Comments (6)
非常感谢,我使用的是nvidia v100 GPU, 上面提问时第三行有个错误,把YOLO V4写成YOLOV3了。 V4 keras版本我训练时的batchsize配置单GPU: freeze 32 unfreeze 8 ,超出时就会报错OOM。 8块GPU时 freeze 168 unfreeze 18,超过时就会报错。 这是在input_shape = (608,608)情况下, 在(416,416)情况下 unfreeze 可以配置为 16 或者 28 。 我在使用YOLOV3训练时,8块GPU (416,416)情况下, freeze 1288 unfreeze 32*8 由于batchsize配置较大,训练速度非常快,比YOLOV4快了10倍以上。
from keras-yolo4.
This usally means that data you are supplying is overwhelming the GPU. Can you tell us what GPU you have, watch batch settings you are using, if this happens during training (I assume so), and any other relevant information.
from keras-yolo4.
from keras-yolo4.
我尝试了很多次。 frozen to just 3, batchsize one GPU可以设置到32, 8 GPU设置为 4X8,更大的batchsize就会报错 OOM, unfrozen后,one GPU 必须设置8 以及以下, 8GPU必须设置1X8 或者更小,。
我目前最大的问题是YOLOV3 完成一个任务的训练只需要20多个小时,因为batchsize配置的很大, YOLOV4 unfrozen后batchsize只能配置8,训练时间非常的长,40万数据1个epoch需要10个小时,通常我需要训练50个epoch, 2个月才能完成一个模型训练,
from keras-yolo4.
我尝试了很多次。 frozen to just 3, batchsize one GPU可以设置到32, 8 GPU设置为 4X8,更大的batchsize就会报错 OOM, unfrozen后,one GPU 必须设置8 以及以下, 8GPU必须设置1X8 或者更小,。
我目前最大的问题是YOLOV3 完成一个任务的训练只需要20多个小时,因为batchsize配置的很大, YOLOV4 unfrozen后batchsize只能配置8,训练时间非常的长,40万数据1个epoch需要10个小时,通常我需要训练50个epoch, 2个月才能完成一个模型训练,
I understand. I too am confused about why train.py works like this. Its confusing. I am starting to look for a tensorflow - yolo4 implementation to see how it performs. Also... I wish I had your GPU's! :-)
from keras-yolo4.
我尝试了很多次。 frozen to just 3, batchsize one GPU可以设置到32, 8 GPU设置为 4X8,更大的batchsize就会报错 OOM, unfrozen后,one GPU 必须设置8 以及以下, 8GPU必须设置1X8 或者更小,。
我目前最大的问题是YOLOV3 完成一个任务的训练只需要20多个小时,因为batchsize配置的很大, YOLOV4 unfrozen后batchsize只能配置8,训练时间非常的长,40万数据1个epoch需要10个小时,通常我需要训练50个epoch, 2个月才能完成一个模型训练,
I am also surprised the author of the code does not respond. He must be busy.
from keras-yolo4.
Related Issues (20)
- 恕我直言qqwweee那版yolo3就是个废仓库 HOT 13
- Poor detection when trained on custom dataset HOT 19
- Error while converting to keras model HOT 11
- MemoryError
- keras和tensorflow的版本? HOT 1
- Your some layers are different from miemie2013's. HOT 1
- How about tiny yolov4?
- Error while running test.py
- Did you test accuracy of converted YOLOv4-model on MSCOCO?
- What do the four parameters after loss mean during training, loss of each category or other else?
- [BUG] iou computation in preprocess_true_boxes() should not use scaled bbox
- ValueError: Shapes (1, 1, 1024, 75) and (255, 1024, 1, 1) are incompatible HOT 5
- Batching with inference
- train.py
- Can't seems to be working with yolo tiny
- 网络结构层数对应问题
- convert h5 file
- 你好 ! 我不能導入 yolo4.py 的類別 YOLO
- ValueError: height and width must be > 0 - stable diffusion animation
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from keras-yolo4.