Upload train_config.json with huggingface_hub
Browse files- train_config.json +16 -14
train_config.json
CHANGED
|
@@ -1,6 +1,7 @@
|
|
| 1 |
{
|
| 2 |
"dataset": {
|
| 3 |
"repo_id": null,
|
|
|
|
| 4 |
"root": null,
|
| 5 |
"episodes": null,
|
| 6 |
"image_transforms": {
|
|
@@ -135,26 +136,27 @@
|
|
| 135 |
"freeze_vision_encoder": true,
|
| 136 |
"train_expert_only": false,
|
| 137 |
"train_state_proj": true,
|
| 138 |
-
"optimizer_lr":
|
| 139 |
"optimizer_betas": [
|
| 140 |
0.9,
|
| 141 |
0.95
|
| 142 |
],
|
| 143 |
"optimizer_eps": 1e-08,
|
| 144 |
"optimizer_weight_decay": 1e-10,
|
| 145 |
-
"scheduler_warmup_steps":
|
| 146 |
-
"scheduler_decay_steps":
|
| 147 |
"scheduler_decay_lr": 1e-05
|
| 148 |
},
|
| 149 |
"compile": true,
|
| 150 |
-
"output_dir": "outputs/train/2025-
|
| 151 |
-
"job_name": "
|
| 152 |
-
"resume":
|
|
|
|
| 153 |
"seed": 1000,
|
| 154 |
"num_workers": 4,
|
| 155 |
"batch_size": 6,
|
| 156 |
-
"gradient_accumulation_steps":
|
| 157 |
-
"steps":
|
| 158 |
"eval_freq": 20000,
|
| 159 |
"log_freq": 200,
|
| 160 |
"save_checkpoint": true,
|
|
@@ -162,7 +164,7 @@
|
|
| 162 |
"use_policy_training_preset": true,
|
| 163 |
"optimizer": {
|
| 164 |
"type": "adamw",
|
| 165 |
-
"lr":
|
| 166 |
"weight_decay": 1e-10,
|
| 167 |
"grad_clip_norm": 10.0,
|
| 168 |
"betas": [
|
|
@@ -173,9 +175,9 @@
|
|
| 173 |
},
|
| 174 |
"scheduler": {
|
| 175 |
"type": "cosine_decay_with_warmup",
|
| 176 |
-
"num_warmup_steps":
|
| 177 |
-
"num_decay_steps":
|
| 178 |
-
"peak_lr":
|
| 179 |
"decay_lr": 1e-05
|
| 180 |
},
|
| 181 |
"eval": {
|
|
@@ -184,10 +186,10 @@
|
|
| 184 |
"use_async_envs": false
|
| 185 |
},
|
| 186 |
"wandb": {
|
| 187 |
-
"enable":
|
| 188 |
"disable_artifact": true,
|
| 189 |
"project": "lerobot",
|
| 190 |
-
"entity":
|
| 191 |
"notes": null,
|
| 192 |
"run_id": null,
|
| 193 |
"mode": null
|
|
|
|
| 1 |
{
|
| 2 |
"dataset": {
|
| 3 |
"repo_id": null,
|
| 4 |
+
"repo_ids": null,
|
| 5 |
"root": null,
|
| 6 |
"episodes": null,
|
| 7 |
"image_transforms": {
|
|
|
|
| 136 |
"freeze_vision_encoder": true,
|
| 137 |
"train_expert_only": false,
|
| 138 |
"train_state_proj": true,
|
| 139 |
+
"optimizer_lr": 1e-05,
|
| 140 |
"optimizer_betas": [
|
| 141 |
0.9,
|
| 142 |
0.95
|
| 143 |
],
|
| 144 |
"optimizer_eps": 1e-08,
|
| 145 |
"optimizer_weight_decay": 1e-10,
|
| 146 |
+
"scheduler_warmup_steps": 1000,
|
| 147 |
+
"scheduler_decay_steps": 30000,
|
| 148 |
"scheduler_decay_lr": 1e-05
|
| 149 |
},
|
| 150 |
"compile": true,
|
| 151 |
+
"output_dir": "outputs/train/2025-06-01/11-37-32_pi0",
|
| 152 |
+
"job_name": "pi0",
|
| 153 |
+
"resume": false,
|
| 154 |
+
"resume_scheduler": true,
|
| 155 |
"seed": 1000,
|
| 156 |
"num_workers": 4,
|
| 157 |
"batch_size": 6,
|
| 158 |
+
"gradient_accumulation_steps": 4,
|
| 159 |
+
"steps": 100000,
|
| 160 |
"eval_freq": 20000,
|
| 161 |
"log_freq": 200,
|
| 162 |
"save_checkpoint": true,
|
|
|
|
| 164 |
"use_policy_training_preset": true,
|
| 165 |
"optimizer": {
|
| 166 |
"type": "adamw",
|
| 167 |
+
"lr": 1e-05,
|
| 168 |
"weight_decay": 1e-10,
|
| 169 |
"grad_clip_norm": 10.0,
|
| 170 |
"betas": [
|
|
|
|
| 175 |
},
|
| 176 |
"scheduler": {
|
| 177 |
"type": "cosine_decay_with_warmup",
|
| 178 |
+
"num_warmup_steps": 1000,
|
| 179 |
+
"num_decay_steps": 30000,
|
| 180 |
+
"peak_lr": 1e-05,
|
| 181 |
"decay_lr": 1e-05
|
| 182 |
},
|
| 183 |
"eval": {
|
|
|
|
| 186 |
"use_async_envs": false
|
| 187 |
},
|
| 188 |
"wandb": {
|
| 189 |
+
"enable": false,
|
| 190 |
"disable_artifact": true,
|
| 191 |
"project": "lerobot",
|
| 192 |
+
"entity": null,
|
| 193 |
"notes": null,
|
| 194 |
"run_id": null,
|
| 195 |
"mode": null
|