| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.0, | |
| "eval_steps": 1000, | |
| "global_step": 19113, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.026160205096007954, | |
| "grad_norm": 2.7528529167175293, | |
| "learning_rate": 4.869198974519961e-05, | |
| "loss": 4.403, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.05232041019201591, | |
| "grad_norm": 2.5874102115631104, | |
| "learning_rate": 4.7383979490399206e-05, | |
| "loss": 2.92, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.05232041019201591, | |
| "eval_accuracy": 0.4470485610347474, | |
| "eval_loss": 2.4565796852111816, | |
| "eval_runtime": 53.659, | |
| "eval_samples_per_second": 114.631, | |
| "eval_steps_per_second": 3.597, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.07848061528802386, | |
| "grad_norm": 2.103119134902954, | |
| "learning_rate": 4.607596923559881e-05, | |
| "loss": 2.2955, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.10464082038403182, | |
| "grad_norm": 2.2540347576141357, | |
| "learning_rate": 4.476795898079841e-05, | |
| "loss": 1.9818, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.10464082038403182, | |
| "eval_accuracy": 0.5672200738882789, | |
| "eval_loss": 1.8089067935943604, | |
| "eval_runtime": 53.7114, | |
| "eval_samples_per_second": 114.52, | |
| "eval_steps_per_second": 3.593, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.13080102548003977, | |
| "grad_norm": 1.4521691799163818, | |
| "learning_rate": 4.345994872599801e-05, | |
| "loss": 1.83, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.15696123057604772, | |
| "grad_norm": 1.3930741548538208, | |
| "learning_rate": 4.2151938471197614e-05, | |
| "loss": 1.7321, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.15696123057604772, | |
| "eval_accuracy": 0.6020537553359386, | |
| "eval_loss": 1.613312005996704, | |
| "eval_runtime": 53.6022, | |
| "eval_samples_per_second": 114.753, | |
| "eval_steps_per_second": 3.601, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.18312143567205566, | |
| "grad_norm": 1.4211032390594482, | |
| "learning_rate": 4.084392821639722e-05, | |
| "loss": 1.6624, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 0.20928164076806363, | |
| "grad_norm": 1.275313138961792, | |
| "learning_rate": 3.953591796159682e-05, | |
| "loss": 1.6137, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.20928164076806363, | |
| "eval_accuracy": 0.6195058763065014, | |
| "eval_loss": 1.5170681476593018, | |
| "eval_runtime": 53.3002, | |
| "eval_samples_per_second": 115.403, | |
| "eval_steps_per_second": 3.621, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.23544184586407158, | |
| "grad_norm": 1.2738176584243774, | |
| "learning_rate": 3.8227907706796424e-05, | |
| "loss": 1.5721, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 0.26160205096007955, | |
| "grad_norm": 1.2026641368865967, | |
| "learning_rate": 3.691989745199603e-05, | |
| "loss": 1.5353, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.26160205096007955, | |
| "eval_accuracy": 0.6311744206133264, | |
| "eval_loss": 1.4516141414642334, | |
| "eval_runtime": 53.8002, | |
| "eval_samples_per_second": 114.33, | |
| "eval_steps_per_second": 3.587, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.28776225605608746, | |
| "grad_norm": 1.1667968034744263, | |
| "learning_rate": 3.561188719719563e-05, | |
| "loss": 1.5091, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 0.31392246115209543, | |
| "grad_norm": 1.4095014333724976, | |
| "learning_rate": 3.430387694239523e-05, | |
| "loss": 1.4845, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.31392246115209543, | |
| "eval_accuracy": 0.6400102153795495, | |
| "eval_loss": 1.4056029319763184, | |
| "eval_runtime": 53.6423, | |
| "eval_samples_per_second": 114.667, | |
| "eval_steps_per_second": 3.598, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.3400826662481034, | |
| "grad_norm": 1.2620683908462524, | |
| "learning_rate": 3.299586668759483e-05, | |
| "loss": 1.4606, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 0.3662428713441113, | |
| "grad_norm": 1.2207545042037964, | |
| "learning_rate": 3.168785643279443e-05, | |
| "loss": 1.4443, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 0.3662428713441113, | |
| "eval_accuracy": 0.646570593151532, | |
| "eval_loss": 1.371826171875, | |
| "eval_runtime": 53.6024, | |
| "eval_samples_per_second": 114.752, | |
| "eval_steps_per_second": 3.601, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 0.3924030764401193, | |
| "grad_norm": 1.1445685625076294, | |
| "learning_rate": 3.0379846177994036e-05, | |
| "loss": 1.4255, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 0.41856328153612726, | |
| "grad_norm": 1.1666637659072876, | |
| "learning_rate": 2.907183592319364e-05, | |
| "loss": 1.4118, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.41856328153612726, | |
| "eval_accuracy": 0.6524860734404423, | |
| "eval_loss": 1.3419677019119263, | |
| "eval_runtime": 53.4631, | |
| "eval_samples_per_second": 115.051, | |
| "eval_steps_per_second": 3.61, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.4447234866321352, | |
| "grad_norm": 1.2419095039367676, | |
| "learning_rate": 2.776382566839324e-05, | |
| "loss": 1.4015, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 0.47088369172814315, | |
| "grad_norm": 1.1277480125427246, | |
| "learning_rate": 2.6455815413592845e-05, | |
| "loss": 1.3878, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 0.47088369172814315, | |
| "eval_accuracy": 0.6569193066064805, | |
| "eval_loss": 1.3188564777374268, | |
| "eval_runtime": 53.2918, | |
| "eval_samples_per_second": 115.421, | |
| "eval_steps_per_second": 3.622, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 0.4970438968241511, | |
| "grad_norm": 1.0969843864440918, | |
| "learning_rate": 2.5147805158792447e-05, | |
| "loss": 1.3773, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 0.5232041019201591, | |
| "grad_norm": 1.1453758478164673, | |
| "learning_rate": 2.383979490399205e-05, | |
| "loss": 1.3661, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 0.5232041019201591, | |
| "eval_accuracy": 0.6608493989564993, | |
| "eval_loss": 1.2987806797027588, | |
| "eval_runtime": 53.4371, | |
| "eval_samples_per_second": 115.107, | |
| "eval_steps_per_second": 3.612, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 0.549364307016167, | |
| "grad_norm": 1.1494508981704712, | |
| "learning_rate": 2.253178464919165e-05, | |
| "loss": 1.3563, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 0.5755245121121749, | |
| "grad_norm": 1.1540991067886353, | |
| "learning_rate": 2.1223774394391253e-05, | |
| "loss": 1.3485, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 0.5755245121121749, | |
| "eval_accuracy": 0.6638841358556485, | |
| "eval_loss": 1.2833938598632812, | |
| "eval_runtime": 53.55, | |
| "eval_samples_per_second": 114.865, | |
| "eval_steps_per_second": 3.604, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 0.601684717208183, | |
| "grad_norm": 1.1493678092956543, | |
| "learning_rate": 1.9915764139590855e-05, | |
| "loss": 1.3393, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 0.6278449223041909, | |
| "grad_norm": 1.1292744874954224, | |
| "learning_rate": 1.8607753884790457e-05, | |
| "loss": 1.3326, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 0.6278449223041909, | |
| "eval_accuracy": 0.6669099732330993, | |
| "eval_loss": 1.2675042152404785, | |
| "eval_runtime": 54.1653, | |
| "eval_samples_per_second": 113.56, | |
| "eval_steps_per_second": 3.563, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 0.6540051274001988, | |
| "grad_norm": 1.1094636917114258, | |
| "learning_rate": 1.729974362999006e-05, | |
| "loss": 1.3249, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 0.6801653324962068, | |
| "grad_norm": 1.1196491718292236, | |
| "learning_rate": 1.5991733375189664e-05, | |
| "loss": 1.319, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 0.6801653324962068, | |
| "eval_accuracy": 0.6694385498356529, | |
| "eval_loss": 1.2554900646209717, | |
| "eval_runtime": 53.6683, | |
| "eval_samples_per_second": 114.611, | |
| "eval_steps_per_second": 3.596, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 0.7063255375922147, | |
| "grad_norm": 1.1508065462112427, | |
| "learning_rate": 1.4683723120389265e-05, | |
| "loss": 1.3118, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 0.7324857426882226, | |
| "grad_norm": 1.0860583782196045, | |
| "learning_rate": 1.3375712865588865e-05, | |
| "loss": 1.3068, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 0.7324857426882226, | |
| "eval_accuracy": 0.6718871896629512, | |
| "eval_loss": 1.2440038919448853, | |
| "eval_runtime": 53.2979, | |
| "eval_samples_per_second": 115.408, | |
| "eval_steps_per_second": 3.621, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 0.7586459477842307, | |
| "grad_norm": 1.126051902770996, | |
| "learning_rate": 1.2067702610788469e-05, | |
| "loss": 1.3014, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 0.7848061528802386, | |
| "grad_norm": 1.100677490234375, | |
| "learning_rate": 1.075969235598807e-05, | |
| "loss": 1.2932, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 0.7848061528802386, | |
| "eval_accuracy": 0.6737417864168825, | |
| "eval_loss": 1.2349998950958252, | |
| "eval_runtime": 53.6159, | |
| "eval_samples_per_second": 114.723, | |
| "eval_steps_per_second": 3.6, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 0.8109663579762465, | |
| "grad_norm": 1.1026638746261597, | |
| "learning_rate": 9.451682101187674e-06, | |
| "loss": 1.2913, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 0.8371265630722545, | |
| "grad_norm": 1.152256727218628, | |
| "learning_rate": 8.143671846387275e-06, | |
| "loss": 1.2868, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 0.8371265630722545, | |
| "eval_accuracy": 0.6755100895943456, | |
| "eval_loss": 1.2262605428695679, | |
| "eval_runtime": 53.3362, | |
| "eval_samples_per_second": 115.325, | |
| "eval_steps_per_second": 3.619, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 0.8632867681682624, | |
| "grad_norm": 1.1048023700714111, | |
| "learning_rate": 6.835661591586878e-06, | |
| "loss": 1.2809, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 0.8894469732642704, | |
| "grad_norm": 1.1195833683013916, | |
| "learning_rate": 5.527651336786481e-06, | |
| "loss": 1.2791, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 0.8894469732642704, | |
| "eval_accuracy": 0.6771291668633302, | |
| "eval_loss": 1.219257116317749, | |
| "eval_runtime": 53.5262, | |
| "eval_samples_per_second": 114.916, | |
| "eval_steps_per_second": 3.606, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 0.9156071783602784, | |
| "grad_norm": 1.1175094842910767, | |
| "learning_rate": 4.219641081986083e-06, | |
| "loss": 1.2743, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 0.9417673834562863, | |
| "grad_norm": 1.1169252395629883, | |
| "learning_rate": 2.911630827185685e-06, | |
| "loss": 1.2725, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 0.9417673834562863, | |
| "eval_accuracy": 0.6780440694779302, | |
| "eval_loss": 1.214146614074707, | |
| "eval_runtime": 53.7753, | |
| "eval_samples_per_second": 114.383, | |
| "eval_steps_per_second": 3.589, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 0.9679275885522942, | |
| "grad_norm": 1.1183288097381592, | |
| "learning_rate": 1.6036205723852876e-06, | |
| "loss": 1.2693, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 0.9940877936483022, | |
| "grad_norm": 1.10493803024292, | |
| "learning_rate": 2.956103175848899e-07, | |
| "loss": 1.2711, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 0.9940877936483022, | |
| "eval_accuracy": 0.6788121299845069, | |
| "eval_loss": 1.2107646465301514, | |
| "eval_runtime": 53.4355, | |
| "eval_samples_per_second": 115.111, | |
| "eval_steps_per_second": 3.612, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 19113, | |
| "total_flos": 3.19615549046784e+17, | |
| "train_loss": 1.5581016430754755, | |
| "train_runtime": 7493.9423, | |
| "train_samples_per_second": 81.613, | |
| "train_steps_per_second": 2.55 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 19113, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 1000, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 3.19615549046784e+17, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |