| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 9.562841530054644, | |
| "eval_steps": 500, | |
| "global_step": 28000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.17, | |
| "grad_norm": 1.2606134414672852, | |
| "learning_rate": 6.25e-05, | |
| "loss": 6.1193, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "grad_norm": 1.0857479572296143, | |
| "learning_rate": 0.000125, | |
| "loss": 3.6291, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "grad_norm": 0.9009315967559814, | |
| "learning_rate": 0.0001875, | |
| "loss": 3.3499, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "grad_norm": 1.0317490100860596, | |
| "learning_rate": 0.00025, | |
| "loss": 3.1662, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "grad_norm": 0.9014867544174194, | |
| "learning_rate": 0.0003125, | |
| "loss": 3.0436, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.43863580178730166, | |
| "eval_loss": 3.012347936630249, | |
| "eval_runtime": 3.3529, | |
| "eval_samples_per_second": 1339.456, | |
| "eval_steps_per_second": 10.737, | |
| "step": 2928 | |
| }, | |
| { | |
| "epoch": 1.02, | |
| "grad_norm": 0.8881410956382751, | |
| "learning_rate": 0.000375, | |
| "loss": 2.9386, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "grad_norm": 0.9484242796897888, | |
| "learning_rate": 0.00043750000000000006, | |
| "loss": 2.8566, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 1.37, | |
| "grad_norm": 0.8347967863082886, | |
| "learning_rate": 0.0005, | |
| "loss": 2.8143, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.54, | |
| "grad_norm": 0.8291537761688232, | |
| "learning_rate": 0.0005625000000000001, | |
| "loss": 2.7705, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 1.71, | |
| "grad_norm": 0.8921717405319214, | |
| "learning_rate": 0.000625, | |
| "loss": 2.7471, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.88, | |
| "grad_norm": 0.8225255608558655, | |
| "learning_rate": 0.0006875, | |
| "loss": 2.7076, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.4602170219704501, | |
| "eval_loss": 2.7808685302734375, | |
| "eval_runtime": 3.378, | |
| "eval_samples_per_second": 1329.483, | |
| "eval_steps_per_second": 10.657, | |
| "step": 5856 | |
| }, | |
| { | |
| "epoch": 2.05, | |
| "grad_norm": 0.8256505131721497, | |
| "learning_rate": 0.00075, | |
| "loss": 2.6746, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 2.22, | |
| "grad_norm": 0.7672858834266663, | |
| "learning_rate": 0.0008125, | |
| "loss": 2.6239, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 2.39, | |
| "grad_norm": 0.7460688948631287, | |
| "learning_rate": 0.0008750000000000001, | |
| "loss": 2.62, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 2.56, | |
| "grad_norm": 0.6622563004493713, | |
| "learning_rate": 0.0009375, | |
| "loss": 2.6093, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 2.73, | |
| "grad_norm": 0.6603254675865173, | |
| "learning_rate": 0.001, | |
| "loss": 2.5886, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 2.9, | |
| "grad_norm": 0.6143876910209656, | |
| "learning_rate": 0.0010625, | |
| "loss": 2.5717, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.47059823934833794, | |
| "eval_loss": 2.6882147789001465, | |
| "eval_runtime": 3.4363, | |
| "eval_samples_per_second": 1306.937, | |
| "eval_steps_per_second": 10.476, | |
| "step": 8784 | |
| }, | |
| { | |
| "epoch": 3.07, | |
| "grad_norm": 0.5893853902816772, | |
| "learning_rate": 0.0011250000000000001, | |
| "loss": 2.5493, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 3.24, | |
| "grad_norm": 0.5745663642883301, | |
| "learning_rate": 0.0011875, | |
| "loss": 2.5131, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 3.42, | |
| "grad_norm": 0.5763567686080933, | |
| "learning_rate": 0.00125, | |
| "loss": 2.5234, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 3.59, | |
| "grad_norm": 0.5464815497398376, | |
| "learning_rate": 0.0013125, | |
| "loss": 2.5211, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 3.76, | |
| "grad_norm": 0.4787265658378601, | |
| "learning_rate": 0.001375, | |
| "loss": 2.5021, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 3.93, | |
| "grad_norm": 0.474524587392807, | |
| "learning_rate": 0.0014375000000000002, | |
| "loss": 2.5132, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.47702228604190006, | |
| "eval_loss": 2.6347815990448, | |
| "eval_runtime": 3.3697, | |
| "eval_samples_per_second": 1332.743, | |
| "eval_steps_per_second": 10.683, | |
| "step": 11712 | |
| }, | |
| { | |
| "epoch": 4.1, | |
| "grad_norm": 0.5468990802764893, | |
| "learning_rate": 0.0015, | |
| "loss": 2.4629, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 4.27, | |
| "grad_norm": 0.46725761890411377, | |
| "learning_rate": 0.0015625, | |
| "loss": 2.4516, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 4.44, | |
| "grad_norm": 0.44199100136756897, | |
| "learning_rate": 0.001625, | |
| "loss": 2.4517, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 4.61, | |
| "grad_norm": 0.3936983048915863, | |
| "learning_rate": 0.0016875, | |
| "loss": 2.4622, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 4.78, | |
| "grad_norm": 0.3848307430744171, | |
| "learning_rate": 0.0017500000000000003, | |
| "loss": 2.4646, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 4.95, | |
| "grad_norm": 0.40697723627090454, | |
| "learning_rate": 0.0018124999999999999, | |
| "loss": 2.4683, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.48105835467961294, | |
| "eval_loss": 2.604564666748047, | |
| "eval_runtime": 3.3835, | |
| "eval_samples_per_second": 1327.329, | |
| "eval_steps_per_second": 10.64, | |
| "step": 14640 | |
| }, | |
| { | |
| "epoch": 5.12, | |
| "grad_norm": 0.356196790933609, | |
| "learning_rate": 0.001875, | |
| "loss": 2.4136, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 5.29, | |
| "grad_norm": 0.39348748326301575, | |
| "learning_rate": 0.0019375000000000002, | |
| "loss": 2.4034, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 5.46, | |
| "grad_norm": 0.3885677456855774, | |
| "learning_rate": 0.002, | |
| "loss": 2.4123, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 5.64, | |
| "grad_norm": 0.36985746026039124, | |
| "learning_rate": 0.0020625, | |
| "loss": 2.4175, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 5.81, | |
| "grad_norm": 0.3625586926937103, | |
| "learning_rate": 0.002125, | |
| "loss": 2.4222, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 5.98, | |
| "grad_norm": 0.3486401438713074, | |
| "learning_rate": 0.0021874999999999998, | |
| "loss": 2.4265, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.48315178037615036, | |
| "eval_loss": 2.5843474864959717, | |
| "eval_runtime": 3.4306, | |
| "eval_samples_per_second": 1309.113, | |
| "eval_steps_per_second": 10.494, | |
| "step": 17568 | |
| }, | |
| { | |
| "epoch": 6.15, | |
| "grad_norm": 0.3697022497653961, | |
| "learning_rate": 0.0022500000000000003, | |
| "loss": 2.3574, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 6.32, | |
| "grad_norm": 0.33410874009132385, | |
| "learning_rate": 0.0023125000000000003, | |
| "loss": 2.3638, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 6.49, | |
| "grad_norm": 0.31180012226104736, | |
| "learning_rate": 0.002375, | |
| "loss": 2.3769, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 6.66, | |
| "grad_norm": 0.3748941421508789, | |
| "learning_rate": 0.0024375, | |
| "loss": 2.3866, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 6.83, | |
| "grad_norm": 0.33421775698661804, | |
| "learning_rate": 0.0025, | |
| "loss": 2.3937, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.4853363770410462, | |
| "eval_loss": 2.570960283279419, | |
| "eval_runtime": 3.3663, | |
| "eval_samples_per_second": 1334.095, | |
| "eval_steps_per_second": 10.694, | |
| "step": 20496 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "grad_norm": 0.3082391917705536, | |
| "learning_rate": 0.0025625, | |
| "loss": 2.3981, | |
| "step": 20500 | |
| }, | |
| { | |
| "epoch": 7.17, | |
| "grad_norm": 0.33345505595207214, | |
| "learning_rate": 0.002625, | |
| "loss": 2.3179, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 7.34, | |
| "grad_norm": 0.3456808924674988, | |
| "learning_rate": 0.0026875000000000002, | |
| "loss": 2.3446, | |
| "step": 21500 | |
| }, | |
| { | |
| "epoch": 7.51, | |
| "grad_norm": 0.33925849199295044, | |
| "learning_rate": 0.00275, | |
| "loss": 2.3468, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 7.68, | |
| "grad_norm": 0.31841370463371277, | |
| "learning_rate": 0.0028125, | |
| "loss": 2.3532, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 7.86, | |
| "grad_norm": 0.30321213603019714, | |
| "learning_rate": 0.0028750000000000004, | |
| "loss": 2.3639, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.4873789573898453, | |
| "eval_loss": 2.5551645755767822, | |
| "eval_runtime": 3.3748, | |
| "eval_samples_per_second": 1330.731, | |
| "eval_steps_per_second": 10.667, | |
| "step": 23424 | |
| }, | |
| { | |
| "epoch": 8.03, | |
| "grad_norm": 0.3192748427391052, | |
| "learning_rate": 0.0029375, | |
| "loss": 2.3562, | |
| "step": 23500 | |
| }, | |
| { | |
| "epoch": 8.2, | |
| "grad_norm": 0.3336254060268402, | |
| "learning_rate": 0.003, | |
| "loss": 2.2971, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 8.37, | |
| "grad_norm": 0.3261987566947937, | |
| "learning_rate": 0.002715909090909091, | |
| "loss": 2.304, | |
| "step": 24500 | |
| }, | |
| { | |
| "epoch": 8.54, | |
| "grad_norm": 0.27662965655326843, | |
| "learning_rate": 0.0024318181818181817, | |
| "loss": 2.3079, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 8.71, | |
| "grad_norm": 0.3111967444419861, | |
| "learning_rate": 0.002147727272727273, | |
| "loss": 2.3027, | |
| "step": 25500 | |
| }, | |
| { | |
| "epoch": 8.88, | |
| "grad_norm": 0.29187533259391785, | |
| "learning_rate": 0.0018636363636363638, | |
| "loss": 2.2858, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_accuracy": 0.4944184081198267, | |
| "eval_loss": 2.5044519901275635, | |
| "eval_runtime": 3.3686, | |
| "eval_samples_per_second": 1333.188, | |
| "eval_steps_per_second": 10.687, | |
| "step": 26352 | |
| }, | |
| { | |
| "epoch": 9.05, | |
| "grad_norm": 0.30923736095428467, | |
| "learning_rate": 0.0015795454545454546, | |
| "loss": 2.241, | |
| "step": 26500 | |
| }, | |
| { | |
| "epoch": 9.22, | |
| "grad_norm": 0.3139377534389496, | |
| "learning_rate": 0.0012954545454545456, | |
| "loss": 2.1769, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 9.39, | |
| "grad_norm": 0.2745420038700104, | |
| "learning_rate": 0.0010113636363636364, | |
| "loss": 2.171, | |
| "step": 27500 | |
| }, | |
| { | |
| "epoch": 9.56, | |
| "grad_norm": 0.3075203001499176, | |
| "learning_rate": 0.0007272727272727273, | |
| "loss": 2.159, | |
| "step": 28000 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 29280, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 2000, | |
| "total_flos": 2173869015564288.0, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |