| { | |
| "best_metric": 0.4078540802001953, | |
| "best_model_checkpoint": "mikhail-panzo/fil_b128_le4_s8000/checkpoint-1000", | |
| "epoch": 353.5911602209945, | |
| "eval_steps": 500, | |
| "global_step": 4000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 4.419889502762431, | |
| "grad_norm": 2.1200406551361084, | |
| "learning_rate": 2.5e-06, | |
| "loss": 0.77, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 8.839779005524862, | |
| "grad_norm": 0.8919622302055359, | |
| "learning_rate": 5e-06, | |
| "loss": 0.695, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 13.259668508287293, | |
| "grad_norm": 0.9864558577537537, | |
| "learning_rate": 7.5e-06, | |
| "loss": 0.6278, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 17.679558011049725, | |
| "grad_norm": 0.9195986986160278, | |
| "learning_rate": 1e-05, | |
| "loss": 0.5431, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 22.099447513812155, | |
| "grad_norm": 1.0175281763076782, | |
| "learning_rate": 1.25e-05, | |
| "loss": 0.5155, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 26.519337016574585, | |
| "grad_norm": 1.0303566455841064, | |
| "learning_rate": 1.5e-05, | |
| "loss": 0.502, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 30.939226519337016, | |
| "grad_norm": 1.005897879600525, | |
| "learning_rate": 1.75e-05, | |
| "loss": 0.4889, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 35.35911602209945, | |
| "grad_norm": 1.0381360054016113, | |
| "learning_rate": 2e-05, | |
| "loss": 0.477, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 39.77900552486188, | |
| "grad_norm": 1.2148911952972412, | |
| "learning_rate": 2.25e-05, | |
| "loss": 0.4711, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 44.19889502762431, | |
| "grad_norm": 1.5404785871505737, | |
| "learning_rate": 2.5e-05, | |
| "loss": 0.463, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 44.19889502762431, | |
| "eval_loss": 0.42523157596588135, | |
| "eval_runtime": 15.6706, | |
| "eval_samples_per_second": 10.274, | |
| "eval_steps_per_second": 1.34, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 48.61878453038674, | |
| "grad_norm": 1.1097612380981445, | |
| "learning_rate": 2.7500000000000004e-05, | |
| "loss": 0.4601, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 53.03867403314917, | |
| "grad_norm": 1.2796109914779663, | |
| "learning_rate": 3e-05, | |
| "loss": 0.454, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 57.4585635359116, | |
| "grad_norm": 1.0656301975250244, | |
| "learning_rate": 3.2500000000000004e-05, | |
| "loss": 0.4498, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 61.87845303867403, | |
| "grad_norm": 1.3971511125564575, | |
| "learning_rate": 3.5e-05, | |
| "loss": 0.4439, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 66.29834254143647, | |
| "grad_norm": 1.9344899654388428, | |
| "learning_rate": 3.7500000000000003e-05, | |
| "loss": 0.4418, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 70.7182320441989, | |
| "grad_norm": 1.080334186553955, | |
| "learning_rate": 4e-05, | |
| "loss": 0.4379, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 75.13812154696133, | |
| "grad_norm": 1.9891581535339355, | |
| "learning_rate": 4.25e-05, | |
| "loss": 0.4334, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 79.55801104972376, | |
| "grad_norm": 1.3854402303695679, | |
| "learning_rate": 4.5e-05, | |
| "loss": 0.4331, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 83.97790055248619, | |
| "grad_norm": 1.5818183422088623, | |
| "learning_rate": 4.75e-05, | |
| "loss": 0.428, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 88.39779005524862, | |
| "grad_norm": 1.8474106788635254, | |
| "learning_rate": 5e-05, | |
| "loss": 0.4238, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 88.39779005524862, | |
| "eval_loss": 0.4078540802001953, | |
| "eval_runtime": 14.8436, | |
| "eval_samples_per_second": 10.846, | |
| "eval_steps_per_second": 1.415, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 92.81767955801105, | |
| "grad_norm": 3.1851751804351807, | |
| "learning_rate": 5.25e-05, | |
| "loss": 0.4234, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 97.23756906077348, | |
| "grad_norm": 1.3754140138626099, | |
| "learning_rate": 5.500000000000001e-05, | |
| "loss": 0.4209, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 101.65745856353591, | |
| "grad_norm": 1.6143475770950317, | |
| "learning_rate": 5.7499999999999995e-05, | |
| "loss": 0.4228, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 106.07734806629834, | |
| "grad_norm": 1.9961720705032349, | |
| "learning_rate": 6e-05, | |
| "loss": 0.4162, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 110.49723756906077, | |
| "grad_norm": 1.9125347137451172, | |
| "learning_rate": 6.25e-05, | |
| "loss": 0.4178, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 114.9171270718232, | |
| "grad_norm": 3.1913959980010986, | |
| "learning_rate": 6.500000000000001e-05, | |
| "loss": 0.4143, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 119.33701657458563, | |
| "grad_norm": 2.1010069847106934, | |
| "learning_rate": 6.750000000000001e-05, | |
| "loss": 0.4118, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 123.75690607734806, | |
| "grad_norm": 1.3079431056976318, | |
| "learning_rate": 7e-05, | |
| "loss": 0.4091, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 128.1767955801105, | |
| "grad_norm": 2.3663132190704346, | |
| "learning_rate": 7.25e-05, | |
| "loss": 0.4092, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 132.59668508287294, | |
| "grad_norm": 1.031698226928711, | |
| "learning_rate": 7.500000000000001e-05, | |
| "loss": 0.4083, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 132.59668508287294, | |
| "eval_loss": 0.4082767069339752, | |
| "eval_runtime": 15.1068, | |
| "eval_samples_per_second": 10.657, | |
| "eval_steps_per_second": 1.39, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 137.01657458563537, | |
| "grad_norm": 2.7781426906585693, | |
| "learning_rate": 7.75e-05, | |
| "loss": 0.4079, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 141.4364640883978, | |
| "grad_norm": 1.136169672012329, | |
| "learning_rate": 8e-05, | |
| "loss": 0.4028, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 145.85635359116023, | |
| "grad_norm": 1.7299158573150635, | |
| "learning_rate": 8.25e-05, | |
| "loss": 0.4034, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 150.27624309392266, | |
| "grad_norm": 2.4281983375549316, | |
| "learning_rate": 8.5e-05, | |
| "loss": 0.3992, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 154.6961325966851, | |
| "grad_norm": 3.6834373474121094, | |
| "learning_rate": 8.75e-05, | |
| "loss": 0.3997, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 159.11602209944752, | |
| "grad_norm": 3.1484293937683105, | |
| "learning_rate": 9e-05, | |
| "loss": 0.396, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 163.53591160220995, | |
| "grad_norm": 3.978259563446045, | |
| "learning_rate": 9.250000000000001e-05, | |
| "loss": 0.3969, | |
| "step": 1850 | |
| }, | |
| { | |
| "epoch": 167.95580110497238, | |
| "grad_norm": 2.906266212463379, | |
| "learning_rate": 9.5e-05, | |
| "loss": 0.4017, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 172.3756906077348, | |
| "grad_norm": 1.1569880247116089, | |
| "learning_rate": 9.75e-05, | |
| "loss": 0.396, | |
| "step": 1950 | |
| }, | |
| { | |
| "epoch": 176.79558011049724, | |
| "grad_norm": 2.2872982025146484, | |
| "learning_rate": 0.0001, | |
| "loss": 0.3928, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 176.79558011049724, | |
| "eval_loss": 0.414584219455719, | |
| "eval_runtime": 15.2895, | |
| "eval_samples_per_second": 10.53, | |
| "eval_steps_per_second": 1.373, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 181.21546961325967, | |
| "grad_norm": 1.1768990755081177, | |
| "learning_rate": 9.916666666666667e-05, | |
| "loss": 0.3928, | |
| "step": 2050 | |
| }, | |
| { | |
| "epoch": 185.6353591160221, | |
| "grad_norm": 1.2087491750717163, | |
| "learning_rate": 9.833333333333333e-05, | |
| "loss": 0.3901, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 190.05524861878453, | |
| "grad_norm": 1.17588210105896, | |
| "learning_rate": 9.75e-05, | |
| "loss": 0.3877, | |
| "step": 2150 | |
| }, | |
| { | |
| "epoch": 194.47513812154696, | |
| "grad_norm": 1.5979018211364746, | |
| "learning_rate": 9.666666666666667e-05, | |
| "loss": 0.3879, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 198.8950276243094, | |
| "grad_norm": 1.0489567518234253, | |
| "learning_rate": 9.583333333333334e-05, | |
| "loss": 0.3869, | |
| "step": 2250 | |
| }, | |
| { | |
| "epoch": 203.31491712707182, | |
| "grad_norm": 1.833038091659546, | |
| "learning_rate": 9.501666666666668e-05, | |
| "loss": 0.3859, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 207.73480662983425, | |
| "grad_norm": 2.075849771499634, | |
| "learning_rate": 9.418333333333334e-05, | |
| "loss": 0.3843, | |
| "step": 2350 | |
| }, | |
| { | |
| "epoch": 212.15469613259668, | |
| "grad_norm": 2.8613317012786865, | |
| "learning_rate": 9.335e-05, | |
| "loss": 0.381, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 216.5745856353591, | |
| "grad_norm": 1.6291433572769165, | |
| "learning_rate": 9.251666666666667e-05, | |
| "loss": 0.3802, | |
| "step": 2450 | |
| }, | |
| { | |
| "epoch": 220.99447513812154, | |
| "grad_norm": 1.1548006534576416, | |
| "learning_rate": 9.168333333333333e-05, | |
| "loss": 0.381, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 220.99447513812154, | |
| "eval_loss": 0.4127160906791687, | |
| "eval_runtime": 15.1026, | |
| "eval_samples_per_second": 10.66, | |
| "eval_steps_per_second": 1.39, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 225.41436464088397, | |
| "grad_norm": 1.3956973552703857, | |
| "learning_rate": 9.085e-05, | |
| "loss": 0.3795, | |
| "step": 2550 | |
| }, | |
| { | |
| "epoch": 229.8342541436464, | |
| "grad_norm": 1.6613599061965942, | |
| "learning_rate": 9.001666666666667e-05, | |
| "loss": 0.3757, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 234.25414364640883, | |
| "grad_norm": 1.4951831102371216, | |
| "learning_rate": 8.918333333333334e-05, | |
| "loss": 0.3763, | |
| "step": 2650 | |
| }, | |
| { | |
| "epoch": 238.67403314917127, | |
| "grad_norm": 3.0197391510009766, | |
| "learning_rate": 8.834999999999999e-05, | |
| "loss": 0.3765, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 243.0939226519337, | |
| "grad_norm": 0.9016386866569519, | |
| "learning_rate": 8.751666666666668e-05, | |
| "loss": 0.3724, | |
| "step": 2750 | |
| }, | |
| { | |
| "epoch": 247.51381215469613, | |
| "grad_norm": 1.6974703073501587, | |
| "learning_rate": 8.668333333333334e-05, | |
| "loss": 0.3729, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 251.93370165745856, | |
| "grad_norm": 2.8571081161499023, | |
| "learning_rate": 8.585000000000001e-05, | |
| "loss": 0.3715, | |
| "step": 2850 | |
| }, | |
| { | |
| "epoch": 256.353591160221, | |
| "grad_norm": 2.8940412998199463, | |
| "learning_rate": 8.501666666666667e-05, | |
| "loss": 0.3716, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 260.77348066298345, | |
| "grad_norm": 1.1075533628463745, | |
| "learning_rate": 8.418333333333334e-05, | |
| "loss": 0.3691, | |
| "step": 2950 | |
| }, | |
| { | |
| "epoch": 265.1933701657459, | |
| "grad_norm": 1.145349383354187, | |
| "learning_rate": 8.335e-05, | |
| "loss": 0.3712, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 265.1933701657459, | |
| "eval_loss": 0.40815719962120056, | |
| "eval_runtime": 15.0404, | |
| "eval_samples_per_second": 10.705, | |
| "eval_steps_per_second": 1.396, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 269.6132596685083, | |
| "grad_norm": 0.788912832736969, | |
| "learning_rate": 8.251666666666668e-05, | |
| "loss": 0.3691, | |
| "step": 3050 | |
| }, | |
| { | |
| "epoch": 274.03314917127074, | |
