Smart seeding

Enable "Smart Seeding" which allows a prefix to be provided during line generation. The generator will complete the line based on the provided seed. When training on structured data (DataFrames) this enables the first N column values to be pre-provided and then remaining columns will be generated based on the initial values.

v0.14.0: Jm/syn 21 (#58)

• Introduce Keras Early Stopping and Save Best Model features. Set default number of epochs to 100 which should allow most training sequences to automatically stop without potential over-fitting.

• Provide better tracking of which epoch's model was used as the best one in the model history table

• Temporarily disable DP mode

TF Updates

🏃 🏃‍♀️ Speedups! Utilized TensorFlow functions and batch predictions for massive speedups in text generation!

Large Dataset and MP Optimizations

• Included SentencePiece support for tokenizing large datasets. By default, if a dataset is larger than 100,000 lines, we will utilize a sample of 100k to build the tokenizer.

• Switched to a loky based backend for multi-processing. This fixes several run-time bugs on Windows platforms and increases general stability for consistent multi-processing text generation.

Multi-Processing Enhancements

⏩ ⏩ Multi-Processing, now with more Multi! This release has some updates and bugfixes around multi-processing that makes it faster and more efficient with use of CPUs.

Updated numpy and pandas libraries

Updates:

• Updated pandas and numpy libraries to automatically select latest minor version and bug fix.
• Better compatibility with notebook environments with pre-installed runtimes such as Google Colaboratory.

v0.11.0

Last major release before the Gretel Beta 🚀 !!!

Updates:

• Data generation, by default, will utilize CPU and multi-processing, using as many logical CPUs as possible on the system. This change should be automatic with no specific interface changes to have it work.

• Batch mode has a read-only interface for quickly re-loading a model and generating new data. Please see our example Notebook for a usage example!

RC5

set generator back to previous state, update var name (#43)

Co-authored-by: John Myers <[email protected]>

RC4

Jm/bugfix generate (#42)

* Bugfix on generation, RC2 prep

* Use SCM for install

* test

* Revert to full list decoding for gen chars

* Remove parallelism kwarg

Co-authored-by: John Myers <[email protected]>

RC3

Bugfix on generation, RC2 prep (#41)

* Bugfix on generation, RC2 prep

* Use SCM for install

Co-authored-by: John Myers <[email protected]>