wangcker

Results 12 comments of wangcker

也没有找到scikit-learn py3.8/py3.9 arm版本

自行拉取release/2.1代码,修改paddlex=>scikit-learn lap=>numpy版本依赖,并打包安装,现已经可以自行调用paddlex api进行推理。

> 您好,能说的详细一些吗,我也遇到了这个问题 一,修改lap 1 下载lap最新源文件 2 将setup.py中的numpy依赖去掉或将numpy版本校验代码去掉(后一种比较粗暴,我用的后一种) 3 重新打包安装 二, 修改paddlex 1 git clone paddlex release/2.1 2 requirements.txt和setup.py将scikit-learn==0.23.2修改为scikit-learn>=0.23.2 3 python setup.py bdist_wheel 4 pip install dist/xx.wheel 然后就可以愉快的推理了,不过调用picodet目标检测耗时200多ms,不知道是硬件问题还是适配问题

the similar bugs in release mode: ``` [ +7 ms] I/flutter (28140): time 77.142 [ +10 ms] E/flutter (28140): [ERROR:flutter/impeller/renderer/backend/vulkan/allocator_vk.cc(528)] Impeller validation: Unable to allocate a device buffer: ErrorOutOfDeviceMemory [...

the whole code : ``` import 'dart:math'; import 'package:flutter/material.dart'; import 'package:flutter_scene/camera.dart'; import 'package:flutter_scene/node.dart'; import 'package:flutter_scene/scene.dart'; import 'package:vector_math/vector_math.dart'; void main() { runApp(const MyApp()); } class MyApp extends StatefulWidget { const MyApp({super.key});...

> @wangcker did you run using flutter scene locally like in #38 ? I'm curious if were you able to resolve that issue when you ran the demo no,the problem...

> @wangcker did you run using flutter scene locally like in #38 ? I'm curious if were you able to resolve that issue when you ran the demo i fix...

* it is hardware * testing on xiaomi 14 pro it`s may be producted by memory leak