Buckets:

Pranav2748's picture
download
raw
37 kB
`torch_dtype` is deprecated! Use `dtype` instead!
{"event": "start", "train_count": 40000, "dev_count": 1000, "total_steps": 1250, "n_params": 600248320, "n_new": 4198400, "dropped_overlong": {}}
{"event": "dev", "step": 0, "seconds": 6.3, "gate": null, "K1": {"ce": 0.29219, "kl": 0.26018, "acc": 0.90316, "n": 1000}}
W1106 13:41:48.031000 16132 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 13:41:51.632000 16132 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": 1, "lr_bb": 1.5873015873015873e-06, "grad_norm": 8.5476, "loss": 0.06537, "ce": 0.06536, "kl": 0.04453, "rps": 2e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 11.1, "tok_per_s": 6338, "pass_tok_per_s": 6338, "peak_gib": 43.42, "epoch": 1, "head_o": 0.0185}
{"step": 20, "total": 1250, "K": 1, "lr_bb": 3.1746031746031746e-06, "grad_norm": 12.9231, "loss": 0.1804, "ce": 0.18039, "kl": 0.18039, "rps": 5e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 17.6, "tok_per_s": 8123, "pass_tok_per_s": 8123, "peak_gib": 46.85, "epoch": 1, "head_o": 0.0546}
{"step": 30, "total": 1250, "K": 1, "lr_bb": 4.761904761904762e-06, "grad_norm": 16.2937, "loss": 0.27766, "ce": 0.27766, "kl": 0.23481, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 24.2, "tok_per_s": 8921, "pass_tok_per_s": 8921, "peak_gib": 51.0, "epoch": 1, "head_o": 0.0733}
{"step": 40, "total": 1250, "K": 1, "lr_bb": 6.349206349206349e-06, "grad_norm": 9.8252, "loss": 0.15461, "ce": 0.15346, "kl": 0.15346, "rps": 0.00461, "l1": 0.0, "dkl": 0.0, "elapsed": 32.3, "tok_per_s": 9157, "pass_tok_per_s": 9157, "peak_gib": 52.17, "epoch": 1, "head_o": 0.1139}
{"step": 50, "total": 1250, "K": 1, "lr_bb": 7.936507936507936e-06, "grad_norm": 10.267, "loss": 0.34837, "ce": 0.34837, "kl": 0.34837, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 39.0, "tok_per_s": 9378, "pass_tok_per_s": 9378, "peak_gib": 52.17, "epoch": 1, "head_o": 0.094}
{"step": 60, "total": 1250, "K": 1, "lr_bb": 9.523809523809525e-06, "grad_norm": 8.0984, "loss": 0.23017, "ce": 0.23017, "kl": 0.20542, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 46.4, "tok_per_s": 9575, "pass_tok_per_s": 9575, "peak_gib": 52.17, "epoch": 1, "head_o": 0.1406}
{"step": 70, "total": 1250, "K": 1, "lr_bb": 9.999432619768909e-06, "grad_norm": 25.1559, "loss": 0.41199, "ce": 0.40801, "kl": 0.40801, "rps": 0.01589, "l1": 0.0, "dkl": 0.0, "elapsed": 53.9, "tok_per_s": 9630, "pass_tok_per_s": 9630, "peak_gib": 57.91, "epoch": 1, "head_o": 0.1494}
{"step": 80, "total": 1250, "K": 1, "lr_bb": 9.995965814255133e-06, "grad_norm": 8.2417, "loss": 0.26281, "ce": 0.26265, "kl": 0.20051, "rps": 0.00067, "l1": 0.0, "dkl": 0.0, "elapsed": 60.3, "tok_per_s": 9812, "pass_tok_per_s": 9812, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1435}
{"step": 90, "total": 1250, "K": 1, "lr_bb": 9.989349839823577e-06, "grad_norm": 12.7841, "loss": 0.37896, "ce": 0.37876, "kl": 0.26663, "rps": 0.00079, "l1": 0.0, "dkl": 0.0, "elapsed": 67.8, "tok_per_s": 9816, "pass_tok_per_s": 9816, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1601}
{"step": 100, "total": 1250, "K": 1, "lr_bb": 9.979589330588889e-06, "grad_norm": 8.3805, "loss": 0.31586, "ce": 0.31586, "kl": 0.27518, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 75.0, "tok_per_s": 9930, "pass_tok_per_s": 9930, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1321}
{"step": 110, "total": 1250, "K": 1, "lr_bb": 9.966691123233898e-06, "grad_norm": 13.4612, "loss": 0.24876, "ce": 0.24826, "kl": 0.24826, "rps": 0.00202, "l1": 0.0, "dkl": 0.0, "elapsed": 81.4, "tok_per_s": 10041, "pass_tok_per_s": 10041, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1518}
{"step": 120, "total": 1250, "K": 1, "lr_bb": 9.95066425222092e-06, "grad_norm": 14.0159, "loss": 0.27359, "ce": 0.27359, "kl": 0.2537, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 88.1, "tok_per_s": 10108, "pass_tok_per_s": 10108, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1698}
{"step": 130, "total": 1250, "K": 1, "lr_bb": 9.931519943463613e-06, "grad_norm": 10.2511, "loss": 0.29798, "ce": 0.29796, "kl": 0.24065, "rps": 0.00011, "l1": 0.0, "dkl": 0.0, "elapsed": 94.2, "tok_per_s": 10224, "pass_tok_per_s": 10224, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1723}
{"step": 140, "total": 1250, "K": 1, "lr_bb": 9.909271606463866e-06, "grad_norm": 47.1557, "loss": 0.66684, "ce": 0.66515, "kl": 0.63025, "rps": 0.00678, "l1": 0.0, "dkl": 0.0, "elapsed": 100.3, "tok_per_s": 10345, "pass_tok_per_s": 10345, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1652}
{"step": 150, "total": 1250, "K": 1, "lr_bb": 9.883934824919209e-06, "grad_norm": 17.499, "loss": 0.59316, "ce": 0.59315, "kl": 0.59315, "rps": 3e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 106.9, "tok_per_s": 10415, "pass_tok_per_s": 10415, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1959}
{"step": 160, "total": 1250, "K": 1, "lr_bb": 9.855527345807316e-06, "grad_norm": 7.8649, "loss": 0.31848, "ce": 0.31848, "kl": 0.24617, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 113.4, "tok_per_s": 10463, "pass_tok_per_s": 10463, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1625}
{"step": 170, "total": 1250, "K": 1, "lr_bb": 9.824069066955246e-06, "grad_norm": 18.7939, "loss": 0.51677, "ce": 0.51401, "kl": 0.49364, "rps": 0.01106, "l1": 0.0, "dkl": 0.0, "elapsed": 120.0, "tok_per_s": 10490, "pass_tok_per_s": 10490, "peak_gib": 61.4, "epoch": 1, "head_o": 0.171}
{"step": 180, "total": 1250, "K": 1, "lr_bb": 9.789582023102157e-06, "grad_norm": 19.0382, "loss": 0.33252, "ce": 0.33252, "kl": 0.33252, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 126.7, "tok_per_s": 10548, "pass_tok_per_s": 10548, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1595}
{"step": 190, "total": 1250, "K": 1, "lr_bb": 9.75209037046521e-06, "grad_norm": 6.9348, "loss": 0.25742, "ce": 0.25279, "kl": 0.25279, "rps": 0.01852, "l1": 0.0, "dkl": 0.0, "elapsed": 133.8, "tok_per_s": 10552, "pass_tok_per_s": 10552, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1711}
{"step": 200, "total": 1250, "K": 1, "lr_bb": 9.711620369819521e-06, "grad_norm": 7.5667, "loss": 0.35246, "ce": 0.35244, "kl": 0.32096, "rps": 9e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 140.0, "tok_per_s": 10642, "pass_tok_per_s": 10642, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1795}
{"step": 210, "total": 1250, "K": 1, "lr_bb": 9.668200368103964e-06, "grad_norm": 8.2447, "loss": 0.32593, "ce": 0.32593, "kl": 0.30786, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 146.7, "tok_per_s": 10662, "pass_tok_per_s": 10662, "peak_gib": 61.4, "epoch": 1, "head_o": 0.2248}
{"step": 220, "total": 1250, "K": 1, "lr_bb": 9.621860778565765e-06, "grad_norm": 7.6411, "loss": 0.15891, "ce": 0.15891, "kl": 0.14013, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 154.6, "tok_per_s": 10617, "pass_tok_per_s": 10617, "peak_gib": 61.4, "epoch": 1, "head_o": 0.2282}
{"step": 230, "total": 1250, "K": 1, "lr_bb": 9.572634059457735e-06, "grad_norm": 10.1675, "loss": 0.30912, "ce": 0.30912, "kl": 0.27279, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 161.4, "tok_per_s": 10618, "pass_tok_per_s": 10618, "peak_gib": 61.4, "epoch": 1, "head_o": 0.2323}
{"step": 240, "total": 1250, "K": 1, "lr_bb": 9.520554691303113e-06, "grad_norm": 18.6369, "loss": 0.34762, "ce": 0.34762, "kl": 0.3135, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 167.6, "tok_per_s": 10665, "pass_tok_per_s": 10665, "peak_gib": 61.4, "epoch": 1, "head_o": 0.2287}
{"step": 250, "total": 1250, "K": 1, "lr_bb": 9.465659152743914e-06, "grad_norm": 7.0155, "loss": 0.20418, "ce": 0.20056, "kl": 0.14899, "rps": 0.01449, "l1": 0.0, "dkl": 0.0, "elapsed": 173.9, "tok_per_s": 10703, "pass_tok_per_s": 10703, "peak_gib": 61.4, "epoch": 1, "head_o": 0.2117}
{"event": "dev", "step": 250, "seconds": 2.7, "gate": null, "K1": {"ce": 0.3593, "kl": 0.32728, "acc": 0.87579, "n": 1000}}
{"step": 260, "total": 1250, "K": 1, "lr_bb": 9.407985894989713e-06, "grad_norm": 12.1242, "loss": 0.2093, "ce": 0.20497, "kl": 0.20497, "rps": 0.01731, "l1": 0.0, "dkl": 0.0, "elapsed": 184.3, "tok_per_s": 10511, "pass_tok_per_s": 10511, "peak_gib": 61.4, "epoch": 1, "head_o": 0.2174}
{"step": 270, "total": 1250, "K": 1, "lr_bb": 9.347575314884746e-06, "grad_norm": 21.6832, "loss": 0.51471, "ce": 0.51456, "kl": 0.51456, "rps": 0.0006, "l1": 0.0, "dkl": 0.0, "elapsed": 192.1, "tok_per_s": 10489, "pass_tok_per_s": 10489, "peak_gib": 61.4, "epoch": 1, "head_o": 0.2046}
{"step": 280, "total": 1250, "K": 1, "lr_bb": 9.284469726612224e-06, "grad_norm": 13.5253, "loss": 0.24188, "ce": 0.24188, "kl": 0.21759, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 198.5, "tok_per_s": 10519, "pass_tok_per_s": 10519, "peak_gib": 61.4, "epoch": 1, "head_o": 0.191}
{"step": 290, "total": 1250, "K": 1, "lr_bb": 9.218713332055646e-06, "grad_norm": 6.6585, "loss": 0.33694, "ce": 0.33665, "kl": 0.315, "rps": 0.00116, "l1": 0.0, "dkl": 0.0, "elapsed": 205.1, "tok_per_s": 10543, "pass_tok_per_s": 10543, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1875}
{"step": 300, "total": 1250, "K": 1, "lr_bb": 9.150352189837881e-06, "grad_norm": 8.3801, "loss": 0.34574, "ce": 0.34574, "kl": 0.29323, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 212.3, "tok_per_s": 10537, "pass_tok_per_s": 10537, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1915}
{"step": 310, "total": 1250, "K": 1, "lr_bb": 9.079434183059728e-06, "grad_norm": 28.2715, "loss": 0.35755, "ce": 0.35483, "kl": 0.24971, "rps": 0.01091, "l1": 0.0, "dkl": 0.0, "elapsed": 218.4, "tok_per_s": 10589, "pass_tok_per_s": 10589, "peak_gib": 61.4, "epoch": 1, "head_o": 0.2087}
{"step": 320, "total": 1250, "K": 1, "lr_bb": 9.006008985760507e-06, "grad_norm": 13.3196, "loss": 0.30293, "ce": 0.29571, "kl": 0.23175, "rps": 0.02891, "l1": 0.0, "dkl": 0.0, "elapsed": 225.8, "tok_per_s": 10599, "pass_tok_per_s": 10599, "peak_gib": 61.4, "epoch": 1, "head_o": 0.188}
{"step": 330, "total": 1250, "K": 1, "lr_bb": 8.930128028124227e-06, "grad_norm": 12.1154, "loss": 0.28603, "ce": 0.28594, "kl": 0.20735, "rps": 0.00035, "l1": 0.0, "dkl": 0.0, "elapsed": 231.9, "tok_per_s": 10629, "pass_tok_per_s": 10629, "peak_gib": 61.4, "epoch": 1, "head_o": 0.2189}
{"step": 340, "total": 1250, "K": 1, "lr_bb": 8.851844460455654e-06, "grad_norm": 9.1275, "loss": 0.35137, "ce": 0.34983, "kl": 0.33124, "rps": 0.00617, "l1": 0.0, "dkl": 0.0, "elapsed": 238.7, "tok_per_s": 10645, "pass_tok_per_s": 10645, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1854}
{"step": 350, "total": 1250, "K": 1, "lr_bb": 8.771213115951537e-06, "grad_norm": 4.385, "loss": 0.08134, "ce": 0.08131, "kl": 0.04854, "rps": 0.00012, "l1": 0.0, "dkl": 0.0, "elapsed": 245.9, "tok_per_s": 10661, "pass_tok_per_s": 10661, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1834}
{"step": 360, "total": 1250, "K": 1, "lr_bb": 8.688290472293068e-06, "grad_norm": 6.9915, "loss": 0.27678, "ce": 0.27194, "kl": 0.1742, "rps": 0.01934, "l1": 0.0, "dkl": 0.0, "elapsed": 252.6, "tok_per_s": 10670, "pass_tok_per_s": 10670, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1749}
{"step": 370, "total": 1250, "K": 1, "lr_bb": 8.603134612086474e-06, "grad_norm": 10.5265, "loss": 0.25377, "ce": 0.2537, "kl": 0.23204, "rps": 0.00028, "l1": 0.0, "dkl": 0.0, "elapsed": 259.1, "tok_per_s": 10677, "pass_tok_per_s": 10677, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1733}
{"step": 380, "total": 1250, "K": 1, "lr_bb": 8.515805182179437e-06, "grad_norm": 13.0197, "loss": 0.16994, "ce": 0.16896, "kl": 0.13531, "rps": 0.00394, "l1": 0.0, "dkl": 0.0, "elapsed": 265.9, "tok_per_s": 10679, "pass_tok_per_s": 10679, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1631}
{"step": 390, "total": 1250, "K": 1, "lr_bb": 8.426363351881879e-06, "grad_norm": 9.2825, "loss": 0.20671, "ce": 0.20663, "kl": 0.18257, "rps": 0.00034, "l1": 0.0, "dkl": 0.0, "elapsed": 273.2, "tok_per_s": 10663, "pass_tok_per_s": 10663, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1696}
{"step": 400, "total": 1250, "K": 1, "lr_bb": 8.33487177012032e-06, "grad_norm": 8.3511, "loss": 0.21005, "ce": 0.21004, "kl": 0.15531, "rps": 1e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 280.1, "tok_per_s": 10665, "pass_tok_per_s": 10665, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1792}
{"step": 410, "total": 1250, "K": 1, "lr_bb": 8.241394521555854e-06, "grad_norm": 17.8931, "loss": 0.24541, "ce": 0.2454, "kl": 0.21963, "rps": 1e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 287.2, "tok_per_s": 10673, "pass_tok_per_s": 10673, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1816}
{"step": 420, "total": 1250, "K": 1, "lr_bb": 8.145997081696497e-06, "grad_norm": 13.5442, "loss": 0.29418, "ce": 0.29418, "kl": 0.29418, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 293.6, "tok_per_s": 10679, "pass_tok_per_s": 10679, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1789}
{"step": 430, "total": 1250, "K": 1, "lr_bb": 8.048746271035295e-06, "grad_norm": 3.2099, "loss": 0.0944, "ce": 0.09437, "kl": 0.05918, "rps": 0.00013, "l1": 0.0, "dkl": 0.0, "elapsed": 300.1, "tok_per_s": 10688, "pass_tok_per_s": 10688, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1905}
{"step": 440, "total": 1250, "K": 1, "lr_bb": 7.949710208246379e-06, "grad_norm": 7.3221, "loss": 0.22613, "ce": 0.22613, "kl": 0.12831, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 308.0, "tok_per_s": 10659, "pass_tok_per_s": 10659, "peak_gib": 61.4, "epoch": 1, "head_o": 0.2596}
{"step": 450, "total": 1250, "K": 1, "lr_bb": 7.848958262471667e-06, "grad_norm": 17.761, "loss": 0.38083, "ce": 0.38063, "kl": 0.38063, "rps": 0.00079, "l1": 0.0, "dkl": 0.0, "elapsed": 314.6, "tok_per_s": 10669, "pass_tok_per_s": 10669, "peak_gib": 61.4, "epoch": 1, "head_o": 0.195}
{"step": 460, "total": 1250, "K": 1, "lr_bb": 7.74656100473175e-06, "grad_norm": 14.1929, "loss": 0.30567, "ce": 0.30332, "kl": 0.29855, "rps": 0.00943, "l1": 0.0, "dkl": 0.0, "elapsed": 322.2, "tok_per_s": 10671, "pass_tok_per_s": 10671, "peak_gib": 61.4, "epoch": 1, "head_o": 0.2008}
{"step": 470, "total": 1250, "K": 1, "lr_bb": 7.642590158494882e-06, "grad_norm": 5.4206, "loss": 0.1127, "ce": 0.11269, "kl": 0.11269, "rps": 2e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 329.3, "tok_per_s": 10650, "pass_tok_per_s": 10650, "peak_gib": 61.4, "epoch": 1, "head_o": 0.2114}
{"step": 480, "total": 1250, "K": 1, "lr_bb": 7.537118549438758e-06, "grad_norm": 10.7665, "loss": 0.17474, "ce": 0.17474, "kl": 0.11525, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 336.0, "tok_per_s": 10671, "pass_tok_per_s": 10671, "peak_gib": 61.4, "epoch": 1, "head_o": 0.2084}
{"step": 490, "total": 1250, "K": 1, "lr_bb": 7.430220054440267e-06, "grad_norm": 6.167, "loss": 0.16946, "ce": 0.16946, "kl": 0.16946, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 343.5, "tok_per_s": 10671, "pass_tok_per_s": 10671, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1959}
{"step": 500, "total": 1250, "K": 1, "lr_bb": 7.321969549828923e-06, "grad_norm": 13.4391, "loss": 0.17718, "ce": 0.17718, "kl": 0.17718, "rps": 1e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 351.3, "tok_per_s": 10658, "pass_tok_per_s": 10658, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1982}
{"event": "dev", "step": 500, "seconds": 2.9, "gate": null, "K1": {"ce": 0.35938, "kl": 0.32737, "acc": 0.89474, "n": 1000}}
{"step": 510, "total": 1250, "K": 1, "lr_bb": 7.212442858940257e-06, "grad_norm": 16.5254, "loss": 0.2662, "ce": 0.2662, "kl": 0.21918, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 361.1, "tok_per_s": 10583, "pass_tok_per_s": 10583, "peak_gib": 61.4, "epoch": 1, "head_o": 0.251}
{"step": 520, "total": 1250, "K": 1, "lr_bb": 7.101716699005861e-06, "grad_norm": 15.0952, "loss": 0.20699, "ce": 0.20699, "kl": 0.18744, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 368.5, "tok_per_s": 10575, "pass_tok_per_s": 10575, "peak_gib": 61.4, "epoch": 1, "head_o": 0.2465}
{"step": 530, "total": 1250, "K": 1, "lr_bb": 6.989868627417334e-06, "grad_norm": 19.0595, "loss": 0.21596, "ce": 0.21591, "kl": 0.21591, "rps": 0.00018, "l1": 0.0, "dkl": 0.0, "elapsed": 375.5, "tok_per_s": 10576, "pass_tok_per_s": 10576, "peak_gib": 61.4, "epoch": 1, "head_o": 0.2022}
{"step": 540, "total": 1250, "K": 1, "lr_bb": 6.876976987401713e-06, "grad_norm": 13.9295, "loss": 0.49516, "ce": 0.49516, "kl": 0.49516, "rps": 1e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 383.0, "tok_per_s": 10569, "pass_tok_per_s": 10569, "peak_gib": 61.4, "epoch": 1, "head_o": 0.218}
{"step": 550, "total": 1250, "K": 1, "lr_bb": 6.763120853146505e-06, "grad_norm": 21.8203, "loss": 0.33662, "ce": 0.33635, "kl": 0.31318, "rps": 0.00109, "l1": 0.0, "dkl": 0.0, "elapsed": 389.4, "tok_per_s": 10572, "pass_tok_per_s": 10572, "peak_gib": 61.4, "epoch": 1, "head_o": 0.2713}
{"step": 560, "total": 1250, "K": 1, "lr_bb": 6.648379974412702e-06, "grad_norm": 5.1401, "loss": 0.16976, "ce": 0.16974, "kl": 0.13773, "rps": 9e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 395.8, "tok_per_s": 10589, "pass_tok_per_s": 10589, "peak_gib": 61.4, "epoch": 1, "head_o": 0.3036}
{"step": 570, "total": 1250, "K": 1, "lr_bb": 6.532834720674585e-06, "grad_norm": 11.7574, "loss": 0.06894, "ce": 0.06894, "kl": 0.06894, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 402.1, "tok_per_s": 10606, "pass_tok_per_s": 10606, "peak_gib": 61.4, "epoch": 1, "head_o": 0.3457}
{"step": 580, "total": 1250, "K": 1, "lr_bb": 6.416566024825488e-06, "grad_norm": 16.2517, "loss": 0.41927, "ce": 0.4192, "kl": 0.37338, "rps": 0.00027, "l1": 0.0, "dkl": 0.0, "elapsed": 408.2, "tok_per_s": 10626, "pass_tok_per_s": 10626, "peak_gib": 61.4, "epoch": 1, "head_o": 0.3503}
{"step": 590, "total": 1250, "K": 1, "lr_bb": 6.2996553264888904e-06, "grad_norm": 11.2067, "loss": 0.29123, "ce": 0.29113, "kl": 0.22307, "rps": 0.00039, "l1": 0.0, "dkl": 0.0, "elapsed": 414.3, "tok_per_s": 10646, "pass_tok_per_s": 10646, "peak_gib": 61.4, "epoch": 1, "head_o": 0.3234}
{"step": 600, "total": 1250, "K": 1, "lr_bb": 6.182184514974578e-06, "grad_norm": 9.8334, "loss": 0.18542, "ce": 0.18542, "kl": 0.14592, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 420.4, "tok_per_s": 10664, "pass_tok_per_s": 10664, "peak_gib": 61.4, "epoch": 1, "head_o": 0.2614}
{"step": 610, "total": 1250, "K": 1, "lr_bb": 6.064235871919834e-06, "grad_norm": 3.5117, "loss": 0.13576, "ce": 0.13557, "kl": 0.07628, "rps": 0.00077, "l1": 0.0, "dkl": 0.0, "elapsed": 427.3, "tok_per_s": 10661, "pass_tok_per_s": 10661, "peak_gib": 61.4, "epoch": 1, "head_o": 0.2452}
{"step": 620, "total": 1250, "K": 1, "lr_bb": 5.945892013655822e-06, "grad_norm": 21.0622, "loss": 0.50998, "ce": 0.50998, "kl": 0.4893, "rps": 1e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 434.1, "tok_per_s": 10672, "pass_tok_per_s": 10672, "peak_gib": 61.4, "epoch": 1, "head_o": 0.221}
{"step": 630, "total": 1250, "K": 1, "lr_bb": 5.8272358333395295e-06, "grad_norm": 7.5544, "loss": 0.06323, "ce": 0.06323, "kl": 0.06323, "rps": 2e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 440.3, "tok_per_s": 10683, "pass_tok_per_s": 10683, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1907}
