ByteTrack
ByteTrack copied to clipboard
ifzhang
AssertionError
Traceback (most recent call last):
File "tools/train.py", line 112, in
"yolox/data/datasets/mot.py", line 102, in pull_item assert img is not None AssertionError
Traceback (most recent call last): File "tools/train.py", line 112, in assert num_gpu <= torch.cuda.device_count() AssertionError
I've got same error when train the model.
- I've checked the 'torch.cuda.**' and there's nothing wrong happend. ONLY at this happend. 2.I also have right cuda version
Any helps?
USED: cd <ByteTrack_HOME> python3 tools/train.py -f exps/example/mot/yolox_x_ablation.py -d 8 -b 48 --fp16 -o -c pretrained/yolox_x.pth
Check your machine has GPU OR make sure your pytorch has cuda version
I've had got this problem too, and I had solved by downloading pytorch with cuda version and aslo try the with gpu device 0. instead of -d use -d 0
I've had got this problem too, and I had solved by downloading pytorch with cuda version and aslo try the with gpu device 0. instead of -d use -d 0
Thx a lot.
I've tried following you. and I could see the "exp value" table.
but still get an error as bellow.
did this happend to you ?
Sorry, I didnt got this
https://github.com/ifzhang/ByteTrack/issues/57
Guess this was the problem. Thx. good day :)
Sorry, I didnt got this
#57
Guess this was the problem. Thx. good day :)
Good
But I've got this problem
"yolox/data/datasets/mot.py", line 102, in pull_item assert img is not None AssertionError
As for my experience, those 'AsserstionError's occured by 'yolox' version or packages. So, I'd solved by deleting yolox and packages. or simply just delelting the whole 'Bytetrack' folder and re-download and re install by the following steps noted at the 'readme.md'.
I guess there could be some collision between install steps.
and beaware that this 'Bytetrack' uses the old version of yolox
https://github.com/Megvii-BaseDetection/YOLOX/tree/0.1.0
"yolox/data/datasets/mot.py", line 102, in pull_item assert img is not None AssertionError
I also ran into this problem, were you able to solve it?
"yolox/data/datasets/mot.py", line 102, in pull_item assert img is not None AssertionError
I also ran into this problem, were you able to solve it?
If you get warnnings, try to solve them
it is generally the issue of wrong image directoies, you can print out the dir for debugging.
Any other update on Solving this problem? I have tried pytorch with cuda version and it still doesn't work. Any other way round it or should I simply reinstall ByteTrack.
@parthmalpathak In the dataset directory(../Bytetrack/datasets) , name your folder for your dataset “mot”
@TheMadScientiist just to confirm, the datasets folder already has 'mot' which contains MOT17- train, test and annotations. Apart from that can you tell me where to store mix_det, mix_mot_20_ch and mix_mot_ch? Because I have just kept them as directories along with ETHZ, cityscapes and other datasets
@parthmalpathak In the ETHZ and cityscapes directories are there train, test, and annotations like the “mot” folder?
datasets |——————mot | └——————train | └——————test └——————crowdhuman | └——————Crowdhuman_train | └——————Crowdhuman_val | └——————annotation_train.odgt | └——————annotation_val.odgt └——————MOT20 | └——————train | └——————test └——————Cityscapes | └——————images | └——————labels_with_ids └——————ETHZ └——————eth01 └——————... └——————eth07
The folders are made the same way as suggested in the README. They have the same content. Each folder has annotations extra when you convert to COCO by running the .py files.
“yolox/data/datasets/mot.py”，第 102 行，在 pull_item 中 断言 img 不是 None AssertionError
我也遇到了这个问题，请问你解决了吗？
“yolox/data/datasets/mot.py”，第 102 行，在 pull_item 中 断言 img 不是 None AssertionError
请问你解决了吗
I've had got this problem too, and I had solved by downloading pytorch with cuda version and aslo try the with gpu device 0. instead of -d use -d 0
Thx a lot. I've tried following you. and I could see the "exp value" table.
