| { | |
| "best_metric": 0.6750838756561279, | |
| "best_model_checkpoint": "data/Gemma-2-2B_task-2_120-samples_config-1/checkpoint-66", | |
| "epoch": 13.0, | |
| "eval_steps": 500, | |
| "global_step": 143, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.09090909090909091, | |
| "grad_norm": 0.8511618375778198, | |
| "learning_rate": 1.818181818181818e-06, | |
| "loss": 0.9874, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.18181818181818182, | |
| "grad_norm": 1.144727349281311, | |
| "learning_rate": 3.636363636363636e-06, | |
| "loss": 1.2222, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.36363636363636365, | |
| "grad_norm": 0.8819023370742798, | |
| "learning_rate": 7.272727272727272e-06, | |
| "loss": 0.9639, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.5454545454545454, | |
| "grad_norm": 0.8682326078414917, | |
| "learning_rate": 1.0909090909090909e-05, | |
| "loss": 1.026, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.7272727272727273, | |
| "grad_norm": 0.9766391515731812, | |
| "learning_rate": 1.4545454545454545e-05, | |
| "loss": 1.1319, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.9090909090909091, | |
| "grad_norm": 0.8171083927154541, | |
| "learning_rate": 1.8181818181818182e-05, | |
| "loss": 0.9426, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_loss": 1.0277127027511597, | |
| "eval_runtime": 9.4047, | |
| "eval_samples_per_second": 2.552, | |
| "eval_steps_per_second": 2.552, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 1.0909090909090908, | |
| "grad_norm": 0.8393822312355042, | |
| "learning_rate": 2.1818181818181818e-05, | |
| "loss": 0.962, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 1.2727272727272727, | |
| "grad_norm": 0.7643381953239441, | |
| "learning_rate": 2.5454545454545454e-05, | |
| "loss": 0.9501, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 1.4545454545454546, | |
| "grad_norm": 0.6771353483200073, | |
| "learning_rate": 2.909090909090909e-05, | |
| "loss": 0.9694, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 1.6363636363636362, | |
| "grad_norm": 0.6344677209854126, | |
| "learning_rate": 3.272727272727273e-05, | |
| "loss": 0.9353, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 1.8181818181818183, | |
| "grad_norm": 0.5558416247367859, | |
| "learning_rate": 3.6363636363636364e-05, | |
| "loss": 0.8041, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.4717422127723694, | |
| "learning_rate": 4e-05, | |
| "loss": 0.7193, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 0.8336329460144043, | |
| "eval_runtime": 9.1236, | |
| "eval_samples_per_second": 2.631, | |
| "eval_steps_per_second": 2.631, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 2.1818181818181817, | |
| "grad_norm": 0.4751701056957245, | |
| "learning_rate": 4.3636363636363636e-05, | |
| "loss": 0.7583, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 2.3636363636363638, | |
| "grad_norm": 0.5473897457122803, | |
| "learning_rate": 4.7272727272727275e-05, | |
| "loss": 0.8181, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 2.5454545454545454, | |
| "grad_norm": 0.48169592022895813, | |
| "learning_rate": 5.090909090909091e-05, | |
| "loss": 0.702, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 2.7272727272727275, | |
| "grad_norm": 0.45491889119148254, | |
| "learning_rate": 5.4545454545454546e-05, | |
| "loss": 0.6957, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 2.909090909090909, | |
| "grad_norm": 0.4473062753677368, | |
| "learning_rate": 5.818181818181818e-05, | |
| "loss": 0.654, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 0.7473955154418945, | |
| "eval_runtime": 9.1284, | |
| "eval_samples_per_second": 2.629, | |
| "eval_steps_per_second": 2.629, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 3.090909090909091, | |
| "grad_norm": 0.4919629991054535, | |
| "learning_rate": 6.181818181818182e-05, | |
| "loss": 0.7357, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 3.2727272727272725, | |
| "grad_norm": 0.46526703238487244, | |
| "learning_rate": 6.545454545454546e-05, | |
| "loss": 0.6668, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 3.4545454545454546, | |
| "grad_norm": 0.40764591097831726, | |
| "learning_rate": 6.90909090909091e-05, | |
| "loss": 0.6524, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 3.6363636363636362, | |
| "grad_norm": 0.36564379930496216, | |
| "learning_rate": 7.272727272727273e-05, | |
| "loss": 0.6019, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 3.8181818181818183, | |
| "grad_norm": 0.5233597755432129, | |
| "learning_rate": 7.636363636363637e-05, | |
| "loss": 0.663, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.39497652649879456, | |
| "learning_rate": 8e-05, | |
| "loss": 0.6017, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 0.7071707248687744, | |
| "eval_runtime": 9.1216, | |
| "eval_samples_per_second": 2.631, | |
| "eval_steps_per_second": 2.631, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 4.181818181818182, | |
| "grad_norm": 0.44469550251960754, | |
| "learning_rate": 8.363636363636364e-05, | |
| "loss": 0.6389, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 4.363636363636363, | |
| "grad_norm": 0.48129987716674805, | |
| "learning_rate": 8.727272727272727e-05, | |
