Closed jacobdanovitch closed 6 years ago
Hi Jacob,
It seems that you trained your model with IS_DEV env flag. Filename gives you a clue about params of your NN. I see two ways to solve the problem
1) rename you file processed_dialogs_gru_hd7_drop0.2_encd2_decd2_il7_cs3_ansl9_lr1.0_gc_5.0_learnemb_cdim128_window10_voc11786_vec15_sgTrue_pp_free2926.32_sensitive3066.80
-> processed_dialogs_gru_hd7_drop0.2_encd2_decd2_il7_cs3_ansl9_lr1.0_gc_5.0_learnemb_cdim128_window10_voc11786_vec15_sgTrue
and run IS_DEV=1 python bin/cakechat_server.py
2) Train your model initially without IS_DEV flag
Thank you, I'd forgotten I used IS_DEV, and I didn't realize I had to precede the server call with that with that as well. Works perfectly.
cool! let us know if you have other questions about cakechat
Hi,
I've loaded my own training and validation corpus, ran
prepare_index_files.py
, and trained it with no issue. Afterwards, when I ranpython bin/cakechat_server.py
, it continually threw this error:I messed around with
get_nn_model()
indialog_model/model.py
a bit and realized it was looking for a file named:processed_dialogs_gru_hd512_drop0.2_encd2_decd2_il30_cs3_ansl32_lr1.0_gc_5.0_learnemb_cdim128_window10_voc11786_vec128_sgTrue
My file in
data/nn_models
was called:processed_dialogs_gru_hd7_drop0.2_encd2_decd2_il7_cs3_ansl9_lr1.0_gc_5.0_learnemb_cdim128_window10_voc11786_vec15_sgTrue_pp_free2926.32_sensitive3066.80
I made a copy and renamed it. I then tried to both run the server and train it again, and as soon as it tried to load the model, in both instances I got:
ValueError: mismatch: parameter has shape (11786L, 128L) but value to set has shape (11786L, 15L)
Not really sure where to go from here; thanks in advance. Running Windows / Anaconda with py2.7. All dependencies installed and everything else is running fine thus far. I had it working with the pre-trained model. Tried running the server through both Git Bash and cmd, if that makes a difference. Trained through Bash.