| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 10.0, | |
| "eval_steps": 500, | |
| "global_step": 380, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.02631578947368421, | |
| "grad_norm": 170.0, | |
| "learning_rate": 5.263157894736842e-06, | |
| "loss": 33.7533, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.13157894736842105, | |
| "grad_norm": 137.0, | |
| "learning_rate": 2.6315789473684212e-05, | |
| "loss": 34.5932, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.2631578947368421, | |
| "grad_norm": 67.5, | |
| "learning_rate": 5.2631578947368424e-05, | |
| "loss": 30.1164, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.39473684210526316, | |
| "grad_norm": 17.25, | |
| "learning_rate": 7.894736842105263e-05, | |
| "loss": 21.7002, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.5263157894736842, | |
| "grad_norm": 14.3125, | |
| "learning_rate": 0.00010526315789473685, | |
| "loss": 17.9593, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.6578947368421053, | |
| "grad_norm": 5.375, | |
| "learning_rate": 0.00013157894736842108, | |
| "loss": 15.8965, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.7894736842105263, | |
| "grad_norm": 3.203125, | |
| "learning_rate": 0.00015789473684210527, | |
| "loss": 14.3005, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.9210526315789473, | |
| "grad_norm": 3.90625, | |
| "learning_rate": 0.00018421052631578948, | |
| "loss": 13.7073, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_loss": 7.137722015380859, | |
| "eval_runtime": 0.2537, | |
| "eval_samples_per_second": 39.423, | |
| "eval_steps_per_second": 3.942, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 1.0526315789473684, | |
| "grad_norm": 7.375, | |
| "learning_rate": 0.00019998312416333227, | |
| "loss": 12.9513, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 1.1842105263157894, | |
| "grad_norm": 12.0, | |
| "learning_rate": 0.00019979333640833947, | |
| "loss": 10.9396, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 1.3157894736842106, | |
| "grad_norm": 20.25, | |
| "learning_rate": 0.00019939306773179497, | |
| "loss": 8.1563, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 1.4473684210526316, | |
| "grad_norm": 13.8125, | |
| "learning_rate": 0.00019878316236762196, | |
| "loss": 4.3271, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 1.5789473684210527, | |
| "grad_norm": 5.84375, | |
| "learning_rate": 0.0001979649067087574, | |
| "loss": 2.3733, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.7105263157894737, | |
| "grad_norm": 4.96875, | |
| "learning_rate": 0.00019694002659393305, | |
| "loss": 2.0099, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 1.8421052631578947, | |
| "grad_norm": 1.171875, | |
| "learning_rate": 0.00019571068366759143, | |
| "loss": 1.7688, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.973684210526316, | |
| "grad_norm": 1.046875, | |
| "learning_rate": 0.00019427947082061432, | |
| "loss": 1.6173, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 2.9756951332092285, | |
| "eval_runtime": 0.2386, | |
| "eval_samples_per_second": 41.915, | |
| "eval_steps_per_second": 4.191, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 2.1052631578947367, | |
| "grad_norm": 0.9453125, | |
| "learning_rate": 0.00019264940672148018, | |
| "loss": 1.5155, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 2.236842105263158, | |
| "grad_norm": 0.84765625, | |
| "learning_rate": 0.00019082392944938466, | |
| "loss": 1.4557, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 2.3684210526315788, | |
| "grad_norm": 0.7421875, | |
| "learning_rate": 0.00018880688924275378, | |
| "loss": 1.4092, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "grad_norm": 0.43359375, | |
| "learning_rate": 0.00018660254037844388, | |
| "loss": 1.3523, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 2.6315789473684212, | |
| "grad_norm": 0.58203125, | |
| "learning_rate": 0.00018421553219875658, | |
| "loss": 1.3176, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 2.763157894736842, | |
| "grad_norm": 0.58203125, | |
| "learning_rate": 0.0001816508993051943, | |
| "loss": 1.2811, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 2.8947368421052633, | |
| "grad_norm": 0.82421875, | |
| "learning_rate": 0.00017891405093963938, | |
| "loss": 1.2585, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 2.7225849628448486, | |
| "eval_runtime": 0.2375, | |
| "eval_samples_per_second": 42.102, | |
| "eval_steps_per_second": 4.21, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 3.026315789473684, | |
| "grad_norm": 0.9453125, | |
| "learning_rate": 0.00017601075957535364, | |
| "loss": 1.2406, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 3.1578947368421053, | |
| "grad_norm": 0.984375, | |
| "learning_rate": 0.0001729471487418621, | |
| "loss": 1.209, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 3.2894736842105265, | |
| "grad_norm": 0.85546875, | |
| "learning_rate": 0.00016972968010939954, | |
| "loss": 1.1876, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 3.4210526315789473, | |
| "grad_norm": 0.953125, | |
| "learning_rate": 0.00016636513986016213, | |
| "loss": 1.1913, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 3.5526315789473686, | |
| "grad_norm": 0.6015625, | |
| "learning_rate": 0.0001628606243751082, | |
| "loss": 1.17, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 3.6842105263157894, | |
| "grad_norm": 0.546875, | |
| "learning_rate": 0.00015922352526649803, | |
| "loss": 1.1573, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 3.8157894736842106, | |
| "grad_norm": 0.734375, | |
| "learning_rate": 0.00015546151378774086, | |
| "loss": 1.157, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 3.9473684210526314, | |
| "grad_norm": 0.81640625, | |
| "learning_rate": 0.00015158252465343242, | |
| "loss": 1.1493, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 2.631202220916748, | |
| "eval_runtime": 0.238, | |
| "eval_samples_per_second": 42.017, | |
| "eval_steps_per_second": 4.202, | |
| "step": 152 | |
| }, | |
| { | |
| "epoch": 4.078947368421052, | |
| "grad_norm": 1.21875, | |
| "learning_rate": 0.00014759473930370736, | |
| "loss": 1.1375, | |
| "step": 155 | |
| }, | |
| { | |
| "epoch": 4.2105263157894735, | |
| "grad_norm": 1.1015625, | |
| "learning_rate": 0.00014350656864820733, | |
| "loss": 1.1162, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 4.342105263157895, | |
| "grad_norm": 1.2734375, | |
| "learning_rate": 0.0001393266353260583, | |
| "loss": 1.1148, | |
| "step": 165 | |
| }, | |
| { | |
| "epoch": 4.473684210526316, | |
| "grad_norm": 1.0078125, | |
| "learning_rate": 0.00013506375551927547, | |
| "loss": 1.1125, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 4.605263157894737, | |
| "grad_norm": 0.625, | |
| "learning_rate": 0.00013072692035795305, | |
| "loss": 1.1057, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 4.7368421052631575, | |
| "grad_norm": 0.7421875, | |
| "learning_rate": 0.00012632527695645993, | |
| "loss": 1.1023, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 4.868421052631579, | |
| "grad_norm": 0.703125, | |
| "learning_rate": 0.0001218681091206376, | |
| "loss": 1.0983, | |
| "step": 185 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 0.6796875, | |
| "learning_rate": 0.00011736481776669306, | |
| "loss": 1.0934, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_loss": 2.6199984550476074, | |
| "eval_runtime": 0.2365, | |
| "eval_samples_per_second": 42.281, | |
| "eval_steps_per_second": 4.228, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 5.131578947368421, | |
| "grad_norm": 0.74609375, | |
| "learning_rate": 0.00011282490109308633, | |
| "loss": 1.0826, | |
| "step": 195 | |
| }, | |
| { | |
| "epoch": 5.2631578947368425, | |
| "grad_norm": 0.81640625, | |
| "learning_rate": 0.00010825793454723325, | |
| "loss": 1.0788, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 5.394736842105263, | |
| "grad_norm": 0.68359375, | |
| "learning_rate": 0.00010367355062927726, | |
| "loss": 1.0782, | |
| "step": 205 | |
| }, | |
| { | |
| "epoch": 5.526315789473684, | |
| "grad_norm": 0.95703125, | |
| "learning_rate": 9.908141857552737e-05, | |
| "loss": 1.0606, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 5.657894736842105, | |
| "grad_norm": 0.65234375, | |
| "learning_rate": 9.449122396441345e-05, | |
| "loss": 1.0564, | |
| "step": 215 | |
| }, | |
| { | |
| "epoch": 5.7894736842105265, | |
| "grad_norm": 0.5546875, | |
| "learning_rate": 8.991264828797319e-05, | |
| "loss": 1.0509, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 5.921052631578947, | |
| "grad_norm": 0.6328125, | |
| "learning_rate": 8.535534853195786e-05, | |
| "loss": 1.0587, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 2.6019885540008545, | |
| "eval_runtime": 0.2397, | |
| "eval_samples_per_second": 41.723, | |
| "eval_steps_per_second": 4.172, | |
| "step": 228 | |
| }, | |
| { | |
| "epoch": 6.052631578947368, | |
| "grad_norm": 0.6640625, | |
| "learning_rate": 8.082893680762619e-05, | |
| "loss": 1.0543, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 6.184210526315789, | |
| "grad_norm": 0.72265625, | |
| "learning_rate": 7.634296007818576e-05, | |
| "loss": 1.0452, | |
| "step": 235 | |
| }, | |
| { | |
| "epoch": 6.315789473684211, | |
| "grad_norm": 0.84375, | |
| "learning_rate": 7.190688002264308e-05, | |
| "loss": 1.042, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 6.447368421052632, | |
| "grad_norm": 0.66796875, | |
| "learning_rate": 6.753005307953167e-05, | |
| "loss": 1.0413, | |
| "step": 245 | |
| }, | |
| { | |
| "epoch": 6.578947368421053, | |
| "grad_norm": 0.80078125, | |
| "learning_rate": 6.322171071261071e-05, | |
| "loss": 1.0436, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 6.7105263157894735, | |
| "grad_norm": 0.77734375, | |
| "learning_rate": 5.8990939940156e-05, | |
| "loss": 1.0367, | |
| "step": 255 | |
| }, | |
| { | |
| "epoch": 6.842105263157895, | |
| "grad_norm": 0.5703125, | |
| "learning_rate": 5.484666416891109e-05, | |
| "loss": 1.0299, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 6.973684210526316, | |
| "grad_norm": 0.70703125, | |
| "learning_rate": 5.079762437312219e-05, | |
| "loss": 1.0289, | |
| "step": 265 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_loss": 2.5995545387268066, | |
| "eval_runtime": 0.2387, | |
| "eval_samples_per_second": 41.887, | |
| "eval_steps_per_second": 4.189, | |
| "step": 266 | |
| }, | |
| { | |
| "epoch": 7.105263157894737, | |
| "grad_norm": 0.53125, | |
| "learning_rate": 4.685236065835443e-05, | |
| "loss": 1.0249, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 7.2368421052631575, | |
| "grad_norm": 0.671875, | |
| "learning_rate": 4.301919424897338e-05, | |
| "loss": 1.0192, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 7.368421052631579, | |
| "grad_norm": 0.486328125, | |
| "learning_rate": 3.9306209937284346e-05, | |
| "loss": 1.0285, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 7.5, | |
| "grad_norm": 0.72265625, | |
| "learning_rate": 3.5721239031346066e-05, | |
| "loss": 1.025, | |
| "step": 285 | |
| }, | |
| { | |
| "epoch": 7.631578947368421, | |
| "grad_norm": 0.58984375, | |
| "learning_rate": 3.227184283742591e-05, | |
| "loss": 1.0347, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 7.7631578947368425, | |
| "grad_norm": 0.51171875, | |
| "learning_rate": 2.89652967119336e-05, | |
