yangyangzhao

> 我想问一下这样的结果正常么，精度上的差距主要是因为缺失第三步骤么？还有这里的border可以带来多少的精度提升呢 ：） > > https://github.com/chenjun2hao/CenterFace.pytorch/blob/4ea7b08255f282d8ddbbf6c6011114018de07c74/src/lib/datasets/sample/multi_pose_2.py#L48 请问，您是否有修改过作者的代码，主要指的是数据增强那块，为什么我按着作者代码的训练调参方式，训练出来的结果精度只有70呢

> 这是我的调优结果  您好，您用v2的精度最高达到多少呢，还有就是K=600是哪个参数，期待您的回复

> 这是我的调优结果  还有就是是否能够给出您的训练步骤，和关键超参的设置值，比如学习率 batch，是否就是是否按着作者提供的训练步骤进行的训练，谢谢

> i've been using controlNET daily after last Auto1111 pull i'm having this problem i tryed reinstall everything, is still there Did you fix the problem, now my controlnet is...

> same error here i tryed: -reinstall the extention -redownload models from various sources -delete venv and regenerate -commit a older branch > > **none of this solved** me too,now...

> hi we cannot reproduce this problem. help wanted if anyone can provide steps to get the broken environment Thank you for your reply,I just updated the [AUTOMATIC1111/stable diffsuion webui]...

> removing [sd-webui-segment-anything] solved someone also sayd that removing [SD-CN-Animation] may solve too thanks,this works for me