Inomjon

Same error? did you find the solution?

@Xccelrator

> 想请教一下，我运行create_traindata生成了5000张图片，batch_size是256训练了20轮，但是在运行test_nn.py文件输出的结果却是完全错误的，想请教一下大家有遇到这种情况吗 Hi brother, did you solve the problem? I encountered same issue

> Yeah,because the time you run is too short. I run the code for a whole night,the recognized result was fine. | | buptwanghao | | 邮箱：[email protected] | Signature is...

> Yeah,train the model as long as possible.You can try to run the code before you sleep,and check the performance when you get up. | | buptwanghao | | 邮箱：[email protected]...

Hi @YonghaoHe ! I'm trying to run your head detection model on Jetson nano, it could run only 480p and scale=0.25 , I couldn't understand reason , please guide

You can try 'git clone' the project.

Hi, Please use SDK version of firmware to run the demo, https://github.com/simshineaicamera/SIMCAM_SDK/blob/master/src/rootfs Our commercial version of firmware does not support some dependencies, it is slightly different from public SDK version...

is it same error? Please make sure that SDK version of firmware updated correctly, Note* if the firmware updated to SDK version you cannot connect the camera to your Simcam...

> I think the batch size = 1 can't not be split to multi-gpu training in test phase , so you can close the test phase and start training ....