AttributeError: 'NoneType' object has no attribute 'group'

[LLF-afk]

Traceback (most recent call last):

File "/22085400415/leilanfang/smoke_fusion_theraml/YOLOV-master3/tools/vid_train.py", line 131, in
launch(
└ <function launch at 0x7fbdcd768670>

File "/22085400415/leilanfang/smoke_fusion_theraml/YOLOV-master3/yolox/core/launch.py", line 98, in launch
main_func(*args)
│ └ (╒════════════════════╤══════════════════════════════════════════════════════════════════════════════════════════════════════...
└ <function main at 0x7fbdc45c75e0>

File "/22085400415/leilanfang/smoke_fusion_theraml/YOLOV-master3/tools/vid_train.py", line 115, in main
trainer.train()
│ └ <function Trainer.train at 0x7fbdc45c5d30>
└ <yolox.core.vid_trainer.Trainer object at 0x7fbe446d6640>

File "/22085400415/leilanfang/smoke_fusion_theraml/YOLOV-master3/yolox/core/vid_trainer.py", line 122, in train
self.train_in_epoch()
│ └ <function Trainer.train_in_epoch at 0x7fbdc45c5dc0>
└ <yolox.core.vid_trainer.Trainer object at 0x7fbe446d6640>

File "/22085400415/leilanfang/smoke_fusion_theraml/YOLOV-master3/yolox/core/vid_trainer.py", line 132, in train_in_epoch
self.after_epoch()
│ └ <function Trainer.after_epoch at 0x7fbdc45c7160>
└ <yolox.core.vid_trainer.Trainer object at 0x7fbe446d6640>

File "/22085400415/leilanfang/smoke_fusion_theraml/YOLOV-master3/yolox/core/vid_trainer.py", line 278, in after_epoch
self.evaluate_and_save_model()
│ └ <function Trainer.evaluate_and_save_model at 0x7fbdc45c7430>
└ <yolox.core.vid_trainer.Trainer object at 0x7fbe446d6640>

File "/22085400415/leilanfang/smoke_fusion_theraml/YOLOV-master3/yolox/core/vid_trainer.py", line 387, in evaluate_and_save_model
detail_info = extract_values(summary_info)
│ └ 'Average forward time: 0.00 ms, Average NMS time: 0.00 ms, Average inference time: 0.00 ms\n'
└ <function extract_values at 0x7fbdc45c5b80>

File "/22085400415/leilanfang/smoke_fusion_theraml/YOLOV-master3/yolox/core/vid_trainer.py", line 42, in extract_values
AP75 = re.search(r'Average Precision (AP) @[ IoU=0.75.*? ] = (\d+.\d+)', text).group(1)
│ │ └ 'Average forward time: 0.00 ms, Average NMS time: 0.00 ms, Average inference time: 0.00 ms\n'
│ └ <function search at 0x7fbe42e63af0>
└ <module 're' from '/opt/conda/lib/python3.8/re.py'>
AttributeError: 'NoneType' object has no attribute 'group'
在自制的数据集上出现这个问题，这个问题如何解决，是不是一个视频的帧数太少了

[auggieFT]

Its probably because data is not loading correctly.

[AngleBeats147753]

This may be because the number of frames in all of your videos is less than lframe + gframe.

[LLF-afk]

This may be because the number of frames in all of your videos is less than lframe + gframe.

请问你如何解决这个问题的？如果已经解决能否给我提供一些方案

[AngleBeats147753]

我是将默认的gframe从16 / 32 设置为 4 / 8。或者你可以让每个视频的帧数量大于32。（主要原因是dataset不会读取帧数量小于lframe + gframe的视频）

修改exps/customed_example/v++_SwinTiny_example.py中的self.gframe、self.gframe_val

self.gframe = 4
self.gframe_val = 8 #config your gframe_val and gframe here

[LLF-afk]

我是将默认的gframe从16 / 32 设置为 4 / 8。或者你可以让每个视频的帧数量大于32。（主要原因是dataset不会读取帧数量小于lframe + gframe的视频）

修改exps/customed_example/v++_SwinTiny_example.py中的self.gframe、self.gframe_val

self.gframe = 4
self.gframe_val = 8 #config your gframe_val and gframe here

能否加一个联系方式呢，进一步回答呢，联系方式为qq：2926968068

[LLF-afk]

This may be because the number of frames in all of your videos is less than lframe + gframe.

因为我现在主要是学习这个方向的

[LLF-afk]

= 8 #config your gframe_val and gframe here

你好，我问一下你在自制的数据集上进行训练，效果如何？

[AngleBeats147753]

= 8 #config your gframe_val and gframe here

你好，我问一下你在自制的数据集上进行训练，效果如何？

不太好，可能是我数据集问题

[dilwolf]

I got the same error. But my video frames vary between 200 and 300 frames for each video. I am trying to fine-tune YOLOv++ on a single class. Can this error be because of a single class (I think it is not a reason)? Any help or idea is really appreciated.

[LLF-afk]

I got the same error. But my video frames vary between 200 and 300 frames for each video. I am trying to fine-tune YOLOv++ on a single class. Can this error be because of a single class (I think it is not a reason)? Any help or idea is really appreciated.

我是将默认的gframe从16 / 32 设置为 4 / 8。或者你可以让每个视频的帧数量大于32。（主要原因是dataset不会读取帧数量小于lframe + gframe的视频）

修改exps/customed_example/v++_SwinTiny_example.py中的self.gframe、self.gframe_val

self.gframe = 4
self.gframe_val = 8 #config your gframe_val and gframe here

你可以参考上面的，我也是按照这个进行改进的

[dilwolf]

I got the same error. But my video frames vary between 200 and 300 frames for each video. I am trying to fine-tune YOLOv++ on a single class. Can this error be because of a single class (I think it is not a reason)? Any help or idea is really appreciated.

我是将默认的gframe从16 / 32 设置为 4 / 8。或者你可以让每个视频的帧数量大于32。（主要原因是dataset不会读取帧数量小于lframe + gframe的视频）

修改exps/customed_example/v++_SwinTiny_example.py中的self.gframe、self.gframe_val

self.gframe = 4 self.gframe_val = 8 #config your gframe_val and gframe here

你可以参考上面的，我也是按照这个进行改进的

Thank you so much for your reply, but when I got this issue first time, I tried above as you suggested. Unfortunately, it did not work for me. That's why I asked here. I am getting a similar error....

Anyway, below is my dataset:

Totally 95 videos, frames range from 200 to 300 for each video and I have single class. I splitted 20 for val set and 75 video sequences into train set. I am stuck into this problem. There is no progress.

I also re-prepared my json file structure based on json file in #115 but it did not work.

Do you think, guys is there any issue in my dataset?

[LLF-afk]

我想了解的是你用什么配置文件进行训练的，还有我没有用到json文件，你如果方便可以加我qq:2926968068&nbsp; 雷兰芳 ***@***.*** &nbsp;

…

------------------&nbsp;原始邮件&nbsp;------------------ 发件人: "Dilshod Bazarov ***@***.***&gt;; 发送时间: 2024年12月6日(星期五) 下午3:47 收件人: ***@***.***&gt;; 抄送: ***@***.***&gt;; ***@***.***&gt;; 主题: Re: [YuHengsss/YOLOV] AttributeError: &#39;NoneType&#39; object has no attribute &#39;group&#39; (Issue #110) I got the same error. But my video frames vary between 200 and 300 frames for each video. I am trying to fine-tune YOLOv++ on a single class. Can this error be because of a single class (I think it is not a reason)? Any help or idea is really appreciated. 我是将默认的gframe从16 / 32 设置为 4 / 8。或者你可以让每个视频的帧数量大于32。（主要原因是dataset不会读取帧数量小于lframe + gframe的视频） 修改exps/customed_example/v++_SwinTiny_example.py中的self.gframe、self.gframe_val self.gframe = 4 self.gframe_val = 8 #config your gframe_val and gframe here 你可以参考上面的，我也是按照这个进行改进的 Thank you so much for your reply, but when I got this issue first time, I tried above as you suggested. Unfortunately, it did not work for me. That's why I asked here. I am getting a similar error.... Anyway, below is my dataset: Totally 95 videos, frames range from 200 to 300 for each video and I have single class. I splitted 20 for val set and 75 video sequences into train set. I am stuck into this problem. There is no progress. I also re-prepared my json file structure based on json file in #115 but it did not work. Do you think, guys is there any issue in my dataset? — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: ***@***.***&gt;

[dilwolf]

What I want to know is what configuration file you used for training, and I didn't use the json file. If it's convenient for you, you can add me on QQ: 2926968068 Lei Lanfang @.***
…
------------------ Original Message ------------------ From: "Dilshod Bazarov @.>; Sent: Friday, December 6, 2024, 3:47 PM To: @.>; Cc: @.>; @.>; Subject: Re: [YuHengsss/YOLOV] AttributeError: ' NoneType' #39 ; object has no attribute ' group' #39 ; (Issue #110 ) I got the same error. But my video frames vary between 200 and 300 frames for each video. I am trying to fine-tune YOLOv++ on a single class. Can this error be because of a single class (I think it is not a reason)? Any help or idea is really appreciated. Thank you so much for your reply, but when I got this issue first time, I tried above as you suggested. Unfortunately, it did not work for me. That's why I asked here. I am getting a similar error.... Anyway, below is my dataset: Totally 95 videos, frames range from 200 to 300 for each video and I have single class. I splitted 20 for val set and 75 video sequences into train set. I am stuck into this problem. There is no progress. I also re-prepared my json file structure based on json file in #115 but it did not work. Do you think, guys is there any issue in my dataset? — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>

Usually I do not use QQ. I will try to contact you soon,
But quick question? Do you mean exp file by saying configuration file? I am using the below CLI to train:

python tools/vid_train.py -f exps/customed_example/v++_SwinTiny_example.py -c v++_swinTiny.pth --fp16

Below is my exp file:

v++_SwinTiny_example.txt

Thank you for your reply in advance !!!

[LLF-afk]

    你用的是json格式，而我用的是npy格式，我没有使用过json格式        

   self.data_dir = "YoloVOD/FIRE_VID/FIRE/"
    self.train_ann = "YoloVOD/FIRE_VID/FIRE/train_video.json"
    self.val_ann = "YoloVOD/FIRE_VID/FIRE/val_video.json"

这是我的
self.data_dir = '/smoke_fusion_theraml/tar_gz'
self.vid_train_path = '/smoke_fusion_theraml/tar_gz/yolox/data/datasets/train_seq.npy'
self.vid_val_path = '/smoke_fusion_theraml/tar_gz/yolox/data/datasets/val_seq.npy'

你可以参考这个，npy文件（在YOLOv代码的主界面中Quick Start中的：Download our COCO-style annotations for training, FGFA version training annotation and video sequences. Then, put them in these two directories: 这段里面的可以下载npy文件）：https://drive.google.com/file/d/1vJs8rLl_2oZOWCMJtk3a9ZJmdNn8cu-G/view

The train_seq.npy and val_seq.npy files are numpy arrays of lists. They can be loaded using the following command:
numpy.load('./yolox/data/datasets/train_seq.npy',allow_pickle=True)

[LLF-afk]

What I want to know is what configuration file you used for training, and I didn't use the json file. If it's convenient for you, you can add me on QQ: 2926968068 Lei Lanfang @.***
…
------------------ Original Message ------------------ From: "Dilshod Bazarov @.>; Sent: Friday, December 6, 2024, 3:47 PM To: _@**._>; Cc: _@.>; @._>; Subject: Re: [YuHengsss/YOLOV] AttributeError: ' NoneType' #39 ; object has no attribute ' group' #39 ; (Issue #110 ) I got the same error. But my video frames vary between 200 and 300 frames for each video. I am trying to fine-tune YOLOv++ on a single class. Can this error be because of a single class (I think it is not a reason)? Any help or idea is really appreciated. Thank you so much for your reply, but when I got this issue first time, I tried above as you suggested. Unfortunately, it did not work for me. That's why I asked here. I am getting a similar error.... Anyway, below is my dataset: Totally 95 videos, frames range from 200 to 300 for each video and I have single class. I splitted 20 for val set and 75 video sequences into train set. I am stuck into this problem. There is no progress. I also re-prepared my json file structure based on json file in #115 but it did not work. Do you think, guys is there any issue in my dataset? — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: _@_.*>

Usually I do not use QQ. I will try to contact you soon, But quick question? Do you mean exp file by saying configuration file? I am using the below CLI to train:

python tools/vid_train.py -f exps/customed_example/v++_SwinTiny_example.py -c v++_swinTiny.pth --fp16

Below is my exp file:

v++_SwinTiny_example.txt

Thank you for your reply in advance !!!

可以看这个，#110 (comment) 我刚刚进行评论的