Commit
·
fe9a2e9
1
Parent(s):
21958b5
End of training
Browse files
all_results.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"eval_loss": 3.0140793323516846,
|
| 4 |
+
"eval_runtime": 22.3191,
|
| 5 |
+
"eval_samples": 991,
|
| 6 |
+
"eval_samples_per_second": 44.401,
|
| 7 |
+
"eval_steps_per_second": 22.223,
|
| 8 |
+
"train_loss": 3.5130721891769254,
|
| 9 |
+
"train_runtime": 755.7574,
|
| 10 |
+
"train_samples": 3960,
|
| 11 |
+
"train_samples_per_second": 5.24,
|
| 12 |
+
"train_steps_per_second": 2.62
|
| 13 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"eval_loss": 3.0140793323516846,
|
| 4 |
+
"eval_runtime": 22.3191,
|
| 5 |
+
"eval_samples": 991,
|
| 6 |
+
"eval_samples_per_second": 44.401,
|
| 7 |
+
"eval_steps_per_second": 22.223
|
| 8 |
+
}
|
runs/Oct20_17-43-48_f2ba06383ef5/events.out.tfevents.1666291586.f2ba06383ef5.200.2
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:406f42ce8a2c89ac5393e8f1ddab2ca9c1568194e6d903992b47d8908ccafc0d
|
| 3 |
+
size 311
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 3.5130721891769254,
|
| 4 |
+
"train_runtime": 755.7574,
|
| 5 |
+
"train_samples": 3960,
|
| 6 |
+
"train_samples_per_second": 5.24,
|
| 7 |
+
"train_steps_per_second": 2.62
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"global_step": 1980,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.25,
|
| 12 |
+
"learning_rate": 8.314606741573035e-06,
|
| 13 |
+
"loss": 4.2194,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.51,
|
| 18 |
+
"learning_rate": 5.50561797752809e-06,
|
| 19 |
+
"loss": 3.3918,
|
| 20 |
+
"step": 1000
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.51,
|
| 24 |
+
"eval_loss": 3.1021862030029297,
|
| 25 |
+
"eval_runtime": 23.3073,
|
| 26 |
+
"eval_samples_per_second": 42.519,
|
| 27 |
+
"eval_steps_per_second": 21.281,
|
| 28 |
+
"step": 1000
|
| 29 |
+
},
|
| 30 |
+
{
|
| 31 |
+
"epoch": 0.76,
|
| 32 |
+
"learning_rate": 2.696629213483146e-06,
|
| 33 |
+
"loss": 3.2545,
|
| 34 |
+
"step": 1500
|
| 35 |
+
},
|
| 36 |
+
{
|
| 37 |
+
"epoch": 1.0,
|
| 38 |
+
"step": 1980,
|
| 39 |
+
"total_flos": 306526787665920.0,
|
| 40 |
+
"train_loss": 3.5130721891769254,
|
| 41 |
+
"train_runtime": 755.7574,
|
| 42 |
+
"train_samples_per_second": 5.24,
|
| 43 |
+
"train_steps_per_second": 2.62
|
| 44 |
+
}
|
| 45 |
+
],
|
| 46 |
+
"max_steps": 1980,
|
| 47 |
+
"num_train_epochs": 1,
|
| 48 |
+
"total_flos": 306526787665920.0,
|
| 49 |
+
"trial_name": null,
|
| 50 |
+
"trial_params": null
|
| 51 |
+
}
|