-
Referring the Doc, I tried to build the kernal to fit rkllm toolkit. But there are too many errors.
Firstly, there are too many implicit declaration of function such as 'vm_flags_clear', 'vm_flags_se…
-
1. Some blogs said that the pt model should be exported to onnx with the parameter opt=12, but I found that you did not set the parameter.
2. Can you provide some script to export the rknn from onnx?…
-
On the board whose features I have given below, it takes more than 2 minutes for the retina face to examine 1 single photo, and it takes a very long time for the camera to be activated. The only thing…
-
使用的是1.7.3版本,是自己跑的lighttrack的onnx转成的rknn模型,在精度分析这里卡住了。不知道是什么错误
```D Using CPPUTILS: True
E Catch exception when snapshot: AttributeError("'Namespace' object has no attribute 'can_snapshot'",)
E Trac…
7yzx updated
6 months ago
-
请问一下能否适配一下yolov8-pose,对我们很有用,谢谢。
-
![Screenshot from 2024-03-27 17-02-24](https://github.com/airockchip/rknn-llm/assets/45985213/9d8e5536-1e7e-4337-8ba6-fae57888f830)
system memory : 64GB
IDEA memory : 16GB
model name : LLaMA-7b
-
**请将下面信息填写完整,便于我们快速解决问题,谢谢!**
**问题描述**
请在此处详细的描述报错信息
[Paddle2ONNX] Oops, there are some operators not supported yet, including deformable_conv,matrix_nms,
[ERROR] Due to the unsupported operator…
-
*********************************************
温馨提示:根据社区不完全统计,按照模板提问,可以加快回复和解决问题的速度
*********************************************
## 环境
- 【FastDeploy版本】: 说明具体的版本,如fastdeploy-linux-gpu-0.8.0
- …
-
**Describe the bug**
PhotonVision does not Run on Arm32. It looks like 64bit JNIs are used in the 32 bit build.
**To Reproduce**
I created my own Pi image to run on a Pi Zero W. (using the RPi …
-
ubuntu20,mediapipe=0.10.11
task:pose landmark detection
problem:The inference results of tflite and Python APIs are inconsistent
expected:The inference results of tflite and Python APIs should be c…