tf-eager-text-generation

Code from tensorflow tutorial about text generation modified to be more flexible + including fun examples.

Fun examples

You can give your own text file in order to train a text generator :D.

1. run: python train.py --text-path <path to txt>, check all the available arguments with python train.py --help.
2. run: python sample.py --text-path <path to txt>, including all arguments passed during train.

usage: train.py [-h] --text-path TEXT_PATH [--text-type {full-text,sentences}]
                [--max-length MAX_LENGTH] [--buffer-size BUFFER_SIZE]
                [--batch-size BATCH_SIZE] [--to-lower] [--model {gru}]
                [--embedding-dim EMBEDDING_DIM] [--units UNITS]
                [--checkpoint-dir CHECKPOINT_DIR] [--epochs EPOCHS]
                [--verbose] [--num-char-generate NUM_CHAR_GENERATE]
                [--start-string START_STRING] [--temperature TEMPERATURE]

optional arguments:
  -h, --help            show this help message and exit
  --text-path TEXT_PATH
                        Path to text file.
  --text-type {full-text,sentences}
                        Type of text senteces = consider each line of file as
                        a piece of text. full-text = consider text file as a
                        single long text.
  --max-length MAX_LENGTH
                        Max length of a piece of text used to train model.
  --buffer-size BUFFER_SIZE
                        Buffer size to shuffle dataset.
  --batch-size BATCH_SIZE
                        Batch size for training.
  --to-lower            Convert all chars to lower case.
  --model {gru}         RNN cell, for now only GRU is available.
  --embedding-dim EMBEDDING_DIM
                        Embedding dimension in model.
  --units UNITS         RNN units.
  --checkpoint-dir CHECKPOINT_DIR
                        Path to save models checkpoints.
  --epochs EPOCHS       Train epochs.
  --verbose             Generate text as it train.
  --num-char-generate NUM_CHAR_GENERATE
                        Number of chars to generate as it train.
  --start-string START_STRING
                        Start string to generate text. If None a random char
                        will be used instead.
  --temperature TEMPERATURE
                        Higher = more creative text, lower = boring text.

Bands and Artists

1. get dataset
2. python train.py --text-path datasets/bands/bands.txt --text-type sentences --num-char-generate 40

The Muriscerd
Tune
Bredie Jamesson
Creed Underson
May
Mol Neic & Gistol
Aika Gray
Indele
Ipu8t
Ignes
R.E.T.H.
Sumpy
Fuster Live

Colors

1. get dataset
2. python train.py --text-path datasets/colors/colors.txt --text-type sentences --num-char-generate 40 --max-length 40

Names

1. get dataset
2. python train.py --text-path datasets/names/names.txt --text-type sentences --max-length 40 --to-lower

anezenina
amaxdo
pedras
phaudo
bertho
rodomira
ratu
edjulia
eledete
zirdes

Shakespeare

This example was taken from TensorFlow tutorial.

1. get dataset
2. python train.py --text-path datasets/shakespeare.txt --num-char-generate 100

were to the death of him
And nothing of the field in the view of hell,
When I said, banish him, I will not burn thee that would live.

HENRY BOLINGBROKE:
My gracious uncle--

DUKE OF YORK:
As much disgraced to the court, the gods them speak,
And now in peace himself excuse thee in the world.

HORTENSIO:
Madam, 'tis not the cause of the counterfeit of the earth,
And leave me to the sun that set them on the earth
And leave the world and are revenged for thee.

GLOUCESTER:
I would they were talking with the very name of means
To make a puppet of a guest, and therefore, good Grumio,
Nor arm'd to prison, o' the clouds, of the whole field,
With the admire
With the feeding of thy chair, and we have heard it so,
I thank you, sir, he is a visor friendship with your silly your bed.

SAMPSON:
I do desire to live, I pray: some stand of the minds, make thee remedies
With the enemies of my soul.

MENENIUS:
I'll keep the cause of my mistress.

POLIXENES:
My brother Marcius!

Second Servant:
Will't ple

---

The code is orinally from the Tensorflow repository and can be accessed at https://github.com/tensorflow/tensorflow/blob/master/tensorflow/contrib/eager/python/examples/generative_examples/text_generation.ipynb, I just did small modifications.