Upload 7 files
Browse filesprovide the optimize state and parameters
checkpoint-3000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:74b1f584b40bb9bfdae608a7e43eed55e33f344ba8bf07d40a465cacc3ed3f16
|
| 3 |
+
size 52523141
|
checkpoint-3000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5e128f6b283f767dfa41535a396455b9fd6d89bb74a291ee40a8155eb49ae2e0
|
| 3 |
+
size 26271757
|
checkpoint-3000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:341848bc473600d46ffeb92fdd2cb6b58ff12fbc736eec38882150dfa66f396e
|
| 3 |
+
size 14575
|
checkpoint-3000/scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fb1f9398b77268202e8e1465734a63d123b1ef11c27f20f2473677e9883a6869
|
| 3 |
+
size 557
|
checkpoint-3000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8837f70c64b67c82e0c313c718f914155fd661e9ebab740551b7fb4b967b285f
|
| 3 |
+
size 627
|
checkpoint-3000/trainer_state.json
ADDED
|
@@ -0,0 +1,916 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.1066027296200664,
|
| 5 |
+
"global_step": 3000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.01,
|
| 12 |
+
"learning_rate": 9.975408828230666e-05,
|
| 13 |
+
"loss": 0.9349,
|
| 14 |
+
"step": 20
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.01,
|
| 18 |
+
"learning_rate": 9.950817656461331e-05,
|
| 19 |
+
"loss": 0.9288,
|
| 20 |
+
"step": 40
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.02,
|
| 24 |
+
"learning_rate": 9.926226484691996e-05,
|
| 25 |
+
"loss": 0.929,
|
| 26 |
+
"step": 60
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.03,
|
| 30 |
+
"learning_rate": 9.901635312922661e-05,
|
| 31 |
+
"loss": 0.9254,
|
| 32 |
+
"step": 80
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.04,
|
| 36 |
+
"learning_rate": 9.877044141153327e-05,
|
| 37 |
+
"loss": 0.9303,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.04,
|
| 42 |
+
"learning_rate": 9.852452969383992e-05,
|
| 43 |
+
"loss": 0.9255,
|
| 44 |
+
"step": 120
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.05,
|
| 48 |
+
"learning_rate": 9.827861797614657e-05,
|
| 49 |
+
"loss": 0.9197,
|
| 50 |
+
"step": 140
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.06,
|
| 54 |
+
"learning_rate": 9.803270625845322e-05,
|
| 55 |
+
"loss": 0.9275,
|
| 56 |
+
"step": 160
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.07,
|
| 60 |
+
"learning_rate": 9.778679454075987e-05,
|
| 61 |
+
"loss": 0.919,
|
| 62 |
+
"step": 180
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.07,
|
| 66 |
+
"learning_rate": 9.754088282306653e-05,
|
| 67 |
+
"loss": 0.9246,
|
| 68 |
+
"step": 200
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.08,
|
| 72 |
+
"learning_rate": 9.729497110537318e-05,
|
| 73 |
+
"loss": 0.917,
|
| 74 |
+
"step": 220
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.09,
|
| 78 |
+
"learning_rate": 9.704905938767983e-05,
|
| 79 |
+
"loss": 0.9211,
|
| 80 |
+
"step": 240
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.1,
|
| 84 |
+
"learning_rate": 9.680314766998648e-05,
|
| 85 |
+
"loss": 0.9204,
|
| 86 |
+
"step": 260
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.1,
|
| 90 |
+
"learning_rate": 9.655723595229314e-05,
|
| 91 |
+
"loss": 0.9127,
|
| 92 |
+
"step": 280
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.11,
|
| 96 |
+
"learning_rate": 9.631132423459979e-05,
|
| 97 |
+
"loss": 0.9077,
|
| 98 |
+
"step": 300
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.12,
|
| 102 |
+
"learning_rate": 9.606541251690644e-05,
|
| 103 |
+
"loss": 0.9176,
|
| 104 |
+
"step": 320
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.13,
|
| 108 |
+
"learning_rate": 9.581950079921309e-05,
|
| 109 |
+
"loss": 0.9119,
|
| 110 |
+
"step": 340
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.13,
|
| 114 |
+
"learning_rate": 9.557358908151974e-05,
|
| 115 |
+
"loss": 0.9135,
|
| 116 |
+
"step": 360
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.14,
|
| 120 |
+
"learning_rate": 9.53276773638264e-05,
|
| 121 |
+
"loss": 0.9098,
|
| 122 |
+
"step": 380
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.15,
|
| 126 |
+
"learning_rate": 9.508176564613305e-05,
|
| 127 |
+
"loss": 0.9043,
|
| 128 |
+
"step": 400
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.15,
|
| 132 |
+
"learning_rate": 9.48358539284397e-05,
|
| 133 |
+
"loss": 0.9167,
|
| 134 |
+
"step": 420
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.16,
|
| 138 |
+
"learning_rate": 9.458994221074635e-05,
|
| 139 |
+
"loss": 0.9076,
|
| 140 |
+
"step": 440
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.17,
|
| 144 |
+
"learning_rate": 9.4344030493053e-05,
|
| 145 |
+
"loss": 0.9039,
|
| 146 |
+
"step": 460
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.18,
|
| 150 |
+
"learning_rate": 9.409811877535964e-05,
|
| 151 |
+
"loss": 0.9147,
|
| 152 |
+
"step": 480
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.18,
|
| 156 |
+
"learning_rate": 9.38522070576663e-05,
|
| 157 |
+
"loss": 0.9108,
|
| 158 |
+
"step": 500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.19,
|
| 162 |
+
"learning_rate": 9.360629533997295e-05,
|
| 163 |
+
"loss": 0.9068,
|
| 164 |
+
"step": 520
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.2,
|
| 168 |
+
"learning_rate": 9.33603836222796e-05,
|
| 169 |
+
"loss": 0.9123,
|
| 170 |
+
"step": 540
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.21,
|
| 174 |
+
"learning_rate": 9.311447190458625e-05,
|
| 175 |
+
"loss": 0.9084,
|
| 176 |
+
"step": 560
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.21,
|
| 180 |
+
"learning_rate": 9.28685601868929e-05,
|
| 181 |
+
"loss": 0.9049,
|
| 182 |
+
"step": 580
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.22,
|
| 186 |
+
"learning_rate": 9.262264846919956e-05,
|
| 187 |
+
"loss": 0.9044,
|
| 188 |
+
"step": 600
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.23,
|
| 192 |
+
"learning_rate": 9.237673675150621e-05,
|
| 193 |
+
"loss": 0.9074,
|
| 194 |
+
"step": 620
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.24,
|
| 198 |
+
"learning_rate": 9.213082503381286e-05,
|
| 199 |
+
"loss": 0.9024,
|
| 200 |
+
"step": 640
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.24,
|
| 204 |
+
"learning_rate": 9.188491331611951e-05,
|
| 205 |
+
"loss": 0.9017,
|
| 206 |
+
"step": 660
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.25,
|
| 210 |
+
"learning_rate": 9.163900159842617e-05,
|
| 211 |
+
"loss": 0.895,
|
| 212 |
+
"step": 680
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.26,
|
| 216 |
+
"learning_rate": 9.139308988073282e-05,
|
| 217 |
+
"loss": 0.8943,
|
| 218 |
+
"step": 700
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.27,
|
| 222 |
+
"learning_rate": 9.114717816303947e-05,
|
| 223 |
+
"loss": 0.9036,
|
| 224 |
+
"step": 720
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.27,
|
| 228 |
+
"learning_rate": 9.090126644534612e-05,
|
| 229 |
+
"loss": 0.8968,
|
| 230 |
+
"step": 740
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.28,
|
| 234 |
+
"learning_rate": 9.065535472765277e-05,
|
| 235 |
+
"loss": 0.8997,
|
| 236 |
+
"step": 760
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.29,
|
| 240 |
+
"learning_rate": 9.040944300995943e-05,
|
| 241 |
+
"loss": 0.8935,
|
| 242 |
+
"step": 780
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.3,
|
| 246 |
+
"learning_rate": 9.016353129226608e-05,
|
| 247 |
+
"loss": 0.891,
|
| 248 |
+
"step": 800
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.3,
|
| 252 |
+
"learning_rate": 8.991761957457273e-05,
|
| 253 |
+
"loss": 0.8986,
|
| 254 |
+
"step": 820
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.31,
|
| 258 |
+
"learning_rate": 8.967170785687938e-05,
|
| 259 |
+
"loss": 0.8976,
|
| 260 |
+
"step": 840
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.32,
|
| 264 |
+
"learning_rate": 8.942579613918604e-05,
|
| 265 |
+
"loss": 0.9008,
|
| 266 |
+
"step": 860
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.32,
|
| 270 |
+
"learning_rate": 8.917988442149269e-05,
|
| 271 |
+
"loss": 0.9023,
|
| 272 |
+
"step": 880
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.33,
|
| 276 |
+
"learning_rate": 8.893397270379934e-05,
|
| 277 |
+
"loss": 0.8904,
|
| 278 |
+
"step": 900
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 0.34,
|
| 282 |
+
"learning_rate": 8.868806098610599e-05,
|
| 283 |
+
"loss": 0.8892,
|
| 284 |
+
"step": 920
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.35,
|
| 288 |
+
"learning_rate": 8.844214926841264e-05,
|
| 289 |
+
"loss": 0.8949,
|
| 290 |
+
"step": 940
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.35,
|
| 294 |
+
"learning_rate": 8.81962375507193e-05,
|
| 295 |
+
"loss": 0.8982,
|
| 296 |
+
"step": 960
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.36,
|
| 300 |
+
"learning_rate": 8.795032583302595e-05,
|
| 301 |
+
"loss": 0.888,
|
| 302 |
+
"step": 980
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.37,
|
| 306 |
+
"learning_rate": 8.77044141153326e-05,
|
| 307 |
+
"loss": 0.8866,
|
| 308 |
+
"step": 1000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.38,
|
| 312 |
+
"learning_rate": 8.745850239763925e-05,
|
| 313 |
+
"loss": 0.8859,
|
| 314 |
+
"step": 1020
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.38,
|
| 318 |
+
"learning_rate": 8.72125906799459e-05,
|
| 319 |
+
"loss": 0.8952,
|
| 320 |
+
"step": 1040
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 0.39,
|
| 324 |
+
"learning_rate": 8.696667896225256e-05,
|
| 325 |
+
"loss": 0.8855,
|
| 326 |
+
"step": 1060
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 0.4,
|
| 330 |
+
"learning_rate": 8.672076724455921e-05,
|
| 331 |
+
"loss": 0.8951,
|
| 332 |
+
"step": 1080
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.41,
|
| 336 |
+
"learning_rate": 8.647485552686586e-05,
|
| 337 |
+
"loss": 0.8875,
|
| 338 |
+
"step": 1100
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.41,
|
| 342 |
+
"learning_rate": 8.622894380917251e-05,
|
| 343 |
+
"loss": 0.8919,
|
| 344 |
+
"step": 1120
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 0.42,
|
| 348 |
+
"learning_rate": 8.598303209147917e-05,
|
| 349 |
+
"loss": 0.8889,
|
| 350 |
+
"step": 1140
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 0.43,
|
| 354 |
+
"learning_rate": 8.573712037378582e-05,
|
| 355 |
+
"loss": 0.8857,
|
| 356 |
+
"step": 1160
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.44,
|
| 360 |
+
"learning_rate": 8.549120865609247e-05,
|
| 361 |
+
"loss": 0.8876,
|
| 362 |
+
"step": 1180
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 0.44,
|
| 366 |
+
"learning_rate": 8.524529693839912e-05,
|
| 367 |
+
"loss": 0.8905,
|
| 368 |
+
"step": 1200
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 0.45,
|
| 372 |
+
"learning_rate": 8.499938522070577e-05,
|
| 373 |
+
"loss": 0.886,
|
| 374 |
+
"step": 1220
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.46,
|
| 378 |
+
"learning_rate": 8.475347350301243e-05,
|
| 379 |
+
"loss": 0.8839,
|
| 380 |
+
"step": 1240
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.46,
|
| 384 |
+
"learning_rate": 8.450756178531908e-05,
|
| 385 |
+
"loss": 0.8851,
|
| 386 |
+
"step": 1260
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.47,
|
| 390 |
+
"learning_rate": 8.426165006762573e-05,
|
| 391 |
+
"loss": 0.8784,
|
| 392 |
+
"step": 1280
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.48,
|
| 396 |
+
"learning_rate": 8.401573834993238e-05,
|
| 397 |
+
"loss": 0.8897,
|
| 398 |
+
"step": 1300
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 0.49,
|
| 402 |
+
"learning_rate": 8.376982663223904e-05,
|
| 403 |
+
"loss": 0.883,
|
| 404 |
+
"step": 1320
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 0.49,
|
| 408 |
+
"learning_rate": 8.352391491454569e-05,
|
| 409 |
+
"loss": 0.8819,
|
| 410 |
+
"step": 1340
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.5,
|
| 414 |
+
"learning_rate": 8.327800319685234e-05,
|
| 415 |
+
"loss": 0.883,
|
| 416 |
+
"step": 1360
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.51,
|
| 420 |
+
"learning_rate": 8.303209147915899e-05,
|
| 421 |
+
"loss": 0.8864,
|
| 422 |
+
"step": 1380
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.52,
|
| 426 |
+
"learning_rate": 8.278617976146564e-05,
|
| 427 |
+
"loss": 0.8875,
|
| 428 |
+
"step": 1400
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 0.52,
|
| 432 |
+
"learning_rate": 8.254026804377228e-05,
|
| 433 |
+
"loss": 0.8749,
|
| 434 |
+
"step": 1420
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 0.53,
|
| 438 |
+
"learning_rate": 8.229435632607894e-05,
|
| 439 |
+
"loss": 0.878,
|
| 440 |
+
"step": 1440
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 0.54,
|
| 444 |
+
"learning_rate": 8.204844460838559e-05,
|
| 445 |
+
"loss": 0.8786,
|
| 446 |
+
"step": 1460
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.55,
|
| 450 |
+
"learning_rate": 8.180253289069224e-05,
|
| 451 |
+
"loss": 0.8844,
|
| 452 |
+
"step": 1480
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 0.55,
|
| 456 |
+
"learning_rate": 8.155662117299889e-05,
|
| 457 |
+
"loss": 0.8768,
|
| 458 |
+
"step": 1500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.56,
|
| 462 |
+
"learning_rate": 8.131070945530554e-05,
|
| 463 |
+
"loss": 0.8751,
|
| 464 |
+
"step": 1520
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.57,
|
| 468 |
+
"learning_rate": 8.10647977376122e-05,
|
| 469 |
+
"loss": 0.8795,
|
| 470 |
+
"step": 1540
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 0.58,
|
| 474 |
+
"learning_rate": 8.081888601991885e-05,
|
| 475 |
+
"loss": 0.8782,
|
| 476 |
+
"step": 1560
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 0.58,
|
| 480 |
+
"learning_rate": 8.05729743022255e-05,
|
| 481 |
+
"loss": 0.8789,
|
| 482 |
+
"step": 1580
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 0.59,
|
| 486 |
+
"learning_rate": 8.032706258453215e-05,
|
| 487 |
+
"loss": 0.8801,
|
| 488 |
+
"step": 1600
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 0.6,
|
| 492 |
+
"learning_rate": 8.00811508668388e-05,
|
| 493 |
+
"loss": 0.8772,
|
| 494 |
+
"step": 1620
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.6,
|
| 498 |
+
"learning_rate": 7.983523914914546e-05,
|
| 499 |
+
"loss": 0.8722,
|
| 500 |
+
"step": 1640
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 0.61,
|
| 504 |
+
"learning_rate": 7.958932743145211e-05,
|
| 505 |
+
"loss": 0.8749,
|
| 506 |
+
"step": 1660
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.62,
|
| 510 |
+
"learning_rate": 7.934341571375876e-05,
|
| 511 |
+
"loss": 0.8728,
|
| 512 |
+
"step": 1680
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 0.63,
|
| 516 |
+
"learning_rate": 7.909750399606541e-05,
|
| 517 |
+
"loss": 0.8811,
|
| 518 |
+
"step": 1700
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 0.63,
|
| 522 |
+
"learning_rate": 7.885159227837207e-05,
|
| 523 |
+
"loss": 0.8741,
|
| 524 |
+
"step": 1720
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 0.64,
|
| 528 |
+
"learning_rate": 7.860568056067872e-05,
|
| 529 |
+
"loss": 0.8697,
|
| 530 |
+
"step": 1740
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 0.65,
|
| 534 |
+
"learning_rate": 7.835976884298537e-05,
|
| 535 |
+
"loss": 0.8746,
|
| 536 |
+
"step": 1760
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 0.66,
|
| 540 |
+
"learning_rate": 7.811385712529202e-05,
|
| 541 |
+
"loss": 0.878,
|
| 542 |
+
"step": 1780
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.66,
|
| 546 |
+
"learning_rate": 7.786794540759867e-05,
|
| 547 |
+
"loss": 0.8766,
|
| 548 |
+
"step": 1800
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.67,
|
| 552 |
+
"learning_rate": 7.762203368990533e-05,
|
| 553 |
+
"loss": 0.8694,
|
| 554 |
+
"step": 1820
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 0.68,
|
| 558 |
+
"learning_rate": 7.737612197221198e-05,
|
| 559 |
+
"loss": 0.8821,
|
| 560 |
+
"step": 1840
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 0.69,
|
| 564 |
+
"learning_rate": 7.713021025451863e-05,
|
| 565 |
+
"loss": 0.8776,
|
| 566 |
+
"step": 1860
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 0.69,
|
| 570 |
+
"learning_rate": 7.688429853682528e-05,
|
| 571 |
+
"loss": 0.8678,
|
| 572 |
+
"step": 1880
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 0.7,
|
| 576 |
+
"learning_rate": 7.663838681913194e-05,
|
| 577 |
+
"loss": 0.8709,
|
| 578 |
+
"step": 1900
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 0.71,
|
| 582 |
+
"learning_rate": 7.639247510143859e-05,
|
| 583 |
+
"loss": 0.8683,
|
| 584 |
+
"step": 1920
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 0.72,
|
| 588 |
+
"learning_rate": 7.614656338374524e-05,
|
| 589 |
+
"loss": 0.8754,
|
| 590 |
+
"step": 1940
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.72,
|
| 594 |
+
"learning_rate": 7.590065166605189e-05,
|
| 595 |
+
"loss": 0.8724,
|
| 596 |
+
"step": 1960
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 0.73,
|
| 600 |
+
"learning_rate": 7.565473994835854e-05,
|
| 601 |
+
"loss": 0.8679,
|
| 602 |
+
"step": 1980
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 0.74,
|
| 606 |
+
"learning_rate": 7.54088282306652e-05,
|
| 607 |
+
"loss": 0.8747,
|
| 608 |
+
"step": 2000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 0.75,
|
| 612 |
+
"learning_rate": 7.516291651297185e-05,
|
| 613 |
+
"loss": 0.8669,
|
| 614 |
+
"step": 2020
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 0.75,
|
| 618 |
+
"learning_rate": 7.49170047952785e-05,
|
| 619 |
+
"loss": 0.8712,
|
| 620 |
+
"step": 2040
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 0.76,
|
| 624 |
+
"learning_rate": 7.467109307758515e-05,
|
| 625 |
+
"loss": 0.8656,
|
| 626 |
+
"step": 2060
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 0.77,
|
| 630 |
+
"learning_rate": 7.44251813598918e-05,
|
| 631 |
+
"loss": 0.8736,
|
| 632 |
+
"step": 2080
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.77,
|
| 636 |
+
"learning_rate": 7.417926964219846e-05,
|
| 637 |
+
"loss": 0.868,
|
| 638 |
+
"step": 2100
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 0.78,
|
| 642 |
+
"learning_rate": 7.393335792450511e-05,
|
| 643 |
+
"loss": 0.8747,
|
| 644 |
+
"step": 2120
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 0.79,
|
| 648 |
+
"learning_rate": 7.368744620681176e-05,
|
| 649 |
+
"loss": 0.8693,
|
| 650 |
+
"step": 2140
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 0.8,
|
| 654 |
+
"learning_rate": 7.344153448911841e-05,
|
| 655 |
+
"loss": 0.8704,
|
| 656 |
+
"step": 2160
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 0.8,
|
| 660 |
+
"learning_rate": 7.319562277142507e-05,
|
| 661 |
+
"loss": 0.8616,
|
| 662 |
+
"step": 2180
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 0.81,
|
| 666 |
+
"learning_rate": 7.294971105373172e-05,
|
| 667 |
+
"loss": 0.8692,
|
| 668 |
+
"step": 2200
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 0.82,
|
| 672 |
+
"learning_rate": 7.270379933603837e-05,
|
| 673 |
+
"loss": 0.8721,
|
| 674 |
+
"step": 2220
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.83,
|
| 678 |
+
"learning_rate": 7.245788761834502e-05,
|
| 679 |
+
"loss": 0.8632,
|
| 680 |
+
"step": 2240
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 0.83,
|
| 684 |
+
"learning_rate": 7.221197590065168e-05,
|
| 685 |
+
"loss": 0.8698,
|
| 686 |
+
"step": 2260
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 0.84,
|
| 690 |
+
"learning_rate": 7.196606418295833e-05,
|
| 691 |
+
"loss": 0.8606,
|
| 692 |
+
"step": 2280
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 0.85,
|
| 696 |
+
"learning_rate": 7.172015246526498e-05,
|
| 697 |
+
"loss": 0.8626,
|
| 698 |
+
"step": 2300
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 0.86,
|
| 702 |
+
"learning_rate": 7.147424074757163e-05,
|
| 703 |
+
"loss": 0.865,
|
| 704 |
+
"step": 2320
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 0.86,
|
| 708 |
+
"learning_rate": 7.122832902987828e-05,
|
| 709 |
+
"loss": 0.8651,
|
| 710 |
+
"step": 2340
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 0.87,
|
| 714 |
+
"learning_rate": 7.098241731218494e-05,
|
| 715 |
+
"loss": 0.8676,
|
| 716 |
+
"step": 2360
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.88,
|
| 720 |
+
"learning_rate": 7.073650559449157e-05,
|
| 721 |
+
"loss": 0.8629,
|
| 722 |
+
"step": 2380
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 0.89,
|
| 726 |
+
"learning_rate": 7.049059387679823e-05,
|
| 727 |
+
"loss": 0.8636,
|
| 728 |
+
"step": 2400
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 0.89,
|
| 732 |
+
"learning_rate": 7.024468215910488e-05,
|
| 733 |
+
"loss": 0.8632,
|
| 734 |
+
"step": 2420
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 0.9,
|
| 738 |
+
"learning_rate": 6.999877044141153e-05,
|
| 739 |
+
"loss": 0.8632,
|
| 740 |
+
"step": 2440
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 0.91,
|
| 744 |
+
"learning_rate": 6.975285872371818e-05,
|
| 745 |
+
"loss": 0.8655,
|
| 746 |
+
"step": 2460
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 0.91,
|
| 750 |
+
"learning_rate": 6.950694700602484e-05,
|
| 751 |
+
"loss": 0.8659,
|
| 752 |
+
"step": 2480
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 0.92,
|
| 756 |
+
"learning_rate": 6.926103528833149e-05,
|
| 757 |
+
"loss": 0.8581,
|
| 758 |
+
"step": 2500
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.93,
|
| 762 |
+
"learning_rate": 6.901512357063814e-05,
|
| 763 |
+
"loss": 0.8608,
|
| 764 |
+
"step": 2520
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 0.94,
|
| 768 |
+
"learning_rate": 6.876921185294479e-05,
|
| 769 |
+
"loss": 0.8623,
|
| 770 |
+
"step": 2540
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 0.94,
|
| 774 |
+
"learning_rate": 6.852330013525144e-05,
|
| 775 |
+
"loss": 0.8644,
|
| 776 |
+
"step": 2560
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 0.95,
|
| 780 |
+
"learning_rate": 6.82773884175581e-05,
|
| 781 |
+
"loss": 0.8604,
|
| 782 |
+
"step": 2580
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 0.96,
|
| 786 |
+
"learning_rate": 6.803147669986475e-05,
|
| 787 |
+
"loss": 0.8708,
|
| 788 |
+
"step": 2600
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 0.97,
|
| 792 |
+
"learning_rate": 6.77855649821714e-05,
|
| 793 |
+
"loss": 0.8585,
|
| 794 |
+
"step": 2620
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 0.97,
|
| 798 |
+
"learning_rate": 6.753965326447805e-05,
|
| 799 |
+
"loss": 0.8596,
|
| 800 |
+
"step": 2640
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.98,
|
| 804 |
+
"learning_rate": 6.72937415467847e-05,
|
| 805 |
+
"loss": 0.8582,
|
| 806 |
+
"step": 2660
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 0.99,
|
| 810 |
+
"learning_rate": 6.704782982909136e-05,
|
| 811 |
+
"loss": 0.8587,
|
| 812 |
+
"step": 2680
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 1.0,
|
| 816 |
+
"learning_rate": 6.680191811139801e-05,
|
| 817 |
+
"loss": 0.8598,
|
| 818 |
+
"step": 2700
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 1.0,
|
| 822 |
+
"learning_rate": 6.655600639370466e-05,
|
| 823 |
+
"loss": 0.8617,
|
| 824 |
+
"step": 2720
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 1.01,
|
| 828 |
+
"learning_rate": 6.631009467601131e-05,
|
| 829 |
+
"loss": 0.8589,
|
| 830 |
+
"step": 2740
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 1.02,
|
| 834 |
+
"learning_rate": 6.606418295831797e-05,
|
| 835 |
+
"loss": 0.8582,
|
| 836 |
+
"step": 2760
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 1.03,
|
| 840 |
+
"learning_rate": 6.581827124062462e-05,
|
| 841 |
+
"loss": 0.8576,
|
| 842 |
+
"step": 2780
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 1.03,
|
| 846 |
+
"learning_rate": 6.557235952293127e-05,
|
| 847 |
+
"loss": 0.8613,
|
| 848 |
+
"step": 2800
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 1.04,
|
| 852 |
+
"learning_rate": 6.532644780523792e-05,
|
| 853 |
+
"loss": 0.8517,
|
| 854 |
+
"step": 2820
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 1.05,
|
| 858 |
+
"learning_rate": 6.508053608754458e-05,
|
| 859 |
+
"loss": 0.8579,
|
| 860 |
+
"step": 2840
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 1.05,
|
| 864 |
+
"learning_rate": 6.483462436985123e-05,
|
| 865 |
+
"loss": 0.8588,
|
| 866 |
+
"step": 2860
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 1.06,
|
| 870 |
+
"learning_rate": 6.458871265215788e-05,
|
| 871 |
+
"loss": 0.8554,
|
| 872 |
+
"step": 2880
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 1.07,
|
| 876 |
+
"learning_rate": 6.434280093446453e-05,
|
| 877 |
+
"loss": 0.8542,
|
| 878 |
+
"step": 2900
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 1.08,
|
| 882 |
+
"learning_rate": 6.409688921677118e-05,
|
| 883 |
+
"loss": 0.8501,
|
| 884 |
+
"step": 2920
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 1.08,
|
| 888 |
+
"learning_rate": 6.385097749907784e-05,
|
| 889 |
+
"loss": 0.8523,
|
| 890 |
+
"step": 2940
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 1.09,
|
| 894 |
+
"learning_rate": 6.360506578138449e-05,
|
| 895 |
+
"loss": 0.8506,
|
| 896 |
+
"step": 2960
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 1.1,
|
| 900 |
+
"learning_rate": 6.335915406369114e-05,
|
| 901 |
+
"loss": 0.8613,
|
| 902 |
+
"step": 2980
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 1.11,
|
| 906 |
+
"learning_rate": 6.311324234599779e-05,
|
| 907 |
+
"loss": 0.8468,
|
| 908 |
+
"step": 3000
|
| 909 |
+
}
|
| 910 |
+
],
|
| 911 |
+
"max_steps": 8133,
|
| 912 |
+
"num_train_epochs": 3,
|
| 913 |
+
"total_flos": 1.5168022953677292e+19,
|
| 914 |
+
"trial_name": null,
|
| 915 |
+
"trial_params": null
|
| 916 |
+
}
|
checkpoint-3000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d0c3111c6ecafee1a5c2564b172597e0bc4912398407304187ecbd7fde603e10
|
| 3 |
+
size 3579
|