yxtay / char-rnn-text-generation Goto Github PK

I'm trying to use this project to generate text with non-ascii symbols (cyrilic) and keep getting spaces/tabs/commas but no text, is this related to string.printable which part i should modify to enable support of non-ascii symbols?

Thanks!

Output

python tf_model.py train --text-path Vojna_i_mir._Kniga_1.txt --checkpoint-path /home/norn/src/char-rnn-text-generation/tf_checkpoint/chk
2017-12-16 15:10:56,234 - main - INFO - corpus length: 1433026.
2017-12-16 15:10:56,235 - main - INFO - building model: {'clip_norm': 5.0, 'batch_size': 64, 'num_layers': 2, 'vocab_size': 98, 'rnn_size': 128, 'p_keep': 1.0, 'learning_rate': 0.001, 'embedding_size': 32}.
2017-12-16 15:10:57.231924: I tensorflow/core/platform/cpu_feature_guard.cc:137] Your CPU supports instructions that this TensorFlow binary was not compiled to use: SSE4.1 SSE4.2 AVX AVX2 FMA
2017-12-16 15:10:57,523 - main - INFO - model saved: /home/norn/src/char-rnn-text-generation/tf_checkpoint/chk.
2017-12-16 15:10:57,739 - main - INFO - tensorboard set up.
2017-12-16 15:10:57,740 - main - INFO - building model: {'clip_norm': 5.0, 'batch_size': 1, 'num_layers': 2, 'vocab_size': 98, 'rnn_size': 128, 'p_keep': 1.0, 'learning_rate': 0.001, 'embedding_size': 32}.
2017-12-16 15:10:57,996 - main - INFO - inference model loaded: /home/norn/src/char-rnn-text-generation/tf_checkpoint/chk.
2017-12-16 15:10:58,530 - main - INFO - start of training.
epoch 1/32: 0%| | 0/349 [00:00<?, ?it/s]2017-12-16 15:10:58,532 - utils - INFO - number of batches: 349.
2017-12-16 15:10:58,532 - utils - INFO - effective text length: 1429504.
2017-12-16 15:10:58,532 - utils - INFO - x shape: (64, 22336).
2017-12-16 15:10:58,532 - utils - INFO - y shape: (64, 22336).
epoch 1/32: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████| 349/349 [02:53<00:00, 1.97it/s]
2017-12-16 15:13:51,930 - main - INFO - epoch: 1, duration: 173s, loss: 0.93437.
2017-12-16 15:13:52,082 - main - INFO - model saved: /home/norn/src/char-rnn-text-generation/tf_checkpoint/chk.
2017-12-16 15:13:52,114 - main - INFO - generating 512 characters from top 10 choices.
2017-12-16 15:13:52,115 - main - INFO - generating with seed: "а помощь к брату, кто бы он ни б".
2017-12-16 15:13:53,027 - main - INFO - generated text:
а помощь к брату, кто бы он ни б , , ,
, ,-
, , .
, . ,

epoch 2/32:

Hi there!

Im a noob exploring your text generation project. Its been excellent so far, but I was hoping that you could help me with one question: what is the approach used to pick the best number of epochs? for example, if I were to assign a random number of epochs when I run the script, does the model intelligently stop at a point where it can no longer learn?
Is there a way for me to find out how many epochs are best, even though I may have assigned and run a large number of epochs at start?

Hey,
Nice work!!

I just have one query.
I am working on character-level text generation with data as novels.

In Novels, there are character names, which occur too many times, so will that affect my model while generating text?

Thanks in Advance!