| { | |
| "best_metric": 0.7979517910648003, | |
| "best_model_checkpoint": "checkpoint/cross_encoder_20250522_full_data/checkpoint-16219", | |
| "epoch": 7.0, | |
| "eval_steps": 500, | |
| "global_step": 16219, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "grad_norm": 0.9638161659240723, | |
| "learning_rate": 4.315925766076824e-09, | |
| "loss": 0.6827, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "grad_norm": 2.8890857696533203, | |
| "learning_rate": 8.631851532153649e-07, | |
| "loss": 0.6745, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "grad_norm": 2.8610997200012207, | |
| "learning_rate": 1.7263703064307298e-06, | |
| "loss": 0.5921, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 4.058478832244873, | |
| "learning_rate": 2.5895554596460943e-06, | |
| "loss": 0.5238, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "grad_norm": 2.51100492477417, | |
| "learning_rate": 3.4527406128614595e-06, | |
| "loss": 0.4943, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "grad_norm": 3.381241798400879, | |
| "learning_rate": 4.3159257660768235e-06, | |
| "loss": 0.4882, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "grad_norm": 3.5782713890075684, | |
| "learning_rate": 5.179110919292189e-06, | |
| "loss": 0.4782, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "grad_norm": 5.436559200286865, | |
| "learning_rate": 6.042296072507553e-06, | |
| "loss": 0.4719, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "grad_norm": 2.6600098609924316, | |
| "learning_rate": 6.905481225722919e-06, | |
| "loss": 0.4593, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "grad_norm": 4.711264610290527, | |
| "learning_rate": 7.768666378938283e-06, | |
| "loss": 0.4629, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "grad_norm": 7.427570819854736, | |
| "learning_rate": 8.631851532153647e-06, | |
| "loss": 0.4549, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "grad_norm": 2.386420965194702, | |
| "learning_rate": 9.495036685369013e-06, | |
| "loss": 0.4572, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_f2": 0.7404627946070191, | |
| "eval_loss": 0.4705376923084259, | |
| "eval_precision": 0.47348360206208184, | |
| "eval_recall": 0.8619708466453674, | |
| "eval_runtime": 147.2641, | |
| "eval_samples_per_second": 816.166, | |
| "eval_steps_per_second": 6.376, | |
| "step": 2317 | |
| }, | |
| { | |
| "epoch": 1.04, | |
| "grad_norm": 3.0445728302001953, | |
| "learning_rate": 9.999609111230013e-06, | |
| "loss": 0.4511, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 1.12, | |
| "grad_norm": 2.2899084091186523, | |
| "learning_rate": 9.995456298935638e-06, | |
| "loss": 0.447, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "grad_norm": 4.465543270111084, | |
| "learning_rate": 9.986768615435655e-06, | |
| "loss": 0.4405, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 1.29, | |
| "grad_norm": 4.2559404373168945, | |
| "learning_rate": 9.973553947402149e-06, | |
| "loss": 0.4436, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.38, | |
| "grad_norm": 4.245833873748779, | |
| "learning_rate": 9.955824291100119e-06, | |
| "loss": 0.4393, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 1.47, | |
| "grad_norm": 4.579006671905518, | |
| "learning_rate": 9.933595741497281e-06, | |
| "loss": 0.4376, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 1.55, | |
| "grad_norm": 7.389001369476318, | |
| "learning_rate": 9.906888477653065e-06, | |
| "loss": 0.4373, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 1.64, | |
| "grad_norm": 4.0490312576293945, | |
| "learning_rate": 9.875726744400081e-06, | |
| "loss": 0.4345, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 1.73, | |
| "grad_norm": 2.065431594848633, | |
| "learning_rate": 9.840138830334685e-06, | |
| "loss": 0.4327, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.81, | |
| "grad_norm": 3.0262961387634277, | |
| "learning_rate": 9.800157042136608e-06, | |
| "loss": 0.4327, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 1.9, | |
