End of training
Browse files- all_results.json +13 -0
- eval_results.json +8 -0
- train_results.json +8 -0
- trainer_state.json +67 -0
all_results.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"eval_accuracy": 0.9904761904761905,
|
| 4 |
+
"eval_loss": 0.08451159298419952,
|
| 5 |
+
"eval_runtime": 1.8446,
|
| 6 |
+
"eval_samples_per_second": 113.843,
|
| 7 |
+
"eval_steps_per_second": 3.795,
|
| 8 |
+
"total_flos": 4.697913185667072e+16,
|
| 9 |
+
"train_loss": 0.6861413399378459,
|
| 10 |
+
"train_runtime": 27.5947,
|
| 11 |
+
"train_samples_per_second": 342.458,
|
| 12 |
+
"train_steps_per_second": 2.718
|
| 13 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"eval_accuracy": 0.9904761904761905,
|
| 4 |
+
"eval_loss": 0.08451159298419952,
|
| 5 |
+
"eval_runtime": 1.8446,
|
| 6 |
+
"eval_samples_per_second": 113.843,
|
| 7 |
+
"eval_steps_per_second": 3.795
|
| 8 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"total_flos": 4.697913185667072e+16,
|
| 4 |
+
"train_loss": 0.6861413399378459,
|
| 5 |
+
"train_runtime": 27.5947,
|
| 6 |
+
"train_samples_per_second": 342.458,
|
| 7 |
+
"train_steps_per_second": 2.718
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,67 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.9904761904761905,
|
| 3 |
+
"best_model_checkpoint": "/kaggle/working/swin-brain-modality-classification/checkpoint-15",
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 15,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.6666666666666666,
|
| 13 |
+
"grad_norm": 12.149126052856445,
|
| 14 |
+
"learning_rate": 4.850746268656717e-05,
|
| 15 |
+
"loss": 0.9131,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 1.0,
|
| 20 |
+
"eval_accuracy": 0.9904761904761905,
|
| 21 |
+
"eval_loss": 0.08451159298419952,
|
| 22 |
+
"eval_runtime": 1.8249,
|
| 23 |
+
"eval_samples_per_second": 115.074,
|
| 24 |
+
"eval_steps_per_second": 3.836,
|
| 25 |
+
"step": 15
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 1.0,
|
| 29 |
+
"step": 15,
|
| 30 |
+
"total_flos": 4.697913185667072e+16,
|
| 31 |
+
"train_loss": 0.6861413399378459,
|
| 32 |
+
"train_runtime": 27.5947,
|
| 33 |
+
"train_samples_per_second": 342.458,
|
| 34 |
+
"train_steps_per_second": 2.718
|
| 35 |
+
}
|
| 36 |
+
],
|
| 37 |
+
"logging_steps": 10,
|
| 38 |
+
"max_steps": 75,
|
| 39 |
+
"num_input_tokens_seen": 0,
|
| 40 |
+
"num_train_epochs": 5,
|
| 41 |
+
"save_steps": 500,
|
| 42 |
+
"stateful_callbacks": {
|
| 43 |
+
"CustomEarlyStoppingCallback": {
|
| 44 |
+
"args": {
|
| 45 |
+
"early_stopping_patience": 1,
|
| 46 |
+
"early_stopping_threshold": 0.0
|
| 47 |
+
},
|
| 48 |
+
"attributes": {
|
| 49 |
+
"early_stopping_patience_counter": 0
|
| 50 |
+
}
|
| 51 |
+
},
|
| 52 |
+
"TrainerControl": {
|
| 53 |
+
"args": {
|
| 54 |
+
"should_epoch_stop": false,
|
| 55 |
+
"should_evaluate": false,
|
| 56 |
+
"should_log": false,
|
| 57 |
+
"should_save": true,
|
| 58 |
+
"should_training_stop": true
|
| 59 |
+
},
|
| 60 |
+
"attributes": {}
|
| 61 |
+
}
|
| 62 |
+
},
|
| 63 |
+
"total_flos": 4.697913185667072e+16,
|
| 64 |
+
"train_batch_size": 32,
|
| 65 |
+
"trial_name": null,
|
| 66 |
+
"trial_params": null
|
| 67 |
+
}
|