xclele/yolov10
1
0
Fork 0
mirror of https://github.com/THU-MIG/yolov10.git synced 2025-11-04 08:56:11 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity
1,096 Commits 1 Branch 2 Tags
Commit Graph

822 Commits

Author SHA1 Message Date
Unai Gurbindo
a974079bd5
Update metrics.py 
## Possible error in the `ap_per_class` function when calculating the precision-recall curve

Hello,

I have identified a possible error in the `ap_per_class` function when calculating the metrics for the precision-recall curve. Below, I provide an example to explain it more clearly:

Let’s assume the `unique_classes` list is `[0, 1, 2]`, and the model never predicted class 1. In this case, on line 583 of the `utils/metrics.py` file, the loop will skip to the next class (class 2) because `n_p == 0` for class 1. 

However, before moving on to the next class, a zero-filled array corresponding to class 1 should be added to the `prec_values` list. Otherwise, the precision-recall curve for class 2 will incorrectly appear as if it belongs to class 1, and the legend will also fail to show the correct label for class 2.

To fix this behavior, before the `continue` statement, the code should be modified as follows:

```python
if n_p == 0 or n_l == 0:
    prec_values.append(np.zeros(1000))
    continue
```

I hope my explanation is clear, and I have correctly identified the issue. While this is a rare case where the model never predicts one of the classes in the dataset, it is still something to consider to avoid errors in the precision-recall curve visualizations.

Best regards.
 2024-10-08 12:32:30 +02:00
inisis
cd2f79c702
fix onnxslim version (#380) 2024-07-30 15:16:55 +08:00
leonnil
0e51495ce6 replace onnxsim with onnxslim 2024-07-30 04:34:28 +00:00
leonnil
aad320dd80 prediction augment (TTA) fix (#288) 2024-06-23 08:21:40 +00:00
leonnil
2c36ab0f10 ddp resume checkpoint fix (#184) 2024-06-15 16:22:32 +00:00
shiraha721
3a53d91ec8
fix cv2.error when use PIL.Image.Image (#241) 2024-06-11 21:55:42 +08:00
wa22
ea93d4f379 update huggingface hub card template 2024-06-07 16:22:42 +08:00
wa22
e161d70ff7 update huggingface hub card template 2024-06-07 16:19:04 +08:00
Pbatch
d8777c1449
Make postprocessing compatible with nobuco (#204) 2024-06-06 07:57:59 +08:00
kangsanha
e6d80f3fa6
Update __init__.py (#191) 
/Desktop/ultralytics/runs/detect/ -> /Desktop/yolov10/runs/detect/
 2024-06-04 19:19:53 +08:00
kangsanha
15bdea757a
fix onnx model val.py AssertionError (#190) 2024-06-04 19:18:04 +08:00
leonnil
ba0e6cbc1a fix onnx validation (#179) 2024-06-04 10:29:21 +00:00
wa22
e5f2f03c55 use yolov10 by default in cli 2024-06-04 11:06:48 +08:00
wa22
dfed201384 add support for HuggingFace Hub for cli 2024-06-03 23:04:27 +08:00
Ekaterina Aidova
060af8b8c1
Small fix in detection head for enabling ov export from scratch (#183) 2024-06-03 17:59:44 +08:00
wa22
1539b5a678 add calculation for flops 2024-06-03 11:21:49 +08:00
NielsRogge
a6e064d331
Track downloads (#172) 
Track downloads from the HuggingFace hub
 2024-06-03 10:23:37 +08:00
wa22
7a0cce17b6 add names when pushing to huggingface hub 2024-06-03 10:21:43 +08:00
NielsRogge
989cdb812d
Add mixin (#168) 2024-06-01 19:05:46 +08:00
wa22
8f0984817d fix val with half 2024-05-30 15:46:16 +08:00
wa22
1f6d1fc7e6 fix YOLO init 2024-05-28 21:13:14 +08:00
wa22
62aaafb3a0 fix export tensorrt with dynamic size 2024-05-28 09:30:11 +08:00
wa22
9f73bc7768 fix predict with class filter 2024-05-27 13:43:44 +08:00
wa22
c617472b55 update README 2024-05-26 23:58:44 +08:00
Thomas Friedel
799ff3be47
fix for predicting with batch_size > 1 (#47) 
* fix for predicting with batch_size > 1
 2024-05-26 19:13:12 +08:00
wa22
7a8f9643d2 fix print val map 2024-05-26 03:49:16 +00:00
Thomas Friedel
2f6208c356
fix for KeyError: 'epoch' (#48) 
* fix for KeyError: 'epoch'
 2024-05-26 10:32:05 +08:00
wa22
f19f3e521f fix val each epoch 2024-05-25 02:05:42 +00:00
wa22
4197c99fbf fix resume training 2024-05-24 15:27:28 +00:00
wa22
66bca50c47 fix non-coco val 2024-05-24 07:38:21 +00:00
wa22
483d7a9050 update 2024-05-24 06:23:19 +00:00
wa22
b714fa8bae update 2024-05-24 05:40:00 +00:00
wa22
8d9f45a955 add yolov10 2024-05-23 14:12:00 +00:00
Glenn Jocher
5be2ffbd13
ultralytics 8.1.34 Inference API robust imgsz checks (#9274) 
Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com>
 2024-03-25 02:00:38 +01:00
Muhammad Rizwan Munawar
dcb953bc41
Add https://youtu.be/uDrn9QZJ2lk and object_counting.py, ai_gym.py updates (#9265) 2024-03-24 18:22:17 +01:00
Kalen Michael
ec1d110689
ultralytics 8.1.33 fix HUB model checks (#9153) 
Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com>
Co-authored-by: Laughing <61612323+Laughing-q@users.noreply.github.com>
Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com>
 2024-03-24 06:14:15 +01:00
Glenn Jocher
fc6c66a4a4
Fix OpenVINO missing batch-size bug (#9264) 2024-03-24 05:56:37 +01:00
Laughing
33607ca278
Add bgr hyperparameter (#9139) 
Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com>
 2024-03-23 23:17:00 +01:00
Glenn Jocher
ebbbdd7e25
Fix IS_PYTHON_3_12 bug (#9258) 
Co-authored-by: UltralyticsAssistant <web@ultralytics.com>
 2024-03-23 21:18:15 +01:00
Glenn Jocher
6de99a29e6
ultralytics 8.1.32 fix CLIP backwards compatibility (#9253) 
Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com>
 2024-03-23 18:39:22 +01:00
Glenn Jocher
7d2e94bbe2
Limit ray<2.9.3 due to bug in ray==2.10.0 (#9254) 
Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com>
 2024-03-23 18:27:14 +01:00
Muhammad Rizwan Munawar
fbb180da36
Add https://youtu.be/aeAX6vWpfR0 to docs & african-wildlife dataset docs page (#9102) 
Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com>
 2024-03-23 16:20:06 +01:00
zhaoruibing
528793f452
Fix TypeError in check_cfg() (#9245) 
Co-authored-by: UltralyticsAssistant <web@ultralytics.com>
Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com>
 2024-03-23 16:09:28 +01:00
Glenn Jocher
3c179f87cb
ultralytics 8.1.31 NCNN and CLIP updates (#9235) 
Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com>
 2024-03-22 23:51:50 +01:00
Mo Li
292e028779
fix yolo classify model loading error (#9196) 2024-03-22 18:04:34 +01:00
Kalen Michael
8617fcf32d
ultralytics 8.1.30 add advanced HUB train arguments (#9110) 
Co-authored-by: UltralyticsAssistant <web@ultralytics.com>
Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com>
 2024-03-20 15:53:00 -07:00
Muhammad Rizwan Munawar
e4afd47a6e
Add brain-tumor dataset docs page (#9041) 
Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com>
 2024-03-18 22:13:01 -07:00
Glenn Jocher
ab712288f7
ultralytics 8.1.29 improved disk space checking on correct path (#8977) 
Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com>
Co-authored-by: Kayzwer <68285002+Kayzwer@users.noreply.github.com>
 2024-03-16 01:54:45 +01:00
Kayzwer
2d513a9e4b
ultralytics 8.1.28 avoid * ops on bool Tensors for RT-DETR OpenVINO export (#8937) 
Co-authored-by: UltralyticsAssistant <web@ultralytics.com>
Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com>
 2024-03-15 01:20:10 +01:00
UltralyticsAssistant
bc34ae560b Auto-format by https://ultralytics.com/actions 2024-03-15 00:12:44 +00:00