| "grad_norm": 1.5521596670150757, | |
| "learning_rate": 8.168333333333333e-05, | |
| "loss": 0.3687, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 278.45303867403317, | |
| "grad_norm": 1.2638827562332153, | |
| "learning_rate": 8.085e-05, | |
| "loss": 0.3681, | |
| "step": 3150 | |
| }, | |
| { | |
| "epoch": 282.8729281767956, | |
| "grad_norm": 1.6849925518035889, | |
| "learning_rate": 8.001666666666667e-05, | |
| "loss": 0.3664, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 287.292817679558, | |
| "grad_norm": 1.9946844577789307, | |
| "learning_rate": 7.918333333333334e-05, | |
| "loss": 0.3671, | |
| "step": 3250 | |
| }, | |
| { | |
| "epoch": 291.71270718232046, | |
| "grad_norm": 1.3560407161712646, | |
| "learning_rate": 7.835000000000001e-05, | |
| "loss": 0.3631, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 296.1325966850829, | |
| "grad_norm": 1.4300700426101685, | |
| "learning_rate": 7.751666666666668e-05, | |
| "loss": 0.3616, | |
| "step": 3350 | |
| }, | |
| { | |
| "epoch": 300.5524861878453, | |
| "grad_norm": 1.0440477132797241, | |
| "learning_rate": 7.668333333333335e-05, | |
| "loss": 0.3627, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 304.97237569060775, | |
| "grad_norm": 0.78243088722229, | |
| "learning_rate": 7.585e-05, | |
| "loss": 0.3624, | |
| "step": 3450 | |
| }, | |
| { | |
| "epoch": 309.3922651933702, | |
| "grad_norm": 1.4992115497589111, | |
| "learning_rate": 7.501666666666667e-05, | |
| "loss": 0.3589, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 309.3922651933702, | |
| "eval_loss": 0.4111907482147217, | |
| "eval_runtime": 15.163, | |
| "eval_samples_per_second": 10.618, | |
| "eval_steps_per_second": 1.385, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 313.8121546961326, | |
| "grad_norm": 1.0604934692382812, | |
| "learning_rate": 7.418333333333334e-05, | |
| "loss": 0.3599, | |
| "step": 3550 | |
| }, | |
| { | |
| "epoch": 318.23204419889504, | |
| "grad_norm": 0.8896504640579224, | |
| "learning_rate": 7.335000000000001e-05, | |
| "loss": 0.361, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 322.65193370165747, | |
| "grad_norm": 1.3247913122177124, | |
| "learning_rate": 7.251666666666666e-05, | |
| "loss": 0.3575, | |
| "step": 3650 | |
| }, | |
| { | |
| "epoch": 327.0718232044199, | |
| "grad_norm": 0.6475388407707214, | |
| "learning_rate": 7.168333333333333e-05, | |
| "loss": 0.3575, | |
| "step": 3700 | |
| }, | |
| { | |
| "epoch": 331.49171270718233, | |
| "grad_norm": 0.6724014282226562, | |
| "learning_rate": 7.085e-05, | |
| "loss": 0.3558, | |
| "step": 3750 | |
| }, | |
| { | |
| "epoch": 335.91160220994476, | |
| "grad_norm": 1.0705429315567017, | |
| "learning_rate": 7.001666666666667e-05, | |
| "loss": 0.3564, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 340.3314917127072, | |
| "grad_norm": 1.0692503452301025, | |
| "learning_rate": 6.918333333333334e-05, | |
| "loss": 0.3601, | |
| "step": 3850 | |
| }, | |
| { | |
| "epoch": 344.7513812154696, | |
| "grad_norm": 0.8828334212303162, | |
| "learning_rate": 6.835000000000001e-05, | |
| "loss": 0.3548, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 349.17127071823205, | |
| "grad_norm": 0.6390587091445923, | |
| "learning_rate": 6.751666666666668e-05, | |
| "loss": 0.3529, | |
| "step": 3950 | |
| }, | |
| { | |
| "epoch": 353.5911602209945, | |
| "grad_norm": 0.7502044439315796, | |
| "learning_rate": 6.668333333333333e-05, | |
| "loss": 0.3544, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 353.5911602209945, | |
| "eval_loss": 0.4128935635089874, | |
| "eval_runtime": 16.0241, | |
| "eval_samples_per_second": 10.047, | |
| "eval_steps_per_second": 1.311, | |
| "step": 4000 | |
| } | |
| ], | |
| "logging_steps": 50, | |
| "max_steps": 8000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 728, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 9.057224003438318e+16, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |