Instructions to use binhpham/molmoact2-leslider with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use binhpham/molmoact2-leslider with Transformers:
# Load model directly from transformers import AutoModelForImageTextToText model = AutoModelForImageTextToText.from_pretrained("binhpham/molmoact2-leslider", trust_remote_code=True, dtype="auto") - Notebooks
- Google Colab
- Kaggle
Upload folder using huggingface_hub
Browse files- config.json +150 -120
- model.safetensors +2 -2
config.json
CHANGED
|
@@ -1,123 +1,153 @@
|
|
| 1 |
{
|
| 2 |
-
|
| 3 |
-
|
| 4 |
-
"
|
| 5 |
-
|
| 6 |
-
|
| 7 |
-
|
| 8 |
-
|
| 9 |
-
|
| 10 |
-
|
| 11 |
-
|
| 12 |
-
|
| 13 |
-
|
| 14 |
-
|
| 15 |
-
|
| 16 |
-
|
| 17 |
-
|
| 18 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 19 |
},
|
| 20 |
-
"
|
| 21 |
-
|
| 22 |
-
|
| 23 |
-
|
| 24 |
-
|
| 25 |
-
|
| 26 |
-
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
|
| 30 |
-
|
| 31 |
-
"
|
| 32 |
-
"
|
| 33 |
-
"
|
| 34 |
-
"
|
| 35 |
-
"
|
| 36 |
-
"
|
| 37 |
-
"
|
| 38 |
-
|
| 39 |
-
|
| 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": null,
|
| 48 |
-
"setup_type": "single so100/so101 robotic arm in molmoact2",
|
| 49 |
-
"control_mode": "absolute joint pose",
|
| 50 |
-
"image_keys": [
|
| 51 |
-
"observation.images.arm_camera"
|
| 52 |
-
],
|
| 53 |
-
"normalize_language": true,
|
| 54 |
-
"add_setup_tokens": true,
|
| 55 |
-
"add_control_tokens": true,
|
| 56 |
-
"normalize_gripper": false,
|
| 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 |
-
"enable_inference_cuda_graph": true,
|
| 69 |
-
"per_episode_seed": false,
|
| 70 |
-
"eval_seed": null,
|
| 71 |
-
"rtc_config": null,
|
| 72 |
-
"enable_lora_vlm": false,
|
| 73 |
-
"lora_rank": 64,
|
| 74 |
-
"lora_alpha": 16,
|
| 75 |
-
"lora_dropout": 0.05,
|
| 76 |
-
"lora_bias": "none",
|
| 77 |
-
"enable_lora_action_expert": false,
|
| 78 |
-
"enable_knowledge_insulation": false,
|
| 79 |
-
"freeze_embedding": true,
|
| 80 |
-
"train_action_expert_only": false,
|
| 81 |
-
"gradient_checkpointing": true,
|
| 82 |
-
"model_dtype": "bfloat16",
|
| 83 |
-
"softmax_auxiliary_loss": true,
|
| 84 |
-
"softmax_auxiliary_loss_scale": 0.0001,
|
| 85 |
-
"discrete_loss_token_weighting": "root_subsegments_root_tokens",
|
| 86 |
-
"optimizer_lr": 1e-05,
|
| 87 |
-
"optimizer_vit_lr": 5e-06,
|
| 88 |
-
"optimizer_connector_lr": 5e-06,
|
| 89 |
-
"optimizer_action_expert_lr": 5e-05,
|
| 90 |
-
"optimizer_betas": [
|
| 91 |
-
0.9,
|
| 92 |
-
0.95
|
| 93 |
],
|
| 94 |
-
"
|
| 95 |
-
"
|
| 96 |
-
"
|
| 97 |
-
"
|
| 98 |
-
"
|
| 99 |
-
"
|
| 100 |
-
"
|
| 101 |
-
|
| 102 |
-
|
| 103 |
-
|
| 104 |
-
|
| 105 |
-
|
| 106 |
-
|
| 107 |
-
|
| 108 |
-
|
| 109 |
-
"elbow_flex.pos",
|
| 110 |
-
"wrist_flex.pos",
|
| 111 |
-
"wrist_roll.pos",
|
| 112 |
-
"gripper.pos"
|
| 113 |
-
],
|
| 114 |
-
"observation.state": [
|
| 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 |
-
}
|
| 123 |
-
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"action_end_token_id": 151933,
|
| 3 |
+
"action_expert_config": {
|
| 4 |
+
"attn_dropout": 0.0,
|
| 5 |
+
"causal_attn": false,
|
| 6 |
+
"context_layer_norm": true,
|
| 7 |
+
"dropout": 0.0,
|
| 8 |
+
"ffn_multiple_of": 256,
|
| 9 |
+
"hidden_size": 768,
|
| 10 |
+
"mlp_ratio": 4.0,
|
| 11 |
+
"model_type": "molmoact2_action_expert",
|
| 12 |
+
"num_heads": 8,
|
| 13 |
+
"num_layers": 36,
|
| 14 |
+
"qk_norm": true,
|
| 15 |
+
"qk_norm_eps": 1e-06,
|
| 16 |
+
"rope": true,
|
| 17 |
+
"timestep_embed_dim": 256
|
| 18 |
+
},
|
| 19 |
+
"action_expert_depth_gate": false,
|
| 20 |
+
"action_expert_depth_gate_init_bias": -4.0,
|
| 21 |
+
"action_expert_depth_gate_per_layer": false,
|
| 22 |
+
"action_mode": "both",
|
| 23 |
+
"max_action_horizon": 30,
|
| 24 |
+
"action_output_token_id": 151931,
|
| 25 |
+
"action_start_token_id": 151932,
|
| 26 |
+
"action_token_start_id": 151934,
|
| 27 |
+
"adapter_config": {
|
| 28 |
+
"attention_dropout": 0.0,
|
| 29 |
+
"attn_implementation": "sdpa",
|
| 30 |
+
"float32_attention": true,
|
| 31 |
+
"head_dim": 72,
|
| 32 |
+
"hidden_act": "silu",
|
| 33 |
+
"hidden_size": 1152,
|
| 34 |
+
"image_feature_dropout": 0.0,
|
| 35 |
+
"initializer_range": 0.02,
|
| 36 |
+
"intermediate_size": 9728,
|
| 37 |
+
"model_type": "molmoact2",
|
| 38 |
+
"num_attention_heads": 16,
|
| 39 |
+
"num_key_value_heads": 16,
|
| 40 |
+
"pooling_attention_mask": true,
|
| 41 |
+
"residual_dropout": 0.0,
|
| 42 |
+
"text_hidden_size": 2560,
|
| 43 |
+
"vit_layers": [
|
| 44 |
+
-3,
|
| 45 |
+
-9
|
| 46 |
+
]
|
| 47 |
+
},
|
| 48 |
+
"add_action_expert": true,
|
| 49 |
+
"add_control_tokens": true,
|
| 50 |
+
"add_setup_tokens": true,
|
| 51 |
+
"architectures": [
|
| 52 |
+
"MolmoAct2ForConditionalGeneration"
|
| 53 |
+
],
|
| 54 |
+
"auto_map": {
|
| 55 |
+
"AutoConfig": "configuration_molmoact2.MolmoAct2Config",
|
| 56 |
+
"AutoModelForImageTextToText": "modeling_molmoact2.MolmoAct2ForConditionalGeneration"
|
| 57 |
+
},
|
| 58 |
+
"depth_end_token_id": null,
|
| 59 |
+
"depth_mode": 2,
|
| 60 |
+
"depth_output_token_id": null,
|
| 61 |
+
"depth_start_token_id": null,
|
| 62 |
+
"depth_token_start_id": null,
|
| 63 |
+
"dtype": "float32",
|
| 64 |
+
"enable_depth_reasoning": false,
|
| 65 |
+
"flow_matching_beta_alpha": 1.0,
|
| 66 |
+
"flow_matching_beta_beta": 1.5,
|
| 67 |
+
"flow_matching_cutoff": 1.0,
|
| 68 |
+
"flow_matching_num_steps": 10,
|
| 69 |
+
"flow_matching_time_offset": 0.001,
|
| 70 |
+
"flow_matching_time_scale": 0.999,
|
| 71 |
+
"frame_end_token_id": 154632,
|
| 72 |
+
"frame_start_token_id": 154631,
|
| 73 |
+
"image_col_id": 154627,
|
| 74 |
+
"image_end_token_id": 154625,
|
| 75 |
+
"image_high_res_id": 154626,
|
| 76 |
+
"image_low_res_id": 154630,
|
| 77 |
+
"image_patch_id": 154626,
|
| 78 |
+
"image_start_token_id": 154624,
|
| 79 |
+
"initializer_range": 0.02,
|
| 80 |
+
"low_res_image_start_token_id": 154628,
|
| 81 |
+
"mask_action_dim_padding": true,
|
| 82 |
+
"max_action_dim": 32,
|
| 83 |
+
"model_type": "molmoact2",
|
| 84 |
+
"n_obs_steps": 1,
|
| 85 |
+
"norm_stats_filename": "norm_stats.json",
|
| 86 |
+
"num_action_tokens": 2048,
|
| 87 |
+
"num_depth_codes": 100,
|
| 88 |
+
"num_depth_tokens": 0,
|
| 89 |
+
"num_state_tokens": 256,
|
| 90 |
+
"state_end_token_id": 151674,
|
| 91 |
+
"state_format": "discrete",
|
| 92 |
+
"state_start_token_id": 151673,
|
| 93 |
+
"state_token_start_id": 151675,
|
| 94 |
+
"text_config": {
|
| 95 |
+
"additional_vocab_size": 128,
|
| 96 |
+
"attention_dropout": 0.0,
|
| 97 |
+
"attn_implementation": "sdpa",
|
| 98 |
+
"embedding_dropout": 0.0,
|
| 99 |
+
"head_dim": 128,
|
| 100 |
+
"hidden_act": "silu",
|
| 101 |
+
"hidden_size": 2560,
|
| 102 |
+
"initializer_range": 0.02,
|
| 103 |
+
"intermediate_size": 9728,
|
| 104 |
+
"layer_norm_eps": 1e-06,
|
| 105 |
+
"max_position_embeddings": 16384,
|
| 106 |
+
"model_type": "molmoact2_text",
|
| 107 |
+
"norm_after": false,
|
| 108 |
+
"num_attention_heads": 32,
|
| 109 |
+
"num_hidden_layers": 36,
|
| 110 |
+
"num_key_value_heads": 8,
|
| 111 |
+
"qk_norm_type": "qwen3",
|
| 112 |
+
"qkv_bias": false,
|
| 113 |
+
"residual_dropout": 0.0,
|
| 114 |
+
"rope_parameters": {
|
| 115 |
+
"rope_theta": 5000000.0,
|
| 116 |
+
"rope_type": "default"
|
| 117 |
},
|
| 118 |
+
"rope_scaling_layers": null,
|
| 119 |
+
"rope_theta": 5000000.0,
|
| 120 |
+
"tie_word_embeddings": false,
|
| 121 |
+
"use_cache": true,
|
| 122 |
+
"use_qk_norm": true,
|
| 123 |
+
"vocab_size": 154624
|
| 124 |
+
},
|
| 125 |
+
"tie_word_embeddings": false,
|
| 126 |
+
"transformers_version": "5.3.0",
|
| 127 |
+
"use_frame_special_tokens": true,
|
| 128 |
+
"vit_config": {
|
| 129 |
+
"attention_dropout": 0.0,
|
| 130 |
+
"attn_implementation": "sdpa",
|
| 131 |
+
"float32_attention": true,
|
| 132 |
+
"head_dim": 72,
|
| 133 |
+
"hidden_act": "gelu_pytorch_tanh",
|
| 134 |
+
"hidden_size": 1152,
|
| 135 |
+
"image_default_input_size": [
|
| 136 |
+
378,
|
| 137 |
+
378
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 138 |
],
|
| 139 |
+
"image_num_pos": 729,
|
| 140 |
+
"image_patch_size": 14,
|
| 141 |
+
"initializer_range": 0.02,
|
| 142 |
+
"intermediate_size": 4304,
|
| 143 |
+
"layer_norm_eps": 1e-06,
|
| 144 |
+
"model_type": "molmoact2",
|
| 145 |
+
"num_attention_heads": 16,
|
| 146 |
+
"num_hidden_layers": 27,
|
| 147 |
+
"num_key_value_heads": 16,
|
| 148 |
+
"residual_dropout": 0.0
|
| 149 |
+
},
|
| 150 |
+
"bos_token_id": 151645,
|
| 151 |
+
"eos_token_id": 151645,
|
| 152 |
+
"pad_token_id": 151643
|
| 153 |
+
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
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 sha256:6d34a2c69ef43bb3f346bacffff8866e8e6f6653f4cfa493aee26b75afdead19
|
| 3 |
+
size 10884565968
|