gptbase_hom2_100mb / trainer_state.json
i-be-snek's picture
Upload local model gptbase_hom2_100mb
bcdbe0e verified
Raw History Blame Contribute Delete
31.2 kB
{ "best_global_step": null, "best_metric": null, "best_model_checkpoint": null, "epoch": 9.999600372985213, "eval_steps": 3753, "global_step": 75067, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.0666045024643666, "grad_norm": 1.4687844514846802, "learning_rate": 6.648014921396217e-06, "loss": 9.4504, "step": 500 }, { "epoch": 0.1332090049287332, "grad_norm": 1.0154616832733154, "learning_rate": 1.3309352517985613e-05, "loss": 7.8097, "step": 1000 }, { "epoch": 0.19981350739309978, "grad_norm": 1.1075133085250854, "learning_rate": 1.9970690114575007e-05, "loss": 7.1387, "step": 1500 }, { "epoch": 0.2664180098574664, "grad_norm": 1.2340620756149292, "learning_rate": 2.6632027711164403e-05, "loss": 6.7753, "step": 2000 }, { "epoch": 0.33302251232183294, "grad_norm": 1.0900161266326904, "learning_rate": 3.32933653077538e-05, "loss": 6.5135, "step": 2500 }, { "epoch": 0.39962701478619955, "grad_norm": 1.099942922592163, "learning_rate": 3.9954702904343194e-05, "loss": 6.2828, "step": 3000 }, { "epoch": 0.46623151725056616, "grad_norm": 1.171620488166809, "learning_rate": 4.6616040500932586e-05, "loss": 6.084, "step": 3500 }, { "epoch": 0.4999333954975356, "eval_loss": 5.858362674713135, "eval_runtime": 40.6156, "eval_samples_per_second": 246.211, "eval_steps_per_second": 3.866, "step": 3753 }, { "epoch": 0.5328360197149328, "grad_norm": 1.062933087348938, "learning_rate": 5.327737809752199e-05, "loss": 5.904, "step": 4000 }, { "epoch": 0.5994405221792993, "grad_norm": 1.0676071643829346, "learning_rate": 5.993871569411138e-05, "loss": 5.759, "step": 4500 }, { "epoch": 0.6660450246436659, "grad_norm": 1.0839391946792603, "learning_rate": 6.660005329070078e-05, "loss": 5.6212, "step": 5000 }, { "epoch": 0.7326495271080325, "grad_norm": 1.1834132671356201, "learning_rate": 7.326139088729018e-05, "loss": 5.4985, "step": 5500 }, { "epoch": 0.7992540295723991, "grad_norm": 1.0901800394058228, "learning_rate": 7.992272848387957e-05, "loss": 5.3911, "step": 6000 }, { "epoch": 0.8658585320367657, "grad_norm": 1.00985848903656, "learning_rate": 8.658406608046896e-05, "loss": 5.2951, "step": 6500 }, { "epoch": 0.9324630345011323, "grad_norm": 1.0541818141937256, "learning_rate": 9.324540367705837e-05, "loss": 5.2063, "step": 7000 }, { "epoch": 0.9990675369654989, "grad_norm": 0.9412537813186646, "learning_rate": 9.990674127364776e-05, "loss": 5.1328, "step": 7500 }, { "epoch": 0.9998667909950713, "eval_loss": 5.0325398445129395, "eval_runtime": 40.6289, "eval_samples_per_second": 246.13, "eval_steps_per_second": 3.864, "step": 7506 }, { "epoch": 1.0656720394298655, "grad_norm": 1.041761040687561, "learning_rate": 9.92702890720978e-05, "loss": 5.0604, "step": 8000 }, { "epoch": 1.132276541894232, "grad_norm": 0.9710037708282471, "learning_rate": 9.853021713710573e-05, "loss": 4.991, "step": 8500 }, { "epoch": 1.1988810443585987, "grad_norm": 0.9740124344825745, "learning_rate": 9.779014520211365e-05, "loss": 4.9298, "step": 9000 }, { "epoch": 1.265485546822965, "grad_norm": 0.9319948554039001, "learning_rate": 9.705007326712156e-05, "loss": 4.8748, "step": 9500 }, { "epoch": 1.3320900492873318, "grad_norm": 0.9285650849342346, "learning_rate": 9.631000133212948e-05, "loss": 4.83, "step": 10000 }, { "epoch": 1.3986945517516984, "grad_norm": 0.9392498135566711, "learning_rate": 9.556992939713741e-05, "loss": 4.7835, "step": 10500 }, { "epoch": 1.4652990542160649, "grad_norm": 0.9685980081558228, "learning_rate": 9.482985746214532e-05, "loss": 4.7453, "step": 11000 }, { "epoch": 1.499800186492607, "eval_loss": 4.637198448181152, "eval_runtime": 40.6211, "eval_samples_per_second": 246.178, "eval_steps_per_second": 3.865, "step": 11259 }, { "epoch": 1.5319035566804315, "grad_norm": 0.9327673316001892, "learning_rate": 9.408978552715324e-05, "loss": 4.7043, "step": 11500 }, { "epoch": 1.5985080591447982, "grad_norm": 0.9539050459861755, "learning_rate": 9.334971359216115e-05, "loss": 4.6766, "step": 12000 }, { "epoch": 1.6651125616091647, "grad_norm": 0.9189325571060181, "learning_rate": 9.260964165716908e-05, "loss": 4.6414, "step": 12500 }, { "epoch": 1.7317170640735313, "grad_norm": 0.9299664497375488, "learning_rate": 9.1869569722177e-05, "loss": 4.6133, "step": 13000 }, { "epoch": 1.798321566537898, "grad_norm": 0.9159539937973022, "learning_rate": 9.112949778718491e-05, "loss": 4.5846, "step": 13500 }, { "epoch": 1.8649260690022644, "grad_norm": 0.9165326952934265, "learning_rate": 9.038942585219283e-05, "loss": 4.5607, "step": 14000 }, { "epoch": 1.931530571466631, "grad_norm": 0.9292281866073608, "learning_rate": 8.964935391720076e-05, "loss": 4.5331, "step": 14500 }, { "epoch": 1.9981350739309978, "grad_norm": 0.9381275773048401, "learning_rate": 8.890928198220867e-05, "loss": 4.5126, "step": 15000 }, { "epoch": 1.9997335819901425, "eval_loss": 4.451350212097168, "eval_runtime": 40.616, "eval_samples_per_second": 246.208, "eval_steps_per_second": 3.865, "step": 15012 }, { "epoch": 2.064739576395364, "grad_norm": 0.8954569697380066, "learning_rate": 8.816921004721659e-05, "loss": 4.4935, "step": 15500 }, { "epoch": 2.131344078859731, "grad_norm": 0.9439771771430969, "learning_rate": 8.742913811222452e-05, "loss": 4.473, "step": 16000 }, { "epoch": 2.1979485813240975, "grad_norm": 0.9190642237663269, "learning_rate": 8.668906617723243e-05, "loss": 4.4535, "step": 16500 }, { "epoch": 2.264553083788464, "grad_norm": 0.9132155179977417, "learning_rate": 8.594899424224036e-05, "loss": 4.4359, "step": 17000 }, { "epoch": 2.331157586252831, "grad_norm": 0.8906153440475464, "learning_rate": 8.520892230724828e-05, "loss": 4.4231, "step": 17500 }, { "epoch": 2.3977620887171973, "grad_norm": 0.8858339190483093, "learning_rate": 8.446885037225619e-05, "loss": 4.4051, "step": 18000 }, { "epoch": 2.4643665911815638, "grad_norm": 0.8869625926017761, "learning_rate": 8.37287784372641e-05, "loss": 4.3923, "step": 18500 }, { "epoch": 2.4996669774876783, "eval_loss": 4.3387041091918945, "eval_runtime": 40.6221, "eval_samples_per_second": 246.172, "eval_steps_per_second": 3.865, "step": 18765 }, { "epoch": 2.53097109364593, "grad_norm": 0.9139703512191772, "learning_rate": 8.298870650227204e-05, "loss": 4.3734, "step": 19000 }, { "epoch": 2.597575596110297, "grad_norm": 0.9062601327896118, "learning_rate": 8.224863456727995e-05, "loss": 4.3634, "step": 19500 }, { "epoch": 2.6641800985746635, "grad_norm": 0.8571125864982605, "learning_rate": 