Upload policy weights, train config and readme
Browse files- README.md +3 -3
- config.json +68 -68
- model.safetensors +2 -2
- train_config.json +174 -174
README.md
CHANGED
|
@@ -1,13 +1,13 @@
|
|
| 1 |
---
|
| 2 |
-
datasets: jvillav/
|
| 3 |
library_name: lerobot
|
| 4 |
license: apache-2.0
|
| 5 |
model_name: act
|
| 6 |
pipeline_tag: robotics
|
| 7 |
tags:
|
| 8 |
-
- act
|
| 9 |
-
- robotics
|
| 10 |
- lerobot
|
|
|
|
|
|
|
| 11 |
---
|
| 12 |
|
| 13 |
# Model Card for act
|
|
|
|
| 1 |
---
|
| 2 |
+
datasets: jvillav/exp_real_4_50ep_1
|
| 3 |
library_name: lerobot
|
| 4 |
license: apache-2.0
|
| 5 |
model_name: act
|
| 6 |
pipeline_tag: robotics
|
| 7 |
tags:
|
|
|
|
|
|
|
| 8 |
- lerobot
|
| 9 |
+
- robotics
|
| 10 |
+
- act
|
| 11 |
---
|
| 12 |
|
| 13 |
# Model Card for act
|
config.json
CHANGED
|
@@ -1,69 +1,69 @@
|
|
| 1 |
-
{
|
| 2 |
-
"type": "act",
|
| 3 |
-
"n_obs_steps": 1,
|
| 4 |
-
"normalization_mapping": {
|
| 5 |
-
"VISUAL": "MEAN_STD",
|
| 6 |
-
"STATE": "MEAN_STD",
|
| 7 |
-
"ACTION": "MEAN_STD"
|
| 8 |
-
},
|
| 9 |
-
"input_features": {
|
| 10 |
-
"observation.state": {
|
| 11 |
-
"type": "STATE",
|
| 12 |
-
"shape": [
|
| 13 |
-
6
|
| 14 |
-
]
|
| 15 |
-
},
|
| 16 |
-
"observation.images.
|
| 17 |
-
"type": "VISUAL",
|
| 18 |
-
"shape": [
|
| 19 |
-
3,
|
| 20 |
-
480,
|
| 21 |
-
640
|
| 22 |
-
]
|
| 23 |
-
},
|
| 24 |
-
"observation.images.
|
| 25 |
-
"type": "VISUAL",
|
| 26 |
-
"shape": [
|
| 27 |
-
3,
|
| 28 |
-
480,
|
| 29 |
-
640
|
| 30 |
-
]
|
| 31 |
-
}
|
| 32 |
-
},
|
| 33 |
-
"output_features": {
|
| 34 |
-
"action": {
|
| 35 |
-
"type": "ACTION",
|
| 36 |
-
"shape": [
|
| 37 |
-
6
|
| 38 |
-
]
|
| 39 |
-
}
|
| 40 |
-
},
|
| 41 |
-
"device": "cuda",
|
| 42 |
-
"use_amp": false,
|
| 43 |
-
"push_to_hub": true,
|
| 44 |
-
"repo_id": "jvillav/my_policy",
|
| 45 |
-
"private": null,
|
| 46 |
-
"tags": null,
|
| 47 |
-
"license": null,
|
| 48 |
-
"chunk_size": 100,
|
| 49 |
-
"n_action_steps": 100,
|
| 50 |
-
"vision_backbone": "resnet18",
|
| 51 |
-
"pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
|
| 52 |
-
"replace_final_stride_with_dilation":
|
| 53 |
-
"pre_norm": false,
|
| 54 |
-
"dim_model": 512,
|
| 55 |
-
"n_heads": 8,
|
| 56 |
-
"dim_feedforward": 3200,
|
| 57 |
-
"feedforward_activation": "relu",
|
| 58 |
-
"n_encoder_layers": 4,
|
| 59 |
-
"n_decoder_layers": 1,
|
| 60 |
-
"use_vae": true,
|
| 61 |
-
"latent_dim": 32,
|
| 62 |
-
"n_vae_encoder_layers": 4,
|
| 63 |
-
"temporal_ensemble_coeff": null,
|
| 64 |
-
"dropout": 0.1,
|
| 65 |
-
"kl_weight": 10.0,
|
| 66 |
-
"optimizer_lr": 1e-05,
|
| 67 |
-
"optimizer_weight_decay": 0.0001,
|
| 68 |
-
"optimizer_lr_backbone": 1e-05
|
| 69 |
}
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"type": "act",
|
| 3 |
+
"n_obs_steps": 1,
|
| 4 |
+
"normalization_mapping": {
|
| 5 |
+
"VISUAL": "MEAN_STD",
|
| 6 |
+
"STATE": "MEAN_STD",
|
| 7 |
+
"ACTION": "MEAN_STD"
|
| 8 |
+
},
|
| 9 |
+
"input_features": {
|
| 10 |
+
"observation.state": {
|
| 11 |
+
"type": "STATE",
|
| 12 |
+
"shape": [
|
| 13 |
+
6
|
| 14 |
+
]
|
| 15 |
+
},
|
| 16 |
+
"observation.images.cam_front": {
|
| 17 |
+
"type": "VISUAL",
|
| 18 |
+
"shape": [
|
| 19 |
+
3,
|
| 20 |
+
480,
|
| 21 |
+
640
|
| 22 |
+
]
|
| 23 |
+
},
|
| 24 |
+
"observation.images.cam_gripper": {
|
| 25 |
+
"type": "VISUAL",
|
| 26 |
+
"shape": [
|
| 27 |
+
3,
|
| 28 |
+
480,
|
| 29 |
+
640
|
| 30 |
+
]
|
| 31 |
+
}
|
| 32 |
+
},
|
| 33 |
+
"output_features": {
|
| 34 |
+
"action": {
|
| 35 |
+
"type": "ACTION",
|
| 36 |
+
"shape": [
|
| 37 |
+
6
|
| 38 |
+
]
|
| 39 |
+
}
|
| 40 |
+
},
|
| 41 |
+
"device": "cuda",
|
| 42 |
+
"use_amp": false,
|
| 43 |
+
"push_to_hub": true,
|
| 44 |
+
"repo_id": "jvillav/my_policy",
|
| 45 |
+
"private": null,
|
| 46 |
+
"tags": null,
|
| 47 |
+
"license": null,
|
| 48 |
+
"chunk_size": 100,
|
| 49 |
+
"n_action_steps": 100,
|
| 50 |
+
"vision_backbone": "resnet18",
|
| 51 |
+
"pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
|
| 52 |
+
"replace_final_stride_with_dilation": 0,
|
| 53 |
+
"pre_norm": false,
|
| 54 |
+
"dim_model": 512,
|
| 55 |
+
"n_heads": 8,
|
| 56 |
+
"dim_feedforward": 3200,
|
| 57 |
+
"feedforward_activation": "relu",
|
| 58 |
+
"n_encoder_layers": 4,
|
| 59 |
+
"n_decoder_layers": 1,
|
| 60 |
+
"use_vae": true,
|
| 61 |
+
"latent_dim": 32,
|
| 62 |
+
"n_vae_encoder_layers": 4,
|
| 63 |
+
"temporal_ensemble_coeff": null,
|
| 64 |
+
"dropout": 0.1,
|
| 65 |
+
"kl_weight": 10.0,
|
| 66 |
+
"optimizer_lr": 1e-05,
|
| 67 |
+
"optimizer_weight_decay": 0.0001,
|
| 68 |
+
"optimizer_lr_backbone": 1e-05
|
| 69 |
}
|
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 sha256:6a1b3833a8d3aa5a8f0a6afcbead41506fcc4f47a8fba1454a84a3affba482e9
|
| 3 |
+
size 206701088
|
train_config.json
CHANGED
|
@@ -1,175 +1,175 @@
|
|
| 1 |
-
{
|
| 2 |
-
"dataset": {
|
| 3 |
-
"repo_id": "jvillav/
|
| 4 |
-
"root": null,
|
| 5 |
-
"episodes": null,
|
| 6 |
-
"image_transforms": {
|
| 7 |
-
"enable": false,
|
| 8 |
-
"max_num_transforms": 3,
|
| 9 |
-
"random_order": false,
|
| 10 |
-
"tfs": {
|
| 11 |
-
"brightness": {
|
| 12 |
-
"weight": 1.0,
|
| 13 |
-
"type": "ColorJitter",
|
| 14 |
-
"kwargs": {
|
| 15 |
-
"brightness": [
|
| 16 |
-
0.8,
|
| 17 |
-
1.2
|
| 18 |
-
]
|
| 19 |
-
}
|
| 20 |
-
},
|
| 21 |
-
"contrast": {
|
| 22 |
-
"weight": 1.0,
|
| 23 |
-
"type": "ColorJitter",
|
| 24 |
-
"kwargs": {
|
| 25 |
-
"contrast": [
|
| 26 |
-
0.8,
|
| 27 |
-
1.2
|
| 28 |
-
]
|
| 29 |
-
}
|
| 30 |
-
},
|
| 31 |
-
"saturation": {
|
| 32 |
-
"weight": 1.0,
|
| 33 |
-
"type": "ColorJitter",
|
| 34 |
-
"kwargs": {
|
| 35 |
-
"saturation": [
|
| 36 |
-
0.5,
|
| 37 |
-
1.5
|
| 38 |
-
]
|
| 39 |
-
}
|
| 40 |
-
},
|
| 41 |
-
"hue": {
|
| 42 |
-
"weight": 1.0,
|
| 43 |
-
"type": "ColorJitter",
|
| 44 |
-
"kwargs": {
|
| 45 |
-
"hue": [
|
| 46 |
-
-0.05,
|
| 47 |
-
0.05
|
| 48 |
-
]
|
| 49 |
-
}
|
| 50 |
-
},
|
| 51 |
-
"sharpness": {
|
| 52 |
-
"weight": 1.0,
|
| 53 |
-
"type": "SharpnessJitter",
|
| 54 |
-
"kwargs": {
|
| 55 |
-
"sharpness": [
|
| 56 |
-
0.5,
|
| 57 |
-
1.5
|
| 58 |
-
]
|
| 59 |
-
}
|
| 60 |
-
}
|
| 61 |
-
}
|
| 62 |
-
},
|
| 63 |
-
"revision": null,
|
| 64 |
-
"use_imagenet_stats": true,
|
| 65 |
-
"video_backend": "
|
| 66 |
-
},
|
| 67 |
-
"env": null,
|
| 68 |
-
"policy": {
|
| 69 |
-
"type": "act",
|
| 70 |
-
"n_obs_steps": 1,
|
| 71 |
-
"normalization_mapping": {
|
| 72 |
-
"VISUAL": "MEAN_STD",
|
| 73 |
-
"STATE": "MEAN_STD",
|
| 74 |
-
"ACTION": "MEAN_STD"
|
| 75 |
-
},
|
| 76 |
-
"input_features": {
|
| 77 |
-
"observation.state": {
|
| 78 |
-
"type": "STATE",
|
| 79 |
-
"shape": [
|
| 80 |
-
6
|
| 81 |
-
]
|
| 82 |
-
},
|
| 83 |
-
"observation.images.
|
| 84 |
-
"type": "VISUAL",
|
| 85 |
-
"shape": [
|
| 86 |
-
3,
|
| 87 |
-
480,
|
| 88 |
-
640
|
| 89 |
-
]
|
| 90 |
-
},
|
| 91 |
-
"observation.images.
|
| 92 |
-
"type": "VISUAL",
|
| 93 |
-
"shape": [
|
| 94 |
-
3,
|
| 95 |
-
480,
|
| 96 |
-
640
|
| 97 |
-
]
|
| 98 |
-
}
|
| 99 |
-
},
|
| 100 |
-
"output_features": {
|
| 101 |
-
"action": {
|
| 102 |
-
"type": "ACTION",
|
| 103 |
-
"shape": [
|
| 104 |
-
6
|
| 105 |
-
]
|
| 106 |
-
}
|
| 107 |
-
},
|
| 108 |
-
"device": "cuda",
|
| 109 |
-
"use_amp": false,
|
| 110 |
-
"push_to_hub": true,
|
| 111 |
-
"repo_id": "jvillav/my_policy",
|
| 112 |
-
"private": null,
|
| 113 |
-
"tags": null,
|
| 114 |
-
"license": null,
|
| 115 |
-
"chunk_size": 100,
|
| 116 |
-
"n_action_steps": 100,
|
| 117 |
-
"vision_backbone": "resnet18",
|
| 118 |
-
"pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
|
| 119 |
-
"replace_final_stride_with_dilation":
|
| 120 |
-
"pre_norm": false,
|
| 121 |
-
"dim_model": 512,
|
| 122 |
-
"n_heads": 8,
|
| 123 |
-
"dim_feedforward": 3200,
|
| 124 |
-
"feedforward_activation": "relu",
|
| 125 |
-
"n_encoder_layers": 4,
|
| 126 |
-
"n_decoder_layers": 1,
|
| 127 |
-
"use_vae": true,
|
| 128 |
-
"latent_dim": 32,
|
| 129 |
-
"n_vae_encoder_layers": 4,
|
| 130 |
-
"temporal_ensemble_coeff": null,
|
| 131 |
-
"dropout": 0.1,
|
| 132 |
-
"kl_weight": 10.0,
|
| 133 |
-
"optimizer_lr": 1e-05,
|
| 134 |
-
"optimizer_weight_decay": 0.0001,
|
| 135 |
-
"optimizer_lr_backbone": 1e-05
|
| 136 |
-
},
|
| 137 |
-
"output_dir": "outputs
|
| 138 |
-
"job_name": "
|
| 139 |
-
"resume":
|
| 140 |
-
"seed": 1000,
|
| 141 |
-
"num_workers":
|
| 142 |
-
"batch_size": 8,
|
| 143 |
-
"steps": 100000,
|
| 144 |
-
"eval_freq": 20000,
|
| 145 |
-
"log_freq":
|
| 146 |
-
"save_checkpoint": true,
|
| 147 |
-
"save_freq": 10000,
|
| 148 |
-
"use_policy_training_preset": true,
|
| 149 |
-
"optimizer": {
|
| 150 |
-
"type": "adamw",
|
| 151 |
-
"lr": 1e-05,
|
| 152 |
-
"weight_decay": 0.0001,
|
| 153 |
-
"grad_clip_norm": 10.0,
|
| 154 |
-
"betas": [
|
| 155 |
-
0.9,
|
| 156 |
-
0.999
|
| 157 |
-
],
|
| 158 |
-
"eps": 1e-08
|
| 159 |
-
},
|
| 160 |
-
"scheduler": null,
|
| 161 |
-
"eval": {
|
| 162 |
-
"n_episodes": 50,
|
| 163 |
-
"batch_size": 50,
|
| 164 |
-
"use_async_envs": false
|
| 165 |
-
},
|
| 166 |
-
"wandb": {
|
| 167 |
-
"enable": false,
|
| 168 |
-
"disable_artifact": false,
|
| 169 |
-
"project": "lerobot",
|
| 170 |
-
"entity": null,
|
| 171 |
-
"notes": null,
|
| 172 |
