59 lines
1.9 KiB
JSON
59 lines
1.9 KiB
JSON
{
|
|
"action_counts": {
|
|
"BACK_STEP": 1280,
|
|
"CROUCH_FB": 1152,
|
|
"CROUCH_GUARD": 1024,
|
|
"FOR_JUMP": 1024,
|
|
"STAND_D_DF_FA": 1024,
|
|
"STAND_D_DF_FB": 1152,
|
|
"STAND_D_DF_FC": 1024,
|
|
"STAND_FA": 1024,
|
|
"STAND_FB": 1024,
|
|
"STAND_F_D_DFA": 1024,
|
|
"STAND_GUARD": 1536
|
|
},
|
|
"adapter_dir": "/mnt/d/DareFightingICE-7.0/zoning_z541_action_balanced_package/z541_artifacts/z541_adapter",
|
|
"base_config_sha256": "c99d22ec8058e2c743ce459347d74fb934083763ce2fb6cd3f03004e1173c5b6",
|
|
"base_model": "/mnt/d/DareFightingICE-7.0/Zoning_Z5_3_1",
|
|
"continuation_strategy": "action_balanced_runtime_single_layer_over_merged_z531",
|
|
"contract": "Z541_ACTION_BALANCED_TRAINING_V3",
|
|
"effective_batch_size": 64,
|
|
"eos_supervised": false,
|
|
"epochs_requested": 1.0,
|
|
"final_internal_evaluation": false,
|
|
"global_step": 192,
|
|
"gradient_accumulation_steps": 16,
|
|
"gradient_checkpointing": true,
|
|
"learning_rate": 3e-06,
|
|
"lora_alpha": 32,
|
|
"lora_dropout": 0.05,
|
|
"lora_rank": 16,
|
|
"lora_targets": [
|
|
"q_proj",
|
|
"k_proj",
|
|
"v_proj",
|
|
"o_proj",
|
|
"gate_proj",
|
|
"up_proj",
|
|
"down_proj"
|
|
],
|
|
"loss_implementation": "selected_action_position_v1",
|
|
"max_action_share": 0.125,
|
|
"micro_batch_size": 4,
|
|
"parent_model_contract": "Z531_MERGED_MODEL_CONTRACT_V1_INTENT_MAPPING_CALIBRATION",
|
|
"periodic_evaluation": false,
|
|
"prompt_bridge_suffix_sha256": "8637249c326b3533df26f5cbda89eb1a8ea1a00ee42d496b33e17a89c5b28340",
|
|
"runtime_output_format": "Output: (no trailing newline)",
|
|
"runtime_prompt_shape": "preformatted_single_layer",
|
|
"supervised_tokens_per_sample": 1,
|
|
"train_metrics": {
|
|
"epoch": 1.0,
|
|
"total_flos": 1.7405402916175872e+17,
|
|
"train_loss": 7.608128234289325,
|
|
"train_runtime": 4857.1371,
|
|
"train_samples_per_second": 2.53,
|
|
"train_steps_per_second": 0.04
|
|
},
|
|
"train_sha256": "e7f6b930569bd89c6756fc1e87cc62bff872eb28e5c31e3b7ced78353f90d98c"
|
|
}
|