Comments (21)
看起来貌似teacher-model没有起到作用 是student-model自己在学习,是不是蒸馏损失的权重太小了?
from fgd.
I am researching this direction, can I exchange some questions about FGD with you?
from fgd.
@renyu-renyu ok,by the way, if you are Chinese, we can talk with native language.
from fgd.
@Fight-hawk 你好,可以一起交流吗
from fgd.
@renyu-renyu 可以啊 我最近也在搞知识蒸馏
from fgd.
@renyu-renyu 可以啊 我最近也在搞知识蒸馏
你有在其他算法实现吗
from fgd.
@renyu-renyu 其他目标检测算法么,目前只在yolox上进行了实验
from fgd.
@renyu-renyu 其他目标检测算法么,目前只在yolox上进行了实验
效果怎么样阿
from fgd.
训练150epoch 提升了AP0.5 提高了1.5 mmdet
from fgd.
@www516717402 我没复现精度,在自己的数据集上试了一下,有效果。作者一直拒绝提供完整的训练log,可能是加了一些trick吧
from fgd.
@Fight-hawk @renyu-renyu mmdet是在V2.15.1版本加入的YOLOX(当时很不稳定,而且只有tiny版本复现出了精度),而这个蒸馏用的是mmdet-V2.11版本。是他自己复现的YOLOX-m的baseline是45.9%???但是官方YOLOX-m的baseline是46.9%,不清楚作者用的是什么版本,还有如何复现的精度,这baseline精度差距1%去哪了。
from fgd.
@Fight-hawk 你自己的数据集尺度有coco那么大吗?还有AP0.5提高了1.5%,其它项有降低吗?一句题外话,蒸馏一般不适用YOLOX-L和YOLOX-Nano配合,Gap差距太大非常容易崩,你能有提高也是奇迹
from fgd.
不明白你在逼逼什么 蒸馏一般不适用YOLOX-L和YOLOX-Nano配合 关于你有实践过么 道听途说也能对别人逼逼赖赖?
from fgd.
@Fight-hawk
1)首先非常抱歉打扰你了。
2)蒸馏过人脸识别,大模型和超小模型有gap,很难训练而且效果不好。
3)仅仅是提出我的疑惑,没有任何别的意思。冒犯到你,再次抱歉。
from fgd.
@Fight-hawk @renyu-renyu mmdet是在V2.15.1版本加入的YOLOX(当时很不稳定,而且只有tiny版本复现出了精度),而这个蒸馏用的是mmdet-V2.11版本。是他自己复现的YOLOX-m的baseline是45.9%???但是官方YOLOX-m的baseline是46.9%,不清楚作者用的是什么版本,还有如何复现的精度,这baseline精度差距1%去哪了。
我用的mmdet==2.19做的yolox,mmdet的config跑出来就是45.9..........,对于yolox的修改是有yolox分支
from fgd.
@www516717402 我没复现精度,在自己的数据集上试了一下,有效果。作者一直拒绝提供完整的训练log,可能是加了一些trick吧
没有其他trick,用mmdet==2.19,再根据yolox分支进行修改一下就可以跑到结果的点
from fgd.
@Fight-hawk @renyu-renyu mmdet是在V2.15.1版本加入的YOLOX(当时很不稳定,而且只有tiny版本复现出了精度),而这个蒸馏用的是mmdet-V2.11版本。是他自己复现的YOLOX-m的baseline是45.9%???但是官方YOLOX-m的baseline是46.9%,不清楚作者用的是什么版本,还有如何复现的精度,这baseline精度差距1%去哪了。
我用的mmdet==2.19做的yolox,mmdet的config跑出来就是45.9..........,对于yolox的修改是有yolox分支
多谢解惑,yolox-m不达标可以提bug了。
from fgd.
@yzd-v 我跑了一下代码没问题(精度确实达不到官方)。感谢开源!
from fgd.
您好,您是怎么跑的?我看FGD/configs/distillers/fgd目录下没有相关yolox的配置?方便告知一下吗?多谢
from fgd.
您好,您是怎么跑的?我看FGD/configs/distillers/fgd目录下没有相关yolox的配置?方便告知一下吗?多谢
在yolox分支
from fgd.
您好,我想问一下你跑的那个版本的mmdet啊,mmdet==2.19的话会报sgd错误
ValueEPror: SGD: some parameters appear in more than one parameter group。方便告知一下这个是怎么解决的嘛。多谢
from fgd.
Related Issues (20)
- Why does inference slow down with weights generated by FGD? HOT 1
- How Can I install mmdet 2.11.0 while its latest version is 2.28.1? HOT 1
- no module named 'mmdet.distillation' HOT 1
- Whether additional monitoring training is required for gc-block? HOT 2
- The pre-trained model for Solo_r101_3x.pth HOT 1
- 1 HOT 1
- ConfigDict has no attribute "model" HOT 6
- 请问作者有在yolov5上实现效果吗 HOT 2
- Welcome update to OpenMMLab 2.0
- hello,I want to ask a question about yolox distillation HOT 8
- KeyError: 'neck.fpn_convs.4.conv'
- invalid pre-training HOT 7
- Why the mAP is lower after distillation in VOC? HOT 1
- Can the network be replicated using the 2023 version of mmdetection? HOT 4
- YoloX-l distill Yolox-s with coco pretrained weights HOT 1
- Drop in mAP after Retinanet Distillation HOT 10
- How to output the Params or FLOPs after FGD distillation? HOT 5
- Some questions about gcblock HOT 1
- 你好,我看到您在GitHub更新了yolox的相关蒸馏的数据 HOT 4
- FGD for Ultralytics frameworks
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 fgd.