Add roundoff to quantization calculation so that precision is not lost in in edge cases or when quantization error propagates through model layes. Signed-off-by: Sayooj K Karun <sayooj@aerlync.com> |
||
|---|---|---|
| .. | ||
| hello_world | ||
| magic_wand | ||
| tflm_ethosu | ||
| tflite-micro.rst | ||