15集終于編譯成功了-個球!編譯TFLite Micro語音識別工程-《MCU嵌入式AI開發筆記》
還是參考這個官方文檔:
https://codelabs.developers.google.cn/codelabs/sparkfun-tensorflow#2
全是干貨!
這里面提到的這個Micro工程已經移開了:
https://github.com/tensorflow/tensorflow/tree/master/tensorflow/lite/micro
之后我們重新git clone 這個microlite
git clone https://github.com/tensorflow/tflite-micro.git
之后我們編譯這個工程:
make -f tensorflow/lite/micro/tools/make/Makefile TARGET=sparkfun_edge micro_speech_bin
又出錯:
Cloning into ‘tensorflow/lite/micro/tools/make/downloads/pigweed’…
fatal: unable to access ‘https://pigweed.googlesource.com/pigweed/pigweed/’: Failed to connect to pigweed.googlesource.com port 443: Connection refused
tensorflow/lite/micro/tools/make/Makefile:614: *** Something went wrong with the pigweed download: . Stop.
這里面pigweed.googlesource.com port 443 我們無法訪問。
我們還要解決一下:
這個make file 的600多行處:
這里面需要修改一下:參考如下文章:
https://bbs.elecfans.com/jishu_2289346_1_1.html
vi tensorflow/lite/micro/tools/make/pigweed_download.sh
我們修改一下這個文件
修改成:
之后在編譯一下,可以了,
但是又出錯啦!
繼續填坑!
休息,帶小孩清華里面逛一圈去,吸取一下養分,也是無語啦!