| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 10.0, | |
| "eval_steps": 500, | |
| "global_step": 1370, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.18248175182481752, | |
| "grad_norm": 23.73174476623535, | |
| "learning_rate": 0.00046, | |
| "loss": 4.0301, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.36496350364963503, | |
| "grad_norm": 2.086200714111328, | |
| "learning_rate": 0.00096, | |
| "loss": 1.0569, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.5474452554744526, | |
| "grad_norm": 0.4636005461215973, | |
| "learning_rate": 0.0009825757575757576, | |
| "loss": 0.4231, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.7299270072992701, | |
| "grad_norm": 0.905696451663971, | |
| "learning_rate": 0.0009636363636363637, | |
| "loss": 0.1587, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.9124087591240876, | |
| "grad_norm": 0.7420445084571838, | |
| "learning_rate": 0.0009446969696969697, | |
| "loss": 0.1302, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_loss": 0.09637967497110367, | |
| "eval_runtime": 192.8711, | |
| "eval_samples_per_second": 4.241, | |
| "eval_steps_per_second": 0.71, | |
| "step": 137 | |
| }, | |
| { | |
| "epoch": 1.094890510948905, | |
| "grad_norm": 0.5315635204315186, | |
| "learning_rate": 0.0009257575757575758, | |
| "loss": 0.1168, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.2773722627737225, | |
| "grad_norm": 2.168321132659912, | |
| "learning_rate": 0.0009068181818181819, | |
| "loss": 0.0957, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 1.4598540145985401, | |
| "grad_norm": 1.4148447513580322, | |
| "learning_rate": 0.000887878787878788, | |
| "loss": 0.0819, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.6423357664233578, | |
| "grad_norm": 1.08322274684906, | |
| "learning_rate": 0.0008689393939393939, | |
| "loss": 0.0967, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 1.8248175182481752, | |
| "grad_norm": 0.34484872221946716, | |
| "learning_rate": 0.00085, | |
| "loss": 0.0897, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 0.05452003329992294, | |
| "eval_runtime": 193.1249, | |
| "eval_samples_per_second": 4.236, | |
| "eval_steps_per_second": 0.709, | |
| "step": 274 | |
| }, | |
| { | |
| "epoch": 2.0072992700729926, | |
| "grad_norm": 0.4588640034198761, | |
| "learning_rate": 0.0008310606060606061, | |
| "loss": 0.078, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 2.18978102189781, | |
| "grad_norm": 0.9430391788482666, | |
| "learning_rate": 0.0008121212121212122, | |
| "loss": 0.047, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 2.372262773722628, | |
| "grad_norm": 1.2605994939804077, | |
| "learning_rate": 0.0007931818181818182, | |
| "loss": 0.0733, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 2.554744525547445, | |
| "grad_norm": 0.4879356026649475, | |
| "learning_rate": 0.0007742424242424244, | |
| "loss": 0.0595, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 2.7372262773722627, | |
| "grad_norm": 0.7468044757843018, | |
| "learning_rate": 0.0007553030303030303, | |
| "loss": 0.0581, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 2.9197080291970803, | |
| "grad_norm": 1.2231853008270264, | |
| "learning_rate": 0.0007363636363636363, | |
| "loss": 0.0691, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 0.039563629776239395, | |
| "eval_runtime": 191.6012, | |
| "eval_samples_per_second": 4.269, | |
| "eval_steps_per_second": 0.715, | |
| "step": 411 | |
| }, | |
| { | |
| "epoch": 3.102189781021898, | |
| "grad_norm": 0.3177865445613861, | |
| "learning_rate": 0.0007174242424242424, | |
| "loss": 0.0421, | |
| "step": 425 | |
| }, | |
| { | |
| "epoch": 3.2846715328467155, | |
| "grad_norm": 1.0183290243148804, | |
| "learning_rate": 0.0006984848484848485, | |
| "loss": 0.0485, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 3.4671532846715327, | |
| "grad_norm": 0.361370712518692, | |
| "learning_rate": 0.0006795454545454546, | |
| "loss": 0.0369, | |
| "step": 475 | |
| }, | |
| { | |
| "epoch": 3.6496350364963503, | |
| "grad_norm": 1.0689235925674438, | |
| "learning_rate": 0.0006606060606060606, | |
| "loss": 0.0456, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 3.832116788321168, | |
| "grad_norm": 0.29606688022613525, | |
| "learning_rate": 0.0006416666666666667, | |
| "loss": 0.0482, | |
| "step": 525 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 0.027906004339456558, | |
| "eval_runtime": 190.9612, | |
| "eval_samples_per_second": 4.284, | |
| "eval_steps_per_second": 0.717, | |
| "step": 548 | |
| }, | |
| { | |
| "epoch": 4.014598540145985, | |
| "grad_norm": 0.1826704740524292, | |
| "learning_rate": 0.0006227272727272727, | |
| "loss": 0.0451, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 4.197080291970803, | |
| "grad_norm": 0.1476697325706482, | |
| "learning_rate": 0.0006037878787878788, | |
| "loss": 0.0324, | |
| "step": 575 | |
| }, | |
| { | |
| "epoch": 4.37956204379562, | |
| "grad_norm": 0.37707987427711487, | |
| "learning_rate": 0.0005848484848484848, | |
| "loss": 0.0277, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 4.562043795620438, | |
| "grad_norm": 0.18928465247154236, | |
| "learning_rate": 0.0005659090909090909, | |
| "loss": 0.0331, | |
| "step": 625 | |
| }, | |
| { | |
| "epoch": 4.744525547445256, | |
| "grad_norm": 0.10301584750413895, | |
| "learning_rate": 0.000546969696969697, | |
| "loss": 0.0189, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 4.927007299270073, | |
| "grad_norm": 0.5264261960983276, | |
| "learning_rate": 0.0005280303030303031, | |
| "loss": 0.0374, | |
| "step": 675 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_loss": 0.017436422407627106, | |
| "eval_runtime": 189.4577, | |
| "eval_samples_per_second": 4.318, | |
| "eval_steps_per_second": 0.723, | |
| "step": 685 | |
| }, | |
| { | |
| "epoch": 5.109489051094891, | |
| "grad_norm": 0.3798004984855652, | |
| "learning_rate": 0.000509090909090909, | |
| "loss": 0.0166, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 5.291970802919708, | |
| "grad_norm": 0.22640006244182587, | |
| "learning_rate": 0.0004901515151515152, | |
| "loss": 0.0224, | |
| "step": 725 | |
| }, | |
| { | |
| "epoch": 5.474452554744525, | |
| "grad_norm": 0.2977535128593445, | |
| "learning_rate": 0.0004712121212121212, | |
| "loss": 0.0178, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 5.656934306569343, | |
| "grad_norm": 0.046165354549884796, | |
| "learning_rate": 0.00045227272727272727, | |
| "loss": 0.0188, | |
| "step": 775 | |
| }, | |
| { | |
| "epoch": 5.839416058394161, | |
| "grad_norm": 0.20036080479621887, | |
| "learning_rate": 0.00043333333333333337, | |
| "loss": 0.0189, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 0.009602065198123455, | |
| "eval_runtime": 190.5013, | |
| "eval_samples_per_second": 4.294, | |
| "eval_steps_per_second": 0.719, | |
| "step": 822 | |
| }, | |
| { | |
| "epoch": 6.021897810218978, | |
| "grad_norm": 0.009559686295688152, | |
| "learning_rate": 0.00041439393939393936, | |
| "loss": 0.0112, | |
| "step": 825 | |
| }, | |
| { | |
| "epoch": 6.204379562043796, | |
| "grad_norm": 0.17128558456897736, | |
| "learning_rate": 0.00039545454545454546, | |
| "loss": 0.0115, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 6.386861313868613, | |
| "grad_norm": 0.12201745808124542, | |
| "learning_rate": 0.0003765151515151515, | |
| "loss": 0.0099, | |
| "step": 875 | |
| }, | |
| { | |
| "epoch": 6.569343065693431, | |
| "grad_norm": 0.221001997590065, | |
| "learning_rate": 0.0003575757575757576, | |
| "loss": 0.0062, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 6.751824817518248, | |
| "grad_norm": 0.1375964730978012, | |
| "learning_rate": 0.00033863636363636366, | |
| "loss": 0.0104, | |
| "step": 925 | |
| }, | |
| { | |
| "epoch": 6.934306569343065, | |
| "grad_norm": 0.18805921077728271, | |
| "learning_rate": 0.0003196969696969697, | |
| "loss": 0.0108, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_loss": 0.006272537633776665, | |
| "eval_runtime": 189.9453, | |
| "eval_samples_per_second": 4.307, | |
| "eval_steps_per_second": 0.721, | |
| "step": 959 | |
| }, | |
| { | |
| "epoch": 7.116788321167883, | |
| "grad_norm": 0.04299464076757431, | |
| "learning_rate": 0.0003007575757575758, | |
| "loss": 0.0065, | |
| "step": 975 | |
| }, | |
| { | |
| "epoch": 7.299270072992701, | |
| "grad_norm": 0.19402629137039185, | |
| "learning_rate": 0.0002818181818181818, | |
| "loss": 0.0089, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 7.481751824817518, | |
| "grad_norm": 0.1514054536819458, | |
| "learning_rate": 0.0002628787878787879, | |
| "loss": 0.007, | |
| "step": 1025 | |
| }, | |
| { | |
| "epoch": 7.664233576642336, | |
| "grad_norm": 0.011143018491566181, | |
| "learning_rate": 0.00024393939393939392, | |
| "loss": 0.0074, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 7.846715328467154, | |
| "grad_norm": 0.037931449711322784, | |
| "learning_rate": 0.00022500000000000002, | |
| "loss": 0.004, | |
| "step": 1075 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 0.003884958801791072, | |
| "eval_runtime": 190.9247, | |
| "eval_samples_per_second": 4.284, | |
| "eval_steps_per_second": 0.718, | |
| "step": 1096 | |
| }, | |
| { | |
| "epoch": 8.02919708029197, | |
| "grad_norm": 0.09535694122314453, | |
| "learning_rate": 0.00020606060606060607, | |
| "loss": 0.0047, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 8.211678832116789, | |
| "grad_norm": 0.1803082972764969, | |
| "learning_rate": 0.00018712121212121214, | |
| "loss": 0.0047, | |
| "step": 1125 | |
| }, | |
| { | |
| "epoch": 8.394160583941606, | |
| "grad_norm": 0.07427278906106949, | |
| "learning_rate": 0.0001681818181818182, | |
| "loss": 0.0035, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 8.576642335766424, | |
| "grad_norm": 0.07938549667596817, | |
| "learning_rate": 0.00014924242424242424, | |
| "loss": 0.0038, | |
| "step": 1175 | |
| }, | |
| { | |
| "epoch": 8.75912408759124, | |
| "grad_norm": 0.047964416444301605, | |
| "learning_rate": 0.0001303030303030303, | |
| "loss": 0.0035, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 8.941605839416058, | |
| "grad_norm": 0.03935278207063675, | |
| "learning_rate": 0.00011136363636363636, | |
| "loss": 0.0039, | |
| "step": 1225 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_loss": 0.002366352593526244, | |
| "eval_runtime": 191.3767, | |
| "eval_samples_per_second": 4.274, | |
| "eval_steps_per_second": 0.716, | |
| "step": 1233 | |
| }, | |
| { | |
| "epoch": 9.124087591240876, | |
| "grad_norm": 0.06493715196847916, | |
| "learning_rate": 9.242424242424242e-05, | |
| "loss": 0.0025, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 9.306569343065693, | |
| "grad_norm": 0.02973032370209694, | |
| "learning_rate": 7.348484848484849e-05, | |
| "loss": 0.0031, | |
| "step": 1275 | |
| }, | |
| { | |
| "epoch": 9.489051094890511, | |
| "grad_norm": 0.03156217932701111, | |
| "learning_rate": 5.4545454545454546e-05, | |
| "loss": 0.0013, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 9.671532846715328, | |
| "grad_norm": 0.01587042771279812, | |
| "learning_rate": 3.560606060606061e-05, | |
| "loss": 0.0036, | |
| "step": 1325 | |
| }, | |
| { | |
| "epoch": 9.854014598540147, | |
| "grad_norm": 0.02338283136487007, | |
| "learning_rate": 1.6666666666666667e-05, | |
| "loss": 0.0026, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_loss": 0.001961252186447382, | |
| "eval_runtime": 191.871, | |
| "eval_samples_per_second": 4.263, | |
| "eval_steps_per_second": 0.714, | |
| "step": 1370 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 1370, | |
| "total_flos": 2.4023145037824e+18, | |
| "train_loss": 0.1333546004319278, | |
| "train_runtime": 4360.7512, | |
| "train_samples_per_second": 1.876, | |
| "train_steps_per_second": 0.314 | |
| } | |
| ], | |
| "logging_steps": 25, | |
| "max_steps": 1370, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "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": 2.4023145037824e+18, | |
| "train_batch_size": 6, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |