Commit 路
a40bd34
1
Parent(s): fa060e4
End of training
Browse files- all_results.json +11 -0
- eval_results.json +6 -0
- train_results.json +8 -0
- trainer_state.json +67 -0
all_results.json
ADDED
|
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.0,
|
| 3 |
+
"eval_exact_match": 20.53926206244087,
|
| 4 |
+
"eval_f1": 32.48117986953082,
|
| 5 |
+
"eval_samples": 10784,
|
| 6 |
+
"train_loss": 3.785540608312353,
|
| 7 |
+
"train_runtime": 234.5103,
|
| 8 |
+
"train_samples": 88524,
|
| 9 |
+
"train_samples_per_second": 754.969,
|
| 10 |
+
"train_steps_per_second": 11.803
|
| 11 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.0,
|
| 3 |
+
"eval_exact_match": 20.53926206244087,
|
| 4 |
+
"eval_f1": 32.48117986953082,
|
| 5 |
+
"eval_samples": 10784
|
| 6 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.0,
|
| 3 |
+
"train_loss": 3.785540608312353,
|
| 4 |
+
"train_runtime": 234.5103,
|
| 5 |
+
"train_samples": 88524,
|
| 6 |
+
"train_samples_per_second": 754.969,
|
| 7 |
+
"train_steps_per_second": 11.803
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,67 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.0,
|
| 5 |
+
"global_step": 2768,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.36,
|
| 12 |
+
"learning_rate": 2.458092485549133e-05,
|
| 13 |
+
"loss": 4.4875,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.72,
|
| 18 |
+
"learning_rate": 1.9161849710982662e-05,
|
| 19 |
+
"loss": 3.9584,
|
| 20 |
+
"step": 1000
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 1.0,
|
| 24 |
+
"eval_exact_match": 14.862819299905393,
|
| 25 |
+
"eval_f1": 26.85226111240762,
|
| 26 |
+
"step": 1384
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 1.08,
|
| 30 |
+
"learning_rate": 1.3742774566473989e-05,
|
| 31 |
+
"loss": 3.7147,
|
| 32 |
+
"step": 1500
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 1.45,
|
| 36 |
+
"learning_rate": 8.323699421965318e-06,
|
| 37 |
+
"loss": 3.5194,
|
| 38 |
+
"step": 2000
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 1.81,
|
| 42 |
+
"learning_rate": 2.9046242774566473e-06,
|
| 43 |
+
"loss": 3.4412,
|
| 44 |
+
"step": 2500
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 2.0,
|
| 48 |
+
"eval_exact_match": 20.53926206244087,
|
| 49 |
+
"eval_f1": 32.48117986953082,
|
| 50 |
+
"step": 2768
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 2.0,
|
| 54 |
+
"step": 2768,
|
| 55 |
+
"total_flos": 161967338160128.0,
|
| 56 |
+
"train_loss": 3.785540608312353,
|
| 57 |
+
"train_runtime": 234.5103,
|
| 58 |
+
"train_samples_per_second": 754.969,
|
| 59 |
+
"train_steps_per_second": 11.803
|
| 60 |
+
}
|
| 61 |
+
],
|
| 62 |
+
"max_steps": 2768,
|
| 63 |
+
"num_train_epochs": 2,
|
| 64 |
+
"total_flos": 161967338160128.0,
|
| 65 |
+
"trial_name": null,
|
| 66 |
+
"trial_params": null
|
| 67 |
+
}
|