Commit ·
6b2e261
1
Parent(s): 684da81
End of training
Browse files
all_results.json
ADDED
|
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"eval_loss": 0.520963728427887,
|
| 4 |
+
"eval_runtime": 229.0131,
|
| 5 |
+
"eval_samples_per_second": 100.483,
|
| 6 |
+
"eval_steps_per_second": 3.144,
|
| 7 |
+
"train_loss": 0.5397671833634377,
|
| 8 |
+
"train_runtime": 281.0136,
|
| 9 |
+
"train_samples_per_second": 3.559,
|
| 10 |
+
"train_steps_per_second": 0.114
|
| 11 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"eval_loss": 0.520963728427887,
|
| 4 |
+
"eval_runtime": 229.0131,
|
| 5 |
+
"eval_samples_per_second": 100.483,
|
| 6 |
+
"eval_steps_per_second": 3.144
|
| 7 |
+
}
|
runs/Mar10_07-42-32_9e4a255d196e/events.out.tfevents.1646899252.9e4a255d196e.755.2
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:711372d0a251a27b50fd34b83b29066de11bdefb317c3fba230dd24d42b40011
|
| 3 |
+
size 306
|
train_results.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.5397671833634377,
|
| 4 |
+
"train_runtime": 281.0136,
|
| 5 |
+
"train_samples_per_second": 3.559,
|
| 6 |
+
"train_steps_per_second": 0.114
|
| 7 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.5208448767662048,
|
| 3 |
+
"best_model_checkpoint": "./outputs/checkpoint-32",
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"global_step": 32,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.31,
|
| 12 |
+
"learning_rate": 1.375e-05,
|
| 13 |
+
"loss": 0.566,
|
| 14 |
+
"step": 10
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.62,
|
| 18 |
+
"learning_rate": 7.500000000000001e-06,
|
| 19 |
+
"loss": 0.5404,
|
| 20 |
+
"step": 20
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.94,
|
| 24 |
+
"learning_rate": 1.25e-06,
|
| 25 |
+
"loss": 0.5198,
|
| 26 |
+
"step": 30
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 1.0,
|
| 30 |
+
"eval_loss": 0.5208448767662048,
|
| 31 |
+
"eval_runtime": 244.3236,
|
| 32 |
+
"eval_samples_per_second": 94.187,
|
| 33 |
+
"eval_steps_per_second": 2.947,
|
| 34 |
+
"step": 32
|
| 35 |
+
},
|
| 36 |
+
{
|
| 37 |
+
"epoch": 1.0,
|
| 38 |
+
"step": 32,
|
| 39 |
+
"total_flos": 7.8024697970688e+16,
|
| 40 |
+
"train_loss": 0.5397671833634377,
|
| 41 |
+
"train_runtime": 281.0136,
|
| 42 |
+
"train_samples_per_second": 3.559,
|
| 43 |
+
"train_steps_per_second": 0.114
|
| 44 |
+
}
|
| 45 |
+
],
|
| 46 |
+
"max_steps": 32,
|
| 47 |
+
"num_train_epochs": 1,
|
| 48 |
+
"total_flos": 7.8024697970688e+16,
|
| 49 |
+
"trial_name": null,
|
| 50 |
+
"trial_params": null
|
| 51 |
+
}
|