{ "config": { "run_name": "dec_v4", "data": "data/lc_mix", "cache": "data/lc_mix/dec_v14", "seed": 0, "model": { "d": 384, "layers": 6, "heads": 6, "ff": 1536, "K": 8, "dropout": 0.05 }, "train": { "steps": 130000, "batch": 512, "lr": 0.0002, "wd": 0.01, "warmup": 2000, "min_lr_frac": 0.05, "schedule": "wsd", "decay_frac": 0.2, "label_smoothing": 0.02, "grad_clip": 1.0, "p_noise": 0.4, "max_angle_cos": 0.99, "logn_jitter": 0.02, "p_dz": 0.3, "rewarm": 500, "log_every": 100, "eval_every": 2500, "eval_n": 1024, "eval_ce_n": 4096, "len_buckets": 16, "max_tokens": 32768 }, "encoder": "runs/v14/best.pt" }, "step": 130000, "d_lat": 256, "best": 0.791015625, "component": "decoder_lc", "description": "decoder fine-tuned on Lample-Charton-type expressions (ODE benchmark only)", "source_checkpoint": "runs/dec_v4/best.pt", "domain": [ -3.141592653589793, 3.141592653589793 ], "format": "func-jepa-1" }