Upload policy weights, train config and readme

README.md

@@ -1,22 +1,21 @@
 ---
-base_model: lerobot/smolvla_base
-datasets: lerobot/svla_so100_stacking
+datasets: lerobot/aloha_sim_insertion_human
 library_name: lerobot
 license: apache-2.0
-model_name: smolvla
+model_name: act
 pipeline_tag: robotics
 tags:
-- robotics
-- smolvla
+- act
 - lerobot
+- robotics
 ---
 
-# Model Card for smolvla
+# Model Card for act
 
 <!-- Provide a quick summary of what the model is/does. -->
 
 
-[SmolVLA](https://huggingface.co/papers/2506.01844) is a compact, efficient vision-language-action model that achieves competitive performance at reduced computational costs and can be deployed on consumer-grade hardware.
+[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

@@ -1,18 +1,12 @@
 {
-    "type": "smolvla",
+    "type": "act",
     "n_obs_steps": 1,
     "normalization_mapping": {
-        "VISUAL": "IDENTITY",
+        "VISUAL": "MEAN_STD",
         "STATE": "MEAN_STD",
         "ACTION": "MEAN_STD"
     },
     "input_features": {
-        "observation.state": {
-            "type": "STATE",
-            "shape": [
-                6
-            ]
-        },
         "observation.images.top": {
             "type": "VISUAL",
             "shape": [
@@ -21,12 +15,10 @@
                 640
             ]
         },
-        "observation.images.wrist": {
-            "type": "VISUAL",
+        "observation.state": {
+            "type": "STATE",
             "shape": [
-                3,
-                480,
-                640
+                14
             ]
         }
     },
@@ -34,7 +26,7 @@
         "action": {
             "type": "ACTION",
             "shape": [
-                6
+                14
             ]
         }
     },
@@ -45,44 +37,25 @@
     "private": null,
     "tags": null,
     "license": null,
-    "chunk_size": 50,
-    "n_action_steps": 50,
-    "max_state_dim": 32,
-    "max_action_dim": 32,
-    "resize_imgs_with_padding": [
-        512,
-        512
-    ],
-    "empty_cameras": 0,
-    "adapt_to_pi_aloha": false,
-    "use_delta_joint_actions_aloha": false,
-    "tokenizer_max_length": 48,
-    "num_steps": 10,
-    "use_cache": true,
-    "freeze_vision_encoder": true,
-    "train_expert_only": true,
-    "train_state_proj": true,
-    "optimizer_lr": 0.0001,
-    "optimizer_betas": [
-        0.9,
-        0.95
-    ],
-    "optimizer_eps": 1e-08,
-    "optimizer_weight_decay": 1e-10,
-    "optimizer_grad_clip_norm": 10.0,
-    "scheduler_warmup_steps": 1000,
-    "scheduler_decay_steps": 30000,
-    "scheduler_decay_lr": 2.5e-06,
-    "vlm_model_name": "HuggingFaceTB/SmolVLM2-500M-Video-Instruct",
-    "load_vlm_weights": true,
-    "add_image_special_tokens": false,
-    "attention_mode": "cross_attn",
-    "prefix_length": 0,
-    "pad_language_to": "max_length",
-    "num_expert_layers": 0,
-    "num_vlm_layers": 16,
-    "self_attn_every_n_layers": 2,
-    "expert_width_multiplier": 0.75,
-    "min_period": 0.004,
-    "max_period": 4.0
+    "chunk_size": 100,
+    "n_action_steps": 100,
+    "vision_backbone": "resnet18",
+    "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 sha256:56e643a86156b47c809ddf091c3552dacc0b049a9c427d3f14af79ea70974d0d
-size 906713296
+oid sha256:928a7d6b23400e62c28416e66ae2b8678ddbd47b335a6d9e18ddb02b4ca7dbb4
+size 206766560

train_config.json

@@ -1,6 +1,6 @@
 {
     "dataset": {
-        "repo_id": "lerobot/svla_so100_stacking",
+        "repo_id": "lerobot/aloha_sim_insertion_human",
         "root": null,
         "episodes": null,
         "image_transforms": {
@@ -64,22 +64,51 @@
         "use_imagenet_stats": true,
         "video_backend": "pyav"
     },
-    "env": null,
+    "env": {
+        "type": "aloha",
+        "task": "AlohaInsertion-v0",
+        "fps": 50,
+        "features": {
+            "action": {
+                "type": "ACTION",
+                "shape": [
+                    14
+                ]
+            },
+            "agent_pos": {
+                "type": "STATE",
+                "shape": [
+                    14
+                ]
+            },
+            "pixels/top": {
+                "type": "VISUAL",
+                "shape": [
+                    480,
+                    640,
+                    3
+                ]
+            }
+        },
+        "features_map": {
+            "action": "action",
+            "agent_pos": "observation.state",
+            "top": "observation.image.top",
+            "pixels/top": "observation.images.top"
+        },
+        "episode_length": 400,
+        "obs_type": "pixels_agent_pos",
+        "render_mode": "rgb_array"
+    },
     "policy": {
-        "type": "smolvla",
+        "type": "act",
         "n_obs_steps": 1,
         "normalization_mapping": {
-            "VISUAL": "IDENTITY",
+            "VISUAL": "MEAN_STD",
             "STATE": "MEAN_STD",
             "ACTION": "MEAN_STD"
         },
         "input_features": {
-            "observation.state": {
-                "type": "STATE",
-                "shape": [
-                    6
-                ]
-            },
             "observation.images.top": {
                 "type": "VISUAL",
                 "shape": [
@@ -88,12 +117,10 @@
                     640
                 ]
             },
-            "observation.images.wrist": {
-                "type": "VISUAL",
+            "observation.state": {
+                "type": "STATE",
                 "shape": [
-                    3,
-                    480,
-                    640
+                    14
                 ]
             }
         },
@@ -101,7 +128,7 @@
             "action": {
                 "type": "ACTION",
                 "shape": [
-                    6
+                    14
                 ]
             }
         },
@@ -112,54 +139,35 @@
         "private": null,
         "tags": null,
         "license": null,
-        "chunk_size": 50,
-        "n_action_steps": 50,
-        "max_state_dim": 32,
-        "max_action_dim": 32,
-        "resize_imgs_with_padding": [
-            512,
-            512
-        ],
-        "empty_cameras": 0,
-        "adapt_to_pi_aloha": false,
-        "use_delta_joint_actions_aloha": false,
-        "tokenizer_max_length": 48,
-        "num_steps": 10,
-        "use_cache": true,
-        "freeze_vision_encoder": true,
-        "train_expert_only": true,
-        "train_state_proj": true,
-        "optimizer_lr": 0.0001,
-        "optimizer_betas": [
-            0.9,
-            0.95
-        ],
-        "optimizer_eps": 1e-08,
-        "optimizer_weight_decay": 1e-10,
-        "optimizer_grad_clip_norm": 10.0,
-        "scheduler_warmup_steps": 1000,
-        "scheduler_decay_steps": 30000,
-        "scheduler_decay_lr": 2.5e-06,
-        "vlm_model_name": "HuggingFaceTB/SmolVLM2-500M-Video-Instruct",
-        "load_vlm_weights": true,
-        "add_image_special_tokens": false,
-        "attention_mode": "cross_attn",
-        "prefix_length": 0,
-        "pad_language_to": "max_length",
-        "num_expert_layers": 0,
-        "num_vlm_layers": 16,
-        "self_attn_every_n_layers": 2,
-        "expert_width_multiplier": 0.75,
-        "min_period": 0.004,
-        "max_period": 4.0
+        "chunk_size": 100,
+        "n_action_steps": 100,
+        "vision_backbone": "resnet18",
+        "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": "outputs/train/2025-07-08/17-12-30_smolvla",
-    "job_name": "smolvla",
+    "output_dir": "outputs/train/act_aloha_insertion",
+    "job_name": "aloha_act",
     "resume": false,
     "seed": 1000,
     "num_workers": 4,
-    "batch_size": 64,
-    "steps": 200000,
+    "batch_size": 8,
+    "steps": 100000,
     "eval_freq": 20000,
     "log_freq": 200,
     "save_checkpoint": true,
@@ -167,22 +175,16 @@
     "use_policy_training_preset": true,
     "optimizer": {
         "type": "adamw",
-        "lr": 0.0001,
-        "weight_decay": 1e-10,
+        "lr": 1e-05,
+        "weight_decay": 0.0001,
         "grad_clip_norm": 10.0,
         "betas": [
             0.9,
-            0.95
+            0.999
         ],
         "eps": 1e-08
     },
-    "scheduler": {
-        "type": "cosine_decay_with_warmup",
-        "num_warmup_steps": 1000,
-        "num_decay_steps": 30000,
-        "peak_lr": 0.0001,
-        "decay_lr": 2.5e-06
-    },
+    "scheduler": null,
     "eval": {
         "n_episodes": 50,
         "batch_size": 50,