Training in progress, step 7000, 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 sha256:320c1a6e527bda174c669c1f951ba0d8511688825b12496cdd957e88966c44d9
|
| 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 sha256:57f4bc2a3c4abe8dda39abccfe59caa42f3c3826a1837f4bf0c5399a3526b447
|
| 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 sha256:6677d2f94c1eeef9fd86043f9720c0ec154224510969ab790d1983791f375b88
|
| 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:c4c2eafdb8010cffd636a540805e21c2ed7a900fa295f85b31d51d1a42084ef2
|
| 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 sha256:243c3ab92fec11f8d6189a1ee0c6563a7d0fa0286f6142dbd247bcce148d357d
|
| 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 sha256:12c16967790d71b2e48b6adefaa270162917f0f86335d85480de9fb6860c7492
|
| 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 sha256:369fdbde0f14abf383325d06785d666884716d0228d7dc8c5e50c13cfb424b27
|
| 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:bb8ed41d9a3d876261e27e3c4c409c9b564db5930687ae4ec83f98c5023be0a7
|
| 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 sha256:5a76fd0f99241be8ebb729e7e60a6572e44b68c9f795cc3e8947a8ca4fa56c56
|
| 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 sha256:45c51dab87a7b7a65f8d23876cdf72e77eeca4ff50002d44cd0db1ecc3a43a63
|
| 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:d2dc388ca3bbc2159f39c9ce16c2f601ff32a9453701746d70cc6d36087ab16b
|
| 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,
|
|
@@ -4591,6 +4591,770 @@
|
|
| 4591 |
"eval_spearman_manhattan": 0.7343750357819732,
|
| 4592 |
"eval_steps_per_second": 7.36,
|
| 4593 |
"step": 6000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4594 |
}
|
| 4595 |
],
|
| 4596 |
"logging_steps": 10,
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 6.560449859418932,
|
| 5 |
"eval_steps": 250,
|
| 6 |
+
"global_step": 7000,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 4591 |
"eval_spearman_manhattan": 0.7343750357819732,
|
| 4592 |
"eval_steps_per_second": 7.36,
|
| 4593 |
"step": 6000
|
| 4594 |
+
},
|
| 4595 |
+
{
|
| 4596 |
+
"epoch": 5.63261480787254,
|
| 4597 |
+
"grad_norm": 1.2122074365615845,
|
| 4598 |
+
"learning_rate": 9.955995841204294e-06,
|
| 4599 |
+
"loss": 0.0665,
|
| 4600 |
+
"step": 6010
|
| 4601 |
+
},
|
| 4602 |
+
{
|
| 4603 |
+
"epoch": 5.641986879100282,
|
| 4604 |
+
"grad_norm": 1.7832310199737549,
|
| 4605 |
+
"learning_rate": 9.955922622970026e-06,
|
| 4606 |
+
"loss": 0.063,
|
| 4607 |
+
"step": 6020
|
| 4608 |
+
},
|
| 4609 |
+
{
|
| 4610 |
+
"epoch": 5.651358950328023,
|
| 4611 |
+
"grad_norm": 1.1854170560836792,
|
| 4612 |
+
"learning_rate": 9.955849404735755e-06,
|
| 4613 |
+
"loss": 0.0573,
|
| 4614 |
+
"step": 6030
|
| 4615 |
+
},
|
| 4616 |
+
{
|
| 4617 |
+
"epoch": 5.660731021555764,
|
| 4618 |
+
"grad_norm": 1.6633968353271484,
|
| 4619 |
+
"learning_rate": 9.955776186501486e-06,
|
| 4620 |
+
"loss": 0.0549,
|
| 4621 |
+
"step": 6040
|
| 4622 |
+
},
|
| 4623 |
+
{
|
| 4624 |
+
"epoch": 5.670103092783505,
|
| 4625 |
+
"grad_norm": 1.31834077835083,
|
| 4626 |
+
"learning_rate": 9.955702968267219e-06,
|
| 4627 |
+
"loss": 0.0478,
|
| 4628 |
+
"step": 6050
|
| 4629 |
+
},
|
| 4630 |
+
{
|
| 4631 |
+
"epoch": 5.679475164011246,
|
| 4632 |
+
"grad_norm": 0.8284873962402344,
|
| 4633 |
+
"learning_rate": 9.95562975003295e-06,
|
| 4634 |
+
"loss": 0.0639,
|
| 4635 |
+
"step": 6060
|
| 4636 |
+
},
|
| 4637 |
+
{
|
| 4638 |
+
"epoch": 5.688847235238988,
|
| 4639 |
+
"grad_norm": 1.2393404245376587,
|
| 4640 |
+
"learning_rate": 9.95555653179868e-06,
|
| 4641 |
+
"loss": 0.0593,
|
| 4642 |
+
"step": 6070
|
| 4643 |
+
},
|
| 4644 |
+
{
|
| 4645 |
+
"epoch": 5.698219306466729,
|
| 4646 |
+
"grad_norm": 1.5327643156051636,
|
| 4647 |
+
"learning_rate": 9.95548331356441e-06,
|
| 4648 |
+
"loss": 0.0644,
|
| 4649 |
+
"step": 6080
|
| 4650 |
+
},
|
| 4651 |
+
{
|
| 4652 |
+
"epoch": 5.70759137769447,
|
| 4653 |
+
"grad_norm": 1.8985389471054077,
|
| 4654 |
+
"learning_rate": 9.955410095330142e-06,
|
| 4655 |
+
"loss": 0.0646,
|
| 4656 |
+
"step": 6090
|
| 4657 |
+
},
|
| 4658 |
+
{
|
| 4659 |
+
"epoch": 5.716963448922212,
|
| 4660 |
+
"grad_norm": 1.5896059274673462,
|
| 4661 |
+
"learning_rate": 9.955336877095872e-06,
|
| 4662 |
+
"loss": 0.0716,
|
| 4663 |
+
"step": 6100
|
| 4664 |
+
},
|
| 4665 |
+
{
|
| 4666 |
+
"epoch": 5.726335520149953,
|
| 4667 |
+
"grad_norm": 1.21624755859375,
|
| 4668 |
+
"learning_rate": 9.955263658861603e-06,
|
| 4669 |
+
"loss": 0.0559,
|
| 4670 |
+
"step": 6110
|
| 4671 |
+
},
|
| 4672 |
+
{
|
| 4673 |
+
"epoch": 5.735707591377695,
|
| 4674 |
+
"grad_norm": 1.3084664344787598,
|
| 4675 |
+
"learning_rate": 9.955190440627336e-06,
|
| 4676 |
+
"loss": 0.065,
|
| 4677 |
+
"step": 6120
|
| 4678 |
+
},
|
| 4679 |
+
{
|
| 4680 |
+
"epoch": 5.745079662605436,
|
| 4681 |
+
"grad_norm": 0.9755469560623169,
|
| 4682 |
+
"learning_rate": 9.955117222393066e-06,
|
| 4683 |
+
"loss": 0.0601,
|
| 4684 |
+
"step": 6130
|
| 4685 |
+
},
|
| 4686 |
+
{
|
| 4687 |
+
"epoch": 5.754451733833177,
|
| 4688 |
+
"grad_norm": 1.1662402153015137,
|
| 4689 |
+
"learning_rate": 9.955044004158797e-06,
|
| 4690 |
+
"loss": 0.0588,
|
| 4691 |
+
"step": 6140
|
| 4692 |
+
},
|
| 4693 |
+
{
|
| 4694 |
+
"epoch": 5.763823805060919,
|
| 4695 |
+
"grad_norm": 1.313323974609375,
|
| 4696 |
+
"learning_rate": 9.954970785924528e-06,
|
| 4697 |
+
"loss": 0.0667,
|
| 4698 |
+
"step": 6150
|
| 4699 |
+
},
|
| 4700 |
+
{
|
| 4701 |
+
"epoch": 5.77319587628866,
|
| 4702 |
+
"grad_norm": 1.4725874662399292,
|
| 4703 |
+
"learning_rate": 9.954897567690259e-06,
|
| 4704 |
+
"loss": 0.0619,
|
| 4705 |
+
"step": 6160
|
| 4706 |
+
},
|
| 4707 |
+
{
|
| 4708 |
+
"epoch": 5.782567947516402,
|
| 4709 |
+
"grad_norm": 1.3176454305648804,
|
| 4710 |
+
"learning_rate": 9.95482434945599e-06,
|
| 4711 |
+
"loss": 0.056,
|
| 4712 |
+
"step": 6170
|
| 4713 |
+
},
|
| 4714 |
+
{
|
| 4715 |
+
"epoch": 5.7919400187441425,
|
| 4716 |
+
"grad_norm": 1.0566222667694092,
|
| 4717 |
+
"learning_rate": 9.95475113122172e-06,
|
| 4718 |
+
"loss": 0.0587,
|
| 4719 |
+
"step": 6180
|
| 4720 |
+
},
|
| 4721 |
+
{
|
| 4722 |
+
"epoch": 5.8013120899718835,
|
| 4723 |
+
"grad_norm": 1.0623878240585327,
|
| 4724 |
+
"learning_rate": 9.95467791298745e-06,
|
| 4725 |
+
"loss": 0.0591,
|
| 4726 |
+
"step": 6190
|
| 4727 |
+
},
|
| 4728 |
+
{
|
| 4729 |
+
"epoch": 5.810684161199625,
|
| 4730 |
+
"grad_norm": 1.6217368841171265,
|
| 4731 |
+
"learning_rate": 9.954604694753183e-06,
|
| 4732 |
+
"loss": 0.0536,
|
| 4733 |
+
"step": 6200
|
| 4734 |
+
},
|
| 4735 |
+
{
|
| 4736 |
+
"epoch": 5.820056232427366,
|
| 4737 |
+
"grad_norm": 1.2574353218078613,
|
| 4738 |
+
"learning_rate": 9.954531476518912e-06,
|
| 4739 |
+
"loss": 0.0552,
|
| 4740 |
+
"step": 6210
|
| 4741 |
+
},
|
| 4742 |
+
{
|
| 4743 |
+
"epoch": 5.829428303655108,
|
| 4744 |
+
"grad_norm": 1.2605924606323242,
|
| 4745 |
+
"learning_rate": 9.954458258284643e-06,
|
| 4746 |
+
"loss": 0.0669,
|
| 4747 |
+
"step": 6220
|
| 4748 |
+
},
|
| 4749 |
+
{
|
| 4750 |
+
"epoch": 5.838800374882849,
|
| 4751 |
+
"grad_norm": 1.8283051252365112,
|
| 4752 |
+
"learning_rate": 9.954385040050375e-06,
|
| 4753 |
+
"loss": 0.0631,
|
| 4754 |
+
"step": 6230
|
| 4755 |
+
},
|
| 4756 |
+
{
|
| 4757 |
+
"epoch": 5.84817244611059,
|
| 4758 |
+
"grad_norm": 1.2457951307296753,
|
| 4759 |
+
"learning_rate": 9.954311821816106e-06,
|
| 4760 |
+
"loss": 0.0578,
|
| 4761 |
+
"step": 6240
|
| 4762 |
+
},
|
| 4763 |
+
{
|
| 4764 |
+
"epoch": 5.857544517338332,
|
| 4765 |
+
"grad_norm": 1.1618739366531372,
|
| 4766 |
+
"learning_rate": 9.954238603581837e-06,
|
| 4767 |
+
"loss": 0.0547,
|
| 4768 |
+
"step": 6250
|
| 4769 |
+
},
|
| 4770 |
+
{
|
| 4771 |
+
"epoch": 5.857544517338332,
|
| 4772 |
+
"eval_loss": 0.03839369863271713,
|
| 4773 |
+
"eval_pearson_cosine": 0.7663547396659851,
|
| 4774 |
+
"eval_pearson_dot": 0.7110079526901245,
|
| 4775 |
+
"eval_pearson_euclidean": 0.7369804978370667,
|
| 4776 |
+
"eval_pearson_manhattan": 0.738224983215332,
|
| 4777 |
+
"eval_runtime": 28.702,
|
| 4778 |
+
"eval_samples_per_second": 52.261,
|
| 4779 |
+
"eval_spearman_cosine": 0.766680322110213,
|
| 4780 |
+
"eval_spearman_dot": 0.7118792296635837,
|
| 4781 |
+
"eval_spearman_euclidean": 0.7420173359570077,
|
| 4782 |
+
"eval_spearman_manhattan": 0.7431811125331302,
|
| 4783 |
+
"eval_steps_per_second": 6.55,
|
| 4784 |
+
"step": 6250
|
| 4785 |
+
},
|
| 4786 |
+
{
|
| 4787 |
+
"epoch": 5.866916588566073,
|
| 4788 |
+
"grad_norm": 1.565491795539856,
|
| 4789 |
+
"learning_rate": 9.954165385347568e-06,
|
| 4790 |
+
"loss": 0.0634,
|
| 4791 |
+
"step": 6260
|
| 4792 |
+
},
|
| 4793 |
+
{
|
| 4794 |
+
"epoch": 5.876288659793815,
|
| 4795 |
+
"grad_norm": 1.412607192993164,
|
| 4796 |
+
"learning_rate": 9.954092167113298e-06,
|
| 4797 |
+
"loss": 0.0641,
|
| 4798 |
+
"step": 6270
|
| 4799 |
+
},
|
| 4800 |
+
{
|
| 4801 |
+
"epoch": 5.885660731021556,
|
| 4802 |
+
"grad_norm": 1.5475645065307617,
|
| 4803 |
+
"learning_rate": 9.95401894887903e-06,
|
| 4804 |
+
"loss": 0.058,
|
| 4805 |
+
"step": 6280
|
| 4806 |
+
},
|
| 4807 |
+
{
|
| 4808 |
+
"epoch": 5.895032802249297,
|
| 4809 |
+
"grad_norm": 1.6942791938781738,
|
| 4810 |
+
"learning_rate": 9.95394573064476e-06,
|
| 4811 |
+
"loss": 0.0668,
|
| 4812 |
+
"step": 6290
|
| 4813 |
+
},
|
| 4814 |
+
{
|
| 4815 |
+
"epoch": 5.904404873477039,
|
| 4816 |
+
"grad_norm": 1.286224603652954,
|
| 4817 |
+
"learning_rate": 9.953872512410492e-06,
|
| 4818 |
+
"loss": 0.058,
|
| 4819 |
+
"step": 6300
|
| 4820 |
+
},
|
| 4821 |
+
{
|
| 4822 |
+
"epoch": 5.91377694470478,
|
| 4823 |
+
"grad_norm": 1.5031893253326416,
|
| 4824 |
+
"learning_rate": 9.953799294176223e-06,
|
| 4825 |
+
"loss": 0.062,
|
| 4826 |
+
"step": 6310
|
| 4827 |
+
},
|
| 4828 |
+
{
|
| 4829 |
+
"epoch": 5.9231490159325215,
|
| 4830 |
+
"grad_norm": 1.416455864906311,
|
| 4831 |
+
"learning_rate": 9.953726075941952e-06,
|
| 4832 |
+
"loss": 0.0596,
|
| 4833 |
+
"step": 6320
|
| 4834 |
+
},
|
| 4835 |
+
{
|
| 4836 |
+
"epoch": 5.9325210871602625,
|
| 4837 |
+
"grad_norm": 1.3160662651062012,
|
| 4838 |
+
"learning_rate": 9.953652857707685e-06,
|
| 4839 |
+
"loss": 0.062,
|
| 4840 |
+
"step": 6330
|
| 4841 |
+
},
|
| 4842 |
+
{
|
| 4843 |
+
"epoch": 5.9418931583880035,
|
| 4844 |
+
"grad_norm": 0.9542105793952942,
