starride-teklia yschneider commited on
Commit
df2c601
1 Parent(s): b32907b

Upload folder using huggingface_hub (#1)

Browse files

- Upload folder using huggingface_hub (a7a9b71009423ceb7bc9860300fc0be521de2eca)
- Update README.md (bb2c0c1d93f899bad410b3b00972083377126f72)


Co-authored-by: Yoann Schneider <yschneider@users.noreply.huggingface.co>

Files changed (2) hide show
  1. README.md +50 -3
  2. model.pt +3 -0
README.md CHANGED
@@ -1,3 +1,50 @@
1
- ---
2
- license: cc0-1.0
3
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: YOLO
3
+ license: cc0-1.0
4
+ tags:
5
+ - YOLO
6
+ - object-detection
7
+ - dla
8
+ - historical
9
+ metrics:
10
+ - P
11
+ - R
12
+ - mAP@50
13
+ - mAP@50-95
14
+ pipeline_tag: object-detection
15
+ version:
16
+ - YOLOv8
17
+ ---
18
+
19
+ # YOLOv8 - PELLET Casimir Marius - Line + Illustration detection
20
+
21
+ This model predicts the following elements from document images:
22
+ - text line
23
+ - illustration
24
+
25
+ ## Model description
26
+
27
+ The model has been trained using the YOLOv8 library on [PELLET Casimir Marius](https://huggingface.co/datasets/Teklia/PELLET-Casimir-Marius-yolov8) dataset.
28
+ It has been trained on images with their dimensions equal to 640 pixels, starting from the YOLOv8m checkpoint.
29
+
30
+ ## Evaluation results
31
+
32
+ The model achieves the following results:
33
+
34
+ | Set | Class | Images | Instances | P | R | mAP@50 | mAP@50-95 |
35
+ | ----- | ------------ | ------ | --------- | ----- | ----- | ------ | --------- |
36
+ | train | all | 80 | 881 | 0.707 | 0.823 | 0.825 | 0.457 |
37
+ | | text line | 80 | 843 | 0.801 | 0.73 | 0.781 | 0.275 |
38
+ | | illustration | 80 | 38 | 0.613 | 0.916 | 0.87 | 0.64 |
39
+ | val | all | 10 | 129 | 0.608 | 0.872 | 0.847 | 0.493 |
40
+ | | text line | 10 | 125 | 0.791 | 0.744 | 0.7 | 0.256 |
41
+ | | illustration | 10 | 4 | 0.424 | 1 | 0.995 | 0.731 |
42
+ | test | all | 10 | 125 | 0.767 | 0.881 | 0.932 | 0.556 |
43
+ | | text line | 10 | 122 | 0.903 | 0.763 | 0.868 | 0.331 |
44
+ | | illustration | 10 | 3 | 0.631 | 1 | 0.995 | 0.78 |
45
+
46
+ ## How to use?
47
+
48
+ - Download the [weights of this model](https://huggingface.co/Teklia/PELLET-CasimirMarius-YOLOv8/resolve/main/model.pt?download=true)
49
+ - Refer to the [Ultralytics documentation](https://docs.ultralytics.com/modes/predict/) to use this model.
50
+
model.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8c25d70ce998283ecbaf6a10fe928fcc871c044cf35f975a94ce8abcc417d63f
3
+ size 54844747