Buckets:

Pranav2748's picture
download
raw
41.9 kB
`torch_dtype` is deprecated! Use `dtype` instead!
{"event": "start", "train_count": 40000, "dev_count": 1000, "total_steps": 1250, "n_params": 601298945, "n_new": 5249025, "dropped_overlong": {}}
{"event": "dev", "step": 0, "seconds": 13.6, "gate": 0.0, "K1": {"ce": 0.29224, "kl": 0.26023, "acc": 0.90947, "n": 1000}, "K4": {"ce": 0.29224, "kl": 0.26023, "acc": 0.90947, "n": 1000}}
W1106 14:08:43.326000 20887 torch/fx/experimental/symbolic_shapes.py:6823] [0/1] _maybe_guard_rel() was called on non-relation expression Eq(s6, s92) | Eq(s92, 1)
W1106 14:08:46.853000 20887 torch/fx/experimental/symbolic_shapes.py:6823] [0/1_1] _maybe_guard_rel() was called on non-relation expression Eq(s6, s92) | Eq(s92, 1)
{"step": 10, "total": 1250, "K": 8, "lr_bb": 1.5873015873015873e-06, "grad_norm": 10.6974, "loss": 0.07389, "ce": 0.07389, "kl": 0.05306, "rps": 2e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 22.6, "tok_per_s": 3619, "pass_tok_per_s": 16459, "peak_gib": 74.76, "epoch": 1, "gate": 2.7e-05, "W_dev": 0.0025, "head_o": 0.0252}
{"step": 20, "total": 1250, "K": 1, "lr_bb": 3.1746031746031746e-06, "grad_norm": 15.7561, "loss": 0.2157, "ce": 0.21566, "kl": 0.21566, "rps": 0.00017, "l1": 0.0, "dkl": 0.0, "elapsed": 35.0, "tok_per_s": 4729, "pass_tok_per_s": 18961, "peak_gib": 74.76, "epoch": 1, "gate": 3.2e-05, "W_dev": 0.0189, "head_o": 0.0704}
{"step": 30, "total": 1250, "K": 1, "lr_bb": 4.761904761904762e-06, "grad_norm": 17.3844, "loss": 0.27658, "ce": 0.27658, "kl": 0.23374, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 48.0, "tok_per_s": 5222, "pass_tok_per_s": 20404, "peak_gib": 74.76, "epoch": 1, "gate": -2.4e-05, "W_dev": 0.0501, "head_o": 0.1107}
{"step": 40, "total": 1250, "K": 8, "lr_bb": 6.349206349206349e-06, "grad_norm": 7.9038, "loss": 0.13691, "ce": 0.13514, "kl": 0.13514, "rps": 0.00709, "l1": 0.0, "dkl": 0.0, "elapsed": 57.9, "tok_per_s": 5897, "pass_tok_per_s": 20872, "peak_gib": 74.76, "epoch": 1, "gate": -5e-05, "W_dev": 0.0706, "head_o": 0.1446}
{"step": 50, "total": 1250, "K": 4, "lr_bb": 7.936507936507936e-06, "grad_norm": 13.4167, "loss": 0.35567, "ce": 0.35567, "kl": 0.35567, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 69.8, "tok_per_s": 6075, "pass_tok_per_s": 21329, "peak_gib": 74.76, "epoch": 1, "gate": -2.4e-05, "W_dev": 0.1156, "head_o": 0.1651}
{"step": 60, "total": 1250, "K": 8, "lr_bb": 9.523809523809525e-06, "grad_norm": 8.0568, "loss": 0.22276, "ce": 0.22276, "kl": 0.19801, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 87.1, "tok_per_s": 5893, "pass_tok_per_s": 21777, "peak_gib": 74.76, "epoch": 1, "gate": -5.5e-05, "W_dev": 0.1635, "head_o": 0.2296}
{"step": 70, "total": 1250, "K": 1, "lr_bb": 9.999432619768909e-06, "grad_norm": 20.4893, "loss": 0.36669, "ce": 0.36567, "kl": 0.36567, "rps": 0.00406, "l1": 0.0, "dkl": 0.0, "elapsed": 99.3, "tok_per_s": 6038, "pass_tok_per_s": 21944, "peak_gib": 74.76, "epoch": 1, "gate": -8.9e-05, "W_dev": 0.2055, "head_o": 0.2359}
{"step": 80, "total": 1250, "K": 4, "lr_bb": 9.995965814255133e-06, "grad_norm": 10.5719, "loss": 0.26455, "ce": 0.26365, "kl": 0.20152, "rps": 0.00357, "l1": 0.0, "dkl": 0.0, "elapsed": 108.1, "tok_per_s": 6331, "pass_tok_per_s": 22159, "peak_gib": 74.76, "epoch": 1, "gate": -0.000128, "W_dev": 0.2509, "head_o": 0.24}
{"step": 90, "total": 1250, "K": 8, "lr_bb": 9.989349839823577e-06, "grad_norm": 32.4216, "loss": 0.52827, "ce": 0.52459, "kl": 0.41247, "rps": 0.01468, "l1": 0.0, "dkl": 0.0, "elapsed": 123.7, "tok_per_s": 6218, "pass_tok_per_s": 22351, "peak_gib": 74.76, "epoch": 1, "gate": -0.000118, "W_dev": 0.3397, "head_o": 0.2749}
{"step": 100, "total": 1250, "K": 1, "lr_bb": 9.979589330588889e-06, "grad_norm": 7.8944, "loss": 0.30216, "ce": 0.30186, "kl": 0.26117, "rps": 0.0012, "l1": 0.0, "dkl": 0.0, "elapsed": 136.2, "tok_per_s": 6317, "pass_tok_per_s": 22667, "peak_gib": 74.76, "epoch": 1, "gate": -0.000125, "W_dev": 0.4175, "head_o": 0.3226}
{"step": 110, "total": 1250, "K": 1, "lr_bb": 9.966691123233898e-06, "grad_norm": 15.2721, "loss": 0.32031, "ce": 0.31834, "kl": 0.31834, "rps": 0.00786, "l1": 0.0, "dkl": 0.0, "elapsed": 149.4, "tok_per_s": 6321, "pass_tok_per_s": 22751, "peak_gib": 74.76, "epoch": 1, "gate": -1.8e-05, "W_dev": 0.4827, "head_o": 0.3988}
{"step": 120, "total": 1250, "K": 2, "lr_bb": 9.95066425222092e-06, "grad_norm": 14.0573, "loss": 0.32298, "ce": 0.32103, "kl": 0.30114, "rps": 0.0078, "l1": 0.0, "dkl": 0.0, "elapsed": 159.1, "tok_per_s": 6468, "pass_tok_per_s": 22699, "peak_gib": 74.76, "epoch": 1, "gate": 7.3e-05, "W_dev": 0.5091, "head_o": 0.3783}
{"step": 130, "total": 1250, "K": 2, "lr_bb": 9.931519943463613e-06, "grad_norm": 30.7538, "loss": 0.46428, "ce": 0.46387, "kl": 0.40657, "rps": 0.00163, "l1": 0.0, "dkl": 0.0, "elapsed": 170.9, "tok_per_s": 6513, "pass_tok_per_s": 22814, "peak_gib": 74.76, "epoch": 1, "gate": 0.000136, "W_dev": 0.5286, "head_o": 0.352}
{"step": 140, "total": 1250, "K": 4, "lr_bb": 9.909271606463866e-06, "grad_norm": 14.6418, "loss": 0.49408, "ce": 0.49404, "kl": 0.45914, "rps": 0.00016, "l1": 0.0, "dkl": 0.0, "elapsed": 186.1, "tok_per_s": 6442, "pass_tok_per_s": 22879, "peak_gib": 74.76, "epoch": 1, "gate": 0.00015, "W_dev": 0.5578, "head_o": 0.3366}
{"step": 150, "total": 1250, "K": 4, "lr_bb": 9.883934824919209e-06, "grad_norm": 34.9038, "loss": 0.69086, "ce": 0.68944, "kl": 0.68944, "rps": 0.00568, "l1": 0.0, "dkl": 0.0, "elapsed": 196.5, "tok_per_s": 6548, "pass_tok_per_s": 22964, "peak_gib": 74.76, "epoch": 1, "gate": -2.6e-05, "W_dev": 0.5983, "head_o": 0.3367}
{"step": 160, "total": 1250, "K": 2, "lr_bb": 9.855527345807316e-06, "grad_norm": 8.9001, "loss": 0.29691, "ce": 0.29691, "kl": 0.2246, "rps": 1e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 208.6, "tok_per_s": 6568, "pass_tok_per_s": 22939, "peak_gib": 74.76, "epoch": 1, "gate": -0.000256, "W_dev": 0.6305, "head_o": 0.321}
{"step": 170, "total": 1250, "K": 1, "lr_bb": 9.824069066955246e-06, "grad_norm": 25.5244, "loss": 0.77666, "ce": 0.77539, "kl": 0.75502, "rps": 0.00509, "l1": 0.0, "dkl": 0.0, "elapsed": 230.5, "tok_per_s": 6313, "pass_tok_per_s": 23046, "peak_gib": 74.76, "epoch": 1, "gate": -0.000526, "W_dev": 0.7042, "head_o": 0.3214}
{"step": 180, "total": 1250, "K": 4, "lr_bb": 9.789582023102157e-06, "grad_norm": 13.7719, "loss": 0.36283, "ce": 0.36283, "kl": 0.36283, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 237.0, "tok_per_s": 6512, "pass_tok_per_s": 23069, "peak_gib": 74.76, "epoch": 1, "gate": -0.000692, "W_dev": 0.7768, "head_o": 0.3194}
{"step": 190, "total": 1250, "K": 4, "lr_bb": 9.75209037046521e-06, "grad_norm": 7.338, "loss": 0.28874, "ce": 0.28528, "kl": 0.28528, "rps": 0.01383, "l1": 0.0, "dkl": 0.0, "elapsed": 257.8, "tok_per_s": 6323, "pass_tok_per_s": 23096, "peak_gib": 74.76, "epoch": 1, "gate": -0.000606, "W_dev": 0.9344, "head_o": 0.3372}
{"step": 200, "total": 1250, "K": 8, "lr_bb": 9.711620369819521e-06, "grad_norm": 8.8276, "loss": 0.41929, "ce": 0.41671, "kl": 0.38523, "rps": 0.0103, "l1": 0.0, "dkl": 0.0, "elapsed": 273.6, "tok_per_s": 6288, "pass_tok_per_s": 23171, "peak_gib": 74.76, "epoch": 1, "gate": -0.00046, "W_dev": 1.0411, "head_o": 0.3067}
{"step": 210, "total": 1250, "K": 2, "lr_bb": 9.668200368103964e-06, "grad_norm": 7.9108, "loss": 0.31251, "ce": 0.31251, "kl": 0.29443, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 287.6, "tok_per_s": 6280, "pass_tok_per_s": 23227, "peak_gib": 74.76, "epoch": 1, "gate": -0.000457, "W_dev": 1.1362, "head_o": 0.3441}
{"step": 220, "total": 1250, "K": 2, "lr_bb": 9.621860778565765e-06, "grad_norm": 9.4802, "loss": 0.15547, "ce": 0.15547, "kl": 0.13668, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 296.3, "tok_per_s": 6397, "pass_tok_per_s": 23241, "peak_gib": 74.76, "epoch": 1, "gate": -0.000214, "W_dev": 1.1893, "head_o": 0.3957}
{"step": 230, "total": 1250, "K": 8, "lr_bb": 9.572634059457735e-06, "grad_norm": 17.8401, "loss": 0.33329, "ce": 0.33328, "kl": 0.29696, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 310.9, "tok_per_s": 6366, "pass_tok_per_s": 23246, "peak_gib": 74.76, "epoch": 1, "gate": -7.9e-05, "W_dev": 1.2183, "head_o": 0.3919}
{"step": 240, "total": 1250, "K": 8, "lr_bb": 9.520554691303113e-06, "grad_norm": 19.7966, "loss": 0.3367, "ce": 0.3367, "kl": 0.30258, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 329.5, "tok_per_s": 6265, "pass_tok_per_s": 23311, "peak_gib": 74.76, "epoch": 1, "gate": 0.000118, "W_dev": 1.2349, "head_o": 0.3844}
{"step": 250, "total": 1250, "K": 1, "lr_bb": 9.465659152743914e-06, "grad_norm": 8.7278, "loss": 0.12269, "ce": 0.122, "kl": 0.07044, "rps": 0.00275, "l1": 0.0, "dkl": 0.0, "elapsed": 342.6, "tok_per_s": 6272, "pass_tok_per_s": 23324, "peak_gib": 74.76, "epoch": 1, "gate": 0.000186, "W_dev": 1.241, "head_o": 0.3578}
{"event": "dev", "step": 250, "seconds": 9.7, "gate": 0.00018599162285681814, "K1": {"ce": 0.32647, "kl": 0.29446, "acc": 0.89895, "n": 1000}, "K4": {"ce": 0.32681, "kl": 0.29479, "acc": 0.89789, "n": 1000}}
{"step": 260, "total": 1250, "K": 1, "lr_bb": 9.407985894989713e-06, "grad_norm": 12.5827, "loss": 0.22487, "ce": 0.22216, "kl": 0.22216, "rps": 0.01086, "l1": 0.0, "dkl": 0.0, "elapsed": 364.9, "tok_per_s": 6129, "pass_tok_per_s": 22738, "peak_gib": 74.76, "epoch": 1, "gate": 0.000323, "W_dev": 1.2467, "head_o": 0.3603}
{"step": 270, "total": 1250, "K": 8, "lr_bb": 9.347575314884746e-06, "grad_norm": 22.3845, "loss": 0.51146, "ce": 0.51142, "kl": 0.51142, "rps": 0.0002, "l1": 0.0, "dkl": 0.0, "elapsed": 378.1, "tok_per_s": 6150, "pass_tok_per_s": 22755, "peak_gib": 74.76, "epoch": 1, "gate": 0.000269, "W_dev": 1.2523, "head_o": 0.3503}
{"step": 280, "total": 1250, "K": 8, "lr_bb": 9.284469726612224e-06, "grad_norm": 27.0708, "loss": 0.23949, "ce": 0.23949, "kl": 0.2152, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 386.7, "tok_per_s": 6234, "pass_tok_per_s": 22750, "peak_gib": 74.76, "epoch": 1, "gate": 0.000165, "W_dev": 1.2605, "head_o": 0.3746}
{"step": 290, "total": 1250, "K": 8, "lr_bb": 9.218713332055646e-06, "grad_norm": 7.9591, "loss": 0.33058, "ce": 0.33056, "kl": 0.30891, "rps": 9e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 396.0, "tok_per_s": 6304, "pass_tok_per_s": 22805, "peak_gib": 74.76, "epoch": 1, "gate": 0.000108, "W_dev": 1.2679, "head_o": 0.3642}
{"step": 300, "total": 1250, "K": 2, "lr_bb": 9.150352189837881e-06, "grad_norm": 14.2781, "loss": 0.32664, "ce": 0.32432, "kl": 0.27182, "rps": 0.00924, "l1": 0.0, "dkl": 0.0, "elapsed": 413.8, "tok_per_s": 6241, "pass_tok_per_s": 22892, "peak_gib": 74.76, "epoch": 1, "gate": -8.2e-05, "W_dev": 1.2768, "head_o": 0.3405}
{"step": 310, "total": 1250, "K": 1, "lr_bb": 9.079434183059728e-06, "grad_norm": 14.9798, "loss": 0.43875, "ce": 0.43062, "kl": 0.3255, "rps": 0.03249, "l1": 0.0, "dkl": 0.0, "elapsed": 428.5, "tok_per_s": 6230, "pass_tok_per_s": 22960, "peak_gib": 74.76, "epoch": 1, "gate": -0.000152, "W_dev": 1.2801, "head_o": 0.3385}
{"step": 320, "total": 1250, "K": 8, "lr_bb": 9.006008985760507e-06, "grad_norm": 13.2177, "loss": 0.38885, "ce": 0.38119, "kl": 0.31723, "rps": 0.03066, "l1": 0.0, "dkl": 0.0, "elapsed": 444.4, "tok_per_s": 6215, "pass_tok_per_s": 22983, "peak_gib": 74.76, "epoch": 1, "gate": -0.000115, "W_dev": 1.282, "head_o": 0.3325}
{"step": 330, "total": 1250, "K": 1, "lr_bb": 8.930128028124227e-06, "grad_norm": 9.6545, "loss": 0.282, "ce": 0.28196, "kl": 0.20336, "rps": 0.00016, "l1": 0.0, "dkl": 0.0, "elapsed": 453.3, "tok_per_s": 6277, "pass_tok_per_s": 23003, "peak_gib": 74.76, "epoch": 1, "gate": 1.3e-05, "W_dev": 1.2816, "head_o": 0.3435}
{"step": 340, "total": 1250, "K": 4, "lr_bb": 8.851844460455654e-06, "grad_norm": 12.1279, "loss": 0.31602, "ce": 0.3159, "kl": 0.29731, "rps": 0.00047, "l1": 0.0, "dkl": 0.0, "elapsed": 464.7, "tok_per_s": 6309, "pass_tok_per_s": 23018, "peak_gib": 74.76, "epoch": 1, "gate": 4.9e-05, "W_dev": 1.2817, "head_o": 0.3423}
{"step": 350, "total": 1250, "K": 1, "lr_bb": 8.771213115951537e-06, "grad_norm": 3.6032, "loss": 0.06338, "ce": 0.06338, "kl": 0.03061, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 480.8, "tok_per_s": 6292, "pass_tok_per_s": 22825, "peak_gib": 74.76, "epoch": 1, "gate": 1.4e-05, "W_dev": 1.2815, "head_o": 0.352}
{"step": 360, "total": 1250, "K": 1, "lr_bb": 8.688290472293068e-06, "grad_norm": 8.0404, "loss": 0.25383, "ce": 0.25044, "kl": 0.15269, "rps": 0.01356, "l1": 0.0, "dkl": 0.0, "elapsed": 503.5, "tok_per_s": 6175, "pass_tok_per_s": 22432, "peak_gib": 74.76, "epoch": 1, "gate": -0.000106, "W_dev": 1.281, "head_o": 0.3652}
{"step": 370, "total": 1250, "K": 8, "lr_bb": 8.603134612086474e-06, "grad_norm": 10.0308, "loss": 0.23776, "ce": 0.23775, "kl": 0.21609, "rps": 3e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 517.3, "tok_per_s": 6172, "pass_tok_per_s": 22425, "peak_gib": 74.76, "epoch": 1, "gate": -0.000203, "W_dev": 1.2814, "head_o": 0.3542}
{"step": 380, "total": 1250, "K": 8, "lr_bb": 8.515805182179437e-06, "grad_norm": 14.7454, "loss": 0.15377, "ce": 0.1535, "kl": 0.11985, "rps": 0.00106, "l1": 0.0, "dkl": 0.0, "elapsed": 533.2, "tok_per_s": 6147, "pass_tok_per_s": 22453, "peak_gib": 74.76, "epoch": 1, "gate": -0.000127, "W_dev": 1.2842, "head_o": 0.3388}
{"step": 390, "total": 1250, "K": 1, "lr_bb": 8.426363351881879e-06, "grad_norm": 9.0879, "loss": 0.18602, "ce": 0.18601, "kl": 0.16196, "rps": 5e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 552.5, "tok_per_s": 6086, "pass_tok_per_s": 22477, "peak_gib": 74.76, "epoch": 1, "gate": -4.3e-05, "W_dev": 1.2866, "head_o": 0.3146}
{"step": 400, "total": 1250, "K": 1, "lr_bb": 8.33487177012032e-06, "grad_norm": 9.1608, "loss": 0.20347, "ce": 0.20247, "kl": 0.14774, "rps": 0.00397, "l1": 0.0, "dkl": 0.0, "elapsed": 571.0, "tok_per_s": 6039, "pass_tok_per_s": 22522, "peak_gib": 74.76, "epoch": 1, "gate": 8e-06, "W_dev": 1.288, "head_o": 0.3129}
{"step": 410, "total": 1250, "K": 8, "lr_bb": 8.241394521555854e-06, "grad_norm": 12.1419, "loss": 0.20581, "ce": 0.20581, "kl": 0.18003, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 584.2, "tok_per_s": 6055, "pass_tok_per_s": 22559, "peak_gib": 74.76, "epoch": 1, "gate": 0.000166, "W_dev": 1.2893, "head_o": 0.3417}
{"step": 420, "total": 1250, "K": 2, "lr_bb": 8.145997081696497e-06, "grad_norm": 14.9311, "loss": 0.2866, "ce": 0.2866, "kl": 0.2866, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 593.2, "tok_per_s": 6101, "pass_tok_per_s": 22561, "peak_gib": 74.76, "epoch": 1, "gate": 0.000281, "W_dev": 1.2916, "head_o": 0.3289}
{"step": 430, "total": 1250, "K": 4, "lr_bb": 8.048746271035295e-06, "grad_norm": 2.8701, "loss": 0.09631, "ce": 0.0963, "kl": 0.06111, "rps": 5e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 604.0, "tok_per_s": 6130, "pass_tok_per_s": 22581, "peak_gib": 74.76, "epoch": 1, "gate": 0.000414, "W_dev": 1.2917, "head_o": 0.3252}
{"step": 440, "total": 1250, "K": 1, "lr_bb": 7.949710208246379e-06, "grad_norm": 9.8138, "loss": 0.23018, "ce": 0.23018, "kl": 0.13236, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 616.5, "tok_per_s": 6147, "pass_tok_per_s": 22584, "peak_gib": 74.76, "epoch": 1, "gate": 0.000552, "W_dev": 1.3035, "head_o": 0.3508}
{"step": 450, "total": 1250, "K": 8, "lr_bb": 7.848958262471667e-06, "grad_norm": 35.0812, "loss": 0.56207, "ce": 0.56206, "kl": 0.56206, "rps": 4e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 630.4, "tok_per_s": 6148, "pass_tok_per_s": 22612, "peak_gib": 74.76, "epoch": 1, "gate": 0.000603, "W_dev": 1.3373, "head_o": 0.3523}
{"step": 460, "total": 1250, "K": 4, "lr_bb": 7.74656100473175e-06, "grad_norm": 11.6279, "loss": 0.45063, "ce": 0.44443, "kl": 0.43966, "rps": 0.02482, "l1": 0.0, "dkl": 0.0, "elapsed": 647.9, "tok_per_s": 6124, "pass_tok_per_s": 22675, "peak_gib": 74.76, "epoch": 1, "gate": 0.000531, "W_dev": 1.3863, "head_o": 0.336}
{"step": 470, "total": 1250, "K": 4, "lr_bb": 7.642590158494882e-06, "grad_norm": 7.5349, "loss": 0.13984, "ce": 0.13984, "kl": 0.13984, "rps": 1e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 659.6, "tok_per_s": 6138, "pass_tok_per_s": 22676, "peak_gib": 74.76, "epoch": 1, "gate": 0.000548, "W_dev": 1.4198, "head_o": 0.3338}
{"step": 480, "total": 1250, "K": 4, "lr_bb": 7.537118549438758e-06, "grad_norm": 24.8411, "loss": 0.23612, "ce": 0.23612, "kl": 0.17662, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 673.4, "tok_per_s": 6146, "pass_tok_per_s": 22711, "peak_gib": 74.76, "epoch": 1, "gate": 0.000642, "W_dev": 1.4501, "head_o": 0.3488}
{"step": 490, "total": 1250, "K": 8, "lr_bb": 7.430220054440267e-06, "grad_norm": 6.8962, "loss": 0.16624, "ce": 0.16624, "kl": 0.16624, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 688.9, "tok_per_s": 6140, "pass_tok_per_s": 22727, "peak_gib": 74.76, "epoch": 1, "gate": 0.000706, "W_dev": 1.4696, "head_o": 0.378}
{"step": 500, "total": 1250, "K": 4, "lr_bb": 7.321969549828923e-06, "grad_norm": 26.8872, "loss": 0.32834, "ce": 0.32833, "kl": 0.32833, "rps": 2e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 701.2, "tok_per_s": 6161, "pass_tok_per_s": 22735, "peak_gib": 74.76, "epoch": 1, "gate": 0.000793, "W_dev": 1.4906, "head_o": 0.3892}
{"event": "dev", "step": 500, "seconds": 9.7, "gate": 0.0007933251326903701, "K1": {"ce": 0.38919, "kl": 0.35718, "acc": 0.89474, "n": 1000}, "K4": {"ce": 0.38871, "kl": 0.35669, "acc": 0.89474, "n": 1000}}
{"step": 510, "total": 1250, "K": 4, "lr_bb": 7.212442858940257e-06, "grad_norm": 14.363, "loss": 0.2853, "ce": 0.2853, "kl": 0.23828, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 725.2, "tok_per_s": 6079, "pass_tok_per_s": 22440, "peak_gib": 74.76, "epoch": 1, "gate": 0.000762, "W_dev": 1.5053, "head_o": 0.3819}
{"step": 520, "total": 1250, "K": 8, "lr_bb": 7.101716699005861e-06, "grad_norm": 9.4986, "loss": 0.10194, "ce": 0.10194, "kl": 0.0824, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 737.2, "tok_per_s": 6099, "pass_tok_per_s": 22453, "peak_gib": 74.76, "epoch": 1, "gate": 0.000851, "W_dev": 1.5185, "head_o": 0.366}
{"step": 530, "total": 1250, "K": 8, "lr_bb": 6.989868627417334e-06, "grad_norm": 13.6444, "loss": 0.2755, "ce": 0.27529, "kl": 0.27529, "rps": 0.00083, "l1": 0.0, "dkl": 0.0, "elapsed": 755.9, "tok_per_s": 6061, "pass_tok_per_s": 22490, "peak_gib": 74.76, "epoch": 1, "gate": 0.000936, "W_dev": 1.5436, "head_o": 0.3384}
{"step": 540, "total": 1250, "K": 2, "lr_bb": 6.876976987401713e-06, "grad_norm": 13.8142, "loss": 0.32645, "ce": 0.32644, "kl": 0.32644, "rps": 3e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 769.9, "tok_per_s": 6066, "pass_tok_per_s": 22521, "peak_gib": 74.76, "epoch": 1, "gate": 0.000981, "W_dev": 1.5726, "head_o": 0.2935}
{"step": 550, "total": 1250, "K": 4, "lr_bb": 6.763120853146505e-06, "grad_norm": 19.0128, "loss": 0.3879, "ce": 0.38775, "kl": 0.36458, "rps": 0.00063, "l1": 0.0, "dkl": 0.0, "elapsed": 781.6, "tok_per_s": 6078, "pass_tok_per_s": 22539, "peak_gib": 74.76, "epoch": 1, "gate": 0.00102, "W_dev": 1.6057, "head_o": 0.2899}
{"step": 560, "total": 1250, "K": 8, "lr_bb": 6.648379974412702e-06, "grad_norm": 6.181, "loss": 0.16288, "ce": 0.16272, "kl": 0.13071, "rps": 0.00065, "l1": 0.0, "dkl": 0.0, "elapsed": 797.2, "tok_per_s": 6066, "pass_tok_per_s": 22593, "peak_gib": 74.76, "epoch": 1, "gate": 0.001024, "W_dev": 1.6361, "head_o": 0.3131}
{"step": 570, "total": 1250, "K": 4, "lr_bb": 6.532834720674585e-06, "grad_norm": 12.0985, "loss": 0.09244, "ce": 0.09244, "kl": 0.09244, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 811.6, "tok_per_s": 6064, "pass_tok_per_s": 22630, "peak_gib": 74.76, "epoch": 1, "gate": 0.001062, "W_dev": 1.6795, "head_o": 0.3529}
{"step": 580, "total": 1250, "K": 1, "lr_bb": 6.416566024825488e-06, "grad_norm": 10.7785, "loss": 0.44835, "ce": 0.44828, "kl": 0.40246, "rps": 0.00028, "l1": 0.0, "dkl": 0.0, "elapsed": 827.4, "tok_per_s": 6050, "pass_tok_per_s": 22641, "peak_gib": 74.76, "epoch": 1, "gate": 0.001085, "W_dev": 1.7003, "head_o": 0.3834}
{"step": 590, "total": 1250, "K": 8, "lr_bb": 6.2996553264888904e-06, "grad_norm": 22.012, "loss": 0.18524, "ce": 0.18499, "kl": 0.11692, "rps": 0.00102, "l1": 0.0, "dkl": 0.0, "elapsed": 838.3, "tok_per_s": 6073, "pass_tok_per_s": 22660, "peak_gib": 74.76, "epoch": 1, "gate": 0.001111, "W_dev": 1.7127, "head_o": 0.3854}
{"step": 600, "total": 1250, "K": 2, "lr_bb": 6.182184514974578e-06, "grad_norm": 14.2585, "loss": 0.21121, "ce": 0.21121, "kl": 0.17171, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 851.8, "tok_per_s": 6075, "pass_tok_per_s": 22666, "peak_gib": 74.76, "epoch": 1, "gate": 0.001159, "W_dev": 1.7308, "head_o": 0.3702}
{"step": 610, "total": 1250, "K": 8, "lr_bb": 6.064235871919834e-06, "grad_norm": 3.4369, "loss": 0.14024, "ce": 0.14001, "kl": 0.08072, "rps": 0.00093, "l1": 0.0, "dkl": 0.0, "elapsed": 865.3, "tok_per_s": 6076, "pass_tok_per_s": 22674, "peak_gib": 74.76, "epoch": 1, "gate": 0.001186, "W_dev": 1.7489, "head_o": 0.3502}
{"step": 620, "total": 1250, "K": 4, "lr_bb": 5.945892013655822e-06, "grad_norm": 15.2255, "loss": 0.47349, "ce": 0.47349, "kl": 0.45281, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 874.4, "tok_per_s": 6115, "pass_tok_per_s": 22674, "peak_gib": 74.76, "epoch": 1, "gate": 0.001136, "W_dev": 1.7601, "head_o": 0.3364}
{"step": 630, "total": 1250, "K": 2, "lr_bb": 5.8272358333395295e-06, "grad_norm": 2.6139, "loss": 0.05869, "ce": 0.05731, "kl": 0.05731, "rps": 0.00553, "l1": 0.0, "dkl": 0.0, "elapsed": 885.6, "tok_per_s": 6130, "pass_tok_per_s": 22696, "peak_gib": 74.76, "epoch": 1, "gate": 0.001249, "W_dev": 1.7941, "head_o": 0.3348}
{"step": 640, "total": 1250, "K": 2, "lr_bb": 5.708350442891818e-06, "grad_norm": 1.7385, "loss": 0.05435, "ce": 0.05426, "kl": 0.01627, "rps": 0.00035, "l1": 0.0, "dkl": 0.0, "elapsed": 896.2, "tok_per_s": 6151, "pass_tok_per_s": 22709, "peak_gib": 74.76, "epoch": 1, "gate": 0.001442, "W_dev": 1.8303, "head_o": 0.3254}
{"step": 650, "total": 1250, "K": 1, "lr_bb": 5.589319114782241e-06, "grad_norm": 11.2372, "loss": 0.41633, "ce": 0.41632, "kl": 0.41632, "rps": 4e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 914.6, "tok_per_s": 6118, "pass_tok_per_s": 22744, "peak_gib": 74.76, "epoch": 1, "gate": 0.001613, "W_dev": 1.8607, "head_o": 0.3174}
{"step": 660, "total": 1250, "K": 1, "lr_bb": 5.470225223701402e-06, "grad_norm": 6.5435, "loss": 0.15883, "ce": 0.15871, "kl": 0.11093, "rps": 0.00049, "l1": 0.0, "dkl": 0.0, "elapsed": 929.9, "tok_per_s": 6105, "pass_tok_per_s": 22744, "peak_gib": 74.76, "epoch": 1, "gate": 0.001613, "W_dev": 1.8759, "head_o": 0.3214}
{"step": 670, "total": 1250, "K": 8, "lr_bb": 5.351152188161713e-06, "grad_norm": 4.0226, "loss": 0.08672, "ce": 0.08672, "kl": 0.05971, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 941.2, "tok_per_s": 6121, "pass_tok_per_s": 22756, "peak_gib": 74.76, "epoch": 1, "gate": 0.001637, "W_dev": 1.8983, "head_o": 0.3227}
{"step": 680, "total": 1250, "K": 1, "lr_bb": 5.232183412067466e-06, "grad_norm": 7.3303, "loss": 0.16304, "ce": 0.16304, "kl": 0.16074, "rps": 1e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 952.9, "tok_per_s": 6133, "pass_tok_per_s": 22763, "peak_gib": 74.76, "epoch": 1, "gate": 0.001655, "W_dev": 1.9151, "head_o": 0.3294}
{"step": 690, "total": 1250, "K": 1, "lr_bb": 5.113402226295127e-06, "grad_norm": 3.6239, "loss": 0.1248, "ce": 0.1248, "kl": 0.07174, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 965.5, "tok_per_s": 6143, "pass_tok_per_s": 22779, "peak_gib": 74.76, "epoch": 1, "gate": 0.001809, "W_dev": 1.9413, "head_o": 0.3254}
{"step": 700, "total": 1250, "K": 4, "lr_bb": 4.994891830324788e-06, "grad_norm": 11.719, "loss": 0.15382, "ce": 0.15382, "kl": 0.13217, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 978.8, "tok_per_s": 6150, "pass_tok_per_s": 22795, "peak_gib": 74.76, "epoch": 1, "gate": 0.001895, "W_dev": 1.9588, "head_o": 0.332}
{"step": 710, "total": 1250, "K": 2, "lr_bb": 4.876735233963648e-06, "grad_norm": 16.3446, "loss": 0.24979, "ce": 0.24975, "kl": 0.21913, "rps": 0.00014, "l1": 0.0, "dkl": 0.0, "elapsed": 991.6, "tok_per_s": 6158, "pass_tok_per_s": 22808, "peak_gib": 74.76, "epoch": 1, "gate": 0.001937, "W_dev": 1.9669, "head_o": 0.3353}
{"step": 720, "total": 1250, "K": 2, "lr_bb": 4.759015199202359e-06, "grad_norm": 8.3614, "loss": 0.20983, "ce": 0.20983, "kl": 0.18085, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 1005.2, "tok_per_s": 6162, "pass_tok_per_s": 22831, "peak_gib": 74.76, "epoch": 1, "gate": 0.001893, "W_dev": 1.9695, "head_o": 0.3425}
{"step": 730, "total": 1250, "K": 8, "lr_bb": 4.641814182244947e-06, "grad_norm": 10.2764, "loss": 0.04635, "ce": 0.04635, "kl": 0.04635, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 1027.0, "tok_per_s": 6116, "pass_tok_per_s": 22858, "peak_gib": 74.76, "epoch": 1, "gate": 0.001867, "W_dev": 1.9826, "head_o": 0.3485}
{"step": 740, "total": 1250, "K": 2, "lr_bb": 4.52521427575292e-06, "grad_norm": 5.6833, "loss": 0.11908, "ce": 0.11908, "kl": 0.08442, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 1038.4, "tok_per_s": 6133, "pass_tok_per_s": 22866, "peak_gib": 74.76, "epoch": 1, "gate": 0.001776, "W_dev": 1.9887, "head_o": 0.3186}
{"step": 750, "total": 1250, "K": 4, "lr_bb": 4.409297151344013e-06, "grad_norm": 1.5415, "loss": 0.01917, "ce": 0.01917, "kl": 0.01917, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 1049.2, "tok_per_s": 6152, "pass_tok_per_s": 22862, "peak_gib": 74.76, "epoch": 1, "gate": 0.001733, "W_dev": 1.9963, "head_o": 0.3094}
{"event": "dev", "step": 750, "seconds": 9.8, "gate": 0.0017330661648884416, "K1": {"ce": 0.32943, "kl": 0.29741, "acc": 0.91368, "n": 1000}, "K4": {"ce": 0.32952, "kl": 0.29751, "acc": 0.91263, "n": 1000}}
{"step": 760, "total": 1250, "K": 4, "lr_bb": 4.29414400238586e-06, "grad_norm": 9.4239, "loss": 0.23025, "ce": 0.23025, "kl": 0.21797, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 1067.9, "tok_per_s": 6120, "pass_tok_per_s": 22663, "peak_gib": 74.76, "epoch": 1, "gate": 0.001696, "W_dev": 2.0, "head_o": 0.3122}
{"step": 770, "total": 1250, "K": 4, "lr_bb": 4.1798354871246475e-06, "grad_norm": 6.097, "loss": 0.06267, "ce": 0.06267, "kl": 0.06267, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 1077.8, "tok_per_s": 6143, "pass_tok_per_s": 22662, "peak_gib": 74.76, "epoch": 1, "gate": 0.001676, "W_dev": 2.004, "head_o": 0.3122}
{"step": 780, "total": 1250, "K": 1, "lr_bb": 4.066451672188585e-06, "grad_norm": 4.4717, "loss": 0.08893, "ce": 0.08893, "kl": 0.03636, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 1089.4, "tok_per_s": 6154, "pass_tok_per_s": 22677, "peak_gib": 74.76, "epoch": 1, "gate": 0.001646, "W_dev": 2.0043, "head_o": 0.3051}
{"step": 790, "total": 1250, "K": 1, "lr_bb": 3.95407197650578e-06, "grad_norm": 26.4629, "loss": 0.31751, "ce": 0.31521, "kl": 0.29851, "rps": 0.00921, "l1": 0.0, "dkl": 0.0, "elapsed": 1099.2, "tok_per_s": 6175, "pass_tok_per_s": 22669, "peak_gib": 74.76, "epoch": 1, "gate": 0.001698, "W_dev": 2.0089, "head_o": 0.287}
{"step": 800, "total": 1250, "K": 1, "lr_bb": 3.842775115675784e-06, "grad_norm": 3.1288, "loss": 0.03894, "ce": 0.03892, "kl": 0.03645, "rps": 8e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 1109.8, "tok_per_s": 6188, "pass_tok_per_s": 22662, "peak_gib": 74.76, "epoch": 1, "gate": 0.001814, "W_dev": 2.0234, "head_o": 0.2815}
{"step": 810, "total": 1250, "K": 4, "lr_bb": 3.732639046833784e-06, "grad_norm": 55.7977, "loss": 0.23467, "ce": 0.23461, "kl": 0.23461, "rps": 0.00024, "l1": 0.0, "dkl": 0.0, "elapsed": 1125.4, "tok_per_s": 6179, "pass_tok_per_s": 22673, "peak_gib": 74.76, "epoch": 1, "gate": 0.001827, "W_dev": 2.0317, "head_o": 0.2796}
{"step": 820, "total": 1250, "K": 1, "lr_bb": 3.6237409140460507e-06, "grad_norm": 22.8614, "loss": 0.23126, "ce": 0.23117, "kl": 0.19012, "rps": 0.0004, "l1": 0.0, "dkl": 0.0, "elapsed": 1142.2, "tok_per_s": 6163, "pass_tok_per_s": 22693, "peak_gib": 74.76, "epoch": 1, "gate": 0.001978, "W_dev": 2.0485, "head_o": 0.2789}
{"step": 830, "total": 1250, "K": 8, "lr_bb": 3.516156994274895e-06, "grad_norm": 11.7498, "loss": 0.3983, "ce": 0.3983, "kl": 0.39133, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 1156.9, "tok_per_s": 6159, "pass_tok_per_s": 22719, "peak_gib": 74.76, "epoch": 1, "gate": 0.0022, "W_dev": 2.0783, "head_o": 0.2798}
{"step": 840, "total": 1250, "K": 4, "lr_bb": 3.4099626439509966e-06, "grad_norm": 8.6236, "loss": 0.11256, "ce": 0.11255, "kl": 0.06768, "rps": 3e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 1169.8, "tok_per_s": 6162, "pass_tok_per_s": 22734, "peak_gib": 74.76, "epoch": 1, "gate": 0.00232, "W_dev": 2.0959, "head_o": 0.2832}
{"step": 850, "total": 1250, "K": 4, "lr_bb": 3.3052322461904816e-06, "grad_norm": 6.4696, "loss": 0.11735, "ce": 0.11735, "kl": 0.08729, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 1180.9, "tok_per_s": 6175, "pass_tok_per_s": 22736, "peak_gib": 74.76, "epoch": 1, "gate": 0.002383, "W_dev": 2.1063, "head_o": 0.2864}
{"step": 860, "total": 1250, "K": 2, "lr_bb": 3.202039158693789e-06, "grad_norm": 16.9692, "loss": 0.47597, "ce": 0.47597, "kl": 0.44757, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 1190.4, "tok_per_s": 6194, "pass_tok_per_s": 22726, "peak_gib": 74.76, "epoch": 1, "gate": 0.002408, "W_dev": 2.1117, "head_o": 0.2873}
{"step": 870, "total": 1250, "K": 4, "lr_bb": 3.1004556623627547e-06, "grad_norm": 12.7119, "loss": 0.19703, "ce": 0.19703, "kl": 0.17966, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 1205.4, "tok_per_s": 6187, "pass_tok_per_s": 22737, "peak_gib": 74.76, "epoch": 1, "gate": 0.002397, "W_dev": 2.1167, "head_o": 0.2878}
{"step": 880, "total": 1250, "K": 1, "lr_bb": 3.0005529106719404e-06, "grad_norm": 3.1476, "loss": 0.04254, "ce": 0.04254, "kl": 0.04254, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 1216.2, "tok_per_s": 6206, "pass_tok_per_s": 22734, "peak_gib": 74.76, "epoch": 1, "gate": 0.002409, "W_dev": 2.1214, "head_o": 0.289}
{"step": 890, "total": 1250, "K": 4, "lr_bb": 2.902400879829668e-06, "grad_norm": 1.1426, "loss": 0.05927, "ce": 0.05927, "kl": 0.01023, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 1234.3, "tok_per_s": 6182, "pass_tok_per_s": 22755, "peak_gib": 74.76, "epoch": 1, "gate": 0.002417, "W_dev": 2.1267, "head_o": 0.2887}
{"step": 900, "total": 1250, "K": 2, "lr_bb": 2.8060683197636575e-06, "grad_norm": 10.2233, "loss": 0.17886, "ce": 0.17886, "kl": 0.17886, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 1248.2, "tok_per_s": 6182, "pass_tok_per_s": 22764, "peak_gib": 74.76, "epoch": 1, "gate": 0.002357, "W_dev": 2.1254, "head_o": 0.2877}
{"step": 910, "total": 1250, "K": 1, "lr_bb": 2.711622705965592e-06, "grad_norm": 10.4194, "loss": 0.2547, "ce": 0.2547, "kl": 0.23364, "rps": 1e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 1260.9, "tok_per_s": 6189, "pass_tok_per_s": 22779, "peak_gib": 74.76, "epoch": 1, "gate": 0.002286, "W_dev": 2.1221, "head_o": 0.2894}
{"step": 920, "total": 1250, "K": 4, "lr_bb": 2.619130192228383e-06, "grad_norm": 10.4455, "loss": 0.15267, "ce": 0.14932, "kl": 0.14932, "rps": 0.01344, "l1": 0.0, "dkl": 0.0, "elapsed": 1275.5, "tok_per_s": 6181, "pass_tok_per_s": 22784, "peak_gib": 74.76, "epoch": 1, "gate": 0.002248, "W_dev": 2.1218, "head_o": 0.2927}
{"step": 930, "total": 1250, "K": 8, "lr_bb": 2.528655564309177e-06, "grad_norm": 11.8104, "loss": 0.35956, "ce": 0.35952, "kl": 0.31169, "rps": 0.00014, "l1": 0.0, "dkl": 0.0, "elapsed": 1290.7, "tok_per_s": 6174, "pass_tok_per_s": 22793, "peak_gib": 74.76, "epoch": 1, "gate": 0.002219, "W_dev": 2.1235, "head_o": 0.2979}
{"step": 940, "total": 1250, "K": 1, "lr_bb": 2.440262194550615e-06, "grad_norm": 23.7336, "loss": 0.36875, "ce": 0.36875, "kl": 0.36875, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 1309.8, "tok_per_s": 6150, "pass_tok_per_s": 22808, "peak_gib": 74.76, "epoch": 1, "gate": 0.002173, "W_dev": 2.1212, "head_o": 0.2965}
{"step": 950, "total": 1250, "K": 4, "lr_bb": 2.354011997492107e-06, "grad_norm": 8.5685, "loss": 0.486, "ce": 0.486, "kl": 0.43258, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 1318.7, "tok_per_s": 6168, "pass_tok_per_s": 22811, "peak_gib": 74.76, "epoch": 1, "gate": 0.00214, "W_dev": 2.1189, "head_o": 0.2933}
{"step": 960, "total": 1250, "K": 8, "lr_bb": 2.2699653865022036e-06, "grad_norm": 11.9557, "loss": 0.2195, "ce": 0.21707, "kl": 0.20624, "rps": 0.00972, "l1": 0.0, "dkl": 0.0, "elapsed": 1336.2, "tok_per_s": 6151, "pass_tok_per_s": 22822, "peak_gib": 74.76, "epoch": 1, "gate": 0.002111, "W_dev": 2.1188, "head_o": 0.2949}
{"step": 970, "total": 1250, "K": 1, "lr_bb": 2.188181231462461e-06, "grad_norm": 4.6931, "loss": 0.10282, "ce": 0.10281, "kl": 0.07172, "rps": 5e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 1351.2, "tok_per_s": 6149, "pass_tok_per_s": 22835, "peak_gib": 74.76, "epoch": 1, "gate": 0.002084, "W_dev": 2.1188, "head_o": 0.2921}
{"step": 980, "total": 1250, "K": 8, "lr_bb": 2.108716817532436e-06, "grad_norm": 7.772, "loss": 0.26441, "ce": 0.26434, "kl": 0.22846, "rps": 0.00031, "l1": 0.0, "dkl": 0.0, "elapsed": 1363.4, "tok_per_s": 6158, "pass_tok_per_s": 22841, "peak_gib": 74.76, "epoch": 1, "gate": 0.002065, "W_dev": 2.1184, "head_o": 0.2908}
{"step": 990, "total": 1250, "K": 4, "lr_bb": 2.031627805024673e-06, "grad_norm": 4.8972, "loss": 0.19456, "ce": 0.19456, "kl": 0.13528, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 1375.0, "tok_per_s": 6169, "pass_tok_per_s": 22843, "peak_gib": 74.76, "epoch": 1, "gate": 0.002035, "W_dev": 2.1171, "head_o": 0.2898}
{"step": 1000, "total": 1250, "K": 1, "lr_bb": 1.9569681904178194e-06, "grad_norm": 11.4761, "loss": 0.3203, "ce": 0.3201, "kl": 0.3201, "rps": 0.00081, "l1": 0.0, "dkl": 0.0, "elapsed": 1387.9, "tok_per_s": 6170, "pass_tok_per_s": 22847, "peak_gib": 74.76, "epoch": 1, "gate": 0.002027, "W_dev": 2.1165, "head_o": 0.2912}
{"event": "dev", "step": 1000, "seconds": 9.8, "gate": 0.0020267649088054895, "K1": {"ce": 0.31848, "kl": 0.28646, "acc": 0.92, "n": 1000}, "K4": {"ce": 0.31852, "kl": 0.28651, "acc": 0.92, "n": 1000}}
{"step": 1010, "total": 1250, "K": 8, "lr_bb": 1.8847902685351571e-06, "grad_norm": 5.9367, "loss": 0.17271, "ce": 0.17267, "kl": 0.07999, "rps": 0.00014, "l1": 0.0, "dkl": 0.0, "elapsed": 1407.6, "tok_per_s": 6143, "pass_tok_per_s": 22696, "peak_gib": 74.76, "epoch": 1, "gate": 0.002018, "W_dev": 2.1161, "head_o": 0.2926}
{"step": 1020, "total": 1250, "K": 4, "lr_bb": 1.8151445959150434e-06, "grad_norm": 16.4381, "loss": 0.19621, "ce": 0.19621, "kl": 0.1438, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 1425.2, "tok_per_s": 6131, "pass_tok_per_s": 22705, "peak_gib": 74.9, "epoch": 1, "gate": 0.001992, "W_dev": 2.1154, "head_o": 0.2906}
{"step": 1030, "total": 1250, "K": 4, "lr_bb": 1.7480799553989276e-06, "grad_norm": 22.2788, "loss": 0.54281, "ce": 0.54265, "kl": 0.53514, "rps": 0.00065, "l1": 0.0, "dkl": 0.0, "elapsed": 1445.0, "tok_per_s": 6108, "pass_tok_per_s": 22738, "peak_gib": 74.9, "epoch": 1, "gate": 0.001964, "W_dev": 2.1134, "head_o": 0.2901}
{"step": 1040, "total": 1250, "K": 4, "lr_bb": 1.683643321961735e-06, "grad_norm": 14.5364, "loss": 0.22011, "ce": 0.22011, "kl": 0.19908, "rps": 1e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 1460.7, "tok_per_s": 6102, "pass_tok_per_s": 22744, "peak_gib": 74.9, "epoch": 1, "gate": 0.001956, "W_dev": 2.113, "head_o": 0.2891}
{"step": 1050, "total": 1250, "K": 4, "lr_bb": 1.6218798298085659e-06, "grad_norm": 7.1001, "loss": 0.36109, "ce": 0.36109, "kl": 0.36109, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 1473.2, "tok_per_s": 6107, "pass_tok_per_s": 22759, "peak_gib": 74.9, "epoch": 1, "gate": 0.001961, "W_dev": 2.114, "head_o": 0.2917}
{"step": 1060, "total": 1250, "K": 1, "lr_bb": 1.5628327407607492e-06, "grad_norm": 17.9922, "loss": 0.66574, "ce": 0.66065, "kl": 0.66065, "rps": 0.02036, "l1": 0.0, "dkl": 0.0, "elapsed": 1490.9, "tok_per_s": 6092, "pass_tok_per_s": 22780, "peak_gib": 74.9, "epoch": 1, "gate": 0.00197, "W_dev": 2.115, "head_o": 0.2946}
{"step": 1070, "total": 1250, "K": 4, "lr_bb": 1.5065434139533798e-06, "grad_norm": 5.7585, "loss": 0.12618, "ce": 0.12618, "kl": 0.12618, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 1501.7, "tok_per_s": 6107, "pass_tok_per_s": 22783, "peak_gib": 74.9, "epoch": 1, "gate": 0.001969, "W_dev": 2.115, "head_o": 0.2947}
{"step": 1080, "total": 1250, "K": 1, "lr_bb": 1.453051276865607e-06, "grad_norm": 3.9083, "loss": 0.04741, "ce": 0.04574, "kl": 0.04574, "rps": 0.00668, "l1": 0.0, "dkl": 0.0, "elapsed": 1517.4, "tok_per_s": 6097, "pass_tok_per_s": 22795, "peak_gib": 74.9, "epoch": 1, "gate": 0.001957, "W_dev": 2.1144, "head_o": 0.2933}
{"step": 1090, "total": 1250, "K": 4, "lr_bb": 1.4023937977039125e-06, "grad_norm": 22.0592, "loss": 0.48202, "ce": 0.4815, "kl": 0.4815, "rps": 0.00206, "l1": 0.0, "dkl": 0.0, "elapsed": 1527.4, "tok_per_s": 6113, "pass_tok_per_s": 22798, "peak_gib": 74.9, "epoch": 1, "gate": 0.001952, "W_dev": 2.1142, "head_o": 0.2924}
{"step": 1100, "total": 1250, "K": 4, "lr_bb": 1.354606459157759e-06, "grad_norm": 9.4965, "loss": 0.31698, "ce": 0.31666, "kl": 0.24984, "rps": 0.00128, "l1": 0.0, "dkl": 0.0, "elapsed": 1541.1, "tok_per_s": 6111, "pass_tok_per_s": 22805, "peak_gib": 74.9, "epoch": 1, "gate": 0.001961, "W_dev": 2.1148, "head_o": 0.2924}
{"step": 1110, "total": 1250, "K": 4, "lr_bb": 1.3097227335459846e-06, "grad_norm": 20.4306, "loss": 0.42912, "ce": 0.42903, "kl": 0.38743, "rps": 0.00035, "l1": 0.0, "dkl": 0.0, "elapsed": 1553.5, "tok_per_s": 6115, "pass_tok_per_s": 22815, "peak_gib": 74.9, "epoch": 1, "gate": 0.001968, "W_dev": 2.1156, "head_o": 0.2927}
{"step": 1120, "total": 1250, "K": 2, "lr_bb": 1.2677740593713373e-06, "grad_norm": 11.7718, "loss": 0.20443, "ce": 0.20442, "kl": 0.11393, "rps": 3e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 1564.2, "tok_per_s": 6126, "pass_tok_per_s": 22826, "peak_gib": 74.9, "epoch": 1, "gate": 0.001967, "W_dev": 2.1157, "head_o": 0.2932}
{"step": 1130, "total": 1250, "K": 8, "lr_bb": 1.2287898192995924e-06, "grad_norm": 16.8813, "loss": 0.42781, "ce": 0.42781, "kl": 0.42781, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 1580.3, "tok_per_s": 6121, "pass_tok_per_s": 22840, "peak_gib": 74.9, "epoch": 1, "gate": 0.001962, "W_dev": 2.1153, "head_o": 0.2936}
{"step": 1140, "total": 1250, "K": 2, "lr_bb": 1.1927973195786588e-06, "grad_norm": 3.2939, "loss": 0.07857, "ce": 0.07857, "kl": 0.04573, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 1592.6, "tok_per_s": 6127, "pass_tok_per_s": 22852, "peak_gib": 74.9, "epoch": 1, "gate": 0.001957, "W_dev": 2.1149, "head_o": 0.2938}
{"step": 1150, "total": 1250, "K": 1, "lr_bb": 1.1598217709121034e-06, "grad_norm": 10.1837, "loss": 0.18505, "ce": 0.18499, "kl": 0.13882, "rps": 0.00022, "l1": 0.0, "dkl": 0.0, "elapsed": 1604.2, "tok_per_s": 6136, "pass_tok_per_s": 22857, "peak_gib": 74.9, "epoch": 1, "gate": 0.001954, "W_dev": 2.1146, "head_o": 0.294}
{"step": 1160, "total": 1250, "K": 1, "lr_bb": 1.1298862708004876e-06, "grad_norm": 11.1177, "loss": 0.12997, "ce": 0.12997, "kl": 0.1132, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 1612.1, "tok_per_s": 6162, "pass_tok_per_s": 22863, "peak_gib": 74.9, "epoch": 1, "gate": 0.001951, "W_dev": 2.1142, "head_o": 0.2935}
{"step": 1170, "total": 1250, "K": 8, "lr_bb": 1.1030117873628706e-06, "grad_norm": 12.9545, "loss": 0.24351, "ce": 0.24351, "kl": 0.22228, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 1628.7, "tok_per_s": 6155, "pass_tok_per_s": 22885, "peak_gib": 74.9, "epoch": 1, "gate": 0.001944, "W_dev": 2.1136, "head_o": 0.2933}
{"step": 1180, "total": 1250, "K": 2, "lr_bb": 1.07921714464984e-06, "grad_norm": 14.1588, "loss": 0.2687, "ce": 0.2687, "kl": 0.2156, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 1640.4, "tok_per_s": 6163, "pass_tok_per_s": 22892, "peak_gib": 74.9, "epoch": 1, "gate": 0.001937, "W_dev": 2.1131, "head_o": 0.2934}
{"step": 1190, "total": 1250, "K": 1, "lr_bb": 1.0585190094583284e-06, "grad_norm": 4.049, "loss": 0.01403, "ce": 0.01394, "kl": 0.01394, "rps": 0.00038, "l1": 0.0, "dkl": 0.0, "elapsed": 1656.8, "tok_per_s": 6153, "pass_tok_per_s": 22909, "peak_gib": 74.9, "epoch": 1, "gate": 0.001935, "W_dev": 2.113, "head_o": 0.2934}
{"step": 1200, "total": 1250, "K": 4, "lr_bb": 1.040931879657468e-06, "grad_norm": 2.8738, "loss": 0.10561, "ce": 0.10561, "kl": 0.06723, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 1675.0, "tok_per_s": 6136, "pass_tok_per_s": 22927, "peak_gib": 74.9, "epoch": 1, "gate": 0.001931, "W_dev": 2.1129, "head_o": 0.2932}
{"step": 1210, "total": 1250, "K": 8, "lr_bb": 1.0264680740336586e-06, "grad_norm": 18.1411, "loss": 0.18999, "ce": 0.18999, "kl": 0.18999, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 1694.8, "tok_per_s": 6116, "pass_tok_per_s": 22950, "peak_gib": 74.9, "epoch": 1, "gate": 0.001929, "W_dev": 2.113, "head_o": 0.2931}
{"step": 1220, "total": 1250, "K": 1, "lr_bb": 1.0151377236619533e-06, "grad_norm": 10.0622, "loss": 0.24782, "ce": 0.24782, "kl": 0.22745, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 1708.3, "tok_per_s": 6116, "pass_tok_per_s": 22965, "peak_gib": 74.9, "epoch": 1, "gate": 0.001926, "W_dev": 2.1129, "head_o": 0.293}
{"step": 1230, "total": 1250, "K": 4, "lr_bb": 1.0069487648098174e-06, "grad_norm": 22.18, "loss": 0.50829, "ce": 0.50829, "kl": 0.49444, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 1721.3, "tok_per_s": 6122, "pass_tok_per_s": 22970, "peak_gib": 74.9, "epoch": 1, "gate": 0.001926, "W_dev": 2.113, "head_o": 0.2931}
{"step": 1240, "total": 1250, "K": 4, "lr_bb": 1.0019069333782245e-06, "grad_norm": 22.4636, "loss": 0.25645, "ce": 0.25638, "kl": 0.18638, "rps": 0.00024, "l1": 0.0, "dkl": 0.0, "elapsed": 1735.5, "tok_per_s": 6122, "pass_tok_per_s": 22983, "peak_gib": 74.9, "epoch": 1, "gate": 0.001925, "W_dev": 2.113, "head_o": 0.2932}
{"step": 1250, "total": 1250, "K": 8, "lr_bb": 1.0000157608839794e-06, "grad_norm": 3.9753, "loss": 0.04013, "ce": 0.04013, "kl": 0.01976, "rps": 1e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 1752.8, "tok_per_s": 6109, "pass_tok_per_s": 22998, "peak_gib": 74.9, "epoch": 1, "gate": 0.001924, "W_dev": 2.1129, "head_o": 0.2933}
{"event": "dev", "step": 1250, "seconds": 9.8, "gate": 0.001924028154462576, "K1": {"ce": 0.32304, "kl": 0.29103, "acc": 0.91368, "n": 1000}, "K4": {"ce": 0.32302, "kl": 0.29101, "acc": 0.91368, "n": 1000}}
{"event": "finished", "step": 1250, "elapsed": 1762.6, "oom_retries": 0, "peak_gib": 74.9}
{"event": "saved", "path": "runs/e2/model.safetensors"}

Xet Storage Details

Size:
41.9 kB
·
Xet hash:
6639c49ebd52e7a32dfaf350a947d7cf8b83e163d737338804bf62ef1c9d6463

Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.