| "loss": 0.5528, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 4.545454545454545, | |
| "grad_norm": 0.4613235294818878, | |
| "learning_rate": 9.090909090909092e-05, | |
| "loss": 0.586, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 4.7272727272727275, | |
| "grad_norm": 0.5269588232040405, | |
| "learning_rate": 9.454545454545455e-05, | |
| "loss": 0.6002, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 4.909090909090909, | |
| "grad_norm": 0.43463221192359924, | |
| "learning_rate": 9.818181818181818e-05, | |
| "loss": 0.5602, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_loss": 0.6801621913909912, | |
| "eval_runtime": 9.1227, | |
| "eval_samples_per_second": 2.631, | |
| "eval_steps_per_second": 2.631, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 5.090909090909091, | |
| "grad_norm": 0.3945789933204651, | |
| "learning_rate": 9.999899300364532e-05, | |
| "loss": 0.493, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 5.2727272727272725, | |
| "grad_norm": 0.5101122856140137, | |
| "learning_rate": 9.99909372761763e-05, | |
| "loss": 0.5523, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 5.454545454545454, | |
| "grad_norm": 0.49218466877937317, | |
| "learning_rate": 9.997482711915927e-05, | |
| "loss": 0.4964, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 5.636363636363637, | |
| "grad_norm": 1.013070821762085, | |
| "learning_rate": 9.99506651282272e-05, | |
| "loss": 0.4928, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 5.818181818181818, | |
| "grad_norm": 0.5383065342903137, | |
| "learning_rate": 9.991845519630678e-05, | |
| "loss": 0.4931, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 0.5603281259536743, | |
| "learning_rate": 9.987820251299122e-05, | |
| "loss": 0.4885, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 0.6750838756561279, | |
| "eval_runtime": 9.1243, | |
| "eval_samples_per_second": 2.63, | |
| "eval_steps_per_second": 2.63, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 6.181818181818182, | |
| "grad_norm": 0.5732660889625549, | |
| "learning_rate": 9.982991356370404e-05, | |
| "loss": 0.4438, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 6.363636363636363, | |
| "grad_norm": 0.587360143661499, | |
| "learning_rate": 9.977359612865423e-05, | |
| "loss": 0.4396, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 6.545454545454545, | |
| "grad_norm": 0.6737551093101501, | |
| "learning_rate": 9.970925928158274e-05, | |
| "loss": 0.3876, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 6.7272727272727275, | |
| "grad_norm": 0.6717156767845154, | |
| "learning_rate": 9.963691338830044e-05, | |
| "loss": 0.4048, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 6.909090909090909, | |
| "grad_norm": 0.7882419228553772, | |
| "learning_rate": 9.955657010501806e-05, | |
| "loss": 0.4015, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_loss": 0.7140398025512695, | |
| "eval_runtime": 9.1283, | |
| "eval_samples_per_second": 2.629, | |
| "eval_steps_per_second": 2.629, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 7.090909090909091, | |
| "grad_norm": 0.7362752556800842, | |
| "learning_rate": 9.946824237646824e-05, | |
| "loss": 0.372, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 7.2727272727272725, | |
| "grad_norm": 0.707953691482544, | |
| "learning_rate": 9.937194443381972e-05, | |
| "loss": 0.3116, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 7.454545454545454, | |
| "grad_norm": 0.763971209526062, | |
| "learning_rate": 9.926769179238466e-05, | |
| "loss": 0.3098, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 7.636363636363637, | |
| "grad_norm": 0.9585027694702148, | |
| "learning_rate": 9.915550124911866e-05, | |
| "loss": 0.3044, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 7.818181818181818, | |
| "grad_norm": 1.039747714996338, | |
| "learning_rate": 9.903539087991462e-05, | |
| "loss": 0.3262, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 0.9194966554641724, | |
| "learning_rate": 9.890738003669029e-05, | |
| "loss": 0.3378, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 0.7642681002616882, | |
| "eval_runtime": 9.1268, | |
| "eval_samples_per_second": 2.63, | |
| "eval_steps_per_second": 2.63, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 8.181818181818182, | |
| "grad_norm": 0.7590727210044861, | |
| "learning_rate": 9.877148934427037e-05, | |
| "loss": 0.247, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 8.363636363636363, | |
| "grad_norm": 0.8521314263343811, | |
| "learning_rate": 9.862774069706346e-05, | |
| "loss": 0.2163, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 8.545454545454545, | |
| "grad_norm": 1.1055257320404053, | |
| "learning_rate": 9.847615725553456e-05, | |
| "loss": 0.2174, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 8.727272727272727, | |
| "grad_norm": 1.1664984226226807, | |
| "learning_rate": 9.831676344247342e-05, | |
| "loss": 0.1914, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 8.909090909090908, | |
| "grad_norm": 1.4830482006072998, | |
| "learning_rate": 9.814958493905963e-05, | |
| "loss": 0.2011, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_loss": 0.9955660700798035, | |
| "eval_runtime": 9.1314, | |
| "eval_samples_per_second": 2.628, | |
| "eval_steps_per_second": 2.628, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 9.090909090909092, | |
| "grad_norm": 1.1150389909744263, | |
| "learning_rate": 9.797464868072488e-05, | |
| "loss": 0.1937, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 9.272727272727273, | |
| "grad_norm": 1.0970076322555542, | |
| "learning_rate": 9.779198285281325e-05, | |
| "loss": 0.1124, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 9.454545454545455, | |
| "grad_norm": 1.2145295143127441, | |
| "learning_rate": 9.760161688604008e-05, | |
| "loss": 0.1201, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 9.636363636363637, | |
| "grad_norm": 1.368381381034851, | |
| "learning_rate": 9.740358145174998e-05, | |
| "loss": 0.1269, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 9.818181818181818, | |
| "grad_norm": 1.2358267307281494, | |
| "learning_rate": 9.719790845697533e-05, | |
| "loss": 0.1135, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 1.257324457168579, | |
| "learning_rate": 9.698463103929542e-05, | |
| "loss": 0.1133, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_loss": 1.1606191396713257, | |
| "eval_runtime": 9.1274, | |
| "eval_samples_per_second": 2.629, | |
| "eval_steps_per_second": 2.629, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 10.181818181818182, | |
| "grad_norm": 0.8183621764183044, | |
| "learning_rate": 9.676378356149734e-05, | |
| "loss": 0.0612, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 10.363636363636363, | |
| "grad_norm": 1.148886799812317, | |
| "learning_rate": 9.653540160603956e-05, | |
| "loss": 0.0622, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 10.545454545454545, | |
| "grad_norm": 1.9670130014419556, | |
| "learning_rate": 9.629952196931901e-05, | |
| "loss": 0.0736, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 10.727272727272727, | |
| "grad_norm": 1.28963041305542, | |
| "learning_rate": 9.60561826557425e-05, | |
| "loss": 0.0648, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 10.909090909090908, | |
| "grad_norm": 1.0956385135650635, | |
| "learning_rate": 9.580542287160348e-05, | |
| "loss": 0.0573, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_loss": 1.3293719291687012, | |
| "eval_runtime": 9.1317, | |
| "eval_samples_per_second": 2.628, | |
| "eval_steps_per_second": 2.628, | |
| "step": 121 | |
| }, | |
| { | |
| "epoch": 11.090909090909092, | |
| "grad_norm": 0.7760180830955505, | |
| "learning_rate": 9.554728301876526e-05, | |
| "loss": 0.0503, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 11.272727272727273, | |
| "grad_norm": 1.211250901222229, | |
| "learning_rate": 9.528180468815155e-05, | |
| "loss": 0.037, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 11.454545454545455, | |
| "grad_norm": 1.1533076763153076, | |
| "learning_rate": 9.50090306530454e-05, | |
| "loss": 0.0397, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 11.636363636363637, | |
| "grad_norm": 1.3236202001571655, | |
| "learning_rate": 9.472900486219769e-05, | |
| "loss": 0.0403, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 11.818181818181818, | |
| "grad_norm": 0.8378921151161194, | |
| "learning_rate": 9.444177243274618e-05, | |
| "loss": 0.0341, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "grad_norm": 0.8957329988479614, | |
| "learning_rate": 9.414737964294636e-05, | |
| "loss": 0.0337, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_loss": 1.3806809186935425, | |
| "eval_runtime": 9.1248, | |
| "eval_samples_per_second": 2.63, | |
| "eval_steps_per_second": 2.63, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 12.181818181818182, | |
| "grad_norm": 0.7634783387184143, | |
| "learning_rate": 9.384587392471515e-05, | |
| "loss": 0.0211, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 12.363636363636363, | |
| "grad_norm": 0.6416231393814087, | |
| "learning_rate": 9.353730385598887e-05, | |
| "loss": 0.0161, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 12.545454545454545, | |
| "grad_norm": 1.1831228733062744, | |
| "learning_rate": 9.322171915289635e-05, | |
| "loss": 0.0179, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 12.727272727272727, | |
| "grad_norm": 1.486878514289856, | |
| "learning_rate": 9.289917066174886e-05, | |
| "loss": 0.0276, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 12.909090909090908, | |
| "grad_norm": 1.0545414686203003, | |
| "learning_rate": 9.256971035084785e-05, | |
| "loss": 0.0259, | |
| "step": 142 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "eval_loss": 1.5131946802139282, | |
| "eval_runtime": 9.1272, | |
| "eval_samples_per_second": 2.629, | |
| "eval_steps_per_second": 2.629, | |
| "step": 143 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "step": 143, | |
| "total_flos": 3.728219651964928e+16, | |
| "train_loss": 0.42762258671015707, | |
| "train_runtime": 1593.2725, | |
| "train_samples_per_second": 2.762, | |
| "train_steps_per_second": 0.345 | |
| } | |
| ], | |
| "logging_steps": 2, | |
| "max_steps": 550, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 50, | |
| "save_steps": 25, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 7, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 0 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 3.728219651964928e+16, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |