你代码的分享~
我在train的时候觉得，在训练的时候是不是还是应该固定decode？
还有reward的计算用乘法还是减法？我看你在另一个repository用的是乘法

Prepare Chitchat data in ./grl_data/
Reading development and training data (limit: 0).
b_set length: 0
b_set length: 6
b_set length: 2
b_set length: 0
Creating st_model model with fresh parameters
Created st_model model with fresh parameters
Creating bk_model model with fresh parameters
Created bk_model model with fresh parameters
Creating cc_model model with fresh parameters
Created cc_model model with fresh parameters
Traceback (most recent call last):
File "/IR/Deep-Reinforcement-Learning-for-Dialogue-Generation-in-tensorflow-master/grl_train.py", line 473, in
tf.app.run()
File "/Library/Python/2.7/site-packages/tensorflow/python/platform/app.py", line 43, in run
sys.exit(main(sys.argv[:1] + flags_passthrough))
File "/IR/Deep-Reinforcement-Learning-for-Dialogue-Generation-in-tensorflow-master/grl_train.py", line 467, in main
train()
File "/IR/Deep-Reinforcement-Learning-for-Dialogue-Generation-in-tensorflow-master/grl_train.py", line 261, in train
rl_model = create_rl_model(sess, grl_config, False, grl_config.name_model)
File "/IR/Deep-Reinforcement-Learning-for-Dialogue-Generation-in-tensorflow-master/grl_train.py", line 99, in create_rl_model
rl_model = grl_rnn_model.grl_model(grl_config=rl_config, name_scope=name_scope, forward=forward_only)
File "/IR/Deep-Reinforcement-Learning-for-Dialogue-Generation-in-tensorflow-master/grl_rnn_model.py", line 122, in init
softmax_loss_function=softmax_loss_function)
File "/IR/Deep-Reinforcement-Learning-for-Dialogue-Generation-in-tensorflow-master/grl_seq2seq.py", line 477, in model_with_buckets
decoder_inputs[:bucket[1]])
File "/IR/Deep-Reinforcement-Learning-for-Dialogue-Generation-in-tensorflow-master/grl_rnn_model.py", line 121, in
seq2seq=lambda x, y: seq2seq_f(x, y, tf.select(self.forward_only, True, False)),
File "/IR/Deep-Reinforcement-Learning-for-Dialogue-Generation-in-tensorflow-master/grl_rnn_model.py", line 98, in seq2seq_f
dtype=dtype
File "/IR/Deep-Reinforcement-Learning-for-Dialogue-Generation-in-tensorflow-master/grl_seq2seq.py", line 401, in embedding_attention_seq2seq
lambda: decoder(False))
File "/Library/Python/2.7/site-packages/tensorflow/python/ops/control_flow_ops.py", line 1776, in cond
orig_res, res_t = context_t.BuildCondBranch(fn1)
File "/Library/Python/2.7/site-packages/tensorflow/python/ops/control_flow_ops.py", line 1677, in BuildCondBranch
r = fn()
File "/IR/Deep-Reinforcement-Learning-for-Dialogue-Generation-in-tensorflow-master/grl_seq2seq.py", line 400, in
lambda: decoder(True),
File "/IR/Deep-Reinforcement-Learning-for-Dialogue-Generation-in-tensorflow-master/grl_seq2seq.py", line 397, in decoder
return outputs + state_list
File "/Library/Python/2.7/site-packages/tensorflow/python/ops/math_ops.py", line 813, in binary_op_wrapper
y = ops.convert_to_tensor(y, dtype=x.dtype.base_dtype, name="y")
File "/Library/Python/2.7/site-packages/tensorflow/python/framework/ops.py", line 669, in convert_to_tensor
ret = conversion_func(value, dtype=dtype, name=name, as_ref=as_ref)
File "/Library/Python/2.7/site-packages/tensorflow/python/framework/constant_op.py", line 176, in _constant_tensor_conversion_function
return constant(v, dtype=dtype, name=name)
File "/Library/Python/2.7/site-packages/tensorflow/python/framework/constant_op.py", line 165, in constant
tensor_util.make_tensor_proto(value, dtype=dtype, shape=shape, verify_shape=verify_shape))
File "/Library/Python/2.7/site-packages/tensorflow/python/framework/tensor_util.py", line 367, in make_tensor_proto
_AssertCompatible(values, dtype)
File "/Library/Python/2.7/site-packages/tensorflow/python/framework/tensor_util.py", line 302, in _AssertCompatible
(dtype.name, repr(mismatch), type(mismatch).name))
TypeError: Expected int32, got list containing Tensors of type '_Message' instead.

I don't know how to solve it. Can you give me some idea?

Thank you for helping me a lot, due to your code. Sorry, but I met some code problems that I can't solve. Looking forward to your updating!

can you give me any link of some dataset to train the model? Thanks.

你好，数据可以给个百度网盘地址吗？ 非常感谢，跪求数据 qq 1206401181

The below code is defined in embedding_attention_decoder method, file name grl_seq2seq.py.
` if beam_search:

    return beam_rnn_decoder(emb_inp, initial_state, cell,
                            embedding=embedding,
                            output_projection=output_projection,
                            beam_size=beam_size,
                            scope=scope)
else:
    loop_function = _extract_argmax_and_embed(embedding, output_projection,
                                              update_embedding_for_previous) if feed_previous else None

return attention_decoder(emb_inp, initial_state, attention_states, cell,
    					output_size=output_size,
  						num_heads=num_heads,
    					loop_function=loop_function,
    					initial_state_attention=initial_state_attention,
    					scope=scope)`

Here if the beam search is activated it initializes normal decoder. I would have expected beam search along with attention decoder?

我只用了 Ease of answering 作为reward，但是随着训练这一项从-2.x开始一直减小到负无穷。

at

python grl_train.py
Prepare Chitchat data in ./grl_data/
Tokenizing data in ./grl_data/chitchat.train.answer
Tokenizing data in ./grl_data/chitchat.train.query
Tokenizing data in ./grl_data/chitchat.dev.answer
Tokenizing data in ./grl_data/chitchat.dev.query
Reading development and training data (limit: 0).
b_set length: 0
b_set length: 6
b_set length: 2
b_set length: 0
Creating st_model model with fresh parameters
Created st_model model with fresh parameters
Creating bk_model model with fresh parameters
Created bk_model model with fresh parameters
Creating cc_model model with fresh parameters
Created cc_model model with fresh parameters
Traceback (most recent call last):
File "grl_train.py", line 473, in
tf.app.run()
File "/Users/venv/lib/python3.6/site-packages/tensorflow/python/platform/app.py", line 43, in run
sys.exit(main(sys.argv[:1] + flags_passthrough))
File "grl_train.py", line 467, in main
train()
File "grl_train.py", line 261, in train
rl_model = create_rl_model(sess, grl_config, False, grl_config.name_model)
File "grl_train.py", line 99, in create_rl_model
rl_model = grl_rnn_model.grl_model(grl_config=rl_config, name_scope=name_scope, forward=forward_only)
File "/Users/Deep-Reinforcement-Learning-for-Dialogue-Generation-in-tensorflow/grl_rnn_model.py", line 123, in init
softmax_loss_function=softmax_loss_function)
File "/Users/Deep-Reinforcement-Learning-for-Dialogue-Generation-in-tensorflow/grl_seq2seq.py", line 477, in model_with_buckets
decoder_inputs[:bucket[1]])
File "/Users/Deep-Reinforcement-Learning-for-Dialogue-Generation-in-tensorflow/grl_rnn_model.py", line 122, in
seq2seq=lambda x, y: seq2seq_f(x, y, tf.select(self.forward_only, True, False)),
File "/Users/Deep-Reinforcement-Learning-for-Dialogue-Generation-in-tensorflow/grl_rnn_model.py", line 99, in seq2seq_f
dtype=dtype
File "/Users/Deep-Reinforcement-Learning-for-Dialogue-Generation-in-tensorflow/grl_seq2seq.py", line 401, in embedding_attention_seq2seq
lambda: decoder(False))
File "/Users/venv/lib/python3.6/site-packages/tensorflow/python/ops/control_flow_ops.py", line 1776, in cond
orig_res, res_t = context_t.BuildCondBranch(fn1)
File "/Users/venv/lib/python3.6/site-packages/tensorflow/python/ops/control_flow_ops.py", line 1677, in BuildCondBranch
r = fn()
File "/Users/Deep-Reinforcement-Learning-for-Dialogue-Generation-in-tensorflow/grl_seq2seq.py", line 400, in
lambda: decoder(True),
File "/Users/Deep-Reinforcement-Learning-for-Dialogue-Generation-in-tensorflow/grl_seq2seq.py", line 397, in decoder
return outputs + state_list
File "/Users/venv/lib/python3.6/site-packages/tensorflow/python/ops/math_ops.py", line 813, in binary_op_wrapper
y = ops.convert_to_tensor(y, dtype=x.dtype.base_dtype, name="y")
File "/Users/venv/lib/python3.6/site-packages/tensorflow/python/framework/ops.py", line 669, in convert_to_tensor
ret = conversion_func(value, dtype=dtype, name=name, as_ref=as_ref)
File "/Users/venv/lib/python3.6/site-packages/tensorflow/python/framework/constant_op.py", line 176, in _constant_tensor_conversion_function
return constant(v, dtype=dtype, name=name)
File "/Users/venv/lib/python3.6/site-packages/tensorflow/python/framework/constant_op.py", line 165, in constant
tensor_util.make_tensor_proto(value, dtype=dtype, shape=shape, verify_shape=verify_shape))
File "/Users/venv/lib/python3.6/site-packages/tensorflow/python/framework/tensor_util.py", line 367, in make_tensor_proto
_AssertCompatible(values, dtype)
File "/Users/venv/lib/python3.6/site-packages/tensorflow/python/framework/tensor_util.py", line 302, in _AssertCompatible
(dtype.name, repr(mismatch), type(mismatch).name))
TypeError: Expected int32, got list containing Tensors of type '_Message' instead.