"NotFoundError" when test pre-trained model

[Zminghua]

hello, I have downloaded pre-trained model, when i follow the instructions in README.md to test the pre-trained model then I got the error info.

NotFoundError (see above for traceback): Tensor name "embedding_rnn_seq2seq/rnn/embedding_wrapper/multi_rnn_cell/cell_1/basic_lstm_cell/weights" not found in checkpoint files /home/zmh/github/DeepQA/save/model-pretrainedv2/model.ckpt

my python is 3.4.2 and my tensorflow is 1.1, this is run in Debian 3.2.78 with x64。

[Conchylicultor]

The pre-trained model has been trained with a older​ version of tensorflow so it is possible that there is some compatibility issues with TF 1.1

[jinjiaji512]

I also got a similar error.

NotFoundError (see above for traceback): Tensor name "embedding_rnn_seq2seq/embedding_rnn_decoder/rnn_decoder/output_projection_wrapper/multi_rnn_cell/cell_1/basic_lstm_cell/bias" not found in checkpoint files /Users/jingyongxia/git/DeepQA/save/model-pretrainedv2/model.ckpt

my python is 3.6.1 and my tensorflow is 1.2.1, this is run in MacOs 10.12.5 with x64。

[svakulenk0]

confirm. works with pip install tensorflow==1.0.0

[EMCP]

According to TF 1.2 release notes

tensorflow/contrib/rnn undergoes RNN cell variable renaming for consistency with Keras layers. Specifically, the previous variable names "weights" and "biases" are changed to "kernel" and "bias", respectively. This may cause backward incompatibility with regard to your old checkpoints containing such RNN cells, in which case you can use the checkpoint_convert script

https://github.com/tensorflow/tensorflow/blob/master/tensorflow/contrib/rnn/python/tools/checkpoint_convert.py

Update : I ran this script.. passed the V1 flag to indicate these checkpoint files are the version 1 type... (I assumed this based on the files in the zip) .. still get an error.

https://stackoverflow.com/questions/45633165/tried-out-the-convert-checkpoint-py-unsure-if-it-worked

[stiger104]

Pleas use TF 1.0.0, it works!

[wiwengweng]

if you want to use TF1.2 or 1.3, please check Nicolas C.'s work here. https://mcastedu-my.sharepoint.com/personal/nicholas_cutajar_a100636_mcast_edu_mt/_layouts/15/guestaccess.aspx?folderid=077576c4cf9854642a968f67909380f45&authkey=AVt2JWMPkf2R_mWBpI1eAUY

[snorlaxse]

如果你想使用TF1.2或1.3，请在这里查看Nicolas C.的工作。 https://mcastedu-my.sharepoint.com/personal/nicholas_cutajar_a100636_mcast_edu_mt/_layouts/15/guestaccess.aspx?folderid=077576c4cf9854642a968f67909380f45&authkey=AVt2JWMPkf2R_mWBpI1eAUY

sorry, this link is nothing. I want to know its context. What can I do ?