Files
zagreus-0.4B-ita-elicit/train_args.json

37 lines
864 B
JSON
Raw Normal View History

{
"config": "config/baseline.yaml",
"teacher": "results/teacher_model_lang_pool_v2.jsonl",
"model": "model",
"out": "runs/elicit_A2_cleanteacher",
"epochs": 6,
"lr": 0.0001,
"weight_decay": 0.0,
"warmup_ratio": 0.03,
"batch_size": 16,
"grad_accum": 2,
"max_len": 1024,
"temp": 1.0,
"lambda_ce": 1.0,
"kl_weight": 0.3,
"ce_target": "gold",
"mu_format": 0.1,
"gate_gamma": 0,
"permute": true,
"perm_k": 1,
"lambda_cons": 0.0,
"cons_conf_gamma": 0.0,
"shots": 0,
"shot_file": "/workspace/mii-zagreus-post-training-challenge/data/raw/train_shots.jsonl",
"dtype": "bfloat16",
"grad_ckpt": false,
"device": "cuda",
"limit": null,
"log_every": 10,
"save_every": 300,
"keep_last": 6,
"eval_every": 50,
"eval_data": "data/processed/lang_dev_v2.jsonl",
"eval_n": 600,
"eval_shots": null,
"smoke": false
}