| { |
| "best_global_step": null, |
| "best_metric": null, |
| "best_model_checkpoint": null, |
| "epoch": 1.0, |
| "eval_steps": 500, |
| "global_step": 54, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 0.018604651162790697, |
| "grad_norm": 51.7990608215332, |
| "learning_rate": 0.0, |
| "loss": 0.71331787109375, |
| "step": 1 |
| }, |
| { |
| "epoch": 0.037209302325581395, |
| "grad_norm": 56.753963470458984, |
| "learning_rate": 7.8125e-08, |
| "loss": 0.71588134765625, |
| "step": 2 |
| }, |
| { |
| "epoch": 0.05581395348837209, |
| "grad_norm": 45.70652389526367, |
| "learning_rate": 1.5625e-07, |
| "loss": 0.7108154296875, |
| "step": 3 |
| }, |
| { |
| "epoch": 0.07441860465116279, |
| "grad_norm": 53.1850700378418, |
| "learning_rate": 2.3437500000000003e-07, |
| "loss": 0.7137451171875, |
| "step": 4 |
| }, |
| { |
| "epoch": 0.09302325581395349, |
| "grad_norm": 49.760826110839844, |
| "learning_rate": 3.125e-07, |
| "loss": 0.71337890625, |
| "step": 5 |
| }, |
| { |
| "epoch": 0.11162790697674418, |
| "grad_norm": 57.31139373779297, |
| "learning_rate": 3.90625e-07, |
| "loss": 0.71978759765625, |
| "step": 6 |
| }, |
| { |
| "epoch": 0.13023255813953488, |
| "grad_norm": 56.42349624633789, |
| "learning_rate": 4.6875000000000006e-07, |
| "loss": 0.7164306640625, |
| "step": 7 |
| }, |
| { |
| "epoch": 0.14883720930232558, |
| "grad_norm": 54.40520477294922, |
| "learning_rate": 5.468750000000001e-07, |
| "loss": 0.71539306640625, |
| "step": 8 |
| }, |
| { |
| "epoch": 0.16744186046511628, |
| "grad_norm": 50.69281768798828, |
| "learning_rate": 6.25e-07, |
| "loss": 0.71405029296875, |
| "step": 9 |
| }, |
| { |
| "epoch": 0.18604651162790697, |
| "grad_norm": 58.078224182128906, |
| "learning_rate": 7.03125e-07, |
| "loss": 0.71514892578125, |
| "step": 10 |
| }, |
| { |
| "epoch": 0.20465116279069767, |
| "grad_norm": 60.195770263671875, |
| "learning_rate": 7.8125e-07, |
| "loss": 0.71466064453125, |
| "step": 11 |
| }, |
| { |
| "epoch": 0.22325581395348837, |
| "grad_norm": 53.095191955566406, |
| "learning_rate": 8.59375e-07, |
| "loss": 0.71002197265625, |
| "step": 12 |
| }, |
| { |
| "epoch": 0.24186046511627907, |
| "grad_norm": 49.17010498046875, |
| "learning_rate": 9.375000000000001e-07, |
| "loss": 0.70770263671875, |
| "step": 13 |
| }, |
| { |
| "epoch": 0.26046511627906976, |
| "grad_norm": 50.612850189208984, |
| "learning_rate": 1.0156250000000001e-06, |
| "loss": 0.7069091796875, |
| "step": 14 |
| }, |
| { |
| "epoch": 0.27906976744186046, |
| "grad_norm": 52.69828414916992, |
| "learning_rate": 1.0937500000000001e-06, |
| "loss": 0.71014404296875, |
| "step": 15 |
| }, |
| { |
| "epoch": 0.29767441860465116, |
| "grad_norm": 59.4550895690918, |
| "learning_rate": 1.1718750000000001e-06, |
| "loss": 0.701416015625, |
| "step": 16 |
| }, |
| { |
| "epoch": 0.31627906976744186, |
| "grad_norm": 55.300506591796875, |
| "learning_rate": 1.25e-06, |
| "loss": 0.6976318359375, |
| "step": 17 |
| }, |
| { |
| "epoch": 0.33488372093023255, |
| "grad_norm": 47.445289611816406, |
| "learning_rate": 1.328125e-06, |
| "loss": 0.69403076171875, |
| "step": 18 |
| }, |
| { |
| "epoch": 0.35348837209302325, |
| "grad_norm": 52.989654541015625, |
| "learning_rate": 1.40625e-06, |
| "loss": 0.69329833984375, |
| "step": 19 |
| }, |
| { |
| "epoch": 0.37209302325581395, |
| "grad_norm": 45.23550033569336, |
| "learning_rate": 1.484375e-06, |
| "loss": 0.692138671875, |
| "step": 20 |
| }, |
| { |
| "epoch": 0.39069767441860465, |
| "grad_norm": 51.775882720947266, |
| "learning_rate": 1.5625e-06, |
| "loss": 0.69146728515625, |
| "step": 21 |
| }, |
| { |
| "epoch": 0.40930232558139534, |
| "grad_norm": 44.17149353027344, |
| "learning_rate": 1.640625e-06, |
| "loss": 0.66351318359375, |
| "step": 22 |
| }, |
| { |
| "epoch": 0.42790697674418604, |
| "grad_norm": 44.85835266113281, |
| "learning_rate": 1.71875e-06, |
| "loss": 0.65997314453125, |
| "step": 23 |
| }, |
| { |
| "epoch": 0.44651162790697674, |
| "grad_norm": 43.70960998535156, |
| "learning_rate": 1.796875e-06, |
| "loss": 0.65948486328125, |
| "step": 24 |
| }, |
| { |
| "epoch": 0.46511627906976744, |
| "grad_norm": 43.53929138183594, |
| "learning_rate": 1.8750000000000003e-06, |
| "loss": 0.6571044921875, |
| "step": 25 |
| }, |
| { |
| "epoch": 0.48372093023255813, |
| "grad_norm": 41.27622985839844, |
| "learning_rate": 1.953125e-06, |
| "loss": 0.65191650390625, |
| "step": 26 |
| }, |
| { |
| "epoch": 0.5023255813953489, |
| "grad_norm": 45.456512451171875, |
| "learning_rate": 2.0312500000000002e-06, |
| "loss": 0.6461181640625, |
| "step": 27 |
| }, |
| { |
| "epoch": 0.5209302325581395, |
| "grad_norm": 47.72049331665039, |
| "learning_rate": 2.109375e-06, |
| "loss": 0.639892578125, |
| "step": 28 |
| }, |
| { |
| "epoch": 0.5395348837209303, |
| "grad_norm": 39.66932678222656, |
| "learning_rate": 2.1875000000000002e-06, |
| "loss": 0.64697265625, |
| "step": 29 |
| }, |
| { |
| "epoch": 0.5581395348837209, |
| "grad_norm": 32.290470123291016, |
| "learning_rate": 2.265625e-06, |
| "loss": 0.61138916015625, |
| "step": 30 |
| }, |
| { |
| "epoch": 0.5767441860465117, |
| "grad_norm": 33.039894104003906, |
| "learning_rate": 2.3437500000000002e-06, |
| "loss": 0.6019287109375, |
| "step": 31 |
| }, |
| { |
| "epoch": 0.5953488372093023, |
| "grad_norm": 38.35325622558594, |
| "learning_rate": 2.421875e-06, |
| "loss": 0.5806884765625, |
| "step": 32 |
| }, |
| { |
| "epoch": 0.6139534883720931, |
| "grad_norm": 33.649505615234375, |
| "learning_rate": 2.5e-06, |
| "loss": 0.58551025390625, |
| "step": 33 |
| }, |
| { |
| "epoch": 0.6325581395348837, |
| "grad_norm": 30.9255428314209, |
| "learning_rate": 2.5781250000000004e-06, |
| "loss": 0.58953857421875, |
| "step": 34 |
| }, |
| { |
| "epoch": 0.6511627906976745, |
| "grad_norm": 35.72899627685547, |
| "learning_rate": 2.65625e-06, |
| "loss": 0.572021484375, |
| "step": 35 |
| }, |
| { |
| "epoch": 0.6697674418604651, |
| "grad_norm": 35.09605026245117, |
| "learning_rate": 2.7343750000000004e-06, |
| "loss": 0.555633544921875, |
| "step": 36 |
| }, |
| { |
| "epoch": 0.6883720930232559, |
| "grad_norm": 29.248594284057617, |
| "learning_rate": 2.8125e-06, |
| "loss": 0.567535400390625, |
| "step": 37 |
| }, |
| { |
| "epoch": 0.7069767441860465, |
| "grad_norm": 31.37779426574707, |
| "learning_rate": 2.8906250000000004e-06, |
| "loss": 0.5550537109375, |
| "step": 38 |
| }, |
| { |
| "epoch": 0.7255813953488373, |
| "grad_norm": 30.69561767578125, |
| "learning_rate": 2.96875e-06, |
| "loss": 0.540008544921875, |
| "step": 39 |
| }, |
| { |
| "epoch": 0.7441860465116279, |
| "grad_norm": 30.249013900756836, |
| "learning_rate": 3.0468750000000004e-06, |
| "loss": 0.526031494140625, |
| "step": 40 |
| }, |
| { |
| "epoch": 0.7627906976744186, |
| "grad_norm": 30.84266471862793, |
| "learning_rate": 3.125e-06, |
| "loss": 0.4935302734375, |
| "step": 41 |
| }, |
| { |
| "epoch": 0.7813953488372093, |
| "grad_norm": 23.530658721923828, |
| "learning_rate": 3.2031250000000004e-06, |
| "loss": 0.506683349609375, |
| "step": 42 |
| }, |
| { |
| "epoch": 0.8, |
| "grad_norm": 26.80575180053711, |
| "learning_rate": 3.28125e-06, |
| "loss": 0.476959228515625, |
| "step": 43 |
| }, |
| { |
| "epoch": 0.8186046511627907, |
| "grad_norm": 27.92676544189453, |
| "learning_rate": 3.3593750000000003e-06, |
| "loss": 0.452056884765625, |
| "step": 44 |
| }, |
| { |
| "epoch": 0.8372093023255814, |
| "grad_norm": 14.391761779785156, |
| "learning_rate": 3.4375e-06, |
| "loss": 0.516387939453125, |
| "step": 45 |
| }, |
| { |
| "epoch": 0.8558139534883721, |
| "grad_norm": 21.352672576904297, |
| "learning_rate": 3.5156250000000003e-06, |
| "loss": 0.4576416015625, |
| "step": 46 |
| }, |
| { |
| "epoch": 0.8744186046511628, |
| "grad_norm": 17.331666946411133, |
| "learning_rate": 3.59375e-06, |
| "loss": 0.467803955078125, |
| "step": 47 |
| }, |
| { |
| "epoch": 0.8930232558139535, |
| "grad_norm": 18.73495864868164, |
| "learning_rate": 3.6718750000000003e-06, |
| "loss": 0.446685791015625, |
| "step": 48 |
| }, |
| { |
| "epoch": 0.9116279069767442, |
| "grad_norm": 16.758777618408203, |
| "learning_rate": 3.7500000000000005e-06, |
| "loss": 0.442779541015625, |
| "step": 49 |
| }, |
| { |
| "epoch": 0.9302325581395349, |
| "grad_norm": 17.156814575195312, |
| "learning_rate": 3.828125000000001e-06, |
| "loss": 0.4298095703125, |
| "step": 50 |
| }, |
| { |
| "epoch": 0.9488372093023256, |
| "grad_norm": 11.690564155578613, |
| "learning_rate": 3.90625e-06, |
| "loss": 0.448089599609375, |
| "step": 51 |
| }, |
| { |
| "epoch": 0.9674418604651163, |
| "grad_norm": 12.62314510345459, |
| "learning_rate": 3.984375e-06, |
| "loss": 0.424957275390625, |
| "step": 52 |
| }, |
| { |
| "epoch": 0.986046511627907, |
| "grad_norm": 8.785666465759277, |
| "learning_rate": 4.0625000000000005e-06, |
| "loss": 0.44073486328125, |
| "step": 53 |
| }, |
| { |
| "epoch": 1.0, |
| "grad_norm": 9.357829093933105, |
| "learning_rate": 4.140625000000001e-06, |
| "loss": 0.413665771484375, |
| "step": 54 |
| }, |
| { |
| "epoch": 1.0, |
| "step": 54, |
| "total_flos": 1.4367917178722714e+17, |
| "train_loss": 0.6073235405815972, |
| "train_runtime": 2167.6199, |
| "train_samples_per_second": 3.168, |
| "train_steps_per_second": 0.025 |
| } |
| ], |
| "logging_steps": 1.0, |
| "max_steps": 54, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 1, |
| "save_steps": 1000, |
| "stateful_callbacks": { |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": true |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 1.4367917178722714e+17, |
| "train_batch_size": 2, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
