v1.3.1

• Fix Linux release tag

v1.3.0

• Optimized concat, batched softmax, pad, sub, broadcast, along with other ops
• Refactored and improved operator splitting
• Cleaned up loading from flash to support loading weights from flash, tile, or DDR
• Added memory analysis report
• Added more examples
• Updated documentation

v1.2.0

• Added support for int16 ops
• Optimized transpose conv, softmax, slice, along with other ops
• Added denoise audio example
• Updated tensorflow to 2.15
• Updated documentation

v1.1.1

• Fix library inclusion bug

v1.1.0

• Include compiled library to simplify building for xcore
• Better offline memory planning
• Add example apps
• Updated documentation

v1.0.0

• Updated tensorflow to 2.12.0
• Better offline memory planning
• Operation split support
• Multiple models in flash
• Fast flash library
• Bug fixes
• Updated documentation

v0.2.1

• Faster flash
• Multithreading in compiled models
• Add support for more TFlite ops
• Bug fixes

v0.2.0

• Integrated tflite-micro-compiler with xformer
• Multithreading support
• Generalized load from flash
• Optimizations and bugfixes

v0.1.9

• Update to Tensorflow 2.10.rc0 and larq 6aa942272db79dabe076821f87826e2c0223d372

v0.1.8

• Added Windows build support
• Refactor Python interface to support both host and device
• Update Tensorflow to 2.8