| { | |
| "best_metric": 74.67606866807955, | |
| "best_model_checkpoint": "/root/turkic_qa/ru_kaz_models/ru_kaz_xlm_roberta_base_model/checkpoint-4416", | |
| "epoch": 10.0, | |
| "eval_steps": 500, | |
| "global_step": 5520, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "step": 552, | |
| "train_exact_match": 24.275724275724276, | |
| "train_f1": 40.76875114599373, | |
| "train_runtime": 11.4713, | |
| "train_samples_per_second": 88.569, | |
| "train_steps_per_second": 3.225 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 50.65106201171875, | |
| "learning_rate": 5e-06, | |
| "loss": 4.6936, | |
| "step": 552 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_exact_match": 24.4375, | |
| "eval_f1": 40.89175449073484, | |
| "eval_runtime": 36.7427, | |
| "eval_samples_per_second": 89.297, | |
| "eval_steps_per_second": 3.212, | |
| "step": 552 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "step": 1104, | |
| "train_exact_match": 56.54345654345654, | |
| "train_f1": 72.7903191757748, | |
| "train_runtime": 11.4595, | |
| "train_samples_per_second": 89.707, | |
| "train_steps_per_second": 3.229 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 61.147613525390625, | |
| "learning_rate": 1e-05, | |
| "loss": 2.162, | |
| "step": 1104 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_exact_match": 48.28125, | |
| "eval_f1": 67.10891985034017, | |
| "eval_runtime": 36.689, | |
| "eval_samples_per_second": 89.427, | |
| "eval_steps_per_second": 3.216, | |
| "step": 1104 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 1656, | |
| "train_exact_match": 60.33966033966034, | |
| "train_f1": 77.53847505132588, | |
| "train_runtime": 11.4178, | |
| "train_samples_per_second": 89.072, | |
| "train_steps_per_second": 3.241 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 43.21469497680664, | |
| "learning_rate": 8.750000000000001e-06, | |
| "loss": 1.4851, | |
| "step": 1656 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_exact_match": 54.59375, | |
| "eval_f1": 72.32243072865627, | |
| "eval_runtime": 36.5153, | |
| "eval_samples_per_second": 89.853, | |
| "eval_steps_per_second": 3.232, | |
| "step": 1656 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 2208, | |
| "train_exact_match": 68.13186813186813, | |
| "train_f1": 83.71981586843339, | |
| "train_runtime": 11.3106, | |
| "train_samples_per_second": 89.65, | |
| "train_steps_per_second": 3.271 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 39.521087646484375, | |
| "learning_rate": 7.500000000000001e-06, | |
| "loss": 1.1864, | |
| "step": 2208 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_exact_match": 55.6875, | |
| "eval_f1": 73.33848461420202, | |
| "eval_runtime": 36.4825, | |
| "eval_samples_per_second": 89.934, | |
| "eval_steps_per_second": 3.234, | |
| "step": 2208 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 2760, | |
| "train_exact_match": 72.82717282717283, | |
| "train_f1": 87.11695605040637, | |
| "train_runtime": 11.478, | |
| "train_samples_per_second": 89.04, | |
| "train_steps_per_second": 3.224 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 14.2109956741333, | |
| "learning_rate": 6.25e-06, | |
| "loss": 1.0004, | |
| "step": 2760 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_exact_match": 57.5, | |
| "eval_f1": 74.39778409830554, | |
| "eval_runtime": 36.7472, | |
| "eval_samples_per_second": 89.286, | |
| "eval_steps_per_second": 3.211, | |
| "step": 2760 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "step": 3312, | |
| "train_exact_match": 76.62337662337663, | |
| "train_f1": 89.20109871672176, | |
| "train_runtime": 11.3687, | |
| "train_samples_per_second": 89.896, | |
| "train_steps_per_second": 3.255 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 37.63686752319336, | |
| "learning_rate": 5e-06, | |
| "loss": 0.8721, | |
| "step": 3312 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_exact_match": 57.84375, | |
| "eval_f1": 74.55139589840982, | |
| "eval_runtime": 36.3591, | |
| "eval_samples_per_second": 90.239, | |
| "eval_steps_per_second": 3.245, | |
| "step": 3312 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "step": 3864, | |
| "train_exact_match": 78.42157842157842, | |
| "train_f1": 91.01806696135736, | |
| "train_runtime": 11.4573, | |
| "train_samples_per_second": 89.725, | |
| "train_steps_per_second": 3.229 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "grad_norm": 21.571279525756836, | |
| "learning_rate": 3.7500000000000005e-06, | |
| "loss": 0.7755, | |
| "step": 3864 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_exact_match": 57.6875, | |
| "eval_f1": 74.44520627477633, | |
| "eval_runtime": 36.4488, | |
| "eval_samples_per_second": 90.017, | |
| "eval_steps_per_second": 3.237, | |
| "step": 3864 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "step": 4416, | |
| "train_exact_match": 79.12087912087912, | |
| "train_f1": 91.65370167331513, | |
| "train_runtime": 11.3648, | |
| "train_samples_per_second": 90.103, | |
| "train_steps_per_second": 3.256 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 57.543540954589844, | |
| "learning_rate": 2.5e-06, | |
| "loss": 0.7089, | |
| "step": 4416 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_exact_match": 58.09375, | |
| "eval_f1": 74.67606866807955, | |
| "eval_runtime": 36.2701, | |
| "eval_samples_per_second": 90.46, | |
| "eval_steps_per_second": 3.253, | |
| "step": 4416 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "step": 4968, | |
| "train_exact_match": 80.21978021978022, | |
| "train_f1": 92.17230025305071, | |
| "train_runtime": 11.629, | |
| "train_samples_per_second": 88.916, | |
| "train_steps_per_second": 3.182 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "grad_norm": 21.385902404785156, | |
| "learning_rate": 1.25e-06, | |
| "loss": 0.6522, | |
| "step": 4968 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_exact_match": 58.3125, | |
| "eval_f1": 74.6689953301795, | |
| "eval_runtime": 36.9735, | |
| "eval_samples_per_second": 88.739, | |
| "eval_steps_per_second": 3.191, | |
| "step": 4968 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 5520, | |
| "train_exact_match": 80.21978021978022, | |
| "train_f1": 91.90698139768105, | |
| "train_runtime": 11.5232, | |
| "train_samples_per_second": 89.472, | |
| "train_steps_per_second": 3.211 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 32.45027160644531, | |
| "learning_rate": 0.0, | |
| "loss": 0.6146, | |
| "step": 5520 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_exact_match": 57.75, | |
| "eval_f1": 74.50735271348769, | |
| "eval_runtime": 36.529, | |
| "eval_samples_per_second": 89.819, | |
| "eval_steps_per_second": 3.23, | |
| "step": 5520 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 5520, | |
| "total_flos": 3.025228525300224e+16, | |
| "train_loss": 1.4150862375895181, | |
| "train_runtime": 3533.8403, | |
| "train_samples_per_second": 43.683, | |
| "train_steps_per_second": 1.562 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 5520, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "total_flos": 3.025228525300224e+16, | |
| "train_batch_size": 28, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |