-
Notifications
You must be signed in to change notification settings - Fork 5.3k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
单尺度0.791! InsightFace赛高!RetinaFace MobileNet0.25同人模型公开 #669
Comments
非常感谢 |
Great! Thanks for sharing! |
@yangfly how can we download mobileNet model ? Dropbox better in Europe ? Can you share the model link? |
@MyraBaba Update Google Drive |
@yangfly thanks a lot is the detector line :
is this correct ? because it gives error:
|
@yangfly Thanks for uploading the models. Could you explain what the difference is between |
mobilenet_0_25-0000 is the pre-trained model from Gluon Model Zoo and mnet.25-0000 is the detector that I trained. |
@yangfly Ah... I get it... thanks... (I was thinking they BOTH are Detector models) |
@yangfly quick question: When I try mobile model mode cpu is using 4 cores and perform very well and fast, but normal model using only one core and pretty slooooow in cpu. any idea why ? how can I push to use all available cores ? Best |
has anyone convert it to caffe or tensorflow pb ,will u share it? |
Hi, @yangfly can you share with us the config.py used for training ? Thank you |
The config.py is added to BaiduYun and GoogleDrive. |
+1 |
1 测试 (WIDER Face Hard 单尺度测试:0.791) 这个时的scale和阈值是怎么设定的。 |
参考test_widerface, 短边1600 @ytt1790579195 |
我没有修改 test_widerface.py 和 retinaface.py 中的任何参数 |
@nttstar Do you know where to get the pretrained model named "mobilenet025fd0"? |
|
mnet0.25的anchor设置貌似没有scale step,默认使用了RAC_SSH的配置,anchor的大小与论文中Table2不一致,对吗(默默过来确认一下 |
配置参数,在 train.log 中都打印出来了。 |
嗯~好滴谢谢啦 |
@yangfly |
这个MobileNet0.25是v1还是v2? |
你好,请问为什么我将你训练好的mnet.25在cpu上做推断时得到的结果与gpu不一致,且差别很大呢? |
|
I think we can run this model on TensorRT. Mobilenet Support is available. @yangfly https://mxnet.incubator.apache.org/versions/master/tutorials/tensorrt/inference_with_trt.html |
hello @yangfly 我用的gpu是titan xp。 |
|
command 2>&1 | tee train.log |
@yangfly Good job!!When I test this model,I found that it consumes so much memory about 3G,could you tell me why? |
请问下mnet.25模型在手机上测试过吗? |
@yangfly Did you only use wider face dataset when training?What is the batch size when you training?Thanks a lot! |
Did you use different learning rate & learning rate schedule while training on mobilenet (compared to resnet)? I don't see these values in the config files... |
@yangfly 你好,参考了你提供的训练日志,我发现FG acc 开始会往下降到0,然后在回升。这个现象是正常的么,我遇到的问题是FG acc下降到0以后就不会再回升,你有遇到过这问题么 |
抱歉,没有遇到这种现象。------------------ 原始邮件 ------------------
发件人: "gripdestiny"<notifications@github.com>
发送时间: 2019年12月5日(星期四) 下午5:30
收件人: "deepinsight/insightface"<insightface@noreply.github.com>;
抄送: "杨帆"<yf.dl@qq.com>;"Mention"<mention@noreply.github.com>;
主题: Re: [deepinsight/insightface] 单尺度0.791! InsightFace赛高!RetinaFace MobileNet0.25同人模型公开 (#669)
@yangfly 你好,参考了你提供的训练日志,我发现FG acc 开始会往下降到0,然后在回升。这个现象是正常的么,我遇到的问题是FG acc下降到0以后就不会再回升,你有遇到过这问题么
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub, or unsubscribe.
|
https://github.com/lbin/Retinaface_Mobilenet_Pytorch 单尺度相同设置hard达到85.55% |
你好,我想问一下你解决了这个问题嘛 |
老哥,我用你的mobilenet_0_25-symbol.json和最新的retinaface的代码训练,得到的mnet.25-symbol.json和你自己训练得到的网络结构有一点局部区别,请问是怎么回事啊? |
你好,请问你在gpu和cpu下用R50测试图片单张耗时多久, |
+1 Waiting for fine tuned optimized version with tensorRT or TVM. Can be quantize it ? fp16 or int8 with minimal loss in accuracy ? |
@yangfly 您好,有个问题,在使用test_widerface.py时,怎么分开测试easy,medium,hard三种测试集呢?或者怎么分开测试验证集呢?图片中的recall all是不是就是最终参考的指标? 谢谢您的时间 |
@yangfly What is the output shape of this mobilenet model ? |
请问,有测试过mnet0.25使用cpu在VGA分辨率下的时间吗?我在ncnn框架下测试时间为50ms(单线程),和文章相差较大 @leo2105 @yangfly @ashuezy @chenzhengnan |
@yangfly, I executed retinaface based on mobilenet0.25 (you trained) from onnx file.
|
@yangfly What is the license on the models your trained? Is it MIT? |
RetinaFace 牛逼!:two_hearts: InsightFace 赛高!:two_hearts:
开放 mobilenet0.25 版 RetinaFace 同人版模型
更完整的模型、日志和WiderFace测试截图 百度云 提取码:nzof GoogleDrive
模型说明,使用的是Gluon Model Zoo 的标准版 MobileNet0.25 预训练模型。(没有 fast downsampling,模型大小 1.68Mb)
Batch-size 32x2(两块1080Ti),其他参数都是 RetinaFace 默认的。
WIDER Face Hard 单尺度测试:0.791
![val-hard-quick](https://user-images.githubusercontent.com/15797180/57523356-55c78980-7357-11e9-9bbd-2dbbdf073ea6.jpg)
WIDER Face Hard 多尺度测试:0.825
![val-hard-slow](https://user-images.githubusercontent.com/15797180/57523720-4d238300-7358-11e9-9741-0ffca752e7cf.jpg)
欢迎调参大佬,公开更给力的结果。:beer:
The text was updated successfully, but these errors were encountered: