loss 1.0
Browse files- optimizer.pt +1 -1
- pytorch_model.bin +1 -1
- rng_state.pth +1 -1
- scheduler.pt +1 -1
- trainer_state.json +933 -3
optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1376950965
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f76533d218d66acc8b5bba0f1a7679d4b368b7cef88912358d862be08181db2a
|
| 3 |
size 1376950965
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 688496379
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:22c0886ab707b8fb20a1104edd37cdee899a2e4445bb7bafa1d7c325a933a6ad
|
| 3 |
size 688496379
|
rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14503
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9fd43fe4eb5e834fcfb1f491790d59c502c67ed4729f7fbbb3291ddfcd6feac4
|
| 3 |
size 14503
|
scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 623
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:32b51740f55e1aeef4ee3b07911236360a8c4fc355d594709c1ff07a872a7511
|
| 3 |
size 623
|
trainer_state.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch":
|
| 5 |
-
"global_step":
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
|
@@ -36,11 +36,941 @@
|
|
| 36 |
"learning_rate": 0.0006961985445856413,
|
| 37 |
"loss": 1.0191,
|
| 38 |
"step": 500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 39 |
}
|
| 40 |
],
|
| 41 |
"max_steps": 92070,
|
| 42 |
"num_train_epochs": 10,
|
| 43 |
-
"total_flos":
|
| 44 |
"trial_name": null,
|
| 45 |
"trial_params": null
|
| 46 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.7377881560726602,
|
| 5 |
+
"global_step": 16000,
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
|
|
|
| 36 |
"learning_rate": 0.0006961985445856413,
|
| 37 |
"loss": 1.0191,
|
| 38 |
"step": 500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.07,
|
| 42 |
+
"learning_rate": 0.0006954382535027696,
|
| 43 |
+
"loss": 1.0418,
|
| 44 |
+
"step": 600
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.08,
|
| 48 |
+
"learning_rate": 0.0006946779624198979,
|
| 49 |
+
"loss": 1.1059,
|
| 50 |
+
"step": 700
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.09,
|
| 54 |
+
"learning_rate": 0.0006939176713370262,
|
| 55 |
+
"loss": 1.0536,
|
| 56 |
+
"step": 800
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.1,
|
| 60 |
+
"learning_rate": 0.0006931573802541545,
|
| 61 |
+
"loss": 1.0631,
|
| 62 |
+
"step": 900
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.11,
|
| 66 |
+
"learning_rate": 0.0006923970891712827,
|
| 67 |
+
"loss": 1.0861,
|
| 68 |
+
"step": 1000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.12,
|
| 72 |
+
"learning_rate": 0.0006916367980884109,
|
| 73 |
+
"loss": 1.1102,
|
| 74 |
+
"step": 1100
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.13,
|
| 78 |
+
"learning_rate": 0.0006908765070055393,
|
| 79 |
+
"loss": 1.0543,
|
| 80 |
+
"step": 1200
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.14,
|
| 84 |
+
"learning_rate": 0.0006901162159226675,
|
| 85 |
+
"loss": 1.0827,
|
| 86 |
+
"step": 1300
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.15,
|
| 90 |
+
"learning_rate": 0.0006893559248397958,
|
| 91 |
+
"loss": 1.0794,
|
| 92 |
+
"step": 1400
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.16,
|
| 96 |
+
"learning_rate": 0.0006885956337569241,
|
| 97 |
+
"loss": 1.0768,
|
| 98 |
+
"step": 1500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.17,
|
| 102 |
+
"learning_rate": 0.0006878353426740524,
|
| 103 |
+
"loss": 1.1544,
|
| 104 |
+
"step": 1600
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.18,
|
| 108 |
+
"learning_rate": 0.0006870750515911806,
|
| 109 |
+
"loss": 1.1412,
|
| 110 |
+
"step": 1700
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.2,
|
| 114 |
+
"learning_rate": 0.0006863147605083089,
|
| 115 |
+
"loss": 1.0918,
|
| 116 |
+
"step": 1800
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.21,
|
| 120 |
+
"learning_rate": 0.0006855544694254371,
|
| 121 |
+
"loss": 1.0784,
|
| 122 |
+
"step": 1900
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.22,
|
| 126 |
+
"learning_rate": 0.0006847941783425654,
|
| 127 |
+
"loss": 1.085,
|
| 128 |
+
"step": 2000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.23,
|
| 132 |
+
"learning_rate": 0.0006840338872596937,
|
| 133 |
+
"loss": 1.1224,
|
| 134 |
+
"step": 2100
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.24,
|
| 138 |
+
"learning_rate": 0.000683273596176822,
|
| 139 |
+
"loss": 1.1052,
|
| 140 |
+
"step": 2200
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.25,
|
| 144 |
+
"learning_rate": 0.0006825133050939503,
|
| 145 |
+
"loss": 1.0949,
|
| 146 |
+
"step": 2300
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.26,
|
| 150 |
+
"learning_rate": 0.0006817530140110785,
|
| 151 |
+
"loss": 1.0888,
|
| 152 |
+
"step": 2400
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.27,
|
| 156 |
+
"learning_rate": 0.0006809927229282067,
|
| 157 |
+
"loss": 1.0424,
|
| 158 |
+
"step": 2500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.28,
|
| 162 |
+
"learning_rate": 0.0006802324318453351,
|
| 163 |
+
"loss": 1.0191,
|
| 164 |
+
"step": 2600
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.29,
|
| 168 |
+
"learning_rate": 0.0006794721407624634,
|
| 169 |
+
"loss": 1.1374,
|
| 170 |
+
"step": 2700
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.3,
|
| 174 |
+
"learning_rate": 0.0006787118496795916,
|
| 175 |
+
"loss": 1.0963,
|
| 176 |
+
"step": 2800
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.31,
|
| 180 |
+
"learning_rate": 0.0006779515585967199,
|
| 181 |
+
"loss": 1.1217,
|
| 182 |
+
"step": 2900
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.33,
|
| 186 |
+
"learning_rate": 0.0006771912675138482,
|
| 187 |
+
"loss": 1.1168,
|
| 188 |
+
"step": 3000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.34,
|
| 192 |
+
"learning_rate": 0.0006764309764309764,
|
| 193 |
+
"loss": 1.0919,
|
| 194 |
+
"step": 3100
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.35,
|
| 198 |
+
"learning_rate": 0.0006756706853481047,
|
| 199 |
+
"loss": 1.0552,
|
| 200 |
+
"step": 3200
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.36,
|
| 204 |
+
"learning_rate": 0.000674910394265233,
|
| 205 |
+
"loss": 1.1408,
|
| 206 |
+
"step": 3300
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.37,
|
| 210 |
+
"learning_rate": 0.0006741501031823612,
|
| 211 |
+
"loss": 1.1443,
|
| 212 |
+
"step": 3400
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.38,
|
| 216 |
+
"learning_rate": 0.0006733898120994895,
|
| 217 |
+
"loss": 1.137,
|
| 218 |
+
"step": 3500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.39,
|
| 222 |
+
"learning_rate": 0.0006726295210166178,
|
| 223 |
+
"loss": 1.1577,
|
| 224 |
+
"step": 3600
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.4,
|
| 228 |
+
"learning_rate": 0.000671869229933746,
|
| 229 |
+
"loss": 1.1505,
|
| 230 |
+
"step": 3700
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.41,
|
| 234 |
+
"learning_rate": 0.0006711089388508743,
|
| 235 |
+
"loss": 1.1479,
|
| 236 |
+
"step": 3800
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.42,
|
| 240 |
+
"learning_rate": 0.0006703486477680026,
|
| 241 |
+
"loss": 1.0887,
|
| 242 |
+
"step": 3900
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.43,
|
| 246 |
+
"learning_rate": 0.0006695883566851309,
|
| 247 |
+
"loss": 1.1219,
|
| 248 |
+
"step": 4000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.45,
|
| 252 |
+
"learning_rate": 0.0006688280656022592,
|
| 253 |
+
"loss": 1.1188,
|
| 254 |
+
"step": 4100
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.46,
|
| 258 |
+
"learning_rate": 0.0006680677745193874,
|
| 259 |
+
"loss": 1.0956,
|
| 260 |
+
"step": 4200
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.47,
|
| 264 |
+
"learning_rate": 0.0006673074834365157,
|
| 265 |
+
"loss": 1.1544,
|
| 266 |
+
"step": 4300
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.48,
|
| 270 |
+
"learning_rate": 0.0006665471923536439,
|
| 271 |
+
"loss": 1.1014,
|
| 272 |
+
"step": 4400
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.49,
|
| 276 |
+
"learning_rate": 0.0006657869012707722,
|
| 277 |
+
"loss": 1.0761,
|
| 278 |
+
"step": 4500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 0.5,
|
| 282 |
+
"learning_rate": 0.0006650266101879005,
|
| 283 |
+
"loss": 1.1168,
|
| 284 |
+
"step": 4600
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.51,
|
| 288 |
+
"learning_rate": 0.0006642663191050288,
|
| 289 |
+
"loss": 1.0792,
|
| 290 |
+
"step": 4700
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.52,
|
| 294 |
+
"learning_rate": 0.000663506028022157,
|
| 295 |
+
"loss": 1.085,
|
| 296 |
+
"step": 4800
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.53,
|
| 300 |
+
"learning_rate": 0.0006627457369392853,
|
| 301 |
+
"loss": 1.1649,
|
| 302 |
+
"step": 4900
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.54,
|
| 306 |
+
"learning_rate": 0.0006619854458564137,
|
| 307 |
+
"loss": 1.1575,
|
| 308 |
+
"step": 5000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.55,
|
| 312 |
+
"learning_rate": 0.0006612251547735418,
|
| 313 |
+
"loss": 1.098,
|
| 314 |
+
"step": 5100
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.56,
|
| 318 |
+
"learning_rate": 0.0006604648636906701,
|
| 319 |
+
"loss": 1.1442,
|
| 320 |
+
"step": 5200
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 0.58,
|
| 324 |
+
"learning_rate": 0.0006597045726077984,
|
| 325 |
+
"loss": 1.1135,
|
| 326 |
+
"step": 5300
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 0.59,
|
| 330 |
+
"learning_rate": 0.0006589442815249267,
|
| 331 |
+
"loss": 1.1428,
|
| 332 |
+
"step": 5400
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.6,
|
| 336 |
+
"learning_rate": 0.000658183990442055,
|
| 337 |
+
"loss": 1.0998,
|
| 338 |
+
"step": 5500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.61,
|
| 342 |
+
"learning_rate": 0.0006574236993591833,
|
| 343 |
+
"loss": 1.0847,
|
| 344 |
+
"step": 5600
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 0.62,
|
| 348 |
+
"learning_rate": 0.0006566634082763115,
|
| 349 |
+
"loss": 1.0957,
|
| 350 |
+
"step": 5700
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 0.63,
|
| 354 |
+
"learning_rate": 0.0006559031171934397,
|
| 355 |
+
"loss": 1.0979,
|
| 356 |
+
"step": 5800
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.64,
|
| 360 |
+
"learning_rate": 0.000655142826110568,
|
| 361 |
+
"loss": 1.1163,
|
| 362 |
+
"step": 5900
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 0.65,
|
| 366 |
+
"learning_rate": 0.0006543825350276963,
|
| 367 |
+
"loss": 1.1216,
|
| 368 |
+
"step": 6000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 0.66,
|
| 372 |
+
"learning_rate": 0.0006536222439448246,
|
| 373 |
+
"loss": 1.0892,
|
| 374 |
+
"step": 6100
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.67,
|
| 378 |
+
"learning_rate": 0.0006528619528619529,
|
| 379 |
+
"loss": 1.1051,
|
| 380 |
+
"step": 6200
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.68,
|
| 384 |
+
"learning_rate": 0.0006521016617790811,
|
| 385 |
+
"loss": 1.1076,
|
| 386 |
+
"step": 6300
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.7,
|
| 390 |
+
"learning_rate": 0.0006513413706962095,
|
| 391 |
+
"loss": 1.1235,
|
| 392 |
+
"step": 6400
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.71,
|
| 396 |
+
"learning_rate": 0.0006505810796133376,
|
| 397 |
+
"loss": 1.0842,
|
| 398 |
+
"step": 6500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 0.72,
|
| 402 |
+
"learning_rate": 0.0006498207885304659,
|
| 403 |
+
"loss": 1.1257,
|
| 404 |
+
"step": 6600
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 0.73,
|
| 408 |
+
"learning_rate": 0.0006490604974475942,
