smlhd/bert_cn_finetuning

1{2  "best_metric": null,3  "best_model_checkpoint": null,4  "epoch": 1.0,5  "eval_steps": 500,6  "global_step": 8419,7  "is_hyper_param_search": false,8  "is_local_process_zero": true,9  "is_world_process_zero": true,10  "log_history": [11    {12      "epoch": 0.05938947618482005,13      "grad_norm": 3.4233484268188477,14      "learning_rate": 1.88122104763036e-05,15      "loss": 0.666,16      "step": 50017    },18    {19      "epoch": 0.1187789523696401,20      "grad_norm": 21.556442260742188,21      "learning_rate": 1.76244209526072e-05,22      "loss": 0.57,23      "step": 100024    },25    {26      "epoch": 0.17816842855446016,27      "grad_norm": 7.450474739074707,28      "learning_rate": 1.64366314289108e-05,29      "loss": 0.5052,30      "step": 150031    },32    {33      "epoch": 0.2375579047392802,34      "grad_norm": 12.34352970123291,35      "learning_rate": 1.5248841905214398e-05,36      "loss": 0.4877,37      "step": 200038    },39    {40      "epoch": 0.2969473809241002,41      "grad_norm": 16.83225440979004,42      "learning_rate": 1.4061052381517997e-05,43      "loss": 0.4655,44      "step": 250045    },46    {47      "epoch": 0.3563368571089203,48      "grad_norm": 14.64586067199707,49      "learning_rate": 1.2873262857821596e-05,50      "loss": 0.4335,51      "step": 300052    },53    {54      "epoch": 0.41572633329374037,55      "grad_norm": 17.33638572692871,56      "learning_rate": 1.1685473334125195e-05,57      "loss": 0.4348,58      "step": 350059    },60    {61      "epoch": 0.4751158094785604,62      "grad_norm": 22.9376163482666,63      "learning_rate": 1.0497683810428794e-05,64      "loss": 0.413,65      "step": 400066    },67    {68      "epoch": 0.5345052856633804,69      "grad_norm": 21.58479881286621,70      "learning_rate": 9.309894286732392e-06,71      "loss": 0.412,72      "step": 450073    },74    {75      "epoch": 0.5938947618482004,76      "grad_norm": 14.35308837890625,77      "learning_rate": 8.12210476303599e-06,78      "loss": 0.3915,79      "step": 500080    },81    {82      "epoch": 0.6532842380330206,83      "grad_norm": 16.408004760742188,84      "learning_rate": 6.93431523933959e-06,85      "loss": 0.3829,86      "step": 550087    },88    {89      "epoch": 0.7126737142178406,90      "grad_norm": 63.41299819946289,91      "learning_rate": 5.746525715643189e-06,92      "loss": 0.3753,93      "step": 600094    },95    {96      "epoch": 0.7720631904026607,97      "grad_norm": 6.248060703277588,98      "learning_rate": 4.558736191946788e-06,99      "loss": 0.3778,100      "step": 6500101    },102    {103      "epoch": 0.8314526665874807,104      "grad_norm": 15.529841423034668,105      "learning_rate": 3.370946668250386e-06,106      "loss": 0.3709,107      "step": 7000108    },109    {110      "epoch": 0.8908421427723008,111      "grad_norm": 15.034144401550293,112      "learning_rate": 2.183157144553985e-06,113      "loss": 0.3698,114      "step": 7500115    },116    {117      "epoch": 0.9502316189571208,118      "grad_norm": 29.073537826538086,119      "learning_rate": 9.953676208575841e-07,120      "loss": 0.3525,121      "step": 8000122    },123    {124      "epoch": 1.0,125      "step": 8419,126      "total_flos": 4430066616860160.0,127      "train_loss": 0.4344410251191965,128      "train_runtime": 82812.5472,129      "train_samples_per_second": 0.813,130      "train_steps_per_second": 0.102131    }132  ],133  "logging_steps": 500,134  "max_steps": 8419,135  "num_input_tokens_seen": 0,136  "num_train_epochs": 1,137  "save_steps": 500,138  "stateful_callbacks": {139    "TrainerControl": {140      "args": {141        "should_epoch_stop": false,142        "should_evaluate": false,143        "should_log": false,144        "should_save": true,145        "should_training_stop": true146      },147      "attributes": {}148    }149  },150  "total_flos": 4430066616860160.0,151  "train_batch_size": 8,152  "trial_name": null,153  "trial_params": null154}155