Arduino_TensorFlowLite_ESP32
Arduino_TensorFlowLite_ESP32 copied to clipboard
The micro_speech_ESP_EYE example does not work correctly.
I have an Esp32 Wrover-e board with MP34DT05 microphone. I am running the micro_speech example but I am not getting the correct result. It always prints "no" even though it is not said. What do I need to change to get the correct result?
It's difficult to recognize. A new version is out and is in preparation.
Tip. You need to pronounce the words for recognition. It is difficult to pronounce in a normal way.
Yes, I know I have to pronounce words to be recognized. I don't get the same issue for "yes".
I think the preset model is bad. The accuracy is not high. However, since it is an old version, it may be improved if it is the latest version.
When will the new version be released?
The Arduino version is in the works.
https://www.tensorflow.org/lite/microcontrollers The official ESP-IDF version is recommended.