Upload policy weights, train config and readme

README.md

@@ -2,20 +2,20 @@
 datasets: data/banana_plate
 library_name: lerobot
 license: apache-2.0
-model_name: diffusion
+model_name: act
 pipeline_tag: robotics
 tags:
+- act
 - lerobot
 - robotics
-- diffusion
 ---
 
-# Model Card for diffusion
+# Model Card for act
 
 <!-- Provide a quick summary of what the model is/does. -->
 
 
-[Diffusion Policy](https://huggingface.co/papers/2303.04137) treats visuomotor control as a generative diffusion process, producing smooth, multi-step action trajectories that excel at contact-rich manipulation.
+[Action Chunking with Transformers (ACT)](https://huggingface.co/papers/2304.13705) is an imitation-learning method that predicts short action chunks instead of single steps. It learns from teleoperated data and often achieves high success rates.
 
 
 This policy has been trained and pushed to the Hub using [LeRobot](https://github.com/huggingface/lerobot).

config.json

@@ -1,6 +1,6 @@
 {
-    "type": "diffusion",
-    "n_obs_steps": 2,
+    "type": "act",
+    "n_obs_steps": 1,
     "input_features": {
         "observation.state": {
             "type": "STATE",
@@ -50,50 +50,30 @@
     "tags": null,
     "license": null,
     "pretrained_path": null,
-    "horizon": 16,
-    "n_action_steps": 8,
+    "chunk_size": 100,
+    "n_action_steps": 100,
     "normalization_mapping": {
         "VISUAL": "MEAN_STD",
-        "STATE": "MIN_MAX",
-        "ACTION": "MIN_MAX"
+        "STATE": "MEAN_STD",
+        "ACTION": "MEAN_STD"
     },
-    "drop_n_last_frames": 7,
     "vision_backbone": "resnet18",
-    "crop_shape": [
-        84,
-        84
-    ],
-    "crop_is_random": true,
-    "pretrained_backbone_weights": null,
-    "use_group_norm": true,
-    "spatial_softmax_num_keypoints": 32,
-    "use_separate_rgb_encoder_per_camera": false,
-    "down_dims": [
-        512,
-        1024,
-        2048
-    ],
-    "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,
-        0.999
-    ],
-    "optimizer_eps": 1e-08,
-    "optimizer_weight_decay": 1e-06,
-    "scheduler_name": "cosine",
-    "scheduler_warmup_steps": 500
+    "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
+    "replace_final_stride_with_dilation": false,
+    "pre_norm": false,
+    "dim_model": 512,
+    "n_heads": 8,
+    "dim_feedforward": 3200,
+    "feedforward_activation": "relu",
+    "n_encoder_layers": 4,
+    "n_decoder_layers": 1,
+    "use_vae": true,
+    "latent_dim": 32,
+    "n_vae_encoder_layers": 4,
+    "temporal_ensemble_coeff": null,
+    "dropout": 0.1,
+    "kl_weight": 10.0,
+    "optimizer_lr": 1e-05,
+    "optimizer_weight_decay": 0.0001,
+    "optimizer_lr_backbone": 1e-05
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5ab9433255460c5696acd8c2c0a083c57bc4903f60d8f31f3c93342a97f5ebd0
-size 1082658752
+oid sha256:b5772347a9043d9ba7423980491cdfd816450eda8cde5d9631b41a7eb8fa9d4d
+size 206748912

train_config.json

@@ -81,8 +81,8 @@
     },
     "env": null,
     "policy": {
-        "type": "diffusion",
-        "n_obs_steps": 2,
+        "type": "act",
+        "n_obs_steps": 1,
         "input_features": {
             "observation.state": {
                 "type": "STATE",
@@ -132,55 +132,35 @@
         "tags": null,
         "license": null,
         "pretrained_path": null,
-        "horizon": 16,
-        "n_action_steps": 8,
+        "chunk_size": 100,
+        "n_action_steps": 100,
         "normalization_mapping": {
             "VISUAL": "MEAN_STD",
-            "STATE": "MIN_MAX",
-            "ACTION": "MIN_MAX"
+            "STATE": "MEAN_STD",
+            "ACTION": "MEAN_STD"
         },
-        "drop_n_last_frames": 7,
         "vision_backbone": "resnet18",
-        "crop_shape": [
-            84,
-            84
-        ],
-        "crop_is_random": true,
-        "pretrained_backbone_weights": null,
-        "use_group_norm": true,
-        "spatial_softmax_num_keypoints": 32,
-        "use_separate_rgb_encoder_per_camera": false,
-        "down_dims": [
-            512,
-            1024,
-            2048
-        ],
-        "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,
-            0.999
-        ],
-        "optimizer_eps": 1e-08,
-        "optimizer_weight_decay": 1e-06,
-        "scheduler_name": "cosine",
-        "scheduler_warmup_steps": 500
+        "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
+        "replace_final_stride_with_dilation": false,
+        "pre_norm": false,
+        "dim_model": 512,
+        "n_heads": 8,
+        "dim_feedforward": 3200,
+        "feedforward_activation": "relu",
+        "n_encoder_layers": 4,
+        "n_decoder_layers": 1,
+        "use_vae": true,
+        "latent_dim": 32,
+        "n_vae_encoder_layers": 4,
+        "temporal_ensemble_coeff": null,
+        "dropout": 0.1,
+        "kl_weight": 10.0,
+        "optimizer_lr": 1e-05,
+        "optimizer_weight_decay": 0.0001,
+        "optimizer_lr_backbone": 1e-05
     },
-    "output_dir": "/media/choi/HDD/outputs/train/Diffusion_banana_plate",
-    "job_name": "diffusion_merged",
+    "output_dir": "/media/choi/HDD/outputs/train/ACT_banana_plate",
+    "job_name": "act_merged",
     "resume": false,
     "seed": 1000,
     "num_workers": 4,
@@ -193,21 +173,17 @@
     "save_freq": 20000,
     "use_policy_training_preset": true,
     "optimizer": {
-        "type": "adam",
-        "lr": 0.0001,
-        "weight_decay": 1e-06,
+        "type": "adamw",
+        "lr": 1e-05,
+        "weight_decay": 0.0001,
         "grad_clip_norm": 10.0,
         "betas": [
-            0.95,
+            0.9,
             0.999
         ],
         "eps": 1e-08
     },
-    "scheduler": {
-        "type": "diffuser",
-        "num_warmup_steps": 500,
-        "name": "cosine"
-    },
+    "scheduler": null,
     "eval": {
         "n_episodes": 50,
         "batch_size": 50,