| { | |
| "best_metric": 2.9402894973754883, | |
| "best_model_checkpoint": "./models/lora-finetuning/LLaMmlein_1B/checkpoint-26000", | |
| "epoch": 0.25068939583855604, | |
| "eval_steps": 1000, | |
| "global_step": 30000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.016712626389237067, | |
| "grad_norm": 25.078365325927734, | |
| "learning_rate": 4.93e-05, | |
| "loss": 4.0094, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.033425252778474135, | |
| "grad_norm": 13.957054138183594, | |
| "learning_rate": 4.916204915525037e-05, | |
| "loss": 3.8688, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.033425252778474135, | |
| "eval_loss": 3.781656265258789, | |
| "eval_runtime": 249.2772, | |
| "eval_samples_per_second": 17.9, | |
| "eval_steps_per_second": 2.238, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.05013787916771121, | |
| "grad_norm": 13.748096466064453, | |
| "learning_rate": 4.8313900125777615e-05, | |
| "loss": 3.7362, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.06685050555694827, | |
| "grad_norm": 11.765251159667969, | |
| "learning_rate": 4.746405139885101e-05, | |
| "loss": 3.7171, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.06685050555694827, | |
| "eval_loss": 3.6495721340179443, | |
| "eval_runtime": 249.1863, | |
| "eval_samples_per_second": 17.906, | |
| "eval_steps_per_second": 2.239, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.08356313194618534, | |
| "grad_norm": 14.349141120910645, | |
| "learning_rate": 4.66142026719244e-05, | |
| "loss": 3.6213, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.10027575833542242, | |
| "grad_norm": 12.409707069396973, | |
| "learning_rate": 4.576435394499779e-05, | |
| "loss": 3.5643, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.10027575833542242, | |
| "eval_loss": 3.565227508544922, | |
| "eval_runtime": 249.1942, | |
| "eval_samples_per_second": 17.906, | |
| "eval_steps_per_second": 2.239, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.11698838472465949, | |
| "grad_norm": 8.590898513793945, | |
| "learning_rate": 4.4914505218071186e-05, | |
| "loss": 3.5126, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 0.13370101111389654, | |
| "grad_norm": 14.560601234436035, | |
| "learning_rate": 4.406465649114458e-05, | |
| "loss": 3.4777, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.13370101111389654, | |
| "eval_loss": 3.48999285697937, | |
| "eval_runtime": 249.1782, | |
| "eval_samples_per_second": 17.907, | |
| "eval_steps_per_second": 2.239, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.1504136375031336, | |
| "grad_norm": 10.846818923950195, | |
| "learning_rate": 4.321480776421797e-05, | |
| "loss": 3.4575, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 0.16712626389237067, | |
| "grad_norm": 10.20279598236084, | |
| "learning_rate": 4.2364959037291364e-05, | |
| "loss": 3.3958, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.16712626389237067, | |
| "eval_loss": 3.433652639389038, | |
| "eval_runtime": 249.1687, | |
| "eval_samples_per_second": 17.908, | |
| "eval_steps_per_second": 2.239, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.18383889028160774, | |
| "grad_norm": 11.423223495483398, | |
| "learning_rate": 4.1515110310364756e-05, | |
| "loss": 3.3574, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 0.20055151667084484, | |
| "grad_norm": 10.610285758972168, | |
| "learning_rate": 4.066526158343815e-05, | |
| "loss": 3.3517, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.20055151667084484, | |
| "eval_loss": 3.387014150619507, | |
| "eval_runtime": 249.1821, | |
| "eval_samples_per_second": 17.907, | |
| "eval_steps_per_second": 2.239, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.2172641430600819, | |
| "grad_norm": 8.165976524353027, | |
| "learning_rate": 3.981541285651154e-05, | |
| "loss": 3.3294, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 0.23397676944931897, | |
| "grad_norm": 10.196443557739258, | |
| "learning_rate": 3.8965564129584935e-05, | |
| "loss": 3.2805, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 0.23397676944931897, | |
| "eval_loss": 3.3461484909057617, | |
| "eval_runtime": 249.1887, | |
| "eval_samples_per_second": 17.906, | |
| "eval_steps_per_second": 2.239, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 0.25068939583855604, | |
| "grad_norm": 9.405224800109863, | |
| "learning_rate": 3.8115715402658334e-05, | |
| "loss": 3.2766, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 0.2674020222277931, | |
| "grad_norm": 8.490914344787598, | |
| "learning_rate": 3.726586667573172e-05, | |
| "loss": 3.2408, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.2674020222277931, | |
| "eval_loss": 3.315063714981079, | |
| "eval_runtime": 249.1918, | |
| "eval_samples_per_second": 17.906, | |
| "eval_steps_per_second": 2.239, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.2841146486170302, | |
| "grad_norm": 11.21275806427002, | |
| "learning_rate": 3.641601794880511e-05, | |
| "loss": 3.2381, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 0.3008272750062672, | |
| "grad_norm": 10.82959270477295, | |
| "learning_rate": 3.556616922187851e-05, | |
| "loss": 3.1716, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 0.3008272750062672, | |
| "eval_loss": 3.2550790309906006, | |
| "eval_runtime": 249.1754, | |
| "eval_samples_per_second": 17.907, | |
| "eval_steps_per_second": 2.239, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 0.3175399013955043, | |
| "grad_norm": 11.301346778869629, | |
| "learning_rate": 3.47163204949519e-05, | |
| "loss": 3.176, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 0.33425252778474135, | |
| "grad_norm": 10.199508666992188, | |
| "learning_rate": 3.386647176802529e-05, | |
| "loss": 3.1645, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 0.33425252778474135, | |
| "eval_loss": 3.2413389682769775, | |
| "eval_runtime": 249.1913, | |
| "eval_samples_per_second": 17.906, | |
| "eval_steps_per_second": 2.239, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 0.35096515417397844, | |
| "grad_norm": 7.239902019500732, | |
| "learning_rate": 3.301662304109869e-05, | |
| "loss": 3.1457, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 0.3676777805632155, | |
| "grad_norm": 10.78636360168457, | |
| "learning_rate": 3.2166774314172076e-05, | |
| "loss": 3.1221, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 0.3676777805632155, | |
| "eval_loss": 3.1978135108947754, | |
| "eval_runtime": 249.1932, | |
| "eval_samples_per_second": 17.906, | |
| "eval_steps_per_second": 2.239, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 0.3843904069524526, | |
| "grad_norm": 8.892194747924805, | |
| "learning_rate": 3.131692558724547e-05, | |
| "loss": 3.1435, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 0.4011030333416897, | |
| "grad_norm": 9.634190559387207, | |
| "learning_rate": 3.0467076860318865e-05, | |
| "loss": 3.0882, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 0.4011030333416897, | |
| "eval_loss": 3.1803853511810303, | |
| "eval_runtime": 249.2264, | |
| "eval_samples_per_second": 17.903, | |
| "eval_steps_per_second": 2.239, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 0.4178156597309267, | |
| "grad_norm": 8.388688087463379, | |
| "learning_rate": 2.9617228133392254e-05, | |
| "loss": 3.0468, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 0.4345282861201638, | |
| "grad_norm": 7.932670593261719, | |
| "learning_rate": 2.876737940646565e-05, | |
| "loss": 3.0627, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 0.4345282861201638, | |
| "eval_loss": 3.1499178409576416, | |
| "eval_runtime": 249.1889, | |
| "eval_samples_per_second": 17.906, | |
| "eval_steps_per_second": 2.239, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 0.45124091250940085, | |
| "grad_norm": 10.718411445617676, | |
| "learning_rate": 2.7917530679539043e-05, | |
| "loss": 3.0438, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 0.46795353889863794, | |
| "grad_norm": 10.233518600463867, | |
| "learning_rate": 2.7072781044973994e-05, | |
| "loss": 3.0231, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 0.46795353889863794, | |
| "eval_loss": 3.1285717487335205, | |
| "eval_runtime": 249.2148, | |
| "eval_samples_per_second": 17.904, | |
| "eval_steps_per_second": 2.239, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 0.484666165287875, | |
| "grad_norm": 10.557904243469238, | |
| "learning_rate": 2.6222932318047387e-05, | |
| "loss": 3.0278, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 0.5013787916771121, | |
| "grad_norm": 6.9997334480285645, | |
| "learning_rate": 2.5373083591120783e-05, | |
| "loss": 3.019, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 0.5013787916771121, | |
| "eval_loss": 3.1056880950927734, | |
| "eval_runtime": 249.1758, | |
| "eval_samples_per_second": 17.907, | |
| "eval_steps_per_second": 2.239, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 0.5180914180663492, | |
| "grad_norm": 8.922798156738281, | |
| "learning_rate": 2.4523234864194176e-05, | |
| "loss": 2.9907, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 0.5348040444555862, | |
| "grad_norm": 9.043310165405273, | |
| "learning_rate": 2.367338613726757e-05, | |
| "loss": 2.9945, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 0.5348040444555862, | |
| "eval_loss": 3.0864944458007812, | |
| "eval_runtime": 249.2, | |
| "eval_samples_per_second": 17.905, | |
| "eval_steps_per_second": 2.239, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 0.5515166708448233, | |
| "grad_norm": 24.114103317260742, | |
| "learning_rate": 2.2830336200156375e-05, | |
| "loss": 3.8624, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 0.5682292972340603, | |
| "grad_norm": 15.645467758178711, | |
| "learning_rate": 2.1983886868137474e-05, | |
| "loss": 3.5355, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 0.5682292972340603, | |
| "eval_loss": 3.449280261993408, | |
| "eval_runtime": 249.5095, | |
| "eval_samples_per_second": 17.883, | |
| "eval_steps_per_second": 2.236, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 0.5849419236232974, | |
| "grad_norm": 12.026097297668457, | |
| "learning_rate": 2.113743753611857e-05, | |
| "loss": 3.4294, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 0.6016545500125344, | |
| "grad_norm": 15.589037895202637, | |
| "learning_rate": 2.0287588809191966e-05, | |
| "loss": 3.3682, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 0.6016545500125344, | |
| "eval_loss": 3.334568500518799, | |
| "eval_runtime": 249.295, | |
| "eval_samples_per_second": 17.898, | |
| "eval_steps_per_second": 2.238, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 0.6183671764017715, | |
| "grad_norm": 18.867961883544922, | |
| "learning_rate": 1.9437740082265356e-05, | |
| "loss": 3.3126, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 0.6350798027910086, | |
| "grad_norm": 14.147026062011719, | |
| "learning_rate": 1.8587891355338752e-05, | |
| "loss": 3.2718, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 0.6350798027910086, | |
| "eval_loss": 3.245798349380493, | |
| "eval_runtime": 249.2433, | |
| "eval_samples_per_second": 17.902, | |
| "eval_steps_per_second": 2.239, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 0.6517924291802457, | |
| "grad_norm": 15.220115661621094, | |
| "learning_rate": 1.7738042628412145e-05, | |
| "loss": 3.1792, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 0.6685050555694827, | |
| "grad_norm": 17.45062255859375, | |
| "learning_rate": 1.6888193901485537e-05, | |
| "loss": 3.1603, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 0.6685050555694827, | |
| "eval_loss": 3.180778741836548, | |
| "eval_runtime": 249.2332, | |
| "eval_samples_per_second": 17.903, | |
| "eval_steps_per_second": 2.239, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 0.6852176819587198, | |
| "grad_norm": 16.632200241088867, | |
| "learning_rate": 1.603834517455893e-05, | |
| "loss": 3.1225, | |
| "step": 20500 | |
| }, | |
| { | |
| "epoch": 0.7019303083479569, | |
| "grad_norm": 13.793306350708008, | |
| "learning_rate": 1.5188496447632323e-05, | |
| "loss": 3.0798, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 0.7019303083479569, | |
| "eval_loss": 3.1489596366882324, | |
| "eval_runtime": 249.2624, | |
| "eval_samples_per_second": 17.901, | |
| "eval_steps_per_second": 2.239, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 0.718642934737194, | |
| "grad_norm": 18.920320510864258, | |
| "learning_rate": 1.4338647720705714e-05, | |
| "loss": 3.0602, | |
| "step": 21500 | |
| }, | |
| { | |
| "epoch": 0.735355561126431, | |
| "grad_norm": 16.346004486083984, | |
| "learning_rate": 1.3488798993779108e-05, | |
| "loss": 3.0281, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 0.735355561126431, | |
| "eval_loss": 3.083583354949951, | |
| "eval_runtime": 249.1953, | |
| "eval_samples_per_second": 17.906, | |
| "eval_steps_per_second": 2.239, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 0.7520681875156681, | |
| "grad_norm": 12.671531677246094, | |
| "learning_rate": 1.26389502668525e-05, | |
| "loss": 3.0254, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 0.7687808139049052, | |
| "grad_norm": 14.962254524230957, | |
| "learning_rate": 1.1789101539925894e-05, | |
| "loss": 3.023, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 0.7687808139049052, | |
| "eval_loss": 3.0390138626098633, | |
| "eval_runtime": 249.2133, | |
| "eval_samples_per_second": 17.904, | |
| "eval_steps_per_second": 2.239, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 0.7854934402941423, | |
| "grad_norm": 16.1228084564209, | |
| "learning_rate": 1.0939252812999288e-05, | |
| "loss": 2.9347, | |
| "step": 23500 | |
| }, | |
| { | |
| "epoch": 0.8022060666833793, | |
| "grad_norm": 13.979843139648438, | |
| "learning_rate": 1.0089404086072679e-05, | |
| "loss": 2.9604, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 0.8022060666833793, | |
| "eval_loss": 3.0071020126342773, | |
| "eval_runtime": 249.2612, | |
| "eval_samples_per_second": 17.901, | |
| "eval_steps_per_second": 2.239, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 0.8189186930726163, | |
| "grad_norm": 17.260498046875, | |
| "learning_rate": 9.239555359146072e-06, | |
| "loss": 2.899, | |
| "step": 24500 | |
| }, | |
| { | |
| "epoch": 0.8356313194618534, | |
| "grad_norm": 13.376078605651855, | |
| "learning_rate": 8.389706632219466e-06, | |
| "loss": 2.9291, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 0.8356313194618534, | |
| "eval_loss": 2.9687299728393555, | |
| "eval_runtime": 249.2378, | |
| "eval_samples_per_second": 17.903, | |
| "eval_steps_per_second": 2.239, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 0.8523439458510905, | |
| "grad_norm": 20.349105834960938, | |
| "learning_rate": 7.539857905292859e-06, | |
| "loss": 2.8853, | |
| "step": 25500 | |
| }, | |
| { | |
| "epoch": 0.8690565722403276, | |
| "grad_norm": 17.05868911743164, | |
| "learning_rate": 6.690009178366251e-06, | |
| "loss": 2.8542, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 0.8690565722403276, | |
| "eval_loss": 2.9402894973754883, | |
| "eval_runtime": 249.3105, | |
| "eval_samples_per_second": 17.897, | |
| "eval_steps_per_second": 2.238, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 0.8857691986295646, | |
| "grad_norm": 28.65497589111328, | |
| "learning_rate": 5.846959241255057e-06, | |
| "loss": 3.7512, | |
| "step": 26500 | |
| }, | |
| { | |
| "epoch": 0.9024818250188017, | |
| "grad_norm": 41.53907775878906, | |
| "learning_rate": 4.99711051432845e-06, | |
| "loss": 3.4169, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 0.9024818250188017, | |
| "eval_loss": 3.276771068572998, | |
| "eval_runtime": 249.323, | |
| "eval_samples_per_second": 17.896, | |
| "eval_steps_per_second": 2.238, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 0.9191944514080388, | |
| "grad_norm": 26.49399757385254, | |
| "learning_rate": 4.147261787401842e-06, | |
| "loss": 3.3055, | |
| "step": 27500 | |
| }, | |
| { | |
| "epoch": 0.9359070777972759, | |
| "grad_norm": 30.56415557861328, | |
| "learning_rate": 3.297413060475236e-06, | |
| "loss": 3.2131, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 0.9359070777972759, | |
| "eval_loss": 3.140591859817505, | |
| "eval_runtime": 249.261, | |
| "eval_samples_per_second": 17.901, | |
| "eval_steps_per_second": 2.239, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 0.9526197041865129, | |
| "grad_norm": 33.53205871582031, | |
| "learning_rate": 2.4509637284563352e-06, | |
| "loss": 3.1509, | |
| "step": 28500 | |
| }, | |
| { | |
| "epoch": 0.96933233057575, | |
| "grad_norm": 26.73455810546875, | |
| "learning_rate": 1.6011150015297277e-06, | |
| "loss": 3.0998, | |
| "step": 29000 | |
| }, | |
| { | |
| "epoch": 0.96933233057575, | |
| "eval_loss": 3.0688536167144775, | |
| "eval_runtime": 249.2871, | |
| "eval_samples_per_second": 17.899, | |
| "eval_steps_per_second": 2.238, | |
| "step": 29000 | |
| }, | |
| { | |
| "epoch": 0.9860449569649871, | |
| "grad_norm": 26.864349365234375, | |
| "learning_rate": 7.512662746031207e-07, | |
| "loss": 3.0876, | |
| "step": 29500 | |
| }, | |
| { | |
| "epoch": 0.25068939583855604, | |
| "grad_norm": 65.23794555664062, | |
| "learning_rate": 4.794667744453116e-05, | |
| "loss": 4.6117, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 0.25068939583855604, | |
| "eval_loss": 4.784034252166748, | |
| "eval_runtime": 253.6816, | |
| "eval_samples_per_second": 17.589, | |
| "eval_steps_per_second": 2.2, | |
| "step": 30000 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 718020, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 6, | |
| "save_steps": 2000, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 3.063185677140099e+18, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |