| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 3.2771163298743855, | |
| "eval_steps": 500, | |
| "global_step": 15000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.10922992900054615, | |
| "grad_norm": 0.9561710357666016, | |
| "learning_rate": 1.4564520827264783e-05, | |
| "loss": 0.7634, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.10922992900054615, | |
| "eval_loss": 0.3866455852985382, | |
| "eval_runtime": 53.8542, | |
| "eval_samples_per_second": 1.727, | |
| "eval_steps_per_second": 1.727, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.2184598580010923, | |
| "grad_norm": 0.6381384134292603, | |
| "learning_rate": 2.9129041654529566e-05, | |
| "loss": 0.2265, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.2184598580010923, | |
| "eval_loss": 0.28045815229415894, | |
| "eval_runtime": 53.8781, | |
| "eval_samples_per_second": 1.726, | |
| "eval_steps_per_second": 1.726, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.3276897870016384, | |
| "grad_norm": 0.80877685546875, | |
| "learning_rate": 4.3693562481794354e-05, | |
| "loss": 0.13, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.3276897870016384, | |
| "eval_loss": 0.24824748933315277, | |
| "eval_runtime": 53.7906, | |
| "eval_samples_per_second": 1.729, | |
| "eval_steps_per_second": 1.729, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.4369197160021846, | |
| "grad_norm": 0.7288264632225037, | |
| "learning_rate": 5.8228954267404603e-05, | |
| "loss": 0.1067, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.4369197160021846, | |
| "eval_loss": 0.24914385378360748, | |
| "eval_runtime": 53.8469, | |
| "eval_samples_per_second": 1.727, | |
| "eval_steps_per_second": 1.727, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.5461496450027308, | |
| "grad_norm": 1.1100046634674072, | |
| "learning_rate": 7.27934750946694e-05, | |
| "loss": 0.1018, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.5461496450027308, | |
| "eval_loss": 0.27210181951522827, | |
| "eval_runtime": 53.5513, | |
| "eval_samples_per_second": 1.737, | |
| "eval_steps_per_second": 1.737, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.6553795740032768, | |
| "grad_norm": 0.8480390310287476, | |
| "learning_rate": 8.732886688027965e-05, | |
| "loss": 0.0976, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.6553795740032768, | |
| "eval_loss": 0.2787032425403595, | |
| "eval_runtime": 51.6926, | |
| "eval_samples_per_second": 1.799, | |
| "eval_steps_per_second": 1.799, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.764609503003823, | |
| "grad_norm": 0.8974778056144714, | |
| "learning_rate": 0.00010189338770754442, | |
| "loss": 0.0948, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 0.764609503003823, | |
| "eval_loss": 0.26553070545196533, | |
| "eval_runtime": 52.0101, | |
| "eval_samples_per_second": 1.788, | |
| "eval_steps_per_second": 1.788, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 0.8738394320043692, | |
| "grad_norm": 0.4318696856498718, | |
| "learning_rate": 0.00011645790853480921, | |
| "loss": 0.0941, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.8738394320043692, | |
| "eval_loss": 0.2772391140460968, | |
| "eval_runtime": 51.821, | |
| "eval_samples_per_second": 1.795, | |
| "eval_steps_per_second": 1.795, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.9830693610049154, | |
| "grad_norm": 0.22688086330890656, | |
| "learning_rate": 0.00013102242936207398, | |
| "loss": 0.0916, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 0.9830693610049154, | |
| "eval_loss": 0.2516581118106842, | |
| "eval_runtime": 51.8197, | |
| "eval_samples_per_second": 1.795, | |
| "eval_steps_per_second": 1.795, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 1.0922992900054616, | |
| "grad_norm": 0.053704842925071716, | |
| "learning_rate": 0.0001455869501893388, | |
| "loss": 0.084, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.0922992900054616, | |
| "eval_loss": 0.24150623381137848, | |
| "eval_runtime": 50.6724, | |
| "eval_samples_per_second": 1.835, | |