|
| 409 |
+
"loss": 1.1014,
|
| 410 |
+
"step": 6700
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.74,
|
| 414 |
+
"learning_rate": 0.0006483002063647225,
|
| 415 |
+
"loss": 1.0675,
|
| 416 |
+
"step": 6800
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.75,
|
| 420 |
+
"learning_rate": 0.0006475399152818508,
|
| 421 |
+
"loss": 1.0937,
|
| 422 |
+
"step": 6900
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.76,
|
| 426 |
+
"learning_rate": 0.0006467796241989791,
|
| 427 |
+
"loss": 1.1009,
|
| 428 |
+
"step": 7000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 0.77,
|
| 432 |
+
"learning_rate": 0.0006460193331161072,
|
| 433 |
+
"loss": 1.0831,
|
| 434 |
+
"step": 7100
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 0.78,
|
| 438 |
+
"learning_rate": 0.0006452590420332355,
|
| 439 |
+
"loss": 1.0571,
|
| 440 |
+
"step": 7200
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 0.79,
|
| 444 |
+
"learning_rate": 0.0006444987509503639,
|
| 445 |
+
"loss": 1.046,
|
| 446 |
+
"step": 7300
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.8,
|
| 450 |
+
"learning_rate": 0.0006437384598674921,
|
| 451 |
+
"loss": 1.1214,
|
| 452 |
+
"step": 7400
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 0.81,
|
| 456 |
+
"learning_rate": 0.0006429781687846204,
|
| 457 |
+
"loss": 1.0686,
|
| 458 |
+
"step": 7500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.83,
|
| 462 |
+
"learning_rate": 0.0006422178777017487,
|
| 463 |
+
"loss": 1.1346,
|
| 464 |
+
"step": 7600
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.84,
|
| 468 |
+
"learning_rate": 0.0006414575866188769,
|
| 469 |
+
"loss": 1.16,
|
| 470 |
+
"step": 7700
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 0.85,
|
| 474 |
+
"learning_rate": 0.0006406972955360052,
|
| 475 |
+
"loss": 1.1717,
|
| 476 |
+
"step": 7800
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 0.86,
|
| 480 |
+
"learning_rate": 0.0006399370044531335,
|
| 481 |
+
"loss": 1.1601,
|
| 482 |
+
"step": 7900
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 0.87,
|
| 486 |
+
"learning_rate": 0.0006391767133702617,
|
| 487 |
+
"loss": 1.1241,
|
| 488 |
+
"step": 8000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 0.88,
|
| 492 |
+
"learning_rate": 0.00063841642228739,
|
| 493 |
+
"loss": 1.0692,
|
| 494 |
+
"step": 8100
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.89,
|
| 498 |
+
"learning_rate": 0.0006376561312045183,
|
| 499 |
+
"loss": 1.0775,
|
| 500 |
+
"step": 8200
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 0.9,
|
| 504 |
+
"learning_rate": 0.0006368958401216466,
|
| 505 |
+
"loss": 1.126,
|
| 506 |
+
"step": 8300
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.91,
|
| 510 |
+
"learning_rate": 0.0006361355490387749,
|
| 511 |
+
"loss": 1.0801,
|
| 512 |
+
"step": 8400
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 0.92,
|
| 516 |
+
"learning_rate": 0.0006353752579559031,
|
| 517 |
+
"loss": 1.1176,
|
| 518 |
+
"step": 8500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 0.93,
|
| 522 |
+
"learning_rate": 0.0006346149668730313,
|
| 523 |
+
"loss": 1.1101,
|
| 524 |
+
"step": 8600
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 0.94,
|
| 528 |
+
"learning_rate": 0.0006338546757901597,
|
| 529 |
+
"loss": 1.0881,
|
| 530 |
+
"step": 8700
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 0.96,
|
| 534 |
+
"learning_rate": 0.0006330943847072879,
|
| 535 |
+
"loss": 1.1009,
|
| 536 |
+
"step": 8800
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 0.97,
|
| 540 |
+
"learning_rate": 0.0006323340936244162,
|
| 541 |
+
"loss": 1.1158,
|
| 542 |
+
"step": 8900
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.98,
|
| 546 |
+
"learning_rate": 0.0006315738025415445,
|
| 547 |
+
"loss": 1.0991,
|
| 548 |
+
"step": 9000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.99,
|
| 552 |
+
"learning_rate": 0.0006308135114586728,
|
| 553 |
+
"loss": 1.1478,
|
| 554 |
+
"step": 9100
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 1.0,
|
| 558 |
+
"learning_rate": 0.000630053220375801,
|
| 559 |
+
"loss": 1.1408,
|
| 560 |
+
"step": 9200
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 1.01,
|
| 564 |
+
"learning_rate": 0.0006292929292929293,
|
| 565 |
+
"loss": 1.0484,
|
| 566 |
+
"step": 9300
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 1.02,
|
| 570 |
+
"learning_rate": 0.0006285326382100575,
|
| 571 |
+
"loss": 1.103,
|
| 572 |
+
"step": 9400
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 1.03,
|
| 576 |
+
"learning_rate": 0.0006277723471271858,
|
| 577 |
+
"loss": 0.9948,
|
| 578 |
+
"step": 9500
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 1.04,
|
| 582 |
+
"learning_rate": 0.0006270120560443141,
|
| 583 |
+
"loss": 1.0212,
|
| 584 |
+
"step": 9600
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 1.05,
|
| 588 |
+
"learning_rate": 0.0006262517649614424,
|
| 589 |
+
"loss": 1.1029,
|
| 590 |
+
"step": 9700
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 1.06,
|
| 594 |
+
"learning_rate": 0.0006254914738785707,
|
| 595 |
+
"loss": 1.064,
|
| 596 |
+
"step": 9800
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 1.08,
|
| 600 |
+
"learning_rate": 0.0006247311827956989,
|
| 601 |
+
"loss": 1.0066,
|
| 602 |
+
"step": 9900
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 1.09,
|
| 606 |
+
"learning_rate": 0.0006239708917128271,
|
| 607 |
+
"loss": 1.0117,
|
| 608 |
+
"step": 10000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 1.1,
|
| 612 |
+
"learning_rate": 0.0006232106006299555,
|
| 613 |
+
"loss": 1.0196,
|
| 614 |
+
"step": 10100
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 1.11,
|
| 618 |
+
"learning_rate": 0.0006224503095470838,
|
| 619 |
+
"loss": 1.0576,
|
| 620 |
+
"step": 10200
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 1.12,
|
| 624 |
+
"learning_rate": 0.000621690018464212,
|
| 625 |
+
"loss": 1.0636,
|
| 626 |
+
"step": 10300
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 1.13,
|
| 630 |
+
"learning_rate": 0.0006209297273813403,
|
| 631 |
+
"loss": 1.0454,
|
| 632 |
+
"step": 10400
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 1.14,
|
| 636 |
+
"learning_rate": 0.0006201694362984685,
|
| 637 |
+
"loss": 1.0798,
|
| 638 |
+
"step": 10500
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 1.15,
|
| 642 |
+
"learning_rate": 0.0006194091452155968,
|
| 643 |
+
"loss": 1.007,
|
| 644 |
+
"step": 10600
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 1.16,
|
| 648 |
+
"learning_rate": 0.0006186488541327251,
|
| 649 |
+
"loss": 1.0612,
|
| 650 |
+
"step": 10700
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 1.17,
|
| 654 |
+
"learning_rate": 0.0006178885630498534,
|
| 655 |
+
"loss": 1.0791,
|
| 656 |
+
"step": 10800
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 1.18,
|
| 660 |
+
"learning_rate": 0.0006171282719669816,
|
| 661 |
+
"loss": 1.0961,
|
| 662 |
+
"step": 10900
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 1.19,
|
| 666 |
+
"learning_rate": 0.0006163679808841099,
|
| 667 |
+
"loss": 1.0745,
|
| 668 |
+
"step": 11000
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 1.21,
|
| 672 |
+
"learning_rate": 0.0006156076898012382,
|
| 673 |
+
"loss": 0.9832,
|
| 674 |
+
"step": 11100
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 1.22,
|
| 678 |
+
"learning_rate": 0.0006148473987183664,
|
| 679 |
+
"loss": 1.0083,
|
| 680 |
+
"step": 11200
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 1.23,
|
| 684 |
+
"learning_rate": 0.0006140871076354947,
|
| 685 |
+
"loss": 1.0772,
|
| 686 |
+
"step": 11300
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 1.24,
|
| 690 |
+
"learning_rate": 0.000613326816552623,
|
| 691 |
+
"loss": 1.0673,
|
| 692 |
+
"step": 11400
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 1.25,
|
| 696 |
+
"learning_rate": 0.0006125665254697513,
|
| 697 |
+
"loss": 1.1005,
|
| 698 |
+
"step": 11500
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 1.26,
|
| 702 |
+
"learning_rate": 0.0006118062343868796,
|
| 703 |
+
"loss": 1.0219,
|
| 704 |
+
"step": 11600
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 1.27,
|
| 708 |
+
"learning_rate": 0.0006110459433040078,
|
| 709 |
+
"loss": 1.0878,
|
| 710 |
+
"step": 11700
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 1.28,
|
| 714 |
+
"learning_rate": 0.0006102856522211361,
|
| 715 |
+
"loss": 1.047,
|
| 716 |
+
"step": 11800
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 1.29,
|
| 720 |
+
"learning_rate": 0.0006095253611382643,
|
| 721 |
+
"loss": 1.0663,
|
| 722 |
+
"step": 11900
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 1.3,
|
| 726 |
+
"learning_rate": 0.0006087650700553926,
|
| 727 |
+
"loss": 1.0473,
|
| 728 |
+
"step": 12000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 1.31,
|
| 732 |
+
"learning_rate": 0.0006080047789725209,
|
| 733 |
+
"loss": 1.0146,
|
| 734 |
+
"step": 12100
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 1.33,
|
| 738 |
+
"learning_rate": 0.0006072444878896492,
|
| 739 |
+
"loss": 1.0381,
|
| 740 |
+
"step": 12200
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 1.34,
|
| 744 |
+
"learning_rate": 0.0006064841968067774,
|
| 745 |
+
"loss": 1.0991,
|
| 746 |
+
"step": 12300
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 1.35,
|
| 750 |
+
"learning_rate": 0.0006057239057239057,
|
| 751 |
+
"loss": 1.0377,
|
| 752 |
+
"step": 12400
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 1.36,
|
| 756 |
+
"learning_rate": 0.0006049636146410341,
|
| 757 |
+
"loss": 1.0478,
|
| 758 |
+
"step": 12500
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 1.37,
|
| 762 |
+
"learning_rate": 0.0006042033235581622,
|
| 763 |
+
"loss": 1.0773,
|
| 764 |
+
"step": 12600
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 1.38,
|
| 768 |
+
"learning_rate": 0.0006034430324752905,
|
| 769 |
+
"loss": 1.0496,
|
| 770 |
+
"step": 12700
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 1.39,
|
| 774 |
+
"learning_rate": 0.0006026827413924188,
|
| 775 |
+
"loss": 1.0943,
|
| 776 |
+
"step": 12800
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 1.4,
|
| 780 |
+
"learning_rate": 0.0006019224503095471,
|
| 781 |
+
"loss": 1.0303,
|
| 782 |
+
"step": 12900
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 1.41,
|
| 786 |
+
"learning_rate": 0.0006011621592266754,
|
| 787 |
+
"loss": 0.9987,
|
| 788 |
+
"step": 13000
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 1.42,
|
| 792 |
+
"learning_rate": 0.0006004018681438037,
|
| 793 |
+
"loss": 1.0489,
|
| 794 |
+
"step": 13100
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 1.43,
|
| 798 |
+
"learning_rate": 0.0005996415770609318,
|
| 799 |
+
"loss": 1.111,
|
| 800 |
+
"step": 13200
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 1.44,
|
| 804 |
+
"learning_rate": 0.0005988812859780601,
|
| 805 |
+
"loss": 1.0514,
|
| 806 |
+
"step": 13300
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 1.46,
|
| 810 |
+
"learning_rate": 0.0005981209948951884,
|
| 811 |
+
"loss": 1.0931,
|
| 812 |
+
"step": 13400
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 1.47,
|
| 816 |
+
"learning_rate": 0.0005973607038123167,
|
| 817 |
+
"loss": 1.0225,
|
| 818 |
+
"step": 13500
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 1.48,
|
| 822 |
+
"learning_rate": 0.000596600412729445,
|
| 823 |
+
"loss": 1.0295,
|
| 824 |
+
"step": 13600
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 1.49,
|
| 828 |
+
"learning_rate": 0.0005958401216465733,
|
| 829 |
+
"loss": 1.1068,
|
| 830 |
+
"step": 13700
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 1.5,
|
| 834 |
+
"learning_rate": 0.0005950798305637015,
|
| 835 |
+
"loss": 1.0739,
|
| 836 |
+
"step": 13800
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 1.51,
|
| 840 |
+
"learning_rate": 0.0005943195394808299,
|
| 841 |
+
"loss": 1.0453,
|
| 842 |
+
"step": 13900
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 1.52,
|
| 846 |
+
"learning_rate": 0.000593559248397958,
|
| 847 |
+
"loss": 1.0392,
|
| 848 |
+
"step": 14000
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 1.53,
|
| 852 |
+
"learning_rate": 0.0005927989573150863,
|
| 853 |
+
"loss": 1.0482,
|
| 854 |
+
"step": 14100
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 1.54,
|
| 858 |
+
"learning_rate": 0.0005920386662322146,
|
| 859 |
+
"loss": 1.0337,
|
| 860 |
+
"step": 14200
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 1.55,
|
| 864 |
+
"learning_rate": 0.0005912783751493429,
|
| 865 |
+
"loss": 1.0415,
|
| 866 |
+
"step": 14300
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 1.56,
|
| 870 |
+
"learning_rate": 0.0005905180840664712,
|
| 871 |
+
"loss": 1.0534,
|
| 872 |
+
"step": 14400
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 1.57,
|
| 876 |
+
"learning_rate": 0.0005897577929835995,
|
| 877 |
+
"loss": 1.028,
|
| 878 |
+
"step": 14500
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 1.59,
|
| 882 |
+
"learning_rate": 0.0005889975019007276,
|
| 883 |
+
"loss": 0.9978,
|
| 884 |
+
"step": 14600
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 1.6,
|
| 888 |
+
"learning_rate": 0.0005882372108178559,
|
| 889 |
+
"loss": 1.047,
|
| 890 |
+
"step": 14700
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 1.61,
|
| 894 |
+
"learning_rate": 0.0005874769197349843,
|
| 895 |
+
"loss": 1.0501,
|
| 896 |
+
"step": 14800
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 1.62,
|
| 900 |
+
"learning_rate": 0.0005867166286521125,
|
| 901 |
+
"loss": 1.0165,
|
| 902 |
+
"step": 14900
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 1.63,
|
| 906 |
+
"learning_rate": 0.0005859563375692408,
|
| 907 |
+
"loss": 1.0867,
|
| 908 |
+
"step": 15000
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 1.64,
|
| 912 |
+
"learning_rate": 0.0005851960464863691,
|
| 913 |
+
"loss": 1.0614,
|
| 914 |
+
"step": 15100
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 1.65,
|
| 918 |
+
"learning_rate": 0.0005844357554034973,
|
| 919 |
+
"loss": 1.0512,
|
| 920 |
+
"step": 15200
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 1.66,
|
| 924 |
+
"learning_rate": 0.0005836754643206256,
|
| 925 |
+
"loss": 0.9962,
|
| 926 |
+
"step": 15300
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 1.67,
|
| 930 |
+
"learning_rate": 0.0005829151732377539,
|
| 931 |
+
"loss": 1.0325,
|
| 932 |
+
"step": 15400
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 1.68,
|
| 936 |
+
"learning_rate": 0.0005821548821548821,
|
| 937 |
+
"loss": 1.0313,
|
| 938 |
+
"step": 15500
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 1.69,
|
| 942 |
+
"learning_rate": 0.0005813945910720104,
|
| 943 |
+
"loss": 1.0441,
|
| 944 |
+
"step": 15600
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 1.71,
|
| 948 |
+
"learning_rate": 0.0005806342999891387,
|
| 949 |
+
"loss": 1.0705,
|
| 950 |
+
"step": 15700
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 1.72,
|
| 954 |
+
"learning_rate": 0.000579874008906267,
|
| 955 |
+
"loss": 1.0209,
|
| 956 |
+
"step": 15800
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 1.73,
|
| 960 |
+
"learning_rate": 0.0005791137178233953,
|
| 961 |
+
"loss": 1.0627,
|
| 962 |
+
"step": 15900
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 1.74,
|
| 966 |
+
"learning_rate": 0.0005783534267405235,
|
| 967 |
+
"loss": 1.076,
|
| 968 |
+
"step": 16000
|
| 969 |
}
|
| 970 |
],
|
| 971 |
"max_steps": 92070,
|
| 972 |
"num_train_epochs": 10,
|
| 973 |
+
"total_flos": 3.281082781355213e+16,
|
| 974 |
"trial_name": null,
|
| 975 |
"trial_params": null
|
| 976 |
}
|