| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.971014492753623, | |
| "eval_steps": 12, | |
| "global_step": 92, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.02, | |
| "grad_norm": 0.9959812474764307, | |
| "learning_rate": 0.0001999417022366174, | |
| "loss": 2.8926, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "eval_loss": 2.7616989612579346, | |
| "eval_runtime": 62.0214, | |
| "eval_samples_per_second": 322.227, | |
| "eval_steps_per_second": 20.154, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "grad_norm": 0.9553562375664082, | |
| "learning_rate": 0.00019976687691905393, | |
| "loss": 2.7178, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "grad_norm": 1.074690670043139, | |
| "learning_rate": 0.00019947572788580947, | |
| "loss": 2.3477, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "grad_norm": 1.2607308881446835, | |
| "learning_rate": 0.00019906859460363307, | |
| "loss": 2.0869, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "grad_norm": 1.5197691743409516, | |
| "learning_rate": 0.00019854595177171968, | |
| "loss": 1.8477, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 0.7112299977659389, | |
| "learning_rate": 0.00019790840876823232, | |
| "loss": 1.7227, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "grad_norm": 0.7680093267515216, | |
| "learning_rate": 0.00019715670893979414, | |
| "loss": 1.6553, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "grad_norm": 0.7744797513985666, | |
| "learning_rate": 0.00019629172873477995, | |
| "loss": 1.6455, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 0.735518802506217, | |
| "learning_rate": 0.00019531447668141608, | |
| "loss": 1.5508, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "grad_norm": 0.4535717241942791, | |
| "learning_rate": 0.00019422609221188207, | |
| "loss": 1.5039, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 0.29477652700703394, | |
| "learning_rate": 0.0001930278443337833, | |
| "loss": 1.502, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 0.24434475566463373, | |
| "learning_rate": 0.00019172113015054532, | |
| "loss": 1.4502, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "eval_loss": 1.456370234489441, | |
| "eval_runtime": 63.0041, | |
| "eval_samples_per_second": 317.202, | |
| "eval_steps_per_second": 19.84, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "grad_norm": 0.45228370307484317, | |
| "learning_rate": 0.00019030747323245327, | |
| "loss": 1.4277, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "grad_norm": 0.46523254227617566, | |
| "learning_rate": 0.0001887885218402375, | |
| "loss": 1.4097, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "grad_norm": 0.2855872854921807, | |
| "learning_rate": 0.00018716604700327514, | |
| "loss": 1.3833, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "grad_norm": 0.25447863105741503, | |
| "learning_rate": 0.00018544194045464886, | |
| "loss": 1.3652, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "grad_norm": 0.3619444462250106, | |
| "learning_rate": 0.0001836182124254711, | |
| "loss": 1.3677, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "grad_norm": 0.36635956863370217, | |
| "learning_rate": 0.0001816969893010442, | |
| "loss": 1.332, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "grad_norm": 0.39509330975297474, | |
| "learning_rate": 0.00017968051114159047, | |
| "loss": 1.3086, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "grad_norm": 0.3249913054000147, | |
| "learning_rate": 0.000177571129070442, | |
| "loss": 1.3071, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "grad_norm": 0.30914838660163285, | |
| "learning_rate": 0.00017537130253273613, | |
| "loss": 1.2729, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "grad_norm": 0.42289317853499514, | |
| "learning_rate": 0.00017308359642781242, | |
| "loss": 1.2773, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "grad_norm": 0.35414226064550436, | |
| "learning_rate": 0.00017071067811865476, | |
| "loss": 1.2417, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "grad_norm": 1.5532132927649234, | |
| "learning_rate": 0.00016825531432186543, | |
| "loss": 1.7617, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "eval_loss": 1.314684271812439, | |
| "eval_runtime": 63.1261, | |
| "eval_samples_per_second": 316.588, | |
| "eval_steps_per_second": 19.802, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "grad_norm": 0.12961326680655122, | |
| "learning_rate": 0.00016572036788179727, | |
| "loss": 1.2183, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "grad_norm": 0.1963943275104468, | |
| "learning_rate": 0.00016310879443260528, | |
| "loss": 1.2222, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "grad_norm": 0.21010667932420385, | |
| "learning_rate": 0.00016042363895210946, | |
| "loss": 1.2163, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "grad_norm": 0.2259721381332128, | |
| "learning_rate": 0.00015766803221148673, | |
| "loss": 1.2217, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "grad_norm": 0.19095866752454876, | |
| "learning_rate": 0.00015484518712493187, | |
| "loss": 1.2212, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "grad_norm": 0.11496636776764421, | |
| "learning_rate": 0.00015195839500354335, | |
| "loss": 1.2212, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "grad_norm": 0.12075046364081166, | |
| "learning_rate": 0.00014901102171780174, | |
| "loss": 1.1616, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "grad_norm": 0.21352779335608432, | |
| "learning_rate": 0.00014600650377311522, | |
| "loss": 1.2002, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "grad_norm": 0.20188789347169253, | |
| "learning_rate": 0.0001429483443030082, | |
| "loss": 1.2236, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "grad_norm": 0.1547188722220508, | |
| "learning_rate": 0.00013984010898462416, | |
| "loss": 1.1826, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "grad_norm": 0.11634196980431331, | |
| "learning_rate": 0.00013668542188130566, | |
| "loss": 1.209, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "grad_norm": 0.11144391112143442, | |
| "learning_rate": 0.00013348796121709862, | |
| "loss": 1.2051, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "eval_loss": 1.2780728340148926, | |
| "eval_runtime": 63.136, | |
| "eval_samples_per_second": 316.539, | |
| "eval_steps_per_second": 19.799, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 0.16076857117392804, | |
| "learning_rate": 0.0001302514550881076, | |
| "loss": 1.2046, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "grad_norm": 0.15687324678725567, | |
| "learning_rate": 0.00012697967711570242, | |
| "loss": 1.21, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "grad_norm": 0.4130016628721442, | |
| "learning_rate": 0.00012367644204664468, | |
| "loss": 1.5093, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "grad_norm": 0.1343313559360696, | |
| "learning_rate": 0.0001203456013052634, | |
| "loss": 1.2051, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "grad_norm": 0.11359323094812154, | |
| "learning_rate": 0.00011699103850286669, | |
| "loss": 1.2031, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "grad_norm": 0.09939941549415923, | |
| "learning_rate": 0.00011361666490962468, | |
| "loss": 1.1929, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "grad_norm": 0.09916599553279144, | |
| "learning_rate": 0.00011022641489420342, | |
| "loss": 1.1748, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "grad_norm": 0.14209942605129547, | |
| "learning_rate": 0.0001068242413364671, | |
| "loss": 1.1665, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "grad_norm": 0.12442345202001188, | |
| "learning_rate": 0.00010341411101859679, | |
| "loss": 1.1768, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 0.26217962366852043, | |
| "learning_rate": 0.0001, | |
| "loss": 1.5122, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 1.02, | |
| "grad_norm": 0.09589512696932487, | |
| "learning_rate": 9.658588898140322e-05, | |
| "loss": 1.1147, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 1.01, | |
| "grad_norm": 0.08883256552206654, | |
| "learning_rate": 9.317575866353292e-05, | |
| "loss": 1.1353, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 1.01, | |
| "eval_loss": 1.2603201866149902, | |
| "eval_runtime": 63.1562, | |
| "eval_samples_per_second": 316.438, | |
| "eval_steps_per_second": 19.792, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 1.04, | |
| "grad_norm": 0.12476996026721204, | |
| "learning_rate": 8.977358510579657e-05, | |
| "loss": 1.1499, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 1.06, | |
| "grad_norm": 0.1410594758409736, | |
| "learning_rate": 8.638333509037536e-05, | |
| "loss": 1.1636, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 1.08, | |
| "grad_norm": 0.13805158665360767, | |
| "learning_rate": 8.300896149713334e-05, | |
| "loss": 1.1592, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "grad_norm": 0.11182820786743164, | |
| "learning_rate": 7.965439869473664e-05, | |
| "loss": 1.189, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 1.12, | |
| "grad_norm": 0.08415581979953264, | |
| "learning_rate": 7.632355795335533e-05, | |
| "loss": 1.1362, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 1.14, | |
| "grad_norm": 0.08283198190447906, | |
| "learning_rate": 7.302032288429756e-05, | |
| "loss": 1.1543, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 1.17, | |
| "grad_norm": 0.09566331541455832, | |
| "learning_rate": 6.974854491189243e-05, | |
| "loss": 1.1475, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 1.19, | |
| "grad_norm": 0.12102788241266879, | |
| "learning_rate": 6.651203878290139e-05, | |
| "loss": 1.123, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "grad_norm": 0.12120546678963659, | |
| "learning_rate": 6.331457811869437e-05, | |
| "loss": 1.1235, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 1.23, | |
| "grad_norm": 0.10373804201590558, | |
| "learning_rate": 6.015989101537586e-05, | |
| "loss": 1.1987, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "grad_norm": 0.08553749701513035, | |
| "learning_rate": 5.7051655696991826e-05, | |
| "loss": 1.1616, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 1.28, | |
| "grad_norm": 0.07986407048448123, | |
| "learning_rate": 5.399349622688479e-05, | |
| "loss": 1.1787, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.28, | |
| "eval_loss": 1.2497843503952026, | |
| "eval_runtime": 63.2115, | |
| "eval_samples_per_second": 316.161, | |
| "eval_steps_per_second": 19.775, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.3, | |
| "grad_norm": 0.07539804797614923, | |
| "learning_rate": 5.0988978282198305e-05, | |
| "loss": 1.125, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 1.32, | |
| "grad_norm": 0.09159666914568447, | |
| "learning_rate": 4.804160499645667e-05, | |
| "loss": 1.1128, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 1.34, | |
| "grad_norm": 0.1010418577495551, | |
| "learning_rate": 4.515481287506811e-05, | |
| "loss": 1.1729, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 1.36, | |
| "grad_norm": 0.11598213897408995, | |
| "learning_rate": 4.2331967788513295e-05, | |
| "loss": 1.1606, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 1.38, | |
| "grad_norm": 0.09017416588433445, | |
| "learning_rate": 3.9576361047890554e-05, | |
| "loss": 1.1924, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 1.41, | |
| "grad_norm": 0.07860813360953343, | |
| "learning_rate": 3.689120556739475e-05, | |
| "loss": 1.1636, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 1.43, | |
| "grad_norm": 0.07385649237022004, | |
| "learning_rate": 3.427963211820274e-05, | |
| "loss": 1.1133, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 1.45, | |
| "grad_norm": 0.07323818513742975, | |
| "learning_rate": 3.174468567813461e-05, | |
| "loss": 1.1528, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 1.47, | |
| "grad_norm": 0.07313504011051983, | |
| "learning_rate": 2.9289321881345254e-05, | |
| "loss": 1.1528, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 1.49, | |
| "grad_norm": 0.2493580533486283, | |
| "learning_rate": 2.691640357218759e-05, | |
| "loss": 1.4531, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.51, | |
| "grad_norm": 0.08492071745345914, | |
| "learning_rate": 2.4628697467263918e-05, | |
| "loss": 1.1367, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 1.54, | |
| "grad_norm": 0.09336318243714266, | |
| "learning_rate": 2.242887092955801e-05, | |
| "loss": 1.1416, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 1.54, | |
| "eval_loss": 1.2444965839385986, | |
| "eval_runtime": 63.1509, | |
| "eval_samples_per_second": 316.464, | |
| "eval_steps_per_second": 19.794, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 1.56, | |
| "grad_norm": 0.08651448401545622, | |
| "learning_rate": 2.0319488858409553e-05, | |
| "loss": 1.1528, | |
| "step": 73 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "grad_norm": 0.08082722437043596, | |
| "learning_rate": 1.8303010698955804e-05, | |
| "loss": 1.1323, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "grad_norm": 0.07673966178613437, | |
| "learning_rate": 1.638178757452894e-05, | |
| "loss": 1.1548, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 1.62, | |
| "grad_norm": 0.07500837050855058, | |
| "learning_rate": 1.4558059545351143e-05, | |
| "loss": 1.1113, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 1.64, | |
| "grad_norm": 0.07854170828608634, | |
| "learning_rate": 1.2833952996724863e-05, | |
| "loss": 1.1426, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 1.67, | |
| "grad_norm": 0.07331164880963319, | |
| "learning_rate": 1.1211478159762478e-05, | |
| "loss": 1.1299, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 1.69, | |
| "grad_norm": 0.07647451775596137, | |
| "learning_rate": 9.692526767546729e-06, | |
| "loss": 1.1191, | |
| "step": 79 | |
| }, | |
| { | |
| "epoch": 1.71, | |
| "grad_norm": 0.07236874800404586, | |
| "learning_rate": 8.278869849454718e-06, | |
| "loss": 1.1167, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.73, | |
| "grad_norm": 0.06955020330114496, | |
| "learning_rate": 6.972155666216684e-06, | |
| "loss": 1.1191, | |
| "step": 81 | |
| }, | |
| { | |
| "epoch": 1.75, | |
| "grad_norm": 0.07541349884268325, | |
| "learning_rate": 5.77390778811796e-06, | |
| "loss": 1.1294, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 1.78, | |
| "grad_norm": 0.07686237827636627, | |
| "learning_rate": 4.685523318583918e-06, | |
| "loss": 1.146, | |
| "step": 83 | |
| }, | |
| { | |
| "epoch": 1.8, | |
| "grad_norm": 0.0739857032690884, | |
| "learning_rate": 3.7082712652200867e-06, | |
| "loss": 1.1606, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 1.8, | |
| "eval_loss": 1.243038535118103, | |
| "eval_runtime": 63.1702, | |
| "eval_samples_per_second": 316.368, | |
| "eval_steps_per_second": 19.788, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 1.82, | |
| "grad_norm": 0.07147810640105105, | |
| "learning_rate": 2.843291060205855e-06, | |
| "loss": 1.1396, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 1.84, | |
| "grad_norm": 0.07131641333572095, | |
| "learning_rate": 2.091591231767709e-06, | |
| "loss": 1.1255, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 1.86, | |
| "grad_norm": 0.08365332706415897, | |
| "learning_rate": 1.4540482282803137e-06, | |
| "loss": 1.1362, | |
| "step": 87 | |
| }, | |
| { | |
| "epoch": 1.88, | |
| "grad_norm": 0.08023410527314905, | |
| "learning_rate": 9.314053963669245e-07, | |
| "loss": 1.1489, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 1.91, | |
| "grad_norm": 0.07738378998352305, | |
| "learning_rate": 5.24272114190516e-07, | |
| "loss": 1.1167, | |
| "step": 89 | |
| }, | |
| { | |
| "epoch": 1.93, | |
| "grad_norm": 0.07310454468648853, | |
| "learning_rate": 2.3312308094607382e-07, | |
| "loss": 1.1201, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.95, | |
| "grad_norm": 0.07335973208363882, | |
| "learning_rate": 5.8297763382597626e-08, | |
| "loss": 1.1724, | |
| "step": 91 | |
| }, | |
| { | |
| "epoch": 1.97, | |
| "grad_norm": 0.07850704063156043, | |
| "learning_rate": 0.0, | |
| "loss": 1.124, | |
| "step": 92 | |
| } | |
| ], | |
| "logging_steps": 1, | |
| "max_steps": 92, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 2, | |
| "save_steps": 46, | |
| "total_flos": 2.193460016625746e+17, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |