{ "name": "addition_sharp position coupling seed 2", "task": { "op": "add", "max_digits": 20, "min_digits": 1, "carry_heavy": 0.0, "reverse": true, "zero_pad": true, "positions": "coupled", "blanks": 0, "blanks_fixed": true, "offset_max": 100, "test_digits": [ 5, 10, 15, 20, 25, 30, 40, 50, 60, 80, 100 ], "test_examples": 256, "kind": "arithmetic" }, "model": { "arch": "transformer", "d_model": 256, "n_layers": 4, "n_heads": 4, "d_mlp": 1024, "max_positions": 1024, "pos_embed": "absolute", "attn_scale": 2.0, "dropout": 0.0 }, "train": { "steps": 50000, "batch_size": 256, "lr": 0.0003, "weight_decay": 0.1, "betas": [ 0.9, 0.98 ], "warmup_steps": 1000, "cosine": true, "amp": true, "eval_every": 2500, "checkpoint_every": 10000, "seed": 2, "wandb": false } }