Infrastructure to enable deployment of ML models to low-power resource-constrained embedded targets (including microcontrollers and digital signal processors).
Role in this project:
ML Engineer Contributions:100 reviews, 31 commits, 113 PRs in 6 months
Contributions summary:TANMAY primarily contributed to the Xtensa quantize kernel, adding support for float to int16 conversion. They addressed issues related to the SEANet NN-only benchmark by modifying quantization operations. The user made modifications to the quantize and requantize functions, including specific adaptations for the HIFI5 architecture. They also made improvements by addressing typo fixes and refactoring the quantization code.
embeddedmicrocontrollermlmodel
Contributions:1 PR, 1 push, 1 branch in 2 years 10 months