AssertionError: ../coco/images/train2017 is an invalid directory path!

[LiJiahang617]

i met an issue when i trained,this issue was like below: why it is an invalid directory path? i just followed the README file,and i don't know where was i wrong

[LiJiahang617]

i don't think i make a mistake in organizing the datasets.

[Chilicyy]

Maybe you can try changing the relative path to absolute path in data/coco.yaml.

[LiJiahang617]

Sorry to bother you again，i tried to use the absolute path in coco.yaml, but this issue still exists,like this: this is my coco.yaml this is the issue : how to fix this issue.i'm very confused now .

[meituan-gengyifei]

Maybe you can try ‘ls + data-path' to verify the absolute path that is correct？Judging from the current problem, it should still be the problem that the path is not written correctly, thanks

[LiJiahang617]

ok, i have solved it, thanks a lot

[ruoyuchen0903]

ok, i have solved it, thanks a lot

想问问怎么解决的可以嘛？可以用voc数据集的格式嘛？

[LiJiahang617]

不可以，要用coco的。最后还是发现是路径有问题，按照terminal里面的提示把一些参数文件放到指定的位置就可以了，不过这个项目目前有bug，我训练的map还有loss那些全是0，解决不了

---Original--- From: @.> Date: Sun, Jun 26, 2022 23:15 PM To: @.>; Cc: @.**@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is an invalid directory path! (Issue #48)

ok, i have solved it, thanks a lot

想问问怎么解决的可以嘛？可以用voc数据集的格式嘛？

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>

[ruoyuchen0903]

噢噢好的，能具体说说数据集的格式是哪出问题了嘛？要修改哪里？其他yolo训练用这个完全没问题呀。

2022-06-27 00:46:04LiJiahang617 @.***>写道：

不可以，要用coco的。最后还是发现是路径有问题，按照terminal里面的提示把一些参数文件放到指定的位置就可以了，不过这个项目目前有bug，我训练的map还有loss那些全是0，解决不了

---Original--- From: @.> Date: Sun, Jun 26, 2022 23:15 PM To: @.>; Cc: @.**@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is an invalid directory path! (Issue #48)

ok, i have solved it, thanks a lot

想问问怎么解决的可以嘛？可以用voc数据集的格式嘛？

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.***>

[LiJiahang617]

voc的话好像得改很多东西，这个用的应该就是Coco格式的数据集，应该是数据集组织不对，我把数据集文件的组织树关系按照终端上的提示改了一下，就好了，但是跑起来loss就一个有数字，其他都为0，模型跑不出来，也不知道为啥，

---Original--- From: @.> Date: Mon, Jun 27, 2022 11:02 AM To: @.>; Cc: @.**@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is aninvalid directory path! (Issue #48)

噢噢好的，能具体说说数据集的格式是哪出问题了嘛？要修改哪里？其他yolo训练用这个完全没问题呀。

2022-06-27 00:46:04LiJiahang617 @.***>写道：

不可以，要用coco的。最后还是发现是路径有问题，按照terminal里面的提示把一些参数文件放到指定的位置就可以了，不过这个项目目前有bug，我训练的map还有loss那些全是0，解决不了

---Original--- From: @.> Date: Sun, Jun 26, 2022 23:15 PM To: @.>; Cc: @.**@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is an invalid directory path! (Issue #48)

ok, i have solved it, thanks a lot

想问问怎么解决的可以嘛？可以用voc数据集的格式嘛？

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.> — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.>

[shenshi19]

voc的话好像得改很多东西，这个用的应该就是Coco格式的数据集，应该是数据集组织不对，我把数据集文件的组织树关系按照终端上的提示改了一下，就好了，但是跑起来loss就一个有数字，其他都为0，模型跑不出来，也不知道为啥， … ---Original--- From: @.> Date: Mon, Jun 27, 2022 11:02 AM To: @.>; Cc: @.**@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is aninvalid directory path! (Issue #48) 噢噢好的，能具体说说数据集的格式是哪出问题了嘛？要修改哪里？其他yolo训练用这个完全没问题呀。 2022-06-27 00:46:04LiJiahang617 @.>写道： 不可以，要用coco的。最后还是发现是路径有问题，按照terminal里面的提示把一些参数文件放到指定的位置就可以了，不过这个项目目前有bug，我训练的map还有loss那些全是0，解决不了 ---Original--- From: @.> Date: Sun, Jun 26, 2022 23:15 PM To: @.>; Cc: @*.**@*.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is an invalid directory path! (Issue #48) ok, i have solved it, thanks a lot 想问问怎么解决的可以嘛？可以用voc数据集的格式嘛？ — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.> — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.> — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.>

Traceback (most recent call last): File "tools/train.py", line 87, in main(args) File "tools/train.py", line 76, in main trainer = Trainer(args, cfg, device) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/core/engine.py", line 40, in init self.train_loader, self.val_loader = self.get_data_loader(args, cfg, self.data_dict) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/core/engine.py", line 206, in get_data_loader train_loader = create_dataloader(train_path, args.img_size, args.batch_size // args.world_size, grid_size, File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/data_load.py", line 23, in create_dataloader dataset = TrainValDataset(path, img_size, batch_size, File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/datasets.py", line 39, in init self.img_paths, self.labels = self.get_imgs_labels(self.img_dir) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/datasets.py", line 184, in get_imgs_labels assert osp.exists(label_dir), f'{label_dir} is an invalid directory path!' AssertionError: /home/lrq/yolotest/YOLOv6-main/YOLOv6-main/data/coco/labels/train2017 is an invalid directory path! 大神，您可以帮我看一下我需要改哪里吗？

[LiJiahang617]

你看看最后两行，你对应的那个路径里面有相应的文件嘛，那个路径你应该是没放东西，往那个路径里面放对应的东西就可以了

---Original--- From: @.> Date: Mon, Jun 27, 2022 15:20 PM To: @.>; Cc: @.**@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is an invalid directory path! (Issue #48)

voc的话好像得改很多东西，这个用的应该就是Coco格式的数据集，应该是数据集组织不对，我把数据集文件的组织树关系按照终端上的提示改了一下，就好了，但是跑起来loss就一个有数字，其他都为0，模型跑不出来，也不知道为啥， … ---Original--- From: @.> Date: Mon, Jun 27, 2022 11:02 AM To: @.>; Cc: @.@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is aninvalid directory path! (Issue #48) 噢噢好的，能具体说说数据集的格式是哪出问题了嘛？要修改哪里？其他yolo训练用这个完全没问题呀。 2022-06-27 00:46:04LiJiahang617 @.>写道： 不可以，要用coco的。最后还是发现是路径有问题，按照terminal里面的提示把一些参数文件放到指定的位置就可以了，不过这个项目目前有bug，我训练的map还有loss那些全是0，解决不了 ---Original--- From: @.> Date: Sun, Jun 26, 2022 23:15 PM To: @.>; Cc: @.@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is an invalid directory path! (Issue #48) ok, i have solved it, thanks a lot 想问问怎么解决的可以嘛？可以用voc数据集的格式嘛？ — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.> — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.> — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.>

Traceback (most recent call last): File "tools/train.py", line 87, in main(args) File "tools/train.py", line 76, in main trainer = Trainer(args, cfg, device) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/core/engine.py", line 40, in init self.train_loader, self.val_loader = self.get_data_loader(args, cfg, self.data_dict) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/core/engine.py", line 206, in get_data_loader train_loader = create_dataloader(train_path, args.img_size, args.batch_size // args.world_size, grid_size, File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/data_load.py", line 23, in create_dataloader dataset = TrainValDataset(path, img_size, batch_size, File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/datasets.py", line 39, in init self.img_paths, self.labels = self.get_imgs_labels(self.img_dir) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/datasets.py", line 184, in get_imgs_labels assert osp.exists(label_dir), f'{label_dir} is an invalid directory path!' AssertionError: /home/lrq/yolotest/YOLOv6-main/YOLOv6-main/data/coco/labels/train2017 is an invalid directory path! 大神，您可以帮我看一下我需要改哪里吗？

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>

[shenshi19]

你看看最后两行，你对应的那个路径里面有相应的文件嘛，那个路径你应该是没放东西，往那个路径里面放对应的东西就可以了 … ---Original--- From: @.> Date: Mon, Jun 27, 2022 15:20 PM To: @.>; Cc: @.**@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is an invalid directory path! (Issue #48) voc的话好像得改很多东西，这个用的应该就是Coco格式的数据集，应该是数据集组织不对，我把数据集文件的组织树关系按照终端上的提示改了一下，就好了，但是跑起来loss就一个有数字，其他都为0，模型跑不出来，也不知道为啥， … ---Original--- From: @.> Date: Mon, Jun 27, 2022 11:02 AM To: @.>; Cc: @.@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is aninvalid directory path! (Issue #48) 噢噢好的，能具体说说数据集的格式是哪出问题了嘛？要修改哪里？其他yolo训练用这个完全没问题呀。 2022-06-27 00:46:04LiJiahang617 @.>写道： 不可以，要用coco的。最后还是发现是路径有问题，按照terminal里面的提示把一些参数文件放到指定的位置就可以了，不过这个项目目前有bug，我训练的map还有loss那些全是0，解决不了 ---Original--- From: @.> Date: Sun, Jun 26, 2022 23:15 PM To: @.>; Cc: @.@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is an invalid directory path! (Issue #48) ok, i have solved it, thanks a lot 想问问怎么解决的可以嘛？可以用voc数据集的格式嘛？ — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.> — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.> — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.> Traceback (most recent call last): File "tools/train.py", line 87, in main(args) File "tools/train.py", line 76, in main trainer = Trainer(args, cfg, device) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/core/engine.py", line 40, in init self.train_loader, self.val_loader = self.get_data_loader(args, cfg, self.data_dict) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/core/engine.py", line 206, in get_data_loader train_loader = create_dataloader(train_path, args.img_size, args.batch_size // args.world_size, grid_size, File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/data_load.py", line 23, in create_dataloader dataset = TrainValDataset(path, img_size, batch_size, File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/datasets.py", line 39, in init self.img_paths, self.labels = self.get_imgs_labels(self.img_dir) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/datasets.py", line 184, in get_imgs_labels assert osp.exists(label_dir), f'{label_dir} is an invalid directory path!' AssertionError: /home/lrq/yolotest/YOLOv6-main/YOLOv6-main/data/coco/labels/train2017 is an invalid directory path! 大神，您可以帮我看一下我需要改哪里吗？ — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***> 已经解决，感谢感谢

[LiJiahang617]

好的，有时候确实项目很奇怪，不一定是自己的问题，是这个项目估计编的不太好，他把一些路径代码给弄的不行，自己按照他那个改一下路径就行了

---Original--- From: @.> Date: Mon, Jun 27, 2022 16:09 PM To: @.>; Cc: @.**@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is an invalid directory path! (Issue #48)

你看看最后两行，你对应的那个路径里面有相应的文件嘛，那个路径你应该是没放东西，往那个路径里面放对应的东西就可以了 … ---Original--- From: @.> Date: Mon, Jun 27, 2022 15:20 PM To: @.>; Cc: @.@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is an invalid directory path! (Issue #48) voc的话好像得改很多东西，这个用的应该就是Coco格式的数据集，应该是数据集组织不对，我把数据集文件的组织树关系按照终端上的提示改了一下，就好了，但是跑起来loss就一个有数字，其他都为0，模型跑不出来，也不知道为啥， … ---Original--- From: @.> Date: Mon, Jun 27, 2022 11:02 AM To: @.>; Cc: @.@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is aninvalid directory path! (Issue #48) 噢噢好的，能具体说说数据集的格式是哪出问题了嘛？要修改哪里？其他yolo训练用这个完全没问题呀。 2022-06-27 00:46:04LiJiahang617 @.>写道： 不可以，要用coco的。最后还是发现是路径有问题，按照terminal里面的提示把一些参数文件放到指定的位置就可以了，不过这个项目目前有bug，我训练的map还有loss那些全是0，解决不了 ---Original--- From: @.> Date: Sun, Jun 26, 2022 23:15 PM To: @.>; Cc: @.@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is an invalid directory path! (Issue #48) ok, i have solved it, thanks a lot 想问问怎么解决的可以嘛？可以用voc数据集的格式嘛？ — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.> — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.> — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.> Traceback (most recent call last): File "tools/train.py", line 87, in main(args) File "tools/train.py", line 76, in main trainer = Trainer(args, cfg, device) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/core/engine.py", line 40, in init self.train_loader, self.val_loader = self.get_data_loader(args, cfg, self.data_dict) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/core/engine.py", line 206, in get_data_loader train_loader = create_dataloader(train_path, args.img_size, args.batch_size // args.world_size, grid_size, File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/data_load.py", line 23, in create_dataloader dataset = TrainValDataset(path, img_size, batch_size, File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/datasets.py", line 39, in init self.img_paths, self.labels = self.get_imgs_labels(self.img_dir) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/datasets.py", line 184, in get_imgs_labels assert osp.exists(label_dir), f'{label_dir} is an invalid directory path!' AssertionError: /home/lrq/yolotest/YOLOv6-main/YOLOv6-main/data/coco/labels/train2017 is an invalid directory path! 大神，您可以帮我看一下我需要改哪里吗？ — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***> 已经解决，感谢感谢

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>

[ruoyuchen0903]

您好，我的数据集是这么排列的，怎样改呢？ --data --annotations --instance train --instance val --images --train --val --test --labels --train --val --test

2022-06-27 15:20:36shenshi19 @.***>写道：

voc的话好像得改很多东西，这个用的应该就是Coco格式的数据集，应该是数据集组织不对，我把数据集文件的组织树关系按照终端上的提示改了一下，就好了，但是跑起来loss就一个有数字，其他都为0，模型跑不出来，也不知道为啥， … ---Original--- From: @.> Date: Mon, Jun 27, 2022 11:02 AM To: @.>; Cc: @.@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is aninvalid directory path! (Issue #48) 噢噢好的，能具体说说数据集的格式是哪出问题了嘛？要修改哪里？其他yolo训练用这个完全没问题呀。 2022-06-27 00:46:04LiJiahang617 @.>写道： 不可以，要用coco的。最后还是发现是路径有问题，按照terminal里面的提示把一些参数文件放到指定的位置就可以了，不过这个项目目前有bug，我训练的map还有loss那些全是0，解决不了 ---Original--- From: @.> Date: Sun, Jun 26, 2022 23:15 PM To: @.>; Cc: @.@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is an invalid directory path! (Issue #48) ok, i have solved it, thanks a lot 想问问怎么解决的可以嘛？可以用voc数据集的格式嘛？ — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.> — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.> — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.>

Traceback (most recent call last): File "tools/train.py", line 87, in main(args) File "tools/train.py", line 76, in main trainer = Trainer(args, cfg, device) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/core/engine.py", line 40, in init self.train_loader, self.val_loader = self.get_data_loader(args, cfg, self.data_dict) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/core/engine.py", line 206, in get_data_loader train_loader = create_dataloader(train_path, args.img_size, args.batch_size // args.world_size, grid_size, File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/data_load.py", line 23, in create_dataloader dataset = TrainValDataset(path, img_size, batch_size, File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/datasets.py", line 39, in init self.img_paths, self.labels = self.get_imgs_labels(self.img_dir) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/datasets.py", line 184, in get_imgs_labels assert osp.exists(label_dir), f'{label_dir} is an invalid directory path!' AssertionError: /home/lrq/yolotest/YOLOv6-main/YOLOv6-main/data/coco/labels/train2017 is an invalid directory path! 大神，您可以帮我看一下我需要改哪里吗？

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.***>

[LiJiahang617]

你去网上查一下，你这个格式不对，我也不好说，你看看人家的要求是咋样的

---Original--- From: @.> Date: Mon, Jun 27, 2022 16:22 PM To: @.>; Cc: @.**@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is aninvalid directory path! (Issue #48)

您好，我的数据集是这么排列的，怎样改呢？ --data --annotations --instance train --instance val --images --train --val --test --labels --train --val --test

2022-06-27 15:20:36shenshi19 @.***>写道：

voc的话好像得改很多东西，这个用的应该就是Coco格式的数据集，应该是数据集组织不对，我把数据集文件的组织树关系按照终端上的提示改了一下，就好了，但是跑起来loss就一个有数字，其他都为0，模型跑不出来，也不知道为啥， … ---Original--- From: @.> Date: Mon, Jun 27, 2022 11:02 AM To: @.>; Cc: @.@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is aninvalid directory path! (Issue #48) 噢噢好的，能具体说说数据集的格式是哪出问题了嘛？要修改哪里？其他yolo训练用这个完全没问题呀。 2022-06-27 00:46:04LiJiahang617 @.>写道： 不可以，要用coco的。最后还是发现是路径有问题，按照terminal里面的提示把一些参数文件放到指定的位置就可以了，不过这个项目目前有bug，我训练的map还有loss那些全是0，解决不了 ---Original--- From: @.> Date: Sun, Jun 26, 2022 23:15 PM To: @.>; Cc: @.@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is an invalid directory path! (Issue #48) ok, i have solved it, thanks a lot 想问问怎么解决的可以嘛？可以用voc数据集的格式嘛？ — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.> — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.> — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.>

Traceback (most recent call last): File "tools/train.py", line 87, in main(args) File "tools/train.py", line 76, in main trainer = Trainer(args, cfg, device) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/core/engine.py", line 40, in init self.train_loader, self.val_loader = self.get_data_loader(args, cfg, self.data_dict) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/core/engine.py", line 206, in get_data_loader train_loader = create_dataloader(train_path, args.img_size, args.batch_size // args.world_size, grid_size, File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/data_load.py", line 23, in create_dataloader dataset = TrainValDataset(path, img_size, batch_size, File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/datasets.py", line 39, in init self.img_paths, self.labels = self.get_imgs_labels(self.img_dir) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/datasets.py", line 184, in get_imgs_labels assert osp.exists(label_dir), f'{label_dir} is an invalid directory path!' AssertionError: /home/lrq/yolotest/YOLOv6-main/YOLOv6-main/data/coco/labels/train2017 is an invalid directory path! 大神，您可以帮我看一下我需要改哪里吗？

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.> — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.>

[shenshi19]

您好，我的数据集是这么排列的，怎样改呢？ --data --annotations --instance train --instance val --images --train --val --test --labels --train --val --test 2022-06-27 15:20:36shenshi19 @.>写道： voc的话好像得改很多东西，这个用的应该就是Coco格式的数据集，应该是数据集组织不对，我把数据集文件的组织树关系按照终端上的提示改了一下，就好了，但是跑起来loss就一个有数字，其他都为0，模型跑不出来，也不知道为啥， … ---Original--- From: @.> Date: Mon, Jun 27, 2022 11:02 AM To: @.>; Cc: @.@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is aninvalid directory path! (Issue #48) 噢噢好的，能具体说说数据集的格式是哪出问题了嘛？要修改哪里？其他yolo训练用这个完全没问题呀。 2022-06-27 00:46:04LiJiahang617 @.>写道： 不可以，要用coco的。最后还是发现是路径有问题，按照terminal里面的提示把一些参数文件放到指定的位置就可以了，不过这个项目目前有bug，我训练的map还有loss那些全是0，解决不了 ---Original--- From: @.> Date: Sun, Jun 26, 2022 23:15 PM To: @.>; Cc: @.@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is an invalid directory path! (Issue #48) ok, i have solved it, thanks a lot 想问问怎么解决的可以嘛？可以用voc数据集的格式嘛？ — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.> — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.> — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.> Traceback (most recent call last): File "tools/train.py", line 87, in main(args) File "tools/train.py", line 76, in main trainer = Trainer(args, cfg, device) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/core/engine.py", line 40, in init self.train_loader, self.val_loader = self.get_data_loader(args, cfg, self.data_dict) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/core/engine.py", line 206, in get_data_loader train_loader = create_dataloader(train_path, args.img_size, args.batch_size // args.world_size, grid_size, File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/data_load.py", line 23, in create_dataloader dataset = TrainValDataset(path, img_size, batch_size, File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/datasets.py", line 39, in init self.img_paths, self.labels = self.get_imgs_labels(self.img_dir) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/datasets.py", line 184, in get_imgs_labels assert osp.exists(label_dir), f'{label_dir} is an invalid directory path!' AssertionError: /home/lrq/yolotest/YOLOv6-main/YOLOv6-main/data/coco/labels/train2017 is an invalid directory path! 大神，您可以帮我看一下我需要改哪里吗？ — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.>

删掉images试试

[ruoyuchen0903]

好的好的 谢谢大佬

2022-06-27 16:25:02LiJiahang617 @.***>写道：

你去网上查一下，你这个格式不对，我也不好说，你看看人家的要求是咋样的

---Original--- From: @.> Date: Mon, Jun 27, 2022 16:22 PM To: @.>; Cc: @.**@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is aninvalid directory path! (Issue #48)

您好，我的数据集是这么排列的，怎样改呢？ --data --annotations --instance train --instance val --images --train --val --test --labels --train --val --test

2022-06-27 15:20:36shenshi19 @.***>写道：

voc的话好像得改很多东西，这个用的应该就是Coco格式的数据集，应该是数据集组织不对，我把数据集文件的组织树关系按照终端上的提示改了一下，就好了，但是跑起来loss就一个有数字，其他都为0，模型跑不出来，也不知道为啥， … ---Original--- From: @.> Date: Mon, Jun 27, 2022 11:02 AM To: @.>; Cc: @.@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is aninvalid directory path! (Issue #48) 噢噢好的，能具体说说数据集的格式是哪出问题了嘛？要修改哪里？其他yolo训练用这个完全没问题呀。 2022-06-27 00:46:04LiJiahang617 @.>写道： 不可以，要用coco的。最后还是发现是路径有问题，按照terminal里面的提示把一些参数文件放到指定的位置就可以了，不过这个项目目前有bug，我训练的map还有loss那些全是0，解决不了 ---Original--- From: @.> Date: Sun, Jun 26, 2022 23:15 PM To: @.>; Cc: @.@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is an invalid directory path! (Issue #48) ok, i have solved it, thanks a lot 想问问怎么解决的可以嘛？可以用voc数据集的格式嘛？ — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.> — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.> — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.>

Traceback (most recent call last): File "tools/train.py", line 87, in main(args) File "tools/train.py", line 76, in main trainer = Trainer(args, cfg, device) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/core/engine.py", line 40, in init self.train_loader, self.val_loader = self.get_data_loader(args, cfg, self.data_dict) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/core/engine.py", line 206, in get_data_loader train_loader = create_dataloader(train_path, args.img_size, args.batch_size // args.world_size, grid_size, File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/data_load.py", line 23, in create_dataloader dataset = TrainValDataset(path, img_size, batch_size, File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/datasets.py", line 39, in init self.img_paths, self.labels = self.get_imgs_labels(self.img_dir) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/datasets.py", line 184, in get_imgs_labels assert osp.exists(label_dir), f'{label_dir} is an invalid directory path!' AssertionError: /home/lrq/yolotest/YOLOv6-main/YOLOv6-main/data/coco/labels/train2017 is an invalid directory path! 大神，您可以帮我看一下我需要改哪里吗？

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.> — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.>

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.***>

[ruoyuchen0903]

删掉了还是显示AssertionError: Results do not correspond to current coco set。头疼

2022-06-27 16:29:59shenshi19 @.***>写道：

您好，我的数据集是这么排列的，怎样改呢？ --data --annotations --instance train --instance val --images --train --val --test --labels --train --val --test 2022-06-27 15:20:36shenshi19 @.>写道： voc的话好像得改很多东西，这个用的应该就是Coco格式的数据集，应该是数据集组织不对，我把数据集文件的组织树关系按照终端上的提示改了一下，就好了，但是跑起来loss就一个有数字，其他都为0，模型跑不出来，也不知道为啥， … ---Original--- From: @.> Date: Mon, Jun 27, 2022 11:02 AM To: @.>; Cc: @.@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is aninvalid directory path! (Issue #48) 噢噢好的，能具体说说数据集的格式是哪出问题了嘛？要修改哪里？其他yolo训练用这个完全没问题呀。 2022-06-27 00:46:04LiJiahang617 @.>写道： 不可以，要用coco的。最后还是发现是路径有问题，按照terminal里面的提示把一些参数文件放到指定的位置就可以了，不过这个项目目前有bug，我训练的map还有loss那些全是0，解决不了 ---Original--- From: @.> Date: Sun, Jun 26, 2022 23:15 PM To: @.>; Cc: @.@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is an invalid directory path! (Issue #48) ok, i have solved it, thanks a lot 想问问怎么解决的可以嘛？可以用voc数据集的格式嘛？ — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.> — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.> — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.> Traceback (most recent call last): File "tools/train.py", line 87, in main(args) File "tools/train.py", line 76, in main trainer = Trainer(args, cfg, device) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/core/engine.py", line 40, in init self.train_loader, self.val_loader = self.get_data_loader(args, cfg, self.data_dict) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/core/engine.py", line 206, in get_data_loader train_loader = create_dataloader(train_path, args.img_size, args.batch_size // args.world_size, grid_size, File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/data_load.py", line 23, in create_dataloader dataset = TrainValDataset(path, img_size, batch_size, File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/datasets.py", line 39, in init self.img_paths, self.labels = self.get_imgs_labels(self.img_dir) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/datasets.py", line 184, in get_imgs_labels assert osp.exists(label_dir), f'{label_dir} is an invalid directory path!' AssertionError: /home/lrq/yolotest/YOLOv6-main/YOLOv6-main/data/coco/labels/train2017 is an invalid directory path! 大神，您可以帮我看一下我需要改哪里吗？ — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.>

删掉images试试

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.***>

[shenshi19]

删掉了还是显示AssertionError: Results do not correspond to current coco set。头疼 2022-06-27 16:29:59shenshi19 @.>写道： 您好，我的数据集是这么排列的，怎样改呢？ --data --annotations --instance train --instance val --images --train --val --test --labels --train --val --test 2022-06-27 15:20:36shenshi19 @.>写道： voc的话好像得改很多东西，这个用的应该就是Coco格式的数据集，应该是数据集组织不对，我把数据集文件的组织树关系按照终端上的提示改了一下，就好了，但是跑起来loss就一个有数字，其他都为0，模型跑不出来，也不知道为啥， … ---Original--- From: @.> Date: Mon, Jun 27, 2022 11:02 AM To: @.>; Cc: @.@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is aninvalid directory path! (Issue #48) 噢噢好的，能具体说说数据集的格式是哪出问题了嘛？要修改哪里？其他yolo训练用这个完全没问题呀。 2022-06-27 00:46:04LiJiahang617 @.>写道： 不可以，要用coco的。最后还是发现是路径有问题，按照terminal里面的提示把一些参数文件放到指定的位置就可以了，不过这个项目目前有bug，我训练的map还有loss那些全是0，解决不了 ---Original--- From: @.> Date: Sun, Jun 26, 2022 23:15 PM To: @.>; Cc: @.@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is an invalid directory path! (Issue #48) ok, i have solved it, thanks a lot 想问问怎么解决的可以嘛？可以用voc数据集的格式嘛？ — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.> — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.> — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.> Traceback (most recent call last): File "tools/train.py", line 87, in main(args) File "tools/train.py", line 76, in main trainer = Trainer(args, cfg, device) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/core/engine.py", line 40, in init self.train_loader, self.val_loader = self.get_data_loader(args, cfg, self.data_dict) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/core/engine.py", line 206, in get_data_loader train_loader = create_dataloader(train_path, args.img_size, args.batch_size // args.world_size, grid_size, File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/data_load.py", line 23, in create_dataloader dataset = TrainValDataset(path, img_size, batch_size, File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/datasets.py", line 39, in init self.img_paths, self.labels = self.get_imgs_labels(self.img_dir) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/datasets.py", line 184, in get_imgs_labels assert osp.exists(label_dir), f'{label_dir} is an invalid directory path!' AssertionError: /home/lrq/yolotest/YOLOv6-main/YOLOv6-main/data/coco/labels/train2017 is an invalid directory path! 大神，您可以帮我看一下我需要改哪里吗？ — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.> 删掉images试试 — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.> 我记得是train2017

[LiJiahang617]

试试csdn里面把问题关键字复制一下问一下，我这样也没法直接说啊

---Original--- From: @.> Date: Mon, Jun 27, 2022 15:47 PM To: @.>; Cc: @.**@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is an invalid directory path! (Issue #48)

你对应的那个路径里面有相应的文件嘛，那个路径你应该是没放东西，往那个路径里面放对应的东西就可以了 File "tools/train.py", line 87, in main(args) File "tools/train.py", line 76, in main trainer = Trainer(args, cfg, device) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/core/engine.py", line 40, in init self.train_loader, self.val_loader = self.get_data_loader(args, cfg, self.data_dict) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/core/engine.py", line 206, in get_data_loader train_loader = create_dataloader(train_path, args.img_size, args.batch_size // args.world_size, grid_size, File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/data_load.py", line 23, in create_dataloader dataset = TrainValDataset(path, img_size, batch_size, File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/datasets.py", line 39, in init self.img_paths, self.labels = self.get_imgs_labels(self.img_dir) File "/home/lrq/yolotest/YOLOv6-main/YOLOv6-main/yolov6/data/datasets.py", line 217, in get_imgs_labels LOGGER.warning(f'WARNING: No labels found in {osp.dirname(self.img_paths[0])}. ') AttributeError: 'TrainValDataset' object has no attribute 'img_paths' 感谢大神，但是又出现了这个新问题....

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>

[shenshi19]

Maybe you can try ‘ls + data-path' to verify the absolute path that is correct？Judging from the current problem, it should still be the problem that the path is not written correctly, thanks

我遇到了同样的问题，损失，map都是0

[AhmedEmadEldinHussin]

I have the same problem ls directory is working fine AssertionError: /home/ahmed/yolov6/images/labels/ is an invalid directory path!

[noreenanwar]

ok, i have solved it, thanks a lot

how you solve this?

[LiJiahang617]

by reorganizing my datasets files

---Original--- From: "Noreen @.> Date: Thu, Jun 30, 2022 12:32 PM To: @.>; Cc: @.**@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is aninvalid directory path! (Issue #48)

ok, i have solved it, thanks a lot

how you solve this?

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>

[noreenanwar]

I did the same but I am facing error

[geekdreamer04]

python tools/train.py --batch 32 --conf configs/yolov6s.py --data data/coco.yaml --device 0 training args are: Namespace(batch_size=32, check_images=False, check_labels=False, conf_file='configs/yolov6s.py', data_path='data/coco.yaml', device='0', dist_url='env://', epochs=400, eval_final_only=False, eval_interval=20, gpu_count=0, heavy_eval_range=50, img_size=640, local_rank=-1, name='exp', output_dir='./runs/train', rank=-1, resume=None, workers=8, world_size=1)

Using 1 GPU for training... Traceback (most recent call last): File "tools/train.py", line 94, in main(args) File "tools/train.py", line 83, in main trainer = Trainer(args, cfg, device) File "/home/ubuntu/YOLOv6/yolov6/core/engine.py", line 40, in init self.train_loader, self.val_loader = self.get_data_loader(args, cfg, self.data_dict) File "/home/ubuntu/YOLOv6/yolov6/core/engine.py", line 222, in get_data_loader train_loader = create_dataloader(train_path, args.img_size, args.batch_size // args.world_size, grid_size, File "/home/ubuntu/YOLOv6/yolov6/data/data_load.py", line 41, in create_dataloader dataset = TrainValDataset( File "/home/ubuntu/YOLOv6/yolov6/data/datasets.py", line 63, in init self.img_paths, self.labels = self.get_imgs_labels(self.img_dir) File "/home/ubuntu/YOLOv6/yolov6/data/datasets.py", line 256, in get_imgs_labels assert osp.exists(label_dir), f"{label_dir} is an invalid directory path!" AssertionError: ../coco/labels/train2014 is an invalid directory path!

I have the same issue. images are loading but the annotations are not loading. I downloaded the coco dataset from the official website and using it directly. Will I have to convert COCO annotations to .txt format? If I have to do that as well, how am I supposed to do it?

[geekdreamer04]

ok, i have solved it, thanks a lot

How did you solve the issue? I have the same issue. Apart from .json annotations do we also need .txt labels?

[LiJiahang617]

i have not solved it yet，and im confused too. I think yolo format datasets use .txt as their labels and all the things you have to do is to organzie datasets to offical datasets format,and the yolo should recognize the .txt automatically.but why this yolov6 need json file,this belong to coco format datasets

---Original--- From: "Mahathi @.> Date: Tue, Jul 5, 2022 01:56 AM To: @.>; Cc: @.**@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is aninvalid directory path! (Issue #48)

python tools/train.py --batch 32 --conf configs/yolov6s.py --data data/coco.yaml --device 0 training args are: Namespace(batch_size=32, check_images=False, check_labels=False, conf_file='configs/yolov6s.py', data_path='data/coco.yaml', device='0', dist_url='env://', epochs=400, eval_final_only=False, eval_interval=20, gpu_count=0, heavy_eval_range=50, img_size=640, local_rank=-1, name='exp', output_dir='./runs/train', rank=-1, resume=None, workers=8, world_size=1)

Using 1 GPU for training... Traceback (most recent call last): File "tools/train.py", line 94, in main(args) File "tools/train.py", line 83, in main trainer = Trainer(args, cfg, device) File "/home/ubuntu/YOLOv6/yolov6/core/engine.py", line 40, in init self.train_loader, self.val_loader = self.get_data_loader(args, cfg, self.data_dict) File "/home/ubuntu/YOLOv6/yolov6/core/engine.py", line 222, in get_data_loader train_loader = create_dataloader(train_path, args.img_size, args.batch_size // args.world_size, grid_size, File "/home/ubuntu/YOLOv6/yolov6/data/data_load.py", line 41, in create_dataloader dataset = TrainValDataset( File "/home/ubuntu/YOLOv6/yolov6/data/datasets.py", line 63, in init self.img_paths, self.labels = self.get_imgs_labels(self.img_dir) File "/home/ubuntu/YOLOv6/yolov6/data/datasets.py", line 256, in get_imgs_labels assert osp.exists(label_dir), f"{label_dir} is an invalid directory path!" AssertionError: ../coco/labels/train2014 is an invalid directory path!

I have the same issue. images are loading but the annotations are not loading. I downloaded the coco dataset from the official website and using it directly. Will I have to convert COCO annotations to .txt format? If I have to do that as well, how am I supposed to do it?

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>

[geekdreamer04]

After placing the .txt labels in the folder structure, this issue got resolved for me.

[LiJiahang617]

ok i'll try it later .Thanks a lot man.

---Original--- From: "Mahathi @.> Date: Tue, Jul 5, 2022 13:27 PM To: @.>; Cc: @.**@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is an invalid directory path! (Issue #48)

After placing the .txt labels in the folder structure, this issue got resolved for me.

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>

[noreenanwar]

I have .txt in label folder but still have same error.

[LiJiahang617]

you may have to write a python script to generate .txt file from .xml annotations.

---Original--- From: "Johnny @.> Date: Thu, Jul 7, 2022 17:00 PM To: @.>; Cc: @.**@.>; Subject: Re: [meituan/YOLOv6] AssertionError: ../coco/images/train2017 is aninvalid directory path! (Issue #48)

Sorry to interrupt, but how to generate the .txt labels files?

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>