tf_chatbot_seq2seq_antilm
tf_chatbot_seq2seq_antilm copied to clipboard
Marsan-Ma-zz
Seq2seq chatbot with attention and anti-language model to suppress generic response, option for further improve by deep reinforcement learning.
Hi all, I am trying to understand the implementation of the anti-LM model, in particular the meaning of this line: line 128: all_prob_t = model_step(dummy_encoder_inputs, cand['dec_inp'], dptr, target_weights, bucket_id) where...
I bump into the issue while trying to train based on the data from [twitter_en-big](https://github.com/Marsan-Ma/chat_corpus#5-twitter_en-big) `python main.py --mode train --model_name twitter_en_big` My working environment is: Windows 10 64 bit Python...
File "main.py", line 40, in tf.app.run() File "/home/luban/anaconda2/lib/python2.7/site-packages/tensorflow/python/platform/app.py", line 48, in run _sys.exit(main(_sys.argv[:1] + flags_passthrough)) File "main.py", line 30, in main train(args) File "/nfs/private/proj/chatbot/lib/train.py", line 32, in train model =...
reading data line 4600000 reading data line 4700000 reading data line 4800000 reading data line 4900000 reading data line 5000000 I tensorflow/core/common_runtime/gpu/pool_allocator.cc:247] PoolAllocator: After 5476 get requests, put_count=2291 evicted_count=1000 eviction_rate=0.436491...
Tensorflow API changes a lot since tf1.2. Does author consider update those api to 1.3? That would be good to keep the APIs update with official.
in the function "get_predicted_sentence" in the lib/seq2seq_model_utils.py, you loop the dptr and pass a new cand["dec_inp"] to the function model_step. However, when the flag "forward_only" is set to True, any...
I notice the n_bonus type is int in config.py, however the recommended value for n_bonus is 0.05 in readme. Is this a typo error?
Hello, I am just curious how you implemented beam search decoding. Did you use the output logits of the default Seq2Seq model? I am working around these dialogue papers and...
I've found this error occured when trying to test with data set is a part of the training data which has been done. File /chatbot_seq2seq_antilm/lib/seq2seq_model_utils.py", line 49, in dict_lookup word...
Metadata
Owner
Metadata
Seq2seq chatbot with attention and anti-language model to suppress generic response, option for further improve by deep reinforcement learning.