| { | |
| "best_metric": 0.6912539515279241, | |
| "best_model_checkpoint": "../saved_model/tibetan-bert_tusa/checkpoint-3000", | |
| "epoch": 20.0, | |
| "eval_steps": 500, | |
| "global_step": 5000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.646, | |
| "eval_f1": 0.5241935483870969, | |
| "eval_loss": 0.7404859662055969, | |
| "eval_precision": 0.7991803278688525, | |
| "eval_recall": 0.39, | |
| "eval_runtime": 7.8525, | |
| "eval_samples_per_second": 127.348, | |
| "eval_steps_per_second": 4.075, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "learning_rate": 4.5e-05, | |
| "loss": 0.344, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.669, | |
| "eval_f1": 0.6182237600922721, | |
| "eval_loss": 0.9954240322113037, | |
| "eval_precision": 0.7302452316076294, | |
| "eval_recall": 0.536, | |
| "eval_runtime": 7.8266, | |
| "eval_samples_per_second": 127.769, | |
| "eval_steps_per_second": 4.089, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.69, | |
| "eval_f1": 0.6784232365145229, | |
| "eval_loss": 0.9894322752952576, | |
| "eval_precision": 0.7047413793103449, | |
| "eval_recall": 0.654, | |
| "eval_runtime": 7.8356, | |
| "eval_samples_per_second": 127.623, | |
| "eval_steps_per_second": 4.084, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "learning_rate": 4e-05, | |
| "loss": 0.157, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.657, | |
| "eval_f1": 0.5862484921592279, | |
| "eval_loss": 1.347570538520813, | |
| "eval_precision": 0.7386018237082067, | |
| "eval_recall": 0.486, | |
| "eval_runtime": 7.8871, | |
| "eval_samples_per_second": 126.789, | |
| "eval_steps_per_second": 4.057, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.643, | |
| "eval_f1": 0.5486725663716814, | |
| "eval_loss": 1.6362618207931519, | |
| "eval_precision": 0.7457044673539519, | |
| "eval_recall": 0.434, | |
| "eval_runtime": 7.8844, | |
| "eval_samples_per_second": 126.833, | |
| "eval_steps_per_second": 4.059, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "learning_rate": 3.5e-05, | |
| "loss": 0.0727, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.649, | |
| "eval_f1": 0.5755743651753326, | |
| "eval_loss": 1.8187189102172852, | |
| "eval_precision": 0.72782874617737, | |
| "eval_recall": 0.476, | |
| "eval_runtime": 7.8876, | |
| "eval_samples_per_second": 126.781, | |
| "eval_steps_per_second": 4.057, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.678, | |
| "eval_f1": 0.6373873873873873, | |
| "eval_loss": 1.8846988677978516, | |
| "eval_precision": 0.729381443298969, | |
| "eval_recall": 0.566, | |
| "eval_runtime": 7.8344, | |
| "eval_samples_per_second": 127.643, | |
| "eval_steps_per_second": 4.085, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "learning_rate": 3e-05, | |
| "loss": 0.0407, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.667, | |
| "eval_f1": 0.622876557191393, | |
| "eval_loss": 2.593871831893921, | |
| "eval_precision": 0.7180156657963447, | |
| "eval_recall": 0.55, | |
| "eval_runtime": 7.8379, | |
| "eval_samples_per_second": 127.584, | |
| "eval_steps_per_second": 4.083, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_accuracy": 0.663, | |
| "eval_f1": 0.6174801362088536, | |
| "eval_loss": 2.2667245864868164, | |
| "eval_precision": 0.7139107611548556, | |
| "eval_recall": 0.544, | |
| "eval_runtime": 7.8522, | |
| "eval_samples_per_second": 127.353, | |
| "eval_steps_per_second": 4.075, | |
| "step": 2250 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "learning_rate": 2.5e-05, | |
| "loss": 0.0219, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_accuracy": 0.683, | |
| "eval_f1": 0.6458100558659218, | |
| "eval_loss": 2.565606117248535, | |
| "eval_precision": 0.7316455696202532, | |
| "eval_recall": 0.578, | |
| "eval_runtime": 7.8306, | |
| "eval_samples_per_second": 127.705, | |
| "eval_steps_per_second": 4.087, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_accuracy": 0.689, | |
| "eval_f1": 0.645381984036488, | |
| "eval_loss": 2.9553158283233643, | |
| "eval_precision": 0.7506631299734748, | |
| "eval_recall": 0.566, | |
| "eval_runtime": 7.8582, | |
| "eval_samples_per_second": 127.256, | |
| "eval_steps_per_second": 4.072, | |
| "step": 2750 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "learning_rate": 2e-05, | |
| "loss": 0.0066, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_accuracy": 0.707, | |
| "eval_f1": 0.6912539515279241, | |
| "eval_loss": 2.7209701538085938, | |
| "eval_precision": 0.7305122494432071, | |
| "eval_recall": 0.656, | |
| "eval_runtime": 7.9302, | |
| "eval_samples_per_second": 126.101, | |
| "eval_steps_per_second": 4.035, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "eval_accuracy": 0.69, | |
| "eval_f1": 0.6630434782608696, | |
| "eval_loss": 2.795156478881836, | |
| "eval_precision": 0.7261904761904762, | |
| "eval_recall": 0.61, | |
| "eval_runtime": 7.8563, | |
| "eval_samples_per_second": 127.287, | |
| "eval_steps_per_second": 4.073, | |
| "step": 3250 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "learning_rate": 1.5e-05, | |
| "loss": 0.0054, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_accuracy": 0.701, | |
| "eval_f1": 0.6862539349422876, | |
| "eval_loss": 2.847877025604248, | |
| "eval_precision": 0.7218543046357616, | |
| "eval_recall": 0.654, | |
| "eval_runtime": 7.8163, | |
| "eval_samples_per_second": 127.938, | |
| "eval_steps_per_second": 4.094, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "eval_accuracy": 0.696, | |
| "eval_f1": 0.68, | |
| "eval_loss": 2.996243715286255, | |
| "eval_precision": 0.7177777777777777, | |
| "eval_recall": 0.646, | |
| "eval_runtime": 7.8761, | |
| "eval_samples_per_second": 126.966, | |
| "eval_steps_per_second": 4.063, | |
| "step": 3750 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0026, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_accuracy": 0.695, | |
| "eval_f1": 0.6709816612729235, | |
| "eval_loss": 3.150722026824951, | |
| "eval_precision": 0.7283372365339579, | |
| "eval_recall": 0.622, | |
| "eval_runtime": 7.8492, | |
| "eval_samples_per_second": 127.402, | |
| "eval_steps_per_second": 4.077, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 17.0, | |
| "eval_accuracy": 0.688, | |
| "eval_f1": 0.6578947368421052, | |
| "eval_loss": 3.2561535835266113, | |
| "eval_precision": 0.7281553398058253, | |
| "eval_recall": 0.6, | |
| "eval_runtime": 7.8888, | |
| "eval_samples_per_second": 126.762, | |
| "eval_steps_per_second": 4.056, | |
| "step": 4250 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "learning_rate": 5e-06, | |
| "loss": 0.0022, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "eval_accuracy": 0.682, | |
| "eval_f1": 0.671487603305785, | |
| "eval_loss": 3.1687259674072266, | |
| "eval_precision": 0.6944444444444444, | |
| "eval_recall": 0.65, | |
| "eval_runtime": 7.8583, | |
| "eval_samples_per_second": 127.254, | |
| "eval_steps_per_second": 4.072, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 19.0, | |
| "eval_accuracy": 0.685, | |
| "eval_f1": 0.672216441207076, | |
| "eval_loss": 3.162193775177002, | |
| "eval_precision": 0.7006507592190889, | |
| "eval_recall": 0.646, | |
| "eval_runtime": 7.8615, | |
| "eval_samples_per_second": 127.202, | |
| "eval_steps_per_second": 4.07, | |
| "step": 4750 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "learning_rate": 0.0, | |
| "loss": 0.0012, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "eval_accuracy": 0.698, | |
| "eval_f1": 0.6766595289079229, | |
| "eval_loss": 3.1741974353790283, | |
| "eval_precision": 0.728110599078341, | |
| "eval_recall": 0.632, | |
| "eval_runtime": 7.8547, | |
| "eval_samples_per_second": 127.313, | |
| "eval_steps_per_second": 4.074, | |
| "step": 5000 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 5000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 20, | |
| "save_steps": 500, | |
| "total_flos": 4.20977688576e+16, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |