Upload policy weights, train config and readme

README.md

@@ -2,20 +2,20 @@
 datasets: usr256864/lr_ds
 library_name: lerobot
 license: apache-2.0
-model_name: groot
+model_name: act
 pipeline_tag: robotics
 tags:
 - robotics
+- act
 - lerobot
-- groot
 ---
 
-# Model Card for groot
+# Model Card for act
 
 <!-- Provide a quick summary of what the model is/does. -->
 
 
-_Model type not recognized — please update this template._
+[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

@@ -1,5 +1,5 @@
 {
-    "type": "groot",
+    "type": "act",
     "n_obs_steps": 1,
     "input_features": {
         "observation.image": {
@@ -25,7 +25,7 @@
             ]
         }
     },
-    "device": "cuda",
+    "device": "mps",
     "use_amp": false,
     "use_peft": false,
     "push_to_hub": true,
@@ -34,48 +34,30 @@
     "tags": null,
     "license": null,
     "pretrained_path": null,
-    "chunk_size": 50,
-    "n_action_steps": 50,
-    "max_state_dim": 64,
-    "max_action_dim": 32,
+    "chunk_size": 100,
+    "n_action_steps": 100,
     "normalization_mapping": {
-        "VISUAL": "IDENTITY",
+        "VISUAL": "MEAN_STD",
         "STATE": "MEAN_STD",
         "ACTION": "MEAN_STD"
     },
-    "image_size": [
-        224,
-        224
-    ],
-    "base_model_path": "nvidia/GR00T-N1.5-3B",
-    "tokenizer_assets_repo": "lerobot/eagle2hg-processor-groot-n1p5",
-    "embodiment_tag": "new_embodiment",
-    "tune_llm": false,
-    "tune_visual": false,
-    "tune_projector": true,
-    "tune_diffusion_model": true,
-    "lora_rank": 0,
-    "lora_alpha": 16,
-    "lora_dropout": 0.1,
-    "lora_full_model": false,
-    "optimizer_lr": 0.0001,
-    "optimizer_betas": [
-        0.95,
-        0.999
-    ],
-    "optimizer_eps": 1e-08,
-    "optimizer_weight_decay": 1e-05,
-    "warmup_ratio": 0.05,
-    "use_bf16": true,
-    "video_backend": "decord",
-    "balance_dataset_weights": true,
-    "balance_trajectory_weights": true,
-    "dataset_paths": null,
-    "output_dir": "./tmp/gr00t",
-    "save_steps": 1000,
-    "max_steps": 10000,
-    "batch_size": 32,
-    "dataloader_num_workers": 8,
-    "report_to": "wandb",
-    "resume": false
+    "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 sha256:411b556159f0e7518b7ee80cffdc85dc6f5e54c8c5722ae86b30308c349e5b48
-size 6964802952
+oid sha256:cd79073615ee3bed3448cebbc9a06fedc564a3bc620f2d4a8f9801b0806c6d91
+size 206912856

train_config.json

@@ -1,7 +1,7 @@
 {
     "dataset": {
         "repo_id": "usr256864/lr_ds",
-        "root": "../datasets/RH5_dataset_qdot/",
+        "root": "../datasets/RH5_dataset_threads_target_singleStair_doubleStair",
         "episodes": null,
         "image_transforms": {
             "enable": false,
@@ -113,14 +113,30 @@
         },
         "max_parallel_tasks": 1,
         "disable_env_checker": true,
-        "config": null,
-        "episode_length": 1000,
+        "num_envs": 1,
+        "episode_length": 1200,
         "obs_type": "pixels_agent_pos",
         "render_mode": "rgb_array",
-        "use_velocity": false
+        "use_velocity": false,
+        "task_description": "rh5 task",
+        "tasks": null,
+        "horizons": null,
+        "play": false,
+        "add_distractor": false,
+        "seed": 42,
+        "ablation": null,
+        "render_success_only": null,
+        "return_episode_data": false,
+        "eval_ds": "../datasets/RH5_dataset_debug/",
+        "num_eval_samples": 2,
+        "max_episodes_rendered": 1,
+        "absolute": false,
+        "has_distractor": false,
+        "episode_idx": null,
+        "policy_type": null
     },
     "policy": {
-        "type": "groot",
+        "type": "act",
         "n_obs_steps": 1,
         "input_features": {
             "observation.image": {
@@ -146,7 +162,7 @@
                 ]
             }
         },
-        "device": "cuda",
+        "device": "mps",
         "use_amp": false,
         "use_peft": false,
         "push_to_hub": true,
@@ -155,59 +171,41 @@
         "tags": null,
         "license": null,
         "pretrained_path": null,
-        "chunk_size": 50,
-        "n_action_steps": 50,
-        "max_state_dim": 64,
-        "max_action_dim": 32,
+        "chunk_size": 100,
+        "n_action_steps": 100,
         "normalization_mapping": {
-            "VISUAL": "IDENTITY",
+            "VISUAL": "MEAN_STD",
             "STATE": "MEAN_STD",
             "ACTION": "MEAN_STD"
         },
-        "image_size": [
-            224,
-            224
-        ],
-        "base_model_path": "nvidia/GR00T-N1.5-3B",
-        "tokenizer_assets_repo": "lerobot/eagle2hg-processor-groot-n1p5",
-        "embodiment_tag": "new_embodiment",
-        "tune_llm": false,
-        "tune_visual": false,
-        "tune_projector": true,
-        "tune_diffusion_model": true,
-        "lora_rank": 0,
-        "lora_alpha": 16,
-        "lora_dropout": 0.1,
-        "lora_full_model": false,
-        "optimizer_lr": 0.0001,
-        "optimizer_betas": [
-            0.95,
-            0.999
-        ],
-        "optimizer_eps": 1e-08,
-        "optimizer_weight_decay": 1e-05,
-        "warmup_ratio": 0.05,
-        "use_bf16": true,
-        "video_backend": "decord",
-        "balance_dataset_weights": true,
-        "balance_trajectory_weights": true,
-        "dataset_paths": null,
-        "output_dir": "./tmp/gr00t",
-        "save_steps": 1000,
-        "max_steps": 10000,
-        "batch_size": 32,
-        "dataloader_num_workers": 8,
-        "report_to": "wandb",
-        "resume": false
+        "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/2026-02-11/13-18-22_debug_train",
-    "job_name": "debug_train",
+    "output_dir": "outputs/train/2026-07-01/14-59-53_debug_experiment",
+    "job_name": "debug_experiment",
     "resume": false,
     "seed": 1000,
     "num_workers": 4,
-    "batch_size": 1,
+    "batch_size": 2,
     "steps": 2,
-    "eval_freq": 0,
+    "eval_freq": 2,
     "log_freq": 2,
     "tolerance_s": 0.0001,
     "save_checkpoint": true,
@@ -215,22 +213,16 @@
     "use_policy_training_preset": true,
     "optimizer": {
         "type": "adamw",
-        "lr": 0.0001,
-        "weight_decay": 1e-05,
+        "lr": 1e-05,
+        "weight_decay": 0.0001,
         "grad_clip_norm": 10.0,
         "betas": [
-            0.95,
+            0.9,
             0.999
         ],
         "eps": 1e-08
     },
-    "scheduler": {
-        "type": "cosine_decay_with_warmup",
-        "num_warmup_steps": 500,
-        "num_decay_steps": 10000,
-        "peak_lr": 0.0001,
-        "decay_lr": 1e-05
-    },
+    "scheduler": null,
     "eval": {
         "n_episodes": 1,
         "batch_size": 1,
@@ -251,11 +243,8 @@
     "rabc_kappa": 0.01,
     "rabc_epsilon": 1e-06,
     "rabc_head_mode": "sparse",
-    "rename_map": {},
-    "checkpoint_path": null,
-    "eval_ds": "../datasets/RH5_dataset_qdot/",
-    "episode_idx": 0,
-    "absolute": false,
-    "num_eval_samples": 20,
-    "use_velocity": true
+    "rename_map": {
+        "observation.images.top": "observation.image"
+    },
+    "checkpoint_path": null
 }