Upload policy weights, train config and readme
Browse files- README.md +2 -2
- config.json +8 -7
- model.safetensors +2 -2
- train_config.json +17 -16
README.md
CHANGED
|
@@ -1,13 +1,13 @@
|
|
| 1 |
---
|
| 2 |
-
datasets:
|
| 3 |
library_name: lerobot
|
| 4 |
license: apache-2.0
|
| 5 |
model_name: ditflow
|
| 6 |
pipeline_tag: robotics
|
| 7 |
tags:
|
| 8 |
-
- robotics
|
| 9 |
- lerobot
|
| 10 |
- ditflow
|
|
|
|
| 11 |
---
|
| 12 |
|
| 13 |
# Model Card for ditflow
|
|
|
|
| 1 |
---
|
| 2 |
+
datasets: pusht_merged_300_fine
|
| 3 |
library_name: lerobot
|
| 4 |
license: apache-2.0
|
| 5 |
model_name: ditflow
|
| 6 |
pipeline_tag: robotics
|
| 7 |
tags:
|
|
|
|
| 8 |
- lerobot
|
| 9 |
- ditflow
|
| 10 |
+
- robotics
|
| 11 |
---
|
| 12 |
|
| 13 |
# Model Card for ditflow
|
config.json
CHANGED
|
@@ -37,7 +37,8 @@
|
|
| 37 |
"normalization_mapping": {
|
| 38 |
"VISUAL": "MEAN_STD",
|
| 39 |
"STATE": "MIN_MAX",
|
| 40 |
-
"ACTION": "MIN_MAX"
|
|
|
|
| 41 |
},
|
| 42 |
"drop_n_last_frames": 7,
|
| 43 |
"vision_backbone": "resnet18",
|
|
@@ -50,19 +51,19 @@
|
|
| 50 |
"use_group_norm": true,
|
| 51 |
"spatial_softmax_num_keypoints": 32,
|
| 52 |
"use_separate_rgb_encoder_per_camera": true,
|
| 53 |
-
"frequency_embedding_dim":
|
| 54 |
-
"hidden_dim":
|
| 55 |
-
"num_blocks":
|
| 56 |
-
"num_heads":
|
| 57 |
"dropout": 0.1,
|
| 58 |
-
"dim_feedforward":
|
| 59 |
"activation": "gelu",
|
| 60 |
"training_noise_sampling": "uniform",
|
| 61 |
"clip_sample": true,
|
| 62 |
"clip_sample_range": 1.0,
|
| 63 |
"num_inference_steps": 100,
|
| 64 |
"do_mask_loss_for_padding": false,
|
| 65 |
-
"optimizer_lr": 0.
|
| 66 |
"optimizer_betas": [
|
| 67 |
0.95,
|
| 68 |
0.999
|
|
|
|
| 37 |
"normalization_mapping": {
|
| 38 |
"VISUAL": "MEAN_STD",
|
| 39 |
"STATE": "MIN_MAX",
|
| 40 |
+
"ACTION": "MIN_MAX",
|
| 41 |
+
"ENV": "MIN_MAX"
|
| 42 |
},
|
| 43 |
"drop_n_last_frames": 7,
|
| 44 |
"vision_backbone": "resnet18",
|
|
|
|
| 51 |
"use_group_norm": true,
|
| 52 |
"spatial_softmax_num_keypoints": 32,
|
| 53 |
"use_separate_rgb_encoder_per_camera": true,
|
| 54 |
+
"frequency_embedding_dim": 256,
|
| 55 |
+
"hidden_dim": 1024,
|
| 56 |
+
"num_blocks": 12,
|
| 57 |
+
"num_heads": 16,
|
| 58 |
"dropout": 0.1,
|
| 59 |
+
"dim_feedforward": 4096,
|
| 60 |
"activation": "gelu",
|
| 61 |
"training_noise_sampling": "uniform",
|
| 62 |
"clip_sample": true,
|
| 63 |
"clip_sample_range": 1.0,
|
| 64 |
"num_inference_steps": 100,
|
| 65 |
"do_mask_loss_for_padding": false,
|
| 66 |
+
"optimizer_lr": 0.0002,
|
| 67 |
"optimizer_betas": [
|
| 68 |
0.95,
|
| 69 |
0.999
|
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e412546ed36d69e450762729fc4e8ba260df7d4166501a57090c631564496182
|
| 3 |
+
size 920834376
|
train_config.json
CHANGED
|
@@ -1,7 +1,7 @@
|
|
| 1 |
{
|
| 2 |
"dataset": {
|
| 3 |
-
"repo_id": "
|
| 4 |
-
"root": "pusht_teleop_data/
|
| 5 |
"episodes": null,
|
| 6 |
"image_transforms": {
|
| 7 |
"enable": false,
|
|
@@ -169,7 +169,8 @@
|
|
| 169 |
"normalization_mapping": {
|
| 170 |
"VISUAL": "MEAN_STD",
|
| 171 |
"STATE": "MIN_MAX",
|
| 172 |
-
"ACTION": "MIN_MAX"
|
|
|
|
| 173 |
},
|
| 174 |
"drop_n_last_frames": 7,
|
| 175 |
"vision_backbone": "resnet18",
|
|
@@ -182,19 +183,19 @@
|
|
| 182 |
"use_group_norm": true,
|
| 183 |
"spatial_softmax_num_keypoints": 32,
|
| 184 |
"use_separate_rgb_encoder_per_camera": true,
|
| 185 |
-
"frequency_embedding_dim":
|
| 186 |
-
"hidden_dim":
|
| 187 |
-
"num_blocks":
|
| 188 |
-
"num_heads":
|
| 189 |
"dropout": 0.1,
|
| 190 |
-
"dim_feedforward":
|
| 191 |
"activation": "gelu",
|
| 192 |
"training_noise_sampling": "uniform",
|
| 193 |
"clip_sample": true,
|
| 194 |
"clip_sample_range": 1.0,
|
| 195 |
"num_inference_steps": 100,
|
| 196 |
"do_mask_loss_for_padding": false,
|
| 197 |
-
"optimizer_lr": 0.
|
| 198 |
"optimizer_betas": [
|
| 199 |
0.95,
|
| 200 |
0.999
|
|
@@ -209,13 +210,13 @@
|
|
| 209 |
"partial_green_t_cover_processor": false,
|
| 210 |
"features_to_include": null
|
| 211 |
},
|
| 212 |
-
"output_dir": "outputs/train/2025-12-
|
| 213 |
"job_name": "pusht_ditflow_fine",
|
| 214 |
"resume": false,
|
| 215 |
"seed": 1000,
|
| 216 |
-
"num_workers":
|
| 217 |
-
"batch_size":
|
| 218 |
-
"steps":
|
| 219 |
"eval_freq": 5000,
|
| 220 |
"log_freq": 500,
|
| 221 |
"save_checkpoint": true,
|
|
@@ -223,7 +224,7 @@
|
|
| 223 |
"use_policy_training_preset": true,
|
| 224 |
"optimizer": {
|
| 225 |
"type": "adam",
|
| 226 |
-
"lr": 0.
|
| 227 |
"weight_decay": 1e-06,
|
| 228 |
"grad_clip_norm": 10.0,
|
| 229 |
"betas": [
|
|
@@ -238,7 +239,7 @@
|
|
| 238 |
"name": "cosine"
|
| 239 |
},
|
| 240 |
"eval": {
|
| 241 |
-
"n_episodes":
|
| 242 |
"batch_size": 50,
|
| 243 |
"use_async_envs": false
|
| 244 |
},
|
|
@@ -248,7 +249,7 @@
|
|
| 248 |
"project": "policies",
|
| 249 |
"entity": null,
|
| 250 |
"notes": null,
|
| 251 |
-
"run_id": "
|
| 252 |
"mode": null
|
| 253 |
},
|
| 254 |
"checkpoint_path": null,
|
|
|
|
| 1 |
{
|
| 2 |
"dataset": {
|
| 3 |
+
"repo_id": "pusht_merged_300_fine",
|
| 4 |
+
"root": "pusht_teleop_data/pusht_merged_300_fine",
|
| 5 |
"episodes": null,
|
| 6 |
"image_transforms": {
|
| 7 |
"enable": false,
|
|
|
|
| 169 |
"normalization_mapping": {
|
| 170 |
"VISUAL": "MEAN_STD",
|
| 171 |
"STATE": "MIN_MAX",
|
| 172 |
+
"ACTION": "MIN_MAX",
|
| 173 |
+
"ENV": "MIN_MAX"
|
| 174 |
},
|
| 175 |
"drop_n_last_frames": 7,
|
| 176 |
"vision_backbone": "resnet18",
|
|
|
|
| 183 |
"use_group_norm": true,
|
| 184 |
"spatial_softmax_num_keypoints": 32,
|
| 185 |
"use_separate_rgb_encoder_per_camera": true,
|
| 186 |
+
"frequency_embedding_dim": 256,
|
| 187 |
+
"hidden_dim": 1024,
|
| 188 |
+
"num_blocks": 12,
|
| 189 |
+
"num_heads": 16,
|
| 190 |
"dropout": 0.1,
|
| 191 |
+
"dim_feedforward": 4096,
|
| 192 |
"activation": "gelu",
|
| 193 |
"training_noise_sampling": "uniform",
|
| 194 |
"clip_sample": true,
|
| 195 |
"clip_sample_range": 1.0,
|
| 196 |
"num_inference_steps": 100,
|
| 197 |
"do_mask_loss_for_padding": false,
|
| 198 |
+
"optimizer_lr": 0.0002,
|
| 199 |
"optimizer_betas": [
|
| 200 |
0.95,
|
| 201 |
0.999
|
|
|
|
| 210 |
"partial_green_t_cover_processor": false,
|
| 211 |
"features_to_include": null
|
| 212 |
},
|
| 213 |
+
"output_dir": "outputs/train/2025-12-06/12-48-03_pusht_ditflow_fine",
|
| 214 |
"job_name": "pusht_ditflow_fine",
|
| 215 |
"resume": false,
|
| 216 |
"seed": 1000,
|
| 217 |
+
"num_workers": 20,
|
| 218 |
+
"batch_size": 512,
|
| 219 |
+
"steps": 100000,
|
| 220 |
"eval_freq": 5000,
|
| 221 |
"log_freq": 500,
|
| 222 |
"save_checkpoint": true,
|
|
|
|
| 224 |
"use_policy_training_preset": true,
|
| 225 |
"optimizer": {
|
| 226 |
"type": "adam",
|
| 227 |
+
"lr": 0.0002,
|
| 228 |
"weight_decay": 1e-06,
|
| 229 |
"grad_clip_norm": 10.0,
|
| 230 |
"betas": [
|
|
|
|
| 239 |
"name": "cosine"
|
| 240 |
},
|
| 241 |
"eval": {
|
| 242 |
+
"n_episodes": 500,
|
| 243 |
"batch_size": 50,
|
| 244 |
"use_async_envs": false
|
| 245 |
},
|
|
|
|
| 249 |
"project": "policies",
|
| 250 |
"entity": null,
|
| 251 |
"notes": null,
|
| 252 |
+
"run_id": "dp72s3nd",
|
| 253 |
"mode": null
|
| 254 |
},
|
| 255 |
"checkpoint_path": null,
|