|
| 4845 |
+
"learning_rate": 9.953579639473415e-06,
|
| 4846 |
+
"loss": 0.0645,
|
| 4847 |
+
"step": 6340
|
| 4848 |
+
},
|
| 4849 |
+
{
|
| 4850 |
+
"epoch": 5.951265229615745,
|
| 4851 |
+
"grad_norm": 1.4458489418029785,
|
| 4852 |
+
"learning_rate": 9.953506421239146e-06,
|
| 4853 |
+
"loss": 0.0563,
|
| 4854 |
+
"step": 6350
|
| 4855 |
+
},
|
| 4856 |
+
{
|
| 4857 |
+
"epoch": 5.960637300843486,
|
| 4858 |
+
"grad_norm": 1.0310072898864746,
|
| 4859 |
+
"learning_rate": 9.953433203004877e-06,
|
| 4860 |
+
"loss": 0.0567,
|
| 4861 |
+
"step": 6360
|
| 4862 |
+
},
|
| 4863 |
+
{
|
| 4864 |
+
"epoch": 5.970009372071228,
|
| 4865 |
+
"grad_norm": 1.4674971103668213,
|
| 4866 |
+
"learning_rate": 9.95335998477061e-06,
|
| 4867 |
+
"loss": 0.0579,
|
| 4868 |
+
"step": 6370
|
| 4869 |
+
},
|
| 4870 |
+
{
|
| 4871 |
+
"epoch": 5.979381443298969,
|
| 4872 |
+
"grad_norm": 1.229636311531067,
|
| 4873 |
+
"learning_rate": 9.953286766536338e-06,
|
| 4874 |
+
"loss": 0.0589,
|
| 4875 |
+
"step": 6380
|
| 4876 |
+
},
|
| 4877 |
+
{
|
| 4878 |
+
"epoch": 5.98875351452671,
|
| 4879 |
+
"grad_norm": 1.4654268026351929,
|
| 4880 |
+
"learning_rate": 9.95321354830207e-06,
|
| 4881 |
+
"loss": 0.0519,
|
| 4882 |
+
"step": 6390
|
| 4883 |
+
},
|
| 4884 |
+
{
|
| 4885 |
+
"epoch": 5.998125585754452,
|
| 4886 |
+
"grad_norm": 1.276367425918579,
|
| 4887 |
+
"learning_rate": 9.953140330067802e-06,
|
| 4888 |
+
"loss": 0.066,
|
| 4889 |
+
"step": 6400
|
| 4890 |
+
},
|
| 4891 |
+
{
|
| 4892 |
+
"epoch": 6.007497656982193,
|
| 4893 |
+
"grad_norm": 1.0710258483886719,
|
| 4894 |
+
"learning_rate": 9.953067111833532e-06,
|
| 4895 |
+
"loss": 0.0462,
|
| 4896 |
+
"step": 6410
|
| 4897 |
+
},
|
| 4898 |
+
{
|
| 4899 |
+
"epoch": 6.016869728209935,
|
| 4900 |
+
"grad_norm": 0.9316133856773376,
|
| 4901 |
+
"learning_rate": 9.952993893599263e-06,
|
| 4902 |
+
"loss": 0.044,
|
| 4903 |
+
"step": 6420
|
| 4904 |
+
},
|
| 4905 |
+
{
|
| 4906 |
+
"epoch": 6.026241799437676,
|
| 4907 |
+
"grad_norm": 0.8318607211112976,
|
| 4908 |
+
"learning_rate": 9.952920675364994e-06,
|
| 4909 |
+
"loss": 0.0399,
|
| 4910 |
+
"step": 6430
|
| 4911 |
+
},
|
| 4912 |
+
{
|
| 4913 |
+
"epoch": 6.035613870665417,
|
| 4914 |
+
"grad_norm": 0.9682859182357788,
|
| 4915 |
+
"learning_rate": 9.952847457130725e-06,
|
| 4916 |
+
"loss": 0.0371,
|
| 4917 |
+
"step": 6440
|
| 4918 |
+
},
|
| 4919 |
+
{
|
| 4920 |
+
"epoch": 6.044985941893159,
|
| 4921 |
+
"grad_norm": 0.8720560669898987,
|
| 4922 |
+
"learning_rate": 9.952774238896455e-06,
|
| 4923 |
+
"loss": 0.0453,
|
| 4924 |
+
"step": 6450
|
| 4925 |
+
},
|
| 4926 |
+
{
|
| 4927 |
+
"epoch": 6.0543580131209,
|
| 4928 |
+
"grad_norm": 0.7835734486579895,
|
| 4929 |
+
"learning_rate": 9.952701020662186e-06,
|
| 4930 |
+
"loss": 0.0475,
|
| 4931 |
+
"step": 6460
|
| 4932 |
+
},
|
| 4933 |
+
{
|
| 4934 |
+
"epoch": 6.0637300843486415,
|
| 4935 |
+
"grad_norm": 1.4373115301132202,
|
| 4936 |
+
"learning_rate": 9.952627802427917e-06,
|
| 4937 |
+
"loss": 0.0416,
|
| 4938 |
+
"step": 6470
|
| 4939 |
+
},
|
| 4940 |
+
{
|
| 4941 |
+
"epoch": 6.073102155576382,
|
| 4942 |
+
"grad_norm": 1.317517638206482,
|
| 4943 |
+
"learning_rate": 9.95255458419365e-06,
|
| 4944 |
+
"loss": 0.0425,
|
| 4945 |
+
"step": 6480
|
| 4946 |
+
},
|
| 4947 |
+
{
|
| 4948 |
+
"epoch": 6.082474226804123,
|
| 4949 |
+
"grad_norm": 1.1831910610198975,
|
| 4950 |
+
"learning_rate": 9.952481365959378e-06,
|
| 4951 |
+
"loss": 0.0471,
|
| 4952 |
+
"step": 6490
|
| 4953 |
+
},
|
| 4954 |
+
{
|
| 4955 |
+
"epoch": 6.091846298031865,
|
| 4956 |
+
"grad_norm": 1.0449994802474976,
|
| 4957 |
+
"learning_rate": 9.95240814772511e-06,
|
| 4958 |
+
"loss": 0.0476,
|
| 4959 |
+
"step": 6500
|
| 4960 |
+
},
|
| 4961 |
+
{
|
| 4962 |
+
"epoch": 6.091846298031865,
|
| 4963 |
+
"eval_loss": 0.03876839950680733,
|
| 4964 |
+
"eval_pearson_cosine": 0.7637665867805481,
|
| 4965 |
+
"eval_pearson_dot": 0.7007623910903931,
|
| 4966 |
+
"eval_pearson_euclidean": 0.7322614192962646,
|
| 4967 |
+
"eval_pearson_manhattan": 0.7338271141052246,
|
| 4968 |
+
"eval_runtime": 22.3296,
|
| 4969 |
+
"eval_samples_per_second": 67.175,
|
| 4970 |
+
"eval_spearman_cosine": 0.7641548541194557,
|
| 4971 |
+
"eval_spearman_dot": 0.7012776165056044,
|
| 4972 |
+
"eval_spearman_euclidean": 0.7377602855270703,
|
| 4973 |
+
"eval_spearman_manhattan": 0.73918298594716,
|
| 4974 |
+
"eval_steps_per_second": 8.419,
|
| 4975 |
+
"step": 6500
|
| 4976 |
+
},
|
| 4977 |
+
{
|
| 4978 |
+
"epoch": 6.101218369259606,
|
| 4979 |
+
"grad_norm": 0.7369022965431213,
|
| 4980 |
+
"learning_rate": 9.952334929490842e-06,
|
| 4981 |
+
"loss": 0.0364,
|
| 4982 |
+
"step": 6510
|
| 4983 |
+
},
|
| 4984 |
+
{
|
| 4985 |
+
"epoch": 6.110590440487348,
|
| 4986 |
+
"grad_norm": 0.8673484325408936,
|
| 4987 |
+
"learning_rate": 9.952261711256572e-06,
|
| 4988 |
+
"loss": 0.0498,
|
| 4989 |
+
"step": 6520
|
| 4990 |
+
},
|
| 4991 |
+
{
|
| 4992 |
+
"epoch": 6.119962511715089,
|
| 4993 |
+
"grad_norm": 1.5341424942016602,
|
| 4994 |
+
"learning_rate": 9.952188493022303e-06,
|
| 4995 |
+
"loss": 0.045,
|
| 4996 |
+
"step": 6530
|
| 4997 |
+
},
|
| 4998 |
+
{
|
| 4999 |
+
"epoch": 6.12933458294283,
|
| 5000 |
+
"grad_norm": 0.8899186253547668,
|
| 5001 |
+
"learning_rate": 9.952115274788034e-06,
|
| 5002 |
+
"loss": 0.0441,
|
| 5003 |
+
"step": 6540
|
| 5004 |
+
},
|
| 5005 |
+
{
|
| 5006 |
+
"epoch": 6.138706654170572,
|
| 5007 |
+
"grad_norm": 1.0708824396133423,
|
| 5008 |
+
"learning_rate": 9.952042056553765e-06,
|
| 5009 |
+
"loss": 0.0458,
|
| 5010 |
+
"step": 6550
|
| 5011 |
+
},
|
| 5012 |
+
{
|
| 5013 |
+
"epoch": 6.148078725398313,
|
| 5014 |
+
"grad_norm": 1.1551895141601562,
|
| 5015 |
+
"learning_rate": 9.951968838319495e-06,
|
| 5016 |
+
"loss": 0.0421,
|
| 5017 |
+
"step": 6560
|
| 5018 |
+
},
|
| 5019 |
+
{
|
| 5020 |
+
"epoch": 6.157450796626055,
|
| 5021 |
+
"grad_norm": 1.0832526683807373,
|
| 5022 |
+
"learning_rate": 9.951895620085226e-06,
|
| 5023 |
+
"loss": 0.0462,
|
| 5024 |
+
"step": 6570
|
| 5025 |
+
},
|
| 5026 |
+
{
|
| 5027 |
+
"epoch": 6.166822867853796,
|
| 5028 |
+
"grad_norm": 1.303536295890808,
|
| 5029 |
+
"learning_rate": 9.951822401850959e-06,
|
| 5030 |
+
"loss": 0.0423,
|
| 5031 |
+
"step": 6580
|
| 5032 |
+
},
|
| 5033 |
+
{
|
| 5034 |
+
"epoch": 6.176194939081537,
|
| 5035 |
+
"grad_norm": 1.2826794385910034,
|
| 5036 |
+
"learning_rate": 9.95174918361669e-06,
|
| 5037 |
+
"loss": 0.0463,
|
| 5038 |
+
"step": 6590
|
| 5039 |
+
},
|
| 5040 |
+
{
|
| 5041 |
+
"epoch": 6.185567010309279,
|
| 5042 |
+
"grad_norm": 1.0724890232086182,
|
| 5043 |
+
"learning_rate": 9.95167596538242e-06,
|
| 5044 |
+
"loss": 0.043,
|
| 5045 |
+
"step": 6600
|
| 5046 |
+
},
|
| 5047 |
+
{
|
| 5048 |
+
"epoch": 6.1949390815370196,
|
| 5049 |
+
"grad_norm": 0.9407768249511719,
|
| 5050 |
+
"learning_rate": 9.95160274714815e-06,
|
| 5051 |
+
"loss": 0.045,
|
| 5052 |
+
"step": 6610
|
| 5053 |
+
},
|
| 5054 |
+
{
|
| 5055 |
+
"epoch": 6.204311152764761,
|
| 5056 |
+
"grad_norm": 1.1686878204345703,
|
| 5057 |
+
"learning_rate": 9.951529528913882e-06,
|
| 5058 |
+
"loss": 0.0407,
|
| 5059 |
+
"step": 6620
|
| 5060 |
+
},
|
| 5061 |
+
{
|
| 5062 |
+
"epoch": 6.213683223992502,
|
| 5063 |
+
"grad_norm": 1.5972820520401,
|
| 5064 |
+
"learning_rate": 9.951456310679612e-06,
|
| 5065 |
+
"loss": 0.0449,
|
| 5066 |
+
"step": 6630
|
| 5067 |
+
},
|
| 5068 |
+
{
|
| 5069 |
+
"epoch": 6.223055295220243,
|
| 5070 |
+
"grad_norm": 0.7610195875167847,
|
| 5071 |
+
"learning_rate": 9.951383092445343e-06,
|
| 5072 |
+
"loss": 0.0397,
|
| 5073 |
+
"step": 6640
|
| 5074 |
+
},
|
| 5075 |
+
{
|
| 5076 |
+
"epoch": 6.232427366447985,
|
| 5077 |
+
"grad_norm": 1.02704656124115,
|
| 5078 |
+
"learning_rate": 9.951309874211075e-06,
|
| 5079 |
+
"loss": 0.0448,
|
| 5080 |
+
"step": 6650
|
| 5081 |
+
},
|
| 5082 |
+
{
|
| 5083 |
+
"epoch": 6.241799437675726,
|
| 5084 |
+
"grad_norm": 0.8035688400268555,
|
| 5085 |
+
"learning_rate": 9.951236655976805e-06,
|
| 5086 |
+
"loss": 0.0445,
|
| 5087 |
+
"step": 6660
|
| 5088 |
+
},
|
| 5089 |
+
{
|
| 5090 |
+
"epoch": 6.251171508903468,
|
| 5091 |
+
"grad_norm": 1.019539475440979,
|
| 5092 |
+
"learning_rate": 9.951163437742535e-06,
|
| 5093 |
+
"loss": 0.0452,
|
| 5094 |
+
"step": 6670
|
| 5095 |
+
},
|
| 5096 |
+
{
|
| 5097 |
+
"epoch": 6.260543580131209,
|
| 5098 |
+
"grad_norm": 1.662574291229248,
|
| 5099 |
+
"learning_rate": 9.951090219508268e-06,
|
| 5100 |
+
"loss": 0.0517,
|
| 5101 |
+
"step": 6680
|
| 5102 |
+
},
|
| 5103 |
+
{
|
| 5104 |
+
"epoch": 6.26991565135895,
|
| 5105 |
+
"grad_norm": 1.1599600315093994,
|
| 5106 |
+
"learning_rate": 9.951017001273998e-06,
|
| 5107 |
+
"loss": 0.0493,
|
| 5108 |
+
"step": 6690
|
| 5109 |
+
},
|
| 5110 |
+
{
|
| 5111 |
+
"epoch": 6.279287722586692,
|
| 5112 |
+
"grad_norm": 0.7756074070930481,
|
| 5113 |
+
"learning_rate": 9.95094378303973e-06,
|
| 5114 |
+
"loss": 0.048,
|
| 5115 |
+
"step": 6700
|
| 5116 |
+
},
|
| 5117 |
+
{
|
| 5118 |
+
"epoch": 6.288659793814433,
|
| 5119 |
+
"grad_norm": 1.0959285497665405,
|
| 5120 |
+
"learning_rate": 9.95087056480546e-06,
|
| 5121 |
+
"loss": 0.0501,
|
| 5122 |
+
"step": 6710
|
| 5123 |
+
},
|
| 5124 |
+
{
|
| 5125 |
+
"epoch": 6.298031865042175,
|
| 5126 |
+
"grad_norm": 1.2311910390853882,
|
| 5127 |
+
"learning_rate": 9.95079734657119e-06,
|
| 5128 |
+
"loss": 0.0486,
|
| 5129 |
+
"step": 6720
|
| 5130 |
+
},
|
| 5131 |
+
{
|
| 5132 |
+
"epoch": 6.307403936269916,
|
| 5133 |
+
"grad_norm": 1.2149254083633423,
|
| 5134 |
+
"learning_rate": 9.950724128336921e-06,
|
| 5135 |
+
"loss": 0.0389,
|
| 5136 |
+
"step": 6730
|
| 5137 |
+
},
|
| 5138 |
+
{
|
| 5139 |
+
"epoch": 6.316776007497657,
|
| 5140 |
+
"grad_norm": 1.5355291366577148,
|
| 5141 |
+
"learning_rate": 9.950650910102652e-06,
|
| 5142 |
+
"loss": 0.0472,
|
| 5143 |
+
"step": 6740
|
| 5144 |
+
},
|
| 5145 |
+
{
|
| 5146 |
+
"epoch": 6.3261480787253985,
|
| 5147 |
+
"grad_norm": 1.1264081001281738,
|
| 5148 |
+
"learning_rate": 9.950577691868385e-06,
|
| 5149 |
+
"loss": 0.043,
|
| 5150 |
+
"step": 6750
|
| 5151 |
+
},
|
| 5152 |
+
{
|
| 5153 |
+
"epoch": 6.3261480787253985,
|
| 5154 |
+
"eval_loss": 0.03764544054865837,
|
| 5155 |
+
"eval_pearson_cosine": 0.7692497968673706,
|
| 5156 |
+
"eval_pearson_dot": 0.7138222455978394,
|
| 5157 |
+
"eval_pearson_euclidean": 0.7343003749847412,
|
| 5158 |
+
"eval_pearson_manhattan": 0.7356712818145752,
|
| 5159 |
+
"eval_runtime": 22.6897,
|
| 5160 |
+
"eval_samples_per_second": 66.109,
|
| 5161 |
+
"eval_spearman_cosine": 0.7695765922931803,
|
| 5162 |
+
"eval_spearman_dot": 0.7152262336240688,
|
| 5163 |
+
"eval_spearman_euclidean": 0.739557951171161,
|
| 5164 |
+
"eval_spearman_manhattan": 0.7408550126908494,
|
| 5165 |
+
"eval_steps_per_second": 8.286,
|
| 5166 |
+
"step": 6750
|
| 5167 |
+
},
|
| 5168 |
+
{
|
| 5169 |
+
"epoch": 6.3355201499531395,
|
| 5170 |
+
"grad_norm": 0.6277545690536499,
|
| 5171 |
+
"learning_rate": 9.950504473634115e-06,
|
| 5172 |
+
"loss": 0.0406,
|
| 5173 |
+
"step": 6760
|
| 5174 |
+
},
|
| 5175 |
+
{
|
| 5176 |
+
"epoch": 6.344892221180881,
|
| 5177 |
+
"grad_norm": 1.3999137878417969,
|
| 5178 |
+
"learning_rate": 9.950431255399846e-06,
|
| 5179 |
+
"loss": 0.0447,
|
| 5180 |
+
"step": 6770
|
| 5181 |
+
},
|
| 5182 |
+
{
|
| 5183 |
+
"epoch": 6.354264292408622,
|
| 5184 |
+
"grad_norm": 0.7465086579322815,
|
| 5185 |
+
"learning_rate": 9.950358037165577e-06,
|
| 5186 |
+
"loss": 0.0502,
|
| 5187 |
+
"step": 6780
|
| 5188 |
+
},
|
| 5189 |
+
{
|
| 5190 |
+
"epoch": 6.363636363636363,
|
| 5191 |
+
"grad_norm": 1.1154383420944214,
|
| 5192 |
+
"learning_rate": 9.950284818931308e-06,
|
| 5193 |
+
"loss": 0.05,
|
| 5194 |
+
"step": 6790
|
| 5195 |
+
},
|
| 5196 |
+
{
|
| 5197 |
+
"epoch": 6.373008434864105,
|
| 5198 |
+
"grad_norm": 1.1133472919464111,
|
| 5199 |
+
"learning_rate": 9.950211600697038e-06,
|
| 5200 |
+
"loss": 0.0473,
|
| 5201 |
+
"step": 6800
|
| 5202 |
+
},
|
| 5203 |
+
{
|
| 5204 |
+
"epoch": 6.382380506091846,
|
| 5205 |
+
"grad_norm": 1.0995352268218994,
|
| 5206 |
+
"learning_rate": 9.95013838246277e-06,
|
| 5207 |
+
"loss": 0.0414,
|
| 5208 |
+
"step": 6810
|
| 5209 |
+
},
|
| 5210 |
+
{
|
| 5211 |
+
"epoch": 6.391752577319588,
|
| 5212 |
+
"grad_norm": 0.9666862487792969,
|
| 5213 |
+
"learning_rate": 9.9500651642285e-06,
|
| 5214 |
+
"loss": 0.049,
|
| 5215 |
+
"step": 6820
|
| 5216 |
+
},
|
| 5217 |
+
{
|
| 5218 |
+
"epoch": 6.401124648547329,
|
| 5219 |
+
"grad_norm": 1.1517918109893799,
|
| 5220 |
+
"learning_rate": 9.94999194599423e-06,
|
| 5221 |
+
"loss": 0.0413,
|
| 5222 |
+
"step": 6830
|
| 5223 |
+
},
|
| 5224 |
+
{
|
| 5225 |
+
"epoch": 6.41049671977507,
|
| 5226 |
+
"grad_norm": 0.5381759405136108,
|
| 5227 |
+
"learning_rate": 9.949918727759961e-06,
|
| 5228 |
+
"loss": 0.0418,
|
| 5229 |
+
"step": 6840
|
| 5230 |
+
},
|
| 5231 |
+
{
|
| 5232 |
+
"epoch": 6.419868791002812,
|
| 5233 |
+
"grad_norm": 0.973006546497345,
|
| 5234 |
+
"learning_rate": 9.949845509525692e-06,
|
| 5235 |
+
"loss": 0.0495,
|
| 5236 |
+
"step": 6850
|
| 5237 |
+
},
|
| 5238 |
+
{
|
| 5239 |
+
"epoch": 6.429240862230553,
|
| 5240 |
+
"grad_norm": 1.126633882522583,
|
| 5241 |
+
"learning_rate": 9.949772291291425e-06,
|
| 5242 |
+
"loss": 0.0493,
|
| 5243 |
+
"step": 6860
|
| 5244 |
+
},
|
| 5245 |
+
{
|
| 5246 |
+
"epoch": 6.438612933458295,
|
| 5247 |
+
"grad_norm": 0.7894268035888672,
|
| 5248 |
+
"learning_rate": 9.949699073057155e-06,
|
| 5249 |
+
"loss": 0.0436,
|
| 5250 |
+
"step": 6870
|
| 5251 |
+
},
|
| 5252 |
+
{
|
| 5253 |
+
"epoch": 6.447985004686036,
|
| 5254 |
+
"grad_norm": 0.7125422358512878,
|
| 5255 |
+
"learning_rate": 9.949625854822886e-06,
|
| 5256 |
+
"loss": 0.0433,
|
| 5257 |
+
"step": 6880
|
| 5258 |
+
},
|
| 5259 |
+
{
|
| 5260 |
+
"epoch": 6.457357075913777,
|
| 5261 |
+
"grad_norm": 0.9013342261314392,
|
| 5262 |
+
"learning_rate": 9.949552636588617e-06,
|
| 5263 |
+
"loss": 0.0376,
|
| 5264 |
+
"step": 6890
|
| 5265 |
+
},
|
| 5266 |
+
{
|
| 5267 |
+
"epoch": 6.4667291471415185,
|
| 5268 |
+
"grad_norm": 1.132384181022644,
|
| 5269 |
+
"learning_rate": 9.949479418354348e-06,
|
| 5270 |
+
"loss": 0.0482,
|
| 5271 |
+
"step": 6900
|
| 5272 |
+
},
|
| 5273 |
+
{
|
| 5274 |
+
"epoch": 6.4761012183692594,
|
| 5275 |
+
"grad_norm": 1.0104179382324219,
|
| 5276 |
+
"learning_rate": 9.949406200120078e-06,
|
| 5277 |
+
"loss": 0.0485,
|
| 5278 |
+
"step": 6910
|
| 5279 |
+
},
|
| 5280 |
+
{
|
| 5281 |
+
"epoch": 6.485473289597001,
|
| 5282 |
+
"grad_norm": 1.233464241027832,
|
| 5283 |
+
"learning_rate": 9.949332981885809e-06,
|
| 5284 |
+
"loss": 0.0478,
|
| 5285 |
+
"step": 6920
|
| 5286 |
+
},
|
| 5287 |
+
{
|
| 5288 |
+
"epoch": 6.494845360824742,
|
| 5289 |
+
"grad_norm": 0.7077954411506653,
|
| 5290 |
+
"learning_rate": 9.949259763651542e-06,
|
| 5291 |
+
"loss": 0.0464,
|
| 5292 |
+
"step": 6930
|
| 5293 |
+
},
|
| 5294 |
+
{
|
| 5295 |
+
"epoch": 6.504217432052483,
|
| 5296 |
+
"grad_norm": 1.5273882150650024,
|
| 5297 |
+
"learning_rate": 9.949186545417272e-06,
|
| 5298 |
+
"loss": 0.0404,
|
| 5299 |
+
"step": 6940
|
| 5300 |
+
},
|
| 5301 |
+
{
|
| 5302 |
+
"epoch": 6.513589503280225,
|
| 5303 |
+
"grad_norm": 1.2204720973968506,
|
| 5304 |
+
"learning_rate": 9.949113327183001e-06,
|
| 5305 |
+
"loss": 0.0375,
|
| 5306 |
+
"step": 6950
|
| 5307 |
+
},
|
| 5308 |
+
{
|
| 5309 |
+
"epoch": 6.522961574507966,
|
| 5310 |
+
"grad_norm": 0.9539759755134583,
|
| 5311 |
+
"learning_rate": 9.949040108948734e-06,
|
| 5312 |
+
"loss": 0.0397,
|
| 5313 |
+
"step": 6960
|
| 5314 |
+
},
|
| 5315 |
+
{
|
| 5316 |
+
"epoch": 6.532333645735708,
|
| 5317 |
+
"grad_norm": 1.949201226234436,
|
| 5318 |
+
"learning_rate": 9.948966890714465e-06,
|
| 5319 |
+
"loss": 0.0476,
|
| 5320 |
+
"step": 6970
|
| 5321 |
+
},
|
| 5322 |
+
{
|
| 5323 |
+
"epoch": 6.541705716963449,
|
| 5324 |
+
"grad_norm": 1.046915888786316,
|
| 5325 |
+
"learning_rate": 9.948893672480195e-06,
|
| 5326 |
+
"loss": 0.0445,
|
| 5327 |
+
"step": 6980
|
| 5328 |
+
},
|
| 5329 |
+
{
|
| 5330 |
+
"epoch": 6.55107778819119,
|
| 5331 |
+
"grad_norm": 0.8392923474311829,
|
| 5332 |
+
"learning_rate": 9.948820454245926e-06,
|
| 5333 |
+
"loss": 0.0502,
|
| 5334 |
+
"step": 6990
|
| 5335 |
+
},
|
| 5336 |
+
{
|
| 5337 |
+
"epoch": 6.560449859418932,
|
| 5338 |
+
"grad_norm": 1.357014536857605,
|
| 5339 |
+
"learning_rate": 9.948747236011659e-06,
|
| 5340 |
+
"loss": 0.0436,
|
| 5341 |
+
"step": 7000
|
| 5342 |
+
},
|
| 5343 |
+
{
|
| 5344 |
+
"epoch": 6.560449859418932,
|
| 5345 |
+
"eval_loss": 0.03813355416059494,
|
| 5346 |
+
"eval_pearson_cosine": 0.7662351131439209,
|
| 5347 |
+
"eval_pearson_dot": 0.7104849219322205,
|
| 5348 |
+
"eval_pearson_euclidean": 0.7334129810333252,
|
| 5349 |
+
"eval_pearson_manhattan": 0.7350986003875732,
|
| 5350 |
+
"eval_runtime": 22.7512,
|
| 5351 |
+
"eval_samples_per_second": 65.931,
|
| 5352 |
+
"eval_spearman_cosine": 0.7662226343415417,
|
| 5353 |
+
"eval_spearman_dot": 0.7115825441503862,
|
| 5354 |
+
"eval_spearman_euclidean": 0.7384103552275764,
|
| 5355 |
+
"eval_spearman_manhattan": 0.7397995971405482,
|
| 5356 |
+
"eval_steps_per_second": 8.263,
|
| 5357 |
+
"step": 7000
|
| 5358 |
}
|
| 5359 |
],
|
| 5360 |
"logging_steps": 10,
|