Upload trainer_state.json with huggingface_hub
Browse files- trainer_state.json +100 -0
trainer_state.json
ADDED
|
@@ -0,0 +1,100 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 2757,
|
| 3 |
+
"best_metric": 0.8695398280390421,
|
| 4 |
+
"best_model_checkpoint": "outputs/hate-speech-detection/sphobert/checkpoint-2757",
|
| 5 |
+
"epoch": 3.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 2757,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.998911860718172,
|
| 14 |
+
"grad_norm": 4.184622287750244,
|
| 15 |
+
"learning_rate": 1.9996091589167148e-05,
|
| 16 |
+
"loss": 0.4416,
|
| 17 |
+
"step": 918
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 1.0,
|
| 21 |
+
"eval_accuracy": 0.8672093023255814,
|
| 22 |
+
"eval_f1": 0.8518449095766703,
|
| 23 |
+
"eval_loss": 0.3689223527908325,
|
| 24 |
+
"eval_precision": 0.8462192542311449,
|
| 25 |
+
"eval_recall": 0.8672093023255814,
|
| 26 |
+
"eval_runtime": 38.1007,
|
| 27 |
+
"eval_samples_per_second": 112.859,
|
| 28 |
+
"eval_steps_per_second": 3.543,
|
| 29 |
+
"step": 919
|
| 30 |
+
},
|
| 31 |
+
{
|
| 32 |
+
"epoch": 1.997823721436344,
|
| 33 |
+
"grad_norm": 6.81333589553833,
|
| 34 |
+
"learning_rate": 1.9982377977480694e-05,
|
| 35 |
+
"loss": 0.3299,
|
| 36 |
+
"step": 1836
|
| 37 |
+
},
|
| 38 |
+
{
|
| 39 |
+
"epoch": 2.0,
|
| 40 |
+
"eval_accuracy": 0.8806976744186047,
|
| 41 |
+
"eval_f1": 0.8616461626459706,
|
| 42 |
+
"eval_loss": 0.3702216148376465,
|
| 43 |
+
"eval_precision": 0.8647081088093735,
|
| 44 |
+
"eval_recall": 0.8806976744186047,
|
| 45 |
+
"eval_runtime": 40.0641,
|
| 46 |
+
"eval_samples_per_second": 107.328,
|
| 47 |
+
"eval_steps_per_second": 3.37,
|
| 48 |
+
"step": 1838
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"epoch": 2.996735582154516,
|
| 52 |
+
"grad_norm": 7.187419414520264,
|
| 53 |
+
"learning_rate": 1.9958812963917564e-05,
|
| 54 |
+
"loss": 0.2644,
|
| 55 |
+
"step": 2754
|
| 56 |
+
},
|
| 57 |
+
{
|
| 58 |
+
"epoch": 3.0,
|
| 59 |
+
"eval_accuracy": 0.878139534883721,
|
| 60 |
+
"eval_f1": 0.8695398280390421,
|
| 61 |
+
"eval_loss": 0.3386484682559967,
|
| 62 |
+
"eval_precision": 0.8660161855016026,
|
| 63 |
+
"eval_recall": 0.878139534883721,
|
| 64 |
+
"eval_runtime": 40.1821,
|
| 65 |
+
"eval_samples_per_second": 107.013,
|
| 66 |
+
"eval_steps_per_second": 3.36,
|
| 67 |
+
"step": 2757
|
| 68 |
+
}
|
| 69 |
+
],
|
| 70 |
+
"logging_steps": 918,
|
| 71 |
+
"max_steps": 91900,
|
| 72 |
+
"num_input_tokens_seen": 0,
|
| 73 |
+
"num_train_epochs": 100,
|
| 74 |
+
"save_steps": 500,
|
| 75 |
+
"stateful_callbacks": {
|
| 76 |
+
"EarlyStoppingCallback": {
|
| 77 |
+
"args": {
|
| 78 |
+
"early_stopping_patience": 5,
|
| 79 |
+
"early_stopping_threshold": 0.0
|
| 80 |
+
},
|
| 81 |
+
"attributes": {
|
| 82 |
+
"early_stopping_patience_counter": 0
|
| 83 |
+
}
|
| 84 |
+
},
|
| 85 |
+
"TrainerControl": {
|
| 86 |
+
"args": {
|
| 87 |
+
"should_epoch_stop": false,
|
| 88 |
+
"should_evaluate": false,
|
| 89 |
+
"should_log": false,
|
| 90 |
+
"should_save": true,
|
| 91 |
+
"should_training_stop": false
|
| 92 |
+
},
|
| 93 |
+
"attributes": {}
|
| 94 |
+
}
|
| 95 |
+
},
|
| 96 |
+
"total_flos": 0.0,
|
| 97 |
+
"train_batch_size": 32,
|
| 98 |
+
"trial_name": null,
|
| 99 |
+
"trial_params": null
|
| 100 |
+
}
|