tpoisonooo

Thanks for your contribution and we appreciate it a lot. The following instructions would make your pull request more healthy and more easily get feedback. If you do not understand...

Thanks for your contribution and we appreciate it a lot. The following instructions would make your pull request more healthy and more easily get feedback. If you do not understand...

docs(build): use python -m pip instead of pip

4

## Motivation Fix python requirements install description to avoid Python version problem. They are different: ```shell (base) khj@khj:~/mmdeploy$ python3 -m pip --version pip 21.2.4 from /mnt/d/khj/miniconda3/lib/python3.9/site-packages/pip (python 3.9) (base) khj@khj:~/mmdeploy$...

improvement(cmake): simplify build option and doc

1

Thanks for your contribution and we appreciate it a lot. The following instructions would make your pull request more healthy and more easily receiving feedbacks. If you do not understand...

docs(project): sync en and zh docs

3

Thanks for your contribution and we appreciate it a lot. The following instructions would make your pull request more healthy and more easily receiving feedbacks. If you do not understand...

Thanks for your contribution and we appreciate it a lot. The following instructions would make your pull request more healthy and more easily receiving feedbacks. If you do not understand...

mmdeploy 2022 开源夏令营

20

夏令营只有学生能参加 夏令营只有学生能参加 夏令营只有学生能参加 ## 背景 * GLCC 的经费确实有限 * 参加报名的学生确实都很优秀 OpenMMLab mmdeploy repo 单独准备了一些题目和礼品，作为开源 GLCC 夏令营的补充。 有兴趣的同学，请直接评论自己的 **微信** 账号，附加题目名字作为认领方式。 ## 题目 ### 一、改善 arm Linux 部署体验 【难度】⭐⭐⭐ 【描述】jetson nano 有...

fix gcc7.5 compile error

15

## Context I tried to compile [google/uVkCompute](https://github.com/google/uVkCompute) with GCC7.5, got some error: ```shell ... uVkCompute/third_party/benchmark/include/benchmark/benchmark.h:1204:41: error: field ‘benchmark::RegisterBenchmark(const char*, Lambda&&, Args&& ...) [with Lambda = void (&)(benchmark::State&, uvkc::vulkan::Device*, uvkc::benchmark::LatencyMeasureMode, const...

layernorm+add support int8

4

* 实现了 int8 layernorm，imagenet1k 只量化 LN 精度 84.01->83.68 *

improve vit int8 mha opr

5

这个 PR 写了 mha 的 x86 简单优化： * avx512f 指令集， ViT 从 670 -> 645ms * avx2 指令集，ViT int8 从 1662 -> 1240ms 现在整个网络是 GEMM 比较慢 codebase 是 [PR 3940](https://github.com/Tencent/ncnn/pull/3940)...