| { | |
| "best_metric": 0.2758501172065735, | |
| "best_model_checkpoint": "gbert-large-coherence/checkpoint-3002", | |
| "epoch": 7.0, | |
| "eval_steps": 500, | |
| "global_step": 10507, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.3331112591605596, | |
| "grad_norm": 11.811367988586426, | |
| "learning_rate": 1.933377748167888e-05, | |
| "loss": 0.3986, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.6662225183211192, | |
| "grad_norm": 25.986848831176758, | |
| "learning_rate": 1.8667554963357762e-05, | |
| "loss": 0.3176, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.9993337774816788, | |
| "grad_norm": 0.04597562551498413, | |
| "learning_rate": 1.8001332445036644e-05, | |
| "loss": 0.2642, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.9253582139286904, | |
| "eval_loss": 0.341106116771698, | |
| "eval_runtime": 244.1311, | |
| "eval_samples_per_second": 12.293, | |
| "eval_steps_per_second": 1.54, | |
| "step": 1501 | |
| }, | |
| { | |
| "epoch": 1.3324450366422385, | |
| "grad_norm": 0.26154929399490356, | |
| "learning_rate": 1.7335109926715526e-05, | |
| "loss": 0.2189, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.6655562958027983, | |
| "grad_norm": 0.5341408252716064, | |
| "learning_rate": 1.6668887408394405e-05, | |
| "loss": 0.1798, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.9986675549633577, | |
| "grad_norm": 0.0845261663198471, | |
| "learning_rate": 1.6002664890073287e-05, | |
| "loss": 0.1922, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.9450183272242586, | |
| "eval_loss": 0.2758501172065735, | |
| "eval_runtime": 248.4145, | |
| "eval_samples_per_second": 12.081, | |
| "eval_steps_per_second": 1.514, | |
| "step": 3002 | |
| }, | |
| { | |
| "epoch": 2.3317788141239175, | |
| "grad_norm": 0.04522880166769028, | |
| "learning_rate": 1.533644237175217e-05, | |
| "loss": 0.1424, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 2.664890073284477, | |
| "grad_norm": 0.06311460584402084, | |
| "learning_rate": 1.4670219853431047e-05, | |
| "loss": 0.129, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 2.9980013324450367, | |
| "grad_norm": 0.048002395778894424, | |
| "learning_rate": 1.4003997335109927e-05, | |
| "loss": 0.1417, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.9476841052982339, | |
| "eval_loss": 0.30848464369773865, | |
| "eval_runtime": 245.4422, | |
| "eval_samples_per_second": 12.227, | |
| "eval_steps_per_second": 1.532, | |
| "step": 4503 | |
| }, | |
| { | |
| "epoch": 3.331112591605596, | |
| "grad_norm": 0.42216023802757263, | |
| "learning_rate": 1.333777481678881e-05, | |
| "loss": 0.091, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 3.664223850766156, | |
| "grad_norm": 0.04599091783165932, | |
| "learning_rate": 1.2671552298467688e-05, | |
| "loss": 0.0996, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 3.9973351099267154, | |
| "grad_norm": 0.02595396339893341, | |
| "learning_rate": 1.2005329780146572e-05, | |
| "loss": 0.0992, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.9493502165944685, | |
| "eval_loss": 0.29965537786483765, | |
| "eval_runtime": 245.2153, | |
| "eval_samples_per_second": 12.238, | |
| "eval_steps_per_second": 1.533, | |
| "step": 6004 | |
| }, | |
| { | |
| "epoch": 4.330446369087275, | |
| "grad_norm": 0.02373787760734558, | |
| "learning_rate": 1.133910726182545e-05, | |
| "loss": 0.0746, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 4.663557628247835, | |
| "grad_norm": 0.02029305137693882, | |
| "learning_rate": 1.067288474350433e-05, | |
| "loss": 0.0841, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 4.996668887408394, | |
| "grad_norm": 0.014750126749277115, | |
| "learning_rate": 1.0006662225183212e-05, | |
| "loss": 0.0778, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.9450183272242586, | |
| "eval_loss": 0.3702091574668884, | |
| "eval_runtime": 245.2652, | |
| "eval_samples_per_second": 12.236, | |
| "eval_steps_per_second": 1.533, | |
| "step": 7505 | |
| }, | |
| { | |
| "epoch": 5.329780146568954, | |
| "grad_norm": 0.010594181716442108, | |
| "learning_rate": 9.340439706862093e-06, | |
| "loss": 0.0446, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 5.662891405729514, | |
| "grad_norm": 0.00855865329504013, | |
| "learning_rate": 8.674217188540973e-06, | |
| "loss": 0.0314, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 5.9960026648900735, | |
| "grad_norm": 0.1962573230266571, | |
| "learning_rate": 8.007994670219855e-06, | |
| "loss": 0.067, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.9496834388537154, | |
| "eval_loss": 0.3373388946056366, | |
| "eval_runtime": 249.442, | |
| "eval_samples_per_second": 12.031, | |
| "eval_steps_per_second": 1.507, | |
| "step": 9006 | |
| }, | |
| { | |
| "epoch": 6.329113924050633, | |
| "grad_norm": 0.25471004843711853, | |
| "learning_rate": 7.341772151898735e-06, | |
| "loss": 0.0328, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 6.662225183211192, | |
| "grad_norm": 0.008936082012951374, | |
| "learning_rate": 6.675549633577616e-06, | |
| "loss": 0.0372, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 6.995336442371752, | |
| "grad_norm": 0.06890871375799179, | |
| "learning_rate": 6.0093271152564956e-06, | |
| "loss": 0.0383, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.9530156614461847, | |
| "eval_loss": 0.3258904218673706, | |
| "eval_runtime": 251.6126, | |
| "eval_samples_per_second": 11.927, | |
| "eval_steps_per_second": 1.494, | |
| "step": 10507 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 15010, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "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": 7.830180507818189e+16, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |