| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 3.7560975609756095, | |
| "eval_steps": 3, | |
| "global_step": 40, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0975609756097561, | |
| "grad_norm": 0.13416488468647003, | |
| "learning_rate": 2e-05, | |
| "loss": 1.1198, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.0975609756097561, | |
| "eval_loss": 1.0946272611618042, | |
| "eval_runtime": 105.1442, | |
| "eval_samples_per_second": 0.951, | |
| "eval_steps_per_second": 0.476, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.1951219512195122, | |
| "grad_norm": 0.14614562690258026, | |
| "learning_rate": 4e-05, | |
| "loss": 1.1433, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.2926829268292683, | |
| "grad_norm": 0.12903447449207306, | |
| "learning_rate": 6e-05, | |
| "loss": 1.0338, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.2926829268292683, | |
| "eval_loss": 1.0899513959884644, | |
| "eval_runtime": 105.534, | |
| "eval_samples_per_second": 0.948, | |
| "eval_steps_per_second": 0.474, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.3902439024390244, | |
| "grad_norm": 0.12645691633224487, | |
| "learning_rate": 8e-05, | |
| "loss": 1.0196, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.4878048780487805, | |
| "grad_norm": 0.15484200417995453, | |
| "learning_rate": 0.0001, | |
| "loss": 1.0868, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.5853658536585366, | |
| "grad_norm": 0.13987194001674652, | |
| "learning_rate": 0.00012, | |
| "loss": 1.0658, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.5853658536585366, | |
| "eval_loss": 1.0339258909225464, | |
| "eval_runtime": 105.5831, | |
| "eval_samples_per_second": 0.947, | |
| "eval_steps_per_second": 0.474, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.6829268292682927, | |
| "grad_norm": 0.10199978202581406, | |
| "learning_rate": 0.00014, | |
| "loss": 1.0156, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.7804878048780488, | |
| "grad_norm": 0.12288854271173477, | |
| "learning_rate": 0.00016, | |
| "loss": 1.0503, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.8780487804878049, | |
| "grad_norm": 0.12781274318695068, | |
| "learning_rate": 0.00018, | |
| "loss": 1.0521, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.8780487804878049, | |
| "eval_loss": 0.9954264163970947, | |
| "eval_runtime": 105.5993, | |
| "eval_samples_per_second": 0.947, | |
| "eval_steps_per_second": 0.473, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.975609756097561, | |
| "grad_norm": 0.11168772727251053, | |
| "learning_rate": 0.0002, | |
| "loss": 1.0098, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 1.024390243902439, | |
| "grad_norm": 1.0994621515274048, | |
| "learning_rate": 0.00019945218953682734, | |
| "loss": 1.0201, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 1.1219512195121952, | |
| "grad_norm": 0.13263079524040222, | |
| "learning_rate": 0.00019781476007338058, | |
| "loss": 1.0211, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 1.1219512195121952, | |
| "eval_loss": 0.980489194393158, | |
| "eval_runtime": 105.5984, | |
| "eval_samples_per_second": 0.947, | |
| "eval_steps_per_second": 0.473, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 1.2195121951219512, | |
| "grad_norm": 0.10054657608270645, | |
| "learning_rate": 0.00019510565162951537, | |
| "loss": 0.9632, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 1.3170731707317074, | |
| "grad_norm": 0.09211089462041855, | |
| "learning_rate": 0.0001913545457642601, | |
| "loss": 0.9381, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 1.4146341463414633, | |
| "grad_norm": 0.0957920104265213, | |
| "learning_rate": 0.00018660254037844388, | |
| "loss": 0.9945, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 1.4146341463414633, | |
| "eval_loss": 0.9690712094306946, | |
| "eval_runtime": 105.5491, | |
| "eval_samples_per_second": 0.947, | |
| "eval_steps_per_second": 0.474, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 1.5121951219512195, | |
| "grad_norm": 0.11032579094171524, | |
| "learning_rate": 0.00018090169943749476, | |
| "loss": 0.9766, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 1.6097560975609757, | |
| "grad_norm": 0.08608294278383255, | |
| "learning_rate": 0.00017431448254773944, | |
| "loss": 0.9178, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 1.7073170731707317, | |
| "grad_norm": 0.08259813487529755, | |
| "learning_rate": 0.00016691306063588583, | |
| "loss": 0.967, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 1.7073170731707317, | |
| "eval_loss": 0.9576988220214844, | |
| "eval_runtime": 105.558, | |
| "eval_samples_per_second": 0.947, | |
| "eval_steps_per_second": 0.474, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 1.8048780487804879, | |
| "grad_norm": 0.08044092357158661, | |
| "learning_rate": 0.00015877852522924732, | |
| "loss": 1.0114, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 1.9024390243902438, | |
| "grad_norm": 0.08861377835273743, | |
| "learning_rate": 0.00015000000000000001, | |
| "loss": 1.0564, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.07940205931663513, | |
| "learning_rate": 0.00014067366430758004, | |
| "loss": 0.9694, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 0.9524104595184326, | |
| "eval_runtime": 105.5393, | |
| "eval_samples_per_second": 0.948, | |
| "eval_steps_per_second": 0.474, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 2.048780487804878, | |
| "grad_norm": 0.11204800009727478, | |
| "learning_rate": 0.00013090169943749476, | |
| "loss": 0.9271, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 2.1463414634146343, | |
| "grad_norm": 0.08592814207077026, | |
| "learning_rate": 0.00012079116908177593, | |
| "loss": 0.9159, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 2.2439024390243905, | |
| "grad_norm": 0.0843978151679039, | |
| "learning_rate": 0.00011045284632676536, | |
| "loss": 0.9627, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 2.2439024390243905, | |
| "eval_loss": 0.9468244314193726, | |
| "eval_runtime": 105.5461, | |
| "eval_samples_per_second": 0.947, | |
| "eval_steps_per_second": 0.474, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 2.341463414634146, | |
| "grad_norm": 0.08427581936120987, | |
| "learning_rate": 0.0001, | |
| "loss": 0.9047, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 2.4390243902439024, | |
| "grad_norm": 0.10019209235906601, | |
| "learning_rate": 8.954715367323468e-05, | |
| "loss": 0.9646, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 2.5365853658536586, | |
| "grad_norm": 0.09277612715959549, | |
| "learning_rate": 7.920883091822408e-05, | |
| "loss": 0.9598, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 2.5365853658536586, | |
| "eval_loss": 0.944459855556488, | |
| "eval_runtime": 105.5362, | |
| "eval_samples_per_second": 0.948, | |
| "eval_steps_per_second": 0.474, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 2.6341463414634148, | |
| "grad_norm": 0.09026714414358139, | |
| "learning_rate": 6.909830056250527e-05, | |
| "loss": 0.9146, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 2.7317073170731705, | |
| "grad_norm": 0.08944018930196762, | |
| "learning_rate": 5.9326335692419995e-05, | |
| "loss": 0.8949, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 2.8292682926829267, | |
| "grad_norm": 0.09638389945030212, | |
| "learning_rate": 5.000000000000002e-05, | |
| "loss": 0.9156, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 2.8292682926829267, | |
| "eval_loss": 0.9444054961204529, | |
| "eval_runtime": 105.5151, | |
| "eval_samples_per_second": 0.948, | |
| "eval_steps_per_second": 0.474, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 2.926829268292683, | |
| "grad_norm": 0.09340815991163254, | |
| "learning_rate": 4.12214747707527e-05, | |
| "loss": 0.9134, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 3.024390243902439, | |
| "grad_norm": 0.10298282653093338, | |
| "learning_rate": 3.308693936411421e-05, | |
| "loss": 0.9466, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 3.073170731707317, | |
| "grad_norm": 0.11301551759243011, | |
| "learning_rate": 2.5685517452260567e-05, | |
| "loss": 0.8781, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 3.073170731707317, | |
| "eval_loss": 0.9458777904510498, | |
| "eval_runtime": 105.5242, | |
| "eval_samples_per_second": 0.948, | |
| "eval_steps_per_second": 0.474, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 3.1707317073170733, | |
| "grad_norm": 0.09558753669261932, | |
| "learning_rate": 1.9098300562505266e-05, | |
| "loss": 0.904, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 3.2682926829268295, | |
| "grad_norm": 0.09189840406179428, | |
| "learning_rate": 1.339745962155613e-05, | |
| "loss": 0.9318, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 3.3658536585365852, | |
| "grad_norm": 0.09462269395589828, | |
| "learning_rate": 8.645454235739903e-06, | |
| "loss": 0.8903, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 3.3658536585365852, | |
| "eval_loss": 0.9447765946388245, | |
| "eval_runtime": 105.4779, | |
| "eval_samples_per_second": 0.948, | |
| "eval_steps_per_second": 0.474, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 3.4634146341463414, | |
| "grad_norm": 0.09194021672010422, | |
| "learning_rate": 4.8943483704846475e-06, | |
| "loss": 0.8289, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 3.5609756097560976, | |
| "grad_norm": 0.09003977477550507, | |
| "learning_rate": 2.1852399266194314e-06, | |
| "loss": 0.8691, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 3.658536585365854, | |
| "grad_norm": 0.09344612807035446, | |
| "learning_rate": 5.478104631726711e-07, | |
| "loss": 0.8687, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 3.658536585365854, | |
| "eval_loss": 0.9451260566711426, | |
| "eval_runtime": 105.4911, | |
| "eval_samples_per_second": 0.948, | |
| "eval_steps_per_second": 0.474, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 3.7560975609756095, | |
| "grad_norm": 0.0963226780295372, | |
| "learning_rate": 0.0, | |
| "loss": 1.005, | |
| "step": 40 | |
| } | |
| ], | |
| "logging_steps": 1, | |
| "max_steps": 40, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 10, | |
| "total_flos": 5.912132833404518e+16, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |