distant-viewing commited on
Commit
1a21874
·
verified ·
1 Parent(s): fba4173

The terms as their sources state them, and where the TransNetV2 weights come from

Browse files
Files changed (1) hide show
  1. README.md +10 -4
README.md CHANGED
@@ -24,10 +24,16 @@ here; the other models it uses come from their own repositories.
24
  The files do not share one license.
25
 
26
  - `detect_shots.pt` holds the weights of TransNetV2 (Souček and Lokoč), which
27
- its authors released under the MIT license.
28
- - `yolo26l_heads.pt` was trained with the ultralytics library, which is
29
- distributed under the AGPL-3.0 license. Read its terms before using these
30
- weights in anything that is not open source.
 
 
 
 
 
 
31
  - The two laughter files were trained by the Distant Viewing Lab and are
32
  released under the MIT license.
33
 
 
24
  The files do not share one license.
25
 
26
  - `detect_shots.pt` holds the weights of TransNetV2 (Souček and Lokoč), which
27
+ its authors released under the MIT license at
28
+ [soCzech/TransNetV2](https://github.com/soCzech/TransNetV2). It is the
29
+ PyTorch state that the repository's `inference-pytorch/convert_weights.py`
30
+ makes from the weights under `inference/transnetv2-weights`. The copy is
31
+ kept here because that repository stores its weights in Git LFS, and its
32
+ quota is often used up.
33
+ - `yolo26l_heads.pt` was trained with the ultralytics library from its
34
+ `yolo26l.pt`, and is distributed under the AGPL-3.0 license as ultralytics
35
+ is. Ultralytics sells an Enterprise License for use without the AGPL's
36
+ conditions.
37
  - The two laughter files were trained by the Distant Viewing Lab and are
38
  released under the MIT license.
39