| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.0, | |
| "eval_steps": 500, | |
| "global_step": 6921, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.014448779078167894, | |
| "grad_norm": 0.1319790631532669, | |
| "learning_rate": 3.576589595375723e-07, | |
| "loss": 2.9859, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.02889755815633579, | |
| "grad_norm": 0.09819761663675308, | |
| "learning_rate": 7.153179190751446e-07, | |
| "loss": 2.9932, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.043346337234503686, | |
| "grad_norm": 0.1437925547361374, | |
| "learning_rate": 1.0729768786127169e-06, | |
| "loss": 2.9802, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.05779511631267158, | |
| "grad_norm": 0.1754620224237442, | |
| "learning_rate": 1.4306358381502892e-06, | |
| "loss": 2.951, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.07224389539083947, | |
| "grad_norm": 0.16474293172359467, | |
| "learning_rate": 1.791907514450867e-06, | |
| "loss": 2.9437, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.08669267446900737, | |
| "grad_norm": 0.19413641095161438, | |
| "learning_rate": 2.153179190751445e-06, | |
| "loss": 2.909, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.10114145354717527, | |
| "grad_norm": 0.21083518862724304, | |
| "learning_rate": 2.5144508670520235e-06, | |
| "loss": 2.8579, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.11559023262534315, | |
| "grad_norm": 0.16336515545845032, | |
| "learning_rate": 2.8757225433526016e-06, | |
| "loss": 2.8228, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.13003901170351106, | |
| "grad_norm": 0.20088078081607819, | |
| "learning_rate": 3.2369942196531797e-06, | |
| "loss": 2.8003, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.14448779078167895, | |
| "grad_norm": 0.15995082259178162, | |
| "learning_rate": 3.5982658959537574e-06, | |
| "loss": 2.7788, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.15893656985984683, | |
| "grad_norm": 0.15541793406009674, | |
| "learning_rate": 3.9595375722543355e-06, | |
| "loss": 2.731, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 0.17338534893801474, | |
| "grad_norm": 0.16839973628520966, | |
| "learning_rate": 4.320809248554913e-06, | |
| "loss": 2.71, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 0.18783412801618263, | |
| "grad_norm": 0.16297543048858643, | |
| "learning_rate": 4.682080924855492e-06, | |
| "loss": 2.6887, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 0.20228290709435054, | |
| "grad_norm": 0.1565973162651062, | |
| "learning_rate": 5.043352601156069e-06, | |
| "loss": 2.666, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 0.21673168617251842, | |
| "grad_norm": 0.13941773772239685, | |
| "learning_rate": 5.404624277456648e-06, | |
| "loss": 2.6619, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.2311804652506863, | |
| "grad_norm": 0.14650990068912506, | |
| "learning_rate": 5.765895953757226e-06, | |
| "loss": 2.6402, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 0.24562924432885422, | |
| "grad_norm": 0.16068528592586517, | |
| "learning_rate": 6.127167630057804e-06, | |
| "loss": 2.639, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 0.26007802340702213, | |
| "grad_norm": 0.16715380549430847, | |
| "learning_rate": 6.488439306358382e-06, | |
| "loss": 2.6312, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 0.27452680248519, | |
| "grad_norm": 0.2185903787612915, | |
| "learning_rate": 6.84971098265896e-06, | |
| "loss": 2.627, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 0.2889755815633579, | |
| "grad_norm": 0.16029363870620728, | |
| "learning_rate": 7.210982658959538e-06, | |
| "loss": 2.6231, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.3034243606415258, | |
| "grad_norm": 0.16060592234134674, | |
| "learning_rate": 7.5722543352601166e-06, | |
| "loss": 2.6127, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 0.31787313971969366, | |
| "grad_norm": 0.16263119876384735, | |
| "learning_rate": 7.933526011560694e-06, | |
| "loss": 2.6082, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 0.3323219187978616, | |
| "grad_norm": 0.1585131585597992, | |
| "learning_rate": 8.294797687861273e-06, | |
| "loss": 2.6037, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 0.3467706978760295, | |
| "grad_norm": 0.16766057908535004, | |
| "learning_rate": 8.656069364161851e-06, | |
| "loss": 2.6005, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 0.36121947695419737, | |
| "grad_norm": 0.1492636650800705, | |
| "learning_rate": 9.017341040462428e-06, | |
| "loss": 2.5915, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.37566825603236526, | |
| "grad_norm": 0.17886599898338318, | |
| "learning_rate": 9.378612716763007e-06, | |
| "loss": 2.5881, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 0.39011703511053314, | |
| "grad_norm": 0.18520204722881317, | |
| "learning_rate": 9.739884393063585e-06, | |
| "loss": 2.5915, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 0.4045658141887011, | |
| "grad_norm": 0.1520683318376541, | |
| "learning_rate": 1.0101156069364162e-05, | |
| "loss": 2.5806, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 0.41901459326686896, | |
| "grad_norm": 0.1854362040758133, | |
| "learning_rate": 1.046242774566474e-05, | |
| "loss": 2.5711, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 0.43346337234503685, | |
| "grad_norm": 0.18411140143871307, | |
| "learning_rate": 1.0823699421965319e-05, | |
| "loss": 2.5841, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.44791215142320473, | |
| "grad_norm": 0.1857396364212036, | |
| "learning_rate": 1.1184971098265898e-05, | |
| "loss": 2.5727, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 0.4623609305013726, | |
| "grad_norm": 0.23173969984054565, | |
| "learning_rate": 1.1546242774566474e-05, | |
| "loss": 2.5793, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 0.47680970957954055, | |
| "grad_norm": 0.17829596996307373, | |
| "learning_rate": 1.1907514450867053e-05, | |
| "loss": 2.574, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 0.49125848865770844, | |
| "grad_norm": 0.1748592108488083, | |
| "learning_rate": 1.2268786127167632e-05, | |
| "loss": 2.5675, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 0.5057072677358763, | |
| "grad_norm": 0.18341368436813354, | |
| "learning_rate": 1.263005780346821e-05, | |
| "loss": 2.5695, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 0.5201560468140443, | |
| "grad_norm": 0.17201048135757446, | |
| "learning_rate": 1.2991329479768787e-05, | |
| "loss": 2.565, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 0.5346048258922121, | |
| "grad_norm": 0.21022534370422363, | |
| "learning_rate": 1.3352601156069365e-05, | |
| "loss": 2.5662, | |
| "step": 3700 | |
| }, | |
| { | |
| "epoch": 0.54905360497038, | |
| "grad_norm": 0.17963945865631104, | |
| "learning_rate": 1.3713872832369944e-05, | |
| "loss": 2.5645, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 0.5635023840485479, | |
| "grad_norm": 0.18298958241939545, | |
| "learning_rate": 1.4075144508670523e-05, | |
| "loss": 2.5634, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 0.5779511631267158, | |
| "grad_norm": 0.221732035279274, | |
| "learning_rate": 1.4436416184971101e-05, | |
| "loss": 2.5532, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.5923999422048837, | |
| "grad_norm": 0.19882583618164062, | |
| "learning_rate": 1.4797687861271676e-05, | |
| "loss": 2.5595, | |
| "step": 4100 | |
| }, | |
| { | |
| "epoch": 0.6068487212830516, | |
| "grad_norm": 0.21510709822177887, | |
| "learning_rate": 1.5158959537572255e-05, | |
| "loss": 2.5632, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 0.6212975003612194, | |
| "grad_norm": 0.17533738911151886, | |
| "learning_rate": 1.5520231213872833e-05, | |
| "loss": 2.5448, | |
| "step": 4300 | |
| }, | |
| { | |
| "epoch": 0.6357462794393873, | |
| "grad_norm": 0.19394342601299286, | |
| "learning_rate": 1.5881502890173412e-05, | |
| "loss": 2.5597, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 0.6501950585175552, | |
| "grad_norm": 0.196193665266037, | |
| "learning_rate": 1.624277456647399e-05, | |
| "loss": 2.551, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 0.6646438375957232, | |
| "grad_norm": 0.1857462376356125, | |
| "learning_rate": 1.660404624277457e-05, | |
| "loss": 2.5478, | |
| "step": 4600 | |
| }, | |
| { | |
| "epoch": 0.6790926166738911, | |
| "grad_norm": 0.20186574757099152, | |
| "learning_rate": 1.6965317919075147e-05, | |
| "loss": 2.5508, | |
| "step": 4700 | |
| }, | |
| { | |
| "epoch": 0.693541395752059, | |
| "grad_norm": 0.22574672102928162, | |
| "learning_rate": 1.7326589595375726e-05, | |
| "loss": 2.5492, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 0.7079901748302269, | |
| "grad_norm": 0.21713367104530334, | |
| "learning_rate": 1.76878612716763e-05, | |
| "loss": 2.5326, | |
| "step": 4900 | |
| }, | |
| { | |
| "epoch": 0.7224389539083947, | |
| "grad_norm": 0.2183527648448944, | |
| "learning_rate": 1.804913294797688e-05, | |
| "loss": 2.5512, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.7368877329865626, | |
| "grad_norm": 0.23467661440372467, | |
| "learning_rate": 1.8410404624277458e-05, | |
| "loss": 2.5462, | |
| "step": 5100 | |
| }, | |
| { | |
| "epoch": 0.7513365120647305, | |
| "grad_norm": 0.20273762941360474, | |
| "learning_rate": 1.8771676300578037e-05, | |
| "loss": 2.5493, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 0.7657852911428984, | |
| "grad_norm": 0.20707377791404724, | |
| "learning_rate": 1.9132947976878615e-05, | |
| "loss": 2.5425, | |
| "step": 5300 | |
| }, | |
| { | |
| "epoch": 0.7802340702210663, | |
| "grad_norm": 0.20679670572280884, | |
| "learning_rate": 1.949421965317919e-05, | |
| "loss": 2.5424, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 0.7946828492992342, | |
| "grad_norm": 0.22911733388900757, | |
| "learning_rate": 1.985549132947977e-05, | |
| "loss": 2.5369, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 0.8091316283774022, | |
| "grad_norm": 0.2293398678302765, | |
| "learning_rate": 1.990752970486014e-05, | |
| "loss": 2.5414, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 0.82358040745557, | |
| "grad_norm": 0.21145911514759064, | |
| "learning_rate": 1.934861471129627e-05, | |
| "loss": 2.5341, | |
| "step": 5700 | |
| }, | |
| { | |
| "epoch": 0.8380291865337379, | |
| "grad_norm": 0.22642740607261658, | |
| "learning_rate": 1.8310755913154726e-05, | |
| "loss": 2.5312, | |
| "step": 5800 | |
| }, | |
| { | |
| "epoch": 0.8524779656119058, | |
| "grad_norm": 0.22286154329776764, | |
| "learning_rate": 1.6847124401206384e-05, | |
| "loss": 2.5354, | |
| "step": 5900 | |
| }, | |
| { | |
| "epoch": 0.8669267446900737, | |
| "grad_norm": 0.22432397305965424, | |
| "learning_rate": 1.5032704252395315e-05, | |
| "loss": 2.5355, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.8813755237682416, | |
| "grad_norm": 0.21558962762355804, | |
| "learning_rate": 1.2960450981095643e-05, | |
| "loss": 2.5481, | |
| "step": 6100 | |
| }, | |
| { | |
| "epoch": 0.8958243028464095, | |
| "grad_norm": 0.2449900358915329, | |
| "learning_rate": 1.0736529286191087e-05, | |
| "loss": 2.5352, | |
| "step": 6200 | |
| }, | |
| { | |
| "epoch": 0.9102730819245773, | |
| "grad_norm": 0.22407910227775574, | |
| "learning_rate": 8.474874071465144e-06, | |
| "loss": 2.5261, | |
| "step": 6300 | |
| }, | |
| { | |
| "epoch": 0.9247218610027452, | |
| "grad_norm": 0.20298534631729126, | |
| "learning_rate": 6.291353386531074e-06, | |
| "loss": 2.5389, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 0.9391706400809131, | |
| "grad_norm": 0.2112707942724228, | |
| "learning_rate": 4.297832329749687e-06, | |
| "loss": 2.5369, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 0.9536194191590811, | |
| "grad_norm": 0.22108633816242218, | |
| "learning_rate": 2.596442028451194e-06, | |
| "loss": 2.5331, | |
| "step": 6600 | |
| }, | |
| { | |
| "epoch": 0.968068198237249, | |
| "grad_norm": 0.21827644109725952, | |
| "learning_rate": 1.2743473053542842e-06, | |
| "loss": 2.5279, | |
| "step": 6700 | |
| }, | |
| { | |
| "epoch": 0.9825169773154169, | |
| "grad_norm": 0.22307871282100677, | |
| "learning_rate": 3.9928109162008953e-07, | |
| "loss": 2.5346, | |
| "step": 6800 | |
| }, | |
| { | |
| "epoch": 0.9969657563935848, | |
| "grad_norm": 0.22532746195793152, | |
| "learning_rate": 1.6074365378105915e-08, | |
| "loss": 2.5428, | |
| "step": 6900 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 6921, | |
| "total_flos": 1.0085141819823227e+18, | |
| "train_loss": 2.6225888340860033, | |
| "train_runtime": 2201.4938, | |
| "train_samples_per_second": 50.3, | |
| "train_steps_per_second": 3.144 | |
| } | |
| ], | |
| "logging_steps": 100, | |
| "max_steps": 6921, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 1.0085141819823227e+18, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |