End of training
Browse files- all_results.json +12 -0
- eval_results.json +8 -0
- runs/Jul05_13-49-06_u20/events.out.tfevents.1688539432.u20.2776403.1 +3 -0
- train_results.json +7 -0
- trainer_state.json +844 -0
all_results.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"eval_accuracy": 0.8195488721804511,
|
| 4 |
+
"eval_loss": 0.9759521484375,
|
| 5 |
+
"eval_runtime": 59.1239,
|
| 6 |
+
"eval_samples_per_second": 2.25,
|
| 7 |
+
"eval_steps_per_second": 0.575,
|
| 8 |
+
"train_loss": 1.017018985011863,
|
| 9 |
+
"train_runtime": 3198.9902,
|
| 10 |
+
"train_samples_per_second": 1.616,
|
| 11 |
+
"train_steps_per_second": 0.405
|
| 12 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"eval_accuracy": 0.8195488721804511,
|
| 4 |
+
"eval_loss": 0.9759521484375,
|
| 5 |
+
"eval_runtime": 59.1239,
|
| 6 |
+
"eval_samples_per_second": 2.25,
|
| 7 |
+
"eval_steps_per_second": 0.575
|
| 8 |
+
}
|
runs/Jul05_13-49-06_u20/events.out.tfevents.1688539432.u20.2776403.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8efbb53e4c9f7ef2f7c9dc8b8459fad11962e1f7c08e624a3bdd3fedd526b572
|
| 3 |
+
size 363
|
train_results.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"train_loss": 1.017018985011863,
|
| 4 |
+
"train_runtime": 3198.9902,
|
| 5 |
+
"train_samples_per_second": 1.616,
|
| 6 |
+
"train_steps_per_second": 0.405
|
| 7 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,844 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.9759521484375,
|
| 3 |
+
"best_model_checkpoint": "./beans_outputs_1/checkpoint-1295",
|
| 4 |
+
"epoch": 5.0,
|
| 5 |
+
"global_step": 1295,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.04,
|
| 12 |
+
"learning_rate": 1.9845559845559848e-05,
|
| 13 |
+
"loss": 1.106,
|
| 14 |
+
"step": 10
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.08,
|
| 18 |
+
"learning_rate": 1.9691119691119695e-05,
|
| 19 |
+
"loss": 1.1028,
|
| 20 |
+
"step": 20
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.12,
|
| 24 |
+
"learning_rate": 1.9536679536679538e-05,
|
| 25 |
+
"loss": 1.0825,
|
| 26 |
+
"step": 30
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.15,
|
| 30 |
+
"learning_rate": 1.938223938223938e-05,
|
| 31 |
+
"loss": 1.1019,
|
| 32 |
+
"step": 40
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.19,
|
| 36 |
+
"learning_rate": 1.9227799227799228e-05,
|
| 37 |
+
"loss": 1.0847,
|
| 38 |
+
"step": 50
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.23,
|
| 42 |
+
"learning_rate": 1.9073359073359075e-05,
|
| 43 |
+
"loss": 1.0689,
|
| 44 |
+
"step": 60
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.27,
|
| 48 |
+
"learning_rate": 1.891891891891892e-05,
|
| 49 |
+
"loss": 1.0911,
|
| 50 |
+
"step": 70
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.31,
|
| 54 |
+
"learning_rate": 1.8764478764478768e-05,
|
| 55 |
+
"loss": 1.0968,
|
| 56 |
+
"step": 80
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.35,
|
| 60 |
+
"learning_rate": 1.861003861003861e-05,
|
| 61 |
+
"loss": 1.0731,
|
| 62 |
+
"step": 90
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.39,
|
| 66 |
+
"learning_rate": 1.8455598455598458e-05,
|
| 67 |
+
"loss": 1.0753,
|
| 68 |
+
"step": 100
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.42,
|
| 72 |
+
"learning_rate": 1.83011583011583e-05,
|
| 73 |
+
"loss": 1.0795,
|
| 74 |
+
"step": 110
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.46,
|
| 78 |
+
"learning_rate": 1.8146718146718147e-05,
|
| 79 |
+
"loss": 1.0757,
|
| 80 |
+
"step": 120
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.5,
|
| 84 |
+
"learning_rate": 1.7992277992277994e-05,
|
| 85 |
+
"loss": 1.0769,
|
| 86 |
+
"step": 130
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.54,
|
| 90 |
+
"learning_rate": 1.783783783783784e-05,
|
| 91 |
+
"loss": 1.0742,
|
| 92 |
+
"step": 140
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.58,
|
| 96 |
+
"learning_rate": 1.7683397683397684e-05,
|
| 97 |
+
"loss": 1.0649,
|
| 98 |
+
"step": 150
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.62,
|
| 102 |
+
"learning_rate": 1.752895752895753e-05,
|
| 103 |
+
"loss": 1.0669,
|
| 104 |
+
"step": 160
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.66,
|
| 108 |
+
"learning_rate": 1.7374517374517377e-05,
|
| 109 |
+
"loss": 1.0599,
|
| 110 |
+
"step": 170
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.69,
|
| 114 |
+
"learning_rate": 1.722007722007722e-05,
|
| 115 |
+
"loss": 1.057,
|
| 116 |
+
"step": 180
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.73,
|
| 120 |
+
"learning_rate": 1.7065637065637067e-05,
|
| 121 |
+
"loss": 1.0733,
|
| 122 |
+
"step": 190
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.77,
|
| 126 |
+
"learning_rate": 1.6911196911196913e-05,
|
| 127 |
+
"loss": 1.0602,
|
| 128 |
+
"step": 200
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.81,
|
| 132 |
+
"learning_rate": 1.6756756756756757e-05,
|
| 133 |
+
"loss": 1.0651,
|
| 134 |
+
"step": 210
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.85,
|
| 138 |
+
"learning_rate": 1.6602316602316603e-05,
|
| 139 |
+
"loss": 1.0482,
|
| 140 |
+
"step": 220
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.89,
|
| 144 |
+
"learning_rate": 1.644787644787645e-05,
|
| 145 |
+
"loss": 1.0514,
|
| 146 |
+
"step": 230
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.93,
|
| 150 |
+
"learning_rate": 1.6293436293436296e-05,
|
| 151 |
+
"loss": 1.0653,
|
| 152 |
+
"step": 240
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.97,
|
| 156 |
+
"learning_rate": 1.613899613899614e-05,
|
| 157 |
+
"loss": 1.0596,
|
| 158 |
+
"step": 250
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 1.0,
|
| 162 |
+
"eval_accuracy": 0.7142857142857143,
|
| 163 |
+
"eval_loss": 1.0507243871688843,
|
| 164 |
+
"eval_runtime": 60.3908,
|
| 165 |
+
"eval_samples_per_second": 2.202,
|
| 166 |
+
"eval_steps_per_second": 0.563,
|
| 167 |
+
"step": 259
|
| 168 |
+
},
|
| 169 |
+
{
|
| 170 |
+
"epoch": 1.0,
|
| 171 |
+
"learning_rate": 1.5984555984555986e-05,
|
| 172 |
+
"loss": 1.0574,
|
| 173 |
+
"step": 260
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 1.04,
|
| 177 |
+
"learning_rate": 1.583011583011583e-05,
|
| 178 |
+
"loss": 1.0497,
|
| 179 |
+
"step": 270
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 1.08,
|
| 183 |
+
"learning_rate": 1.5675675675675676e-05,
|
| 184 |
+
"loss": 1.0488,
|
| 185 |
+
"step": 280
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 1.12,
|
| 189 |
+
"learning_rate": 1.5521235521235523e-05,
|
| 190 |
+
"loss": 1.0416,
|
| 191 |
+
"step": 290
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 1.16,
|
| 195 |
+
"learning_rate": 1.536679536679537e-05,
|
| 196 |
+
"loss": 1.0478,
|
| 197 |
+
"step": 300
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 1.2,
|
| 201 |
+
"learning_rate": 1.5212355212355214e-05,
|
| 202 |
+
"loss": 1.0349,
|
| 203 |
+
"step": 310
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 1.24,
|
| 207 |
+
"learning_rate": 1.5057915057915061e-05,
|
| 208 |
+
"loss": 1.0518,
|
| 209 |
+
"step": 320
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 1.27,
|
| 213 |
+
"learning_rate": 1.4903474903474904e-05,
|
| 214 |
+
"loss": 1.0354,
|
| 215 |
+
"step": 330
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 1.31,
|
| 219 |
+
"learning_rate": 1.4749034749034749e-05,
|
| 220 |
+
"loss": 1.0478,
|
| 221 |
+
"step": 340
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 1.35,
|
| 225 |
+
"learning_rate": 1.4594594594594596e-05,
|
| 226 |
+
"loss": 1.064,
|
| 227 |
+
"step": 350
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 1.39,
|
| 231 |
+
"learning_rate": 1.444015444015444e-05,
|
| 232 |
+
"loss": 1.0362,
|
| 233 |
+
"step": 360
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 1.43,
|
| 237 |
+
"learning_rate": 1.4285714285714287e-05,
|
| 238 |
+
"loss": 1.0351,
|
| 239 |
+
"step": 370
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 1.47,
|
| 243 |
+
"learning_rate": 1.4131274131274134e-05,
|
| 244 |
+
"loss": 1.0423,
|
| 245 |
+
"step": 380
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 1.51,
|
| 249 |
+
"learning_rate": 1.3976833976833979e-05,
|
| 250 |
+
"loss": 1.0345,
|
| 251 |
+
"step": 390
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 1.54,
|
| 255 |
+
"learning_rate": 1.3822393822393822e-05,
|
| 256 |
+
"loss": 1.0271,
|
| 257 |
+
"step": 400
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 1.58,
|
| 261 |
+
"learning_rate": 1.3667953667953668e-05,
|
| 262 |
+
"loss": 1.0337,
|
| 263 |
+
"step": 410
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 1.62,
|
| 267 |
+
"learning_rate": 1.3513513513513515e-05,
|
| 268 |
+
"loss": 1.0158,
|
| 269 |
+
"step": 420
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 1.66,
|
| 273 |
+
"learning_rate": 1.335907335907336e-05,
|
| 274 |
+
"loss": 1.0421,
|
| 275 |
+
"step": 430
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 1.7,
|
| 279 |
+
"learning_rate": 1.3204633204633207e-05,
|
| 280 |
+
"loss": 1.0309,
|
| 281 |
+
"step": 440
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 1.74,
|
| 285 |
+
"learning_rate": 1.3050193050193052e-05,
|
| 286 |
+
"loss": 1.0097,
|
| 287 |
+
"step": 450
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 1.78,
|
| 291 |
+
"learning_rate": 1.2895752895752898e-05,
|
| 292 |
+
"loss": 1.0157,
|
| 293 |
+
"step": 460
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 1.81,
|
| 297 |
+
"learning_rate": 1.2741312741312741e-05,
|
| 298 |
+
"loss": 1.0195,
|
| 299 |
+
"step": 470
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 1.85,
|
| 303 |
+
"learning_rate": 1.2586872586872588e-05,
|
| 304 |
+
"loss": 1.0279,
|
| 305 |
+
"step": 480
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 1.89,
|
| 309 |
+
"learning_rate": 1.2432432432432433e-05,
|
| 310 |
+
"loss": 1.0187,
|
| 311 |
+
"step": 490
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 1.93,
|
| 315 |
+
"learning_rate": 1.227799227799228e-05,
|
| 316 |
+
"loss": 1.015,
|
| 317 |
+
"step": 500
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 1.97,
|
| 321 |
+
"learning_rate": 1.2123552123552124e-05,
|
| 322 |
+
"loss": 1.0165,
|
| 323 |
+
"step": 510
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 2.0,
|
| 327 |
+
"eval_accuracy": 0.7894736842105263,
|
| 328 |
+
"eval_loss": 1.0165094137191772,
|
| 329 |
+
"eval_runtime": 57.5347,
|
| 330 |
+
"eval_samples_per_second": 2.312,
|
| 331 |
+
"eval_steps_per_second": 0.591,
|
| 332 |
+
"step": 518
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 2.01,
|
| 336 |
+
"learning_rate": 1.1969111969111971e-05,
|
| 337 |
+
"loss": 1.0331,
|
| 338 |
+
"step": 520
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 2.05,
|
| 342 |
+
"learning_rate": 1.1814671814671816e-05,
|
| 343 |
+
"loss": 1.0241,
|
| 344 |
+
"step": 530
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 2.08,
|
| 348 |
+
"learning_rate": 1.166023166023166e-05,
|
| 349 |
+
"loss": 1.0185,
|
| 350 |
+
"step": 540
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 2.12,
|
| 354 |
+
"learning_rate": 1.1505791505791506e-05,
|
| 355 |
+
"loss": 1.0257,
|
| 356 |
+
"step": 550
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 2.16,
|
| 360 |
+
"learning_rate": 1.1351351351351352e-05,
|
| 361 |
+
"loss": 1.0213,
|
| 362 |
+
"step": 560
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 2.2,
|
| 366 |
+
"learning_rate": 1.1196911196911197e-05,
|
| 367 |
+
"loss": 1.0278,
|
| 368 |
+
"step": 570
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 2.24,
|
| 372 |
+
"learning_rate": 1.1042471042471044e-05,
|
| 373 |
+
"loss": 1.0307,
|
| 374 |
+
"step": 580
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 2.28,
|
| 378 |
+
"learning_rate": 1.0888030888030889e-05,
|
| 379 |
+
"loss": 1.0054,
|
| 380 |
+
"step": 590
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 2.32,
|
| 384 |
+
"learning_rate": 1.0733590733590735e-05,
|
| 385 |
+
"loss": 1.0094,
|
| 386 |
+
"step": 600
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 2.36,
|
| 390 |
+
"learning_rate": 1.057915057915058e-05,
|
| 391 |
+
"loss": 0.9994,
|
| 392 |
+
"step": 610
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 2.39,
|
| 396 |
+
"learning_rate": 1.0424710424710425e-05,
|
| 397 |
+
"loss": 1.0235,
|
| 398 |
+
"step": 620
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 2.43,
|
| 402 |
+
"learning_rate": 1.027027027027027e-05,
|
| 403 |
+
"loss": 1.0063,
|
| 404 |
+
"step": 630
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 2.47,
|
| 408 |
+
"learning_rate": 1.0115830115830117e-05,
|
| 409 |
+
"loss": 0.9763,
|
| 410 |
+
"step": 640
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 2.51,
|
| 414 |
+
"learning_rate": 9.961389961389962e-06,
|
| 415 |
+
"loss": 1.0209,
|
| 416 |
+
"step": 650
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 2.55,
|
| 420 |
+
"learning_rate": 9.806949806949808e-06,
|
| 421 |
+
"loss": 0.998,
|
| 422 |
+
"step": 660
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 2.59,
|
| 426 |
+
"learning_rate": 9.652509652509653e-06,
|
| 427 |
+
"loss": 1.0063,
|
| 428 |
+
"step": 670
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 2.63,
|
| 432 |
+
"learning_rate": 9.498069498069498e-06,
|
| 433 |
+
"loss": 1.0366,
|
| 434 |
+
"step": 680
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 2.66,
|
| 438 |
+
"learning_rate": 9.343629343629345e-06,
|
| 439 |
+
"loss": 1.0004,
|
| 440 |
+
"step": 690
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 2.7,
|
| 444 |
+
"learning_rate": 9.189189189189191e-06,
|
| 445 |
+
"loss": 1.0212,
|
| 446 |
+
"step": 700
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 2.74,
|
| 450 |
+
"learning_rate": 9.034749034749034e-06,
|
| 451 |
+
"loss": 1.002,
|
| 452 |
+
"step": 710
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 2.78,
|
| 456 |
+
"learning_rate": 8.880308880308881e-06,
|
| 457 |
+
"loss": 0.9824,
|
| 458 |
+
"step": 720
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 2.82,
|
| 462 |
+
"learning_rate": 8.725868725868728e-06,
|
| 463 |
+
"loss": 1.0043,
|
| 464 |
+
"step": 730
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 2.86,
|
| 468 |
+
"learning_rate": 8.571428571428571e-06,
|
| 469 |
+
"loss": 0.9829,
|
| 470 |
+
"step": 740
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 2.9,
|
| 474 |
+
"learning_rate": 8.416988416988418e-06,
|
| 475 |
+
"loss": 0.9977,
|
| 476 |
+
"step": 750
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 2.93,
|
| 480 |
+
"learning_rate": 8.262548262548264e-06,
|
| 481 |
+
"loss": 0.9996,
|
| 482 |
+
"step": 760
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 2.97,
|
| 486 |
+
"learning_rate": 8.108108108108109e-06,
|
| 487 |
+
"loss": 1.0113,
|
| 488 |
+
"step": 770
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 3.0,
|
| 492 |
+
"eval_accuracy": 0.8045112781954887,
|
| 493 |
+
"eval_loss": 0.9941034317016602,
|
| 494 |
+
"eval_runtime": 57.8525,
|
| 495 |
+
"eval_samples_per_second": 2.299,
|
| 496 |
+
"eval_steps_per_second": 0.588,
|
| 497 |
+
"step": 777
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 3.01,
|
| 501 |
+
"learning_rate": 7.953667953667954e-06,
|
| 502 |
+
"loss": 0.9685,
|
| 503 |
+
"step": 780
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 3.05,
|
| 507 |
+
"learning_rate": 7.7992277992278e-06,
|
| 508 |
+
"loss": 0.998,
|
| 509 |
+
"step": 790
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 3.09,
|
| 513 |
+
"learning_rate": 7.644787644787645e-06,
|
| 514 |
+
"loss": 0.9931,
|
| 515 |
+
"step": 800
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 3.13,
|
| 519 |
+
"learning_rate": 7.49034749034749e-06,
|
| 520 |
+
"loss": 0.9845,
|
| 521 |
+
"step": 810
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 3.17,
|
| 525 |
+
"learning_rate": 7.335907335907336e-06,
|
| 526 |
+
"loss": 0.9952,
|
| 527 |
+
"step": 820
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 3.2,
|
| 531 |
+
"learning_rate": 7.181467181467182e-06,
|
| 532 |
+
"loss": 0.9726,
|
| 533 |
+
"step": 830
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 3.24,
|
| 537 |
+
"learning_rate": 7.027027027027028e-06,
|
| 538 |
+
"loss": 0.9986,
|
| 539 |
+
"step": 840
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 3.28,
|
| 543 |
+
"learning_rate": 6.872586872586873e-06,
|
| 544 |
+
"loss": 0.9956,
|
| 545 |
+
"step": 850
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 3.32,
|
| 549 |
+
"learning_rate": 6.718146718146718e-06,
|
| 550 |
+
"loss": 1.0021,
|
| 551 |
+
"step": 860
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 3.36,
|
| 555 |
+
"learning_rate": 6.563706563706564e-06,
|
| 556 |
+
"loss": 0.9785,
|
| 557 |
+
"step": 870
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 3.4,
|
| 561 |
+
"learning_rate": 6.409266409266411e-06,
|
| 562 |
+
"loss": 1.0007,
|
| 563 |
+
"step": 880
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 3.44,
|
| 567 |
+
"learning_rate": 6.254826254826255e-06,
|
| 568 |
+
"loss": 0.9712,
|
| 569 |
+
"step": 890
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 3.47,
|
| 573 |
+
"learning_rate": 6.1003861003861005e-06,
|
| 574 |
+
"loss": 0.988,
|
| 575 |
+
"step": 900
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 3.51,
|
| 579 |
+
"learning_rate": 5.945945945945947e-06,
|
| 580 |
+
"loss": 0.9954,
|
| 581 |
+
"step": 910
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 3.55,
|
| 585 |
+
"learning_rate": 5.791505791505791e-06,
|
| 586 |
+
"loss": 0.9693,
|
| 587 |
+
"step": 920
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 3.59,
|
| 591 |
+
"learning_rate": 5.637065637065637e-06,
|
| 592 |
+
"loss": 1.0062,
|
| 593 |
+
"step": 930
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 3.63,
|
| 597 |
+
"learning_rate": 5.4826254826254836e-06,
|
| 598 |
+
"loss": 0.9875,
|
| 599 |
+
"step": 940
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 3.67,
|
| 603 |
+
"learning_rate": 5.328185328185329e-06,
|
| 604 |
+
"loss": 1.0029,
|
| 605 |
+
"step": 950
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 3.71,
|
| 609 |
+
"learning_rate": 5.173745173745173e-06,
|
| 610 |
+
"loss": 0.9766,
|
| 611 |
+
"step": 960
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 3.75,
|
| 615 |
+
"learning_rate": 5.01930501930502e-06,
|
| 616 |
+
"loss": 0.9875,
|
| 617 |
+
"step": 970
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 3.78,
|
| 621 |
+
"learning_rate": 4.864864864864866e-06,
|
| 622 |
+
"loss": 0.9631,
|
| 623 |
+
"step": 980
|
| 624 |
+
},
|
| 625 |
+
{
|
| 626 |
+
"epoch": 3.82,
|
| 627 |
+
"learning_rate": 4.710424710424711e-06,
|
| 628 |
+
"loss": 0.9838,
|
| 629 |
+
"step": 990
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 3.86,
|
| 633 |
+
"learning_rate": 4.5559845559845564e-06,
|
| 634 |
+
"loss": 0.9694,
|
| 635 |
+
"step": 1000
|
| 636 |
+
},
|
| 637 |
+
{
|
| 638 |
+
"epoch": 3.9,
|
| 639 |
+
"learning_rate": 4.401544401544402e-06,
|
| 640 |
+
"loss": 1.0125,
|
| 641 |
+
"step": 1010
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 3.94,
|
| 645 |
+
"learning_rate": 4.247104247104247e-06,
|
| 646 |
+
"loss": 0.9848,
|
| 647 |
+
"step": 1020
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 3.98,
|
| 651 |
+
"learning_rate": 4.092664092664093e-06,
|
| 652 |
+
"loss": 1.0067,
|
| 653 |
+
"step": 1030
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 4.0,
|
| 657 |
+
"eval_accuracy": 0.8195488721804511,
|
| 658 |
+
"eval_loss": 0.9804362654685974,
|
| 659 |
+
"eval_runtime": 57.6615,
|
| 660 |
+
"eval_samples_per_second": 2.307,
|
| 661 |
+
"eval_steps_per_second": 0.59,
|
| 662 |
+
"step": 1036
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 4.02,
|
| 666 |
+
"learning_rate": 3.938223938223939e-06,
|
| 667 |
+
"loss": 0.9862,
|
| 668 |
+
"step": 1040
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 4.05,
|
| 672 |
+
"learning_rate": 3.7837837837837844e-06,
|
| 673 |
+
"loss": 0.968,
|
| 674 |
+
"step": 1050
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 4.09,
|
| 678 |
+
"learning_rate": 3.6293436293436297e-06,
|
| 679 |
+
"loss": 0.9788,
|
| 680 |
+
"step": 1060
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 4.13,
|
| 684 |
+
"learning_rate": 3.4749034749034755e-06,
|
| 685 |
+
"loss": 0.9988,
|
| 686 |
+
"step": 1070
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 4.17,
|
| 690 |
+
"learning_rate": 3.320463320463321e-06,
|
| 691 |
+
"loss": 0.9927,
|
| 692 |
+
"step": 1080
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 4.21,
|
| 696 |
+
"learning_rate": 3.166023166023166e-06,
|
| 697 |
+
"loss": 0.9883,
|
| 698 |
+
"step": 1090
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 4.25,
|
| 702 |
+
"learning_rate": 3.011583011583012e-06,
|
| 703 |
+
"loss": 0.9859,
|
| 704 |
+
"step": 1100
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 4.29,
|
| 708 |
+
"learning_rate": 2.8571428571428573e-06,
|
| 709 |
+
"loss": 0.9918,
|
| 710 |
+
"step": 1110
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 4.32,
|
| 714 |
+
"learning_rate": 2.702702702702703e-06,
|
| 715 |
+
"loss": 0.9805,
|
| 716 |
+
"step": 1120
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 4.36,
|
| 720 |
+
"learning_rate": 2.5482625482625484e-06,
|
| 721 |
+
"loss": 0.9869,
|
| 722 |
+
"step": 1130
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 4.4,
|
| 726 |
+
"learning_rate": 2.393822393822394e-06,
|
| 727 |
+
"loss": 1.0008,
|
| 728 |
+
"step": 1140
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 4.44,
|
| 732 |
+
"learning_rate": 2.2393822393822394e-06,
|
| 733 |
+
"loss": 0.9577,
|
| 734 |
+
"step": 1150
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 4.48,
|
| 738 |
+
"learning_rate": 2.084942084942085e-06,
|
| 739 |
+
"loss": 0.9531,
|
| 740 |
+
"step": 1160
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 4.52,
|
| 744 |
+
"learning_rate": 1.9305019305019305e-06,
|
| 745 |
+
"loss": 0.9739,
|
| 746 |
+
"step": 1170
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 4.56,
|
| 750 |
+
"learning_rate": 1.7760617760617763e-06,
|
| 751 |
+
"loss": 0.9914,
|
| 752 |
+
"step": 1180
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 4.59,
|
| 756 |
+
"learning_rate": 1.6216216216216219e-06,
|
| 757 |
+
"loss": 0.9602,
|
| 758 |
+
"step": 1190
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 4.63,
|
| 762 |
+
"learning_rate": 1.4671814671814674e-06,
|
| 763 |
+
"loss": 0.9821,
|
| 764 |
+
"step": 1200
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 4.67,
|
| 768 |
+
"learning_rate": 1.3127413127413127e-06,
|
| 769 |
+
"loss": 0.9576,
|
| 770 |
+
"step": 1210
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 4.71,
|
| 774 |
+
"learning_rate": 1.1583011583011585e-06,
|
| 775 |
+
"loss": 0.9677,
|
| 776 |
+
"step": 1220
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 4.75,
|
| 780 |
+
"learning_rate": 1.0038610038610038e-06,
|
| 781 |
+
"loss": 0.9838,
|
| 782 |
+
"step": 1230
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 4.79,
|
| 786 |
+
"learning_rate": 8.494208494208495e-07,
|
| 787 |
+
"loss": 0.9762,
|
| 788 |
+
"step": 1240
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 4.83,
|
| 792 |
+
"learning_rate": 6.94980694980695e-07,
|
| 793 |
+
"loss": 0.9833,
|
| 794 |
+
"step": 1250
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 4.86,
|
| 798 |
+
"learning_rate": 5.405405405405406e-07,
|
| 799 |
+
"loss": 0.9839,
|
| 800 |
+
"step": 1260
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 4.9,
|
| 804 |
+
"learning_rate": 3.8610038610038613e-07,
|
| 805 |
+
"loss": 0.9802,
|
| 806 |
+
"step": 1270
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 4.94,
|
| 810 |
+
"learning_rate": 2.3166023166023168e-07,
|
| 811 |
+
"loss": 0.991,
|
| 812 |
+
"step": 1280
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 4.98,
|
| 816 |
+
"learning_rate": 7.722007722007723e-08,
|
| 817 |
+
"loss": 0.9746,
|
| 818 |
+
"step": 1290
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 5.0,
|
| 822 |
+
"eval_accuracy": 0.8195488721804511,
|
| 823 |
+
"eval_loss": 0.9759521484375,
|
| 824 |
+
"eval_runtime": 60.3048,
|
| 825 |
+
"eval_samples_per_second": 2.205,
|
| 826 |
+
"eval_steps_per_second": 0.564,
|
| 827 |
+
"step": 1295
|
| 828 |
+
},
|
| 829 |
+
{
|
| 830 |
+
"epoch": 5.0,
|
| 831 |
+
"step": 1295,
|
| 832 |
+
"total_flos": 2.945297577944494e+18,
|
| 833 |
+
"train_loss": 1.017018985011863,
|
| 834 |
+
"train_runtime": 3198.9902,
|
| 835 |
+
"train_samples_per_second": 1.616,
|
| 836 |
+
"train_steps_per_second": 0.405
|
| 837 |
+
}
|
| 838 |
+
],
|
| 839 |
+
"max_steps": 1295,
|
| 840 |
+
"num_train_epochs": 5,
|
| 841 |
+
"total_flos": 2.945297577944494e+18,
|
| 842 |
+
"trial_name": null,
|
| 843 |
+
"trial_params": null
|
| 844 |
+
}
|