{ "alora_invocation_tokens": null, "alpha_pattern": {}, "arrow_config": null, "auto_mapping": { "base_model_class": "PI05Policy", "parent_library": "lerobot.policies.pi05.modeling_pi05" }, "base_model_name_or_path": "/home/shadeform/ckpts/warm_pg_v2_patched", "bias": "none", "corda_config": null, "ensure_weight_tying": false, "eva_config": null, "exclude_modules": null, "fan_in_fan_out": false, "inference_mode": true, "init_lora_weights": true, "layer_replication": null, "layers_pattern": null, "layers_to_transform": null, "loftq_config": {}, "lora_alpha": 64, "lora_bias": false, "lora_dropout": 0.05, "lora_ga_config": null, "megatron_config": null, "megatron_core": "megatron.core", "modules_to_save": null, "peft_type": "LORA", "peft_version": "0.19.1", "qalora_group_size": 16, "r": 16, "rank_pattern": { ".*language_model\\.layers\\.0\\..*\\.down_proj$": 16, ".*language_model\\.layers\\.0\\..*\\.gate_proj$": 16, ".*language_model\\.layers\\.0\\..*\\.k_proj$": 16, ".*language_model\\.layers\\.0\\..*\\.o_proj$": 16, ".*language_model\\.layers\\.0\\..*\\.q_proj$": 16, ".*language_model\\.layers\\.0\\..*\\.up_proj$": 16, ".*language_model\\.layers\\.0\\..*\\.v_proj$": 16, ".*language_model\\.layers\\.10\\..*\\.down_proj$": 64, ".*language_model\\.layers\\.10\\..*\\.gate_proj$": 64, ".*language_model\\.layers\\.10\\..*\\.k_proj$": 64, ".*language_model\\.layers\\.10\\..*\\.o_proj$": 64, ".*language_model\\.layers\\.10\\..*\\.q_proj$": 64, ".*language_model\\.layers\\.10\\..*\\.up_proj$": 64, ".*language_model\\.layers\\.10\\..*\\.v_proj$": 64, ".*language_model\\.layers\\.11\\..*\\.down_proj$": 64, ".*language_model\\.layers\\.11\\..*\\.gate_proj$": 64, ".*language_model\\.layers\\.11\\..*\\.k_proj$": 64, ".*language_model\\.layers\\.11\\..*\\.o_proj$": 64, ".*language_model\\.layers\\.11\\..*\\.q_proj$": 64, ".*language_model\\.layers\\.11\\..*\\.up_proj$": 64, ".*language_model\\.layers\\.11\\..*\\.v_proj$": 64, ".*language_model\\.layers\\.12\\..*\\.down_proj$": 64, ".*language_model\\.layers\\.12\\..*\\.gate_proj$": 64, ".*language_model\\.layers\\.12\\..*\\.k_proj$": 64, ".*language_model\\.layers\\.12\\..*\\.o_proj$": 64, ".*language_model\\.layers\\.12\\..*\\.q_proj$": 64, ".*language_model\\.layers\\.12\\..*\\.up_proj$": 64, ".*language_model\\.layers\\.12\\..*\\.v_proj$": 64, ".*language_model\\.layers\\.13\\..*\\.down_proj$": 32, ".*language_model\\.layers\\.13\\..*\\.gate_proj$": 32, ".*language_model\\.layers\\.13\\..*\\.k_proj$": 32, ".*language_model\\.layers\\.13\\..*\\.o_proj$": 32, ".*language_model\\.layers\\.13\\..*\\.q_proj$": 32, ".*language_model\\.layers\\.13\\..*\\.up_proj$": 32, ".*language_model\\.layers\\.13\\..*\\.v_proj$": 32, ".*language_model\\.layers\\.14\\..*\\.down_proj$": 32, ".*language_model\\.layers\\.14\\..*\\.gate_proj$": 32, ".*language_model\\.layers\\.14\\..*\\.k_proj$": 32, ".*language_model\\.layers\\.14\\..*\\.o_proj$": 32, ".*language_model\\.layers\\.14\\..*\\.q_proj$": 32, ".*language_model\\.layers\\.14\\..*\\.up_proj$": 32, ".*language_model\\.layers\\.14\\..*\\.v_proj$": 32, ".*language_model\\.layers\\.15\\..*\\.down_proj$": 48, ".*language_model\\.layers\\.15\\..*\\.gate_proj$": 48, ".*language_model\\.layers\\.15\\..*\\.k_proj$": 48, ".*language_model\\.layers\\.15\\..*\\.o_proj$": 48, ".*language_model\\.layers\\.15\\..*\\.q_proj$": 48, ".*language_model\\.layers\\.15\\..*\\.up_proj$": 48, ".*language_model\\.layers\\.15\\..*\\.v_proj$": 48, ".*language_model\\.layers\\.16\\..*\\.down_proj$": 48, ".*language_model\\.layers\\.16\\..*\\.gate_proj$": 48, ".*language_model\\.layers\\.16\\..*\\.k_proj$": 48, ".*language_model\\.layers\\.16\\..*\\.o_proj$": 48, ".*language_model\\.layers\\.16\\..*\\.q_proj$": 48, ".*language_model\\.layers\\.16\\..*\\.up_proj$": 48, ".*language_model\\.layers\\.16\\..*\\.v_proj$": 48, ".*language_model\\.layers\\.17\\..*\\.down_proj$": 48, ".*language_model\\.layers\\.17\\..