Instructions to use rslxcvg/molmoact2-banana-frombase-prod-r128-c010-rw3-w8-20260519 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- LeRobot
How to use rslxcvg/molmoact2-banana-frombase-prod-r128-c010-rw3-w8-20260519 with LeRobot:
- Notebooks
- Google Colab
- Kaggle
Upload diagnostic checkpoint 010000
Browse files
config.json
ADDED
|
@@ -0,0 +1,131 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"type": "molmoact2",
|
| 3 |
+
"n_obs_steps": 1,
|
| 4 |
+
"input_features": {
|
| 5 |
+
"observation.state": {
|
| 6 |
+
"type": "STATE",
|
| 7 |
+
"shape": [
|
| 8 |
+
6
|
| 9 |
+
]
|
| 10 |
+
},
|
| 11 |
+
"observation.images.front": {
|
| 12 |
+
"type": "VISUAL",
|
| 13 |
+
"shape": [
|
| 14 |
+
3,
|
| 15 |
+
480,
|
| 16 |
+
640
|
| 17 |
+
]
|
| 18 |
+
}
|
| 19 |
+
},
|
| 20 |
+
"output_features": {
|
| 21 |
+
"action": {
|
| 22 |
+
"type": "ACTION",
|
| 23 |
+
"shape": [
|
| 24 |
+
6
|
| 25 |
+
]
|
| 26 |
+
}
|
| 27 |
+
},
|
| 28 |
+
"device": "cuda",
|
| 29 |
+
"use_amp": false,
|
| 30 |
+
"use_peft": false,
|
| 31 |
+
"push_to_hub": false,
|
| 32 |
+
"repo_id": null,
|
| 33 |
+
"private": null,
|
| 34 |
+
"tags": null,
|
| 35 |
+
"license": null,
|
| 36 |
+
"pretrained_path": null,
|
| 37 |
+
"checkpoint_path": "allenai/MolmoAct2-SO100_101",
|
| 38 |
+
"checkpoint_revision": null,
|
| 39 |
+
"checkpoint_force_download": false,
|
| 40 |
+
"trust_remote_code": true,
|
| 41 |
+
"chunk_size": 30,
|
| 42 |
+
"n_action_steps": 30,
|
| 43 |
+
"action_mode": "continuous",
|
| 44 |
+
"inference_action_mode": null,
|
| 45 |
+
"discrete_action_tokenizer": "allenai/MolmoAct2-FAST-Tokenizer",
|
| 46 |
+
"discrete_generation_max_steps": null,
|
| 47 |
+
"norm_tag": "so100_so101_molmoact2",
|
| 48 |
+
"setup_type": "single so100/so101 robotic arm in molmoact2",
|
| 49 |
+
"control_mode": "absolute joint pose",
|
| 50 |
+
"image_keys": [
|
| 51 |
+
"observation.images.front"
|
| 52 |
+
],
|
| 53 |
+
"normalize_language": true,
|
| 54 |
+
"add_setup_tokens": true,
|
| 55 |
+
"add_control_tokens": true,
|
| 56 |
+
"normalize_gripper": true,
|
| 57 |
+
"num_state_tokens": 256,
|
| 58 |
+
"max_sequence_length": null,
|
| 59 |
+
"expected_max_action_dim": 32,
|
| 60 |
+
"num_flow_timesteps": 8,
|
| 61 |
+
"flow_matching_cutoff": 1.0,
|
| 62 |
+
"flow_matching_time_offset": 0.001,
|
| 63 |
+
"flow_matching_time_scale": 0.999,
|
| 64 |
+
"flow_matching_beta_alpha": 1.0,
|
| 65 |
+
"flow_matching_beta_beta": 1.5,
|
| 66 |
+
"num_inference_steps": null,
|
| 67 |
+
"mask_action_dim_padding": true,
|
| 68 |
+
"flow_loss_action_dim_weights": [
|
| 69 |
+
8.0,
|
| 70 |
+
3.0,
|
| 71 |
+
3.0,
|
| 72 |
+
1.0,
|
| 73 |
+
1.0,
|
| 74 |
+
1.0
|
| 75 |
+
],
|
| 76 |
+
"enable_inference_cuda_graph": true,
|
| 77 |
+
"per_episode_seed": false,
|
| 78 |
+
"eval_seed": null,
|
| 79 |
+
"rtc_config": null,
|
| 80 |
+
"enable_lora_vlm": true,
|
| 81 |
+
"lora_rank": 128,
|
| 82 |
+
"lora_alpha": 16,
|
| 83 |
+
"lora_dropout": 0.05,
|
| 84 |
+
"lora_bias": "none",
|
| 85 |
+
"enable_lora_action_expert": true,
|
| 86 |
+
"enable_knowledge_insulation": false,
|
| 87 |
+
"freeze_embedding": true,
|
| 88 |
+
"train_action_expert_only": false,
|
| 89 |
+
"gradient_checkpointing": true,
|
| 90 |
+
"model_dtype": "bfloat16",
|
| 91 |
+
"softmax_auxiliary_loss": true,
|
| 92 |
+
"softmax_auxiliary_loss_scale": 0.0001,
|
| 93 |
+
"discrete_loss_token_weighting": "root_subsegments_root_tokens",
|
| 94 |
+
"optimizer_lr": 1e-05,
|
| 95 |
+
"optimizer_vit_lr": 5e-06,
|
| 96 |
+
"optimizer_connector_lr": 5e-06,
|
| 97 |
+
"optimizer_action_expert_lr": 5e-05,
|
| 98 |
+
"optimizer_betas": [
|
| 99 |
+
0.9,
|
| 100 |
+
0.95
|
| 101 |
+
],
|
| 102 |
+
"optimizer_eps": 1e-06,
|
| 103 |
+
"optimizer_weight_decay": 0.0,
|
| 104 |
+
"optimizer_grad_clip_norm": 1.0,
|
| 105 |
+
"scheduler_warmup_steps": 200,
|
| 106 |
+
"scheduler_decay_steps": null,
|
| 107 |
+
"scheduler_decay_lr": 1e-06,
|
| 108 |
+
"normalization_mapping": {
|
| 109 |
+
"VISUAL": "IDENTITY",
|
| 110 |
+
"STATE": "QUANTILES",
|
| 111 |
+
"ACTION": "QUANTILES"
|
| 112 |
+
},
|
| 113 |
+
"dataset_feature_names": {
|
| 114 |
+
"action": [
|
| 115 |
+
"shoulder_pan.pos",
|
| 116 |
+
"shoulder_lift.pos",
|
| 117 |
+
"elbow_flex.pos",
|
| 118 |
+
"wrist_flex.pos",
|
| 119 |
+
"wrist_roll.pos",
|
| 120 |
+
"gripper.pos"
|
| 121 |
+
],
|
| 122 |
+
"observation.state": [
|
| 123 |
+
"shoulder_pan.pos",
|
| 124 |
+
"shoulder_lift.pos",
|
| 125 |
+
"elbow_flex.pos",
|
| 126 |
+
"wrist_flex.pos",
|
| 127 |
+
"wrist_roll.pos",
|
| 128 |
+
"gripper.pos"
|
| 129 |
+
]
|
| 130 |
+
}
|
| 131 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cfd9e9e8e515caa1eb1f020322477a2ffc4675338f5f073ad0ce596102f052b7
|
| 3 |
+
size 12583264536
|
policy_postprocessor.json
ADDED
|
@@ -0,0 +1,36 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"name": "policy_postprocessor",
|
| 3 |
+
"steps": [
|
| 4 |
+
{
|
| 5 |
+
"registry_name": "molmoact2_clamp_action",
|
| 6 |
+
"config": {}
|
| 7 |
+
},
|
| 8 |
+
{
|
| 9 |
+
"registry_name": "molmoact2_masked_unnormalizer",
|
| 10 |
+
"config": {
|
| 11 |
+
"eps": 1e-08,
|
| 12 |
+
"features": {
|
| 13 |
+
"action": {
|
| 14 |
+
"type": "ACTION",
|
| 15 |
+
"shape": [
|
| 16 |
+
6
|
| 17 |
+
]
|
| 18 |
+
}
|
| 19 |
+
},
|
| 20 |
+
"norm_map": {
|
| 21 |
+
"VISUAL": "IDENTITY",
|
| 22 |
+
"STATE": "QUANTILES",
|
| 23 |
+
"ACTION": "QUANTILES"
|
| 24 |
+
}
|
| 25 |
+
},
|
| 26 |
+
"state_file": "policy_postprocessor_step_1_molmoact2_masked_unnormalizer.safetensors"
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"registry_name": "device_processor",
|
| 30 |
+
"config": {
|
| 31 |
+
"device": "cpu",
|
| 32 |
+
"float_dtype": null
|
| 33 |
+
}
|
| 34 |
+
}
|
| 35 |
+
]
|
| 36 |
+
}
|
policy_postprocessor_step_1_molmoact2_masked_unnormalizer.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:12d01a380ee3644aa2b3b38d7162cea8caa9f75f1cb8aeac74ea3a91c9fe37a1
|
| 3 |
+
size 6752
|
policy_preprocessor.json
ADDED
|
@@ -0,0 +1,84 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"name": "policy_preprocessor",
|
| 3 |
+
"steps": [
|
| 4 |
+
{
|
| 5 |
+
"registry_name": "rename_observations_processor",
|
| 6 |
+
"config": {
|
| 7 |
+
"rename_map": {}
|
| 8 |
+
}
|
| 9 |
+
},
|
| 10 |
+
{
|
| 11 |
+
"registry_name": "to_batch_processor",
|
| 12 |
+
"config": {}
|
| 13 |
+
},
|
| 14 |
+
{
|
| 15 |
+
"registry_name": "molmoact2_masked_normalizer",
|
| 16 |
+
"config": {
|
| 17 |
+
"eps": 1e-08,
|
| 18 |
+
"features": {
|
| 19 |
+
"observation.state": {
|
| 20 |
+
"type": "STATE",
|
| 21 |
+
"shape": [
|
| 22 |
+
6
|
| 23 |
+
]
|
| 24 |
+
},
|
| 25 |
+
"observation.images.front": {
|
| 26 |
+
"type": "VISUAL",
|
| 27 |
+
"shape": [
|
| 28 |
+
3,
|
| 29 |
+
480,
|
| 30 |
+
640
|
| 31 |
+
]
|
| 32 |
+
},
|
| 33 |
+
"action": {
|
| 34 |
+
"type": "ACTION",
|
| 35 |
+
"shape": [
|
| 36 |
+
6
|
| 37 |
+
]
|
| 38 |
+
}
|
| 39 |
+
},
|
| 40 |
+
"norm_map": {
|
| 41 |
+
"VISUAL": "IDENTITY",
|
| 42 |
+
"STATE": "QUANTILES",
|
| 43 |
+
"ACTION": "QUANTILES"
|
| 44 |
+
}
|
| 45 |
+
},
|
| 46 |
+
"state_file": "policy_preprocessor_step_2_molmoact2_masked_normalizer.safetensors"
|
| 47 |
+
},
|
| 48 |
+
{
|
| 49 |
+
"registry_name": "molmoact2_clamp_normalized",
|
| 50 |
+
"config": {}
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"registry_name": "molmoact2_pack_inputs",
|
| 54 |
+
"config": {
|
| 55 |
+
"checkpoint_path": "allenai/MolmoAct2-SO100_101",
|
| 56 |
+
"checkpoint_revision": null,
|
| 57 |
+
"checkpoint_force_download": false,
|
| 58 |
+
"trust_remote_code": true,
|
| 59 |
+
"action_mode": "continuous",
|
| 60 |
+
"discrete_action_tokenizer": "allenai/MolmoAct2-FAST-Tokenizer",
|
| 61 |
+
"image_keys": [
|
| 62 |
+
"observation.images.front"
|
| 63 |
+
],
|
| 64 |
+
"setup_type": "single so100/so101 robotic arm in molmoact2",
|
| 65 |
+
"control_mode": "absolute joint pose",
|
| 66 |
+
"normalize_language": true,
|
| 67 |
+
"add_setup_tokens": true,
|
| 68 |
+
"add_control_tokens": true,
|
| 69 |
+
"num_state_tokens": 256,
|
| 70 |
+
"max_sequence_length": null,
|
| 71 |
+
"chunk_size": 30,
|
| 72 |
+
"max_action_dim": 32,
|
| 73 |
+
"env_action_dim": 6
|
| 74 |
+
}
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"registry_name": "device_processor",
|
| 78 |
+
"config": {
|
| 79 |
+
"device": "cuda",
|
| 80 |
+
"float_dtype": null
|
| 81 |
+
}
|
| 82 |
+
}
|
| 83 |
+
]
|
| 84 |
+
}
|
policy_preprocessor_step_2_molmoact2_masked_normalizer.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:12d01a380ee3644aa2b3b38d7162cea8caa9f75f1cb8aeac74ea3a91c9fe37a1
|
| 3 |
+
size 6752
|
train_config.json
ADDED
|
@@ -0,0 +1,325 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"dataset": {
|
| 3 |
+
"repo_id": "rslxcvg/banana_act_direct_color_simple_v1_molmo_compat",
|
| 4 |
+
"root": "/mnt/vla_picknplace/outputs/lerobot/banana_act_direct_color_simple_v1_molmo_compat",
|
| 5 |
+
"episodes": null,
|
| 6 |
+
"image_transforms": {
|
| 7 |
+
"enable": true,
|
| 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 |
+
"affine": {
|
| 62 |
+
"weight": 1.0,
|
| 63 |
+
"type": "RandomAffine",
|
| 64 |
+
"kwargs": {
|
| 65 |
+
"degrees": [
|
| 66 |
+
-5.0,
|
| 67 |
+
5.0
|
| 68 |
+
],
|
| 69 |
+
"translate": [
|
| 70 |
+
0.05,
|
| 71 |
+
0.05
|
| 72 |
+
]
|
| 73 |
+
}
|
| 74 |
+
}
|
| 75 |
+
}
|
| 76 |
+
},
|
| 77 |
+
"revision": null,
|
| 78 |
+
"use_imagenet_stats": true,
|
| 79 |
+
"video_backend": "pyav",
|
| 80 |
+
"return_uint8": false,
|
| 81 |
+
"streaming": false
|
| 82 |
+
},
|
| 83 |
+
"env": null,
|
| 84 |
+
"policy": {
|
| 85 |
+
"type": "molmoact2",
|
| 86 |
+
"n_obs_steps": 1,
|
| 87 |
+
"input_features": {
|
| 88 |
+
"observation.state": {
|
| 89 |
+
"type": "STATE",
|
| 90 |
+
"shape": [
|
| 91 |
+
6
|
| 92 |
+
]
|
| 93 |
+
},
|
| 94 |
+
"observation.images.front": {
|
| 95 |
+
"type": "VISUAL",
|
| 96 |
+
"shape": [
|
| 97 |
+
3,
|
| 98 |
+
480,
|
| 99 |
+
640
|
| 100 |
+
]
|
| 101 |
+
}
|
| 102 |
+
},
|
| 103 |
+
"output_features": {
|
| 104 |
+
"action": {
|
| 105 |
+
"type": "ACTION",
|
| 106 |
+
"shape": [
|
| 107 |
+
6
|
| 108 |
+
]
|
| 109 |
+
}
|
| 110 |
+
},
|
| 111 |
+
"device": "cuda",
|
| 112 |
+
"use_amp": false,
|
| 113 |
+
"use_peft": false,
|
| 114 |
+
"push_to_hub": false,
|
| 115 |
+
"repo_id": null,
|
| 116 |
+
"private": null,
|
| 117 |
+
"tags": null,
|
| 118 |
+
"license": null,
|
| 119 |
+
"pretrained_path": null,
|
| 120 |
+
"checkpoint_path": "allenai/MolmoAct2-SO100_101",
|
| 121 |
+
"checkpoint_revision": null,
|
| 122 |
+
"checkpoint_force_download": false,
|
| 123 |
+
"trust_remote_code": true,
|
| 124 |
+
"chunk_size": 30,
|
| 125 |
+
"n_action_steps": 30,
|
| 126 |
+
"action_mode": "continuous",
|
| 127 |
+
"inference_action_mode": null,
|
| 128 |
+
"discrete_action_tokenizer": "allenai/MolmoAct2-FAST-Tokenizer",
|
| 129 |
+
"discrete_generation_max_steps": null,
|
| 130 |
+
"norm_tag": "so100_so101_molmoact2",
|
| 131 |
+
"setup_type": "single so100/so101 robotic arm in molmoact2",
|
| 132 |
+
"control_mode": "absolute joint pose",
|
| 133 |
+
"image_keys": [
|
| 134 |
+
"observation.images.front"
|
| 135 |
+
],
|
| 136 |
+
"normalize_language": true,
|
| 137 |
+
"add_setup_tokens": true,
|
| 138 |
+
"add_control_tokens": true,
|
| 139 |
+
"normalize_gripper": true,
|
| 140 |
+
"num_state_tokens": 256,
|
| 141 |
+
"max_sequence_length": null,
|
| 142 |
+
"expected_max_action_dim": 32,
|
| 143 |
+
"num_flow_timesteps": 8,
|
| 144 |
+
"flow_matching_cutoff": 1.0,
|
| 145 |
+
"flow_matching_time_offset": 0.001,
|
| 146 |
+
"flow_matching_time_scale": 0.999,
|
| 147 |
+
"flow_matching_beta_alpha": 1.0,
|
| 148 |
+
"flow_matching_beta_beta": 1.5,
|
| 149 |
+
"num_inference_steps": null,
|
| 150 |
+
"mask_action_dim_padding": true,
|
| 151 |
+
"flow_loss_action_dim_weights": [
|
| 152 |
+
8.0,
|
| 153 |
+
3.0,
|
| 154 |
+
3.0,
|
| 155 |
+
1.0,
|
| 156 |
+
1.0,
|
| 157 |
+
1.0
|
| 158 |
+
],
|
| 159 |
+
"enable_inference_cuda_graph": true,
|
| 160 |
+
"per_episode_seed": false,
|
| 161 |
+
"eval_seed": null,
|
| 162 |
+
"rtc_config": null,
|
| 163 |
+
"enable_lora_vlm": true,
|
| 164 |
+
"lora_rank": 128,
|
| 165 |
+
"lora_alpha": 16,
|
| 166 |
+
"lora_dropout": 0.05,
|
| 167 |
+
"lora_bias": "none",
|
| 168 |
+
"enable_lora_action_expert": true,
|
| 169 |
+
"enable_knowledge_insulation": false,
|
| 170 |
+
"freeze_embedding": true,
|
| 171 |
+
"train_action_expert_only": false,
|
| 172 |
+
"gradient_checkpointing": true,
|
| 173 |
+
"model_dtype": "bfloat16",
|
| 174 |
+
"softmax_auxiliary_loss": true,
|
| 175 |
+
"softmax_auxiliary_loss_scale": 0.0001,
|
| 176 |
+
"discrete_loss_token_weighting": "root_subsegments_root_tokens",
|
| 177 |
+
"optimizer_lr": 1e-05,
|
| 178 |
+
"optimizer_vit_lr": 5e-06,
|
| 179 |
+
"optimizer_connector_lr": 5e-06,
|
| 180 |
+
"optimizer_action_expert_lr": 5e-05,
|
| 181 |
+
"optimizer_betas": [
|
| 182 |
+
0.9,
|
| 183 |
+
0.95
|
| 184 |
+
],
|
| 185 |
+
"optimizer_eps": 1e-06,
|
| 186 |
+
"optimizer_weight_decay": 0.0,
|
| 187 |
+
"optimizer_grad_clip_norm": 1.0,
|
| 188 |
+
"scheduler_warmup_steps": 200,
|
| 189 |
+
"scheduler_decay_steps": null,
|
| 190 |
+
"scheduler_decay_lr": 1e-06,
|
| 191 |
+
"normalization_mapping": {
|
| 192 |
+
"VISUAL": "IDENTITY",
|
| 193 |
+
"STATE": "QUANTILES",
|
| 194 |
+
"ACTION": "QUANTILES"
|
| 195 |
+
},
|
| 196 |
+
"dataset_feature_names": {
|
| 197 |
+
"action": [
|
| 198 |
+
"shoulder_pan.pos",
|
| 199 |
+
"shoulder_lift.pos",
|
| 200 |
+
"elbow_flex.pos",
|
| 201 |
+
"wrist_flex.pos",
|
| 202 |
+
"wrist_roll.pos",
|
| 203 |
+
"gripper.pos"
|
| 204 |
+
],
|
| 205 |
+
"observation.state": [
|
| 206 |
+
"shoulder_pan.pos",
|
| 207 |
+
"shoulder_lift.pos",
|
| 208 |
+
"elbow_flex.pos",
|
| 209 |
+
"wrist_flex.pos",
|
| 210 |
+
"wrist_roll.pos",
|
| 211 |
+
"gripper.pos"
|
| 212 |
+
]
|
| 213 |
+
}
|
| 214 |
+
},
|
| 215 |
+
"reward_model": null,
|
| 216 |
+
"output_dir": "/mnt/vla_picknplace/outputs/molmoact2/molmoact2_overnight_frombase_prod_r128_c010_rw3_w8_gpu5_20260519_fullcoverage_v1",
|
| 217 |
+
"job_name": "molmoact2_overnight_frombase_prod_r128_c010_rw3_w8_gpu5_20260519_fullcoverage_v1",
|
| 218 |
+
"resume": false,
|
| 219 |
+
"seed": 1000,
|
| 220 |
+
"cudnn_deterministic": false,
|
| 221 |
+
"num_workers": 4,
|
| 222 |
+
"batch_size": 16,
|
| 223 |
+
"prefetch_factor": 4,
|
| 224 |
+
"persistent_workers": true,
|
| 225 |
+
"steps": 10000,
|
| 226 |
+
"eval_freq": -1,
|
| 227 |
+
"log_freq": 20,
|
| 228 |
+
"tolerance_s": 0.0001,
|
| 229 |
+
"save_checkpoint": true,
|
| 230 |
+
"save_freq": 1000,
|
| 231 |
+
"use_policy_training_preset": true,
|
| 232 |
+
"optimizer": {
|
| 233 |
+
"type": "adamw",
|
| 234 |
+
"lr": 1e-05,
|
| 235 |
+
"weight_decay": 0.0,
|
| 236 |
+
"grad_clip_norm": 1.0,
|
| 237 |
+
"betas": [
|
| 238 |
+
0.9,
|
| 239 |
+
0.95
|
| 240 |
+
],
|
| 241 |
+
"eps": 1e-06
|
| 242 |
+
},
|
| 243 |
+
"scheduler": {
|
| 244 |
+
"type": "molmoact2_cosine_decay_with_warmup",
|
| 245 |
+
"num_warmup_steps": 200,
|
| 246 |
+
"num_decay_steps": null,
|
| 247 |
+
"peak_lr": 1e-05,
|
| 248 |
+
"decay_lr": 1e-06
|
| 249 |
+
},
|
| 250 |
+
"eval": {
|
| 251 |
+
"n_episodes": 50,
|
| 252 |
+
"batch_size": 50,
|
| 253 |
+
"use_async_envs": true
|
| 254 |
+
},
