| { |
| "best_metric": 0.02135350927710533, |
| "best_model_checkpoint": "./results/answerdotai/ModernBERT-base/trial-0/checkpoint-1506", |
| "epoch": 1.0, |
| "eval_steps": 500, |
| "global_step": 1506, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 0.033200531208499334, |
| "grad_norm": 11.822611808776855, |
| "learning_rate": 4.4935320035267014e-05, |
| "loss": 0.295, |
| "step": 50 |
| }, |
| { |
| "epoch": 0.06640106241699867, |
| "grad_norm": 0.11557121574878693, |
| "learning_rate": 4.463495024893502e-05, |
| "loss": 0.0808, |
| "step": 100 |
| }, |
| { |
| "epoch": 0.099601593625498, |
| "grad_norm": 0.01743650808930397, |
| "learning_rate": 4.433458046260302e-05, |
| "loss": 0.052, |
| "step": 150 |
| }, |
| { |
| "epoch": 0.13280212483399734, |
| "grad_norm": 4.474731922149658, |
| "learning_rate": 4.4034210676271024e-05, |
| "loss": 0.0491, |
| "step": 200 |
| }, |
| { |
| "epoch": 0.16600265604249667, |
| "grad_norm": 4.205756664276123, |
| "learning_rate": 4.373384088993902e-05, |
| "loss": 0.0344, |
| "step": 250 |
| }, |
| { |
| "epoch": 0.199203187250996, |
| "grad_norm": 4.239188194274902, |
| "learning_rate": 4.343347110360703e-05, |
| "loss": 0.0295, |
| "step": 300 |
| }, |
| { |
| "epoch": 0.23240371845949534, |
| "grad_norm": 0.19662700593471527, |
| "learning_rate": 4.3133101317275027e-05, |
| "loss": 0.0342, |
| "step": 350 |
| }, |
| { |
| "epoch": 0.2656042496679947, |
| "grad_norm": 0.008393031544983387, |
| "learning_rate": 4.2832731530943025e-05, |
| "loss": 0.0245, |
| "step": 400 |
| }, |
| { |
| "epoch": 0.29880478087649404, |
| "grad_norm": 0.06995929777622223, |
| "learning_rate": 4.253236174461103e-05, |
| "loss": 0.0281, |
| "step": 450 |
| }, |
| { |
| "epoch": 0.33200531208499334, |
| "grad_norm": 0.010315222665667534, |
| "learning_rate": 4.223199195827902e-05, |
| "loss": 0.0188, |
| "step": 500 |
| }, |
| { |
| "epoch": 0.3652058432934927, |
| "grad_norm": 3.1021769046783447, |
| "learning_rate": 4.193162217194703e-05, |
| "loss": 0.018, |
| "step": 550 |
| }, |
| { |
| "epoch": 0.398406374501992, |
| "grad_norm": 0.00041495164623484015, |
| "learning_rate": 4.1631252385615027e-05, |
| "loss": 0.0053, |
| "step": 600 |
| }, |
| { |
| "epoch": 0.4316069057104914, |
| "grad_norm": 0.19596342742443085, |
| "learning_rate": 4.133088259928303e-05, |
| "loss": 0.0178, |
| "step": 650 |
| }, |
| { |
| "epoch": 0.4648074369189907, |
| "grad_norm": 0.0566418319940567, |
| "learning_rate": 4.103051281295103e-05, |
| "loss": 0.0101, |
| "step": 700 |
| }, |
| { |
| "epoch": 0.49800796812749004, |
| "grad_norm": 0.005816417746245861, |
| "learning_rate": 4.0730143026619036e-05, |
| "loss": 0.0166, |
| "step": 750 |
| }, |
| { |
| "epoch": 0.5312084993359893, |
| "grad_norm": 2.2474324703216553, |
| "learning_rate": 4.0429773240287035e-05, |
| "loss": 0.0156, |
| "step": 800 |
| }, |
| { |
| "epoch": 0.5644090305444888, |
| "grad_norm": 0.06311876326799393, |
| "learning_rate": 4.0129403453955033e-05, |
| "loss": 0.0166, |
| "step": 850 |
| }, |
| { |
| "epoch": 0.5976095617529881, |
| "grad_norm": 0.012764506973326206, |
| "learning_rate": 3.982903366762304e-05, |
| "loss": 0.0175, |
| "step": 900 |
| }, |
| { |
| "epoch": 0.6308100929614874, |
| "grad_norm": 0.00253055221401155, |
| "learning_rate": 3.952866388129104e-05, |
| "loss": 0.0047, |
| "step": 950 |
| }, |
| { |
| "epoch": 0.6640106241699867, |
| "grad_norm": 0.03604559600353241, |
| "learning_rate": 3.922829409495904e-05, |
| "loss": 0.016, |
| "step": 1000 |
| }, |
| { |
| "epoch": 0.6972111553784861, |
| "grad_norm": 0.006498202681541443, |
| "learning_rate": 3.892792430862704e-05, |
| "loss": 0.0055, |
| "step": 1050 |
| }, |
| { |
| "epoch": 0.7304116865869854, |
| "grad_norm": 0.11296769976615906, |
| "learning_rate": 3.862755452229504e-05, |
| "loss": 0.0122, |
| "step": 1100 |
| }, |
| { |
| "epoch": 0.7636122177954847, |
| "grad_norm": 0.0005851402529515326, |
| "learning_rate": 3.8327184735963046e-05, |
| "loss": 0.01, |
| "step": 1150 |
| }, |
| { |
| "epoch": 0.796812749003984, |
| "grad_norm": 0.018440622836351395, |
| "learning_rate": 3.8026814949631044e-05, |
| "loss": 0.0064, |
| "step": 1200 |
| }, |
| { |
| "epoch": 0.8300132802124834, |
| "grad_norm": 0.0023099363315850496, |
| "learning_rate": 3.772644516329905e-05, |
| "loss": 0.0011, |
| "step": 1250 |
| }, |
| { |
| "epoch": 0.8632138114209827, |
| "grad_norm": 0.07595626264810562, |
| "learning_rate": 3.742607537696705e-05, |
| "loss": 0.0156, |
| "step": 1300 |
| }, |
| { |
| "epoch": 0.896414342629482, |
| "grad_norm": 0.0008996099350042641, |
| "learning_rate": 3.7125705590635054e-05, |
| "loss": 0.0103, |
| "step": 1350 |
| }, |
| { |
| "epoch": 0.9296148738379814, |
| "grad_norm": 3.656134504126385e-05, |
| "learning_rate": 3.682533580430305e-05, |
| "loss": 0.0027, |
| "step": 1400 |
| }, |
| { |
| "epoch": 0.9628154050464808, |
| "grad_norm": 0.2666904032230377, |
| "learning_rate": 3.652496601797105e-05, |
| "loss": 0.0152, |
| "step": 1450 |
| }, |
| { |
| "epoch": 0.9960159362549801, |
| "grad_norm": 0.011590929701924324, |
| "learning_rate": 3.622459623163905e-05, |
| "loss": 0.0115, |
| "step": 1500 |
| }, |
| { |
| "epoch": 1.0, |
| "eval_accuracy": 0.9963024809160306, |
| "eval_f1": 0.9962997469825083, |
| "eval_loss": 0.02135350927710533, |
| "eval_precision": 0.9962971957079396, |
| "eval_recall": 0.9963024809160306, |
| "eval_runtime": 34.0647, |
| "eval_samples_per_second": 246.12, |
| "eval_steps_per_second": 7.691, |
| "step": 1506 |
| } |
| ], |
| "logging_steps": 50, |
| "max_steps": 7530, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 5, |
| "save_steps": 500, |
| "stateful_callbacks": { |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": false |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 1.641430544259072e+16, |
| "train_batch_size": 32, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
