RK3588 Exception: Unsupported run platform: Linux aarch64

[yongju-Shin]

I am using an RK3588 platform and Python 3.8.10, with rknn_toolkit_lite2-2.3.0-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl installed.

I also downloaded the librknnrt.so file from the following link:
https://github.com/rockchip-linux/rknpu2/raw/master/runtime/RK3588/Linux/librknn_api/aarch64/librknnrt.so

Then, I placed it in /usr/lib/ using the following command:

sudo mv librknnrt.so /usr/lib/librknnrt.so

However, when I try to run the YOLO example, I encounter the following error.

/rknn_model_zoo/examples/yolov8/python$ python3 yolov8.py --model_path /home/linaro/yolov8.rknn --target rk3588 --img_show
W rknn-toolkit-lite2 version: 2.3.0
--> Init runtime environment
E Catch exception when init runtime!
E Traceback (most recent call last):
  File "/home/linaro/.local/lib/python3.11/site-packages/rknnlite/api/rknn_lite.py", line 148, in init_runtime
    self.rknn_runtime = RKNNRuntime(root_dir=self.root_dir, target=target, device_id=device_id,
                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "rknnlite/api/rknn_runtime.py", line 363, in rknnlite.api.rknn_runtime.RKNNRuntime.__init__
  File "rknnlite/api/rknn_runtime.py", line 607, in rknnlite.api.rknn_runtime.RKNNRuntime._load_library
  File "rknnlite/api/rknn_runtime.py", line 602, in rknnlite.api.rknn_runtime.RKNNRuntime._get_rknn_api_lib_path
Exception: Unsupported run platform: Linux aarch64

Init runtime environment failed

Os : Debian GNU/Linux 12

Is there any solution for this issue?

[Keele520]

Please try use with this librknnrt.so
https://github.com/airockchip/rknn-toolkit2/blob/master/rknpu2/runtime/Linux/librknn_api/aarch64/librknnrt.so

[yongju-Shin]

Thank you for your response.
However, I am still encountering the same issue.
Are there any other solutions?

[Keele520]

open this filer "rknn_model_zoo/py_utils/rknn_executor.py"

edit from :

print('--> Init runtime environment')
if target==None:
ret = rknn.init_runtime()
else:
ret = rknn.init_runtime(target=target, device_id=device_id)

to :

print('--> Init runtime environment')
#if target==None:
# ret = rknn.init_runtime()
#else:
# ret = rknn.init_runtime(target=target, device_id=device_id)
ret = rknn.init_runtime()

then try run again your code.

[yongju-Shin]

Thank you for your help in resolving the issue.

[KZI-22]

Did you solve this issue? I have met the same question

[lgffllsxx]

Maybe I can share a suggestion. When calling rknn.init_runtime(), the target parameter is mainly used for remote debugging via ADB.
If that doesn’t work in your case, you could try running it directly on the RK3588. In that case, you don’t need to provide the target or device_id parameters.
You can also optionally specify core_mask=RKNNLite.NPU_CORE_AUTO to choose which NPU core to use.
I ran into a similar issue when using rknn_lite, and this approach worked for me — hopefully it helps you as well.

[xinmou0129]

Maybe I can share a suggestion. When calling rknn.init_runtime(), the target parameter is mainly used for remote debugging via ADB. If that doesn’t work in your case, you could try running it directly on the RK3588. In that case, you don’t need to provide the target or device_id parameters. You can also optionally specify core_mask=RKNNLite.NPU_CORE_AUTO to choose which NPU core to use. I ran into a similar issue when using rknn_lite, and this approach worked for me — hopefully it helps you as well.

Thank you!