|
| 255 |
+
"wandb": {
|
| 256 |
+
"enable": true,
|
| 257 |
+
"disable_artifact": true,
|
| 258 |
+
"project": "vla_picknplace",
|
| 259 |
+
"entity": null,
|
| 260 |
+
"notes": null,
|
| 261 |
+
"run_id": "we76vk6a",
|
| 262 |
+
"mode": "online",
|
| 263 |
+
"add_tags": true
|
| 264 |
+
},
|
| 265 |
+
"peft": null,
|
| 266 |
+
"sample_weighting": {
|
| 267 |
+
"type": "manifest",
|
| 268 |
+
"progress_path": null,
|
| 269 |
+
"weights_path": "/mnt/vla_picknplace/outputs/molmoact2/decision_manifests/banana_act_direct_color_simple_v1_molmo_compat_manifest.parquet",
|
| 270 |
+
"weight_column": "loss_weight_normalized",
|
| 271 |
+
"index_column": "index",
|
| 272 |
+
"normalize_batch_mean": false,
|
| 273 |
+
"clip_min": 0.25,
|
| 274 |
+
"clip_max": 4.0,
|
| 275 |
+
"head_mode": "sparse",
|
| 276 |
+
"kappa": 0.01,
|
| 277 |
+
"epsilon": 1e-06,
|
| 278 |
+
"extra_params": {}
|
| 279 |
+
},
|
| 280 |
+
"frame_sampling": {
|
| 281 |
+
"type": "manifest",
|
| 282 |
+
"weights_path": "/mnt/vla_picknplace/outputs/molmoact2/decision_manifests/banana_act_direct_color_simple_v1_molmo_compat_manifest.parquet",
|
| 283 |
+
"weight_column": "sample_weight",
|
| 284 |
+
"index_column": "index",
|
| 285 |
+
"replacement": false,
|
| 286 |
+
"num_samples": null
|
| 287 |
+
},
|
| 288 |
+
"prompt_contrast": {
|
| 289 |
+
"type": "same_layout_reference",
|
| 290 |
+
"manifest_path": "/mnt/vla_picknplace/outputs/molmoact2/decision_manifests/banana_act_direct_color_simple_v1_molmo_compat_manifest.parquet",
|
| 291 |
+
"weight": 0.1,
|
| 292 |
+
"max_anchors_per_batch": 1,
|
| 293 |
+
"min_valid_horizon": 30,
|
| 294 |
+
"phase_min": 0.55,
|
| 295 |
+
"phase_max": 0.72,
|
| 296 |
+
"max_reference_phase_delta": 0.03,
|
| 297 |
+
"max_reference_state_l2": null,
|
| 298 |
+
"max_reference_joint_abs": [
|
| 299 |
+
40.0,
|
| 300 |
+
25.0,
|
| 301 |
+
25.0,
|
| 302 |
+
25.0,
|
| 303 |
+
30.0,
|
| 304 |
+
30.0
|
| 305 |
+
],
|
| 306 |
+
"require_complete_color_set": true,
|
| 307 |
+
"colors": [
|
| 308 |
+
"red",
|
| 309 |
+
"green",
|
| 310 |
+
"blue"
|
| 311 |
+
],
|
| 312 |
+
"same_noise_across_colors": true,
|
| 313 |
+
"low_t_min": 0.001,
|
| 314 |
+
"low_t_max": 0.1,
|
| 315 |
+
"rank_loss_weight": 3.0,
|
| 316 |
+
"rank_action_index": 5,
|
| 317 |
+
"rank_action_dim": 0,
|
| 318 |
+
"rank_margin": 0.03,
|
| 319 |
+
"rank_temperature": 0.05,
|
| 320 |
+
"rank_min_target_delta": 0.0001,
|
| 321 |
+
"seed": 1000
|
| 322 |
+
},
|
| 323 |
+
"rename_map": {},
|
| 324 |
+
"checkpoint_path": null
|
| 325 |
+
}
|