Training in progress, step 10670, 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 +504 -3
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 sha256:3745bf4a87f162fa13cf355e199ab846b0247a904a1213366e15055a6bf2f43a
|
| 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 sha256:7d1ed18d6fe47a085d91a34b52f0b0e9d63181b84ee4c405b93d5c8e09294fad
|
| 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:e73f3d2830bcf37e3d736f23b27e8ce733b3473d4cab28360690d45ce9f8fbaa
|
| 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 sha256:0272fe6cd9dc84504ed48d7b132ccc945c2d44d9831efb836fdb17160ecec1c1
|
| 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 sha256:7b990e0a8c37f6931ac353c659c83107fdc7ef191bb09facaeb9644d6874f096
|
| 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 sha256:4c914e450f9b754db2d81a37a948ec5a8925105ae3c9bf9546862109977afdeb
|
| 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 sha256:1d26231e7958cdd63301effbce7fbbecf14c30b13c5cf7b6c8e00c1f8efd5317
|
| 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:b0d7d94f216995135aa7d9b310b13d4e41fe010a347e4abe474e09fb0b16e836
|
| 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 sha256:1c3ac9632d99d55f1b6c9a1a327344e69abc94e4998731b976e138a7dfbb679e
|
| 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:f742e8fccfd64912166b4c2a12a9757a7223df0c62ca19cb8c5ef3c4dd1595a8
|
| 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:a60d80d2f480f1553b178ffde691b3d8251b3da0f0ce4460a958add33beb93ab
|
| 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,
|
|
@@ -7647,6 +7647,507 @@
|
|
| 7647 |
"eval_spearman_manhattan": 0.7421316928799319,
|
| 7648 |
"eval_steps_per_second": 7.269,
|
| 7649 |
"step": 10000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 7650 |
}
|
| 7651 |
],
|
| 7652 |
"logging_steps": 10,
|
|
@@ -7661,7 +8162,7 @@
|
|
| 7661 |
"should_evaluate": false,
|
| 7662 |
"should_log": false,
|
| 7663 |
"should_save": true,
|
| 7664 |
-
"should_training_stop":
|
| 7665 |
},
|
| 7666 |
"attributes": {}
|
| 7667 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 10.0,
|
| 5 |
"eval_steps": 250,
|
| 6 |
+
"global_step": 10670,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 7647 |
"eval_spearman_manhattan": 0.7421316928799319,
|
| 7648 |
"eval_steps_per_second": 7.269,
|
| 7649 |
"step": 10000
|
| 7650 |
+
},
|
| 7651 |
+
{
|
| 7652 |
+
"epoch": 9.381443298969073,
|
| 7653 |
+
"grad_norm": 0.4994644522666931,
|
| 7654 |
+
"learning_rate": 9.926708547496669e-06,
|
| 7655 |
+
"loss": 0.0226,
|
| 7656 |
+
"step": 10010
|
| 7657 |
+
},
|
| 7658 |
+
{
|
| 7659 |
+
"epoch": 9.390815370196814,
|
| 7660 |
+
"grad_norm": 1.5270389318466187,
|
| 7661 |
+
"learning_rate": 9.926635329262401e-06,
|
| 7662 |
+
"loss": 0.0211,
|
| 7663 |
+
"step": 10020
|
| 7664 |
+
},
|
| 7665 |
+
{
|
| 7666 |
+
"epoch": 9.400187441424555,
|
| 7667 |
+
"grad_norm": 0.47197312116622925,
|
| 7668 |
+
"learning_rate": 9.92656211102813e-06,
|
| 7669 |
+
"loss": 0.0235,
|
| 7670 |
+
"step": 10030
|
| 7671 |
+
},
|
| 7672 |
+
{
|
| 7673 |
+
"epoch": 9.409559512652296,
|
| 7674 |
+
"grad_norm": 1.132454752922058,
|
| 7675 |
+
"learning_rate": 9.926488892793863e-06,
|
| 7676 |
+
"loss": 0.023,
|
| 7677 |
+
"step": 10040
|
| 7678 |
+
},
|
| 7679 |
+
{
|
| 7680 |
+
"epoch": 9.418931583880038,
|
| 7681 |
+
"grad_norm": 0.7693812251091003,
|
| 7682 |
+
"learning_rate": 9.926415674559593e-06,
|
| 7683 |
+
"loss": 0.0247,
|
| 7684 |
+
"step": 10050
|
| 7685 |
+
},
|
| 7686 |
+
{
|
| 7687 |
+
"epoch": 9.42830365510778,
|
| 7688 |
+
"grad_norm": 0.42411306500434875,
|
| 7689 |
+
"learning_rate": 9.926342456325324e-06,
|
| 7690 |
+
"loss": 0.0234,
|
| 7691 |
+
"step": 10060
|
| 7692 |
+
},
|
| 7693 |
+
{
|
| 7694 |
+
"epoch": 9.43767572633552,
|
| 7695 |
+
"grad_norm": 0.9110538959503174,
|
| 7696 |
+
"learning_rate": 9.926269238091055e-06,
|
| 7697 |
+
"loss": 0.0256,
|
| 7698 |
+
"step": 10070
|
| 7699 |
+
},
|
| 7700 |
+
{
|
| 7701 |
+
"epoch": 9.447047797563261,
|
| 7702 |
+
"grad_norm": 0.6932746171951294,
|
| 7703 |
+
"learning_rate": 9.926196019856786e-06,
|
| 7704 |
+
"loss": 0.0288,
|
| 7705 |
+
"step": 10080
|
| 7706 |
+
},
|
| 7707 |
+
{
|
| 7708 |
+
"epoch": 9.456419868791002,
|
| 7709 |
+
"grad_norm": 0.6196317076683044,
|
| 7710 |
+
"learning_rate": 9.926122801622516e-06,
|
| 7711 |
+
"loss": 0.0239,
|
| 7712 |
+
"step": 10090
|
| 7713 |
+
},
|
| 7714 |
+
{
|
| 7715 |
+
"epoch": 9.465791940018745,
|
| 7716 |
+
"grad_norm": 0.6985231637954712,
|
| 7717 |
+
"learning_rate": 9.926049583388247e-06,
|
| 7718 |
+
"loss": 0.0194,
|
| 7719 |
+
"step": 10100
|
| 7720 |
+
},
|
| 7721 |
+
{
|
| 7722 |
+
"epoch": 9.475164011246486,
|
| 7723 |
+
"grad_norm": 0.8828220963478088,
|
| 7724 |
+
"learning_rate": 9.925976365153978e-06,
|
| 7725 |
+
"loss": 0.0282,
|
| 7726 |
+
"step": 10110
|
| 7727 |
+
},
|
| 7728 |
+
{
|
| 7729 |
+
"epoch": 9.484536082474227,
|
| 7730 |
+
"grad_norm": 0.3887142241001129,
|
| 7731 |
+
"learning_rate": 9.92590314691971e-06,
|
| 7732 |
+
"loss": 0.0231,
|
| 7733 |
+
"step": 10120
|
| 7734 |
+
},
|
| 7735 |
+
{
|
| 7736 |
+
"epoch": 9.493908153701968,
|
| 7737 |
+
"grad_norm": 0.696250855922699,
|
| 7738 |
+
"learning_rate": 9.925829928685441e-06,
|
| 7739 |
+
"loss": 0.0241,
|
| 7740 |
+
"step": 10130
|
| 7741 |
+
},
|
| 7742 |
+
{
|
| 7743 |
+
"epoch": 9.503280224929709,
|
| 7744 |
+
"grad_norm": 0.9591291546821594,
|
| 7745 |
+
"learning_rate": 9.925756710451172e-06,
|
| 7746 |
+
"loss": 0.0237,
|
| 7747 |
+
"step": 10140
|
| 7748 |
+
},
|
| 7749 |
+
{
|
| 7750 |
+
"epoch": 9.512652296157452,
|
| 7751 |
+
"grad_norm": 0.6247865557670593,
|
| 7752 |
+
"learning_rate": 9.925683492216903e-06,
|
| 7753 |
+
"loss": 0.0225,
|
| 7754 |
+
"step": 10150
|
| 7755 |
+
},
|
| 7756 |
+
{
|
| 7757 |
+
"epoch": 9.522024367385193,
|
| 7758 |
+
"grad_norm": 0.8061539530754089,
|
| 7759 |
+
"learning_rate": 9.925610273982633e-06,
|
| 7760 |
+
"loss": 0.0248,
|
| 7761 |
+
"step": 10160
|
| 7762 |
+
},
|
| 7763 |
+
{
|
| 7764 |
+
"epoch": 9.531396438612934,
|
| 7765 |
+
"grad_norm": 0.5681460499763489,
|
| 7766 |
+
"learning_rate": 9.925537055748364e-06,
|
| 7767 |
+
"loss": 0.0216,
|
| 7768 |
+
"step": 10170
|
| 7769 |
+
},
|
| 7770 |
+
{
|
| 7771 |
+
"epoch": 9.540768509840674,
|
| 7772 |
+
"grad_norm": 0.7798430323600769,
|
| 7773 |
+
"learning_rate": 9.925463837514095e-06,
|
| 7774 |
+
"loss": 0.0205,
|
| 7775 |
+
"step": 10180
|
| 7776 |
+
},
|
| 7777 |
+
{
|
| 7778 |
+
"epoch": 9.550140581068415,
|
| 7779 |
+
"grad_norm": 0.633307695388794,
|
| 7780 |
+
"learning_rate": 9.925390619279827e-06,
|
| 7781 |
+
"loss": 0.0257,
|
| 7782 |
+
"step": 10190
|
| 7783 |
+
},
|
| 7784 |
+
{
|
| 7785 |
+
"epoch": 9.559512652296158,
|
| 7786 |
+
"grad_norm": 0.5352799892425537,
|
| 7787 |
+
"learning_rate": 9.925317401045558e-06,
|
| 7788 |
+
"loss": 0.0214,
|
| 7789 |
+
"step": 10200
|
| 7790 |
+
},
|
| 7791 |
+
{
|
| 7792 |
+
"epoch": 9.5688847235239,
|
| 7793 |
+
"grad_norm": 1.4367021322250366,
|
| 7794 |
+
"learning_rate": 9.925244182811287e-06,
|
| 7795 |
+
"loss": 0.0245,
|
| 7796 |
+
"step": 10210
|
| 7797 |
+
},
|
| 7798 |
+
{
|
| 7799 |
+
"epoch": 9.57825679475164,
|
| 7800 |
+
"grad_norm": 0.6616729497909546,
|
| 7801 |
+
"learning_rate": 9.92517096457702e-06,
|
| 7802 |
+
"loss": 0.0168,
|
| 7803 |
+
"step": 10220
|
| 7804 |
+
},
|
| 7805 |
+
{
|
| 7806 |
+
"epoch": 9.587628865979381,
|
| 7807 |
+
"grad_norm": 0.5232043862342834,
|
| 7808 |
+
"learning_rate": 9.92509774634275e-06,
|
| 7809 |
+
"loss": 0.0229,
|
| 7810 |
+
"step": 10230
|
| 7811 |
+
},
|
| 7812 |
+
{
|
| 7813 |
+
"epoch": 9.597000937207122,
|
| 7814 |
+
"grad_norm": 0.5471720099449158,
|
| 7815 |
+
"learning_rate": 9.925024528108481e-06,
|
| 7816 |
+
"loss": 0.0244,
|
| 7817 |
+
"step": 10240
|
| 7818 |
+
},
|
| 7819 |
+
{
|
| 7820 |
+
"epoch": 9.606373008434865,
|
| 7821 |
+
"grad_norm": 0.8130425214767456,
|
| 7822 |
+
"learning_rate": 9.924951309874212e-06,
|
| 7823 |
+
"loss": 0.0243,
|
| 7824 |
+
"step": 10250
|
| 7825 |
+
},
|
| 7826 |
+
{
|
| 7827 |
+
"epoch": 9.606373008434865,
|
| 7828 |
+
"eval_loss": 0.037354420870542526,
|
| 7829 |
+
"eval_pearson_cosine": 0.7731273770332336,
|
| 7830 |
+
"eval_pearson_dot": 0.7302557826042175,
|
| 7831 |
+
