Training in progress, step 4000, 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 +711 -3
last-checkpoint/model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 791869518
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2696563487c81338ee1c316d03a16b8cdda3a230f8c30ff43050548155251701
|
| 3 |
size 791869518
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 2375752250
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c48c8f079566e8995d74a2dde5c2bad04c58bdad806c74f723414147f3b56d46
|
| 3 |
size 2375752250
|
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:e735ed11597ed40a2b6854e0229902e1a21fedc0a0dbc608ca905fae57d5b06b
|
| 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 sha256:2ba3815fc0953b1b7f08cea092dfc0a62c4bbc2a2c68780d3f4dd0b5e22582a7
|
| 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 sha256:647ac15563fcad903adbb616e9b2c36b237a3ed5939d088620212da969930f6c
|
| 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 sha256:93e3733c5b180986b7efbec17b663bf5231343d187374d184768fcd913797167
|
| 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 sha256:9820ea4fec1b01f3da091290c3e8b5ddb86a3a3fa17285c248b64910c2d0b4f0
|
| 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 sha256:7413035def085e41776a629afc94fc24fe5a955f1ad83b32f9b370ab60f9a18d
|
| 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 sha256:91e3953bcbf4089415abffbd914fbbe4580121f6c843eabbf70624c5ed144814
|
| 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 sha256:369fde7bff4dfc0d6b9cf773cf9b0352696083f84763999e05a631ee6d52c5e3
|
| 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 sha256:97c307475b202e1cf87916aaf07584ea570dab26688da354222b9b509afaa3e1
|
| 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": 0.
|
| 5 |
"eval_steps": 500,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -4963,6 +4963,714 @@
|
|
| 4963 |
"eval_samples_per_second": 1118.746,
|
| 4964 |
"eval_steps_per_second": 34.961,
|
| 4965 |
"step": 3500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4966 |
}
|
| 4967 |
],
|
| 4968 |
"logging_steps": 5,
|
|
@@ -4982,7 +5690,7 @@
|
|
| 4982 |
"attributes": {}
|
| 4983 |
}
|
| 4984 |
},
|
| 4985 |
-
"total_flos": 1.
|
| 4986 |
"train_batch_size": 4,
|
| 4987 |
"trial_name": null,
|
| 4988 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9872429697502584,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 4000,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 4963 |
"eval_samples_per_second": 1118.746,
|
| 4964 |
"eval_steps_per_second": 34.961,
|
| 4965 |
"step": 3500
|
| 4966 |
+
},
|
| 4967 |
+
{
|
| 4968 |
+
"epoch": 0.8650716522436639,
|
| 4969 |
+
"grad_norm": 105.625,
|
| 4970 |
+
"learning_rate": 2.995884773662551e-07,
|
| 4971 |
+
"loss": 68.0234,
|
| 4972 |
+
"step": 3505
|
| 4973 |
+
},
|
| 4974 |
+
{
|
| 4975 |
+
"epoch": 0.8663057059558518,
|
| 4976 |
+
"grad_norm": 105.4375,
|
| 4977 |
+
"learning_rate": 2.968449931412894e-07,
|
| 4978 |
+
"loss": 67.563,
|
| 4979 |
+
"step": 3510
|
| 4980 |
+
},
|
| 4981 |
+
{
|
| 4982 |
+
"epoch": 0.8675397596680395,
|
| 4983 |
+
"grad_norm": 102.125,
|
| 4984 |
+
"learning_rate": 2.941015089163237e-07,
|
| 4985 |
+
"loss": 68.1967,
|
| 4986 |
+
"step": 3515
|
| 4987 |
+
},
|
| 4988 |
+
{
|
| 4989 |
+
"epoch": 0.8687738133802274,
|
| 4990 |
+
"grad_norm": 105.0,
|
| 4991 |
+
"learning_rate": 2.91358024691358e-07,
|
| 4992 |
+
"loss": 68.515,
|
| 4993 |
+
"step": 3520
|
| 4994 |
+
},
|
| 4995 |
+
{
|
| 4996 |
+
"epoch": 0.8700078670924152,
|
| 4997 |
+
"grad_norm": 103.3125,
|
| 4998 |
+
"learning_rate": 2.886145404663923e-07,
|
| 4999 |
+
"loss": 68.8215,
|
| 5000 |
+
"step": 3525
|
| 5001 |
+
},
|
| 5002 |
+
{
|
| 5003 |
+
"epoch": 0.871241920804603,
|
| 5004 |
+
"grad_norm": 106.8125,
|
| 5005 |
+
"learning_rate": 2.858710562414266e-07,
|
| 5006 |
+
"loss": 67.0145,
|
| 5007 |
+
"step": 3530
|
| 5008 |
+
},
|
| 5009 |
+
{
|
| 5010 |
+
"epoch": 0.8724759745167908,
|
| 5011 |
+
"grad_norm": 107.375,
|
| 5012 |
+
"learning_rate": 2.8312757201646087e-07,
|
| 5013 |
+
"loss": 68.5533,
|
| 5014 |
+
"step": 3535
|
| 5015 |
+
},
|
| 5016 |
+
{
|
| 5017 |
+
"epoch": 0.8737100282289787,
|
| 5018 |
+
"grad_norm": 104.3125,
|
| 5019 |
+
"learning_rate": 2.803840877914952e-07,
|
| 5020 |
+
"loss": 66.9931,
|
| 5021 |
+
"step": 3540
|
| 5022 |
+
},
|
| 5023 |
+
{
|
| 5024 |
+
"epoch": 0.8749440819411665,
|
| 5025 |
+
"grad_norm": 106.5,
|
| 5026 |
+
"learning_rate": 2.7764060356652947e-07,
|
| 5027 |
+
"loss": 68.8769,
|
| 5028 |
+
"step": 3545
|
| 5029 |
+
},
|
| 5030 |
+
{
|
| 5031 |
+
"epoch": 0.8761781356533543,
|
| 5032 |
+
"grad_norm": 106.3125,
|
| 5033 |
+
"learning_rate": 2.748971193415638e-07,
|
| 5034 |
+
"loss": 67.0216,
|
| 5035 |
+
"step": 3550
|
| 5036 |
+
},
|
| 5037 |
+
{
|
| 5038 |
+
"epoch": 0.8774121893655421,
|
| 5039 |
+
"grad_norm": 102.5,
|
| 5040 |
+
"learning_rate": 2.7215363511659807e-07,
|
| 5041 |
+
"loss": 69.8652,
|
| 5042 |
+
"step": 3555
|
| 5043 |
+
},
|
| 5044 |
+
{
|
| 5045 |
+
"epoch": 0.87864624307773,
|
| 5046 |
+
"grad_norm": 104.25,
|
| 5047 |
+
"learning_rate": 2.694101508916324e-07,
|
| 5048 |
+
"loss": 68.6337,
|
| 5049 |
+
"step": 3560
|
| 5050 |
+
},
|
| 5051 |
+
{
|
| 5052 |
+
"epoch": 0.8798802967899177,
|
| 5053 |
+
"grad_norm": 102.625,
|
| 5054 |
+
"learning_rate": 2.6666666666666667e-07,
|
| 5055 |
+
"loss": 66.5836,
|
| 5056 |
+
"step": 3565
|
| 5057 |
+
},
|
| 5058 |
+
{
|
| 5059 |
+
"epoch": 0.8811143505021056,
|
| 5060 |
+
"grad_norm": 102.875,
|
| 5061 |
+
"learning_rate": 2.63923182441701e-07,
|
| 5062 |
+
"loss": 67.1445,
|
| 5063 |
+
"step": 3570
|
| 5064 |
+
},
|
| 5065 |
+
{
|
| 5066 |
+
"epoch": 0.8823484042142934,
|
| 5067 |
+
"grad_norm": 106.375,
|
| 5068 |
+
"learning_rate": 2.611796982167352e-07,
|
| 5069 |
+
"loss": 68.3128,
|
| 5070 |
+
"step": 3575
|
| 5071 |
+
},
|
| 5072 |
+
{
|
| 5073 |
+
"epoch": 0.8835824579264813,
|
| 5074 |
+
"grad_norm": 106.5,
|
| 5075 |
+
"learning_rate": 2.584362139917695e-07,
|
| 5076 |
+
"loss": 67.5264,
|
| 5077 |
+
"step": 3580
|
| 5078 |
+
},
|
| 5079 |
+
{
|
| 5080 |
+
"epoch": 0.884816511638669,
|
| 5081 |
+
"grad_norm": 105.25,
|
| 5082 |
+
"learning_rate": 2.556927297668038e-07,
|
| 5083 |
+
"loss": 68.2184,
|
| 5084 |
+
"step": 3585
|
| 5085 |
+
},
|
| 5086 |
+
{
|
| 5087 |
+
"epoch": 0.8860505653508569,
|
| 5088 |
+
"grad_norm": 107.0,
|
| 5089 |
+
"learning_rate": 2.529492455418381e-07,
|
| 5090 |
+
"loss": 68.3838,
|
| 5091 |
+
"step": 3590
|
| 5092 |
+
},
|
| 5093 |
+
{
|
| 5094 |
+
"epoch": 0.8872846190630447,
|
| 5095 |
+
"grad_norm": 101.25,
|
| 5096 |
+
"learning_rate": 2.502057613168724e-07,
|
| 5097 |
+
"loss": 67.683,
|
| 5098 |
+
"step": 3595
|
| 5099 |
+
},
|
| 5100 |
+
{
|
| 5101 |
+
"epoch": 0.8885186727752326,
|
| 5102 |
+
"grad_norm": 104.125,
|
| 5103 |
+
"learning_rate": 2.474622770919067e-07,
|
| 5104 |
+
"loss": 69.2804,
|
| 5105 |
+
"step": 3600
|
| 5106 |
+
},
|
| 5107 |
+
{
|
| 5108 |
+
"epoch": 0.8897527264874203,
|
| 5109 |
+
"grad_norm": 106.0625,
|
| 5110 |
+
"learning_rate": 2.44718792866941e-07,
|
| 5111 |
+
"loss": 68.8471,
|
| 5112 |
+
"step": 3605
|
| 5113 |
+
},
|
| 5114 |
+
{
|
| 5115 |
+
"epoch": 0.8909867801996082,
|
| 5116 |
+
"grad_norm": 103.125,
|
| 5117 |
+
"learning_rate": 2.419753086419753e-07,
|
| 5118 |
+
"loss": 68.7409,
|
| 5119 |
+
"step": 3610
|
| 5120 |
+
},
|
| 5121 |
+
{
|
| 5122 |
+
"epoch": 0.892220833911796,
|
| 5123 |
+
"grad_norm": 106.4375,
|
| 5124 |
+
"learning_rate": 2.3923182441700957e-07,
|
| 5125 |
+
"loss": 67.3829,
|
| 5126 |
+
"step": 3615
|
| 5127 |
+
},
|
| 5128 |
+
{
|
| 5129 |
+
"epoch": 0.8934548876239838,
|
| 5130 |
+
"grad_norm": 104.5,
|
| 5131 |
+
"learning_rate": 2.364883401920439e-07,
|
| 5132 |
+
"loss": 68.2949,
|
| 5133 |
+
"step": 3620
|
| 5134 |
+
},
|
| 5135 |
+
{
|
| 5136 |
+
"epoch": 0.8946889413361716,
|
| 5137 |
+
"grad_norm": 103.4375,
|
| 5138 |
+
"learning_rate": 2.337448559670782e-07,
|
| 5139 |
+
"loss": 68.3005,
|
| 5140 |
+
"step": 3625
|
| 5141 |
+
},
|
| 5142 |
+
{
|
| 5143 |
+
"epoch": 0.8959229950483595,
|
| 5144 |
+
"grad_norm": 104.0,
|
| 5145 |
+
"learning_rate": 2.3100137174211247e-07,
|
| 5146 |
+
"loss": 65.948,
|
| 5147 |
+
"step": 3630
|
| 5148 |
+
},
|
| 5149 |
+
{
|
| 5150 |
+
"epoch": 0.8971570487605472,
|
| 5151 |
+
"grad_norm": 104.0,
|
| 5152 |
+
"learning_rate": 2.2825788751714677e-07,
|
| 5153 |
+
"loss": 67.4345,
|
| 5154 |
+
"step": 3635
|
| 5155 |
+
},
|
| 5156 |
+
{
|
| 5157 |
+
"epoch": 0.8983911024727351,
|
| 5158 |
+
"grad_norm": 106.125,
|
| 5159 |
+
"learning_rate": 2.2551440329218105e-07,
|
| 5160 |
+
"loss": 68.0186,
|
| 5161 |
+
"step": 3640
|
| 5162 |
+
},
|
| 5163 |
+
{
|
| 5164 |
+
"epoch": 0.899625156184923,
|
| 5165 |
+
"grad_norm": 106.6875,
|
| 5166 |
+
"learning_rate": 2.2277091906721535e-07,
|
| 5167 |
+
"loss": 69.481,
|
| 5168 |
+
"step": 3645
|
| 5169 |
+
},
|
| 5170 |
+
{
|
| 5171 |
+
"epoch": 0.9008592098971108,
|
| 5172 |
+
"grad_norm": 103.875,
|
| 5173 |
+
"learning_rate": 2.2002743484224965e-07,
|
| 5174 |
+
"loss": 68.3314,
|
| 5175 |
+
"step": 3650
|
| 5176 |
+
},
|
| 5177 |
+
{
|
| 5178 |
+
"epoch": 0.9020932636092986,
|
| 5179 |
+
"grad_norm": 103.3125,
|
| 5180 |
+
"learning_rate": 2.1728395061728395e-07,
|
| 5181 |
+
"loss": 66.6179,
|
| 5182 |
+
"step": 3655
|
| 5183 |
+
},
|
| 5184 |
+
{
|
| 5185 |
+
"epoch": 0.9033273173214864,
|
| 5186 |
+
"grad_norm": 102.25,
|
| 5187 |
+
"learning_rate": 2.1454046639231825e-07,
|
| 5188 |
+
"loss": 68.0456,
|
| 5189 |
+
"step": 3660
|
| 5190 |
+
},
|
| 5191 |
+
{
|
| 5192 |
+
"epoch": 0.9045613710336743,
|
| 5193 |
+
"grad_norm": 110.25,
|
| 5194 |
+
"learning_rate": 2.1179698216735252e-07,
|
| 5195 |
+
"loss": 68.8177,
|
| 5196 |
+
"step": 3665
|
| 5197 |
+
},
|
| 5198 |
+
{
|
| 5199 |
+
"epoch": 0.905795424745862,
|
| 5200 |
+
"grad_norm": 104.6875,
|
| 5201 |
+
"learning_rate": 2.0905349794238682e-07,
|
| 5202 |
+
"loss": 67.566,
|
| 5203 |
+
"step": 3670
|
| 5204 |
+
},
|
| 5205 |
+
{
|
| 5206 |
+
"epoch": 0.9070294784580499,
|
| 5207 |
+
"grad_norm": 103.875,
|
| 5208 |
+
"learning_rate": 2.0631001371742112e-07,
|
| 5209 |
+
"loss": 66.4167,
|
| 5210 |
+
"step": 3675
|
| 5211 |
+
},
|
| 5212 |
+
{
|
| 5213 |
+
"epoch": 0.9082635321702377,
|
| 5214 |
+
"grad_norm": 104.8125,
|
| 5215 |
+
"learning_rate": 2.035665294924554e-07,
|
| 5216 |
+
"loss": 68.3844,
|
| 5217 |
+
"step": 3680
|
| 5218 |
+
},
|
| 5219 |
+
{
|
| 5220 |
+
"epoch": 0.9094975858824256,
|
| 5221 |
+
"grad_norm": 104.0625,
|
| 5222 |
+
"learning_rate": 2.008230452674897e-07,
|
| 5223 |
+
"loss": 67.0489,
|
| 5224 |
+
"step": 3685
|
| 5225 |
+
},
|
| 5226 |
+
{
|
| 5227 |
+
"epoch": 0.9107316395946133,
|
| 5228 |
+
"grad_norm": 107.3125,
|
| 5229 |
+
"learning_rate": 1.98079561042524e-07,
|
| 5230 |
+
"loss": 67.8705,
|
| 5231 |
+
"step": 3690
|
| 5232 |
+
},
|
| 5233 |
+
{
|
| 5234 |
+
"epoch": 0.9119656933068012,
|
| 5235 |
+
"grad_norm": 103.375,
|
| 5236 |
+
"learning_rate": 1.953360768175583e-07,
|
| 5237 |
+
"loss": 67.953,
|
| 5238 |
+
"step": 3695
|
| 5239 |
+
},
|
| 5240 |
+
{
|
| 5241 |
+
"epoch": 0.913199747018989,
|
| 5242 |
+
"grad_norm": 105.25,
|
| 5243 |
+
"learning_rate": 1.9259259259259257e-07,
|
| 5244 |
+
"loss": 67.5511,
|
| 5245 |
+
"step": 3700
|
| 5246 |
+
},
|
| 5247 |
+
{
|
| 5248 |
+
"epoch": 0.9144338007311769,
|
| 5249 |
+
"grad_norm": 103.125,
|
| 5250 |
+
"learning_rate": 1.8984910836762687e-07,
|
| 5251 |
+
"loss": 67.6629,
|
| 5252 |
+
"step": 3705
|
| 5253 |
+
},
|
| 5254 |
+
{
|
| 5255 |
+
"epoch": 0.9156678544433646,
|
| 5256 |
+
"grad_norm": 107.0625,
|
| 5257 |
+
"learning_rate": 1.8710562414266117e-07,
|
| 5258 |
+
"loss": 68.8038,
|
| 5259 |
+
"step": 3710
|
| 5260 |
+
},
|
| 5261 |
+
{
|
| 5262 |
+
"epoch": 0.9169019081555525,
|
| 5263 |
+
"grad_norm": 103.6875,
|
| 5264 |
+
"learning_rate": 1.8436213991769547e-07,
|
| 5265 |
+
"loss": 66.1586,
|
| 5266 |
+
"step": 3715
|
| 5267 |
+
},
|
| 5268 |
+
{
|
| 5269 |
+
"epoch": 0.9181359618677403,
|
| 5270 |
+
"grad_norm": 104.8125,
|
| 5271 |
+
"learning_rate": 1.8161865569272977e-07,
|
| 5272 |
+
"loss": 67.8019,
|
| 5273 |
+
"step": 3720
|
| 5274 |
+
},
|
| 5275 |
+
{
|
| 5276 |
+
"epoch": 0.9193700155799281,
|
| 5277 |
+
"grad_norm": 101.6875,
|
| 5278 |
+
"learning_rate": 1.7887517146776405e-07,
|
| 5279 |
+
"loss": 67.1005,
|
| 5280 |
+
"step": 3725
|
| 5281 |
+
},
|
| 5282 |
+
{
|
| 5283 |
+
"epoch": 0.9206040692921159,
|
| 5284 |
+
"grad_norm": 107.9375,
|
| 5285 |
+
"learning_rate": 1.7613168724279835e-07,
|
| 5286 |
+
"loss": 67.2893,
|
| 5287 |
+
"step": 3730
|
| 5288 |
+
},
|
| 5289 |
+
{
|
| 5290 |
+
"epoch": 0.9218381230043038,
|
| 5291 |
+
"grad_norm": 105.0625,
|
| 5292 |
+
"learning_rate": 1.7338820301783262e-07,
|
| 5293 |
+
"loss": 65.8601,
|
| 5294 |
+
"step": 3735
|
| 5295 |
+
},
|
| 5296 |
+
{
|
| 5297 |
+
"epoch": 0.9230721767164916,
|
| 5298 |
+
"grad_norm": 105.8125,
|
| 5299 |
+
"learning_rate": 1.7064471879286692e-07,
|
| 5300 |
+
"loss": 66.3078,
|
| 5301 |
+
"step": 3740
|
| 5302 |
+
},
|
| 5303 |
+
{
|
| 5304 |
+
"epoch": 0.9243062304286794,
|
| 5305 |
+
"grad_norm": 106.1875,
|
| 5306 |
+
"learning_rate": 1.6790123456790122e-07,
|
| 5307 |
+
"loss": 66.7744,
|
| 5308 |
+
"step": 3745
|
| 5309 |
+
},
|
| 5310 |
+
{
|
| 5311 |
+
"epoch": 0.9255402841408672,
|
| 5312 |
+
"grad_norm": 104.625,
|
| 5313 |
+
"learning_rate": 1.6515775034293552e-07,
|
| 5314 |
+
"loss": 68.7563,
|
| 5315 |
+
"step": 3750
|
| 5316 |
+
},
|
| 5317 |
+
{
|
| 5318 |
+
"epoch": 0.9267743378530551,
|
| 5319 |
+
"grad_norm": 104.625,
|
| 5320 |
+
"learning_rate": 1.6241426611796983e-07,
|
| 5321 |
+
"loss": 67.7685,
|
| 5322 |
+
"step": 3755
|
| 5323 |
+
},
|
| 5324 |
+
{
|
| 5325 |
+
"epoch": 0.9280083915652428,
|
| 5326 |
+
"grad_norm": 105.3125,
|
| 5327 |
+
"learning_rate": 1.5967078189300413e-07,
|
| 5328 |
+
"loss": 68.9463,
|
| 5329 |
+
"step": 3760
|
| 5330 |
+
},
|
| 5331 |
+
{
|
| 5332 |
+
"epoch": 0.9292424452774307,
|
| 5333 |
+
"grad_norm": 99.75,
|
| 5334 |
+
"learning_rate": 1.569272976680384e-07,
|
| 5335 |
+
"loss": 68.4013,
|
| 5336 |
+
"step": 3765
|
| 5337 |
+
},
|
| 5338 |
+
{
|
| 5339 |
+
"epoch": 0.9304764989896185,
|
| 5340 |
+
"grad_norm": 102.875,
|
| 5341 |
+
"learning_rate": 1.5418381344307267e-07,
|
| 5342 |
+
"loss": 67.2276,
|
| 5343 |
+
"step": 3770
|
| 5344 |
+
},
|
| 5345 |
+
{
|
| 5346 |
+
"epoch": 0.9317105527018064,
|
| 5347 |
+
"grad_norm": 105.4375,
|
| 5348 |
+
"learning_rate": 1.5144032921810697e-07,
|
| 5349 |
+
"loss": 67.9008,
|
| 5350 |
+
"step": 3775
|
| 5351 |
+
},
|
| 5352 |
+
{
|
| 5353 |
+
"epoch": 0.9329446064139941,
|
| 5354 |
+
"grad_norm": 106.3125,
|
| 5355 |
+
"learning_rate": 1.4869684499314127e-07,
|
| 5356 |
+
"loss": 68.7067,
|
| 5357 |
+
"step": 3780
|
| 5358 |
+
},
|
| 5359 |
+
{
|
| 5360 |
+
"epoch": 0.934178660126182,
|
| 5361 |
+
"grad_norm": 103.625,
|
| 5362 |
+
"learning_rate": 1.4595336076817558e-07,
|
| 5363 |
+
"loss": 66.8319,
|
| 5364 |
+
"step": 3785
|
| 5365 |
+
},
|
| 5366 |
+
{
|
| 5367 |
+
"epoch": 0.9354127138383698,
|
| 5368 |
+
"grad_norm": 106.0,
|
| 5369 |
+
"learning_rate": 1.4320987654320988e-07,
|
| 5370 |
+
"loss": 66.7882,
|
| 5371 |
+
"step": 3790
|
| 5372 |
+
},
|
| 5373 |
+
{
|
| 5374 |
+
"epoch": 0.9366467675505576,
|
| 5375 |
+
"grad_norm": 109.5625,
|
| 5376 |
+
"learning_rate": 1.4046639231824418e-07,
|
| 5377 |
+
"loss": 67.7522,
|
| 5378 |
+
"step": 3795
|
| 5379 |
+
},
|
| 5380 |
+
{
|
| 5381 |
+
"epoch": 0.9378808212627454,
|
| 5382 |
+
"grad_norm": 103.8125,
|
| 5383 |
+
"learning_rate": 1.3772290809327848e-07,
|
| 5384 |
+
"loss": 67.2511,
|
| 5385 |
+
"step": 3800
|
| 5386 |
+
},
|
| 5387 |
+
{
|
| 5388 |
+
"epoch": 0.9391148749749333,
|
| 5389 |
+
"grad_norm": 104.6875,
|
| 5390 |
+
"learning_rate": 1.3497942386831278e-07,
|
| 5391 |
+
"loss": 67.226,
|
| 5392 |
+
"step": 3805
|
| 5393 |
+
},
|
| 5394 |
+
{
|
| 5395 |
+
"epoch": 0.940348928687121,
|
| 5396 |
+
"grad_norm": 101.8125,
|
| 5397 |
+
"learning_rate": 1.3223593964334702e-07,
|
| 5398 |
+
"loss": 66.5479,
|
| 5399 |
+
"step": 3810
|
| 5400 |
+
},
|
| 5401 |
+
{
|
| 5402 |
+
"epoch": 0.9415829823993089,
|
| 5403 |
+
"grad_norm": 104.3125,
|
| 5404 |
+
"learning_rate": 1.2949245541838133e-07,
|
| 5405 |
+
"loss": 67.8139,
|
| 5406 |
+
"step": 3815
|
| 5407 |
+
},
|
| 5408 |
+
{
|
| 5409 |
+
"epoch": 0.9428170361114967,
|
| 5410 |
+
"grad_norm": 104.375,
|
| 5411 |
+
"learning_rate": 1.2674897119341563e-07,
|
| 5412 |
+
"loss": 67.2027,
|
| 5413 |
+
"step": 3820
|
| 5414 |
+
},
|
| 5415 |
+
{
|
| 5416 |
+
"epoch": 0.9440510898236846,
|
| 5417 |
+
"grad_norm": 104.8125,
|
| 5418 |
+
"learning_rate": 1.2400548696844993e-07,
|
| 5419 |
+
"loss": 67.2941,
|
| 5420 |
+
"step": 3825
|
| 5421 |
+
},
|
| 5422 |
+
{
|
| 5423 |
+
"epoch": 0.9452851435358725,
|
| 5424 |
+
"grad_norm": 103.9375,
|
| 5425 |
+
"learning_rate": 1.2126200274348423e-07,
|
| 5426 |
+
"loss": 66.4945,
|
| 5427 |
+
"step": 3830
|
| 5428 |
+
},
|
| 5429 |
+
{
|
| 5430 |
+
"epoch": 0.9465191972480602,
|
| 5431 |
+
"grad_norm": 102.375,
|
| 5432 |
+
"learning_rate": 1.1851851851851851e-07,
|
| 5433 |
+
"loss": 66.7565,
|
| 5434 |
+
"step": 3835
|
| 5435 |
+
},
|
| 5436 |
+
{
|
| 5437 |
+
"epoch": 0.9477532509602481,
|
| 5438 |
+
"grad_norm": 104.75,
|
| 5439 |
+
"learning_rate": 1.157750342935528e-07,
|
| 5440 |
+
"loss": 69.663,
|
| 5441 |
+
"step": 3840
|
| 5442 |
+
},
|
| 5443 |
+
{
|
| 5444 |
+
"epoch": 0.9489873046724359,
|
| 5445 |
+
"grad_norm": 102.0625,
|
| 5446 |
+
"learning_rate": 1.130315500685871e-07,
|
| 5447 |
+
"loss": 67.022,
|
| 5448 |
+
"step": 3845
|
| 5449 |
+
},
|
| 5450 |
+
{
|
| 5451 |
+
"epoch": 0.9502213583846237,
|
| 5452 |
+
"grad_norm": 103.5625,
|
| 5453 |
+
"learning_rate": 1.1028806584362139e-07,
|
| 5454 |
+
"loss": 66.7464,
|
| 5455 |
+
"step": 3850
|
| 5456 |
+
},
|
| 5457 |
+
{
|
| 5458 |
+
"epoch": 0.9514554120968115,
|
| 5459 |
+
"grad_norm": 106.125,
|
| 5460 |
+
"learning_rate": 1.0754458161865569e-07,
|
| 5461 |
+
"loss": 67.5819,
|
| 5462 |
+
"step": 3855
|
| 5463 |
+
},
|
| 5464 |
+
{
|
| 5465 |
+
"epoch": 0.9526894658089994,
|
| 5466 |
+
"grad_norm": 107.1875,
|
| 5467 |
+
"learning_rate": 1.0480109739368999e-07,
|
| 5468 |
+
"loss": 68.8449,
|
| 5469 |
+
"step": 3860
|
| 5470 |
+
},
|
| 5471 |
+
{
|
| 5472 |
+
"epoch": 0.9539235195211871,
|
| 5473 |
+
"grad_norm": 104.25,
|
| 5474 |
+
"learning_rate": 1.0205761316872428e-07,
|
| 5475 |
+
"loss": 67.4359,
|
| 5476 |
+
"step": 3865
|
| 5477 |
+
},
|
| 5478 |
+
{
|
| 5479 |
+
"epoch": 0.955157573233375,
|
| 5480 |
+
"grad_norm": 110.0625,
|
| 5481 |
+
"learning_rate": 9.931412894375856e-08,
|
| 5482 |
+
"loss": 66.3128,
|
| 5483 |
+
"step": 3870
|
| 5484 |
+
},
|
| 5485 |
+
{
|
| 5486 |
+
"epoch": 0.9563916269455628,
|
| 5487 |
+
"grad_norm": 106.5625,
|
| 5488 |
+
"learning_rate": 9.657064471879286e-08,
|
| 5489 |
+
"loss": 67.7631,
|
| 5490 |
+
"step": 3875
|
| 5491 |
+
},
|
| 5492 |
+
{
|
| 5493 |
+
"epoch": 0.9576256806577507,
|
| 5494 |
+
"grad_norm": 104.25,
|
| 5495 |
+
"learning_rate": 9.382716049382716e-08,
|
| 5496 |
+
"loss": 67.0168,
|
| 5497 |
+
"step": 3880
|
| 5498 |
+
},
|
| 5499 |
+
{
|
| 5500 |
+
"epoch": 0.9588597343699384,
|
| 5501 |
+
"grad_norm": 99.5,
|
| 5502 |
+
"learning_rate": 9.108367626886144e-08,
|
| 5503 |
+
"loss": 67.202,
|
| 5504 |
+
"step": 3885
|
| 5505 |
+
},
|
| 5506 |
+
{
|
| 5507 |
+
"epoch": 0.9600937880821263,
|
| 5508 |
+
"grad_norm": 103.0625,
|
| 5509 |
+
"learning_rate": 8.834019204389574e-08,
|
| 5510 |
+
"loss": 66.6468,
|
| 5511 |
+
"step": 3890
|
| 5512 |
+
},
|
| 5513 |
+
{
|
| 5514 |
+
"epoch": 0.9613278417943141,
|
| 5515 |
+
"grad_norm": 104.25,
|
| 5516 |
+
"learning_rate": 8.559670781893004e-08,
|
| 5517 |
+
"loss": 67.3951,
|
| 5518 |
+
"step": 3895
|
| 5519 |
+
},
|
| 5520 |
+
{
|
| 5521 |
+
"epoch": 0.962561895506502,
|
| 5522 |
+
"grad_norm": 103.8125,
|
| 5523 |
+
"learning_rate": 8.285322359396434e-08,
|
| 5524 |
+
"loss": 66.0457,
|
| 5525 |
+
"step": 3900
|
| 5526 |
+
},
|
| 5527 |
+
{
|
| 5528 |
+
"epoch": 0.9637959492186897,
|
| 5529 |
+
"grad_norm": 105.125,
|
| 5530 |
+
"learning_rate": 8.010973936899861e-08,
|
| 5531 |
+
"loss": 66.7826,
|
| 5532 |
+
"step": 3905
|
| 5533 |
+
},
|
| 5534 |
+
{
|
| 5535 |
+
"epoch": 0.9650300029308776,
|
| 5536 |
+
"grad_norm": 104.125,
|
| 5537 |
+
"learning_rate": 7.736625514403291e-08,
|
| 5538 |
+
"loss": 67.4089,
|
| 5539 |
+
"step": 3910
|
| 5540 |
+
},
|
| 5541 |
+
{
|
| 5542 |
+
"epoch": 0.9662640566430654,
|
| 5543 |
+
"grad_norm": 105.4375,
|
| 5544 |
+
"learning_rate": 7.462277091906722e-08,
|
| 5545 |
+
"loss": 67.6246,
|
| 5546 |
+
"step": 3915
|
| 5547 |
+
},
|
| 5548 |
+
{
|
| 5549 |
+
"epoch": 0.9674981103552532,
|
| 5550 |
+
"grad_norm": 101.875,
|
| 5551 |
+
"learning_rate": 7.187928669410152e-08,
|
| 5552 |
+
"loss": 67.3768,
|
| 5553 |
+
"step": 3920
|
| 5554 |
+
},
|
| 5555 |
+
{
|
| 5556 |
+
"epoch": 0.968732164067441,
|
| 5557 |
+
"grad_norm": 104.3125,
|
| 5558 |
+
"learning_rate": 6.913580246913579e-08,
|
| 5559 |
+
"loss": 67.9191,
|
| 5560 |
+
"step": 3925
|
| 5561 |
+
},
|
| 5562 |
+
{
|
| 5563 |
+
"epoch": 0.9699662177796289,
|
| 5564 |
+
"grad_norm": 103.4375,
|
| 5565 |
+
"learning_rate": 6.639231824417009e-08,
|
| 5566 |
+
"loss": 66.7172,
|
| 5567 |
+
"step": 3930
|
| 5568 |
+
},
|
| 5569 |
+
{
|
| 5570 |
+
"epoch": 0.9712002714918166,
|
| 5571 |
+
"grad_norm": 104.875,
|
| 5572 |
+
"learning_rate": 6.364883401920439e-08,
|
| 5573 |
+
"loss": 68.1103,
|
| 5574 |
+
"step": 3935
|
| 5575 |
+
},
|
| 5576 |
+
{
|
| 5577 |
+
"epoch": 0.9724343252040045,
|
| 5578 |
+
"grad_norm": 106.1875,
|
| 5579 |
+
"learning_rate": 6.090534979423868e-08,
|
| 5580 |
+
"loss": 66.7532,
|
| 5581 |
+
"step": 3940
|
| 5582 |
+
},
|
| 5583 |
+
{
|
| 5584 |
+
"epoch": 0.9736683789161923,
|
| 5585 |
+
"grad_norm": 101.9375,
|
| 5586 |
+
"learning_rate": 5.816186556927297e-08,
|
| 5587 |
+
"loss": 66.8054,
|
| 5588 |
+
"step": 3945
|
| 5589 |
+
},
|
| 5590 |
+
{
|
| 5591 |
+
"epoch": 0.9749024326283802,
|
| 5592 |
+
"grad_norm": 107.1875,
|
| 5593 |
+
"learning_rate": 5.541838134430727e-08,
|
| 5594 |
+
"loss": 66.1427,
|
| 5595 |
+
"step": 3950
|
| 5596 |
+
},
|
| 5597 |
+
{
|
| 5598 |
+
"epoch": 0.9761364863405679,
|
| 5599 |
+
"grad_norm": 99.8125,
|
| 5600 |
+
"learning_rate": 5.267489711934156e-08,
|
| 5601 |
+
"loss": 66.5402,
|
| 5602 |
+
"step": 3955
|
| 5603 |
+
},
|
| 5604 |
+
{
|
| 5605 |
+
"epoch": 0.9773705400527558,
|
| 5606 |
+
"grad_norm": 104.0625,
|
| 5607 |
+
"learning_rate": 4.993141289437586e-08,
|
| 5608 |
+
"loss": 68.0075,
|
| 5609 |
+
"step": 3960
|
| 5610 |
+
},
|
| 5611 |
+
{
|
| 5612 |
+
"epoch": 0.9786045937649436,
|
| 5613 |
+
"grad_norm": 106.125,
|
| 5614 |
+
"learning_rate": 4.718792866941015e-08,
|
| 5615 |
+
"loss": 67.5935,
|
| 5616 |
+
"step": 3965
|
| 5617 |
+
},
|
| 5618 |
+
{
|
| 5619 |
+
"epoch": 0.9798386474771315,
|
| 5620 |
+
"grad_norm": 104.9375,
|
| 5621 |
+
"learning_rate": 4.444444444444444e-08,
|
| 5622 |
+
"loss": 66.8711,
|
| 5623 |
+
"step": 3970
|
| 5624 |
+
},
|
| 5625 |
+
{
|
| 5626 |
+
"epoch": 0.9810727011893192,
|
| 5627 |
+
"grad_norm": 104.5,
|
| 5628 |
+
"learning_rate": 4.1700960219478735e-08,
|
| 5629 |
+
"loss": 66.3836,
|
| 5630 |
+
"step": 3975
|
| 5631 |
+
},
|
| 5632 |
+
{
|
| 5633 |
+
"epoch": 0.9823067549015071,
|
| 5634 |
+
"grad_norm": 102.0625,
|
| 5635 |
+
"learning_rate": 3.895747599451303e-08,
|
| 5636 |
+
"loss": 65.9148,
|
| 5637 |
+
"step": 3980
|
| 5638 |
+
},
|
| 5639 |
+
{
|
| 5640 |
+
"epoch": 0.9835408086136949,
|
| 5641 |
+
"grad_norm": 105.375,
|
| 5642 |
+
"learning_rate": 3.621399176954732e-08,
|
| 5643 |
+
"loss": 66.4969,
|
| 5644 |
+
"step": 3985
|
| 5645 |
+
},
|
| 5646 |
+
{
|
| 5647 |
+
"epoch": 0.9847748623258827,
|
| 5648 |
+
"grad_norm": 101.1875,
|
| 5649 |
+
"learning_rate": 3.3470507544581616e-08,
|
| 5650 |
+
"loss": 66.926,
|
| 5651 |
+
"step": 3990
|
| 5652 |
+
},
|
| 5653 |
+
{
|
| 5654 |
+
"epoch": 0.9860089160380705,
|
| 5655 |
+
"grad_norm": 102.4375,
|
| 5656 |
+
"learning_rate": 3.072702331961591e-08,
|
| 5657 |
+
"loss": 65.3679,
|
| 5658 |
+
"step": 3995
|
| 5659 |
+
},
|
| 5660 |
+
{
|
| 5661 |
+
"epoch": 0.9872429697502584,
|
| 5662 |
+
"grad_norm": 107.875,
|
| 5663 |
+
"learning_rate": 2.7983539094650204e-08,
|
| 5664 |
+
"loss": 67.4716,
|
| 5665 |
+
"step": 4000
|
| 5666 |
+
},
|
| 5667 |
+
{
|
| 5668 |
+
"epoch": 0.9872429697502584,
|
| 5669 |
+
"eval_loss": 2.095496892929077,
|
| 5670 |
+
"eval_runtime": 197.7366,
|
| 5671 |
+
"eval_samples_per_second": 1104.317,
|
| 5672 |
+
"eval_steps_per_second": 34.511,
|
| 5673 |
+
"step": 4000
|
| 5674 |
}
|
| 5675 |
],
|
| 5676 |
"logging_steps": 5,
|
|
|
|
| 5690 |
"attributes": {}
|
| 5691 |
}
|
| 5692 |
},
|
| 5693 |
+
"total_flos": 1.7329431971155149e+19,
|
| 5694 |
"train_batch_size": 4,
|
| 5695 |
"trial_name": null,
|
| 5696 |
"trial_params": null
|