{"step": 640, "total": 1250, "K": 1, "lr_bb": 5.708350442891818e-06, "grad_norm": 0.8504, "loss": 0.04838, "ce": 0.04835, "kl": 0.01036, "rps": 0.00014, "l1": 0.0, "dkl": 0.0, "elapsed": 446.1, "tok_per_s": 10703, "pass_tok_per_s": 10703, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1923}
{"step": 650, "total": 1250, "K": 1, "lr_bb": 5.589319114782241e-06, "grad_norm": 14.1013, "loss": 0.52972, "ce": 0.52972, "kl": 0.52972, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 452.0, "tok_per_s": 10722, "pass_tok_per_s": 10722, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1903}
{"step": 660, "total": 1250, "K": 1, "lr_bb": 5.470225223701402e-06, "grad_norm": 13.5305, "loss": 0.21504, "ce": 0.21504, "kl": 0.16726, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 459.0, "tok_per_s": 10711, "pass_tok_per_s": 10711, "peak_gib": 61.4, "epoch": 1, "head_o": 0.2141}
{"step": 670, "total": 1250, "K": 1, "lr_bb": 5.351152188161713e-06, "grad_norm": 4.2132, "loss": 0.09132, "ce": 0.09132, "kl": 0.06431, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 466.0, "tok_per_s": 10708, "pass_tok_per_s": 10708, "peak_gib": 61.4, "epoch": 1, "head_o": 0.2029}
{"step": 680, "total": 1250, "K": 1, "lr_bb": 5.232183412067466e-06, "grad_norm": 8.2756, "loss": 0.18416, "ce": 0.18416, "kl": 0.18186, "rps": 1e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 473.1, "tok_per_s": 10697, "pass_tok_per_s": 10697, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1987}
{"step": 690, "total": 1250, "K": 1, "lr_bb": 5.113402226295127e-06, "grad_norm": 4.0422, "loss": 0.14139, "ce": 0.14139, "kl": 0.08833, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 480.1, "tok_per_s": 10698, "pass_tok_per_s": 10698, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1911}
{"step": 700, "total": 1250, "K": 1, "lr_bb": 4.994891830324788e-06, "grad_norm": 10.3292, "loss": 0.19559, "ce": 0.19559, "kl": 0.17394, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 487.0, "tok_per_s": 10706, "pass_tok_per_s": 10706, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1903}
{"step": 710, "total": 1250, "K": 1, "lr_bb": 4.876735233963648e-06, "grad_norm": 14.2077, "loss": 0.20918, "ce": 0.20901, "kl": 0.17839, "rps": 0.00066, "l1": 0.0, "dkl": 0.0, "elapsed": 493.6, "tok_per_s": 10714, "pass_tok_per_s": 10714, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1898}
{"step": 720, "total": 1250, "K": 1, "lr_bb": 4.759015199202359e-06, "grad_norm": 7.5864, "loss": 0.1865, "ce": 0.18649, "kl": 0.15751, "rps": 3e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 499.7, "tok_per_s": 10735, "pass_tok_per_s": 10735, "peak_gib": 61.4, "epoch": 1, "head_o": 0.2016}
{"step": 730, "total": 1250, "K": 1, "lr_bb": 4.641814182244947e-06, "grad_norm": 16.5523, "loss": 0.06867, "ce": 0.06867, "kl": 0.06867, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 506.5, "tok_per_s": 10742, "pass_tok_per_s": 10742, "peak_gib": 61.4, "epoch": 1, "head_o": 0.2066}
{"step": 740, "total": 1250, "K": 1, "lr_bb": 4.52521427575292e-06, "grad_norm": 6.2301, "loss": 0.09627, "ce": 0.09627, "kl": 0.06162, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 513.7, "tok_per_s": 10738, "pass_tok_per_s": 10738, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1745}
{"step": 750, "total": 1250, "K": 1, "lr_bb": 4.409297151344013e-06, "grad_norm": 1.2509, "loss": 0.01382, "ce": 0.01382, "kl": 0.01382, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 520.8, "tok_per_s": 10734, "pass_tok_per_s": 10734, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1666}
{"event": "dev", "step": 750, "seconds": 2.8, "gate": null, "K1": {"ce": 0.3593, "kl": 0.32729, "acc": 0.91053, "n": 1000}}
{"step": 760, "total": 1250, "K": 1, "lr_bb": 4.29414400238586e-06, "grad_norm": 11.4376, "loss": 0.22143, "ce": 0.22143, "kl": 0.20915, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 529.9, "tok_per_s": 10682, "pass_tok_per_s": 10682, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1672}
{"step": 770, "total": 1250, "K": 1, "lr_bb": 4.1798354871246475e-06, "grad_norm": 4.9659, "loss": 0.05759, "ce": 0.05759, "kl": 0.05759, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 536.7, "tok_per_s": 10683, "pass_tok_per_s": 10683, "peak_gib": 61.4, "epoch": 1, "head_o": 0.169}
{"step": 780, "total": 1250, "K": 1, "lr_bb": 4.066451672188585e-06, "grad_norm": 8.5667, "loss": 0.11578, "ce": 0.11578, "kl": 0.06321, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 543.3, "tok_per_s": 10686, "pass_tok_per_s": 10686, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1698}
{"step": 790, "total": 1250, "K": 1, "lr_bb": 3.95407197650578e-06, "grad_norm": 30.7865, "loss": 0.31432, "ce": 0.31432, "kl": 0.29762, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 549.4, "tok_per_s": 10697, "pass_tok_per_s": 10697, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1687}