| "grad_norm": 2.6107230186462402, | |
| "learning_rate": 9.755817675240981e-06, | |
| "loss": 0.4294, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 1.99, | |
| "grad_norm": 4.621052265167236, | |
| "learning_rate": 9.707160980889367e-06, | |
| "loss": 0.4283, | |
| "step": 4600 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_f2": 0.7717545519035234, | |
| "eval_loss": 0.4515298902988434, | |
| "eval_precision": 0.4774057047330478, | |
| "eval_recall": 0.9123901757188498, | |
| "eval_runtime": 148.5197, | |
| "eval_samples_per_second": 809.266, | |
| "eval_steps_per_second": 6.322, | |
| "step": 4634 | |
| }, | |
| { | |
| "epoch": 2.07, | |
| "grad_norm": 3.324246644973755, | |
| "learning_rate": 9.65423112958972e-06, | |
| "loss": 0.4244, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 2.16, | |
| "grad_norm": 4.125003337860107, | |
| "learning_rate": 9.597076171018426e-06, | |
| "loss": 0.4183, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 2.24, | |
| "grad_norm": 2.1959779262542725, | |
| "learning_rate": 9.535747990400856e-06, | |
| "loss": 0.4164, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 2.33, | |
| "grad_norm": 1.7736337184906006, | |
| "learning_rate": 9.470302261409985e-06, | |
| "loss": 0.4205, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 2.42, | |
| "grad_norm": 4.9517903327941895, | |
| "learning_rate": 9.400798395625894e-06, | |
| "loss": 0.4176, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "grad_norm": 6.079417705535889, | |
| "learning_rate": 9.327299488601976e-06, | |
| "loss": 0.4202, | |
| "step": 5800 | |
| }, | |
| { | |
| "epoch": 2.59, | |
| "grad_norm": 1.3246735334396362, | |
| "learning_rate": 9.249872262586839e-06, | |
| "loss": 0.4218, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 2.68, | |
| "grad_norm": 2.6098692417144775, | |
| "learning_rate": 9.168587005953913e-06, | |
| "loss": 0.4182, | |
| "step": 6200 | |
| }, | |
| { | |
| "epoch": 2.76, | |
| "grad_norm": 3.9474799633026123, | |
| "learning_rate": 9.083517509393716e-06, | |
| "loss": 0.415, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 2.85, | |
| "grad_norm": 3.5382373332977295, | |
| "learning_rate": 8.994740998926724e-06, | |
| "loss": 0.4186, | |
| "step": 6600 | |
| }, | |
| { | |
| "epoch": 2.93, | |
| "grad_norm": 2.33327054977417, | |
| "learning_rate": 8.902338065797648e-06, | |
| "loss": 0.4115, | |
| "step": 6800 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_f2": 0.7773410482632972, | |
| "eval_loss": 0.4485355019569397, | |
| "eval_precision": 0.47962529274004684, | |
| "eval_recall": 0.9201277955271565, | |
| "eval_runtime": 147.9088, | |
| "eval_samples_per_second": 812.609, | |
| "eval_steps_per_second": 6.349, | |
| "step": 6951 | |
| }, | |
| { | |
| "epoch": 3.02, | |
| "grad_norm": 3.269026517868042, | |
| "learning_rate": 8.806392593314781e-06, | |
| "loss": 0.4167, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 3.11, | |
| "grad_norm": 2.844139575958252, | |
| "learning_rate": 8.70699168070078e-06, | |
| "loss": 0.4101, | |
| "step": 7200 | |
| }, | |
| { | |
| "epoch": 3.19, | |
| "grad_norm": 2.2456936836242676, | |
| "learning_rate": 8.604225564024074e-06, | |
| "loss": 0.4095, | |
| "step": 7400 | |
| }, | |
| { | |
| "epoch": 3.28, | |
| "grad_norm": 2.3910915851593018, | |
| "learning_rate": 8.498187534282632e-06, | |
| "loss": 0.407, | |
| "step": 7600 | |
| }, | |
| { | |
| "epoch": 3.37, | |
| "grad_norm": 2.2186901569366455, | |
| "learning_rate": 8.388973852714463e-06, | |
| "loss": 0.4044, | |
| "step": 7800 | |
| }, | |
| { | |
| "epoch": 3.45, | |
| "grad_norm": 3.1750833988189697, | |
| "learning_rate": 8.276683663411758e-06, | |
| "loss": 0.4048, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 3.54, | |
| "grad_norm": 5.709492206573486, | |
| "learning_rate": 8.161418903317936e-06, | |
| "loss": 0.4061, | |
| "step": 8200 | |
| }, | |
| { | |
| "epoch": 3.63, | |
| "grad_norm": 6.942538738250732, | |
| "learning_rate": 8.043284209689402e-06, | |
| "loss": 0.4067, | |
| "step": 8400 | |
| }, | |
| { | |
| "epoch": 3.71, | |
| "grad_norm": 1.3552676439285278, | |
| "learning_rate": 7.922386825105899e-06, | |
| "loss": 0.4046, | |
| "step": 8600 | |
| }, | |
| { | |
| "epoch": 3.8, | |
| "grad_norm": 3.5627899169921875, | |
| "learning_rate": 7.798836500115803e-06, | |
| "loss": 0.407, | |
| "step": 8800 | |
| }, | |
| { | |
| "epoch": 3.88, | |
| "grad_norm": 2.0421602725982666, | |
| "learning_rate": 7.672745393604649e-06, | |
| "loss": 0.4037, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 3.97, | |
| "grad_norm": 4.619537830352783, | |
| "learning_rate": 7.544227970977395e-06, | |
| "loss": 0.4021, | |
| "step": 9200 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_f2": 0.7901536373601518, | |
| "eval_loss": 0.4387025237083435, | |
| "eval_precision": 0.5217129071170085, | |
| "eval_recall": 0.9067991214057508, | |
| "eval_runtime": 148.7755, | |
| "eval_samples_per_second": 807.875, | |
| "eval_steps_per_second": 6.312, | |
| "step": 9268 | |
| }, | |
| { | |
| "epoch": 4.06, | |
| "grad_norm": 1.7471312284469604, | |
| "learning_rate": 7.413400900246815e-06, | |
| "loss": 0.3959, | |
| "step": 9400 | |
| }, | |
| { | |
| "epoch": 4.14, | |
| "grad_norm": 7.655578136444092, | |
| "learning_rate": 7.280382946122369e-06, | |
| "loss": 0.3996, | |
| "step": 9600 | |
| }, | |
| { | |
| "epoch": 4.23, | |
| "grad_norm": 4.281186580657959, | |
| "learning_rate": 7.1452948621957e-06, | |
| "loss": 0.3943, | |
| "step": 9800 | |
| }, | |
| { | |
| "epoch": 4.32, | |
| "grad_norm": 3.821415424346924, | |
| "learning_rate": 7.0082592813206175e-06, | |
| "loss": 0.3972, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 4.4, | |
| "grad_norm": 2.6942179203033447, | |
| "learning_rate": 6.869400604287093e-06, | |
| "loss": 0.3939, | |
| "step": 10200 | |
| }, | |
| { | |
| "epoch": 4.49, | |
| "grad_norm": 4.14546012878418, | |
| "learning_rate": 6.7288448868903225e-06, | |
| "loss": 0.3997, | |
| "step": 10400 | |
| }, | |
| { | |
| "epoch": 4.57, | |
| "grad_norm": 2.8214612007141113, | |
| "learning_rate": 6.586719725497375e-06, | |
| "loss": 0.3993, | |
| "step": 10600 | |
| }, | |
| { | |
| "epoch": 4.66, | |
| "grad_norm": 2.5755832195281982, | |
| "learning_rate": 6.443154141215318e-06, | |
| "loss": 0.3969, | |
| "step": 10800 | |
| }, | |
| { | |
| "epoch": 4.75, | |
| "grad_norm": 1.6153743267059326, | |
| "learning_rate": 6.298278462765959e-06, | |
| "loss": 0.3987, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 4.83, | |
| "grad_norm": 7.3942766189575195, | |
| "learning_rate": 6.152224208173533e-06, | |
| "loss": 0.3946, | |
| "step": 11200 | |
| }, | |
| { | |
| "epoch": 4.92, | |
| "grad_norm": 1.083901047706604, | |
| "learning_rate": 6.005123965372751e-06, | |
| "loss": 0.3918, | |
| "step": 11400 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_f2": 0.7704512576415139, | |
| "eval_loss": 0.44662219285964966, | |
| "eval_precision": 0.6110967168819632, | |
| "eval_recall": 0.8241813099041534, | |
| "eval_runtime": 149.5247, | |
| "eval_samples_per_second": 803.827, | |
| "eval_steps_per_second": 6.28, | |
| "step": 11585 | |
| }, | |
| { | |
| "epoch": 5.01, | |
| "grad_norm": 3.2887141704559326, | |
| "learning_rate": 5.857111271845573e-06, | |
| "loss": 0.3935, | |
| "step": 11600 | |
| }, | |
| { | |
| "epoch": 5.09, | |
| "grad_norm": 2.212423801422119, | |
| "learning_rate": 5.708320493395999e-06, | |
| "loss": 0.3908, | |
| "step": 11800 | |
| }, | |
| { | |
| "epoch": 5.18, | |
| "grad_norm": 0.9465045928955078, | |
| "learning_rate": 5.558886702172891e-06, | |
| "loss": 0.3854, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 5.27, | |
| "grad_norm": 5.766241073608398, | |
| "learning_rate": 5.408945554051591e-06, | |
| "loss": 0.3912, | |
| "step": 12200 | |
| }, | |
| { | |
| "epoch": 5.35, | |
| "grad_norm": 4.334184646606445, | |
| "learning_rate": 5.258633165485625e-06, | |
| "loss": 0.3866, | |
| "step": 12400 | |
| }, | |
| { | |
| "epoch": 5.44, | |
| "grad_norm": 4.286846160888672, | |
| "learning_rate": 5.108085989940292e-06, | |
| "loss": 0.3919, | |
| "step": 12600 | |
| }, | |
| { | |
| "epoch": 5.52, | |
| "grad_norm": 1.167948603630066, | |
| "learning_rate": 4.95744069402033e-06, | |
| "loss": 0.3898, | |
| "step": 12800 | |
| }, | |
| { | |
| "epoch": 5.61, | |
| "grad_norm": 4.312135696411133, | |
| "learning_rate": 4.806834033404065e-06, | |
| "loss": 0.3894, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 5.7, | |
| "grad_norm": 2.726715564727783, | |
| "learning_rate": 4.6564027286967275e-06, | |
| "loss": 0.3918, | |
| "step": 13200 | |
| }, | |
| { | |
| "epoch": 5.78, | |
| "grad_norm": 2.4171202182769775, | |
| "learning_rate": 4.5062833413156e-06, | |
| "loss": 0.3917, | |
| "step": 13400 | |
| }, | |
| { | |
| "epoch": 5.87, | |
| "grad_norm": 5.957132339477539, | |
| "learning_rate": 4.3566121495196656e-06, | |
| "loss": 0.3849, | |
| "step": 13600 | |
| }, | |
| { | |
| "epoch": 5.96, | |
| "grad_norm": 4.632606029510498, | |
| "learning_rate": 4.20752502469631e-06, | |
| "loss": 0.3879, | |
| "step": 13800 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_f2": 0.7946750167423844, | |
| "eval_loss": 0.43369919061660767, | |
| "eval_precision": 0.5782680276588739, | |
| "eval_recall": 0.8766972843450479, | |
| "eval_runtime": 148.7395, | |
| "eval_samples_per_second": 808.07, | |
| "eval_steps_per_second": 6.313, | |
| "step": 13902 | |
| }, | |
| { | |
| "epoch": 6.04, | |
| "grad_norm": 1.9424314498901367, | |
| "learning_rate": 4.0591573080173995e-06, | |
| "loss": 0.3873, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 6.13, | |
| "grad_norm": 2.831543445587158, | |
| "learning_rate": 3.911643687576664e-06, | |
| "loss": 0.3807, | |
| "step": 14200 | |
| }, | |
| { | |
| "epoch": 6.21, | |
| "grad_norm": 2.7500486373901367, | |
| "learning_rate": 3.7651180761199505e-06, | |
| "loss": 0.3801, | |
| "step": 14400 | |
| }, | |
| { | |
| "epoch": 6.3, | |
| "grad_norm": 4.015429496765137, | |
| "learning_rate": 3.619713489479354e-06, | |
| "loss": 0.3824, | |
| "step": 14600 | |
| }, | |
| { | |
| "epoch": 6.39, | |
| "grad_norm": 2.4831278324127197, | |
| "learning_rate": 3.4755619258215407e-06, | |
| "loss": 0.3808, | |
| "step": 14800 | |
| }, | |
| { | |
| "epoch": 6.47, | |
| "grad_norm": 7.442523002624512, | |
| "learning_rate": 3.3327942458199193e-06, | |
| "loss": 0.385, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 6.56, | |
| "grad_norm": 4.242701530456543, | |
| "learning_rate": 3.1915400538594333e-06, | |
| "loss": 0.3832, | |
| "step": 15200 | |
| }, | |
| { | |
| "epoch": 6.65, | |
| "grad_norm": 5.000258922576904, | |
| "learning_rate": 3.0519275803818014e-06, | |
| "loss": 0.3805, | |
| "step": 15400 | |
| }, | |
| { | |
| "epoch": 6.73, | |
| "grad_norm": 2.0560362339019775, | |
| "learning_rate": 2.914083565478024e-06, | |
| "loss": 0.381, | |
| "step": 15600 | |
| }, | |
| { | |
| "epoch": 6.82, | |
| "grad_norm": 1.8777313232421875, | |
| "learning_rate": 2.7781331438338317e-06, | |
| "loss": 0.3831, | |
| "step": 15800 | |
| }, | |
| { | |
| "epoch": 6.91, | |
| "grad_norm": 1.0822844505310059, | |
| "learning_rate": 2.6441997311325196e-06, | |
| "loss": 0.3797, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 6.99, | |
| "grad_norm": 3.2328083515167236, | |
| "learning_rate": 2.5124049120182916e-06, | |
| "loss": 0.383, | |
| "step": 16200 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_f2": 0.7979517910648003, | |
| "eval_loss": 0.4335618019104004, | |
| "eval_precision": 0.5632793509486378, | |
| "eval_recall": 0.8907248402555911, | |
| "eval_runtime": 148.5629, | |
| "eval_samples_per_second": 809.031, | |
| "eval_steps_per_second": 6.321, | |
| "step": 16219 | |
| } | |
| ], | |
| "logging_steps": 200, | |
| "max_steps": 23170, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "total_flos": 5.462269789659464e+17, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |