Commit
•
a51fae6
1
Parent(s):
164f923
End of training
Browse files
all_results.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"eval_accuracy": 0.9806603773584905,
|
| 4 |
+
"eval_loss": 0.06080754101276398,
|
| 5 |
+
"eval_runtime": 47.1994,
|
| 6 |
+
"eval_samples_per_second": 134.748,
|
| 7 |
+
"eval_steps_per_second": 4.216,
|
| 8 |
+
"total_flos": 4.2672264771137864e+18,
|
| 9 |
+
"train_loss": 0.07098129998106373,
|
| 10 |
+
"train_runtime": 22089.0495,
|
| 11 |
+
"train_samples_per_second": 7.773,
|
| 12 |
+
"train_steps_per_second": 0.061
|
| 13 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"eval_accuracy": 0.9806603773584905,
|
| 4 |
+
"eval_loss": 0.06080754101276398,
|
| 5 |
+
"eval_runtime": 47.1994,
|
| 6 |
+
"eval_samples_per_second": 134.748,
|
| 7 |
+
"eval_steps_per_second": 4.216
|
| 8 |
+
}
|
runs/Jun18_18-01-04_d696d3c41904/events.out.tfevents.1655597596.d696d3c41904.81.10
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1df7a54218819ab51d09ae84fb88b11492bbfa4ce6999865506dfcd64caa1160
|
| 3 |
+
size 363
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"total_flos": 4.2672264771137864e+18,
|
| 4 |
+
"train_loss": 0.07098129998106373,
|
| 5 |
+
"train_runtime": 22089.0495,
|
| 6 |
+
"train_samples_per_second": 7.773,
|
| 7 |
+
"train_steps_per_second": 0.061
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,856 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.9806603773584905,
|
| 3 |
+
"best_model_checkpoint": "swin-tiny-patch4-window7-224-finetuned-mri/checkpoint-1341",
|
| 4 |
+
"epoch": 2.999441028507546,
|
| 5 |
+
"global_step": 1341,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.02,
|
| 12 |
+
"learning_rate": 3.7037037037037037e-06,
|
| 13 |
+
"loss": 0.4997,
|
| 14 |
+
"step": 10
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.04,
|
| 18 |
+
"learning_rate": 7.4074074074074075e-06,
|
| 19 |
+
"loss": 0.4757,
|
| 20 |
+
"step": 20
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.07,
|
| 24 |
+
"learning_rate": 1.1111111111111112e-05,
|
| 25 |
+
"loss": 0.4668,
|
| 26 |
+
"step": 30
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.09,
|
| 30 |
+
"learning_rate": 1.4814814814814815e-05,
|
| 31 |
+
"loss": 0.4247,
|
| 32 |
+
"step": 40
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.11,
|
| 36 |
+
"learning_rate": 1.8518518518518518e-05,
|
| 37 |
+
"loss": 0.3901,
|
| 38 |
+
"step": 50
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.13,
|
| 42 |
+
"learning_rate": 2.2222222222222223e-05,
|
| 43 |
+
"loss": 0.3659,
|
| 44 |
+
"step": 60
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.16,
|
| 48 |
+
"learning_rate": 2.5925925925925925e-05,
|
| 49 |
+
"loss": 0.3264,
|
| 50 |
+
"step": 70
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.18,
|
| 54 |
+
"learning_rate": 2.962962962962963e-05,
|
| 55 |
+
"loss": 0.2907,
|
| 56 |
+
"step": 80
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.2,
|
| 60 |
+
"learning_rate": 3.3333333333333335e-05,
|
| 61 |
+
"loss": 0.2786,
|
| 62 |
+
"step": 90
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.22,
|
| 66 |
+
"learning_rate": 3.7037037037037037e-05,
|
| 67 |
+
"loss": 0.2198,
|
| 68 |
+
"step": 100
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.25,
|
| 72 |
+
"learning_rate": 4.074074074074074e-05,
|
| 73 |
+
"loss": 0.2128,
|
| 74 |
+
"step": 110
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.27,
|
| 78 |
+
"learning_rate": 4.4444444444444447e-05,
|
| 79 |
+
"loss": 0.1964,
|
| 80 |
+
"step": 120
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.29,
|
| 84 |
+
"learning_rate": 4.814814814814815e-05,
|
| 85 |
+
"loss": 0.1769,
|
| 86 |
+
"step": 130
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.31,
|
| 90 |
+
"learning_rate": 4.979270315091211e-05,
|
| 91 |
+
"loss": 0.1848,
|
| 92 |
+
"step": 140
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.34,
|
| 96 |
+
"learning_rate": 4.9378109452736324e-05,
|
| 97 |
+
"loss": 0.1667,
|
| 98 |
+
"step": 150
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.36,
|
| 102 |
+
"learning_rate": 4.896351575456054e-05,
|
| 103 |
+
"loss": 0.1766,
|
| 104 |
+
"step": 160
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.38,
|
| 108 |
+
"learning_rate": 4.8548922056384746e-05,
|
| 109 |
+
"loss": 0.1647,
|
| 110 |
+
"step": 170
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.4,
|
| 114 |
+
"learning_rate": 4.817578772802654e-05,
|
| 115 |
+
"loss": 0.1408,
|
| 116 |
+
"step": 180
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.42,
|
| 120 |
+
"learning_rate": 4.7761194029850745e-05,
|
| 121 |
+
"loss": 0.1196,
|
| 122 |
+
"step": 190
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.45,
|
| 126 |
+
"learning_rate": 4.734660033167496e-05,
|
| 127 |
+
"loss": 0.1494,
|
| 128 |
+
"step": 200
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.47,
|
| 132 |
+
"learning_rate": 4.693200663349917e-05,
|
| 133 |
+
"loss": 0.1374,
|
| 134 |
+
"step": 210
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.49,
|
| 138 |
+
"learning_rate": 4.651741293532338e-05,
|
| 139 |
+
"loss": 0.1103,
|
| 140 |
+
"step": 220
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.51,
|
| 144 |
+
"learning_rate": 4.6102819237147596e-05,
|
| 145 |
+
"loss": 0.1022,
|
| 146 |
+
"step": 230
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.54,
|
| 150 |
+
"learning_rate": 4.568822553897181e-05,
|
| 151 |
+
"loss": 0.1162,
|
| 152 |
+
"step": 240
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.56,
|
| 156 |
+
"learning_rate": 4.5273631840796025e-05,
|
| 157 |
+
"loss": 0.1039,
|
| 158 |
+
"step": 250
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.58,
|
| 162 |
+
"learning_rate": 4.485903814262023e-05,
|
| 163 |
+
"loss": 0.1057,
|
| 164 |
+
"step": 260
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.6,
|
| 168 |
+
"learning_rate": 4.4444444444444447e-05,
|
| 169 |
+
"loss": 0.1047,
|
| 170 |
+
"step": 270
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.63,
|
| 174 |
+
"learning_rate": 4.402985074626866e-05,
|
| 175 |
+
"loss": 0.1001,
|
| 176 |
+
"step": 280
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.65,
|
| 180 |
+
"learning_rate": 4.3615257048092875e-05,
|
| 181 |
+
"loss": 0.0874,
|
| 182 |
+
"step": 290
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.67,
|
| 186 |
+
"learning_rate": 4.320066334991708e-05,
|
| 187 |
+
"loss": 0.0888,
|
| 188 |
+
"step": 300
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.69,
|
| 192 |
+
"learning_rate": 4.27860696517413e-05,
|
| 193 |
+
"loss": 0.1011,
|
| 194 |
+
"step": 310
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.72,
|
| 198 |
+
"learning_rate": 4.237147595356551e-05,
|
| 199 |
+
"loss": 0.0868,
|
| 200 |
+
"step": 320
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.74,
|
| 204 |
+
"learning_rate": 4.195688225538972e-05,
|
| 205 |
+
"loss": 0.094,
|
| 206 |
+
"step": 330
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.76,
|
| 210 |
+
"learning_rate": 4.1542288557213934e-05,
|
| 211 |
+
"loss": 0.1121,
|
| 212 |
+
"step": 340
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.78,
|
| 216 |
+
"learning_rate": 4.112769485903815e-05,
|
| 217 |
+
"loss": 0.0703,
|
| 218 |
+
"step": 350
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.8,
|
| 222 |
+
"learning_rate": 4.0713101160862356e-05,
|
| 223 |
+
"loss": 0.0858,
|
| 224 |
+
"step": 360
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.83,
|
| 228 |
+
"learning_rate": 4.029850746268657e-05,
|
| 229 |
+
"loss": 0.0729,
|
| 230 |
+
"step": 370
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.85,
|
| 234 |
+
"learning_rate": 3.9883913764510784e-05,
|
| 235 |
+
"loss": 0.0652,
|
| 236 |
+
"step": 380
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.87,
|
| 240 |
+
"learning_rate": 3.946932006633499e-05,
|
| 241 |
+
"loss": 0.073,
|
| 242 |
+
"step": 390
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.89,
|
| 246 |
+
"learning_rate": 3.9054726368159206e-05,
|
| 247 |
+
"loss": 0.0711,
|
| 248 |
+
"step": 400
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.92,
|
| 252 |
+
"learning_rate": 3.864013266998342e-05,
|
| 253 |
+
"loss": 0.0739,
|
| 254 |
+
"step": 410
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.94,
|
| 258 |
+
"learning_rate": 3.822553897180763e-05,
|
| 259 |
+
"loss": 0.074,
|
| 260 |
+
"step": 420
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.96,
|
| 264 |
+
"learning_rate": 3.781094527363184e-05,
|
| 265 |
+
"loss": 0.066,
|
| 266 |
+
"step": 430
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.98,
|
| 270 |
+
"learning_rate": 3.739635157545606e-05,
|
| 271 |
+
"loss": 0.0592,
|
| 272 |
+
"step": 440
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 1.0,
|
| 276 |
+
"eval_accuracy": 0.9694968553459119,
|
| 277 |
+
"eval_loss": 0.0823180079460144,
|
| 278 |
+
"eval_runtime": 2100.5651,
|
| 279 |
+
"eval_samples_per_second": 3.028,
|
| 280 |
+
"eval_steps_per_second": 0.095,
|
| 281 |
+
"step": 447
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 1.01,
|
| 285 |
+
"learning_rate": 3.6981757877280265e-05,
|
| 286 |
+
"loss": 0.0781,
|
| 287 |
+
"step": 450
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 1.03,
|
| 291 |
+
"learning_rate": 3.656716417910448e-05,
|
| 292 |
+
"loss": 0.039,
|
| 293 |
+
"step": 460
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 1.05,
|
| 297 |
+
"learning_rate": 3.6152570480928693e-05,
|
| 298 |
+
"loss": 0.0402,
|
| 299 |
+
"step": 470
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 1.07,
|
| 303 |
+
"learning_rate": 3.57379767827529e-05,
|
| 304 |
+
"loss": 0.0585,
|
| 305 |
+
"step": 480
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 1.1,
|
| 309 |
+
"learning_rate": 3.5323383084577115e-05,
|
| 310 |
+
"loss": 0.0524,
|
| 311 |
+
"step": 490
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 1.12,
|
| 315 |
+
"learning_rate": 3.490878938640133e-05,
|
| 316 |
+
"loss": 0.0344,
|
| 317 |
+
"step": 500
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 1.14,
|
| 321 |
+
"learning_rate": 3.449419568822554e-05,
|
| 322 |
+
"loss": 0.0346,
|
| 323 |
+
"step": 510
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 1.16,
|
| 327 |
+
"learning_rate": 3.407960199004975e-05,
|
| 328 |
+
"loss": 0.0474,
|
| 329 |
+
"step": 520
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 1.19,
|
| 333 |
+
"learning_rate": 3.3665008291873966e-05,
|
| 334 |
+
"loss": 0.0308,
|
| 335 |
+
"step": 530
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 1.21,
|
| 339 |
+
"learning_rate": 3.3250414593698174e-05,
|
| 340 |
+
"loss": 0.0323,
|
| 341 |
+
"step": 540
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 1.23,
|
| 345 |
+
"learning_rate": 3.283582089552239e-05,
|
| 346 |
+
"loss": 0.035,
|
| 347 |
+
"step": 550
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 1.25,
|
| 351 |
+
"learning_rate": 3.24212271973466e-05,
|
| 352 |
+
"loss": 0.0217,
|
| 353 |
+
"step": 560
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 1.28,
|
| 357 |
+
"learning_rate": 3.200663349917081e-05,
|
| 358 |
+
"loss": 0.0305,
|
| 359 |
+
"step": 570
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 1.3,
|
| 363 |
+
"learning_rate": 3.1592039800995024e-05,
|
| 364 |
+
"loss": 0.0524,
|
| 365 |
+
"step": 580
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 1.32,
|
| 369 |
+
"learning_rate": 3.117744610281924e-05,
|
| 370 |
+
"loss": 0.0368,
|
| 371 |
+
"step": 590
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 1.34,
|
| 375 |
+
"learning_rate": 3.0762852404643446e-05,
|
| 376 |
+
"loss": 0.0292,
|
| 377 |
+
"step": 600
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 1.36,
|
| 381 |
+
"learning_rate": 3.0348258706467664e-05,
|
| 382 |
+
"loss": 0.0304,
|
| 383 |
+
"step": 610
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 1.39,
|
| 387 |
+
"learning_rate": 2.993366500829188e-05,
|
| 388 |
+
"loss": 0.0391,
|
| 389 |
+
"step": 620
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 1.41,
|
| 393 |
+
"learning_rate": 2.9519071310116086e-05,
|
| 394 |
+
"loss": 0.0359,
|
| 395 |
+
"step": 630
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 1.43,
|
| 399 |
+
"learning_rate": 2.9145936981757878e-05,
|
| 400 |
+
"loss": 0.0689,
|
| 401 |
+
"step": 640
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 1.45,
|
| 405 |
+
"learning_rate": 2.8731343283582092e-05,
|
| 406 |
+
"loss": 0.0271,
|
| 407 |
+
"step": 650
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 1.48,
|
| 411 |
+
"learning_rate": 2.8316749585406303e-05,
|
| 412 |
+
"loss": 0.0333,
|
| 413 |
+
"step": 660
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 1.5,
|
| 417 |
+
"learning_rate": 2.7902155887230514e-05,
|
| 418 |
+
"loss": 0.044,
|
| 419 |
+
"step": 670
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 1.52,
|
| 423 |
+
"learning_rate": 2.748756218905473e-05,
|
| 424 |
+
"loss": 0.0326,
|
| 425 |
+
"step": 680
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 1.54,
|
| 429 |
+
"learning_rate": 2.7072968490878943e-05,
|
| 430 |
+
"loss": 0.0378,
|
| 431 |
+
"step": 690
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 1.57,
|
| 435 |
+
"learning_rate": 2.665837479270315e-05,
|
| 436 |
+
"loss": 0.0265,
|
| 437 |
+
"step": 700
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 1.59,
|
| 441 |
+
"learning_rate": 2.6243781094527365e-05,
|
| 442 |
+
"loss": 0.0289,
|
| 443 |
+
"step": 710
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 1.61,
|
| 447 |
+
"learning_rate": 2.582918739635158e-05,
|
| 448 |
+
"loss": 0.0232,
|
| 449 |
+
"step": 720
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 1.63,
|
| 453 |
+
"learning_rate": 2.5414593698175787e-05,
|
| 454 |
+
"loss": 0.0254,
|
| 455 |
+
"step": 730
|
| 456 |
+
},
|
| 457 |
+
{
|
| 458 |
+
"epoch": 1.66,
|
| 459 |
+
"learning_rate": 2.5e-05,
|
| 460 |
+
"loss": 0.033,
|
| 461 |
+
"step": 740
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 1.68,
|
| 465 |
+
"learning_rate": 2.4585406301824212e-05,
|
| 466 |
+
"loss": 0.0195,
|
| 467 |
+
"step": 750
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"epoch": 1.7,
|
| 471 |
+
"learning_rate": 2.4170812603648427e-05,
|
| 472 |
+
"loss": 0.0252,
|
| 473 |
+
"step": 760
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 1.72,
|
| 477 |
+
"learning_rate": 2.3756218905472638e-05,
|
| 478 |
+
"loss": 0.0235,
|
| 479 |
+
"step": 770
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 1.74,
|
| 483 |
+
"learning_rate": 2.334162520729685e-05,
|
| 484 |
+
"loss": 0.0176,
|
| 485 |
+
"step": 780
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 1.77,
|
| 489 |
+
"learning_rate": 2.2927031509121063e-05,
|
| 490 |
+
"loss": 0.026,
|
| 491 |
+
"step": 790
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 1.79,
|
| 495 |
+
"learning_rate": 2.2512437810945274e-05,
|
| 496 |
+
"loss": 0.0224,
|
| 497 |
+
"step": 800
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 1.81,
|
| 501 |
+
"learning_rate": 2.2097844112769485e-05,
|
| 502 |
+
"loss": 0.0294,
|
| 503 |
+
"step": 810
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 1.83,
|
| 507 |
+
"learning_rate": 2.16832504145937e-05,
|
| 508 |
+
"loss": 0.0288,
|
| 509 |
+
"step": 820
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 1.86,
|
| 513 |
+
"learning_rate": 2.126865671641791e-05,
|
| 514 |
+
"loss": 0.0385,
|
| 515 |
+
"step": 830
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 1.88,
|
| 519 |
+
"learning_rate": 2.0854063018242125e-05,
|
| 520 |
+
"loss": 0.021,
|
| 521 |
+
"step": 840
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 1.9,
|
| 525 |
+
"learning_rate": 2.0439469320066336e-05,
|
| 526 |
+
"loss": 0.0235,
|
| 527 |
+
"step": 850
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 1.92,
|
| 531 |
+
"learning_rate": 2.002487562189055e-05,
|
| 532 |
+
"loss": 0.0249,
|
| 533 |
+
"step": 860
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 1.95,
|
| 537 |
+
"learning_rate": 1.961028192371476e-05,
|
| 538 |
+
"loss": 0.0222,
|
| 539 |
+
"step": 870
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 1.97,
|
| 543 |
+
"learning_rate": 1.9195688225538972e-05,
|
| 544 |
+
"loss": 0.0318,
|
| 545 |
+
"step": 880
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 1.99,
|
| 549 |
+
"learning_rate": 1.8781094527363186e-05,
|
| 550 |
+
"loss": 0.0196,
|
| 551 |
+
"step": 890
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 2.0,
|
| 555 |
+
"eval_accuracy": 0.9738993710691823,
|
| 556 |
+
"eval_loss": 0.07613054662942886,
|
| 557 |
+
"eval_runtime": 49.4856,
|
| 558 |
+
"eval_samples_per_second": 128.522,
|
| 559 |
+
"eval_steps_per_second": 4.021,
|
| 560 |
+
"step": 894
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 2.01,
|
| 564 |
+
"learning_rate": 1.8366500829187397e-05,
|
| 565 |
+
"loss": 0.0208,
|
| 566 |
+
"step": 900
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 2.04,
|
| 570 |
+
"learning_rate": 1.795190713101161e-05,
|
| 571 |
+
"loss": 0.013,
|
| 572 |
+
"step": 910
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 2.06,
|
| 576 |
+
"learning_rate": 1.7537313432835823e-05,
|
| 577 |
+
"loss": 0.0136,
|
| 578 |
+
"step": 920
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 2.08,
|
| 582 |
+
"learning_rate": 1.7122719734660034e-05,
|
| 583 |
+
"loss": 0.0072,
|
| 584 |
+
"step": 930
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 2.1,
|
| 588 |
+
"learning_rate": 1.6708126036484245e-05,
|
| 589 |
+
"loss": 0.0128,
|
| 590 |
+
"step": 940
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 2.13,
|
| 594 |
+
"learning_rate": 1.629353233830846e-05,
|
| 595 |
+
"loss": 0.0072,
|
| 596 |
+
"step": 950
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 2.15,
|
| 600 |
+
"learning_rate": 1.587893864013267e-05,
|
| 601 |
+
"loss": 0.0056,
|
| 602 |
+
"step": 960
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 2.17,
|
| 606 |
+
"learning_rate": 1.546434494195688e-05,
|
| 607 |
+
"loss": 0.0151,
|
| 608 |
+
"step": 970
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 2.19,
|
| 612 |
+
"learning_rate": 1.5049751243781095e-05,
|
| 613 |
+
"loss": 0.0124,
|
| 614 |
+
"step": 980
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 2.21,
|
| 618 |
+
"learning_rate": 1.4635157545605308e-05,
|
| 619 |
+
"loss": 0.0157,
|
| 620 |
+
"step": 990
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 2.24,
|
| 624 |
+
"learning_rate": 1.4220563847429519e-05,
|
| 625 |
+
"loss": 0.0095,
|
| 626 |
+
"step": 1000
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 2.26,
|
| 630 |
+
"learning_rate": 1.3805970149253733e-05,
|
| 631 |
+
"loss": 0.0107,
|
| 632 |
+
"step": 1010
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 2.28,
|
| 636 |
+
"learning_rate": 1.3391376451077944e-05,
|
| 637 |
+
"loss": 0.0131,
|
| 638 |
+
"step": 1020
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 2.3,
|
| 642 |
+
"learning_rate": 1.2976782752902155e-05,
|
| 643 |
+
"loss": 0.0072,
|
| 644 |
+
"step": 1030
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 2.33,
|
| 648 |
+
"learning_rate": 1.256218905472637e-05,
|
| 649 |
+
"loss": 0.0148,
|
| 650 |
+
"step": 1040
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 2.35,
|
| 654 |
+
"learning_rate": 1.214759535655058e-05,
|
| 655 |
+
"loss": 0.0086,
|
| 656 |
+
"step": 1050
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 2.37,
|
| 660 |
+
"learning_rate": 1.1733001658374792e-05,
|
| 661 |
+
"loss": 0.0129,
|
| 662 |
+
"step": 1060
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 2.39,
|
| 666 |
+
"learning_rate": 1.1318407960199006e-05,
|
| 667 |
+
"loss": 0.017,
|
| 668 |
+
"step": 1070
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 2.42,
|
| 672 |
+
"learning_rate": 1.0903814262023219e-05,
|
| 673 |
+
"loss": 0.013,
|
| 674 |
+
"step": 1080
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 2.44,
|
| 678 |
+
"learning_rate": 1.048922056384743e-05,
|
| 679 |
+
"loss": 0.0056,
|
| 680 |
+
"step": 1090
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 2.46,
|
| 684 |
+
"learning_rate": 1.0074626865671643e-05,
|
| 685 |
+
"loss": 0.0089,
|
| 686 |
+
"step": 1100
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 2.48,
|
| 690 |
+
"learning_rate": 9.660033167495855e-06,
|
| 691 |
+
"loss": 0.0072,
|
| 692 |
+
"step": 1110
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 2.51,
|
| 696 |
+
"learning_rate": 9.245439469320066e-06,
|
| 697 |
+
"loss": 0.0113,
|
| 698 |
+
"step": 1120
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 2.53,
|
| 702 |
+
"learning_rate": 8.830845771144279e-06,
|
| 703 |
+
"loss": 0.0068,
|
| 704 |
+
"step": 1130
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 2.55,
|
| 708 |
+
"learning_rate": 8.416252072968492e-06,
|
| 709 |
+
"loss": 0.0128,
|
| 710 |
+
"step": 1140
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 2.57,
|
| 714 |
+
"learning_rate": 8.001658374792703e-06,
|
| 715 |
+
"loss": 0.0061,
|
| 716 |
+
"step": 1150
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 2.59,
|
| 720 |
+
"learning_rate": 7.587064676616916e-06,
|
| 721 |
+
"loss": 0.0088,
|
| 722 |
+
"step": 1160
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 2.62,
|
| 726 |
+
"learning_rate": 7.172470978441129e-06,
|
| 727 |
+
"loss": 0.0037,
|
| 728 |
+
"step": 1170
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 2.64,
|
| 732 |
+
"learning_rate": 6.75787728026534e-06,
|
| 733 |
+
"loss": 0.0071,
|
| 734 |
+
"step": 1180
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 2.66,
|
| 738 |
+
"learning_rate": 6.343283582089552e-06,
|
| 739 |
+
"loss": 0.0038,
|
| 740 |
+
"step": 1190
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 2.68,
|
| 744 |
+
"learning_rate": 5.928689883913765e-06,
|
| 745 |
+
"loss": 0.0116,
|
| 746 |
+
"step": 1200
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 2.71,
|
| 750 |
+
"learning_rate": 5.514096185737977e-06,
|
| 751 |
+
"loss": 0.0089,
|
| 752 |
+
"step": 1210
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 2.73,
|
| 756 |
+
"learning_rate": 5.099502487562189e-06,
|
| 757 |
+
"loss": 0.0091,
|
| 758 |
+
"step": 1220
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 2.75,
|
| 762 |
+
"learning_rate": 4.6849087893864014e-06,
|
| 763 |
+
"loss": 0.0035,
|
| 764 |
+
"step": 1230
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 2.77,
|
| 768 |
+
"learning_rate": 4.270315091210614e-06,
|
| 769 |
+
"loss": 0.0066,
|
| 770 |
+
"step": 1240
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 2.8,
|
| 774 |
+
"learning_rate": 3.855721393034826e-06,
|
| 775 |
+
"loss": 0.0033,
|
| 776 |
+
"step": 1250
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 2.82,
|
| 780 |
+
"learning_rate": 3.4411276948590386e-06,
|
| 781 |
+
"loss": 0.0061,
|
| 782 |
+
"step": 1260
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 2.84,
|
| 786 |
+
"learning_rate": 3.0265339966832505e-06,
|
| 787 |
+
"loss": 0.0034,
|
| 788 |
+
"step": 1270
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 2.86,
|
| 792 |
+
"learning_rate": 2.6119402985074627e-06,
|
| 793 |
+
"loss": 0.0075,
|
| 794 |
+
"step": 1280
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 2.89,
|
| 798 |
+
"learning_rate": 2.197346600331675e-06,
|
| 799 |
+
"loss": 0.0094,
|
| 800 |
+
"step": 1290
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 2.91,
|
| 804 |
+
"learning_rate": 1.7827529021558872e-06,
|
| 805 |
+
"loss": 0.0098,
|
| 806 |
+
"step": 1300
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 2.93,
|
| 810 |
+
"learning_rate": 1.3681592039800995e-06,
|
| 811 |
+
"loss": 0.0067,
|
| 812 |
+
"step": 1310
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 2.95,
|
| 816 |
+
"learning_rate": 9.535655058043119e-07,
|
| 817 |
+
"loss": 0.0036,
|
| 818 |
+
"step": 1320
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 2.97,
|
| 822 |
+
"learning_rate": 5.38971807628524e-07,
|
| 823 |
+
"loss": 0.0045,
|
| 824 |
+
"step": 1330
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 3.0,
|
| 828 |
+
"learning_rate": 1.2437810945273633e-07,
|
| 829 |
+
"loss": 0.0058,
|
| 830 |
+
"step": 1340
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 3.0,
|
| 834 |
+
"eval_accuracy": 0.9806603773584905,
|
| 835 |
+
"eval_loss": 0.06080754101276398,
|
| 836 |
+
"eval_runtime": 49.2168,
|
| 837 |
+
"eval_samples_per_second": 129.224,
|
| 838 |
+
"eval_steps_per_second": 4.043,
|
| 839 |
+
"step": 1341
|
| 840 |
+
},
|
| 841 |
+
{
|
| 842 |
+
"epoch": 3.0,
|
| 843 |
+
"step": 1341,
|
| 844 |
+
"total_flos": 4.2672264771137864e+18,
|
| 845 |
+
"train_loss": 0.07098129998106373,
|
| 846 |
+
"train_runtime": 22089.0495,
|
| 847 |
+
"train_samples_per_second": 7.773,
|
| 848 |
+
"train_steps_per_second": 0.061
|
| 849 |
+
}
|
| 850 |
+
],
|
| 851 |
+
"max_steps": 1341,
|
| 852 |
+
"num_train_epochs": 3,
|
| 853 |
+
"total_flos": 4.2672264771137864e+18,
|
| 854 |
+
"trial_name": null,
|
| 855 |
+
"trial_params": null
|
| 856 |
+
}
|