Error: Init runtime environment failed

#1

I guess NPU is found in the lsusb output.

[toybrick@localhost ~]$ sudo lsusb

Bus 008 Device 001: ID 1d6b:0003 Linux Foundation 3.0 root hub
Bus 007 Device 004: ID 2207:180a
Bus 007 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 004 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub
Bus 002 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 003 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub
Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub

But I am not able to run mobilenet_v1 example from the rknn-toolkit-v1.2.0/example/mobilenet_v1/test.py.

[toybrick@localhost mobilenet_v1]$ python3 test.py
–> config model
done
–> Loading model
done
–> Building model
done
–> Export RKNN model
done
–> Init runtime environment
E Using device with adb mode to init runtime, but npu_transfer_proxy is running, it may cause conflict. Please terminate npu_transfer_proxy first.
E Catch exception when init runtime!
T Traceback (most recent call last):
T File “rknn/api/rknn_base.py”, line 769, in rknn.api.rknn_base.RKNNBase.init_runtime
T File “rknn/api/rknn_runtime.py”, line 142, in rknn.api.rknn_runtime.RKNNRuntime.init
T File “rknn/api/rknn_platform_utils.py”, line 265, in rknn.api.rknn_platform_utils.start_ntp_or_adb
T Exception: Init runtime environment failed!
Init runtime environment failed

Edit:
I tried to update the NPU firmware:

[toybrick@localhost mobilenet_v1]$ sudo npu_upgrade MiniLoaderAll.bin uboot.img trust.img boot.img
init gpio: 4
Error writing to /sys/class/gpio/export value=4: Device or resource busy
gpio init
set clk_en 1 to 0
set clk_en 0 to 1
powerup
start to write trust…

Is the firmware update successful? I see an error “writing to /sys/class/gpio/export value=4: Device or resource busy”. Is it fine?

#2

This is the Error:

E Using device with adb mode to init runtime, but npu_transfer_proxy is running, it may cause conflict. Please terminate npu_transfer_proxy first.

I think @setq can explain why it happens.

#3

Hi yokonav,

What OS do you use?

#4

I am using latest fedora image for rockpin10.

#5

Please tell me the content of init_runtime, which is about line 66 of test.py.

#6

Init_runtime has no parameters.

#7

That’s right. I’ll verify that tomorrow.

#8

We have identified the problem and are working to fix and verify it.

#9

Hi yokonav,

We updated the NPU firmware and uploaded the system image.

You can download it and reburn it, or update NPU firmware as follows:

wget https://rock.sh/npu-fw-update -O npu-fw.zip
sudo unzip -o npu-fw.zip -d /
sudo reboot

By the way, you can also use system image, https://dl.radxa.com/rockpin10/images/fedora/rockpin10_fedora_28_server_edition_arm64_20200114_1533-gpt.img.gz.

#10

Thanks @setq, it is fixed now. By the way why loading the python module and building model takes so long to complete?

#11

Model conversion take a long time. You can load the converted model (.rknn) directly and you will find it very fast.