starxhong comments

Results 7 comments of


                                            starxhong

Getting loss_bbox = nan while training on caltech dataset

I have solved this problem! For my case, there are annotation errors of my dataset. I'm training my network with RSOD dataset, which is a aerial image dataset and get...

coco training problem

@sherryxie1 Thank you very much! it perfectly solved my problem. My numpy version is 1.14.1 , higher than 1.10.1，so when i train faster r-cnn i meet "TypeError: 'numpy.float64' object cannot...

coco training problem

@Huangswust182 Yes i do! @sherryxie1's solution works well on my machine~

bbox_loss is always zero when I train R-FCN with OHEM

#85 Here i get the solution!

Very low test accuracy using OHEM.

#85 Maybe here is the answer to your problem

training error

[rbgirshick/py-faster-rcnn#481](https://github.com/rbgirshick/py-faster-rcnn/issues/481) gives two ways to solve the problem perfectly w./wo. reinstall numpy

qwen1.5_7b sft 单机四卡v100，run tokenizer的时候一直卡着，最后报错cuda显存不够，是什么情况？

同样Qwen1.5-7B-Chat和4卡V100，我这边跑tokenizer是OK的，数据是alpaca格式，其他运行参数和环境见https://github.com/hiyouga/LLaMA-Factory/issues/3061。我这的问题主要是OOM，减少cutoff_len可以正常训练，但我不希望截断，考虑开启fp16但最后发现没有生效，有经验的小伙伴也请给点建议