Add diffusion policy (train and eval works, TODO: reproduce results)

2024-02-28 15:21:30 +00:00
parent f1708c8a37
commit cf5063e50e
5 changed files with 125 additions and 31 deletions
--- a/lerobot/configs/policy/tdmpc.yaml
+++ b/lerobot/configs/policy/tdmpc.yaml
@@ -62,7 +62,7 @@ policy:
  A_scaling: 3.0

  # offline->online
-  offline_steps: 25000 # ${train_steps}/2
+  offline_steps: ${offline_steps}
  pretrained_model_path: ""
  # pretrained_model_path: "/home/rcadene/code/fowm/logs/xarm_lift/all/default/2/models/offline.pt"
  # pretrained_model_path: "/home/rcadene/code/fowm/logs/xarm_lift/all/default/2/models/final.pt"
@@ -73,4 +73,4 @@ policy:
  enc_dim: 256
  num_q: 5
  mlp_dim: 512
-  latent_dim: 50
+  latent_dim: 50