l change the d to 0, but l get this problem No such file or directory: 'pretrained/yolox_x.pth' , how to fix it
You can create a new folder in your root directory. Name it pretrained, and put yolox_x.pth in it
------------------ 原始邮件 ------------------ 发件人: "ifzhang/ByteTrack" @.>; 发送时间: 2022年11月29日(星期二) 晚上10:26 @.>; 抄送: "指尖@@.@.>; 主题: Re: [ifzhang/ByteTrack] AssertionError (Issue #174)
I've had got this problem too, and I had solved by downloading pytorch with cuda version and aslo try the with gpu device 0. instead of -d use -d 0
Thx a lot. I've tried following you. and I could see the "exp value" table. but still get an error as bellow. did this happend to you ?
l change the d to 0, but l get this problem No such file or directory: 'pretrained/yolox_x.pth' , how to fix it
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.***>
“yolox/data/datasets/mot.py”，第 102 行，在 pull_item 中 断言 img 不是 None AssertionError
请问你解决了吗
hi ,when i run python3 tools/convert_crowdhuman_to_coco.py, l have this question, [Errno 2] No such file or directory: 'datasets/crowdhuman/CrowdHuman_val/273271,c9db000d5146c15.jpg', but l have cheek it, there have the 273271,c9db000d5146c15.jpg, how to fix it?
这个没有，我也是这个错误，得问作者
------------------ 原始邮件 ------------------ 发件人: @.>; 发送时间: 2022年12月4日(星期天) 上午9:01 收件人: @.>; 抄送: "指尖@@.>; @.>; 主题: Re: [ifzhang/ByteTrack] AssertionError (Issue #174)
“yolox/data/datasets/mot.py”，第 102 行，在 pull_item 中 断言 img 不是 None AssertionError
请问你解决了吗
hi ,when i run python3 tools/convert_crowdhuman_to_coco.py, l have this question, [Errno 2] No such file or directory: 'datasets/crowdhuman/CrowdHuman_val/273271,c9db000d5146c15.jpg', but l have cheek it, there have the 273271,c9db000d5146c15.jpg, how to fix it?
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.***>
这个没有，我也是这个错误，得问作者 … ------------------ 原始邮件 ------------------ 发件人: @.>; 发送时间: 2022年12月4日(星期天) 上午9:01 收件人: @.>; 抄送: "指尖@@.>; @.>; 主题: Re: [ifzhang/ByteTrack] AssertionError (Issue #174) “yolox/data/datasets/mot.py”，第 102 行，在 pull_item 中 断言 img 不是 None AssertionError 请问你解决了吗 hi ,when i run python3 tools/convert_crowdhuman_to_coco.py, l have this question, [Errno 2] No such file or directory: 'datasets/crowdhuman/CrowdHuman_val/273271,c9db000d5146c15.jpg', but l have cheek it, there have the 273271,c9db000d5146c15.jpg, how to fix it? — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.***>
when i run python tools/train.py -f exps/example/mot/yolox_x_ablation.py -d 8 -b 48 --fp16 -o -c pretrained/yolox_x.pth, it assert num_gpu <= torch.cuda.device_count(), any suggestion?
“yolox/data/datasets/mot.py”，第 102 行，在 pull_item 中 断言 img 不是 None AssertionError
请问你解决了吗
now, l get the same question,how you fix it ?
Sorry, I didnt got this
#57 Guess this was the problem. Thx. good day :)
Good
But I've got this problem
"yolox/data/datasets/mot.py", line 102, in pull_item assert img is not None AssertionError
l have the same question, how you fix it ?
I also have the same problem, which has not been solved. Is there a bug in the original code?
------------------ 原始邮件 ------------------ 发件人: "ifzhang/ByteTrack" @.>; 发送时间: 2022年12月4日(星期天) 下午3:36 @.>; 抄送: "指尖@@.@.>; 主题: Re: [ifzhang/ByteTrack] AssertionError (Issue #174)
Sorry, I didnt got this
#57 Guess this was the problem. Thx. good day :)
Good
But I've got this problem
"yolox/data/datasets/mot.py", line 102, in pull_item assert img is not None AssertionError
l have the same question, how you fix it ?
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.***>
“yolox/data/datasets/mot.py”，第 102 行，在 pull_item 中 断言 img 不是 None AssertionError
请问你解决了吗
any result? l have assert img is not None too
maybe you can change the command line (e.g. from "python train_net.py --num-gpus 8(this num is too big) " to "python train_net.py --num-gpus 2(the number is up to your num of gpus, I recommand you to rise it from 1)")
