Push model using huggingface_hub.

config.json

@@ -15,7 +15,6 @@
   "gpt_output_dim": 512,
   "input_normalization_modes": {
     "observation.images.elp0": "mean_std",
-    "observation.images.elp1": "mean_std",
     "observation.state": "mean_std"
   },
   "input_shapes": {
@@ -24,18 +23,13 @@
       600,
       800
     ],
-    "observation.images.elp1": [
-      3,
-      600,
-      800
-    ],
     "observation.state": [
       6
     ]
   },
   "mlp_hidden_dim": 1024,
-  "n_action_pred_token": 3,
-  "n_obs_steps": 1,
+  "n_action_pred_token": 7,
+  "n_obs_steps": 5,
   "n_vqvae_training_steps": 20000,
   "offset_loss_weight": 10000.0,
   "output_normalization_modes": {
@@ -46,12 +40,12 @@
       6
     ]
   },
-  "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
+  "pretrained_backbone_weights": null,
   "primary_code_loss_weight": 5.0,
   "secondary_code_loss_weight": 0.5,
   "sequentially_select": false,
   "spatial_softmax_num_keypoints": 32,
-  "use_group_norm": false,
+  "use_group_norm": true,
   "vision_backbone": "resnet18",
   "vqvae_embedding_dim": 256,
   "vqvae_enc_hidden_dim": 128,

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:15844fef6890c07ebfa164be2bfc3c5ba398eef872db4c0791136085d9bf5cf9
-size 159539962
+oid sha256:3162cc741fc138f042e82c9875f0afc2ae5ef2a09584bc84c7b934657392a491
+size 159494602