hangX/lerobot
1
0
Fork 0
forked from tangger/lerobot
Code Pull Requests Activity

Add online training with TD-MPC as proof of concept (#338)

Browse Source
This commit is contained in:
Alexander Soare
2024-07-25 11:16:38 +01:00
committed by GitHub
parent abbb1d2367
commit f8a6574698
25 changed files with 1291 additions and 233 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
tests/test_policies.py
View File

@@ -357,7 +357,8 @@ def test_normalize(insert_temporal_dim):
# TODO(alexander-soare): `policy.use_mpc=false` was previously the default in the config yaml but it
# was changed to true. For some reason, tests would pass locally, but not in CI. So here we override
# to test with `policy.use_mpc=false`.
("xarm", "tdmpc", ["policy.use_mpc=false"], ""),
("xarm", "tdmpc", ["policy.use_mpc=false"], "use_policy"),
# ("xarm", "tdmpc", ["policy.use_mpc=true"], "use_mpc"),
(
"pusht",
"diffusion",