tanakamasayuki / Arduino_TensorFlowLite_ESP32

Allows you to run machine learning models locally on your ESP32 device.
Apache License 2.0
218 stars 48 forks source link

Didn't find op for builtin opcode 'CONV_2D' version '2' #3

Open vjsyong opened 3 years ago

vjsyong commented 3 years ago

Can't perform inference on a quantized model that was trained on TF2.3. It crashes when it tries to perform a CONV2D layer inference.

Rebooting... ets Jun 8 2016 00:22:57

rst:0xc (SW_CPU_RESET),boot:0x13 (SPI_FAST_FLASH_BOOT) configsip: 0, SPIWP:0xee clk_drv:0x00,q_drv:0x00,d_drv:0x00,cs0_drv:0x00,hd_drv:0x00,wp_drv:0x00 mode:DIO, clock div:1 load:0x3fff0018,len:4 load:0x3fff001c,len:1216 ho 0 tail 12 room 4 load:0x40078000,len:9720 ho 0 tail 12 room 4 load:0x40080400,len:6352 entry 0x400806b8 MPU6050 Found! Didn't find op for builtin opcode 'CONV_2D' version '2'

Failed to get registration from op code d

Guru Meditation Error: Core 1 panic'ed (StoreProhibited). Exception was unhandled. Core 1 register dump: PC : 0x400d1750 PS : 0x00060f30 A0 : 0x800e67b8 A1 : 0x3ffb1f20
A2 : 0x000023a0 A3 : 0x3ffbec00 A4 : 0x3ffb85c8 A5 : 0x80000020
A6 : 0x00000000 A7 : 0x3ffba264 A8 : 0x000002f8 A9 : 0x3ffd92f0
A10 : 0x000023a0 A11 : 0x3ffb1f24 A12 : 0x3ffb1f48 A13 : 0x3ffb1f6c
A14 : 0x00000155 A15 : 0x3ffb85c8 SAR : 0x00000003 EXCCAUSE: 0x0000001d
EXCVADDR: 0x000023a0 LBEG : 0x4000c46c LEND : 0x4000c477 LCOUNT : 0x00000000

Backtrace: 0x400d1750:0x3ffb1f20 0x400e67b5:0x3ffb1fb0 0x40088b49:0x3ffb1fd0

tanakamasayuki commented 3 years ago

https://www.tensorflow.org/lite/microcontrollers/build_convert Have you converted to microcontrollers?

https://github.com/tensorflow/tensorflow/tree/master/tensorflow/lite/micro/examples/hello_world#deploy-to-esp32 The version of the Arduino is old and the latest needs to be built with ESP-IDF4.

amanglarus commented 1 year ago

hey i am also getting same error i am using tf 2.8.0 can you let me know how can resolve this error

amanglarus commented 1 year ago

https://www.tensorflow.org/lite/microcontrollers/build_convert Have you converted to microcontrollers?

https://github.com/tensorflow/tensorflow/tree/master/tensorflow/lite/micro/examples/hello_world#deploy-to-esp32 The version of the Arduino is old and the latest needs to be built with ESP-IDF4.

hey please resolve my problem