Fail to create new dictionary using g2p-seq2seq

[Askarya QS]

i could to build new phonetic dictionary using g2p-seq2seq
i have some trouble with training my dict
before it, i build new dict with languange bahasa indonesia
just 5 words will determine, the command of robotic move in bahasa indonesia
example : maju, mundur, kiri, kanan, berhenti
im bulid phones with 34 phones, determine symblos and vp in languange bahasa indonesia
after that im run i have some trouble like this

Preparing G2P data
Creating vocabulary ind_model/vocab.phoneme
Creating vocabulary ind_model/vocab.grapheme
Reading development and training data.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE3 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE4.1 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE4.2 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX2 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use FMA instructions, but these are available on your machine and could speed up CPU computations.
Creating 2 layers of 64 units.
Created model with fresh parameters.
global step 200 learning rate 0.5000 step-time 0.44 perplexity 1.64
eval: perplexity 1.00
global step 400 learning rate 0.5000 step-time 0.36 perplexity 1.15
eval: perplexity 1.00
global step 600 learning rate 0.5000 step-time 0.43 perplexity 1.01
eval: perplexity 1.00
global step 800 learning rate 0.5000 step-time 0.41 perplexity 1.00
eval: perplexity 1.00
global step 1000 learning rate 0.5000 step-time 0.44 perplexity 1.00
eval: perplexity 1.00
global step 1200 learning rate 0.5000 step-time 0.42 perplexity 1.00
eval: perplexity 1.00
global step 1400 learning rate 0.5000 step-time 0.42 perplexity 1.00
eval: perplexity 1.00
global step 1600 learning rate 0.5000 step-time 0.52 perplexity 1.00
eval: perplexity 1.00
global step 1800 learning rate 0.5000 step-time 0.44 perplexity 1.00
eval: perplexity 1.00
global step 2000 learning rate 0.5000 step-time 0.52 perplexity 1.00
eval: perplexity 1.00
global step 2200 learning rate 0.5000 step-time 0.41 perplexity 1.00
eval: perplexity 1.00
global step 2400 learning rate 0.5000 step-time 0.42 perplexity 1.00
eval: perplexity 1.00
global step 2600 learning rate 0.5000 step-time 0.49 perplexity 1.00
eval: perplexity 1.00
global step 2800 learning rate 0.5000 step-time 0.46 perplexity 1.00
eval: perplexity 1.00
global step 3000 learning rate 0.5000 step-time 0.43 perplexity 1.00
eval: perplexity 1.00
global step 3200 learning rate 0.5000 step-time 0.42 perplexity 1.00
eval: perplexity 1.00
global step 3400 learning rate 0.5000 step-time 0.53 perplexity 1.00
eval: perplexity 1.00
global step 3600 learning rate 0.5000 step-time 0.47 perplexity 1.00
eval: perplexity 1.00
global step 3800 learning rate 0.5000 step-time 0.45 perplexity 1.00
eval: perplexity 1.00
global step 4000 learning rate 0.5000 step-time 0.47 perplexity 1.00
eval: perplexity 1.00
global step 4200 learning rate 0.5000 step-time 0.51 perplexity 1.00
eval: perplexity 1.00
global step 4400 learning rate 0.5000 step-time 0.44 perplexity 1.00
eval: perplexity 1.00
global step 4600 learning rate 0.5000 step-time 0.35 perplexity 1.00
eval: perplexity 1.00
global step 4800 learning rate 0.5000 step-time 0.53 perplexity 1.00
eval: perplexity 1.00
global step 5000 learning rate 0.5000 step-time 0.49 perplexity 1.00
eval: perplexity 1.00
global step 5200 learning rate 0.5000 step-time 0.48 perplexity 1.00
eval: perplexity 1.00
global step 5400 learning rate 0.5000 step-time 0.44 perplexity 1.00
eval: perplexity 1.00
global step 5600 learning rate 0.5000 step-time 0.45 perplexity 1.00
eval: perplexity 1.00
global step 5800 learning rate 0.5000 step-time 0.42 perplexity 1.00
eval: perplexity 1.00
global step 6000 learning rate 0.4950 step-time 0.42 perplexity 1.00
eval: perplexity 1.00
global step 6200 learning rate 0.4950 step-time 0.49 perplexity 1.00
eval: perplexity 1.00
global step 6400 learning rate 0.4950 step-time 0.43 perplexity 1.00
eval: perplexity 1.00
global step 6600 learning rate 0.4950 step-time 0.48 perplexity 1.00
eval: perplexity 1.00
global step 6800 learning rate 0.4950 step-time 0.44 perplexity 1.00
eval: perplexity 1.00
global step 7000 learning rate 0.4950 step-time 0.47 perplexity 1.00
eval: perplexity 1.00
global step 7200 learning rate 0.4950 step-time 0.41 perplexity 1.00
eval: perplexity 1.00
Training done.
Traceback (most recent call last):
File "/usr/local/bin/g2p-seq2seq", line 9, in <module>
load_entry_point('g2p-seq2seq==5.0.0a0', 'console_scripts', 'g2p-seq2seq')()
File "build/bdist.linux-x86_64/egg/g2p_seq2seq/app.py", line 79, in main
File "build/bdist.linux-x86_64/egg/g2p_seq2seq/g2p.py", line 267, in train
File "build/bdist.linux-x86_64/egg/g2p_seq2seq/g2p.py", line 76, in load_decode_model
RuntimeError: Model not found in ind_model</module>

will you help me to fix that? thank you

[Askarya QS]

that index perplexity not changed, just stopped in 1.00, what the meaning of that "perplexity"?

[Nickolay V. Shmyrev]

what the meaning of that "perplexity"?

https://en.wikipedia.org/wiki/Perplexity

[Askarya QS]

wheter the training data in .dict that i used is too less? i'm just have 5 words

[Arseniy Gorin]

you need around 100k words to train this kind of models. for your task you do not need G2P - just do hand-crafted dictionary

[Askarya QS]

thanks for the answer, perhaps i will build around 100k words, that so help to my projects