Tencent/ncnn issues and pull requests

#4593 - 请问ncnn适用生成对抗网络吗

Issue - State: closed - Opened by JWQZ over 1 year ago - 2 comments

#4593 - 请问ncnn适用生成对抗网络吗

Issue - State: closed - Opened by JWQZ over 1 year ago - 2 comments

#4592 - 推理三个output该写多少？

Issue - State: closed - Opened by jia0511 over 1 year ago - 7 comments

#4586 - 如何查看一个.bin是否经过了量化或者fp16优化？

Issue - State: open - Opened by Mactarvish over 1 year ago - 1 comment

#4574 - ncnn在ARMv7交叉编译后上板执行报错Arithmetic exception

Issue - State: closed - Opened by shone666 over 1 year ago - 9 comments
Labels: bug

#4571 - 树莓派编译vulkan过程中，出现了gfxrecon-replay库的错误

Issue - State: closed - Opened by Hao-Ke-Shui over 1 year ago - 1 comment

#4567 - C++17 is now required in order to build the tools with up-to-date dependencies

Issue - State: closed - Opened by davidebeatrici over 1 year ago - 2 comments

#4558 - 加载量化后的int8模型，卷积的weight_data指针为空

Issue - State: closed - Opened by declanwucv over 1 year ago - 6 comments

#4553 - cpu部署问题

Issue - State: open - Opened by Lexiaotian over 1 year ago - 10 comments

#4551 - 请问怎样不使用neon或者其他加速技巧在armv8的板子上跑模型？

Issue - State: open - Opened by Fanchingai over 1 year ago - 5 comments

#4551 - 请问怎样不使用neon或者其他加速技巧在armv8的板子上跑模型？

Issue - State: open - Opened by Fanchingai over 1 year ago - 5 comments

#4542 - vulkan推理结果异常，GPU使用的是高通的Adreno 690，加载模型后会有警告

Issue - State: open - Opened by wyjj over 1 year ago - 3 comments

#4532 - Feature request: int8 for LSTM

Issue - State: closed - Opened by csukuangfj over 1 year ago - 3 comments
Labels: enhancement

#4531 - 双线性插值汇编代码疑问

Issue - State: closed - Opened by fubuzhuirou1 over 1 year ago - 2 comments

#4521 - Read the customized model path and parameters that need to run benchn…

Pull Request - State: closed - Opened by kuanyan9527 over 1 year ago - 2 comments

#4521 - Read the customized model path and parameters that need to run benchn…

Pull Request - State: closed - Opened by kuanyan9527 over 1 year ago - 2 comments

#4520 - is there an example of YOLOV8?

Issue - State: closed - Opened by YFforever2022 over 1 year ago - 5 comments

#4510 - PNNX 如何支持多个input输入？

Issue - State: closed - Opened by zcy618 almost 2 years ago - 1 comment

#4503 - pnnx 导出 nn.Linear 挂掉了

Issue - State: closed - Opened by csukuangfj almost 2 years ago - 5 comments
Labels: bug

#4495 - support yolov8s ncnn

Pull Request - State: closed - Opened by triple-Mu almost 2 years ago - 4 comments

#4494 - Any support for YOLOv8 segemenation along with NCNN???

Issue - State: closed - Opened by apanand14 almost 2 years ago - 4 comments

#4494 - Any support for YOLOv8 segemenation along with NCNN???

Issue - State: closed - Opened by apanand14 almost 2 years ago - 4 comments

#4489 - block issues

Issue - State: closed - Opened by nihui almost 2 years ago - 1 comment
Labels: bug

#4476 - how ncnn support yolov8n?

Issue - State: closed - Opened by superbayes almost 2 years ago - 9 comments

#4476 - how ncnn support yolov8n?

Issue - State: closed - Opened by superbayes almost 2 years ago - 9 comments

#4474 - pt2pnnx转换出现“no attribute value”错误

Issue - State: closed - Opened by Z-Xiong almost 2 years ago - 3 comments

#4471 - [x86] GroupNorm

Pull Request - State: closed - Opened by EdVince almost 2 years ago - 2 comments

#4468 - ncnn::Net net --> global variable doesn't work local variable works

Issue - State: closed - Opened by gedance almost 2 years ago - 6 comments

#4448 - ncnn/pnnx 支持动态shape吗？

Issue - State: closed - Opened by zuowanbushiwo almost 2 years ago - 1 comment

#4447 - Conflict between ncnn and TengineKit

Issue - State: open - Opened by GanchengZhu almost 2 years ago - 2 comments

#4434 - cv::resize from simple_opencv is much more slower than opencv

Issue - State: closed - Opened by luohao123 almost 2 years ago - 1 comment

#4422 - Having issues trying to tile a large image for scaling

Issue - State: open - Opened by davlee1972 almost 2 years ago - 3 comments

#4415 - 为了减少库的体积需要裁剪到一些不用的layer，应该怎么做？

Issue - State: closed - Opened by zuowanbushiwo almost 2 years ago - 6 comments

#4406 - yolov5s模型经过int8量化,无明显加速效果

Issue - State: open - Opened by liyan110120999 almost 2 years ago - 4 comments

#4398 - 构建的网络模型，运行遇到segmentation fault问题

Issue - State: open - Opened by Elasine almost 2 years ago - 6 comments

#4382 - 20190320版本arm与x86推理结果不一致

Issue - State: closed - Opened by ssr920121 almost 2 years ago - 2 comments

#4377 - Slice does not handle negative indexes

Issue - State: closed - Opened by csukuangfj almost 2 years ago - 2 comments
Labels: bug

#4377 - Slice does not handle negative indexes

Issue - State: closed - Opened by csukuangfj almost 2 years ago - 2 comments
Labels: bug

#4373 - Grid sample op

Pull Request - State: closed - Opened by Yoh-Z almost 2 years ago - 17 comments
Labels: tool, pnnx, test, layer, x86, doc

#4370 - fix openmp affinity abort when cpu goes offline

Pull Request - State: closed - Opened by nihui almost 2 years ago - 1 comment

#4369 - 小米10/10Pro 省电模式必现闪退：__kmp_abort_process

Issue - State: closed - Opened by ijero almost 2 years ago - 20 comments
Labels: bug

#4369 - 小米10/10Pro 省电模式必现闪退：__kmp_abort_process

Issue - State: closed - Opened by ijero almost 2 years ago - 20 comments
Labels: bug

#4348 - PNNX转换问题

Issue - State: open - Opened by mawj111 almost 2 years ago - 6 comments

#4348 - PNNX转换问题

Issue - State: open - Opened by mawj111 almost 2 years ago - 6 comments

#4347 - implement MultiheadAttention kdim vdim

Pull Request - State: closed - Opened by nihui almost 2 years ago - 1 comment

#4336 - MSBUILD : error MSB1009: 项目文件不存在。开关:install.vcxproj

Issue - State: open - Opened by yanqianshilu almost 2 years ago - 1 comment

#4331 - 求语义分割的ncnn部署demo，比如unet、DeepLabv3+

Issue - State: open - Opened by mantou69 about 2 years ago - 4 comments

#4329 - 使用vulkan推理vkDestroyDevice时崩溃

Issue - State: closed - Opened by nicheng0019 about 2 years ago - 3 comments

#4327 - 请问ncnn支持nlp模型的应用吗，比如说本文分类任务。

Issue - State: open - Opened by Kevinddddddd about 2 years ago - 5 comments

#4327 - 请问ncnn支持nlp模型的应用吗，比如说本文分类任务。

Issue - State: open - Opened by Kevinddddddd about 2 years ago - 6 comments

#4320 - YOLOV5在特定尺寸下第三层输出异常

Issue - State: closed - Opened by 408550969 about 2 years ago

#4309 - yolov5 6.2 extract "output" crashed: bottom_blob_ref.refcount was nullptr.

Issue - State: closed - Opened by Mactarvish about 2 years ago - 7 comments

#4299 - SwinIR models and RealESRGAN_x2plus cannot be converted to NCNN

Issue - State: open - Opened by theflyingzamboni about 2 years ago - 4 comments

#4294 - Fix linux build error(#4265)

Pull Request - State: closed - Opened by bestpower about 2 years ago - 3 comments

#4290 - 模型在嵌入式设备上的内存占用如何统计？

Issue - State: open - Opened by aureosun about 2 years ago - 1 comment

#4266 - 安卓端最新自编译推理库在部分设备上执行CPU推理时应用崩溃

Issue - State: open - Opened by bestpower about 2 years ago - 9 comments

#4264 - 请问ncnn目前是不支持Mac catalyst吗？

Issue - State: closed - Opened by Mashura about 2 years ago - 1 comment
Labels: enhancement

#4261 - trouble with converted a custom yolov5s.pt

Issue - State: open - Opened by NacerFaraj about 2 years ago - 6 comments

#4241 - 请问怎么构建ncnn的3d mat输入？

Issue - State: open - Opened by Lin-sudo about 2 years ago - 12 comments

#4239 - ld: symbol(s) not found for architecture arm64

Issue - State: closed - Opened by Iron-iii about 2 years ago - 3 comments

#4233 - Many tests fail: LLVM ERROR: Broken function

Issue - State: closed - Opened by yurivict about 2 years ago - 2 comments

#4226 - How to reduce the CPU usage？

Issue - State: closed - Opened by PapaMadeleine2022 about 2 years ago - 1 comment

#4213 - Add qnx.toolchain.cmake

Pull Request - State: closed - Opened by zchrissirhcz about 2 years ago - 19 comments

#4208 - vkCreateDevice failed VK_ERROR_INITIALIZATION_FAILED

Issue - State: closed - Opened by sdnusqy-art about 2 years ago - 2 comments

#4206 - The benchmark shows more lower speed in int8 than float. It cann't satisfy our expection

Issue - State: closed - Opened by aqsc about 2 years ago - 2 comments

#4204 - Update LICENSE.txt

Pull Request - State: open - Opened by victor30608 about 2 years ago - 1 comment

#4203 - Android12 出现闪退问题

Issue - State: open - Opened by gezhongqi1992 about 2 years ago - 15 comments

#4187 - layer_shader_registry.h.in: remove duplicated newline

Pull Request - State: closed - Opened by MollySophia about 2 years ago - 2 comments

#4187 - layer_shader_registry.h.in: remove duplicated newline

Pull Request - State: closed - Opened by MollySophia about 2 years ago - 2 comments

#4175 - raspberry pi 2b

Issue - State: open - Opened by nihui about 2 years ago - 6 comments

#4169 - simpleocv中Mat具有类型转换的ptr方法

Issue - State: closed - Opened by liangshuisanqian about 2 years ago - 2 comments

#4161 - Android编译失败(ncnn-20220729)

Issue - State: closed - Opened by Hi-Jingzhi about 2 years ago - 5 comments

#4158 - 安卓编译问题

Issue - State: open - Opened by Javacr about 2 years ago - 2 comments

#4150 - cmake -DCMAKE_TOOLCHAIN_FILE=../toolchains/himix200.toolchain.cmake ..

Issue - State: open - Opened by fxj8911 about 2 years ago - 2 comments

#4149 - PNNX无法转换torchvision.vit model

Issue - State: open - Opened by JiahangXu about 2 years ago - 5 comments

#4145 - ncnn识别yolov5模型，部署到iOS端，手机发烫，速度变慢。

Issue - State: open - Opened by wutao23yzd about 2 years ago - 6 comments

#4135 - android 10.0版本之前都无法使用GPU

Issue - State: closed - Opened by yuan9778 about 2 years ago - 1 comment

#4130 - android not support vulkan int8 ?

Issue - State: closed - Opened by dachengai about 2 years ago - 3 comments

#4113 - Tip: The proper way to use NCNN with vulkan acceleration in python

Issue - State: closed - Opened by TabuaTambalam over 2 years ago - 3 comments

#4113 - Tip: The proper way to use NCNN with vulkan acceleration in python

Issue - State: closed - Opened by TabuaTambalam over 2 years ago - 3 comments

#4112 - NCNN_VERSION_STRING enhancement

Issue - State: closed - Opened by zchrissirhcz over 2 years ago
Labels: enhancement

#4101 - 加载 PNNX 导出的模型时出现 layer aten::exp not exists or registered

Issue - State: closed - Opened by csukuangfj over 2 years ago - 10 comments

#4097 - 针对YoloV5框架和YoloV7框架添加Python统一调用接口

Issue - State: closed - Opened by Digital2Slave over 2 years ago - 3 comments

#4035 - PNNX inference results are different from TorchScript inference results

Issue - State: open - Opened by yitian8377 over 2 years ago - 1 comment

#4029 - 在intel CPU上测试速度与mnn差距有点大

Issue - State: closed - Opened by ppl21847 over 2 years ago - 5 comments

#4029 - 在intel CPU上测试速度与mnn差距有点大

Issue - State: closed - Opened by ppl21847 over 2 years ago - 5 comments

#4016 - NCNN resize_nearest

Pull Request - State: open - Opened by obtx over 2 years ago - 4 comments

#3998 - x86平台下，int8量化后，modnet模型速度无明显变化

Issue - State: closed - Opened by huaixia30 over 2 years ago - 1 comment

#3990 - 关于使用vulkan时的初始化时间

Issue - State: open - Opened by mike-2020 over 2 years ago - 2 comments
Labels: enhancement

GitHub / Tencent/ncnn issues and pull requests