NIRVLab/mt5-base-tr-en
048
1{2 "best_global_step": 1955,3 "best_metric": 9.755261490997169,4 "best_model_checkpoint": "/content/mt5_base_baseline_tr_en_100k_5ep/checkpoint-1955",5 "epoch": 5.0,6 "eval_steps": 500,7 "global_step": 1955,8 "is_hyper_param_search": false,9 "is_local_process_zero": true,10 "is_world_process_zero": true,11 "log_history": [12 {13 "epoch": 1.0,14 "grad_norm": 3.609375,15 "learning_rate": 9.468731595555683e-05,16 "loss": 9.6576,17 "step": 39118 },19 {20 "epoch": 1.0,21 "eval_bleu": 3.7407176168912133,22 "eval_chrf_pp": 14.33263610594533,23 "eval_empty_rate": 0.0,24 "eval_gen_len": 4.395,25 "eval_loss": 4.268214702606201,26 "eval_runtime": 90.951,27 "eval_samples_per_second": 21.99,28 "eval_steps_per_second": 0.693,29 "step": 39130 },31 {32 "epoch": 2.0,33 "grad_norm": 2.359375,34 "learning_rate": 7.115889418285841e-05,35 "loss": 4.5467,36 "step": 78237 },38 {39 "epoch": 2.0,40 "eval_bleu": 8.008237501117689,41 "eval_chrf_pp": 22.056956001183675,42 "eval_empty_rate": 0.0,43 "eval_gen_len": 6.2845,44 "eval_loss": 3.9604849815368652,45 "eval_runtime": 81.5042,46 "eval_samples_per_second": 24.539,47 "eval_steps_per_second": 0.773,48 "step": 78249 },50 {51 "epoch": 3.0,52 "grad_norm": 2.015625,53 "learning_rate": 3.851697720368138e-05,54 "loss": 4.3588,55 "step": 117356 },57 {58 "epoch": 3.0,59 "eval_bleu": 9.557890582850064,60 "eval_chrf_pp": 24.373378123346228,61 "eval_empty_rate": 0.0,62 "eval_gen_len": 7.1455,63 "eval_loss": 3.9073450565338135,64 "eval_runtime": 79.4762,65 "eval_samples_per_second": 25.165,66 "eval_steps_per_second": 0.793,67 "step": 117368 },69 {70 "epoch": 4.0,71 "grad_norm": 2.03125,72 "learning_rate": 1.0820993235262333e-05,73 "loss": 4.327,74 "step": 156475 },76 {77 "epoch": 4.0,78 "eval_bleu": 9.670726359576152,79 "eval_chrf_pp": 24.688349222485826,80 "eval_empty_rate": 0.0,81 "eval_gen_len": 7.2725,82 "eval_loss": 3.9009361267089844,83 "eval_runtime": 80.8154,84 "eval_samples_per_second": 24.748,85 "eval_steps_per_second": 0.78,86 "step": 156487 },88 {89 "epoch": 5.0,90 "grad_norm": 2.0,91 "learning_rate": 7.311750030436493e-11,92 "loss": 4.3209,93 "step": 195594 },95 {96 "epoch": 5.0,97 "eval_bleu": 9.755261490997169,98 "eval_chrf_pp": 24.817022146989494,99 "eval_empty_rate": 0.0,100 "eval_gen_len": 7.2195,101 "eval_loss": 3.8945579528808594,102 "eval_runtime": 82.5345,103 "eval_samples_per_second": 24.232,104 "eval_steps_per_second": 0.763,105 "step": 1955106 },107 {108 "epoch": 5.0,109 "step": 1955,110 "total_flos": 2.51499926126592e+16,111 "train_loss": 5.442207168618126,112 "train_runtime": 3581.1772,113 "train_samples_per_second": 139.619,114 "train_steps_per_second": 0.546115 }116 ],117 "logging_steps": 500,118 "max_steps": 1955,119 "num_input_tokens_seen": 0,120 "num_train_epochs": 5,121 "save_steps": 500,122 "stateful_callbacks": {123 "TrainerControl": {124 "args": {125 "should_epoch_stop": false,126 "should_evaluate": false,127 "should_log": false,128 "should_save": true,129 "should_training_stop": true130 },131 "attributes": {}132 }133 },134 "total_flos": 2.51499926126592e+16,135 "train_batch_size": 32,136 "trial_name": null,137 "trial_params": null138}139 