FileNotFoundError: [Errno 2] No such file or directory: '../annotations/person_keypoints_train2017.json'

[vishnukv64]

im getting this error, i followed all the steps properly and shifted everything to 80classes.

This error arises in: raceback (most recent call last): File "train.py", line 89, in dataset = ListDataset(train_path, augment=opt.augment,use_mix=opt.mix_up, multiscale=opt.multiscale_training,normalized_labels=True) File "/home/amvar/vishnu/YOLO_NANO/Yolo_Nano/utils/datasets.py", line 77, in init self.annotations = COCO(json_root) File "/home/amvar/vishnu/YOLO_NANO/venv/lib/python3.6/site-packages/pycocotools/coco.py", line 84, in init dataset = json.load(open(annotation_file, 'r')) FileNotFoundError: [Errno 2] No such file or directory: '../annotations/person_keypoints_train2017.json'

i really dont know where it went wrong, and why is it asking for a personkey train.json and not instance json.

Thanks in advance

[lingtengqiu]

you can modify it in config . it‘s traning set for human coco in 2017， espically for pose . of course pose means have human that’s why we use it to train our model.

------------------ Original ------------------ From: vishnukv64 <notifications@github.com> Date: Wed,Feb 12,2020 10:10 PM To: lingtengqiu/Yolo_Nano <Yolo_Nano@noreply.github.com> Cc: Subscribed <subscribed@noreply.github.com> Subject: Re: [lingtengqiu/Yolo_Nano] FileNotFoundError: [Errno 2] No such file or directory: '../annotations/person_keypoints_train2017.json' (#4)

im getting this error, i followed all the steps properly and shifted everything to 80classes.

This error arises in: raceback (most recent call last): File "train.py", line 89, in dataset = ListDataset(train_path, augment=opt.augment,use_mix=opt.mix_up, multiscale=opt.multiscale_training,normalized_labels=True) File "/home/amvar/vishnu/YOLO_NANO/Yolo_Nano/utils/datasets.py", line 77, in init self.annotations = COCO(json_root) File "/home/amvar/vishnu/YOLO_NANO/venv/lib/python3.6/site-packages/pycocotools/coco.py", line 84, in init dataset = json.load(open(annotation_file, 'r')) FileNotFoundError: [Errno 2] No such file or directory: '../annotations/person_keypoints_train2017.json'

i really dont know where it went wrong, and why is it asking for a personkey train.json and not instance json.

Thanks in advance

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or unsubscribe.

[vishnukv64]

the config youre talking about, is it the yolo-nano.cfg or yolo-nan-person.cfg...

coz i dont want only person, i want to train for the whole 80 classes, with coco 2014 dataset and not 2017 dataset..

btw thanks for your fast reply .

[miaochenGuo]

sorry, i can not find where to modify it...

[Xerus-git]

Hi, you can try to modify "data/default_path.py"->coco_train path. ex: "coco_train": { "img_dir": "data/coco/images/train2014/", "annotations":"data/coco/annotations/instances_train2014.json" }

And the path of the dataset must be created in accordance with the coco-format.

data |_coco 　　|__annotations 　　　　 |__instances_train2014.json 　　　　|instances_val2014.json 　　|__images 　　　　 |train2014 　　　　 |val2014

It works for me.

[ExcellentHunter]

Hi, you can try to modify "data/default_path.py"->coco_train path. ex: "coco_train": { "img_dir": "data/coco/images/train2014/", "annotations":"data/coco/annotations/instances_train2014.json" }

And the path of the dataset must be created in accordance with the coco-format.

data |_coco 　　|__annotations 　　　　 |__instances_train2014.json 　　　　|instances_val2014.json 　　|__images 　　　　 |train2014 　　　　 |val2014

It works for me.

我按照你的思路改了一下，在Evaluating Model时报错FileNotFoundError: [Errno 2] No such file or directory: '../annotations/person_keypoints_val2014.json'，但是我在default_path.py中更改了coco_test路径为"data/coco/annotations/person_keypoints_val2014.json"

[ExcellentHunter]

you can modify it in config . it‘s traning set for human coco in 2017， espically for pose . of course pose means have human that’s why we use it to train our model. … ------------------ Original ------------------ From: vishnukv64 <notifications@github.com> Date: Wed,Feb 12,2020 10:10 PM To: lingtengqiu/Yolo_Nano <Yolo_Nano@noreply.github.com> Cc: Subscribed <subscribed@noreply.github.com> Subject: Re: [lingtengqiu/Yolo_Nano] FileNotFoundError: [Errno 2] No such file or directory: '../annotations/person_keypoints_train2017.json' (#4) im getting this error, i followed all the steps properly and shifted everything to 80classes. This error arises in: raceback (most recent call last): File "train.py", line 89, in dataset = ListDataset(train_path, augment=opt.augment,use_mix=opt.mix_up, multiscale=opt.multiscale_training,normalized_labels=True) File "/home/amvar/vishnu/YOLO_NANO/Yolo_Nano/utils/datasets.py", line 77, in init self.annotations = COCO(json_root) File "/home/amvar/vishnu/YOLO_NANO/venv/lib/python3.6/site-packages/pycocotools/coco.py", line 84, in init dataset = json.load(open(annotation_file, 'r')) FileNotFoundError: [Errno 2] No such file or directory: '../annotations/person_keypoints_train2017.json' i really dont know where it went wrong, and why is it asking for a personkey train.json and not instance json. Thanks in advance — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or unsubscribe.

您好，可以发个邮件或者qq请教您一下嘛，qq：956954782

[kulkarnikeerti]

Hi, you can try to modify "data/default_path.py"->coco_train path. ex: "coco_train": { "img_dir": "data/coco/images/train2014/", "annotations":"data/coco/annotations/instances_train2014.json" }

And the path of the dataset must be created in accordance with the coco-format.

data |_coco 　　|__annotations 　　　　 |__instances_train2014.json 　　　　|instances_val2014.json 　　|__images 　　　　 |train2014 　　　　 |val2014

It works for me.

Hi @Xerus-git I did the same changes to train the model with all the 80 classes. But I end up with this error: RuntimeError: The size of tensor a (80) must match the size of tensor b (27) at non-singleton dimension 1 Do you know what might be causing this error? Thanks in advance.