"eval_pearson_euclidean": 0.7300422191619873,
|
| 7832 |
+
"eval_pearson_manhattan": 0.7321226596832275,
|
| 7833 |
+
"eval_runtime": 25.5048,
|
| 7834 |
+
"eval_samples_per_second": 58.813,
|
| 7835 |
+
"eval_spearman_cosine": 0.7727287355752905,
|
| 7836 |
+
"eval_spearman_dot": 0.7305929253470385,
|
| 7837 |
+
"eval_spearman_euclidean": 0.7346168467659768,
|
| 7838 |
+
"eval_spearman_manhattan": 0.7364009847987945,
|
| 7839 |
+
"eval_steps_per_second": 7.371,
|
| 7840 |
+
"step": 10250
|
| 7841 |
+
},
|
| 7842 |
+
{
|
| 7843 |
+
"epoch": 9.615745079662606,
|
| 7844 |
+
"grad_norm": 0.497060626745224,
|
| 7845 |
+
"learning_rate": 9.924878091639943e-06,
|
| 7846 |
+
"loss": 0.0217,
|
| 7847 |
+
"step": 10260
|
| 7848 |
+
},
|
| 7849 |
+
{
|
| 7850 |
+
"epoch": 9.625117150890347,
|
| 7851 |
+
"grad_norm": 0.985636830329895,
|
| 7852 |
+
"learning_rate": 9.924804873405673e-06,
|
| 7853 |
+
"loss": 0.0238,
|
| 7854 |
+
"step": 10270
|
| 7855 |
+
},
|
| 7856 |
+
{
|
| 7857 |
+
"epoch": 9.634489222118088,
|
| 7858 |
+
"grad_norm": 0.8833957314491272,
|
| 7859 |
+
"learning_rate": 9.924731655171404e-06,
|
| 7860 |
+
"loss": 0.0215,
|
| 7861 |
+
"step": 10280
|
| 7862 |
+
},
|
| 7863 |
+
{
|
| 7864 |
+
"epoch": 9.643861293345829,
|
| 7865 |
+
"grad_norm": 0.7223436832427979,
|
| 7866 |
+
"learning_rate": 9.924658436937137e-06,
|
| 7867 |
+
"loss": 0.0257,
|
| 7868 |
+
"step": 10290
|
| 7869 |
+
},
|
| 7870 |
+
{
|
| 7871 |
+
"epoch": 9.653233364573572,
|
| 7872 |
+
"grad_norm": 1.0917994976043701,
|
| 7873 |
+
"learning_rate": 9.924585218702867e-06,
|
| 7874 |
+
"loss": 0.0272,
|
| 7875 |
+
"step": 10300
|
| 7876 |
+
},
|
| 7877 |
+
{
|
| 7878 |
+
"epoch": 9.662605435801312,
|
| 7879 |
+
"grad_norm": 0.79998779296875,
|
| 7880 |
+
"learning_rate": 9.924512000468598e-06,
|
| 7881 |
+
"loss": 0.0232,
|
| 7882 |
+
"step": 10310
|
| 7883 |
+
},
|
| 7884 |
+
{
|
| 7885 |
+
"epoch": 9.671977507029053,
|
| 7886 |
+
"grad_norm": 0.9708638191223145,
|
| 7887 |
+
"learning_rate": 9.924438782234329e-06,
|
| 7888 |
+
"loss": 0.0214,
|
| 7889 |
+
"step": 10320
|
| 7890 |
+
},
|
| 7891 |
+
{
|
| 7892 |
+
"epoch": 9.681349578256794,
|
| 7893 |
+
"grad_norm": 0.5575175881385803,
|
| 7894 |
+
"learning_rate": 9.92436556400006e-06,
|
| 7895 |
+
"loss": 0.0256,
|
| 7896 |
+
"step": 10330
|
| 7897 |
+
},
|
| 7898 |
+
{
|
| 7899 |
+
"epoch": 9.690721649484535,
|
| 7900 |
+
"grad_norm": 1.2645318508148193,
|
| 7901 |
+
"learning_rate": 9.92429234576579e-06,
|
| 7902 |
+
"loss": 0.0276,
|
| 7903 |
+
"step": 10340
|
| 7904 |
+
},
|
| 7905 |
+
{
|
| 7906 |
+
"epoch": 9.700093720712278,
|
| 7907 |
+
"grad_norm": 0.6546396017074585,
|
| 7908 |
+
"learning_rate": 9.924219127531521e-06,
|
| 7909 |
+
"loss": 0.024,
|
| 7910 |
+
"step": 10350