| "eval_steps_per_second": 1.835, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.2015292190060076, | |
| "grad_norm": 0.30040594935417175, | |
| "learning_rate": 0.00016015147101660357, | |
| "loss": 0.0882, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 1.2015292190060076, | |
| "eval_loss": 0.24166980385780334, | |
| "eval_runtime": 50.7195, | |
| "eval_samples_per_second": 1.834, | |
| "eval_steps_per_second": 1.834, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 1.3107591480065537, | |
| "grad_norm": 0.21068286895751953, | |
| "learning_rate": 0.00017471599184386836, | |
| "loss": 0.0868, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 1.3107591480065537, | |
| "eval_loss": 0.2663758397102356, | |
| "eval_runtime": 50.6208, | |
| "eval_samples_per_second": 1.837, | |
| "eval_steps_per_second": 1.837, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 1.4199890770071, | |
| "grad_norm": 0.1310437172651291, | |
| "learning_rate": 0.00018928051267113314, | |
| "loss": 0.087, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 1.4199890770071, | |
| "eval_loss": 0.2654285728931427, | |
| "eval_runtime": 50.5893, | |
| "eval_samples_per_second": 1.838, | |
| "eval_steps_per_second": 1.838, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 1.529219006007646, | |
| "grad_norm": 0.1630202829837799, | |
| "learning_rate": 0.00019988197347556582, | |
| "loss": 0.0841, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 1.529219006007646, | |
| "eval_loss": 0.25736120343208313, | |
| "eval_runtime": 50.5692, | |
| "eval_samples_per_second": 1.839, | |
| "eval_steps_per_second": 1.839, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 1.6385581649371928, | |
| "grad_norm": 0.07697170972824097, | |
| "learning_rate": 0.00019943149055787806, | |
| "loss": 0.0807, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 1.6385581649371928, | |
| "eval_loss": 0.265057772397995, | |
| "eval_runtime": 50.3751, | |
| "eval_samples_per_second": 1.846, | |
| "eval_steps_per_second": 1.846, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 1.7477880939377388, | |
| "grad_norm": 0.10595889389514923, | |
| "learning_rate": 0.0001989810076401903, | |
| "loss": 0.0782, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 1.7477880939377388, | |
| "eval_loss": 0.2732037305831909, | |
| "eval_runtime": 50.5346, | |
| "eval_samples_per_second": 1.84, | |
| "eval_steps_per_second": 1.84, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 1.857018022938285, | |
| "grad_norm": 0.13792632520198822, | |
| "learning_rate": 0.00019853052472250252, | |
| "loss": 0.0757, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 1.857018022938285, | |
| "eval_loss": 0.25934740900993347, | |
| "eval_runtime": 50.3603, | |
| "eval_samples_per_second": 1.847, | |
| "eval_steps_per_second": 1.847, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 1.97, | |
| "grad_norm": 0.1341710239648819, | |
| "learning_rate": 0.00019808004180481476, | |
| "loss": 0.0764, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 1.97, | |
| "eval_loss": 0.25228968262672424, | |
| "eval_runtime": 47.204, | |
| "eval_samples_per_second": 1.97, | |
| "eval_steps_per_second": 1.97, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 2.16, | |
| "grad_norm": 0.011377187445759773, | |
| "learning_rate": 0.00019726869847708111, | |
| "loss": 0.0743, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 2.16, | |
| "eval_loss": 0.1291830837726593, | |
| "eval_runtime": 193.506, | |
| "eval_samples_per_second": 2.393, | |
| "eval_steps_per_second": 2.393, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 2.184817039868924, | |
| "grad_norm": 0.39175206422805786, | |
| "learning_rate": 0.00019717907596943925, | |
| "loss": 0.0717, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 2.184817039868924, | |
| "eval_loss": 0.2626956105232239, | |
| "eval_runtime": 51.1714, | |
| "eval_samples_per_second": 1.817, | |
| "eval_steps_per_second": 1.817, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 2.2940469688694702, | |
| "grad_norm": 0.17580239474773407, | |
| "learning_rate": 0.0001967285930517515, | |
| "loss": 0.0703, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 2.2940469688694702, | |
| "eval_loss": 0.28000086545944214, | |
| "eval_runtime": 51.2031, | |
| "eval_samples_per_second": 1.816, | |
| "eval_steps_per_second": 1.816, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 2.4032768978700165, | |
| "grad_norm": 0.27358973026275635, | |
| "learning_rate": 0.00019627811013406371, | |
| "loss": 0.0707, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 2.4032768978700165, | |
| "eval_loss": 0.27782928943634033, | |
| "eval_runtime": 50.7155, | |
| "eval_samples_per_second": 1.834, | |
| "eval_steps_per_second": 1.834, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 2.5125068268705624, | |
| "grad_norm": 0.3582260012626648, | |
| "learning_rate": 0.00019582762721637599, | |
| "loss": 0.0698, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 2.5125068268705624, | |
| "eval_loss": 0.24289114773273468, | |
| "eval_runtime": 51.1485, | |
| "eval_samples_per_second": 1.818, | |
| "eval_steps_per_second": 1.818, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 2.6217367558711087, | |
| "grad_norm": 0.3518364727497101, | |
| "learning_rate": 0.0001953771442986882, | |
| "loss": 0.0732, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 2.6217367558711087, | |
| "eval_loss": 0.25252565741539, | |
| "eval_runtime": 51.1118, | |
| "eval_samples_per_second": 1.82, | |
| "eval_steps_per_second": 1.82, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 2.730966684871655, | |
| "grad_norm": 0.08524592220783234, | |
| "learning_rate": 0.00019492666138100045, | |
| "loss": 0.0716, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 2.730966684871655, | |
| "eval_loss": 0.2652508020401001, | |
| "eval_runtime": 51.1953, | |
| "eval_samples_per_second": 1.817, | |
| "eval_steps_per_second": 1.817, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 2.840196613872201, | |
| "grad_norm": 0.14637166261672974, | |
| "learning_rate": 0.00019447617846331266, | |
| "loss": 0.0719, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 2.840196613872201, | |
| "eval_loss": 0.26826488971710205, | |
| "eval_runtime": 51.1801, | |
| "eval_samples_per_second": 1.817, | |
| "eval_steps_per_second": 1.817, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 2.949426542872747, | |
| "grad_norm": 0.234298437833786, | |
| "learning_rate": 0.0001940256955456249, | |
| "loss": 0.0696, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 2.949426542872747, | |
| "eval_loss": 0.27446672320365906, | |
| "eval_runtime": 50.6415, | |
| "eval_samples_per_second": 1.836, | |
| "eval_steps_per_second": 1.836, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 3.0586564718732934, | |
| "grad_norm": 0.11118844896554947, | |
| "learning_rate": 0.00019357521262793715, | |
| "loss": 0.0642, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 3.0586564718732934, | |
| "eval_loss": 0.2984607219696045, | |
| "eval_runtime": 50.5244, | |
| "eval_samples_per_second": 1.841, | |
| "eval_steps_per_second": 1.841, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 3.167886400873839, | |
| "grad_norm": 0.2369268238544464, | |
| "learning_rate": 0.0001931247297102494, | |
| "loss": 0.0611, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 3.167886400873839, | |
| "eval_loss": 0.3083478808403015, | |
| "eval_runtime": 50.5839, | |
| "eval_samples_per_second": 1.839, | |
| "eval_steps_per_second": 1.839, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 3.2771163298743855, | |
| "grad_norm": 0.30132102966308594, | |
| "learning_rate": 0.00019267424679256164, | |
| "loss": 0.0648, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 3.2771163298743855, | |
| "eval_loss": 0.29901525378227234, | |
| "eval_runtime": 50.6251, | |
| "eval_samples_per_second": 1.837, | |
| "eval_steps_per_second": 1.837, | |
| "step": 15000 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 228850, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 50, | |
| "save_steps": 500, | |
| "total_flos": 3.1313819597671526e+17, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |