Revision 2c373a27da9bc2574cfbc14ae11ec26a711fcd26 authored by Akif on 09 July 2026, 08:50:27 UTC, committed by Akif on 09 July 2026, 08:50:27 UTC
The LMA Effort Database (Kim, Neff, Lee; ACM TAP 2022) is released under CC0 1.0, so ship the finger-removed 22-joint version directly as data/extracted_fingerless.zip. reproduce.py now runs from a clean clone with no external download and on the exact skeleton the checkpoints were trained on, removing the joint-count mismatch a reviewer hits with a raw (fingered) download. Add the CC0 attribution/acknowledgment to data/README.md.
1 parent dafa2c0
| File | Mode | Size |
|---|---|---|
| assets | ||
| checkpoints | ||
| data | ||
| notebooks | ||
| results | ||
| source | ||
| .gitignore | -rw-r--r-- | 653 bytes |
| LICENSE | -rw-r--r-- | 1.3 KB |
| LICENSE.pytorch3d | -rw-r--r-- | 1.5 KB |
| README.md | -rw-r--r-- | 10.0 KB |
| Sahin_et_al_CAG_2026_Preprint.pdf | -rw-r--r-- | 5.5 MB |
| install.sh | -rwxr-xr-x | 865 bytes |
| pyproject.toml | -rw-r--r-- | 1.1 KB |
| reproduce.py | -rwxr-xr-x | 9.1 KB |
| submission.txt | -rw-r--r-- | 1.2 KB |
Computing file changes ...
README.md
Loading README.md ...
