| { | |
| "best_metric": 33.144610036959996, | |
| "best_model_checkpoint": "/root/turkic_qa/en_uzn_models/orig_uzn_roberta_base_model/checkpoint-5814", | |
| "epoch": 10.0, | |
| "eval_steps": 500, | |
| "global_step": 6460, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "step": 646, | |
| "train_exact_match": 6.993006993006993, | |
| "train_f1": 13.61593527592794, | |
| "train_runtime": 10.7947, | |
| "train_samples_per_second": 109.498, | |
| "train_steps_per_second": 3.983 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 12.138856887817383, | |
| "learning_rate": 5e-06, | |
| "loss": 4.6327, | |
| "step": 646 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_exact_match": 6.59375, | |
| "eval_f1": 13.607989660328585, | |
| "eval_runtime": 35.2774, | |
| "eval_samples_per_second": 107.803, | |
| "eval_steps_per_second": 3.855, | |
| "step": 646 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "step": 1292, | |
| "train_exact_match": 13.986013986013987, | |
| "train_f1": 22.14711239214366, | |
| "train_runtime": 10.8509, | |
| "train_samples_per_second": 109.669, | |
| "train_steps_per_second": 3.963 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 15.265068054199219, | |
| "learning_rate": 1e-05, | |
| "loss": 3.5453, | |
| "step": 1292 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_exact_match": 11.40625, | |
| "eval_f1": 18.47470115180993, | |
| "eval_runtime": 34.4859, | |
| "eval_samples_per_second": 110.277, | |
| "eval_steps_per_second": 3.944, | |
| "step": 1292 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 1938, | |
| "train_exact_match": 23.276723276723278, | |
| "train_f1": 32.096631211616916, | |
| "train_runtime": 11.0171, | |
| "train_samples_per_second": 110.102, | |
| "train_steps_per_second": 3.994 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 14.558008193969727, | |
| "learning_rate": 8.750000000000001e-06, | |
| "loss": 3.1006, | |
| "step": 1938 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_exact_match": 17.84375, | |
| "eval_f1": 25.580935818007237, | |
| "eval_runtime": 34.7591, | |
| "eval_samples_per_second": 109.41, | |
| "eval_steps_per_second": 3.913, | |
| "step": 1938 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 2584, | |
| "train_exact_match": 31.16883116883117, | |
| "train_f1": 39.17139975564133, | |
| "train_runtime": 11.0165, | |
| "train_samples_per_second": 109.381, | |
| "train_steps_per_second": 3.994 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 17.505403518676758, | |
| "learning_rate": 7.500000000000001e-06, | |
| "loss": 2.7652, | |
| "step": 2584 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_exact_match": 19.40625, | |
| "eval_f1": 27.64189036545174, | |
| "eval_runtime": 34.3705, | |
| "eval_samples_per_second": 110.647, | |
| "eval_steps_per_second": 3.957, | |
| "step": 2584 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 3230, | |
| "train_exact_match": 36.663336663336665, | |
| "train_f1": 46.09688684971856, | |
| "train_runtime": 10.8186, | |
| "train_samples_per_second": 109.349, | |
| "train_steps_per_second": 3.975 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 20.703981399536133, | |
| "learning_rate": 6.25e-06, | |
| "loss": 2.5139, | |
| "step": 3230 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_exact_match": 20.4375, | |
| "eval_f1": 29.34032966285343, | |
| "eval_runtime": 34.5691, | |
| "eval_samples_per_second": 110.011, | |
| "eval_steps_per_second": 3.934, | |
| "step": 3230 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "step": 3876, | |
| "train_exact_match": 41.658341658341655, | |
| "train_f1": 50.86006517307416, | |
| "train_runtime": 10.7595, | |
| "train_samples_per_second": 110.042, | |
| "train_steps_per_second": 3.996 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 20.977209091186523, | |
| "learning_rate": 5e-06, | |
| "loss": 2.3079, | |
| "step": 3876 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_exact_match": 21.5, | |
| "eval_f1": 30.58201123722236, | |
| "eval_runtime": 34.5751, | |
| "eval_samples_per_second": 109.992, | |
| "eval_steps_per_second": 3.933, | |
| "step": 3876 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "step": 4522, | |
| "train_exact_match": 47.45254745254745, | |
| "train_f1": 56.694415842055, | |
| "train_runtime": 10.7135, | |
| "train_samples_per_second": 109.021, | |
| "train_steps_per_second": 3.92 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "grad_norm": 22.69150161743164, | |
| "learning_rate": 3.7500000000000005e-06, | |
| "loss": 2.111, | |
| "step": 4522 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_exact_match": 22.03125, | |
| "eval_f1": 31.858315646855917, | |
| "eval_runtime": 34.9801, | |
| "eval_samples_per_second": 108.719, | |
| "eval_steps_per_second": 3.888, | |
| "step": 4522 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "step": 5168, | |
| "train_exact_match": 50.74925074925075, | |
| "train_f1": 60.56020993278913, | |
| "train_runtime": 10.9489, | |
| "train_samples_per_second": 108.595, | |
| "train_steps_per_second": 3.927 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 27.088655471801758, | |
| "learning_rate": 2.5e-06, | |
| "loss": 1.9761, | |
| "step": 5168 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_exact_match": 22.71875, | |
| "eval_f1": 32.80883643467974, | |
| "eval_runtime": 35.0553, | |
| "eval_samples_per_second": 108.486, | |
| "eval_steps_per_second": 3.88, | |
| "step": 5168 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "step": 5814, | |
| "train_exact_match": 53.84615384615385, | |
| "train_f1": 63.610163638389785, | |
| "train_runtime": 11.1443, | |
| "train_samples_per_second": 108.755, | |
| "train_steps_per_second": 3.948 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "grad_norm": 33.555702209472656, | |
| "learning_rate": 1.25e-06, | |
| "loss": 1.8691, | |
| "step": 5814 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_exact_match": 22.8125, | |
| "eval_f1": 33.144610036959996, | |
| "eval_runtime": 35.2983, | |
| "eval_samples_per_second": 107.739, | |
| "eval_steps_per_second": 3.853, | |
| "step": 5814 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 6460, | |
| "train_exact_match": 55.344655344655344, | |
| "train_f1": 64.42442903231023, | |
| "train_runtime": 10.756, | |
| "train_samples_per_second": 110.636, | |
| "train_steps_per_second": 3.998 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 31.045358657836914, | |
| "learning_rate": 0.0, | |
| "loss": 1.8088, | |
| "step": 6460 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_exact_match": 22.84375, | |
| "eval_f1": 33.05957357823847, | |
| "eval_runtime": 34.2632, | |
| "eval_samples_per_second": 110.994, | |
| "eval_steps_per_second": 3.969, | |
| "step": 6460 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 6460, | |
| "total_flos": 1.770872117833728e+16, | |
| "train_loss": 2.663083348702351, | |
| "train_runtime": 2359.5642, | |
| "train_samples_per_second": 76.59, | |
| "train_steps_per_second": 2.738 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 6460, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "total_flos": 1.770872117833728e+16, | |
| "train_batch_size": 28, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |