| { | |
| "best_metric": 80.63612438300382, | |
| "best_model_checkpoint": "/root/turkic_qa/tr_uzn_models/tr_uzn_xlm_roberta_base_squad_model/checkpoint-8120", | |
| "epoch": 10.0, | |
| "eval_steps": 500, | |
| "global_step": 8120, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "step": 812, | |
| "train_exact_match": 56.743256743256744, | |
| "train_f1": 75.16234224627293, | |
| "train_runtime": 18.93, | |
| "train_samples_per_second": 87.692, | |
| "train_steps_per_second": 3.17 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 33.5258674621582, | |
| "learning_rate": 5e-06, | |
| "loss": 1.3302, | |
| "step": 812 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_exact_match": 56.78125, | |
| "eval_f1": 74.92855180045365, | |
| "eval_runtime": 58.0842, | |
| "eval_samples_per_second": 87.993, | |
| "eval_steps_per_second": 3.151, | |
| "step": 812 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "step": 1624, | |
| "train_exact_match": 66.53346653346654, | |
| "train_f1": 81.01468766554503, | |
| "train_runtime": 19.1594, | |
| "train_samples_per_second": 84.919, | |
| "train_steps_per_second": 3.079 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 39.375850677490234, | |
| "learning_rate": 1e-05, | |
| "loss": 0.9867, | |
| "step": 1624 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_exact_match": 60.71875, | |
| "eval_f1": 77.48042477059983, | |
| "eval_runtime": 59.4097, | |
| "eval_samples_per_second": 86.03, | |
| "eval_steps_per_second": 3.08, | |
| "step": 1624 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 2436, | |
| "train_exact_match": 73.32667332667333, | |
| "train_f1": 86.12936978278546, | |
| "train_runtime": 18.141, | |
| "train_samples_per_second": 87.04, | |
| "train_steps_per_second": 3.142 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 43.98362350463867, | |
| "learning_rate": 8.750000000000001e-06, | |
| "loss": 0.7954, | |
| "step": 2436 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_exact_match": 63.53125, | |
| "eval_f1": 79.36767777349903, | |
| "eval_runtime": 58.582, | |
| "eval_samples_per_second": 87.245, | |
| "eval_steps_per_second": 3.124, | |
| "step": 2436 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 3248, | |
| "train_exact_match": 74.72527472527473, | |
| "train_f1": 88.40189950318806, | |
| "train_runtime": 18.0666, | |
| "train_samples_per_second": 88.174, | |
| "train_steps_per_second": 3.155 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 28.4106502532959, | |
| "learning_rate": 7.500000000000001e-06, | |
| "loss": 0.6427, | |
| "step": 3248 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_exact_match": 65.21875, | |
| "eval_f1": 80.19939907948, | |
| "eval_runtime": 57.9356, | |
| "eval_samples_per_second": 88.219, | |
| "eval_steps_per_second": 3.159, | |
| "step": 3248 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 4060, | |
| "train_exact_match": 79.72027972027972, | |
| "train_f1": 90.39293523910672, | |
| "train_runtime": 18.0742, | |
| "train_samples_per_second": 88.303, | |
| "train_steps_per_second": 3.154 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 58.56404495239258, | |
| "learning_rate": 6.25e-06, | |
| "loss": 0.5287, | |
| "step": 4060 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_exact_match": 64.53125, | |
| "eval_f1": 79.76982128782555, | |
| "eval_runtime": 57.84, | |
| "eval_samples_per_second": 88.365, | |
| "eval_steps_per_second": 3.164, | |
| "step": 4060 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "step": 4872, | |
| "train_exact_match": 79.42057942057941, | |
| "train_f1": 91.11407011087398, | |
| "train_runtime": 18.4811, | |
| "train_samples_per_second": 88.252, | |
| "train_steps_per_second": 3.192 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 45.58940124511719, | |
| "learning_rate": 5e-06, | |
| "loss": 0.4506, | |
| "step": 4872 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_exact_match": 65.46875, | |
| "eval_f1": 80.18984733785271, | |
| "eval_runtime": 58.0085, | |
| "eval_samples_per_second": 88.108, | |
| "eval_steps_per_second": 3.155, | |
| "step": 4872 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "step": 5684, | |
| "train_exact_match": 80.41958041958041, | |
| "train_f1": 91.05656126706158, | |
| "train_runtime": 17.8755, | |
| "train_samples_per_second": 89.172, | |
| "train_steps_per_second": 3.189 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "grad_norm": 26.53660774230957, | |
| "learning_rate": 3.7500000000000005e-06, | |
| "loss": 0.3805, | |
| "step": 5684 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_exact_match": 65.125, | |
| "eval_f1": 80.04186683186089, | |
| "eval_runtime": 57.4933, | |
| "eval_samples_per_second": 88.897, | |
| "eval_steps_per_second": 3.183, | |
| "step": 5684 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "step": 6496, | |
| "train_exact_match": 86.31368631368632, | |
| "train_f1": 93.75812824286764, | |
| "train_runtime": 17.4124, | |
| "train_samples_per_second": 88.328, | |
| "train_steps_per_second": 3.159 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 34.22230911254883, | |
| "learning_rate": 2.5e-06, | |
| "loss": 0.3381, | |
| "step": 6496 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_exact_match": 65.90625, | |
| "eval_f1": 80.14333737032665, | |
| "eval_runtime": 57.8402, | |
| "eval_samples_per_second": 88.364, | |
| "eval_steps_per_second": 3.164, | |
| "step": 6496 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "step": 7308, | |
| "train_exact_match": 86.01398601398601, | |
| "train_f1": 94.05045602201064, | |
| "train_runtime": 18.8091, | |
| "train_samples_per_second": 86.767, | |
| "train_steps_per_second": 3.137 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "grad_norm": 87.17084503173828, | |
| "learning_rate": 1.25e-06, | |
| "loss": 0.3032, | |
| "step": 7308 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_exact_match": 66.21875, | |
| "eval_f1": 80.52702562432783, | |
| "eval_runtime": 58.1783, | |
| "eval_samples_per_second": 87.851, | |
| "eval_steps_per_second": 3.146, | |
| "step": 7308 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 8120, | |
| "train_exact_match": 86.41358641358642, | |
| "train_f1": 94.52312811018126, | |
| "train_runtime": 17.2785, | |
| "train_samples_per_second": 87.276, | |
| "train_steps_per_second": 3.125 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 31.117448806762695, | |
| "learning_rate": 0.0, | |
| "loss": 0.2815, | |
| "step": 8120 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_exact_match": 66.4375, | |
| "eval_f1": 80.63612438300382, | |
| "eval_runtime": 58.3935, | |
| "eval_samples_per_second": 87.527, | |
| "eval_steps_per_second": 3.134, | |
| "step": 8120 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 8120, | |
| "total_flos": 4.451320899376128e+16, | |
| "train_loss": 0.6037781062384544, | |
| "train_runtime": 5220.6695, | |
| "train_samples_per_second": 43.508, | |
| "train_steps_per_second": 1.555 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 8120, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "total_flos": 4.451320899376128e+16, | |
| "train_batch_size": 28, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |