Future development ideas

[cpetig]

This issue contains discussion about potential long term improvements of the compiler

[cpetig]

• Find a way to make reshape a no-op (by implementing in-place operators inside tflite?)
  see Inplace operator planning #53
• Find a way to support runtime calculated quantization (really nice feature, but sadly not yet supported by tf lite, see https://arxiv.org/pdf/1906.00532.pdf, especially figure 1 (this uses the combination of Quantize,QuantizedConv2D,RequantizationRange, Requantize...,Dequantize, as supported by graph transform), see also Add Support for Full-Integer TensorFlow Lite models in TensorFlow 2.0 tensorflow/tensorflow#38285 ). https://petewarden.com/2016/05/03/how-to-quantize-neural-networks-with-tensorflow/ also has good explanations of the methods involved.

[cpetig]

• Deploy an LSTM example as a more complex test case (I am currently working on this)
  Update: I was finally able to get an LSTM defined, trained and converted. I will write about the process and how to navigate the many pitfalls in the coming days.
  Update: It works fine 2c21a36