Upload policy weights, train config and readme

README.md

@@ -1,13 +1,13 @@
 ---
-datasets: alexhegit/starai_lab1
+datasets: alexhegit/so101_lab1
 library_name: lerobot
 license: apache-2.0
 model_name: pi0
 pipeline_tag: robotics
 tags:
-- robotics
 - lerobot
 - pi0
+- robotics
 ---
 
 # Model Card for pi0

config.json

@@ -5,7 +5,7 @@
         "observation.state": {
             "type": "STATE",
             "shape": [
-                7
+                6
             ]
         },
         "observation.images.top": {
@@ -29,7 +29,7 @@
         "action": {
             "type": "ACTION",
             "shape": [
-                7
+                6
             ]
         }
     },
@@ -40,10 +40,10 @@
     "private": null,
     "tags": null,
     "license": null,
-    "pretrained_path": "outputs/train/pi0_starai_lab1/checkpoints/last/pretrained_model",
+    "pretrained_path": "lerobot/pi0_base",
     "paligemma_variant": "gemma_2b",
     "action_expert_variant": "gemma_300m",
-    "dtype": "float32",
+    "dtype": "bfloat16",
     "chunk_size": 50,
     "n_action_steps": 50,
     "max_state_dim": 32,
@@ -68,7 +68,7 @@
     "gradient_checkpointing": false,
     "compile_model": false,
     "compile_mode": "max-autotune",
-    "optimizer_lr": 1e-05,
+    "optimizer_lr": 2.5e-05,
     "optimizer_betas": [
         0.9,
         0.95

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6918d5c9ed64247673e84bb570624b83d859a9e5f034a49ba13d91298206f27e
-size 14005623256
+oid sha256:ae7efd956c041810a03b862713d5fcf38260936eeb4ff524b358861d8c7d744e
+size 7011548640

train_config.json

@@ -1,7 +1,7 @@
 {
     "dataset": {
-        "repo_id": "alexhegit/starai_lab1",
-        "root": "/ws/dataset/starai_dataset/",
+        "repo_id": "alexhegit/so101_lab1",
+        "root": "/ws/so101-train/so101_dataset/",
         "episodes": null,
         "image_transforms": {
             "enable": false,
@@ -87,7 +87,7 @@
             "observation.state": {
                 "type": "STATE",
                 "shape": [
-                    7
+                    6
                 ]
             },
             "observation.images.top": {
@@ -111,7 +111,7 @@
             "action": {
                 "type": "ACTION",
                 "shape": [
-                    7
+                    6
                 ]
             }
         },
@@ -122,10 +122,10 @@
         "private": null,
         "tags": null,
         "license": null,
-        "pretrained_path": "outputs/train/pi0_starai_lab1/checkpoints/last/pretrained_model",
+        "pretrained_path": "lerobot/pi0_base",
         "paligemma_variant": "gemma_2b",
         "action_expert_variant": "gemma_300m",
-        "dtype": "float32",
+        "dtype": "bfloat16",
         "chunk_size": 50,
         "n_action_steps": 50,
         "max_state_dim": 32,
@@ -150,7 +150,7 @@
         "gradient_checkpointing": false,
         "compile_model": false,
         "compile_mode": "max-autotune",
-        "optimizer_lr": 1e-05,
+        "optimizer_lr": 2.5e-05,
         "optimizer_betas": [
             0.9,
             0.95
@@ -163,21 +163,21 @@
         "scheduler_decay_lr": 2.5e-06,
         "tokenizer_max_length": 48
     },
-    "output_dir": "/ws/dataset/outputs/train/pi0_starai_lab1",
-    "job_name": "pi0_starai_lab1",
-    "resume": true,
+    "output_dir": "/ws/so101-train/outputs/train/pi0_so101_lab1",
+    "job_name": "pi0_so101_lab1",
+    "resume": false,
     "seed": 1000,
     "num_workers": 4,
-    "batch_size": 16,
+    "batch_size": 32,
     "steps": 20000,
     "eval_freq": 20000,
     "log_freq": 100,
     "save_checkpoint": true,
-    "save_freq": 5000,
+    "save_freq": 10000,
     "use_policy_training_preset": true,
     "optimizer": {
         "type": "adamw",
-        "lr": 1e-05,
+        "lr": 2.5e-05,
         "weight_decay": 0.01,
         "grad_clip_norm": 1.0,
         "betas": [
@@ -190,7 +190,7 @@
         "type": "cosine_decay_with_warmup",
         "num_warmup_steps": 1000,
         "num_decay_steps": 30000,
-        "peak_lr": 1e-05,
+        "peak_lr": 2.5e-05,
         "decay_lr": 2.5e-06
     },
     "eval": {
@@ -201,12 +201,12 @@
     "wandb": {
         "enable": true,
         "disable_artifact": true,
-        "project": "pi0_starai_lab1-1110",
+        "project": "pi0_so101_lab1-1118",
         "entity": null,
         "notes": null,
-        "run_id": "9r007sbe",
+        "run_id": "izior0oc",
         "mode": null
     },
-    "checkpoint_path": "outputs/train/pi0_starai_lab1/checkpoints/last",
+    "checkpoint_path": null,
     "rename_map": {}
 }