Link the AMD-AGI faster-action-models repository.
Browse files- ASSET_SHA256SUMS +1 -1
- README.md +2 -0
ASSET_SHA256SUMS
CHANGED
|
@@ -2,7 +2,7 @@ c485bd2ee62dbb1828eea5c5ea2bb7b0178f60d809329ebfd15bdf731f592f9a action_vae.pt
|
|
| 2 |
cfc7749b96f63bd31c3c42b5c471bf756814053e847c10f3eb003417bc523d30 LICENSE
|
| 3 |
07b1f41737ff5020427958e08e62c07a176ae395e39d7e41e1f237defa38378a NOTICE
|
| 4 |
b2cd77e9402b7897ba0d680eb45e4a38fb752e89641e3a9487a7cd4e947dad6c policy.pt
|
| 5 |
-
|
| 6 |
835a23bad8ec61cf2c27c2f9440fe1f902ae5365c29f0c1858c0cb646e883594 stats/libero_10.json
|
| 7 |
a648e545ad63d0d398ebf5735aa0e7426efdaf3fa55907c5ec50b72df6e273df stats/libero_goal.json
|
| 8 |
4eed60c361514c0f95f4ad24ec271972197e30463be07d82ec6c764cba3f6f83 stats/libero_object.json
|
|
|
|
| 2 |
cfc7749b96f63bd31c3c42b5c471bf756814053e847c10f3eb003417bc523d30 LICENSE
|
| 3 |
07b1f41737ff5020427958e08e62c07a176ae395e39d7e41e1f237defa38378a NOTICE
|
| 4 |
b2cd77e9402b7897ba0d680eb45e4a38fb752e89641e3a9487a7cd4e947dad6c policy.pt
|
| 5 |
+
09396548a125a5f604c00ea16fb622825d5c06702ef9349f0358045eaabb1fe7 README.md
|
| 6 |
835a23bad8ec61cf2c27c2f9440fe1f902ae5365c29f0c1858c0cb646e883594 stats/libero_10.json
|
| 7 |
a648e545ad63d0d398ebf5735aa0e7426efdaf3fa55907c5ec50b72df6e273df stats/libero_goal.json
|
| 8 |
4eed60c361514c0f95f4ad24ec271972197e30463be07d82ec6c764cba3f6f83 stats/libero_object.json
|
README.md
CHANGED
|
@@ -14,6 +14,8 @@ tags:
|
|
| 14 |
|
| 15 |
A latent world-action policy for the [LIBERO](https://github.com/Lifelong-Robot-Learning/LIBERO) benchmark, fine-tuned by Advanced Micro Devices, Inc. from [Wan2.2-TI2V-5B](https://huggingface.co/Wan-AI/Wan2.2-TI2V-5B).
|
| 16 |
|
|
|
|
|
|
|
| 17 |
Each 32-step action chunk is encoded as 8 latent tokens of width 16.
|
| 18 |
|
| 19 |
## Files
|
|
|
|
| 14 |
|
| 15 |
A latent world-action policy for the [LIBERO](https://github.com/Lifelong-Robot-Learning/LIBERO) benchmark, fine-tuned by Advanced Micro Devices, Inc. from [Wan2.2-TI2V-5B](https://huggingface.co/Wan-AI/Wan2.2-TI2V-5B).
|
| 16 |
|
| 17 |
+
Code: [AMD-AGI/faster-action-models](https://github.com/AMD-AGI/faster-action-models).
|
| 18 |
+
|
| 19 |
Each 32-step action chunk is encoded as 8 latent tokens of width 16.
|
| 20 |
|
| 21 |
## Files
|