TF 2.1 support

Support for Tensorflow 2.1 added! (thanks to #165 by @ZerxXxes!)

TF 2.1 is now the minimum version in light of this migration. (which has both CPU and GPU support). If you need to use TF 1.X, use an older version.

Synthesis + Generate Progress bar

Two major features:

Synthesis (beta)

Generate text using two (or more!) trained models simultaneously. See this notebook for a demo.

The results are messier than usual so a lower temperature is recommended. It should work on both char-level and word-level models, or a mix of both. (however, I do not recommending mixing line-delimited and full text models!)

Please file issues if there are errors!

Generate Progress Bar

Thanks to tqdm, all generate functions show a progress bar! You can override this by passing progress=False to the function.

Additionally, the default generate temperature is now [1.0, 0.5, 0.2, 0.2]!

Emergency Case Fix

v1.4.1

tensorflow dependency note #76

Interactive Mode + Bug Fixes

Features

• Interactive mode, which lets you control which text is added. (#52, thanks @juanets !)
• Allow backends other than TensorFlow (#44, thanks @torokati44 !)
• Allow periodic weights saving (#37, thanks @IrekRybark !)
• Multi-GPU support (beta: see #62 )

Fixes

• Handle prefix in word-level models correctly.

1.3.2

Emergency bug fix to address #57 which occurs in newer Keras versions.

Temperature Cycling + Fixes

• Added ability to cycle temperatures during training (see this notebook for more information
• Added utf-8 encoding for vocab export.
• Added alias for train_on_texts(new_model=True) to train_new_model.
• Fixed an issue where specifying dropout could cause issues.

First implementation of encoding text

• Added encode_text_vectors to encode text using the trained network.
• Added similarity to quickly calculate cosine similarity and return the most similar texts.

See this notebook for details.

Minor Fixes

• Make is_csv work for real downstream.
• Description tweaks

CSV Utility

• Added validation to disable validation training for speed.
• Added is_csv: Use with train_from_file if the source file is a one-column CSV (e.g. an export from BigQuery or Google Sheets) for proper quote/newline escaping.
• README tweaks

Handle Overfitting

• Renamed prop_keep to train_size, and will use the remaining data for validation.
• Added dropout, which randomly excludes input tokens each epoch.