autrainer

New Features

• Save Preprocessing Pipeline for Inference by @ramppdev in https://github.com/autrainer/autrainer/pull/21

Fixes

• Fix Multi-Label Classification Threshold by @ramppdev in https://github.com/autrainer/autrainer/pull/22
• Fix Inconsistent Scheduler Step Call Positions by @ramppdev in https://github.com/autrainer/autrainer/pull/23
• Fix Hugging Face Model Local Path Directory by @ramppdev in https://github.com/autrainer/autrainer/pull/25
• Fix Audio File Handler Deserialization by @ramppdev in https://github.com/autrainer/autrainer/pull/26

Full Changelog: https://github.com/autrainer/autrainer/compare/v0.2.0...v0.3.0

New Features

• Run Hydra CLI Commands only in Notebook as Subprocess by @ramppdev in https://github.com/autrainer/autrainer/pull/11
• Allow custom Callbacks in Main Configuration by @ramppdev in https://github.com/autrainer/autrainer/pull/12
• Add Audiomentations Augmentation Wrapper by @ramppdev in https://github.com/autrainer/autrainer/pull/13
• Modularize CUDA-enabled Device Selection by @ramppdev in https://github.com/autrainer/autrainer/pull/14
• Load Model Checkpoints with Arbitrary Output Dimensions by @ramppdev in https://github.com/autrainer/autrainer/pull/5

Fixes

• Fix README Header Image and Add Missing PyPI Links by @ramppdev in https://github.com/autrainer/autrainer/pull/6
• Fix Inference with custom Classes by @ramppdev in https://github.com/autrainer/autrainer/pull/7

Full Changelog: https://github.com/autrainer/autrainer/compare/v0.1.0...v0.2.0

v0.1.0 - Initial Release

This marks the initial release of autrainer, a modular and extensible deep learning toolkit for computer audition tasks.