Closed schmiph2 closed 2 years ago
Yes. Danilo Pau's team from ST has been using QKeras for similar applications. For example, https://cms.tinyml.org/wp-content/uploads/talks2020/tinyML_Talks_Danilo_Pietro_Pau_210720.pdf. He also presented work in this year tinyml submit, which showed promising results!
Hi everyone
Thank you for this fantastic framework! I came here, after i found out, that TF lite doesn't support quantization-aware training for RNNs. However, I'm interested on deploying the quantized model on a microcontroller instead of an FPGA.
Do you think your framework is suitable for this? Do you see any problems with it?