| "loss": 1.0223, | |
| "step": 295 | |
| }, | |
| { | |
| "epoch": 7.894736842105263, | |
| "grad_norm": 0.52734375, | |
| "learning_rate": 2.5808574716471856e-05, | |
| "loss": 1.0197, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 2.602214813232422, | |
| "eval_runtime": 0.2375, | |
| "eval_samples_per_second": 42.1, | |
| "eval_steps_per_second": 4.21, | |
| "step": 304 | |
| }, | |
| { | |
| "epoch": 8.026315789473685, | |
| "grad_norm": 0.5546875, | |
| "learning_rate": 2.2808334908367914e-05, | |
| "loss": 1.023, | |
| "step": 305 | |
| }, | |
| { | |
| "epoch": 8.157894736842104, | |
| "grad_norm": 0.498046875, | |
| "learning_rate": 1.9970905297711606e-05, | |
| "loss": 1.0158, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 8.289473684210526, | |
| "grad_norm": 0.51953125, | |
| "learning_rate": 1.7302270500518182e-05, | |
| "loss": 1.0183, | |
| "step": 315 | |
| }, | |
| { | |
| "epoch": 8.421052631578947, | |
| "grad_norm": 0.46875, | |
| "learning_rate": 1.4808059116167305e-05, | |
| "loss": 1.0081, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 8.552631578947368, | |
| "grad_norm": 0.53125, | |
| "learning_rate": 1.2493531855740625e-05, | |
| "loss": 1.0149, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 8.68421052631579, | |
| "grad_norm": 0.494140625, | |
| "learning_rate": 1.0363570446297999e-05, | |
| "loss": 1.0197, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 8.81578947368421, | |
| "grad_norm": 0.46484375, | |
| "learning_rate": 8.422667334494249e-06, | |
| "loss": 1.02, | |
| "step": 335 | |
| }, | |
| { | |
| "epoch": 8.947368421052632, | |
| "grad_norm": 0.5, | |
| "learning_rate": 6.674916211254289e-06, | |
| "loss": 1.0221, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_loss": 2.605945110321045, | |
| "eval_runtime": 0.2375, | |
| "eval_samples_per_second": 42.099, | |
| "eval_steps_per_second": 4.21, | |
| "step": 342 | |
| }, | |
| { | |
| "epoch": 9.078947368421053, | |
| "grad_norm": 0.45703125, | |
| "learning_rate": 5.124003377490582e-06, | |
| "loss": 1.0218, | |
| "step": 345 | |
| }, | |
| { | |
| "epoch": 9.210526315789474, | |
| "grad_norm": 0.46484375, | |
| "learning_rate": 3.7731999690749585e-06, | |
| "loss": 1.0087, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 9.342105263157896, | |
| "grad_norm": 0.4609375, | |
| "learning_rate": 2.6253550574632303e-06, | |
| "loss": 1.0178, | |
| "step": 355 | |
| }, | |
| { | |
| "epoch": 9.473684210526315, | |
| "grad_norm": 0.453125, | |
| "learning_rate": 1.6828896405244988e-06, | |
| "loss": 1.0225, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 9.605263157894736, | |
| "grad_norm": 0.455078125, | |
| "learning_rate": 9.477915362496758e-07, | |
| "loss": 1.0206, | |
| "step": 365 | |
| }, | |
| { | |
| "epoch": 9.736842105263158, | |
| "grad_norm": 0.490234375, | |
| "learning_rate": 4.216111901092501e-07, | |
| "loss": 1.017, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 9.868421052631579, | |
| "grad_norm": 0.46484375, | |
| "learning_rate": 1.0545840490313596e-07, | |
| "loss": 1.0248, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 0.57421875, | |
| "learning_rate": 0.0, | |
| "loss": 1.0175, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_loss": 2.6068081855773926, | |
| "eval_runtime": 0.235, | |
| "eval_samples_per_second": 42.55, | |
| "eval_steps_per_second": 4.255, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 380, | |
| "total_flos": 1.158687595912233e+18, | |
| "train_loss": 3.413645140748275, | |
| "train_runtime": 927.0586, | |
| "train_samples_per_second": 26.201, | |
| "train_steps_per_second": 0.41 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 380, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 100, | |
| "total_flos": 1.158687595912233e+18, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |