Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

v0.5.0 (2024-02-26)

Added

Output streaming in the text generation task (#32)

Changed

Update Whisper model list
Moved seed to the prediction form in Stable Diffusion (#41)
Bumped bumblebee and kino dependency requirements

Removed

Removed the conversational task (#41)

v0.4.0 (2023-10-06)

Added

Output streaming in the text generation task (#32)
Output streaming in the speech-to-text task (#32)
Timestamps to the speech-to-text task output (#32)

Changed

Updated code generation with audio/image inputs to read from file (#36)
Bumped bumblebee and kino dependency requirements

Fixed

Respect seed in the text-to-image task

v0.3.0 (2023-04-14)

Added

Image-to-text task (#29)

Changed

Model loading to no longer use :log_params_diff
Updated generation examples to use generation config (#30)

v0.2.1 (2023-03-23)

Added

Conversation task (#23)
Question answering task (#23)

Changed

Updated generated source to rely on enumerable Kino.Control (#27)

v0.2.0 (2023-03-16)

Added

XLM-RoBERTa models for punctuation and language detection (#16)
Zero-shot text classification task (#17)
Speech-to-text task (#19)

v0.1.4 (2022-12-15)

Fixed

Numeric parameters that were inserted as strings in the generated code

v0.1.3 (2022-12-15)

Fixed

Smart cell attributes import for the token classification task (#12)

v0.1.2 (2022-12-15)

Changed

Bumped bumblebee dependency requirement

v0.1.1 (2022-12-14)

Added

Fill-mask to the Neural Network task cell (#9)
Token classification to the Neural Network task cell (#10)

v0.1.0 (2022-12-08)

Initial release.