-
"run_id": null,
|
| 173 |
-
"mode": null
|
| 174 |
-
}
|
| 175 |
}
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"dataset": {
|
| 3 |
+
"repo_id": "jvillav/exp_real_4_50ep_1",
|
| 4 |
+
"root": null,
|
| 5 |
+
"episodes": null,
|
| 6 |
+
"image_transforms": {
|
| 7 |
+
"enable": false,
|
| 8 |
+
"max_num_transforms": 3,
|
| 9 |
+
"random_order": false,
|
| 10 |
+
"tfs": {
|
| 11 |
+
"brightness": {
|
| 12 |
+
"weight": 1.0,
|
| 13 |
+
"type": "ColorJitter",
|
| 14 |
+
"kwargs": {
|
| 15 |
+
"brightness": [
|
| 16 |
+
0.8,
|
| 17 |
+
1.2
|
| 18 |
+
]
|
| 19 |
+
}
|
| 20 |
+
},
|
| 21 |
+
"contrast": {
|
| 22 |
+
"weight": 1.0,
|
| 23 |
+
"type": "ColorJitter",
|
| 24 |
+
"kwargs": {
|
| 25 |
+
"contrast": [
|
| 26 |
+
0.8,
|
| 27 |
+
1.2
|
| 28 |
+
]
|
| 29 |
+
}
|
| 30 |
+
},
|
| 31 |
+
"saturation": {
|
| 32 |
+
"weight": 1.0,
|
| 33 |
+
"type": "ColorJitter",
|
| 34 |
+
"kwargs": {
|
| 35 |
+
"saturation": [
|
| 36 |
+
0.5,
|
| 37 |
+
1.5
|
| 38 |
+
]
|
| 39 |
+
}
|
| 40 |
+
},
|
| 41 |
+
"hue": {
|
| 42 |
+
"weight": 1.0,
|
| 43 |
+
"type": "ColorJitter",
|
| 44 |
+
"kwargs": {
|
| 45 |
+
"hue": [
|
| 46 |
+
-0.05,
|
| 47 |
+
0.05
|
| 48 |
+
]
|
| 49 |
+
}
|
| 50 |
+
},
|
| 51 |
+
"sharpness": {
|
| 52 |
+
"weight": 1.0,
|
| 53 |
+
"type": "SharpnessJitter",
|
| 54 |
+
"kwargs": {
|
| 55 |
+
"sharpness": [
|
| 56 |
+
0.5,
|
| 57 |
+
1.5
|
| 58 |
+
]
|
| 59 |
+
}
|
| 60 |
+
}
|
| 61 |
+
}
|
| 62 |
+
},
|
| 63 |
+
"revision": null,
|
| 64 |
+
"use_imagenet_stats": true,
|
| 65 |
+
"video_backend": "pyav"
|
| 66 |
+
},
|
| 67 |
+
"env": null,
|
| 68 |
+
"policy": {
|
| 69 |
+
"type": "act",
|
| 70 |
+
"n_obs_steps": 1,
|
| 71 |
+
"normalization_mapping": {
|
| 72 |
+
"VISUAL": "MEAN_STD",
|
| 73 |
+
"STATE": "MEAN_STD",
|
| 74 |
+
"ACTION": "MEAN_STD"
|
| 75 |
+
},
|
| 76 |
+
"input_features": {
|
| 77 |
+
"observation.state": {
|
| 78 |
+
"type": "STATE",
|
| 79 |
+
"shape": [
|
| 80 |
+
6
|
| 81 |
+
]
|
| 82 |
+
},
|
| 83 |
+
"observation.images.cam_front": {
|
| 84 |
+
"type": "VISUAL",
|
| 85 |
+
"shape": [
|
| 86 |
+
3,
|
| 87 |
+
480,
|
| 88 |
+
640
|
| 89 |
+
]
|
| 90 |
+
},
|
| 91 |
+
"observation.images.cam_gripper": {
|
| 92 |
+
"type": "VISUAL",
|
| 93 |
+
"shape": [
|
| 94 |
+
3,
|
| 95 |
+
480,
|
| 96 |
+
640
|
| 97 |
+
]
|
| 98 |
+
}
|
| 99 |
+
},
|
| 100 |
+
"output_features": {
|
| 101 |
+
"action": {
|
| 102 |
+
"type": "ACTION",
|
| 103 |
+
"shape": [
|
| 104 |
+
6
|
| 105 |
+
]
|
| 106 |
+
}
|
| 107 |
+
},
|
| 108 |
+
"device": "cuda",
|
| 109 |
+
"use_amp": false,
|
| 110 |
+
"push_to_hub": true,
|
| 111 |
+
"repo_id": "jvillav/my_policy",
|
| 112 |
+
"private": null,
|
| 113 |
+
"tags": null,
|
| 114 |
+
"license": null,
|
| 115 |
+
"chunk_size": 100,
|
| 116 |
+
"n_action_steps": 100,
|
| 117 |
+
"vision_backbone": "resnet18",
|
| 118 |
+
"pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
|
| 119 |
+
"replace_final_stride_with_dilation": 0,
|
| 120 |
+
"pre_norm": false,
|
| 121 |
+
"dim_model": 512,
|
| 122 |
+
"n_heads": 8,
|
| 123 |
+
"dim_feedforward": 3200,
|
| 124 |
+
"feedforward_activation": "relu",
|
| 125 |
+
"n_encoder_layers": 4,
|
| 126 |
+
"n_decoder_layers": 1,
|
| 127 |
+
"use_vae": true,
|
| 128 |
+
"latent_dim": 32,
|
| 129 |
+
"n_vae_encoder_layers": 4,
|
| 130 |
+
"temporal_ensemble_coeff": null,
|
| 131 |
+
"dropout": 0.1,
|
| 132 |
+
"kl_weight": 10.0,
|
| 133 |
+
"optimizer_lr": 1e-05,
|
| 134 |
+
"optimizer_weight_decay": 0.0001,
|
| 135 |
+
"optimizer_lr_backbone": 1e-05
|
| 136 |
+
},
|
| 137 |
+
"output_dir": "outputs\\train\\act_real_exp4_07_31_jorge",
|
| 138 |
+
"job_name": "act_so101_exp4",
|
| 139 |
+
"resume": true,
|
| 140 |
+
"seed": 1000,
|
| 141 |
+
"num_workers": 1,
|
| 142 |
+
"batch_size": 8,
|
| 143 |
+
"steps": 100000,
|
| 144 |
+
"eval_freq": 20000,
|
| 145 |
+
"log_freq": 100,
|
| 146 |
+
"save_checkpoint": true,
|
| 147 |
+
"save_freq": 10000,
|
| 148 |
+
"use_policy_training_preset": true,
|
| 149 |
+
"optimizer": {
|
| 150 |
+
"type": "adamw",
|
| 151 |
+
"lr": 1e-05,
|
| 152 |
+
"weight_decay": 0.0001,
|
| 153 |
+
"grad_clip_norm": 10.0,
|
| 154 |
+
"betas": [
|
| 155 |
+
0.9,
|
| 156 |
+
0.999
|
| 157 |
+
],
|
| 158 |
+
"eps": 1e-08
|
| 159 |
+
},
|
| 160 |
+
"scheduler": null,
|
| 161 |
+
"eval": {
|
| 162 |
+
"n_episodes": 50,
|
| 163 |
+
"batch_size": 50,
|
| 164 |
+
"use_async_envs": false
|
| 165 |
+
},
|
| 166 |
+
"wandb": {
|
| 167 |
+
"enable": false,
|
| 168 |
+
"disable_artifact": false,
|
| 169 |
+
"project": "lerobot",
|
| 170 |
+
"entity": null,
|
| 171 |
+
"notes": null,
|
| 172 |
+
"run_id": null,
|
| 173 |
+
"mode": null
|
| 174 |
+
}
|
| 175 |
}
|