{"step": 800, "total": 1250, "K": 1, "lr_bb": 3.842775115675784e-06, "grad_norm": 11.7607, "loss": 0.05563, "ce": 0.0555, "kl": 0.05303, "rps": 0.00049, "l1": 0.0, "dkl": 0.0, "elapsed": 555.5, "tok_per_s": 10703, "pass_tok_per_s": 10703, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1741}
{"step": 810, "total": 1250, "K": 1, "lr_bb": 3.732639046833784e-06, "grad_norm": 7.9912, "loss": 0.18937, "ce": 0.18937, "kl": 0.18937, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 563.3, "tok_per_s": 10688, "pass_tok_per_s": 10688, "peak_gib": 61.4, "epoch": 1, "head_o": 0.169}
{"step": 820, "total": 1250, "K": 1, "lr_bb": 3.6237409140460507e-06, "grad_norm": 35.9347, "loss": 0.3089, "ce": 0.3088, "kl": 0.26776, "rps": 0.00039, "l1": 0.0, "dkl": 0.0, "elapsed": 569.4, "tok_per_s": 10703, "pass_tok_per_s": 10703, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1713}
{"step": 830, "total": 1250, "K": 1, "lr_bb": 3.516156994274895e-06, "grad_norm": 12.2562, "loss": 0.4155, "ce": 0.4155, "kl": 0.40853, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 575.8, "tok_per_s": 10715, "pass_tok_per_s": 10715, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1725}
{"step": 840, "total": 1250, "K": 1, "lr_bb": 3.4099626439509966e-06, "grad_norm": 5.6097, "loss": 0.08926, "ce": 0.08926, "kl": 0.04438, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 581.9, "tok_per_s": 10725, "pass_tok_per_s": 10725, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1738}
{"step": 850, "total": 1250, "K": 1, "lr_bb": 3.3052322461904816e-06, "grad_norm": 7.591, "loss": 0.13415, "ce": 0.13415, "kl": 0.10409, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 589.3, "tok_per_s": 10711, "pass_tok_per_s": 10711, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1731}
{"step": 860, "total": 1250, "K": 1, "lr_bb": 3.202039158693789e-06, "grad_norm": 15.1372, "loss": 0.4315, "ce": 0.4315, "kl": 0.4031, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 597.0, "tok_per_s": 10691, "pass_tok_per_s": 10691, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1729}
{"step": 870, "total": 1250, "K": 1, "lr_bb": 3.1004556623627547e-06, "grad_norm": 14.3503, "loss": 0.24488, "ce": 0.24488, "kl": 0.22751, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 604.2, "tok_per_s": 10685, "pass_tok_per_s": 10685, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1759}
{"step": 880, "total": 1250, "K": 1, "lr_bb": 3.0005529106719404e-06, "grad_norm": 3.6209, "loss": 0.04943, "ce": 0.04943, "kl": 0.04943, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 611.5, "tok_per_s": 10686, "pass_tok_per_s": 10686, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1773}
{"step": 890, "total": 1250, "K": 1, "lr_bb": 2.902400879829668e-06, "grad_norm": 1.5789, "loss": 0.05964, "ce": 0.05964, "kl": 0.01059, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 617.6, "tok_per_s": 10694, "pass_tok_per_s": 10694, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1856}
{"step": 900, "total": 1250, "K": 1, "lr_bb": 2.8060683197636575e-06, "grad_norm": 21.1187, "loss": 0.19116, "ce": 0.19116, "kl": 0.19116, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 624.6, "tok_per_s": 10695, "pass_tok_per_s": 10695, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1753}
{"step": 910, "total": 1250, "K": 1, "lr_bb": 2.711622705965592e-06, "grad_norm": 8.7102, "loss": 0.19501, "ce": 0.19501, "kl": 0.17395, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 631.7, "tok_per_s": 10694, "pass_tok_per_s": 10694, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1662}
{"step": 920, "total": 1250, "K": 1, "lr_bb": 2.619130192228383e-06, "grad_norm": 11.056, "loss": 0.12449, "ce": 0.12353, "kl": 0.12353, "rps": 0.00386, "l1": 0.0, "dkl": 0.0, "elapsed": 638.0, "tok_per_s": 10697, "pass_tok_per_s": 10697, "peak_gib": 61.4, "epoch": 1, "head_o": 0.16}
{"step": 930, "total": 1250, "K": 1, "lr_bb": 2.528655564309177e-06, "grad_norm": 11.7216, "loss": 0.31771, "ce": 0.31762, "kl": 0.26979, "rps": 0.00033, "l1": 0.0, "dkl": 0.0, "elapsed": 645.5, "tok_per_s": 10686, "pass_tok_per_s": 10686, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1591}
{"step": 940, "total": 1250, "K": 1, "lr_bb": 2.440262194550615e-06, "grad_norm": 22.8638, "loss": 0.36637, "ce": 0.36637, "kl": 0.36637, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 651.3, "tok_per_s": 10704, "pass_tok_per_s": 10704, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1573}
{"step": 950, "total": 1250, "K": 1, "lr_bb": 2.354011997492107e-06, "grad_norm": 9.585, "loss": 0.48367, "ce": 0.48367, "kl": 0.43025, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 657.8, "tok_per_s": 10703, "pass_tok_per_s": 10703, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1566}
{"step": 960, "total": 1250, "K": 1, "lr_bb": 2.2699653865022036e-06, "grad_norm": 8.9508, "loss": 0.15111, "ce": 0.15108, "kl": 0.14026, "rps": 0.00011, "l1": 0.0, "dkl": 0.0, "elapsed": 665.1, "tok_per_s": 10696, "pass_tok_per_s": 10696, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1595}
{"step": 970, "total": 1250, "K": 1, "lr_bb": 2.188181231462461e-06, "grad_norm": 7.4316, "loss": 0.13057, "ce": 0.13055, "kl": 0.09945, "rps": 9e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 672.1, "tok_per_s": 10698, "pass_tok_per_s": 10698, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1536}
{"step": 980, "total": 1250, "K": 1, "lr_bb": 2.108716817532436e-06, "grad_norm": 8.3285, "loss": 0.2658, "ce": 0.26574, "kl": 0.22986, "rps": 0.00022, "l1": 0.0, "dkl": 0.0, "elapsed": 679.7, "tok_per_s": 10691, "pass_tok_per_s": 10691, "peak_gib": 61.4, "epoch": 1, "head_o": 0.154}
{"step": 990, "total": 1250, "K": 1, "lr_bb": 2.031627805024673e-06, "grad_norm": 7.1209, "loss": 0.2136, "ce": 0.2136, "kl": 0.15432, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 687.0, "tok_per_s": 10687, "pass_tok_per_s": 10687, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1573}
{"step": 1000, "total": 1250, "K": 1, "lr_bb": 1.9569681904178194e-06, "grad_norm": 12.5336, "loss": 0.33262, "ce": 0.33262, "kl": 0.33262, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 693.1, "tok_per_s": 10692, "pass_tok_per_s": 10692, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1575}
{"event": "dev", "step": 1000, "seconds": 2.8, "gate": null, "K1": {"ce": 0.33756, "kl": 0.30554, "acc": 0.91053, "n": 1000}}
{"step": 1010, "total": 1250, "K": 1, "lr_bb": 1.8847902685351571e-06, "grad_norm": 6.8318, "loss": 0.17892, "ce": 0.17887, "kl": 0.08619, "rps": 0.00017, "l1": 0.0, "dkl": 0.0, "elapsed": 702.7, "tok_per_s": 10651, "pass_tok_per_s": 10651, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1589}
{"step": 1020, "total": 1250, "K": 1, "lr_bb": 1.8151445959150434e-06, "grad_norm": 15.6122, "loss": 0.20655, "ce": 0.20655, "kl": 0.15415, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 709.7, "tok_per_s": 10656, "pass_tok_per_s": 10656, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1583}
{"step": 1030, "total": 1250, "K": 1, "lr_bb": 1.7480799553989276e-06, "grad_norm": 24.9895, "loss": 0.35325, "ce": 0.35319, "kl": 0.34568, "rps": 0.00024, "l1": 0.0, "dkl": 0.0, "elapsed": 717.3, "tok_per_s": 10649, "pass_tok_per_s": 10649, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1598}
{"step": 1040, "total": 1250, "K": 1, "lr_bb": 1.683643321961735e-06, "grad_norm": 13.0689, "loss": 0.1932, "ce": 0.1932, "kl": 0.17216, "rps": 1e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 724.5, "tok_per_s": 10648, "pass_tok_per_s": 10648, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1593}
{"step": 1050, "total": 1250, "K": 1, "lr_bb": 1.6218798298085659e-06, "grad_norm": 7.4999, "loss": 0.34102, "ce": 0.34102, "kl": 0.34102, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 730.6, "tok_per_s": 10659, "pass_tok_per_s": 10659, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1592}
{"step": 1060, "total": 1250, "K": 1, "lr_bb": 1.5628327407607492e-06, "grad_norm": 18.4718, "loss": 0.48742, "ce": 0.48742, "kl": 0.48742, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 737.1, "tok_per_s": 10665, "pass_tok_per_s": 10665, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1593}
{"step": 1070, "total": 1250, "K": 1, "lr_bb": 1.5065434139533798e-06, "grad_norm": 7.5725, "loss": 0.15453, "ce": 0.15453, "kl": 0.15453, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 744.1, "tok_per_s": 10668, "pass_tok_per_s": 10668, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1574}
{"step": 1080, "total": 1250, "K": 1, "lr_bb": 1.453051276865607e-06, "grad_norm": 2.5731, "loss": 0.03481, "ce": 0.03458, "kl": 0.03458, "rps": 0.00091, "l1": 0.0, "dkl": 0.0, "elapsed": 749.8, "tok_per_s": 10679, "pass_tok_per_s": 10679, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1557}
{"step": 1090, "total": 1250, "K": 1, "lr_bb": 1.4023937977039125e-06, "grad_norm": 20.9868, "loss": 0.5372, "ce": 0.53697, "kl": 0.53697, "rps": 0.00091, "l1": 0.0, "dkl": 0.0, "elapsed": 756.0, "tok_per_s": 10689, "pass_tok_per_s": 10689, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1538}
{"step": 1100, "total": 1250, "K": 1, "lr_bb": 1.354606459157759e-06, "grad_norm": 16.393, "loss": 0.3175, "ce": 0.31743, "kl": 0.2506, "rps": 0.00026, "l1": 0.0, "dkl": 0.0, "elapsed": 762.6, "tok_per_s": 10687, "pass_tok_per_s": 10687, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1531}
{"step": 1110, "total": 1250, "K": 1, "lr_bb": 1.3097227335459846e-06, "grad_norm": 15.3346, "loss": 0.39494, "ce": 0.39488, "kl": 0.35328, "rps": 0.00025, "l1": 0.0, "dkl": 0.0, "elapsed": 769.2, "tok_per_s": 10688, "pass_tok_per_s": 10688, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1531}
{"step": 1120, "total": 1250, "K": 1, "lr_bb": 1.2677740593713373e-06, "grad_norm": 12.2471, "loss": 0.20244, "ce": 0.20243, "kl": 0.11193, "rps": 5e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 776.2, "tok_per_s": 10682, "pass_tok_per_s": 10682, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1537}
{"step": 1130, "total": 1250, "K": 1, "lr_bb": 1.2287898192995924e-06, "grad_norm": 14.0985, "loss": 0.48424, "ce": 0.48424, "kl": 0.48424, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 783.9, "tok_per_s": 10678, "pass_tok_per_s": 10678, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1542}
{"step": 1140, "total": 1250, "K": 1, "lr_bb": 1.1927973195786588e-06, "grad_norm": 9.343, "loss": 0.08696, "ce": 0.08696, "kl": 0.05413, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 789.4, "tok_per_s": 10697, "pass_tok_per_s": 10697, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1538}
{"step": 1150, "total": 1250, "K": 1, "lr_bb": 1.1598217709121034e-06, "grad_norm": 9.2996, "loss": 0.18147, "ce": 0.18145, "kl": 0.13528, "rps": 5e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 796.5, "tok_per_s": 10695, "pass_tok_per_s": 10695, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1538}
{"step": 1160, "total": 1250, "K": 1, "lr_bb": 1.1298862708004876e-06, "grad_norm": 12.6529, "loss": 0.18673, "ce": 0.18673, "kl": 0.16996, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 803.1, "tok_per_s": 10705, "pass_tok_per_s": 10705, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1537}
{"step": 1170, "total": 1250, "K": 1, "lr_bb": 1.1030117873628706e-06, "grad_norm": 11.5281, "loss": 0.2957, "ce": 0.2957, "kl": 0.27446, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 809.7, "tok_per_s": 10717, "pass_tok_per_s": 10717, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1536}
{"step": 1180, "total": 1250, "K": 1, "lr_bb": 1.07921714464984e-06, "grad_norm": 16.9586, "loss": 0.26815, "ce": 0.26815, "kl": 0.21505, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 815.5, "tok_per_s": 10730, "pass_tok_per_s": 10730, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1534}
{"step": 1190, "total": 1250, "K": 1, "lr_bb": 1.0585190094583284e-06, "grad_norm": 3.2164, "loss": 0.01123, "ce": 0.01121, "kl": 0.01121, "rps": 7e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 822.0, "tok_per_s": 10734, "pass_tok_per_s": 10734, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1533}
{"step": 1200, "total": 1250, "K": 1, "lr_bb": 1.040931879657468e-06, "grad_norm": 3.2353, "loss": 0.10115, "ce": 0.10115, "kl": 0.06276, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 828.5, "tok_per_s": 10736, "pass_tok_per_s": 10736, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1532}
{"step": 1210, "total": 1250, "K": 1, "lr_bb": 1.0264680740336586e-06, "grad_norm": 20.1945, "loss": 0.21819, "ce": 0.21819, "kl": 0.21819, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 835.0, "tok_per_s": 10743, "pass_tok_per_s": 10743, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1532}
{"step": 1220, "total": 1250, "K": 1, "lr_bb": 1.0151377236619533e-06, "grad_norm": 17.1059, "loss": 0.32904, "ce": 0.32904, "kl": 0.30867, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 841.9, "tok_per_s": 10740, "pass_tok_per_s": 10740, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1532}
{"step": 1230, "total": 1250, "K": 1, "lr_bb": 1.0069487648098174e-06, "grad_norm": 18.6411, "loss": 0.55828, "ce": 0.55828, "kl": 0.54443, "rps": 0.0, "l1": 0.0, "dkl": 0.0, "elapsed": 849.1, "tok_per_s": 10742, "pass_tok_per_s": 10742, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1533}
{"step": 1240, "total": 1250, "K": 1, "lr_bb": 1.0019069333782245e-06, "grad_norm": 20.0273, "loss": 0.26025, "ce": 0.26021, "kl": 0.19021, "rps": 0.00014, "l1": 0.0, "dkl": 0.0, "elapsed": 855.4, "tok_per_s": 10751, "pass_tok_per_s": 10751, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1533}
{"step": 1250, "total": 1250, "K": 1, "lr_bb": 1.0000157608839794e-06, "grad_norm": 7.1242, "loss": 0.05296, "ce": 0.05296, "kl": 0.03259, "rps": 2e-05, "l1": 0.0, "dkl": 0.0, "elapsed": 861.4, "tok_per_s": 10760, "pass_tok_per_s": 10760, "peak_gib": 61.4, "epoch": 1, "head_o": 0.1533}
{"event": "dev", "step": 1250, "seconds": 2.8, "gate": null, "K1": {"ce": 0.33499, "kl": 0.30298, "acc": 0.91474, "n": 1000}}
{"event": "finished", "step": 1250, "elapsed": 864.2, "oom_retries": 0, "peak_gib": 61.4}
{"event": "saved", "path": "runs/e1/model.safetensors"}

Xet Storage Details

Size:
37 kB
·
Xet hash:
ad0ed1e672dab5e6507e1a7567067487a7639f50b8c65570c7b80f47a2a902b2

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