Training in progress, step 2000, checkpoint
Browse files- last-checkpoint/model.safetensors +1 -1
- last-checkpoint/optimizer.pt +1 -1
- last-checkpoint/rng_state_0.pth +1 -1
- last-checkpoint/rng_state_1.pth +1 -1
- last-checkpoint/rng_state_2.pth +1 -1
- last-checkpoint/rng_state_3.pth +1 -1
- last-checkpoint/rng_state_4.pth +1 -1
- last-checkpoint/rng_state_5.pth +1 -1
- last-checkpoint/rng_state_6.pth +1 -1
- last-checkpoint/rng_state_7.pth +1 -1
- last-checkpoint/scheduler.pt +1 -1
- last-checkpoint/trainer_state.json +766 -2
last-checkpoint/model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 738367848
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ba21b9e70b08c8157a0be6f7f43a461e7115ad8add195cfd8d7ecd41a4a559ba
|
| 3 |
size 738367848
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1476823354
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f32f2ca1f450a8a345cef7a5a79a872c2bc4a1ba658bcfe499421b034c07bf73
|
| 3 |
size 1476823354
|
last-checkpoint/rng_state_0.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6ca51d4b33edcedf9568d5202767b896d828b5aeca18f2cdd82617688464b784
|
| 3 |
size 15984
|
last-checkpoint/rng_state_1.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:10e6ea705ea5a1704cd5773090c827a2013c8caab967a116ff24a5f57ce3ce90
|
| 3 |
size 15984
|
last-checkpoint/rng_state_2.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f6ff0b4da12dd0cdcb6e90b04160e41685d9ccc1fa1cc74bb7949edf700200d4
|
| 3 |
size 15984
|
last-checkpoint/rng_state_3.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:322470b09ac4f5d9443d55c37c8b8e7d0e8a1702208c81e52e3a58a8de515b5b
|
| 3 |
size 15984
|
last-checkpoint/rng_state_4.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ababe8505205ca2bcb959a2abbe2fbc8b6ad677bd43b1f2ee9055b3cb400b061
|
| 3 |
size 15984
|
last-checkpoint/rng_state_5.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ed1c3ba656fdb40a72824e366a08e148a30e1089a6ecd019eaf28fa4a17859fa
|
| 3 |
size 15984
|
last-checkpoint/rng_state_6.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:042c62210b9e1c9ed394e0a3362b1c773c07591d94f2716a8e928676134742b7
|
| 3 |
size 15984
|
last-checkpoint/rng_state_7.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bb54f0278c663494261026658652f845bae43245e75ccc213c6897de179f542a
|
| 3 |
size 15984
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1000
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5fe8eacf1bb6bef5171d133da4e7818c1e7ae0192bd879d86e40a8d923aeb81f
|
| 3 |
size 1000
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 250,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -771,6 +771,770 @@
|
|
| 771 |
"eval_spearman_manhattan": 0.7778181970888292,
|
| 772 |
"eval_steps_per_second": 8.583,
|
| 773 |
"step": 1000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 774 |
}
|
| 775 |
],
|
| 776 |
"logging_steps": 10,
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.8744142455482662,
|
| 5 |
"eval_steps": 250,
|
| 6 |
+
"global_step": 2000,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 771 |
"eval_spearman_manhattan": 0.7778181970888292,
|
| 772 |
"eval_steps_per_second": 8.583,
|
| 773 |
"step": 1000
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"epoch": 0.9465791940018744,
|
| 777 |
+
"grad_norm": 1.1828556060791016,
|
| 778 |
+
"learning_rate": 9.992604958338825e-06,
|
| 779 |
+
"loss": 0.2168,
|
| 780 |
+
"step": 1010
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 0.9559512652296157,
|
| 784 |
+
"grad_norm": 1.2189664840698242,
|
| 785 |
+
"learning_rate": 9.992531740104556e-06,
|
| 786 |
+
"loss": 0.2072,
|
| 787 |
+
"step": 1020
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"epoch": 0.9653233364573571,
|
| 791 |
+
"grad_norm": 1.6102409362792969,
|
| 792 |
+
"learning_rate": 9.992458521870287e-06,
|
| 793 |
+
"loss": 0.2228,
|
| 794 |
+
"step": 1030
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 0.9746954076850984,
|
| 798 |
+
"grad_norm": 1.6891916990280151,
|
| 799 |
+
"learning_rate": 9.99238530363602e-06,
|
| 800 |
+
"loss": 0.2404,
|
| 801 |
+
"step": 1040
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 0.9840674789128397,
|
| 805 |
+
"grad_norm": 1.2274008989334106,
|
| 806 |
+
"learning_rate": 9.99231208540175e-06,
|
| 807 |
+
"loss": 0.2225,
|
| 808 |
+
"step": 1050
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 0.993439550140581,
|
| 812 |
+
"grad_norm": 1.2388169765472412,
|
| 813 |
+
"learning_rate": 9.992238867167479e-06,
|
| 814 |
+
"loss": 0.2215,
|
| 815 |
+
"step": 1060
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"epoch": 1.0028116213683225,
|
| 819 |
+
"grad_norm": 1.2347650527954102,
|
| 820 |
+
"learning_rate": 9.992165648933211e-06,
|
| 821 |
+
"loss": 0.2239,
|
| 822 |
+
"step": 1070
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"epoch": 1.0121836925960637,
|
| 826 |
+
"grad_norm": 1.1266793012619019,
|
| 827 |
+
"learning_rate": 9.992092430698942e-06,
|
| 828 |
+
"loss": 0.1932,
|
| 829 |
+
"step": 1080
|
| 830 |
+
},
|
| 831 |
+
{
|
| 832 |
+
"epoch": 1.021555763823805,
|
| 833 |
+
"grad_norm": 1.5187146663665771,
|
| 834 |
+
"learning_rate": 9.992019212464673e-06,
|
| 835 |
+
"loss": 0.205,
|
| 836 |
+
"step": 1090
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 1.0309278350515463,
|
| 840 |
+
"grad_norm": 1.4463717937469482,
|
| 841 |
+
"learning_rate": 9.991945994230404e-06,
|
| 842 |
+
"loss": 0.1818,
|
| 843 |
+
"step": 1100
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 1.0402999062792877,
|
| 847 |
+
"grad_norm": 1.6186790466308594,
|
| 848 |
+
"learning_rate": 9.991872775996136e-06,
|
| 849 |
+
"loss": 0.2076,
|
| 850 |
+
"step": 1110
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 1.0496719775070291,
|
| 854 |
+
"grad_norm": 1.3895883560180664,
|
| 855 |
+
"learning_rate": 9.991799557761865e-06,
|
| 856 |
+
"loss": 0.2096,
|
| 857 |
+
"step": 1120
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 1.0590440487347703,
|
| 861 |
+
"grad_norm": 1.296912670135498,
|
| 862 |
+
"learning_rate": 9.991726339527596e-06,
|
| 863 |
+
"loss": 0.2046,
|
| 864 |
+
"step": 1130
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 1.0684161199625117,
|
| 868 |
+
"grad_norm": 1.5527839660644531,
|
| 869 |
+
"learning_rate": 9.991653121293328e-06,
|
| 870 |
+
"loss": 0.1972,
|
| 871 |
+
"step": 1140
|
| 872 |
+
},
|
| 873 |
+
{
|
| 874 |
+
"epoch": 1.077788191190253,
|
| 875 |
+
"grad_norm": 1.4777096509933472,
|
| 876 |
+
"learning_rate": 9.99157990305906e-06,
|
| 877 |
+
"loss": 0.2086,
|
| 878 |
+
"step": 1150
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 1.0871602624179943,
|
| 882 |
+
"grad_norm": 1.3155533075332642,
|
| 883 |
+
"learning_rate": 9.99150668482479e-06,
|
| 884 |
+
"loss": 0.1969,
|
| 885 |
+
"step": 1160
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 1.0965323336457358,
|
| 889 |
+
"grad_norm": 1.5277265310287476,
|
| 890 |
+
"learning_rate": 9.99143346659052e-06,
|
| 891 |
+
"loss": 0.1923,
|
| 892 |
+
"step": 1170
|
| 893 |
+
},
|
| 894 |
+
{
|
| 895 |
+
"epoch": 1.105904404873477,
|
| 896 |
+
"grad_norm": 1.3764179944992065,
|
| 897 |
+
"learning_rate": 9.991360248356251e-06,
|
| 898 |
+
"loss": 0.1916,
|
| 899 |
+
"step": 1180
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 1.1152764761012184,
|
| 903 |
+
"grad_norm": 1.6024688482284546,
|
| 904 |
+
"learning_rate": 9.991287030121982e-06,
|
| 905 |
+
"loss": 0.185,
|
| 906 |
+
"step": 1190
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"epoch": 1.1246485473289598,
|
| 910 |
+
"grad_norm": 1.2752821445465088,
|
| 911 |
+
"learning_rate": 9.991213811887713e-06,
|
| 912 |
+
"loss": 0.1829,
|
| 913 |
+
"step": 1200
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 1.134020618556701,
|
| 917 |
+
"grad_norm": 1.4704368114471436,
|
| 918 |
+
"learning_rate": 9.991140593653444e-06,
|
| 919 |
+
"loss": 0.2006,
|
| 920 |
+
"step": 1210
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 1.1433926897844424,
|
| 924 |
+
"grad_norm": 1.3614213466644287,
|
| 925 |
+
"learning_rate": 9.991067375419176e-06,
|
| 926 |
+
"loss": 0.1776,
|
| 927 |
+
"step": 1220
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 1.1527647610121836,
|
| 931 |
+
"grad_norm": 1.2852075099945068,
|
| 932 |
+
"learning_rate": 9.990994157184905e-06,
|
| 933 |
+
"loss": 0.2116,
|
| 934 |
+
"step": 1230
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 1.162136832239925,
|
| 938 |
+
"grad_norm": 1.1774332523345947,
|
| 939 |
+
"learning_rate": 9.990920938950636e-06,
|
| 940 |
+
"loss": 0.1909,
|
| 941 |
+
"step": 1240
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 1.1715089034676662,
|
| 945 |
+
"grad_norm": 1.0442605018615723,
|
| 946 |
+
"learning_rate": 9.990847720716368e-06,
|
| 947 |
+
"loss": 0.1933,
|
| 948 |
+
"step": 1250
|
| 949 |
+
},
|
| 950 |
+
{
|
| 951 |
+
"epoch": 1.1715089034676662,
|
| 952 |
+
"eval_loss": 0.08017747104167938,
|
| 953 |
+
"eval_pearson_cosine": 0.7703680992126465,
|
| 954 |
+
"eval_pearson_dot": 0.6808142066001892,
|
| 955 |
+
"eval_pearson_euclidean": 0.7676056623458862,
|
| 956 |
+
"eval_pearson_manhattan": 0.7677772045135498,
|
| 957 |
+
"eval_runtime": 22.1599,
|
| 958 |
+
"eval_samples_per_second": 67.69,
|
| 959 |
+
"eval_spearman_cosine": 0.7790172740054649,
|
| 960 |
+
"eval_spearman_dot": 0.6796557194170769,
|
| 961 |
+
"eval_spearman_euclidean": 0.7739566900498013,
|
| 962 |
+
"eval_spearman_manhattan": 0.7741509176342483,
|
| 963 |
+
"eval_steps_per_second": 8.484,
|
| 964 |
+
"step": 1250
|
| 965 |
+
},
|
| 966 |
+
{
|
| 967 |
+
"epoch": 1.1808809746954076,
|
| 968 |
+
"grad_norm": 1.3561466932296753,
|
| 969 |
+
"learning_rate": 9.990774502482099e-06,
|
| 970 |
+
"loss": 0.1921,
|
| 971 |
+
"step": 1260
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"epoch": 1.190253045923149,
|
| 975 |
+
"grad_norm": 1.2151105403900146,
|
| 976 |
+
"learning_rate": 9.99070128424783e-06,
|
| 977 |
+
"loss": 0.1865,
|
| 978 |
+
"step": 1270
|
| 979 |
+
},
|
| 980 |
+
{
|
| 981 |
+
"epoch": 1.1996251171508903,
|
| 982 |
+
"grad_norm": 1.4363489151000977,
|
| 983 |
+
"learning_rate": 9.99062806601356e-06,
|
| 984 |
+
"loss": 0.2071,
|
| 985 |
+
"step": 1280
|
| 986 |
+
},
|
| 987 |
+
{
|
| 988 |
+
"epoch": 1.2089971883786317,
|
| 989 |
+
"grad_norm": 1.1078994274139404,
|
| 990 |
+
"learning_rate": 9.990554847779291e-06,
|
| 991 |
+
"loss": 0.1984,
|
| 992 |
+
"step": 1290
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 1.218369259606373,
|
| 996 |
+
"grad_norm": 1.4608142375946045,
|
| 997 |
+
"learning_rate": 9.990481629545022e-06,
|
| 998 |
+
"loss": 0.1926,
|
| 999 |
+
"step": 1300
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 1.2277413308341143,
|
| 1003 |
+
"grad_norm": 1.5290361642837524,
|
| 1004 |
+
"learning_rate": 9.990408411310753e-06,
|
| 1005 |
+
"loss": 0.1935,
|
| 1006 |
+
"step": 1310
|
| 1007 |
+
},
|
| 1008 |
+
{
|
| 1009 |
+
"epoch": 1.2371134020618557,
|
| 1010 |
+
"grad_norm": 1.09344482421875,
|
| 1011 |
+
"learning_rate": 9.990335193076485e-06,
|
| 1012 |
+
"loss": 0.2026,
|
| 1013 |
+
"step": 1320
|
| 1014 |
+
},
|
| 1015 |
+
{
|
| 1016 |
+
"epoch": 1.246485473289597,
|
| 1017 |
+
"grad_norm": 1.5567576885223389,
|
| 1018 |
+
"learning_rate": 9.990261974842216e-06,
|
| 1019 |
+
"loss": 0.1968,
|
| 1020 |
+
"step": 1330
|
| 1021 |
+
},
|
| 1022 |
+
{
|
| 1023 |
+
"epoch": 1.2558575445173383,
|
| 1024 |
+
"grad_norm": 1.243221402168274,
|
| 1025 |
+
"learning_rate": 9.990188756607947e-06,
|
| 1026 |
+
"loss": 0.1859,
|
| 1027 |
+
"step": 1340
|
| 1028 |
+
},
|
| 1029 |
+
{
|
| 1030 |
+
"epoch": 1.2652296157450795,
|
| 1031 |
+
"grad_norm": 1.5287493467330933,
|
| 1032 |
+
"learning_rate": 9.990115538373678e-06,
|
| 1033 |
+
"loss": 0.2067,
|
| 1034 |
+
"step": 1350
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 1.274601686972821,
|
| 1038 |
+
"grad_norm": 1.1587677001953125,
|
| 1039 |
+
"learning_rate": 9.990042320139408e-06,
|
| 1040 |
+
"loss": 0.1848,
|
| 1041 |
+
"step": 1360
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 1.2839737582005624,
|
| 1045 |
+
"grad_norm": 1.3521069288253784,
|
| 1046 |
+
"learning_rate": 9.989969101905139e-06,
|
| 1047 |
+
"loss": 0.1975,
|
| 1048 |
+
"step": 1370
|
| 1049 |
+
},
|
| 1050 |
+
{
|
| 1051 |
+
"epoch": 1.2933458294283038,
|
| 1052 |
+
"grad_norm": 1.1655584573745728,
|
| 1053 |
+
"learning_rate": 9.98989588367087e-06,
|
| 1054 |
+
"loss": 0.1963,
|
| 1055 |
+
"step": 1380
|
| 1056 |
+
},
|
| 1057 |
+
{
|
| 1058 |
+
"epoch": 1.302717900656045,
|
| 1059 |
+
"grad_norm": 1.1636890172958374,
|
| 1060 |
+
"learning_rate": 9.989822665436602e-06,
|
| 1061 |
+
"loss": 0.1768,
|
| 1062 |
+
"step": 1390
|
| 1063 |
+
},
|
| 1064 |
+
{
|
| 1065 |
+
"epoch": 1.3120899718837864,
|
| 1066 |
+
"grad_norm": 1.3106030225753784,
|
| 1067 |
+
"learning_rate": 9.989749447202333e-06,
|
| 1068 |
+
"loss": 0.1918,
|
| 1069 |
+
"step": 1400
|
| 1070 |
+
},
|
| 1071 |
+
{
|
| 1072 |
+
"epoch": 1.3214620431115276,
|
| 1073 |
+
"grad_norm": 1.314274787902832,
|
| 1074 |
+
"learning_rate": 9.989676228968062e-06,
|
| 1075 |
+
"loss": 0.1733,
|
| 1076 |
+
"step": 1410
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 1.330834114339269,
|
| 1080 |
+
"grad_norm": 1.646234393119812,
|
| 1081 |
+
"learning_rate": 9.989603010733795e-06,
|
| 1082 |
+
"loss": 0.1797,
|
| 1083 |
+
"step": 1420
|
| 1084 |
+
},
|
| 1085 |
+
{
|
| 1086 |
+
"epoch": 1.3402061855670104,
|
| 1087 |
+
"grad_norm": 1.3321646451950073,
|
| 1088 |
+
"learning_rate": 9.989529792499525e-06,
|
| 1089 |
+
"loss": 0.1726,
|
| 1090 |
+
"step": 1430
|
| 1091 |
+
},
|
| 1092 |
+
{
|
| 1093 |
+
"epoch": 1.3495782567947516,
|
| 1094 |
+
"grad_norm": 1.3959871530532837,
|
| 1095 |
+
"learning_rate": 9.989456574265256e-06,
|
| 1096 |
+
"loss": 0.1889,
|
| 1097 |
+
"step": 1440
|
| 1098 |
+
},
|
| 1099 |
+
{
|
| 1100 |
+
"epoch": 1.358950328022493,
|
| 1101 |
+
"grad_norm": 1.1790053844451904,
|
| 1102 |
+
"learning_rate": 9.989383356030987e-06,
|
| 1103 |
+
"loss": 0.1779,
|
| 1104 |
+
"step": 1450
|
| 1105 |
+
},
|
| 1106 |
+
{
|
| 1107 |
+
"epoch": 1.3683223992502342,
|
| 1108 |
+
"grad_norm": 1.7612881660461426,
|
| 1109 |
+
"learning_rate": 9.989310137796718e-06,
|
| 1110 |
+
"loss": 0.1834,
|
| 1111 |
+
"step": 1460
|
| 1112 |
+
},
|
| 1113 |
+
{
|
| 1114 |
+
"epoch": 1.3776944704779757,
|
| 1115 |
+
"grad_norm": 1.2366232872009277,
|
| 1116 |
+
"learning_rate": 9.989236919562448e-06,
|
| 1117 |
+
"loss": 0.1996,
|
| 1118 |
+
"step": 1470
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 1.387066541705717,
|
| 1122 |
+
"grad_norm": 1.550465703010559,
|
| 1123 |
+
"learning_rate": 9.989163701328179e-06,
|
| 1124 |
+
"loss": 0.1991,
|
| 1125 |
+
"step": 1480
|
| 1126 |
+
},
|
| 1127 |
+
{
|
| 1128 |
+
"epoch": 1.3964386129334583,
|
| 1129 |
+
"grad_norm": 1.2935107946395874,
|
| 1130 |
+
"learning_rate": 9.98909048309391e-06,
|
| 1131 |
+
"loss": 0.1956,
|
| 1132 |
+
"step": 1490
|
| 1133 |
+
},
|
| 1134 |
+
{
|
| 1135 |
+
"epoch": 1.4058106841611997,
|
| 1136 |
+
"grad_norm": 0.9709776639938354,
|
| 1137 |
+
"learning_rate": 9.989017264859642e-06,
|
| 1138 |
+
"loss": 0.1872,
|
| 1139 |
+
"step": 1500
|
| 1140 |
+
},
|
| 1141 |
+
{
|
| 1142 |
+
"epoch": 1.4058106841611997,
|
| 1143 |
+
"eval_loss": 0.07902642339468002,
|
| 1144 |
+
"eval_pearson_cosine": 0.7684531211853027,
|
| 1145 |
+
"eval_pearson_dot": 0.6580111980438232,
|
| 1146 |
+
"eval_pearson_euclidean": 0.768983006477356,
|
| 1147 |
+
"eval_pearson_manhattan": 0.7692690491676331,
|
| 1148 |
+
"eval_runtime": 23.5462,
|
| 1149 |
+
"eval_samples_per_second": 63.704,
|
| 1150 |
+
"eval_spearman_cosine": 0.7777241764238451,
|
| 1151 |
+
"eval_spearman_dot": 0.6568945327389543,
|
| 1152 |
+
"eval_spearman_euclidean": 0.7752386276211667,
|
| 1153 |
+
"eval_spearman_manhattan": 0.7755204438878311,
|
| 1154 |
+
"eval_steps_per_second": 7.984,
|
| 1155 |
+
"step": 1500
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 1.415182755388941,
|
| 1159 |
+
"grad_norm": 1.5001726150512695,
|
| 1160 |
+
"learning_rate": 9.988944046625373e-06,
|
| 1161 |
+
"loss": 0.2094,
|
| 1162 |
+
"step": 1510
|
| 1163 |
+
},
|
| 1164 |
+
{
|
| 1165 |
+
"epoch": 1.4245548266166823,
|
| 1166 |
+
"grad_norm": 1.1697657108306885,
|
| 1167 |
+
"learning_rate": 9.988870828391102e-06,
|
| 1168 |
+
"loss": 0.1862,
|
| 1169 |
+
"step": 1520
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"epoch": 1.4339268978444237,
|
| 1173 |
+
"grad_norm": 1.3496723175048828,
|
| 1174 |
+
"learning_rate": 9.988797610156834e-06,
|
| 1175 |
+
"loss": 0.1863,
|
| 1176 |
+
"step": 1530
|
| 1177 |
+
},
|
| 1178 |
+
{
|
| 1179 |
+
"epoch": 1.443298969072165,
|
| 1180 |
+
"grad_norm": 1.3314088582992554,
|
| 1181 |
+
"learning_rate": 9.988724391922565e-06,
|
| 1182 |
+
"loss": 0.1809,
|
| 1183 |
+
"step": 1540
|
| 1184 |
+
},
|
| 1185 |
+
{
|
| 1186 |
+
"epoch": 1.4526710402999063,
|
| 1187 |
+
"grad_norm": 1.2966681718826294,
|
| 1188 |
+
"learning_rate": 9.988651173688296e-06,
|
| 1189 |
+
"loss": 0.1799,
|
| 1190 |
+
"step": 1550
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 1.4620431115276475,
|
| 1194 |
+
"grad_norm": 1.141318917274475,
|
| 1195 |
+
"learning_rate": 9.988577955454027e-06,
|
| 1196 |
+
"loss": 0.1983,
|
| 1197 |
+
"step": 1560
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 1.471415182755389,
|
| 1201 |
+
"grad_norm": 1.1170287132263184,
|
| 1202 |
+
"learning_rate": 9.98850473721976e-06,
|
| 1203 |
+
"loss": 0.1823,
|
| 1204 |
+
"step": 1570
|
| 1205 |
+
},
|
| 1206 |
+
{
|
| 1207 |
+
"epoch": 1.4807872539831304,
|
| 1208 |
+
"grad_norm": 1.4531837701797485,
|
| 1209 |
+
"learning_rate": 9.988431518985488e-06,
|
| 1210 |
+
"loss": 0.1693,
|
| 1211 |
+
"step": 1580
|
| 1212 |
+
},
|
| 1213 |
+
{
|
| 1214 |
+
"epoch": 1.4901593252108716,
|
| 1215 |
+
"grad_norm": 1.5249556303024292,
|
| 1216 |
+
"learning_rate": 9.988358300751219e-06,
|
| 1217 |
+
"loss": 0.2014,
|
| 1218 |
+
"step": 1590
|
| 1219 |
+
},
|
| 1220 |
+
{
|
| 1221 |
+
"epoch": 1.499531396438613,
|
| 1222 |
+
"grad_norm": 1.319170594215393,
|
| 1223 |
+
"learning_rate": 9.988285082516951e-06,
|
| 1224 |
+
"loss": 0.1841,
|
| 1225 |
+
"step": 1600
|
| 1226 |
+
},
|
| 1227 |
+
{
|
| 1228 |
+
"epoch": 1.5089034676663542,
|
| 1229 |
+
"grad_norm": 1.2907928228378296,
|
| 1230 |
+
"learning_rate": 9.988211864282682e-06,
|
| 1231 |
+
"loss": 0.1778,
|
| 1232 |
+
"step": 1610
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 1.5182755388940956,
|
| 1236 |
+
"grad_norm": 1.170284628868103,
|
| 1237 |
+
"learning_rate": 9.988138646048413e-06,
|
| 1238 |
+
"loss": 0.1668,
|
| 1239 |
+
"step": 1620
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 1.527647610121837,
|
| 1243 |
+
"grad_norm": 1.4182498455047607,
|
| 1244 |
+
"learning_rate": 9.988065427814144e-06,
|
| 1245 |
+
"loss": 0.1968,
|
| 1246 |
+
"step": 1630
|
| 1247 |
+
},
|
| 1248 |
+
{
|
| 1249 |
+
"epoch": 1.5370196813495782,
|
| 1250 |
+
"grad_norm": 1.3137290477752686,
|
| 1251 |
+
"learning_rate": 9.987992209579874e-06,
|
| 1252 |
+
"loss": 0.1734,
|
| 1253 |
+
"step": 1640
|
| 1254 |
+
},
|
| 1255 |
+
{
|
| 1256 |
+
"epoch": 1.5463917525773194,
|
| 1257 |
+
"grad_norm": 1.458721399307251,
|
| 1258 |
+
"learning_rate": 9.987918991345605e-06,
|
| 1259 |
+
"loss": 0.209,
|
| 1260 |
+
"step": 1650
|
| 1261 |
+
},
|
| 1262 |
+
{
|
| 1263 |
+
"epoch": 1.5557638238050608,
|
| 1264 |
+
"grad_norm": 1.1368082761764526,
|
| 1265 |
+
"learning_rate": 9.987845773111336e-06,
|
| 1266 |
+
"loss": 0.1831,
|
| 1267 |
+
"step": 1660
|
| 1268 |
+
},
|
| 1269 |
+
{
|
| 1270 |
+
"epoch": 1.5651358950328023,
|
| 1271 |
+
"grad_norm": 1.0743663311004639,
|
| 1272 |
+
"learning_rate": 9.987772554877068e-06,
|
| 1273 |
+
"loss": 0.1883,
|
| 1274 |
+
"step": 1670
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 1.5745079662605437,
|
| 1278 |
+
"grad_norm": 1.4294681549072266,
|
| 1279 |
+
"learning_rate": 9.987699336642799e-06,
|
| 1280 |
+
"loss": 0.1851,
|
| 1281 |
+
"step": 1680
|
| 1282 |
+
},
|
| 1283 |
+
{
|
| 1284 |
+
"epoch": 1.5838800374882849,
|
| 1285 |
+
"grad_norm": 1.0537577867507935,
|
| 1286 |
+
"learning_rate": 9.987626118408528e-06,
|
| 1287 |
+
"loss": 0.1818,
|
| 1288 |
+
"step": 1690
|
| 1289 |
+
},
|
| 1290 |
+
{
|
| 1291 |
+
"epoch": 1.5932521087160263,
|
| 1292 |
+
"grad_norm": 1.3930073976516724,
|
| 1293 |
+
"learning_rate": 9.98755290017426e-06,
|
| 1294 |
+
"loss": 0.1876,
|
| 1295 |
+
"step": 1700
|
| 1296 |
+
},
|
| 1297 |
+
{
|
| 1298 |
+
"epoch": 1.6026241799437675,
|
| 1299 |
+
"grad_norm": 1.3290959596633911,
|
| 1300 |
+
"learning_rate": 9.987479681939991e-06,
|
| 1301 |
+
"loss": 0.1777,
|
| 1302 |
+
"step": 1710
|
| 1303 |
+
},
|
| 1304 |
+
{
|
| 1305 |
+
"epoch": 1.611996251171509,
|
| 1306 |
+
"grad_norm": 1.3895900249481201,
|
| 1307 |
+
"learning_rate": 9.987406463705722e-06,
|
| 1308 |
+
"loss": 0.1728,
|
| 1309 |
+
"step": 1720
|
| 1310 |
+
},
|
| 1311 |
+
{
|
| 1312 |
+
"epoch": 1.6213683223992503,
|
| 1313 |
+
"grad_norm": 1.336679220199585,
|
| 1314 |
+
"learning_rate": 9.987333245471453e-06,
|
| 1315 |
+
"loss": 0.202,
|
| 1316 |
+
"step": 1730
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 1.6307403936269915,
|
| 1320 |
+
"grad_norm": 1.4338617324829102,
|
| 1321 |
+
"learning_rate": 9.987260027237184e-06,
|
| 1322 |
+
"loss": 0.1745,
|
| 1323 |
+
"step": 1740
|
| 1324 |
+
},
|
| 1325 |
+
{
|
| 1326 |
+
"epoch": 1.640112464854733,
|
| 1327 |
+
"grad_norm": 1.1854125261306763,
|
| 1328 |
+
"learning_rate": 9.987186809002914e-06,
|
| 1329 |
+
"loss": 0.1628,
|
| 1330 |
+
"step": 1750
|
| 1331 |
+
},
|
