Version 0.6.0

What's Changed

• [BUGFIX] Reject conditioned/gated configs in the A2 fast-path detector by @rhaist in #301
• Fix A2 fast-path prewarm count to match the generic WaveNet by @rhaist in #300
• Fix LSTM real-time safety (#218) by @rhaist in #299
• Add dilation for WaveNet layer array head by @mikeoliphant in #296
• Bumped NEURAL_AMP_MODELER_DSP_VERSION_PATCH by @benediktadams in #302
• Add .nam file validation by @sdatkinson in #315
• Remove a per-call allocation and two redundant copies from the hot path by @constk in #312
• Support slimmable (SlimmableContainer) files in bench_a2_fast by @jfsantos in #286
• [BUGFIX] Fix testing for WaveNet configurable gating by @sdatkinson in #316
• Formatting by @sdatkinson in #317
• Cache A2 prewarm state by @sdatkinson in #319
• Fix layer1x1 post FiLM for non-blended gating by @Helshr in #322
• Apply layer1x1_post_film under every gating mode by @phillipmself in #321
• [FEATURE] Sequential model by @sdatkinson in #323
• Clarify A2 feature-test fixture by @sdatkinson in #325
• [PERFORMANCE] Optimize long linear convolution scheduling by @sdatkinson in #324
• [FEATURE] Support models variable sample rate by @sdatkinson in #333
• [FEATURE, BREAKING] Support 1-to-n and n-to-1 channel configurations in Linear models by @sdatkinson in #335
• Bump version to 0.6.0 by @sdatkinson in #336
• [FEATURE] Instantiate (linear) models from .wav files by @sdatkinson in #337

New Contributors

• @benediktadams made their first contribution in #302
• @constk made their first contribution in #312
• @Helshr made their first contribution in #322
• @phillipmself made their first contribution in #321

Full Changelog: v0.5.4...v0.6.0