Upload policy weights, train config and readme

README.md

@@ -1,21 +1,21 @@
 ---
-datasets: U-RIL/record-test
+datasets: U-RIL/record-Pastery
 library_name: lerobot
 license: apache-2.0
-model_name: act
+model_name: diffusion
 pipeline_tag: robotics
 tags:
-- act
+- diffusion
 - lerobot
 - robotics
 ---
 
-# Model Card for act
+# Model Card for diffusion
 
 <!-- Provide a quick summary of what the model is/does. -->
 
 
-[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.
+[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.
 
 
 This policy has been trained and pushed to the Hub using [LeRobot](https://github.com/huggingface/lerobot).

config.json

@@ -1,10 +1,10 @@
 {
-    "type": "act",
-    "n_obs_steps": 1,
+    "type": "diffusion",
+    "n_obs_steps": 2,
     "normalization_mapping": {
         "VISUAL": "MEAN_STD",
-        "STATE": "MEAN_STD",
-        "ACTION": "MEAN_STD"
+        "STATE": "MIN_MAX",
+        "ACTION": "MIN_MAX"
     },
     "input_features": {
         "observation.state": {
@@ -13,11 +13,11 @@
                 6
             ]
         },
-        "observation.images.laptop": {
+        "observation.images.front": {
             "type": "VISUAL",
             "shape": [
                 3,
-                360,
+                480,
                 640
             ]
         }
@@ -37,25 +37,45 @@
     "private": null,
     "tags": null,
     "license": null,
-    "chunk_size": 100,
-    "n_action_steps": 100,
+    "horizon": 16,
+    "n_action_steps": 8,
+    "drop_n_last_frames": 7,
     "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
+    "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
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:880348cf1ba910f0f60f84c489dcf0a135576cb69ce05b3132d54dffe84fda2e
-size 206700800
+oid sha256:b1593d16c4d0b089b31d9c4c9d03d99bb0560893b727631085c348ecf6c64af9
+size 1051837384

train_config.json

@@ -1,6 +1,6 @@
 {
     "dataset": {
-        "repo_id": "U-RIL/record-test",
+        "repo_id": "U-RIL/record-Pastery",
         "root": null,
         "episodes": null,
         "image_transforms": {
@@ -66,12 +66,12 @@
     },
     "env": null,
     "policy": {
-        "type": "act",
-        "n_obs_steps": 1,
+        "type": "diffusion",
+        "n_obs_steps": 2,
         "normalization_mapping": {
             "VISUAL": "MEAN_STD",
-            "STATE": "MEAN_STD",
-            "ACTION": "MEAN_STD"
+            "STATE": "MIN_MAX",
+            "ACTION": "MIN_MAX"
         },
         "input_features": {
             "observation.state": {
@@ -80,11 +80,11 @@
                     6
                 ]
             },
-            "observation.images.laptop": {
+            "observation.images.front": {
                 "type": "VISUAL",
                 "shape": [
                     3,
-                    360,
+                    480,
                     640
                 ]
             }
@@ -104,52 +104,76 @@
         "private": null,
         "tags": null,
         "license": null,
-        "chunk_size": 100,
-        "n_action_steps": 100,
+        "horizon": 16,
+        "n_action_steps": 8,
+        "drop_n_last_frames": 7,
         "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
+        "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
     },
-    "output_dir": "outputs/train/act_record_test",
-    "job_name": "act_record_test",
+    "output_dir": "outputs/train/diffusion_so100_test",
+    "job_name": "diffusion_92_ep",
     "resume": false,
     "seed": 1000,
     "num_workers": 4,
-    "batch_size": 8,
-    "steps": 10000,
+    "batch_size": 16,
+    "steps": 22000,
     "eval_freq": 2000,
     "log_freq": 200,
     "save_checkpoint": true,
-    "save_freq": 20000,
+    "save_freq": 2000,
     "use_policy_training_preset": true,
     "optimizer": {
-        "type": "adamw",
-        "lr": 1e-05,
-        "weight_decay": 0.0001,
+        "type": "adam",
+        "lr": 0.0001,
+        "weight_decay": 1e-06,
         "grad_clip_norm": 10.0,
         "betas": [
-            0.9,
+            0.95,
             0.999
         ],
         "eps": 1e-08
     },
-    "scheduler": null,
+    "scheduler": {
+        "type": "diffuser",
+        "num_warmup_steps": 500,
+        "name": "cosine"
+    },
     "eval": {
         "n_episodes": 50,
         "batch_size": 50,