8.150856263228787e-05, "loss": 4.3486, "step": 20000 }, { "epoch": 2.73078460103903, "grad_norm": 0.8905714750289917, "learning_rate": 8.076849069729578e-05, "loss": 4.3346, "step": 20500 }, { "epoch": 2.797389103503397, "grad_norm": 0.8740853667259216, "learning_rate": 8.00284187623037e-05, "loss": 4.3209, "step": 21000 }, { "epoch": 2.8639936059677633, "grad_norm": 0.8983705639839172, "learning_rate": 7.928834682731163e-05, "loss": 4.3117, "step": 21500 }, { "epoch": 2.9305981084321298, "grad_norm": 0.8932312726974487, "learning_rate": 7.854827489231954e-05, "loss": 4.2944, "step": 22000 }, { "epoch": 2.9972026108964966, "grad_norm": 0.8963338732719421, "learning_rate": 7.780820295732746e-05, "loss": 4.2871, "step": 22500 }, { "epoch": 2.999600372985214, "eval_loss": 4.264514923095703, "eval_runtime": 40.6227, "eval_samples_per_second": 246.168, "eval_steps_per_second": 3.865, "step": 22518 }, { "epoch": 3.063807113360863, "grad_norm": 0.8693841695785522, "learning_rate": 7.706813102233537e-05, "loss": 4.2786, "step": 23000 }, { "epoch": 3.13041161582523, "grad_norm": 0.9077854752540588, "learning_rate": 7.63280590873433e-05, "loss": 4.2667, "step": 23500 }, { "epoch": 3.1970161182895964, "grad_norm": 0.8860239386558533, "learning_rate": 7.558798715235121e-05, "loss": 4.2561, "step": 24000 }, { "epoch": 3.263620620753963, "grad_norm": 0.8960526585578918, "learning_rate": 7.484791521735913e-05, "loss": 4.2464, "step": 24500 }, { "epoch": 3.3302251232183298, "grad_norm": 0.9195749759674072, "learning_rate": 7.410784328236705e-05, "loss": 4.2393, "step": 25000 }, { "epoch": 3.396829625682696, "grad_norm": 0.8777968883514404, "learning_rate": 7.336777134737497e-05, "loss": 4.2298, "step": 25500 }, { "epoch": 3.4634341281470626, "grad_norm": 0.8967837691307068, "learning_rate": 7.262769941238289e-05, "loss": 4.2223, "step": 26000 }, { "epoch": 3.4995337684827494, "eval_loss": 4.2124176025390625, "eval_runtime": 40.6314, "eval_samples_per_second": 246.115, "eval_steps_per_second": 3.864, "step": 26271 }, { "epoch": 3.530038630611429, "grad_norm": 0.8778949975967407, "learning_rate": 7.18876274773908e-05, "loss": 4.2093, "step": 26500 }, { "epoch": 3.596643133075796, "grad_norm": 0.8900518417358398, "learning_rate": 7.114755554239872e-05, "loss": 4.2048, "step": 27000 }, { "epoch": 3.6632476355401624, "grad_norm": 0.8986549377441406, "learning_rate": 7.040748360740665e-05, "loss": 4.1953, "step": 27500 }, { "epoch": 3.729852138004529, "grad_norm": 0.8861429691314697, "learning_rate": 6.966741167241456e-05, "loss": 4.1859, "step": 28000 }, { "epoch": 3.7964566404688957, "grad_norm": 0.9005793333053589, "learning_rate": 6.892733973742248e-05, "loss": 4.1769, "step": 28500 }, { "epoch": 3.863061142933262, "grad_norm": 0.8939288258552551, "learning_rate": 6.81872678024304e-05, "loss": 4.1722, "step": 29000 }, { "epoch": 3.9296656453976286, "grad_norm": 0.9196659326553345, "learning_rate": 6.744719586743832e-05, "loss": 4.1589, "step": 29500 }, { "epoch": 3.9962701478619955, "grad_norm": 0.897608757019043, "learning_rate": 6.670712393244624e-05, "loss": 4.1556, "step": 30000 }, { "epoch": 3.999467163980285, "eval_loss": 4.167255878448486, "eval_runtime": 40.6359, "eval_samples_per_second": 246.088, "eval_steps_per_second": 3.864, "step": 30024 }, { "epoch": 4.062874650326362, "grad_norm": 0.900414228439331, "learning_rate": 6.596705199745415e-05, "loss": 4.1499, "step": 30500 }, { "epoch": 4.129479152790728, "grad_norm": 0.8857269287109375, "learning_rate": 6.522698006246207e-05, "loss": 4.1424, "step": 31000 }, { "epoch": 4.196083655255095, "grad_norm": 0.8792352676391602, "learning_rate": 6.448690812747e-05, "loss": 4.1346, "step": 31500 }, { "epoch": 4.262688157719462, "grad_norm": 0.895592987537384, "learning_rate": 6.374683619247791e-05, "loss": 4.1286, "step": 32000 }, { "epoch": 4.329292660183828, "grad_norm": 0.8855577707290649, "learning_rate": 6.300676425748583e-05, "loss": 4.1234, "step": 32500 }, { "epoch": 4.395897162648195, "grad_norm": 0.885655403137207, "learning_rate": 6.226669232249374e-05, "loss": 4.1175, "step": 33000 }, { "epoch": 4.462501665112562, "grad_norm": 0.9013127684593201, "learning_rate": 6.152662038750167e-05, "loss": 4.1113, "step": 33500 }, { "epoch": 4.499400559477821, "eval_loss": 4.135107040405273, "eval_runtime": 40.6287, "eval_samples_per_second": 246.131, "eval_steps_per_second": 3.864, "step": 33777 }, { "epoch": 4.529106167576928, "grad_norm": 0.8668416142463684, "learning_rate": 6.078654845250958e-05, "loss": 4.1016, "step": 34000 }, { "epoch": 4.595710670041295, "grad_norm": 0.8878803849220276, "learning_rate": 6.00464765175175e-05, "loss": 4.1001, "step": 34500 }, { "epoch": 4.662315172505662, "grad_norm": 0.9031999707221985, "learning_rate": 5.930640458252542e-05, "loss": 4.092, "step": 35000 }, { "epoch": 4.728919674970028, "grad_norm": 0.8797792792320251, "learning_rate": 5.856633264753334e-05, "loss": 4.0854, "step": 35500 }, { "epoch": 4.795524177434395, "grad_norm": 0.9113759994506836, "learning_rate": 5.7826260712541255e-05, "loss": 4.0777, "step": 36000 }, { "epoch": 4.8621286798987615, "grad_norm": 0.8960673213005066, "learning_rate": 5.708618877754918e-05, "loss": 4.0758, "step": 36500 }, { "epoch": 4.9287331823631275, "grad_norm": 0.9208846688270569, "learning_rate": 5.634611684255709e-05, "loss": 4.064, "step": 37000 }, { "epoch": 4.995337684827494, "grad_norm": 0.9248899817466736, "learning_rate": 5.5606044907565015e-05, "loss": 4.0624, "step": 37500 }, { "epoch": 4.999333954975357, "eval_loss": 4.10422945022583, "eval_runtime": 40.6253, "eval_samples_per_second": 246.152, "eval_steps_per_second": 3.865, "step": 37530 }, { "epoch": 5.061942187291861, "grad_norm": 0.8935125470161438, "learning_rate": 5.486597297257293e-05, "loss": 4.0594, "step": 38000 }, { "epoch": 5.128546689756227, "grad_norm": 0.9079642295837402, "learning_rate": 5.412590103758085e-05, "loss": 4.0529, "step": 38500 }, { "epoch": 5.195151192220594, "grad_norm": 0.897830605506897, "learning_rate": 5.338582910258877e-05, "loss": 4.047, "step": 39000 }, { "epoch": 5.261755694684961, "grad_norm": 0.9031784534454346, "learning_rate": 5.264575716759669e-05, "loss": 4.0425, "step": 39500 }, { "epoch": 5.328360197149327, "grad_norm": 0.9324586987495422, "learning_rate": 5.1905685232604604e-05, "loss": 4.0394, "step": 40000 }, { "epoch": 5.394964699613694, "grad_norm": 0.9265775084495544, "learning_rate": 5.116561329761253e-05, "loss": 4.0346, "step": 40500 }, { "epoch": 5.461569202078061, "grad_norm": 0.9390904307365417, "learning_rate": 5.0425541362620455e-05, "loss": 4.0291, "step": 41000 }, { "epoch": 5.499267350472892, "eval_loss": 4.082469463348389, "eval_runtime": 40.6283, "eval_samples_per_second": 246.134, "eval_steps_per_second": 3.864, "step": 41283 }, { "epoch": 5.528173704542427, "grad_norm": 0.929745078086853, "learning_rate": 4.968546942762837e-05, "loss": 4.0217, "step": 41500 }, { "epoch": 5.594778207006794, "grad_norm": 0.9384097456932068, "learning_rate": 4.8945397492636286e-05, "loss": 4.0212, "step": 42000 }, { "epoch": 5.661382709471161, "grad_norm": 0.915998637676239, "learning_rate": 4.82053255576442e-05, "loss": 4.0145, "step": 42500 }, { "epoch": 5.727987211935527, "grad_norm": 0.9456606507301331, "learning_rate": 4.746525362265212e-05, "loss": 4.0096, "step": 43000 }, { "epoch": 5.7945917143998935, "grad_norm": 0.904883086681366, "learning_rate": 4.672518168766004e-05, "loss": 4.003, "step": 43500 }, { "epoch": 5.86119621686426, "grad_norm": 0.9260241985321045, "learning_rate": 4.598510975266796e-05, "loss": 4.002, "step": 44000 }, { "epoch": 5.927800719328626, "grad_norm": 0.9535908699035645, "learning_rate": 4.5245037817675875e-05, "loss": 3.9917, "step": 44500 }, { "epoch": 5.994405221792993, "grad_norm": 0.9389090538024902, "learning_rate": 4.45049658826838e-05, "loss": 3.9902, "step": 45000 }, { "epoch": 5.999200745970428, "eval_loss": 4.061779975891113, "eval_runtime": 40.6344, "eval_samples_per_second": 246.097, "eval_steps_per_second": 3.864, "step": 45036 }, { "epoch": 6.06100972425736, "grad_norm": 0.9462276101112366, "learning_rate": 4.376489394769171e-05, "loss": 3.9887, "step": 45500 }, { "epoch": 6.127614226721726, "grad_norm": 0.948383092880249, "learning_rate": 4.3024822012699635e-05, "loss": 3.9841, "step": 46000 }, { "epoch": 6.194218729186093, "grad_norm": 0.9472470879554749, "learning_rate": 4.228475007770756e-05, "loss": 3.9789, "step": 46500 }, { "epoch": 6.26082323165046, "grad_norm": 0.9430816769599915, "learning_rate": 4.154467814271548e-05, "loss": 3.975, "step": 47000 }, { "epoch": 6.327427734114826, "grad_norm": 0.9511832594871521, "learning_rate": 4.0804606207723394e-05, "loss": 3.9732, "step": 47500 }, { "epoch": 6.394032236579193, "grad_norm": 0.9368382692337036, "learning_rate": 4.0064534272731316e-05, "loss": 3.9695, "step": 48000 }, { "epoch": 6.46063673904356, "grad_norm": 0.9653155207633972, "learning_rate": 3.932446233773923e-05, "loss": 3.9644, "step": 48500 }, { "epoch": 6.499134141467963, "eval_loss": 4.045886516571045, "eval_runtime": 40.6206, "eval_samples_per_second": 246.181, "eval_steps_per_second": 3.865, "step": 48789 }, { "epoch": 6.527241241507926, "grad_norm": 0.9259459972381592, "learning_rate": 3.858439040274715e-05, "loss": 3.9589, "step": 49000 }, { "epoch": 6.593845743972293, "grad_norm": 0.9427011013031006, "learning_rate": 3.784431846775507e-05, "loss": 3.9586, "step": 49500 }, { "epoch": 6.6604502464366595, "grad_norm": 0.9405387043952942, "learning_rate": 3.710424653276299e-05, "loss": 3.9526, "step": 50000 }, { "epoch": 6.7270547489010255, "grad_norm": 0.9505395889282227, "learning_rate": 3.6364174597770906e-05, "loss": 3.9483, "step": 50500 }, { "epoch": 6.793659251365392, "grad_norm": 0.9568143486976624, "learning_rate": 3.562410266277882e-05, "loss": 3.943, "step": 51000 }, { "epoch": 6.860263753829759, "grad_norm": 0.9627091288566589, "learning_rate": 3.488403072778674e-05, "loss": 3.9422, "step": 51500 }, { "epoch": 6.926868256294125, "grad_norm": 0.9769137501716614, "learning_rate": 3.414395879279466e-05, "loss": 3.9333, "step": 52000 }, { "epoch": 6.993472758758492, "grad_norm": 0.9454171061515808, "learning_rate": 3.340388685780258e-05, "loss": 3.9325, "step": 52500 }, { "epoch": 6.999067536965499, "eval_loss": 4.02980375289917, "eval_runtime": 40.6315, "eval_samples_per_second": 246.114, "eval_steps_per_second": 3.864, "step": 52542 }, { "epoch": 7.060077261222859, "grad_norm": 0.954601526260376, "learning_rate": 3.2663814922810496e-05, "loss": 3.9315, "step": 53000 }, { "epoch": 7.126681763687225, "grad_norm": 1.0312658548355103, "learning_rate": 3.192374298781842e-05, "loss": 3.928, "step": 53500 }, { "epoch": 7.193286266151592, "grad_norm": 0.9952969551086426, "learning_rate": 3.118367105282633e-05, "loss": 3.9232, "step": 54000 }, { "epoch": 7.259890768615959, "grad_norm": 0.975281834602356, "learning_rate": 3.0443599117834255e-05, "loss": 3.9203, "step": 54500 }, { "epoch": 7.326495271080325, "grad_norm": 0.9661374092102051, "learning_rate": 2.9703527182842174e-05, "loss": 3.9191, "step": 55000 }, { "epoch": 7.393099773544692, "grad_norm": 0.9646915793418884, "learning_rate": 2.8963455247850092e-05, "loss": 3.916, "step": 55500 }, { "epoch": 7.459704276009059, "grad_norm": 0.978786826133728, "learning_rate": 2.822338331285801e-05, "loss": 3.9117, "step": 56000 }, { "epoch": 7.4990009324630345, "eval_loss": 4.01784610748291, "eval_runtime": 40.6307, "eval_samples_per_second": 246.119, "eval_steps_per_second": 3.864, "step": 56295 }, { "epoch": 7.526308778473425, "grad_norm": 0.9720600247383118, "learning_rate": 2.748331137786593e-05, "loss": 3.9065, "step": 56500 }, { "epoch": 7.5929132809377915, "grad_norm": 0.9529099464416504, "learning_rate": 2.6743239442873848e-05, "loss": 3.9067, "step": 57000 }, { "epoch": 7.659517783402158, "grad_norm": 0.9640529155731201, "learning_rate": 2.6003167507881763e-05, "loss": 3.9018, "step": 57500 }, { "epoch": 7.726122285866524, "grad_norm": 0.97901451587677, "learning_rate": 2.5263095572889682e-05, "loss": 3.898, "step": 58000 }, { "epoch": 7.792726788330891, "grad_norm": 0.99736487865448, "learning_rate": 2.4523023637897604e-05, "loss": 3.8929, "step": 58500 }, { "epoch": 7.859331290795257, "grad_norm": 0.9964234232902527, "learning_rate": 2.3782951702905523e-05, "loss": 3.8932, "step": 59000 }, { "epoch": 7.925935793259624, "grad_norm": 0.9927836060523987, "learning_rate": 2.304287976791344e-05, "loss": 3.8849, "step": 59500 }, { "epoch": 7.992540295723991, "grad_norm": 0.9759431481361389, "learning_rate": 2.230280783292136e-05, "loss": 3.8848, "step": 60000 }, { "epoch": 7.99893432796057, "eval_loss": 4.007284164428711, "eval_runtime": 40.6218, "eval_samples_per_second": 246.173, "eval_steps_per_second": 3.865, "step": 60048 }, { "epoch": 8.059144798188358, "grad_norm": 0.9723028540611267, "learning_rate": 2.156273589792928e-05, "loss": 3.8837, "step": 60500 }, { "epoch": 8.125749300652725, "grad_norm": 0.9921424388885498, "learning_rate": 2.0822663962937197e-05, "loss": 3.8815, "step": 61000 }, { "epoch": 8.19235380311709, "grad_norm": 1.0018419027328491, "learning_rate": 2.008259202794512e-05, "loss": 3.877, "step": 61500 }, { "epoch": 8.258958305581457, "grad_norm": 1.0002388954162598, "learning_rate": 1.9342520092953038e-05, "loss": 3.8742, "step": 62000 }, { "epoch": 8.325562808045824, "grad_norm": 1.0242356061935425, "learning_rate": 1.8602448157960956e-05, "loss": 3.8754, "step": 62500 }, { "epoch": 8.39216731051019, "grad_norm": 1.014195442199707, "learning_rate": 1.7862376222968875e-05, "loss": 3.8709, "step": 63000 }, { "epoch": 8.458771812974557, "grad_norm": 1.0065619945526123, "learning_rate": 1.7122304287976794e-05, "loss": 3.8686, "step": 63500 }, { "epoch": 8.498867723458106, "eval_loss": 3.9953315258026123, "eval_runtime": 40.6301, "eval_samples_per_second": 246.123, "eval_steps_per_second": 3.864, "step": 63801 }, { "epoch": 8.525376315438924, "grad_norm": 1.0098967552185059, "learning_rate": 1.6382232352984712e-05, "loss": 3.8625, "step": 64000 }, { "epoch": 8.59198081790329, "grad_norm": 0.9921189546585083, "learning_rate": 1.564216041799263e-05, "loss": 3.8647, "step": 64500 }, { "epoch": 8.658585320367656, "grad_norm": 0.9988556504249573, "learning_rate": 1.4902088483000548e-05, "loss": 3.8596, "step": 65000 }, { "epoch": 8.725189822832023, "grad_norm": 1.003258466720581, "learning_rate": 1.4162016548008467e-05, "loss": 3.8559, "step": 65500 }, { "epoch": 8.79179432529639, "grad_norm": 0.9999053478240967, "learning_rate": 1.3421944613016385e-05, "loss": 3.852, "step": 66000 }, { "epoch": 8.858398827760757, "grad_norm": 1.0318782329559326, "learning_rate": 1.2681872678024304e-05, "loss": 3.8525, "step": 66500 }, { "epoch": 8.925003330225124, "grad_norm": 1.0089142322540283, "learning_rate": 1.1941800743032224e-05, "loss": 3.845, "step": 67000 }, { "epoch": 8.991607832689489, "grad_norm": 1.0043953657150269, "learning_rate": 1.1201728808040143e-05, "loss": 3.8455, "step": 67500 }, { "epoch": 8.998801118955642, "eval_loss": 3.9886233806610107, "eval_runtime": 40.6223, "eval_samples_per_second": 246.17, "eval_steps_per_second": 3.865, "step": 67554 }, { "epoch": 9.058212335153856, "grad_norm": 1.0068721771240234, "learning_rate": 1.046165687304806e-05, "loss": 3.844, "step": 68000 }, { "epoch": 9.124816837618223, "grad_norm": 1.0410212278366089, "learning_rate": 9.721584938055978e-06, "loss": 3.8437, "step": 68500 }, { "epoch": 9.19142134008259, "grad_norm": 1.007806658744812, "learning_rate": 8.981513003063899e-06, "loss": 3.839, "step": 69000 }, { "epoch": 9.258025842546957, "grad_norm": 1.0123896598815918, "learning_rate": 8.241441068071817e-06, "loss": 3.8371, "step": 69500 }, { "epoch": 9.324630345011323, "grad_norm": 1.0174599885940552, "learning_rate": 7.501369133079736e-06, "loss": 3.8384, "step": 70000 }, { "epoch": 9.391234847475689, "grad_norm": 1.015404224395752, "learning_rate": 6.7612971980876545e-06, "loss": 3.8355, "step": 70500 }, { "epoch": 9.457839349940055, "grad_norm": 1.0169411897659302, "learning_rate": 6.021225263095573e-06, "loss": 3.8332, "step": 71000 }, { "epoch": 9.498734514453178, "eval_loss": 3.9816718101501465, "eval_runtime": 40.6341, "eval_samples_per_second": 246.099, "eval_steps_per_second": 3.864, "step": 71307 }, { "epoch": 9.524443852404422, "grad_norm": 1.0108140707015991, "learning_rate": 5.281153328103492e-06, "loss": 3.8277, "step": 71500 }, { "epoch": 9.59104835486879, "grad_norm": 1.0033539533615112, "learning_rate": 4.5410813931114104e-06, "loss": 3.83, "step": 72000 }, { "epoch": 9.657652857333156, "grad_norm": 1.0251843929290771, "learning_rate": 3.801009458119329e-06, "loss": 3.8262, "step": 72500 }, { "epoch": 9.724257359797523, "grad_norm": 1.016353964805603, "learning_rate": 3.060937523127248e-06, "loss": 3.822, "step": 73000 }, { "epoch": 9.790861862261888, "grad_norm": 1.0144379138946533, "learning_rate": 2.3208655881351668e-06, "loss": 3.8201, "step": 73500 }, { "epoch": 9.857466364726255, "grad_norm": 1.030078411102295, "learning_rate": 1.5807936531430856e-06, "loss": 3.8211, "step": 74000 }, { "epoch": 9.924070867190622, "grad_norm": 1.0287569761276245, "learning_rate": 8.407217181510043e-07, "loss": 3.814, "step": 74500 }, { "epoch": 9.990675369654989, "grad_norm": 1.0243396759033203, "learning_rate": 1.0064978315892306e-07, "loss": 3.8159, "step": 75000 }, { "epoch": 9.998667909950713, "eval_loss": 3.9775593280792236, "eval_runtime": 40.8011, "eval_samples_per_second": 245.091, "eval_steps_per_second": 3.848, "step": 75060 } ], "logging_steps": 500, "max_steps": 75067, "num_input_tokens_seen": 0, "num_train_epochs": 10, "save_steps": 3753, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": true }, "attributes": {} } }, "total_flos": 6.27639922262016e+17, "train_batch_size": 32, "trial_name": null, "trial_params": null }
{
"best_global_step": null,
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 9.999600372985213,
"eval_steps": 3753,
"global_step": 75067,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.0666045024643666,
"grad_norm": 1.4687844514846802,
"learning_rate": 6.648014921396217e-06,
"loss": 9.4504,
"step": 500
},
{
"epoch": 0.1332090049287332,
"grad_norm": 1.0154616832733154,
"learning_rate": 1.3309352517985613e-05,
"loss": 7.8097,
"step": 1000
},
{
"epoch": 0.19981350739309978,
"grad_norm": 1.1075133085250854,
"learning_rate": 1.9970690114575007e-05,
"loss": 7.1387,
"step": 1500
},
{
"epoch": 0.2664180098574664,
"grad_norm": 1.2340620756149292,
"learning_rate": 2.6632027711164403e-05,
"loss": 6.7753,
"step": 2000
},
{
"epoch": 0.33302251232183294,
"grad_norm": 1.0900161266326904,
"learning_rate": 3.32933653077538e-05,
"loss": 6.5135,
"step": 2500
},
{
"epoch": 0.39962701478619955,
"grad_norm": 1.099942922592163,
"learning_rate": 3.9954702904343194e-05,
"loss": 6.2828,
"step": 3000
},
{
"epoch": 0.46623151725056616,
"grad_norm": 1.171620488166809,
"learning_rate": 4.6616040500932586e-05,
"loss": 6.084,
"step": 3500
},
{
"epoch": 0.4999333954975356,
"eval_loss": 5.858362674713135,
"eval_runtime": 40.6156,
"eval_samples_per_second": 246.211,
"eval_steps_per_second": 3.866,
"step": 3753
},
{
"epoch": 0.5328360197149328,
"grad_norm": 1.062933087348938,
"learning_rate": 5.327737809752199e-05,
"loss": 5.904,
"step": 4000
},
{
"epoch": 0.5994405221792993,
"grad_norm": 1.0676071643829346,
"learning_rate": 5.993871569411138e-05,
"loss": 5.759,
"step": 4500
},
{
"epoch": 0.6660450246436659,
"grad_norm": 1.0839391946792603,
"learning_rate": 6.660005329070078e-05,
"loss": 5.6212,
"step": 5000
},
{
"epoch": 0.7326495271080325,
"grad_norm": 1.1834132671356201,
"learning_rate": 7.326139088729018e-05,
"loss": 5.4985,
"step": 5500
},
{
"epoch": 0.7992540295723991,
"grad_norm": 1.0901800394058228,
"learning_rate": 7.992272848387957e-05,