search

AlexeyAB / darknet

YOLOv4 / Scaled-YOLOv4 / YOLO - Neural Networks for Object Detection (Windows and Linux version of Darknet )
http://pjreddie.com/darknet/
Other
21.69k stars 7.96k forks source link

How to training YOLO by RTX 3080 #6976

Open liucy2020 opened 3 years ago

liucy2020 commented 3 years ago

If something doesn’t work for you, then show 2 screenshots:

  1. screenshots of your issue
  2. screenshots with such information 
    ./darknet detector test cfg/coco.data cfg/yolov4.cfg yolov4.weights data/dog.jpg
CUDA-version: 10000 (10000), cuDNN: 7.4.2, CUDNN_HALF=1, GPU count: 1
CUDNN_HALF=1
OpenCV version: 4.2.0
0 : compute_capability = 750, cudnn_half = 1, GPU: GeForce RTX 2070
net.optimized_memory = 0
mini_batch = 1, batch = 8, time_steps = 1, train = 0
layer   filters  size/strd(dil)      input                output
0 conv     32       3 x 3/ 1    608 x 608 x   3 ->  608 x 608 x  32 0.639 BF

If you do not get an answer for a long time, try to find the answer among Issues with a Solved label: https://github.com/AlexeyAB/darknet/issues?q=is%3Aopen+is%3Aissue+label%3ASolved

liucy2020 commented 3 years ago

When I training test darknet by RTX3080,it dosenot work. Is the GPU RTX3080 not compatible with darknet?How to resolve it? win 10 ,RTX 308, vs 2017，CUDA10.0 CUDNN7.6.5 ,OPENCV 3.4

bulatnv commented 3 years ago

Hello @liucy2020.

CUDA 10 do not suppost RTX30-series GPU ! #6970

Reinstall everything from scratch. For Ampere install latest driver 450+, install at least CUDA 11.0, cuDnn 8.0+. For Windows build from scratch vpkg, darknet.

Best of luck!

liucy2020 commented 3 years ago

problem is solved,thank you very much !发自我的iPhone------------------ Original ------------------From: Bulat notifications@github.comDate: Sat,Nov 14,2020 6:14 PMTo: AlexeyAB/darknet darknet@noreply.github.comCc: liucy2020 liuchuanyang608@163.com, Mention mention@noreply.github.comSubject: Re: [AlexeyAB/darknet] How to training YOLO by RTX 3080 (#6976) Hello @liucy2020. RTX30-series GPU do not suppost CUDA 10! Reinstall everything from scratch. For Ampere install latest driver 450+, install at least CUDA 11.0, cuDnn 8.0+. For Windows build from scratch vpkg, darknet. Best of luck!

—You are receiving this because you were mentioned.Reply to this email directly, view it on GitHub, or unsubscribe. [ { "@context": "http://schema.org", "@type": "EmailMessage", "potentialAction": { "@type": "ViewAction", "target": "https://github.com/AlexeyAB/darknet/issues/6976#issuecomment-727181227", "url": "https://github.com/AlexeyAB/darknet/issues/6976#issuecomment-727181227", "name": "View Issue" }, "description": "View this Issue on GitHub", "publisher": { "@type": "Organization", "name": "GitHub", "url": "https://github.com" } } ]

AlgirdasKartavicius commented 3 years ago

stall latest driver 450+, install at least CUDA 11.0, cuD

Have you successfully build vcpkg and CUDA 11.1 on windows?

liucy2020 commented 3 years ago

I only use CUDA 11.1,cudnn 8.0,restore is Ok.发自我的iPhone------------------ Original ------------------From: AlgirdasKartavicius notifications@github.comDate: Sat,Nov 14,2020 11:57 PMTo: AlexeyAB/darknet darknet@noreply.github.comCc: liucy2020 liuchuanyang608@163.com, Mention mention@noreply.github.comSubject: Re: [AlexeyAB/darknet] How to training YOLO by RTX 3080 (#6976)

stall latest driver 450+, install at least CUDA 11.0, cuD

Have you successfully build vcpkg and CUDA 11.1 on windows?

—You are receiving this because you were mentioned.Reply to this email directly, view it on GitHub, or unsubscribe. [ { "@context": "http://schema.org", "@type": "EmailMessage", "potentialAction": { "@type": "ViewAction", "target": "https://github.com/AlexeyAB/darknet/issues/6976#issuecomment-727227037", "url": "https://github.com/AlexeyAB/darknet/issues/6976#issuecomment-727227037", "name": "View Issue" }, "description": "View this Issue on GitHub", "publisher": { "@type": "Organization", "name": "GitHub", "url": "https://github.com" } } ]

wentao-yolo-1996 commented 3 years ago

I meet the same problem,the difference is that i use a rtx3070,but i still couldn't figure it out after reinstall cuda11.1. this is my qq number:1058574635,could you please share more?#6970

liucy2020 commented 3 years ago

我的QQ280176079发自我的iPhone------------------ Original ------------------From: wentao-yolo-1996 notifications@github.comDate: Sun,Nov 15,2020 9:38 AMTo: AlexeyAB/darknet darknet@noreply.github.comCc: liucy2020 liuchuanyang608@163.com, Mention mention@noreply.github.comSubject: Re: [AlexeyAB/darknet] How to training YOLO by RTX 3080 (#6976) I meet the same problem,the difference is that i use a rtx3070,but i still couldn't figure it out after reinstall cuda11.1. this is my qq number:1058574635,could you please share more?#6970

—You are receiving this because you were mentioned.Reply to this email directly, view it on GitHub, or unsubscribe. [ { "@context": "http://schema.org", "@type": "EmailMessage", "potentialAction": { "@type": "ViewAction", "target": "https://github.com/AlexeyAB/darknet/issues/6976#issuecomment-727292749", "url": "https://github.com/AlexeyAB/darknet/issues/6976#issuecomment-727292749", "name": "View Issue" }, "description": "View this Issue on GitHub", "publisher": { "@type": "Organization", "name": "GitHub", "url": "https://github.com" } } ]

bulatnv commented 3 years ago

Have you successfully build vcpkg and CUDA 11.1 on windows?

@AlgirdasKartavicius, @liucy2020, @wentao-yolo-1996 Yes i have. This instructions explains how to build darknet for rtx 3080, 3090.

liaobear commented 3 years ago

Hello @liucy2020.

CUDA 10 do not suppost RTX30-series GPU ! #6970

Reinstall everything from scratch. For Ampere install latest driver 450+, install at least CUDA 11.0, cuDnn 8.0+. For Windows build from scratch vpkg, darknet.

Best of luck!

it can test with GPU but when i training it use cpu. have any idea about solve training by GPU? RTX3080 cuda 11.0 cuDnn 8.0+ windows 10 darknet.

bulatnv commented 3 years ago

test with GPU

Hello, a sure that you testing with gpu?

liucy2020 commented 3 years ago

Yes 发自我的iPhone------------------ Original ------------------From: Bulat @.>Date: Fri,Jul 2,2021 8:07 PMTo: AlexeyAB/darknet @.>Cc: liucy2020 @.>, Mention @.>Subject: Re: [AlexeyAB/darknet] How to training YOLO by RTX 3080 (#6976)

test with GPU

Hello, a sure that you testing with gpu?

—You are receiving this because you were mentioned.Reply to this email directly, view it on GitHub, or unsubscribe. [ { @.": "http://schema.org", @.": "EmailMessage", "potentialAction": { @.": "ViewAction", "target": "https://github.com/AlexeyAB/darknet/issues/6976#issuecomment-872948157", "url": "https://github.com/AlexeyAB/darknet/issues/6976#issuecomment-872948157", "name": "View Issue" }, "description": "View this Issue on GitHub", "publisher": { @.": "Organization", "name": "GitHub", "url": "https://github.com" } } ]

liucy2020 commented 3 years ago

CUDA of darknet.vcxproj is revised to CUDA 11.1发自我的iPhone------------------ Original ------------------From: Bulat @.>Date: Fri,Jul 2,2021 8:07 PMTo: AlexeyAB/darknet @.>Cc: liucy2020 @.>, Mention @.>Subject: Re: [AlexeyAB/darknet] How to training YOLO by RTX 3080 (#6976)

test with GPU

Hello, a sure that you testing with gpu?

—You are receiving this because you were mentioned.Reply to this email directly, view it on GitHub, or unsubscribe. [ { @.": "http://schema.org", @.": "EmailMessage", "potentialAction": { @.": "ViewAction", "target": "https://github.com/AlexeyAB/darknet/issues/6976#issuecomment-872948157", "url": "https://github.com/AlexeyAB/darknet/issues/6976#issuecomment-872948157", "name": "View Issue" }, "description": "View this Issue on GitHub", "publisher": { @.": "Organization", "name": "GitHub", "url": "https://github.com" } } ]

liaobear commented 3 years ago

Uh,i wrong,no test. I use demo , when i demo, the cuda usage rate up to 80%