*\\.gate_proj$": 48, ".*language_model\\.layers\\.17\\..*\\.k_proj$": 48, ".*language_model\\.layers\\.17\\..*\\.o_proj$": 48, ".*language_model\\.layers\\.17\\..*\\.q_proj$": 48, ".*language_model\\.layers\\.17\\..*\\.up_proj$": 48, ".*language_model\\.layers\\.17\\..*\\.v_proj$": 48, ".*language_model\\.layers\\.1\\..*\\.down_proj$": 16, ".*language_model\\.layers\\.1\\..*\\.gate_proj$": 16, ".*language_model\\.layers\\.1\\..*\\.k_proj$": 16, ".*language_model\\.layers\\.1\\..*\\.o_proj$": 16, ".*language_model\\.layers\\.1\\..*\\.q_proj$": 16, ".*language_model\\.layers\\.1\\..*\\.up_proj$": 16, ".*language_model\\.layers\\.1\\..*\\.v_proj$": 16, ".*language_model\\.layers\\.2\\..*\\.down_proj$": 16, ".*language_model\\.layers\\.2\\..*\\.gate_proj$": 16, ".*language_model\\.layers\\.2\\..*\\.k_proj$": 16, ".*language_model\\.layers\\.2\\..*\\.o_proj$": 16, ".*language_model\\.layers\\.2\\..*\\.q_proj$": 16, ".*language_model\\.layers\\.2\\..*\\.up_proj$": 16, ".*language_model\\.layers\\.2\\..*\\.v_proj$": 16, ".*language_model\\.layers\\.3\\..*\\.down_proj$": 16, ".*language_model\\.layers\\.3\\..*\\.gate_proj$": 16, ".*language_model\\.layers\\.3\\..*\\.k_proj$": 16, ".*language_model\\.layers\\.3\\..*\\.o_proj$": 16, ".*language_model\\.layers\\.3\\..*\\.q_proj$": 16, ".*language_model\\.layers\\.3\\..*\\.up_proj$": 16, ".*language_model\\.layers\\.3\\..*\\.v_proj$": 16, ".*language_model\\.layers\\.4\\..*\\.down_proj$": 16, ".*language_model\\.layers\\.4\\..*\\.gate_proj$": 16, ".*language_model\\.layers\\.4\\..*\\.k_proj$": 16, ".*language_model\\.layers\\.4\\..*\\.o_proj$": 16, ".*language_model\\.layers\\.4\\..*\\.q_proj$": 16, ".*language_model\\.layers\\.4\\..*\\.up_proj$": 16, ".*language_model\\.layers\\.4\\..*\\.v_proj$": 16, ".*language_model\\.layers\\.5\\..*\\.down_proj$": 32, ".*language_model\\.layers\\.5\\..*\\.gate_proj$": 32, ".*language_model\\.layers\\.5\\..*\\.k_proj$": 32, ".*language_model\\.layers\\.5\\..*\\.o_proj$": 32, ".*language_model\\.layers\\.5\\..*\\.q_proj$": 32, ".*language_model\\.layers\\.5\\..*\\.up_proj$": 32, ".*language_model\\.layers\\.5\\..*\\.v_proj$": 32, ".*language_model\\.layers\\.6\\..*\\.down_proj$": 32, ".*language_model\\.layers\\.6\\..*\\.gate_proj$": 32, ".*language_model\\.layers\\.6\\..*\\.k_proj$": 32, ".*language_model\\.layers\\.6\\..*\\.o_proj$": 32, ".*language_model\\.layers\\.6\\..*\\.q_proj$": 32, ".*language_model\\.layers\\.6\\..*\\.up_proj$": 32, ".*language_model\\.layers\\.6\\..*\\.v_proj$": 32, ".*language_model\\.layers\\.7\\..*\\.down_proj$": 32, ".*language_model\\.layers\\.7\\..*\\.gate_proj$": 32, ".*language_model\\.layers\\.7\\..*\\.k_proj$": 32, ".*language_model\\.layers\\.7\\..*\\.o_proj$": 32, ".*language_model\\.layers\\.7\\..*\\.q_proj$": 32, ".*language_model\\.layers\\.7\\..*\\.up_proj$": 32, ".*language_model\\.layers\\.7\\..*\\.v_proj$": 32, ".*language_model\\.layers\\.8\\..*\\.down_proj$": 64, ".*language_model\\.layers\\.8\\..*\\.gate_proj$": 64, ".*language_model\\.layers\\.8\\..*\\.k_proj$": 64, ".*language_model\\.layers\\.8\\..*\\.o_proj$": 64, ".*language_model\\.layers\\.8\\..*\\.q_proj$": 64, ".*language_model\\.layers\\.8\\..*\\.up_proj$": 64, ".*language_model\\.layers\\.8\\..*\\.v_proj$": 64, ".*language_model\\.layers\\.9\\..*\\.down_proj$": 64, ".*language_model\\.layers\\.9\\..*\\.gate_proj$": 64, ".*language_model\\.layers\\.9\\..*\\.k_proj$": 64, ".*language_model\\.layers\\.9\\..*\\.o_proj$": 64, ".*language_model\\.layers\\.9\\..*\\.q_proj$": 64, ".*language_model\\.layers\\.9\\..*\\.up_proj$": 64, ".*language_model\\.layers\\.9\\..*\\.v_proj$": 64 }, "revision": null, "target_modules": [ "v_proj", "up_proj", "down_proj", "o_proj", "q_proj", "k_proj", "gate_proj" ], "target_parameters": null, "task_type": null, "trainable_token_indices": null, "use_bdlora": null, "use_dora": false, "use_qalora": false, "use_rslora": false }