End of training
Browse files
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 4.2,
|
| 3 |
+
"eval_accuracy": 0.9161290322580645,
|
| 4 |
+
"eval_loss": 0.2619659900665283,
|
| 5 |
+
"eval_runtime": 43.4165,
|
| 6 |
+
"eval_samples_per_second": 3.57,
|
| 7 |
+
"eval_steps_per_second": 3.57
|
| 8 |
+
}
|
runs/Aug10_21-50-28_f00da3a1a99f/events.out.tfevents.1691705446.f00da3a1a99f.152.1
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:092fc6a228c87e5644a1bbe390a9cb7fd591fb578a97048b134c31b6e78e1300
|
| 3 |
+
size 734
|
test_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 4.2,
|
| 3 |
+
"eval_accuracy": 0.9161290322580645,
|
| 4 |
+
"eval_loss": 0.2619659900665283,
|
| 5 |
+
"eval_runtime": 43.4165,
|
| 6 |
+
"eval_samples_per_second": 3.57,
|
| 7 |
+
"eval_steps_per_second": 3.57
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,988 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.9,
|
| 3 |
+
"best_model_checkpoint": "videomae-base-finetuned-ucf101-subset/checkpoint-1500",
|
| 4 |
+
"epoch": 4.2,
|
| 5 |
+
"global_step": 1500,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.01,
|
| 12 |
+
"learning_rate": 3.3333333333333333e-06,
|
| 13 |
+
"loss": 2.306,
|
| 14 |
+
"step": 10
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.01,
|
| 18 |
+
"learning_rate": 6.666666666666667e-06,
|
| 19 |
+
"loss": 2.1606,
|
| 20 |
+
"step": 20
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.02,
|
| 24 |
+
"learning_rate": 1e-05,
|
| 25 |
+
"loss": 2.4268,
|
| 26 |
+
"step": 30
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.03,
|
| 30 |
+
"learning_rate": 1.3333333333333333e-05,
|
| 31 |
+
"loss": 2.3848,
|
| 32 |
+
"step": 40
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.03,
|
| 36 |
+
"learning_rate": 1.6666666666666667e-05,
|
| 37 |
+
"loss": 2.3762,
|
| 38 |
+
"step": 50
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.04,
|
| 42 |
+
"learning_rate": 2e-05,
|
| 43 |
+
"loss": 2.274,
|
| 44 |
+
"step": 60
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.05,
|
| 48 |
+
"learning_rate": 2.3333333333333336e-05,
|
| 49 |
+
"loss": 2.0968,
|
| 50 |
+
"step": 70
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.05,
|
| 54 |
+
"learning_rate": 2.6666666666666667e-05,
|
| 55 |
+
"loss": 2.2914,
|
| 56 |
+
"step": 80
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.06,
|
| 60 |
+
"learning_rate": 3e-05,
|
| 61 |
+
"loss": 2.063,
|
| 62 |
+
"step": 90
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.07,
|
| 66 |
+
"learning_rate": 3.3333333333333335e-05,
|
| 67 |
+
"loss": 2.2059,
|
| 68 |
+
"step": 100
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.07,
|
| 72 |
+
"learning_rate": 3.6666666666666666e-05,
|
| 73 |
+
"loss": 2.3758,
|
| 74 |
+
"step": 110
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.08,
|
| 78 |
+
"learning_rate": 4e-05,
|
| 79 |
+
"loss": 2.221,
|
| 80 |
+
"step": 120
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.09,
|
| 84 |
+
"learning_rate": 4.3333333333333334e-05,
|
| 85 |
+
"loss": 2.268,
|
| 86 |
+
"step": 130
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.09,
|
| 90 |
+
"learning_rate": 4.666666666666667e-05,
|
| 91 |
+
"loss": 2.1193,
|
| 92 |
+
"step": 140
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.1,
|
| 96 |
+
"learning_rate": 5e-05,
|
| 97 |
+
"loss": 1.8585,
|
| 98 |
+
"step": 150
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.11,
|
| 102 |
+
"learning_rate": 4.962962962962963e-05,
|
| 103 |
+
"loss": 1.8749,
|
| 104 |
+
"step": 160
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.11,
|
| 108 |
+
"learning_rate": 4.925925925925926e-05,
|
| 109 |
+
"loss": 1.8721,
|
| 110 |
+
"step": 170
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.12,
|
| 114 |
+
"learning_rate": 4.888888888888889e-05,
|
| 115 |
+
"loss": 2.1232,
|
| 116 |
+
"step": 180
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.13,
|
| 120 |
+
"learning_rate": 4.851851851851852e-05,
|
| 121 |
+
"loss": 2.2728,
|
| 122 |
+
"step": 190
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.13,
|
| 126 |
+
"learning_rate": 4.814814814814815e-05,
|
| 127 |
+
"loss": 1.5833,
|
| 128 |
+
"step": 200
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.14,
|
| 132 |
+
"learning_rate": 4.7777777777777784e-05,
|
| 133 |
+
"loss": 2.4943,
|
| 134 |
+
"step": 210
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.15,
|
| 138 |
+
"learning_rate": 4.740740740740741e-05,
|
| 139 |
+
"loss": 1.4056,
|
| 140 |
+
"step": 220
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.15,
|
| 144 |
+
"learning_rate": 4.703703703703704e-05,
|
| 145 |
+
"loss": 1.9793,
|
| 146 |
+
"step": 230
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.16,
|
| 150 |
+
"learning_rate": 4.666666666666667e-05,
|
| 151 |
+
"loss": 2.2575,
|
| 152 |
+
"step": 240
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.17,
|
| 156 |
+
"learning_rate": 4.62962962962963e-05,
|
| 157 |
+
"loss": 1.6757,
|
| 158 |
+
"step": 250
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.17,
|
| 162 |
+
"learning_rate": 4.592592592592593e-05,
|
| 163 |
+
"loss": 2.4101,
|
| 164 |
+
"step": 260
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.18,
|
| 168 |
+
"learning_rate": 4.555555555555556e-05,
|
| 169 |
+
"loss": 1.5145,
|
| 170 |
+
"step": 270
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.19,
|
| 174 |
+
"learning_rate": 4.518518518518519e-05,
|
| 175 |
+
"loss": 1.7479,
|
| 176 |
+
"step": 280
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.19,
|
| 180 |
+
"learning_rate": 4.481481481481482e-05,
|
| 181 |
+
"loss": 1.5142,
|
| 182 |
+
"step": 290
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.2,
|
| 186 |
+
"learning_rate": 4.4444444444444447e-05,
|
| 187 |
+
"loss": 1.7238,
|
| 188 |
+
"step": 300
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.2,
|
| 192 |
+
"eval_accuracy": 0.22857142857142856,
|
| 193 |
+
"eval_loss": 2.0761706829071045,
|
| 194 |
+
"eval_runtime": 19.2571,
|
| 195 |
+
"eval_samples_per_second": 3.635,
|
| 196 |
+
"eval_steps_per_second": 3.635,
|
| 197 |
+
"step": 300
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 1.01,
|
| 201 |
+
"learning_rate": 4.4074074074074076e-05,
|
| 202 |
+
"loss": 1.7489,
|
| 203 |
+
"step": 310
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 1.01,
|
| 207 |
+
"learning_rate": 4.3703703703703705e-05,
|
| 208 |
+
"loss": 1.4939,
|
| 209 |
+
"step": 320
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 1.02,
|
| 213 |
+
"learning_rate": 4.3333333333333334e-05,
|
| 214 |
+
"loss": 1.6296,
|
| 215 |
+
"step": 330
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 1.03,
|
| 219 |
+
"learning_rate": 4.296296296296296e-05,
|
| 220 |
+
"loss": 1.3562,
|
| 221 |
+
"step": 340
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 1.03,
|
| 225 |
+
"learning_rate": 4.259259259259259e-05,
|
| 226 |
+
"loss": 0.5818,
|
| 227 |
+
"step": 350
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 1.04,
|
| 231 |
+
"learning_rate": 4.222222222222222e-05,
|
| 232 |
+
"loss": 1.1949,
|
| 233 |
+
"step": 360
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 1.05,
|
| 237 |
+
"learning_rate": 4.185185185185185e-05,
|
| 238 |
+
"loss": 1.2505,
|
| 239 |
+
"step": 370
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 1.05,
|
| 243 |
+
"learning_rate": 4.148148148148148e-05,
|
| 244 |
+
"loss": 1.0854,
|
| 245 |
+
"step": 380
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 1.06,
|
| 249 |
+
"learning_rate": 4.111111111111111e-05,
|
| 250 |
+
"loss": 1.732,
|
| 251 |
+
"step": 390
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 1.07,
|
| 255 |
+
"learning_rate": 4.074074074074074e-05,
|
| 256 |
+
"loss": 1.0014,
|
| 257 |
+
"step": 400
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 1.07,
|
| 261 |
+
"learning_rate": 4.0370370370370374e-05,
|
| 262 |
+
"loss": 0.8794,
|
| 263 |
+
"step": 410
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 1.08,
|
| 267 |
+
"learning_rate": 4e-05,
|
| 268 |
+
"loss": 1.1037,
|
| 269 |
+
"step": 420
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 1.09,
|
| 273 |
+
"learning_rate": 3.962962962962963e-05,
|
| 274 |
+
"loss": 1.4254,
|
| 275 |
+
"step": 430
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 1.09,
|
| 279 |
+
"learning_rate": 3.925925925925926e-05,
|
| 280 |
+
"loss": 0.7539,
|
| 281 |
+
"step": 440
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 1.1,
|
| 285 |
+
"learning_rate": 3.888888888888889e-05,
|
| 286 |
+
"loss": 1.0575,
|
| 287 |
+
"step": 450
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 1.11,
|
| 291 |
+
"learning_rate": 3.851851851851852e-05,
|
| 292 |
+
"loss": 0.9441,
|
| 293 |
+
"step": 460
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 1.11,
|
| 297 |
+
"learning_rate": 3.814814814814815e-05,
|
| 298 |
+
"loss": 0.8284,
|
| 299 |
+
"step": 470
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 1.12,
|
| 303 |
+
"learning_rate": 3.777777777777778e-05,
|
| 304 |
+
"loss": 1.5265,
|
| 305 |
+
"step": 480
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 1.13,
|
| 309 |
+
"learning_rate": 3.740740740740741e-05,
|
| 310 |
+
"loss": 1.0127,
|
| 311 |
+
"step": 490
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 1.13,
|
| 315 |
+
"learning_rate": 3.7037037037037037e-05,
|
| 316 |
+
"loss": 0.9213,
|
| 317 |
+
"step": 500
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 1.14,
|
| 321 |
+
"learning_rate": 3.6666666666666666e-05,
|
| 322 |
+
"loss": 1.0182,
|
| 323 |
+
"step": 510
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 1.15,
|
| 327 |
+
"learning_rate": 3.62962962962963e-05,
|
| 328 |
+
"loss": 0.9835,
|
| 329 |
+
"step": 520
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 1.15,
|
| 333 |
+
"learning_rate": 3.592592592592593e-05,
|
| 334 |
+
"loss": 1.1437,
|
| 335 |
+
"step": 530
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 1.16,
|
| 339 |
+
"learning_rate": 3.555555555555556e-05,
|
| 340 |
+
"loss": 0.4694,
|
| 341 |
+
"step": 540
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 1.17,
|
| 345 |
+
"learning_rate": 3.518518518518519e-05,
|
| 346 |
+
"loss": 0.6354,
|
| 347 |
+
"step": 550
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 1.17,
|
| 351 |
+
"learning_rate": 3.481481481481482e-05,
|
| 352 |
+
"loss": 0.2423,
|
| 353 |
+
"step": 560
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 1.18,
|
| 357 |
+
"learning_rate": 3.444444444444445e-05,
|
| 358 |
+
"loss": 0.748,
|
| 359 |
+
"step": 570
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 1.19,
|
| 363 |
+
"learning_rate": 3.4074074074074077e-05,
|
| 364 |
+
"loss": 1.6204,
|
| 365 |
+
"step": 580
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 1.19,
|
| 369 |
+
"learning_rate": 3.3703703703703706e-05,
|
| 370 |
+
"loss": 1.1294,
|
| 371 |
+
"step": 590
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 1.2,
|
| 375 |
+
"learning_rate": 3.3333333333333335e-05,
|
| 376 |
+
"loss": 1.7966,
|
| 377 |
+
"step": 600
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 1.2,
|
| 381 |
+
"eval_accuracy": 0.6571428571428571,
|
| 382 |
+
"eval_loss": 1.0164680480957031,
|
| 383 |
+
"eval_runtime": 19.5843,
|
| 384 |
+
"eval_samples_per_second": 3.574,
|
| 385 |
+
"eval_steps_per_second": 3.574,
|
| 386 |
+
"step": 600
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 2.01,
|
| 390 |
+
"learning_rate": 3.2962962962962964e-05,
|
| 391 |
+
"loss": 0.6007,
|
| 392 |
+
"step": 610
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 2.01,
|
| 396 |
+
"learning_rate": 3.25925925925926e-05,
|
| 397 |
+
"loss": 0.158,
|
| 398 |
+
"step": 620
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 2.02,
|
| 402 |
+
"learning_rate": 3.222222222222223e-05,
|
| 403 |
+
"loss": 0.023,
|
| 404 |
+
"step": 630
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 2.03,
|
| 408 |
+
"learning_rate": 3.185185185185185e-05,
|
| 409 |
+
"loss": 0.527,
|
| 410 |
+
"step": 640
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 2.03,
|
| 414 |
+
"learning_rate": 3.148148148148148e-05,
|
| 415 |
+
"loss": 0.3728,
|
| 416 |
+
"step": 650
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 2.04,
|
| 420 |
+
"learning_rate": 3.111111111111111e-05,
|
| 421 |
+
"loss": 0.9129,
|
| 422 |
+
"step": 660
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 2.05,
|
| 426 |
+
"learning_rate": 3.074074074074074e-05,
|
| 427 |
+
"loss": 0.2727,
|
| 428 |
+
"step": 670
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 2.05,
|
| 432 |
+
"learning_rate": 3.037037037037037e-05,
|
| 433 |
+
"loss": 0.3488,
|
| 434 |
+
"step": 680
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 2.06,
|
| 438 |
+
"learning_rate": 3e-05,
|
| 439 |
+
"loss": 0.4387,
|
| 440 |
+
"step": 690
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 2.07,
|
| 444 |
+
"learning_rate": 2.962962962962963e-05,
|
| 445 |
+
"loss": 0.9726,
|
| 446 |
+
"step": 700
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 2.07,
|
| 450 |
+
"learning_rate": 2.925925925925926e-05,
|
| 451 |
+
"loss": 1.0291,
|
| 452 |
+
"step": 710
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 2.08,
|
| 456 |
+
"learning_rate": 2.8888888888888888e-05,
|
| 457 |
+
"loss": 0.3092,
|
| 458 |
+
"step": 720
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 2.09,
|
| 462 |
+
"learning_rate": 2.851851851851852e-05,
|
| 463 |
+
"loss": 1.0766,
|
| 464 |
+
"step": 730
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 2.09,
|
| 468 |
+
"learning_rate": 2.814814814814815e-05,
|
| 469 |
+
"loss": 0.2486,
|
| 470 |
+
"step": 740
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 2.1,
|
| 474 |
+
"learning_rate": 2.777777777777778e-05,
|
| 475 |
+
"loss": 0.7074,
|
| 476 |
+
"step": 750
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 2.11,
|
| 480 |
+
"learning_rate": 2.7407407407407408e-05,
|
| 481 |
+
"loss": 0.1293,
|
| 482 |
+
"step": 760
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 2.11,
|
| 486 |
+
"learning_rate": 2.7037037037037037e-05,
|
| 487 |
+
"loss": 0.2645,
|
| 488 |
+
"step": 770
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 2.12,
|
| 492 |
+
"learning_rate": 2.6666666666666667e-05,
|
| 493 |
+
"loss": 0.3538,
|
| 494 |
+
"step": 780
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 2.13,
|
| 498 |
+
"learning_rate": 2.6296296296296296e-05,
|
| 499 |
+
"loss": 0.3856,
|
| 500 |
+
"step": 790
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 2.13,
|
| 504 |
+
"learning_rate": 2.5925925925925925e-05,
|
| 505 |
+
"loss": 0.0705,
|
| 506 |
+
"step": 800
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 2.14,
|
| 510 |
+
"learning_rate": 2.5555555555555554e-05,
|
| 511 |
+
"loss": 0.5857,
|
| 512 |
+
"step": 810
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 2.15,
|
| 516 |
+
"learning_rate": 2.5185185185185183e-05,
|
| 517 |
+
"loss": 0.0951,
|
| 518 |
+
"step": 820
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 2.15,
|
| 522 |
+
"learning_rate": 2.4814814814814816e-05,
|
| 523 |
+
"loss": 0.3052,
|
| 524 |
+
"step": 830
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 2.16,
|
| 528 |
+
"learning_rate": 2.4444444444444445e-05,
|
| 529 |
+
"loss": 0.2961,
|
| 530 |
+
"step": 840
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 2.17,
|
| 534 |
+
"learning_rate": 2.4074074074074074e-05,
|
| 535 |
+
"loss": 0.2497,
|
| 536 |
+
"step": 850
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 2.17,
|
| 540 |
+
"learning_rate": 2.3703703703703707e-05,
|
| 541 |
+
"loss": 0.6132,
|
| 542 |
+
"step": 860
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 2.18,
|
| 546 |
+
"learning_rate": 2.3333333333333336e-05,
|
| 547 |
+
"loss": 0.2086,
|
| 548 |
+
"step": 870
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 2.19,
|
| 552 |
+
"learning_rate": 2.2962962962962965e-05,
|
| 553 |
+
"loss": 0.0316,
|
| 554 |
+
"step": 880
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 2.19,
|
| 558 |
+
"learning_rate": 2.2592592592592594e-05,
|
| 559 |
+
"loss": 0.0949,
|
| 560 |
+
"step": 890
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 2.2,
|
| 564 |
+
"learning_rate": 2.2222222222222223e-05,
|
| 565 |
+
"loss": 0.4519,
|
| 566 |
+
"step": 900
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 2.2,
|
| 570 |
+
"eval_accuracy": 0.7857142857142857,
|
| 571 |
+
"eval_loss": 0.9175742864608765,
|
| 572 |
+
"eval_runtime": 18.8555,
|
| 573 |
+
"eval_samples_per_second": 3.712,
|
| 574 |
+
"eval_steps_per_second": 3.712,
|
| 575 |
+
"step": 900
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 3.01,
|
| 579 |
+
"learning_rate": 2.1851851851851852e-05,
|
| 580 |
+
"loss": 0.353,
|
| 581 |
+
"step": 910
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 3.01,
|
| 585 |
+
"learning_rate": 2.148148148148148e-05,
|
| 586 |
+
"loss": 0.0182,
|
| 587 |
+
"step": 920
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 3.02,
|
| 591 |
+
"learning_rate": 2.111111111111111e-05,
|
| 592 |
+
"loss": 0.3075,
|
| 593 |
+
"step": 930
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 3.03,
|
| 597 |
+
"learning_rate": 2.074074074074074e-05,
|
| 598 |
+
"loss": 0.0136,
|
| 599 |
+
"step": 940
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 3.03,
|
| 603 |
+
"learning_rate": 2.037037037037037e-05,
|
| 604 |
+
"loss": 0.2276,
|
| 605 |
+
"step": 950
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 3.04,
|
| 609 |
+
"learning_rate": 2e-05,
|
| 610 |
+
"loss": 0.011,
|
| 611 |
+
"step": 960
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 3.05,
|
| 615 |
+
"learning_rate": 1.962962962962963e-05,
|
| 616 |
+
"loss": 0.195,
|
| 617 |
+
"step": 970
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 3.05,
|
| 621 |
+
"learning_rate": 1.925925925925926e-05,
|
| 622 |
+
"loss": 0.0206,
|
| 623 |
+
"step": 980
|
| 624 |
+
},
|
| 625 |
+
{
|
| 626 |
+
"epoch": 3.06,
|
| 627 |
+
"learning_rate": 1.888888888888889e-05,
|
| 628 |
+
"loss": 0.1989,
|
| 629 |
+
"step": 990
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 3.07,
|
| 633 |
+
"learning_rate": 1.8518518518518518e-05,
|
| 634 |
+
"loss": 0.0081,
|
| 635 |
+
"step": 1000
|
| 636 |
+
},
|
| 637 |
+
{
|
| 638 |
+
"epoch": 3.07,
|
| 639 |
+
"learning_rate": 1.814814814814815e-05,
|
| 640 |
+
"loss": 0.6501,
|
| 641 |
+
"step": 1010
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 3.08,
|
| 645 |
+
"learning_rate": 1.777777777777778e-05,
|
| 646 |
+
"loss": 0.0752,
|
| 647 |
+
"step": 1020
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 3.09,
|
| 651 |
+
"learning_rate": 1.740740740740741e-05,
|
| 652 |
+
"loss": 0.8654,
|
| 653 |
+
"step": 1030
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 3.09,
|
| 657 |
+
"learning_rate": 1.7037037037037038e-05,
|
| 658 |
+
"loss": 0.0042,
|
| 659 |
+
"step": 1040
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 3.1,
|
| 663 |
+
"learning_rate": 1.6666666666666667e-05,
|
| 664 |
+
"loss": 0.6607,
|
| 665 |
+
"step": 1050
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"epoch": 3.11,
|
| 669 |
+
"learning_rate": 1.62962962962963e-05,
|
| 670 |
+
"loss": 0.0183,
|
| 671 |
+
"step": 1060
|
| 672 |
+
},
|
| 673 |
+
{
|
| 674 |
+
"epoch": 3.11,
|
| 675 |
+
"learning_rate": 1.5925925925925926e-05,
|
| 676 |
+
"loss": 0.1674,
|
| 677 |
+
"step": 1070
|
| 678 |
+
},
|
| 679 |
+
{
|
| 680 |
+
"epoch": 3.12,
|
| 681 |
+
"learning_rate": 1.5555555555555555e-05,
|
| 682 |
+
"loss": 0.0218,
|
| 683 |
+
"step": 1080
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"epoch": 3.13,
|
| 687 |
+
"learning_rate": 1.5185185185185186e-05,
|
| 688 |
+
"loss": 0.5633,
|
| 689 |
+
"step": 1090
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 3.13,
|
| 693 |
+
"learning_rate": 1.4814814814814815e-05,
|
| 694 |
+
"loss": 0.0141,
|
| 695 |
+
"step": 1100
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 3.14,
|
| 699 |
+
"learning_rate": 1.4444444444444444e-05,
|
| 700 |
+
"loss": 0.6692,
|
| 701 |
+
"step": 1110
|
| 702 |
+
},
|
| 703 |
+
{
|
| 704 |
+
"epoch": 3.15,
|
| 705 |
+
"learning_rate": 1.4074074074074075e-05,
|
| 706 |
+
"loss": 0.4001,
|
| 707 |
+
"step": 1120
|
| 708 |
+
},
|
| 709 |
+
{
|
| 710 |
+
"epoch": 3.15,
|
| 711 |
+
"learning_rate": 1.3703703703703704e-05,
|
| 712 |
+
"loss": 0.0032,
|
| 713 |
+
"step": 1130
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 3.16,
|
| 717 |
+
"learning_rate": 1.3333333333333333e-05,
|
| 718 |
+
"loss": 0.8661,
|
| 719 |
+
"step": 1140
|
| 720 |
+
},
|
| 721 |
+
{
|
| 722 |
+
"epoch": 3.17,
|
| 723 |
+
"learning_rate": 1.2962962962962962e-05,
|
| 724 |
+
"loss": 0.0253,
|
| 725 |
+
"step": 1150
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"epoch": 3.17,
|
| 729 |
+
"learning_rate": 1.2592592592592592e-05,
|
| 730 |
+
"loss": 0.1007,
|
| 731 |
+
"step": 1160
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 3.18,
|
| 735 |
+
"learning_rate": 1.2222222222222222e-05,
|
| 736 |
+
"loss": 0.3541,
|
| 737 |
+
"step": 1170
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 3.19,
|
| 741 |
+
"learning_rate": 1.1851851851851853e-05,
|
| 742 |
+
"loss": 0.5522,
|
| 743 |
+
"step": 1180
|
| 744 |
+
},
|
| 745 |
+
{
|
| 746 |
+
"epoch": 3.19,
|
| 747 |
+
"learning_rate": 1.1481481481481482e-05,
|
| 748 |
+
"loss": 0.5345,
|
| 749 |
+
"step": 1190
|
| 750 |
+
},
|
| 751 |
+
{
|
| 752 |
+
"epoch": 3.2,
|
| 753 |
+
"learning_rate": 1.1111111111111112e-05,
|
| 754 |
+
"loss": 1.8337,
|
| 755 |
+
"step": 1200
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"epoch": 3.2,
|
| 759 |
+
"eval_accuracy": 0.8,
|
| 760 |
+
"eval_loss": 0.9078142642974854,
|
| 761 |
+
"eval_runtime": 20.2891,
|
| 762 |
+
"eval_samples_per_second": 3.45,
|
| 763 |
+
"eval_steps_per_second": 3.45,
|
| 764 |
+
"step": 1200
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 4.01,
|
| 768 |
+
"learning_rate": 1.074074074074074e-05,
|
| 769 |
+
"loss": 0.6171,
|
| 770 |
+
"step": 1210
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 4.01,
|
| 774 |
+
"learning_rate": 1.037037037037037e-05,
|
| 775 |
+
"loss": 0.058,
|
| 776 |
+
"step": 1220
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 4.02,
|
| 780 |
+
"learning_rate": 1e-05,
|
| 781 |
+
"loss": 0.1281,
|
| 782 |
+
"step": 1230
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 4.03,
|
| 786 |
+
"learning_rate": 9.62962962962963e-06,
|
| 787 |
+
"loss": 0.1206,
|
| 788 |
+
"step": 1240
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 4.03,
|
| 792 |
+
"learning_rate": 9.259259259259259e-06,
|
| 793 |
+
"loss": 0.5738,
|
| 794 |
+
"step": 1250
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 4.04,
|
| 798 |
+
"learning_rate": 8.88888888888889e-06,
|
| 799 |
+
"loss": 0.005,
|
| 800 |
+
"step": 1260
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 4.05,
|
| 804 |
+
"learning_rate": 8.518518518518519e-06,
|
| 805 |
+
"loss": 0.573,
|
| 806 |
+
"step": 1270
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 4.05,
|
| 810 |
+
"learning_rate": 8.14814814814815e-06,
|
| 811 |
+
"loss": 0.4713,
|
| 812 |
+
"step": 1280
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 4.06,
|
| 816 |
+
"learning_rate": 7.777777777777777e-06,
|
| 817 |
+
"loss": 0.0032,
|
| 818 |
+
"step": 1290
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 4.07,
|
| 822 |
+
"learning_rate": 7.4074074074074075e-06,
|
| 823 |
+
"loss": 0.0057,
|
| 824 |
+
"step": 1300
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 4.07,
|
| 828 |
+
"learning_rate": 7.0370370370370375e-06,
|
| 829 |
+
"loss": 0.0834,
|
| 830 |
+
"step": 1310
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 4.08,
|
| 834 |
+
"learning_rate": 6.666666666666667e-06,
|
| 835 |
+
"loss": 0.0153,
|
| 836 |
+
"step": 1320
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 4.09,
|
| 840 |
+
"learning_rate": 6.296296296296296e-06,
|
| 841 |
+
"loss": 0.0073,
|
| 842 |
+
"step": 1330
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 4.09,
|
| 846 |
+
"learning_rate": 5.925925925925927e-06,
|
| 847 |
+
"loss": 0.4686,
|
| 848 |
+
"step": 1340
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 4.1,
|
| 852 |
+
"learning_rate": 5.555555555555556e-06,
|
| 853 |
+
"loss": 0.5656,
|
| 854 |
+
"step": 1350
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 4.11,
|
| 858 |
+
"learning_rate": 5.185185185185185e-06,
|
| 859 |
+
"loss": 0.0061,
|
| 860 |
+
"step": 1360
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 4.11,
|
| 864 |
+
"learning_rate": 4.814814814814815e-06,
|
| 865 |
+
"loss": 0.0059,
|
| 866 |
+
"step": 1370
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 4.12,
|
| 870 |
+
"learning_rate": 4.444444444444445e-06,
|
| 871 |
+
"loss": 0.0169,
|
| 872 |
+
"step": 1380
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 4.13,
|
| 876 |
+
"learning_rate": 4.074074074074075e-06,
|
| 877 |
+
"loss": 0.1485,
|
| 878 |
+
"step": 1390
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 4.13,
|
| 882 |
+
"learning_rate": 3.7037037037037037e-06,
|
| 883 |
+
"loss": 0.0036,
|
| 884 |
+
"step": 1400
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 4.14,
|
| 888 |
+
"learning_rate": 3.3333333333333333e-06,
|
| 889 |
+
"loss": 0.2698,
|
| 890 |
+
"step": 1410
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 4.15,
|
| 894 |
+
"learning_rate": 2.9629629629629633e-06,
|
| 895 |
+
"loss": 0.012,
|
| 896 |
+
"step": 1420
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 4.15,
|
| 900 |
+
"learning_rate": 2.5925925925925925e-06,
|
| 901 |
+
"loss": 0.0598,
|
| 902 |
+
"step": 1430
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 4.16,
|
| 906 |
+
"learning_rate": 2.2222222222222225e-06,
|
| 907 |
+
"loss": 0.0041,
|
| 908 |
+
"step": 1440
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 4.17,
|
| 912 |
+
"learning_rate": 1.8518518518518519e-06,
|
| 913 |
+
"loss": 0.0035,
|
| 914 |
+
"step": 1450
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 4.17,
|
| 918 |
+
"learning_rate": 1.4814814814814817e-06,
|
| 919 |
+
"loss": 0.0048,
|
| 920 |
+
"step": 1460
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 4.18,
|
| 924 |
+
"learning_rate": 1.1111111111111112e-06,
|
| 925 |
+
"loss": 0.022,
|
| 926 |
+
"step": 1470
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 4.19,
|
| 930 |
+
"learning_rate": 7.407407407407408e-07,
|
| 931 |
+
"loss": 0.4452,
|
| 932 |
+
"step": 1480
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 4.19,
|
| 936 |
+
"learning_rate": 3.703703703703704e-07,
|
| 937 |
+
"loss": 0.0023,
|
| 938 |
+
"step": 1490
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 4.2,
|
| 942 |
+
"learning_rate": 0.0,
|
| 943 |
+
"loss": 0.006,
|
| 944 |
+
"step": 1500
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 4.2,
|
| 948 |
+
"eval_accuracy": 0.9,
|
| 949 |
+
"eval_loss": 0.38240665197372437,
|
| 950 |
+
"eval_runtime": 17.0083,
|
| 951 |
+
"eval_samples_per_second": 4.116,
|
| 952 |
+
"eval_steps_per_second": 4.116,
|
| 953 |
+
"step": 1500
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"epoch": 4.2,
|
| 957 |
+
"step": 1500,
|
| 958 |
+
"total_flos": 1.86923023515648e+18,
|
| 959 |
+
"train_loss": 0.8117700910270214,
|
| 960 |
+
"train_runtime": 1160.0655,
|
| 961 |
+
"train_samples_per_second": 1.293,
|
| 962 |
+
"train_steps_per_second": 1.293
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 4.2,
|
| 966 |
+
"eval_accuracy": 0.9161290322580645,
|
| 967 |
+
"eval_loss": 0.2619660198688507,
|
| 968 |
+
"eval_runtime": 42.6134,
|
| 969 |
+
"eval_samples_per_second": 3.637,
|
| 970 |
+
"eval_steps_per_second": 3.637,
|
| 971 |
+
"step": 1500
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"epoch": 4.2,
|
| 975 |
+
"eval_accuracy": 0.9161290322580645,
|
| 976 |
+
"eval_loss": 0.2619659900665283,
|
| 977 |
+
"eval_runtime": 43.4165,
|
| 978 |
+
"eval_samples_per_second": 3.57,
|
| 979 |
+
"eval_steps_per_second": 3.57,
|
| 980 |
+
"step": 1500
|
| 981 |
+
}
|
| 982 |
+
],
|
| 983 |
+
"max_steps": 1500,
|
| 984 |
+
"num_train_epochs": 9223372036854775807,
|
| 985 |
+
"total_flos": 1.86923023515648e+18,
|
| 986 |
+
"trial_name": null,
|
| 987 |
+
"trial_params": null
|
| 988 |
+
}
|