bad-laya / decision_config.json
flydexo's picture
Publish evaluated stage-16 Safetensors checkpoint and model card
2fa7a93 verified
Raw History Blame Contribute Delete
1.26 kB
{
"format_version": 1,
"base_model": "answerdotai/ModernBERT-large",
"base_revision": "45bb4654a4d5aaff24dd11d4781fa46d39bf8c13",
"model": {
"encoder": "answerdotai/ModernBERT-large",
"revision": "45bb4654a4d5aaff24dd11d4781fa46d39bf8c13",
"nhead": 16,
"num_layers": 2,
"dim_feedforward": 4096,
"dropout": 0.1,
"max_len": 1024,
"head_max_len": 384,
"option_max_len": 64,
"preserve_options": true,
"balanced_state": true,
"norm_first": true,
"independent_init": true,
"train_encoder": true,
"encoder_last_n_layers": null,
"gradient_checkpointing": true,
"mixed_precision": "bf16",
"fp16_initial_scale": 1024,
"autocast_cache_enabled": false
},
"ablation": {
"objective": "sampled_reward",
"no_question_type": false,
"no_transformer": false,
"linear_scorer": false,
"reward_weights": {
"log": 1.0,
"spherical": 0.5,
"rps": 1.0
},
"name": "baseline"
},
"checkpoint_step": 60928,
"completed_stages": 15,
"partial_stage_rows": 136392,
"source_checkpoint_sha256": "e009fe101e9846279b26be6b292a33f5672894a1b319bd1210a6b71632ea5dee",
"weights_sha256": "9955e523eff4e838f294ec1c3170c422353a1eb626aaeb931e848788c5931324"
}