|
| 7911 |
+
},
|
| 7912 |
+
{
|
| 7913 |
+
"epoch": 9.70946579194002,
|
| 7914 |
+
"grad_norm": 0.8439049124717712,
|
| 7915 |
+
"learning_rate": 9.924145909297252e-06,
|
| 7916 |
+
"loss": 0.0259,
|
| 7917 |
+
"step": 10360
|
| 7918 |
+
},
|
| 7919 |
+
{
|
| 7920 |
+
"epoch": 9.71883786316776,
|
| 7921 |
+
"grad_norm": 0.9637166261672974,
|
| 7922 |
+
"learning_rate": 9.924072691062984e-06,
|
| 7923 |
+
"loss": 0.0225,
|
| 7924 |
+
"step": 10370
|
| 7925 |
+
},
|
| 7926 |
+
{
|
| 7927 |
+
"epoch": 9.728209934395501,
|
| 7928 |
+
"grad_norm": 0.6104253530502319,
|
| 7929 |
+
"learning_rate": 9.923999472828713e-06,
|
| 7930 |
+
"loss": 0.0254,
|
| 7931 |
+
"step": 10380
|
| 7932 |
+
},
|
| 7933 |
+
{
|
| 7934 |
+
"epoch": 9.737582005623242,
|
| 7935 |
+
"grad_norm": 0.5664217472076416,
|
| 7936 |
+
"learning_rate": 9.923926254594444e-06,
|
| 7937 |
+
"loss": 0.0192,
|
| 7938 |
+
"step": 10390
|
| 7939 |
+
},
|
| 7940 |
+
{
|
| 7941 |
+
"epoch": 9.746954076850985,
|
| 7942 |
+
"grad_norm": 0.6904122233390808,
|
| 7943 |
+
"learning_rate": 9.923853036360176e-06,
|
| 7944 |
+
"loss": 0.0213,
|
| 7945 |
+
"step": 10400
|
| 7946 |
+
},
|
| 7947 |
+
{
|
| 7948 |
+
"epoch": 9.756326148078726,
|
| 7949 |
+
"grad_norm": 1.0864416360855103,
|
| 7950 |
+
"learning_rate": 9.923779818125907e-06,
|
| 7951 |
+
"loss": 0.0254,
|
| 7952 |
+
"step": 10410
|
| 7953 |
+
},
|
| 7954 |
+
{
|
| 7955 |
+
"epoch": 9.765698219306467,
|
| 7956 |
+
"grad_norm": 0.791348397731781,
|
| 7957 |
+
"learning_rate": 9.923706599891638e-06,
|
| 7958 |
+
"loss": 0.0264,
|
| 7959 |
+
"step": 10420
|
| 7960 |
+
},
|
| 7961 |
+
{
|
| 7962 |
+
"epoch": 9.775070290534208,
|
| 7963 |
+
"grad_norm": 0.7972745895385742,
|
| 7964 |
+
"learning_rate": 9.923633381657369e-06,
|
| 7965 |
+
"loss": 0.0206,
|
| 7966 |
+
"step": 10430
|
| 7967 |
+
},
|
| 7968 |
+
{
|
| 7969 |
+
"epoch": 9.784442361761949,
|
| 7970 |
+
"grad_norm": 0.6930385231971741,
|
| 7971 |
+
"learning_rate": 9.9235601634231e-06,
|
| 7972 |
+
"loss": 0.0283,
|
| 7973 |
+
"step": 10440
|
| 7974 |
+
},
|
| 7975 |
+
{
|
| 7976 |
+
"epoch": 9.793814432989691,
|
| 7977 |
+
"grad_norm": 0.5096721053123474,
|
| 7978 |
+
"learning_rate": 9.92348694518883e-06,
|
| 7979 |
+
"loss": 0.0263,
|
| 7980 |
+
"step": 10450
|
| 7981 |
+
},
|
| 7982 |
+
{
|
| 7983 |
+
"epoch": 9.803186504217432,
|
| 7984 |
+
"grad_norm": 0.7492228150367737,
|
| 7985 |
+
"learning_rate": 9.923413726954561e-06,
|
| 7986 |
+
"loss": 0.0237,
|
| 7987 |
+
"step": 10460
|
| 7988 |
+
},
|
| 7989 |
+
{
|
| 7990 |
+
"epoch": 9.812558575445173,
|
| 7991 |
+
"grad_norm": 0.8097043037414551,
|
| 7992 |
+
"learning_rate": 9.923340508720293e-06,
|
| 7993 |
+
"loss": 0.0225,
|
| 7994 |
+
"step": 10470
|
| 7995 |
+
},
|
| 7996 |
+
{
|
| 7997 |
+
"epoch": 9.821930646672914,
|
| 7998 |
+
"grad_norm": 0.45464569330215454,
|
| 7999 |
+
"learning_rate": 9.923267290486024e-06,
|
| 8000 |
+
"loss": 0.0175,
|
| 8001 |
+
"step": 10480
|
| 8002 |
+
},
|
| 8003 |
+
{
|
| 8004 |
+
"epoch": 9.831302717900655,
|
| 8005 |
+
"grad_norm": 0.6172147393226624,
|
| 8006 |
+
"learning_rate": 9.923194072251753e-06,
|
| 8007 |
+
"loss": 0.0272,
|
| 8008 |
+
"step": 10490
|
| 8009 |
+
},
|
| 8010 |
+
{
|
| 8011 |
+
"epoch": 9.840674789128398,
|
| 8012 |
+
"grad_norm": 0.9826374650001526,
|
| 8013 |
+
"learning_rate": 9.923120854017486e-06,
|
| 8014 |
+
"loss": 0.0233,
|
| 8015 |
+
"step": 10500
|
| 8016 |
+
},
|
| 8017 |
+
{
|
| 8018 |
+
"epoch": 9.840674789128398,
|
| 8019 |
+
"eval_loss": 0.03700366988778114,
|
| 8020 |
+
"eval_pearson_cosine": 0.7760223746299744,
|
| 8021 |
+
"eval_pearson_dot": 0.7342942953109741,
|
| 8022 |
+
"eval_pearson_euclidean": 0.7316151857376099,
|
| 8023 |
+
"eval_pearson_manhattan": 0.7336723804473877,
|
| 8024 |
+
"eval_runtime": 22.135,
|
| 8025 |
+
"eval_samples_per_second": 67.766,
|
| 8026 |
+
"eval_spearman_cosine": 0.7753394120917871,
|
| 8027 |
+
"eval_spearman_dot": 0.7356003834746606,
|
| 8028 |
+
"eval_spearman_euclidean": 0.7371167930939387,
|
| 8029 |
+
"eval_spearman_manhattan": 0.7388623589601665,
|
| 8030 |
+
"eval_steps_per_second": 8.493,
|
| 8031 |
+
"step": 10500
|
| 8032 |
+
},
|
| 8033 |
+
{
|
| 8034 |
+
"epoch": 9.850046860356139,
|
| 8035 |
+
"grad_norm": 0.5944278240203857,
|
| 8036 |
+
"learning_rate": 9.923047635783216e-06,
|
| 8037 |
+
"loss": 0.0245,
|
| 8038 |
+
"step": 10510
|
| 8039 |
+
},
|
| 8040 |
+
{
|
| 8041 |
+
"epoch": 9.85941893158388,
|
| 8042 |
+
"grad_norm": 0.4207167625427246,
|
| 8043 |
+
"learning_rate": 9.922974417548947e-06,
|
| 8044 |
+
"loss": 0.0236,
|
| 8045 |
+
"step": 10520
|
| 8046 |
+
},
|
| 8047 |
+
{
|
| 8048 |
+
"epoch": 9.868791002811621,
|
| 8049 |
+
"grad_norm": 1.185616374015808,
|
| 8050 |
+
"learning_rate": 9.922901199314678e-06,
|
| 8051 |
+
"loss": 0.025,
|
| 8052 |
+
"step": 10530
|
| 8053 |
+
},
|
| 8054 |
+
{
|
| 8055 |
+
"epoch": 9.878163074039362,
|
| 8056 |
+
"grad_norm": 0.6041834354400635,
|
| 8057 |
+
"learning_rate": 9.92282798108041e-06,
|
| 8058 |
+
"loss": 0.0229,
|
| 8059 |
+
"step": 10540
|
| 8060 |
+
},
|
| 8061 |
+
{
|
| 8062 |
+
"epoch": 9.887535145267105,
|
| 8063 |
+
"grad_norm": 1.3135936260223389,
|
| 8064 |
+
"learning_rate": 9.92275476284614e-06,
|
| 8065 |
+
"loss": 0.022,
|
| 8066 |
+
"step": 10550
|
| 8067 |
+
},
|
| 8068 |
+
{
|
| 8069 |
+
"epoch": 9.896907216494846,
|
| 8070 |
+
"grad_norm": 0.7592184543609619,
|
| 8071 |
+
"learning_rate": 9.92268154461187e-06,
|
| 8072 |
+
"loss": 0.0251,
|
| 8073 |
+
"step": 10560
|
| 8074 |
+
},
|
| 8075 |
+
{
|
| 8076 |
+
"epoch": 9.906279287722587,
|
| 8077 |
+
"grad_norm": 0.5679847002029419,
|
| 8078 |
+
"learning_rate": 9.922608326377603e-06,
|
| 8079 |
+
"loss": 0.0218,
|
| 8080 |
+
"step": 10570
|
| 8081 |
+
},
|
| 8082 |
+
{
|
| 8083 |
+
"epoch": 9.915651358950328,
|
| 8084 |
+
"grad_norm": 1.1727142333984375,
|
| 8085 |
+
"learning_rate": 9.922535108143333e-06,
|
| 8086 |
+
"loss": 0.0266,
|
| 8087 |
+
"step": 10580
|
| 8088 |
+
},
|
| 8089 |
+
{
|
| 8090 |
+
"epoch": 9.925023430178069,
|
| 8091 |
+
"grad_norm": 1.2769267559051514,
|
| 8092 |
+
"learning_rate": 9.922461889909064e-06,
|
| 8093 |
+
"loss": 0.0237,
|
| 8094 |
+
"step": 10590
|
| 8095 |
+
},
|
| 8096 |
+
{
|
| 8097 |
+
"epoch": 9.934395501405811,
|
| 8098 |
+
"grad_norm": 0.6604001522064209,
|
| 8099 |
+
"learning_rate": 9.922388671674795e-06,
|
| 8100 |
+
"loss": 0.0206,
|
| 8101 |
+
"step": 10600
|
| 8102 |
+
},
|
| 8103 |
+
{
|
| 8104 |
+
"epoch": 9.943767572633552,
|
| 8105 |
+
"grad_norm": 0.8065370321273804,
|
| 8106 |
+
"learning_rate": 9.922315453440526e-06,
|
| 8107 |
+
"loss": 0.0272,
|
| 8108 |
+
"step": 10610
|
| 8109 |
+
},
|
| 8110 |
+
{
|
| 8111 |
+
"epoch": 9.953139643861293,
|
| 8112 |
+
"grad_norm": 1.0085433721542358,
|
| 8113 |
+
"learning_rate": 9.922242235206256e-06,
|
| 8114 |
+
"loss": 0.019,
|
| 8115 |
+
"step": 10620
|
| 8116 |
+
},
|
| 8117 |
+
{
|
| 8118 |
+
"epoch": 9.962511715089034,
|
| 8119 |
+
"grad_norm": 0.9662045240402222,
|
| 8120 |
+
"learning_rate": 9.922169016971987e-06,
|
| 8121 |
+
"loss": 0.0218,
|
| 8122 |
+
"step": 10630
|
| 8123 |
+
},
|
| 8124 |
+
{
|
| 8125 |
+
"epoch": 9.971883786316775,
|
| 8126 |
+
"grad_norm": 0.49303632974624634,
|
| 8127 |
+
"learning_rate": 9.922095798737718e-06,
|
| 8128 |
+
"loss": 0.0223,
|
| 8129 |
+
"step": 10640
|
| 8130 |
+
},
|
| 8131 |
+
{
|
| 8132 |
+
"epoch": 9.981255857544518,
|
| 8133 |
+
"grad_norm": 0.7215604186058044,
|
| 8134 |
+
"learning_rate": 9.92202258050345e-06,
|
| 8135 |
+
"loss": 0.0259,
|
| 8136 |
+
"step": 10650
|
| 8137 |
+
},
|
| 8138 |
+
{
|
| 8139 |
+
"epoch": 9.990627928772259,
|
| 8140 |
+
"grad_norm": 0.6104753017425537,
|
| 8141 |
+
"learning_rate": 9.92194936226918e-06,
|
| 8142 |
+
"loss": 0.0232,
|
| 8143 |
+
"step": 10660
|
| 8144 |
+
},
|
| 8145 |
+
{
|
| 8146 |
+
"epoch": 10.0,
|
| 8147 |
+
"grad_norm": 1.011549949645996,
|
| 8148 |
+
"learning_rate": 9.92187614403491e-06,
|
| 8149 |
+
"loss": 0.0234,
|
| 8150 |
+
"step": 10670
|
| 8151 |
}
|
| 8152 |
],
|
| 8153 |
"logging_steps": 10,
|
|
|
|
| 8162 |
"should_evaluate": false,
|
| 8163 |
"should_log": false,
|
| 8164 |
"should_save": true,
|
| 8165 |
+
"should_training_stop": true
|
| 8166 |
},
|
| 8167 |
"attributes": {}
|
| 8168 |
}
|