| 1332 |
+
{
|
| 1333 |
+
"epoch": 1.640112464854733,
|
| 1334 |
+
"eval_loss": 0.07191870361566544,
|
| 1335 |
+
"eval_pearson_cosine": 0.7651911973953247,
|
| 1336 |
+
"eval_pearson_dot": 0.6584045886993408,
|
| 1337 |
+
"eval_pearson_euclidean": 0.7615811228752136,
|
| 1338 |
+
"eval_pearson_manhattan": 0.7618914842605591,
|
| 1339 |
+
"eval_runtime": 22.2177,
|
| 1340 |
+
"eval_samples_per_second": 67.514,
|
| 1341 |
+
"eval_spearman_cosine": 0.7733826669765486,
|
| 1342 |
+
"eval_spearman_dot": 0.6574446699366203,
|
| 1343 |
+
"eval_spearman_euclidean": 0.7678793093449918,
|
| 1344 |
+
"eval_spearman_manhattan": 0.7684997409854779,
|
| 1345 |
+
"eval_steps_per_second": 8.462,
|
| 1346 |
+
"step": 1750
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 1.6494845360824741,
|
| 1350 |
+
"grad_norm": 1.468126654624939,
|
| 1351 |
+
"learning_rate": 9.987113590768645e-06,
|
| 1352 |
+
"loss": 0.1714,
|
| 1353 |
+
"step": 1760
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 1.6588566073102156,
|
| 1357 |
+
"grad_norm": 1.3639568090438843,
|
| 1358 |
+
"learning_rate": 9.987040372534378e-06,
|
| 1359 |
+
"loss": 0.1839,
|
| 1360 |
+
"step": 1770
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 1.668228678537957,
|
| 1364 |
+
"grad_norm": 1.2494312524795532,
|
| 1365 |
+
"learning_rate": 9.986967154300108e-06,
|
| 1366 |
+
"loss": 0.1753,
|
| 1367 |
+
"step": 1780
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 1.6776007497656982,
|
| 1371 |
+
"grad_norm": 1.2897909879684448,
|
| 1372 |
+
"learning_rate": 9.986893936065839e-06,
|
| 1373 |
+
"loss": 0.1704,
|
| 1374 |
+
"step": 1790
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 1.6869728209934396,
|
| 1378 |
+
"grad_norm": 1.413866400718689,
|
| 1379 |
+
"learning_rate": 9.98682071783157e-06,
|
| 1380 |
+
"loss": 0.1868,
|
| 1381 |
+
"step": 1800
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 1.6963448922211808,
|
| 1385 |
+
"grad_norm": 1.093849778175354,
|
| 1386 |
+
"learning_rate": 9.9867474995973e-06,
|
| 1387 |
+
"loss": 0.1889,
|
| 1388 |
+
"step": 1810
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 1.7057169634489222,
|
| 1392 |
+
"grad_norm": 1.3857814073562622,
|
| 1393 |
+
"learning_rate": 9.986674281363031e-06,
|
| 1394 |
+
"loss": 0.1818,
|
| 1395 |
+
"step": 1820
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 1.7150890346766636,
|
| 1399 |
+
"grad_norm": 1.3772344589233398,
|
| 1400 |
+
"learning_rate": 9.986601063128762e-06,
|
| 1401 |
+
"loss": 0.1683,
|
| 1402 |
+
"step": 1830
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 1.7244611059044048,
|
| 1406 |
+
"grad_norm": 1.3299206495285034,
|
| 1407 |
+
"learning_rate": 9.986527844894493e-06,
|
| 1408 |
+
"loss": 0.1865,
|
| 1409 |
+
"step": 1840
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 1.7338331771321462,
|
| 1413 |
+
"grad_norm": 1.3139843940734863,
|
| 1414 |
+
"learning_rate": 9.986454626660225e-06,
|
| 1415 |
+
"loss": 0.169,
|
| 1416 |
+
"step": 1850
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 1.7432052483598874,
|
| 1420 |
+
"grad_norm": 1.3562296628952026,
|
| 1421 |
+
"learning_rate": 9.986381408425954e-06,
|
| 1422 |
+
"loss": 0.2012,
|
| 1423 |
+
"step": 1860
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 1.7525773195876289,
|
| 1427 |
+
"grad_norm": 1.2332826852798462,
|
| 1428 |
+
"learning_rate": 9.986308190191685e-06,
|
| 1429 |
+
"loss": 0.1877,
|
| 1430 |
+
"step": 1870
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 1.7619493908153703,
|
| 1434 |
+
"grad_norm": 1.083622932434082,
|
| 1435 |
+
"learning_rate": 9.986234971957418e-06,
|
| 1436 |
+
"loss": 0.2026,
|
| 1437 |
+
"step": 1880
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 1.7713214620431117,
|
| 1441 |
+
"grad_norm": 1.6391818523406982,
|
| 1442 |
+
"learning_rate": 9.986161753723148e-06,
|
| 1443 |
+
"loss": 0.1902,
|
| 1444 |
+
"step": 1890
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 1.780693533270853,
|
| 1448 |
+
"grad_norm": 1.0985593795776367,
|
| 1449 |
+
"learning_rate": 9.986088535488879e-06,
|
| 1450 |
+
"loss": 0.1845,
|
| 1451 |
+
"step": 1900
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 1.790065604498594,
|
| 1455 |
+
"grad_norm": 1.609025001525879,
|
| 1456 |
+
"learning_rate": 9.98601531725461e-06,
|
| 1457 |
+
"loss": 0.1939,
|
| 1458 |
+
"step": 1910
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 1.7994376757263355,
|
| 1462 |
+
"grad_norm": 1.0637205839157104,
|
| 1463 |
+
"learning_rate": 9.98594209902034e-06,
|
| 1464 |
+
"loss": 0.1775,
|
| 1465 |
+
"step": 1920
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 1.808809746954077,
|
| 1469 |
+
"grad_norm": 1.159469723701477,
|
| 1470 |
+
"learning_rate": 9.985868880786071e-06,
|
| 1471 |
+
"loss": 0.161,
|
| 1472 |
+
"step": 1930
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 1.8181818181818183,
|
| 1476 |
+
"grad_norm": 1.1251918077468872,
|
| 1477 |
+
"learning_rate": 9.985795662551802e-06,
|
| 1478 |
+
"loss": 0.1965,
|
| 1479 |
+
"step": 1940
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 1.8275538894095595,
|
| 1483 |
+
"grad_norm": 1.3804899454116821,
|
| 1484 |
+
"learning_rate": 9.985722444317534e-06,
|
| 1485 |
+
"loss": 0.1768,
|
| 1486 |
+
"step": 1950
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 1.8369259606373007,
|
| 1490 |
+
"grad_norm": 1.194275140762329,
|
| 1491 |
+
"learning_rate": 9.985649226083265e-06,
|
| 1492 |
+
"loss": 0.1782,
|
| 1493 |
+
"step": 1960
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 1.8462980318650422,
|
| 1497 |
+
"grad_norm": 1.5173845291137695,
|
| 1498 |
+
"learning_rate": 9.985576007848996e-06,
|
| 1499 |
+
"loss": 0.193,
|
| 1500 |
+
"step": 1970
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 1.8556701030927836,
|
| 1504 |
+
"grad_norm": 1.7733920812606812,
|
| 1505 |
+
"learning_rate": 9.985502789614727e-06,
|
| 1506 |
+
"loss": 0.1804,
|
| 1507 |
+
"step": 1980
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 1.865042174320525,
|
| 1511 |
+
"grad_norm": 1.1430355310440063,
|
| 1512 |
+
"learning_rate": 9.985429571380457e-06,
|
| 1513 |
+
"loss": 0.1869,
|
| 1514 |
+
"step": 1990
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 1.8744142455482662,
|
| 1518 |
+
"grad_norm": 1.3633067607879639,
|
| 1519 |
+
"learning_rate": 9.985356353146188e-06,
|
| 1520 |
+
"loss": 0.1983,
|
| 1521 |
+
"step": 2000
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 1.8744142455482662,
|
| 1525 |
+
"eval_loss": 0.07371454685926437,
|
| 1526 |
+
"eval_pearson_cosine": 0.7772414684295654,
|
| 1527 |
+
"eval_pearson_dot": 0.660416841506958,
|
| 1528 |
+
"eval_pearson_euclidean": 0.7648824453353882,
|
| 1529 |
+
"eval_pearson_manhattan": 0.7654331922531128,
|
| 1530 |
+
"eval_runtime": 22.1973,
|
| 1531 |
+
"eval_samples_per_second": 67.576,
|
| 1532 |
+
"eval_spearman_cosine": 0.7863920785446639,
|
| 1533 |
+
"eval_spearman_dot": 0.6607574545837009,
|
| 1534 |
+
"eval_spearman_euclidean": 0.7740511645049805,
|
| 1535 |
+
"eval_spearman_manhattan": 0.7747616492851076,
|
| 1536 |
+
"eval_steps_per_second": 8.47,
|
| 1537 |
+
"step": 2000
|
| 1538 |
}
|
| 1539 |
],
|
| 1540 |
"logging_steps": 10,
|