IgorAlexey
commited on
Commit
•
0f3d64c
1
Parent(s):
3dd5a51
Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,99 @@
|
|
1 |
-
---
|
2 |
-
license: lgpl-3.0
|
3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: lgpl-3.0
|
3 |
+
language:
|
4 |
+
- en
|
5 |
+
pipeline_tag: image-feature-extraction
|
6 |
+
---
|
7 |
+
# Model Card for BoardCNN
|
8 |
+
|
9 |
+
BoardCNN implements a Convolutional Neural Network (CNN) to recognize the position from images of chess boards.
|
10 |
+
|
11 |
+
The model expects a board image as input and returns the expected positions of the pieces on the board.
|
12 |
+
|
13 |
+
## Model Details
|
14 |
+
|
15 |
+
Custom CNN architecture was implemented via pytorch
|
16 |
+
|
17 |
+
|
18 |
+
|
19 |
+
**Developed by:** Igor Alexey <br>
|
20 |
+
**Model type:** Safetensors <br>
|
21 |
+
**License:** GNU GPL v3 <br>
|
22 |
+
|
23 |
+
### Model Sources
|
24 |
+
|
25 |
+
- **Repository:** [More Information Needed]
|
26 |
+
- **Demo:** [More Information Needed]
|
27 |
+
|
28 |
+
## Uses
|
29 |
+
|
30 |
+
The model can be used to make predictions on new chess board images. The output is a 8x8 grid of chess piece symbols, representing the predicted position of pieces on the board.
|
31 |
+
|
32 |
+
|
33 |
+
### Out-of-Scope Use
|
34 |
+
|
35 |
+
The pre-trained models are not made for scanning 3D boards, although it's likely the architecture should scale well for this task with a proper training set.
|
36 |
+
|
37 |
+
## Limitations
|
38 |
+
|
39 |
+
Might not always give 100% correct output, especially on varying piece sets and board themes.
|
40 |
+
|
41 |
+
|
42 |
+
## Getting started
|
43 |
+
|
44 |
+
Use the code below to get started with the model.
|
45 |
+
|
46 |
+
[More Information Needed]
|
47 |
+
|
48 |
+
## Training Details
|
49 |
+
|
50 |
+
### Training Data
|
51 |
+
|
52 |
+
The models are trained on 5k gnerated images of valid random board positions with reasonable piece sets from lichess.
|
53 |
+
|
54 |
+
### Training Procedure
|
55 |
+
|
56 |
+
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
57 |
+
|
58 |
+
|
59 |
+
#### Training Hyperparameters
|
60 |
+
|
61 |
+
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
|
62 |
+
|
63 |
+
#### Speeds, Sizes, Times [optional]
|
64 |
+
|
65 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
66 |
+
|
67 |
+
[More Information Needed]
|
68 |
+
|
69 |
+
## Evaluation
|
70 |
+
|
71 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
72 |
+
|
73 |
+
### Testing Data, Factors & Metrics
|
74 |
+
|
75 |
+
#### Testing Data
|
76 |
+
|
77 |
+
<!-- This should link to a Dataset Card if possible. -->
|
78 |
+
|
79 |
+
[More Information Needed]
|
80 |
+
|
81 |
+
#### Factors
|
82 |
+
|
83 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
84 |
+
|
85 |
+
[More Information Needed]
|
86 |
+
|
87 |
+
#### Metrics
|
88 |
+
|
89 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
90 |
+
|
91 |
+
[More Information Needed]
|
92 |
+
|
93 |
+
### Results
|
94 |
+
|
95 |
+
[More Information Needed]
|
96 |
+
|
97 |
+
#### Summary
|
98 |
+
|
99 |
+
|