Upload policy weights, train config and readme

README.md

@@ -5,9 +5,9 @@ license: apache-2.0
 model_name: diffusion
 pipeline_tag: robotics
 tags:
-- diffusion
-- lerobot
 - robotics
+- lerobot
+- diffusion
 ---
 
 # Model Card for diffusion

config.json

@@ -1,5 +1,5 @@
 {
-    "type": "transformer_diffusion",
+    "type": "diffusion",
     "n_obs_steps": 2,
     "input_features": {
         "observation.state": {
@@ -37,40 +37,29 @@
     "use_amp": false,
     "use_peft": false,
     "push_to_hub": true,
-    "repo_id": null,
+    "repo_id": "ISdept/piper_arm",
     "private": null,
     "tags": null,
     "license": null,
     "pretrained_path": null,
-    "horizon": 16,
+    "horizon": 64,
     "n_action_steps": 8,
     "normalization_mapping": {
-        "VISUAL": "IDENTITY",
-        "STATE": "IDENTITY",
-        "ACTION": "IDENTITY"
+        "ACTION": "MEAN_STD",
+        "STATE": "MEAN_STD",
+        "VISUAL": "IDENTITY"
     },
+    "drop_n_last_frames": 7,
     "vision_backbone": "resnet18",
-    "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
-    "input_image_size": [
-        320,
-        320
-    ],
-    "state_dim": 7,
-    "action_dim": 7,
-    "joint_weights": [
-        1.0,
-        1.0,
-        1.0,
-        0.0,
-        1.0,
-        1.0,
-        1.0
+    "crop_shape": [
+        84,
+        84
     ],
-    "d_model": 512,
-    "nhead": 8,
-    "num_encoder_layers": 4,
-    "dim_feedforward": 512,
-    "diffusion_step_embed_dim": 256,
+    "crop_is_random": true,
+    "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
+    "use_group_norm": false,
+    "spatial_softmax_num_keypoints": 32,
+    "use_separate_rgb_encoder_per_camera": false,
     "down_dims": [
         512,
         1024,
@@ -78,7 +67,18 @@
     ],
     "kernel_size": 5,
     "n_groups": 8,
+    "diffusion_step_embed_dim": 128,
     "use_film_scale_modulation": true,
+    "noise_scheduler_type": "DDPM",
+    "num_train_timesteps": 100,
+    "beta_schedule": "squaredcos_cap_v2",
+    "beta_start": 0.0001,
+    "beta_end": 0.02,
+    "prediction_type": "epsilon",
+    "clip_sample": true,
+    "clip_sample_range": 1.0,
+    "num_inference_steps": null,
+    "do_mask_loss_for_padding": false,
     "optimizer_lr": 0.0001,
     "optimizer_betas": [
         0.95,

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:87cc8b667b825a9517b27b6195218c6bcc24c54ac2afcc9a2391a22b34b39d20
-size 1188798492
+oid sha256:c02e8fa4613c9c30fd60022532c5a7bdb162cc806cefb64845c0443005f4d654
+size 1066805588

train_config.json

@@ -127,8 +127,8 @@
         "horizon": 64,
         "n_action_steps": 8,
         "normalization_mapping": {
-            "ACTION": "IDENTITY",
-            "STATE": "IDENTITY",
+            "ACTION": "MEAN_STD",
+            "STATE": "MEAN_STD",
             "VISUAL": "IDENTITY"
         },
         "drop_n_last_frames": 7,