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:d93859b3f54bac8daa5949d1dd355a2a00d83ee3a118f96100fa3ebb9da8bc60
|
| 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 sha256:7937948b64a0cbcfba3f2fac38861165be5e6be90d0b693ca42ac510f2a90bef
|
| 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 sha256:1bef6a7bf53166ec3a9709e315e5a7afc807cf01be6b61a09c96b7113cbb6fd6
|
| 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": 605.018,
|
| 4964 |
"eval_steps_per_second": 37.814,
|
| 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": 0.9247018197842845,
|
| 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": 605.018,
|
| 4964 |
"eval_steps_per_second": 37.814,
|
| 4965 |
"step": 3500
|
| 4966 |
+
},
|
| 4967 |
+
{
|
| 4968 |
+
"epoch": 0.8102699695859792,
|
| 4969 |
+
"grad_norm": 143.625,
|
| 4970 |
+
"learning_rate": 2.1068859198355603e-06,
|
| 4971 |
+
"loss": 84.304,
|
| 4972 |
+
"step": 3505
|
| 4973 |
+
},
|
| 4974 |
+
{
|
| 4975 |
+
"epoch": 0.8114258468607095,
|
| 4976 |
+
"grad_norm": 150.75,
|
| 4977 |
+
"learning_rate": 2.094039054470709e-06,
|
| 4978 |
+
"loss": 84.0018,
|
| 4979 |
+
"step": 3510
|
| 4980 |
+
},
|
| 4981 |
+
{
|
| 4982 |
+
"epoch": 0.8125817241354399,
|
| 4983 |
+
"grad_norm": 142.0,
|
| 4984 |
+
"learning_rate": 2.0811921891058584e-06,
|
| 4985 |
+
"loss": 84.433,
|
| 4986 |
+
"step": 3515
|
| 4987 |
+
},
|
| 4988 |
+
{
|
| 4989 |
+
"epoch": 0.8137376014101703,
|
| 4990 |
+
"grad_norm": 143.25,
|
| 4991 |
+
"learning_rate": 2.0683453237410072e-06,
|
| 4992 |
+
"loss": 84.3377,
|
| 4993 |
+
"step": 3520
|
| 4994 |
+
},
|
| 4995 |
+
{
|
| 4996 |
+
"epoch": 0.8148934786849006,
|
| 4997 |
+
"grad_norm": 146.125,
|
| 4998 |
+
"learning_rate": 2.0554984583761565e-06,
|
| 4999 |
+
"loss": 84.5472,
|
| 5000 |
+
"step": 3525
|
| 5001 |
+
},
|
| 5002 |
+
{
|
| 5003 |
+
"epoch": 0.816049355959631,
|
| 5004 |
+
"grad_norm": 151.5,
|
| 5005 |
+
"learning_rate": 2.0426515930113053e-06,
|
| 5006 |
+
"loss": 82.4523,
|
| 5007 |
+
"step": 3530
|
| 5008 |
+
},
|
| 5009 |
+
{
|
| 5010 |
+
"epoch": 0.8172052332343613,
|
| 5011 |
+
"grad_norm": 141.125,
|
| 5012 |
+
"learning_rate": 2.0298047276464546e-06,
|
| 5013 |
+
"loss": 84.5754,
|
| 5014 |
+
"step": 3535
|
| 5015 |
+
},
|
| 5016 |
+
{
|
| 5017 |
+
"epoch": 0.8183611105090917,
|
| 5018 |
+
"grad_norm": 147.5,
|
| 5019 |
+
"learning_rate": 2.0169578622816035e-06,
|
| 5020 |
+
"loss": 82.441,
|
| 5021 |
+
"step": 3540
|
| 5022 |
+
},
|
| 5023 |
+
{
|
| 5024 |
+
"epoch": 0.8195169877838221,
|
| 5025 |
+
"grad_norm": 146.625,
|
| 5026 |
+
"learning_rate": 2.0041109969167523e-06,
|
| 5027 |
+
"loss": 86.3065,
|
| 5028 |
+
"step": 3545
|
| 5029 |
+
},
|
| 5030 |
+
{
|
| 5031 |
+
"epoch": 0.8206728650585524,
|
| 5032 |
+
"grad_norm": 152.5,
|
| 5033 |
+
"learning_rate": 1.9912641315519016e-06,
|
| 5034 |
+
"loss": 83.517,
|
| 5035 |
+
"step": 3550
|
| 5036 |
+
},
|
| 5037 |
+
{
|
| 5038 |
+
"epoch": 0.8218287423332827,
|
| 5039 |
+
"grad_norm": 145.875,
|
| 5040 |
+
"learning_rate": 1.9784172661870504e-06,
|
| 5041 |
+
"loss": 84.9357,
|
| 5042 |
+
"step": 3555
|
| 5043 |
+
},
|
| 5044 |
+
{
|
| 5045 |
+
"epoch": 0.8229846196080132,
|
| 5046 |
+
"grad_norm": 134.75,
|
| 5047 |
+
"learning_rate": 1.9655704008221997e-06,
|
| 5048 |
+
"loss": 83.7317,
|
| 5049 |
+
"step": 3560
|
| 5050 |
+
},
|
| 5051 |
+
{
|
| 5052 |
+
"epoch": 0.8241404968827435,
|
| 5053 |
+
"grad_norm": 152.0,
|
| 5054 |
+
"learning_rate": 1.9527235354573485e-06,
|
| 5055 |
+
"loss": 81.6789,
|
| 5056 |
+
"step": 3565
|
| 5057 |
+
},
|
| 5058 |
+
{
|
| 5059 |
+
"epoch": 0.8252963741574738,
|
| 5060 |
+
"grad_norm": 165.0,
|
| 5061 |
+
"learning_rate": 1.9398766700924974e-06,
|
| 5062 |
+
"loss": 83.5837,
|
| 5063 |
+
"step": 3570
|
| 5064 |
+
},
|
| 5065 |
+
{
|
| 5066 |
+
"epoch": 0.8264522514322041,
|
| 5067 |
+
"grad_norm": 143.375,
|
| 5068 |
+
"learning_rate": 1.9270298047276467e-06,
|
| 5069 |
+
"loss": 84.4118,
|
| 5070 |
+
"step": 3575
|
| 5071 |
+
},
|
| 5072 |
+
{
|
| 5073 |
+
"epoch": 0.8276081287069346,
|
| 5074 |
+
"grad_norm": 141.375,
|
| 5075 |
+
"learning_rate": 1.9141829393627955e-06,
|
| 5076 |
+
"loss": 81.7501,
|
| 5077 |
+
"step": 3580
|
| 5078 |
+
},
|
| 5079 |
+
{
|
| 5080 |
+
"epoch": 0.8287640059816649,
|
| 5081 |
+
"grad_norm": 136.375,
|
| 5082 |
+
"learning_rate": 1.9013360739979448e-06,
|
| 5083 |
+
"loss": 83.1388,
|
| 5084 |
+
"step": 3585
|
| 5085 |
+
},
|
| 5086 |
+
{
|
| 5087 |
+
"epoch": 0.8299198832563952,
|
| 5088 |
+
"grad_norm": 152.25,
|
| 5089 |
+
"learning_rate": 1.8884892086330936e-06,
|
| 5090 |
+
"loss": 82.2967,
|
| 5091 |
+
"step": 3590
|
| 5092 |
+
},
|
| 5093 |
+
{
|
| 5094 |
+
"epoch": 0.8310757605311256,
|
| 5095 |
+
"grad_norm": 141.875,
|
| 5096 |
+
"learning_rate": 1.8756423432682427e-06,
|
| 5097 |
+
"loss": 83.2457,
|
| 5098 |
+
"step": 3595
|
| 5099 |
+
},
|
| 5100 |
+
{
|
| 5101 |
+
"epoch": 0.832231637805856,
|
| 5102 |
+
"grad_norm": 146.625,
|
| 5103 |
+
"learning_rate": 1.8627954779033915e-06,
|
| 5104 |
+
"loss": 84.0741,
|
| 5105 |
+
"step": 3600
|
| 5106 |
+
},
|
| 5107 |
+
{
|
| 5108 |
+
"epoch": 0.8333875150805863,
|
| 5109 |
+
"grad_norm": 142.875,
|
| 5110 |
+
"learning_rate": 1.8499486125385408e-06,
|
| 5111 |
+
"loss": 84.0243,
|
| 5112 |
+
"step": 3605
|
| 5113 |
+
},
|
| 5114 |
+
{
|
| 5115 |
+
"epoch": 0.8345433923553167,
|
| 5116 |
+
"grad_norm": 143.0,
|
| 5117 |
+
"learning_rate": 1.8371017471736899e-06,
|
| 5118 |
+
"loss": 84.5223,
|
| 5119 |
+
"step": 3610
|
| 5120 |
+
},
|
| 5121 |
+
{
|
| 5122 |
+
"epoch": 0.835699269630047,
|
| 5123 |
+
"grad_norm": 139.25,
|
| 5124 |
+
"learning_rate": 1.8242548818088387e-06,
|
| 5125 |
+
"loss": 82.374,
|
| 5126 |
+
"step": 3615
|
| 5127 |
+
},
|
| 5128 |
+
{
|
| 5129 |
+
"epoch": 0.8368551469047774,
|
| 5130 |
+
"grad_norm": 135.875,
|
| 5131 |
+
"learning_rate": 1.811408016443988e-06,
|
| 5132 |
+
"loss": 83.3824,
|
| 5133 |
+
"step": 3620
|
| 5134 |
+
},
|
| 5135 |
+
{
|
| 5136 |
+
"epoch": 0.8380110241795078,
|
| 5137 |
+
"grad_norm": 146.875,
|
| 5138 |
+
"learning_rate": 1.7985611510791368e-06,
|
| 5139 |
+
"loss": 82.5953,
|
| 5140 |
+
"step": 3625
|
| 5141 |
+
},
|
| 5142 |
+
{
|
| 5143 |
+
"epoch": 0.8391669014542381,
|
| 5144 |
+
"grad_norm": 140.125,
|
| 5145 |
+
"learning_rate": 1.7857142857142859e-06,
|
| 5146 |
+
"loss": 83.3932,
|
| 5147 |
+
"step": 3630
|
| 5148 |
+
},
|
| 5149 |
+
{
|
| 5150 |
+
"epoch": 0.8403227787289684,
|
| 5151 |
+
"grad_norm": 145.125,
|
| 5152 |
+
"learning_rate": 1.7728674203494347e-06,
|
| 5153 |
+
"loss": 84.7285,
|
| 5154 |
+
"step": 3635
|
| 5155 |
+
},
|
| 5156 |
+
{
|
| 5157 |
+
"epoch": 0.8414786560036988,
|
| 5158 |
+
"grad_norm": 147.5,
|
| 5159 |
+
"learning_rate": 1.760020554984584e-06,
|
| 5160 |
+
"loss": 83.1239,
|
| 5161 |
+
"step": 3640
|
| 5162 |
+
},
|
| 5163 |
+
{
|
| 5164 |
+
"epoch": 0.8426345332784292,
|
| 5165 |
+
"grad_norm": 146.125,
|
| 5166 |
+
"learning_rate": 1.747173689619733e-06,
|
| 5167 |
+
"loss": 85.1945,
|
| 5168 |
+
"step": 3645
|
| 5169 |
+
},
|
| 5170 |
+
{
|
| 5171 |
+
"epoch": 0.8437904105531595,
|
| 5172 |
+
"grad_norm": 156.125,
|
| 5173 |
+
"learning_rate": 1.734326824254882e-06,
|
| 5174 |
+
"loss": 83.2979,
|
| 5175 |
+
"step": 3650
|
| 5176 |
+
},
|
| 5177 |
+
{
|
| 5178 |
+
"epoch": 0.8449462878278898,
|
| 5179 |
+
"grad_norm": 146.0,
|
| 5180 |
+
"learning_rate": 1.721479958890031e-06,
|
| 5181 |
+
"loss": 82.1503,
|
| 5182 |
+
"step": 3655
|
| 5183 |
+
},
|
| 5184 |
+
{
|
| 5185 |
+
"epoch": 0.8461021651026203,
|
| 5186 |
+
"grad_norm": 148.0,
|
| 5187 |
+
"learning_rate": 1.7086330935251798e-06,
|
| 5188 |
+
"loss": 83.1019,
|
| 5189 |
+
"step": 3660
|
| 5190 |
+
},
|
| 5191 |
+
{
|
| 5192 |
+
"epoch": 0.8472580423773506,
|
| 5193 |
+
"grad_norm": 143.625,
|
| 5194 |
+
"learning_rate": 1.695786228160329e-06,
|
| 5195 |
+
"loss": 83.5616,
|
| 5196 |
+
"step": 3665
|
| 5197 |
+
},
|
| 5198 |
+
{
|
| 5199 |
+
"epoch": 0.8484139196520809,
|
| 5200 |
+
"grad_norm": 142.75,
|
| 5201 |
+
"learning_rate": 1.682939362795478e-06,
|
| 5202 |
+
"loss": 84.0701,
|
| 5203 |
+
"step": 3670
|
| 5204 |
+
},
|
| 5205 |
+
{
|
| 5206 |
+
"epoch": 0.8495697969268113,
|
| 5207 |
+
"grad_norm": 138.25,
|
| 5208 |
+
"learning_rate": 1.670092497430627e-06,
|
| 5209 |
+
"loss": 82.2154,
|
| 5210 |
+
"step": 3675
|
| 5211 |
+
},
|
| 5212 |
+
{
|
| 5213 |
+
"epoch": 0.8507256742015417,
|
| 5214 |
+
"grad_norm": 137.0,
|
| 5215 |
+
"learning_rate": 1.6572456320657763e-06,
|
| 5216 |
+
"loss": 82.9864,
|
| 5217 |
+
"step": 3680
|
| 5218 |
+
},
|
| 5219 |
+
{
|
| 5220 |
+
"epoch": 0.851881551476272,
|
| 5221 |
+
"grad_norm": 146.125,
|
| 5222 |
+
"learning_rate": 1.6443987667009251e-06,
|
| 5223 |
+
"loss": 82.6712,
|
| 5224 |
+
"step": 3685
|
| 5225 |
+
},
|
| 5226 |
+
{
|
| 5227 |
+
"epoch": 0.8530374287510024,
|
| 5228 |
+
"grad_norm": 142.25,
|
| 5229 |
+
"learning_rate": 1.6315519013360742e-06,
|
| 5230 |
+
"loss": 83.6857,
|
| 5231 |
+
"step": 3690
|
| 5232 |
+
},
|
| 5233 |
+
{
|
| 5234 |
+
"epoch": 0.8541933060257327,
|
| 5235 |
+
"grad_norm": 141.875,
|
| 5236 |
+
"learning_rate": 1.618705035971223e-06,
|
| 5237 |
+
"loss": 82.5429,
|
| 5238 |
+
"step": 3695
|
| 5239 |
+
},
|
| 5240 |
+
{
|
| 5241 |
+
"epoch": 0.8553491833004631,
|
| 5242 |
+
"grad_norm": 152.0,
|
| 5243 |
+
"learning_rate": 1.6058581706063723e-06,
|
| 5244 |
+
"loss": 82.8123,
|
| 5245 |
+
"step": 3700
|
| 5246 |
+
},
|
| 5247 |
+
{
|
| 5248 |
+
"epoch": 0.8565050605751934,
|
| 5249 |
+
"grad_norm": 143.375,
|
| 5250 |
+
"learning_rate": 1.5930113052415211e-06,
|
| 5251 |
+
"loss": 83.5665,
|
| 5252 |
+
"step": 3705
|
| 5253 |
+
},
|
| 5254 |
+
{
|
| 5255 |
+
"epoch": 0.8576609378499238,
|
| 5256 |
+
"grad_norm": 150.875,
|
| 5257 |
+
"learning_rate": 1.5801644398766702e-06,
|
| 5258 |
+
"loss": 83.4201,
|
| 5259 |
+
"step": 3710
|
| 5260 |
+
},
|
| 5261 |
+
{
|
| 5262 |
+
"epoch": 0.8588168151246541,
|
| 5263 |
+
"grad_norm": 153.0,
|
| 5264 |
+
"learning_rate": 1.5673175745118195e-06,
|
| 5265 |
+
"loss": 81.9219,
|
| 5266 |
+
"step": 3715
|
| 5267 |
+
},
|
| 5268 |
+
{
|
| 5269 |
+
"epoch": 0.8599726923993845,
|
| 5270 |
+
"grad_norm": 147.875,
|
| 5271 |
+
"learning_rate": 1.5544707091469683e-06,
|
| 5272 |
+
"loss": 83.8904,
|
| 5273 |
+
"step": 3720
|
| 5274 |
+
},
|
| 5275 |
+
{
|
| 5276 |
+
"epoch": 0.8611285696741149,
|
| 5277 |
+
"grad_norm": 137.125,
|
| 5278 |
+
"learning_rate": 1.5416238437821174e-06,
|
| 5279 |
+
"loss": 82.7793,
|
| 5280 |
+
"step": 3725
|
| 5281 |
+
},
|
| 5282 |
+
{
|
| 5283 |
+
"epoch": 0.8622844469488452,
|
| 5284 |
+
"grad_norm": 146.25,
|
| 5285 |
+
"learning_rate": 1.5287769784172662e-06,
|
| 5286 |
+
"loss": 83.6293,
|
| 5287 |
+
"step": 3730
|
| 5288 |
+
},
|
| 5289 |
+
{
|
| 5290 |
+
"epoch": 0.8634403242235755,
|
| 5291 |
+
"grad_norm": 139.625,
|
| 5292 |
+
"learning_rate": 1.5159301130524153e-06,
|
| 5293 |
+
"loss": 81.7332,
|
| 5294 |
+
"step": 3735
|
| 5295 |
+
},
|
| 5296 |
+
{
|
| 5297 |
+
"epoch": 0.864596201498306,
|
| 5298 |
+
"grad_norm": 152.125,
|
| 5299 |
+
"learning_rate": 1.5030832476875643e-06,
|
| 5300 |
+
"loss": 82.7963,
|
| 5301 |
+
"step": 3740
|
| 5302 |
+
},
|
| 5303 |
+
{
|
| 5304 |
+
"epoch": 0.8657520787730363,
|
| 5305 |
+
"grad_norm": 149.875,
|
| 5306 |
+
"learning_rate": 1.4902363823227134e-06,
|
| 5307 |
+
"loss": 83.1981,
|
| 5308 |
+
"step": 3745
|
| 5309 |
+
},
|
| 5310 |
+
{
|
| 5311 |
+
"epoch": 0.8669079560477666,
|
| 5312 |
+
"grad_norm": 136.0,
|
| 5313 |
+
"learning_rate": 1.4773895169578625e-06,
|
| 5314 |
+
"loss": 83.8111,
|
| 5315 |
+
"step": 3750
|
| 5316 |
+
},
|
| 5317 |
+
{
|
| 5318 |
+
"epoch": 0.868063833322497,
|
| 5319 |
+
"grad_norm": 150.5,
|
| 5320 |
+
"learning_rate": 1.4645426515930113e-06,
|
| 5321 |
+
"loss": 83.2902,
|
| 5322 |
+
"step": 3755
|
| 5323 |
+
},
|
| 5324 |
+
{
|
| 5325 |
+
"epoch": 0.8692197105972274,
|
| 5326 |
+
"grad_norm": 138.375,
|
| 5327 |
+
"learning_rate": 1.4516957862281606e-06,
|
| 5328 |
+
"loss": 84.2568,
|
| 5329 |
+
"step": 3760
|
| 5330 |
+
},
|
| 5331 |
+
{
|
| 5332 |
+
"epoch": 0.8703755878719577,
|
| 5333 |
+
"grad_norm": 139.125,
|
| 5334 |
+
"learning_rate": 1.4388489208633094e-06,
|
| 5335 |
+
"loss": 82.5677,
|
| 5336 |
+
"step": 3765
|
| 5337 |
+
},
|
| 5338 |
+
{
|
| 5339 |
+
"epoch": 0.871531465146688,
|
| 5340 |
+
"grad_norm": 147.0,
|
| 5341 |
+
"learning_rate": 1.4260020554984585e-06,
|
| 5342 |
+
"loss": 82.9927,
|
| 5343 |
+
"step": 3770
|
| 5344 |
+
},
|
| 5345 |
+
{
|
| 5346 |
+
"epoch": 0.8726873424214184,
|
| 5347 |
+
"grad_norm": 149.0,
|
| 5348 |
+
"learning_rate": 1.4131551901336073e-06,
|
| 5349 |
+
"loss": 82.7173,
|
| 5350 |
+
"step": 3775
|
| 5351 |
+
},
|
| 5352 |
+
{
|
| 5353 |
+
"epoch": 0.8738432196961488,
|
| 5354 |
+
"grad_norm": 147.5,
|
| 5355 |
+
"learning_rate": 1.4003083247687566e-06,
|
| 5356 |
+
"loss": 83.3356,
|
| 5357 |
+
"step": 3780
|
| 5358 |
+
},
|
| 5359 |
+
{
|
| 5360 |
+
"epoch": 0.8749990969708791,
|
| 5361 |
+
"grad_norm": 143.625,
|
| 5362 |
+
"learning_rate": 1.3874614594039057e-06,
|
| 5363 |
+
"loss": 81.6103,
|
| 5364 |
+
"step": 3785
|
| 5365 |
+
},
|
| 5366 |
+
{
|
| 5367 |
+
"epoch": 0.8761549742456095,
|
| 5368 |
+
"grad_norm": 140.0,
|
| 5369 |
+
"learning_rate": 1.3746145940390545e-06,
|
| 5370 |
+
"loss": 81.7399,
|
| 5371 |
+
"step": 3790
|
| 5372 |
+
},
|
| 5373 |
+
{
|
| 5374 |
+
"epoch": 0.8773108515203398,
|
| 5375 |
+
"grad_norm": 142.125,
|
| 5376 |
+
"learning_rate": 1.3617677286742038e-06,
|
| 5377 |
+
"loss": 82.5157,
|
| 5378 |
+
"step": 3795
|
| 5379 |
+
},
|
| 5380 |
+
{
|
| 5381 |
+
"epoch": 0.8784667287950702,
|
| 5382 |
+
"grad_norm": 143.625,
|
| 5383 |
+
"learning_rate": 1.3489208633093526e-06,
|
| 5384 |
+
"loss": 83.8571,
|
| 5385 |
+
"step": 3800
|
| 5386 |
+
},
|
| 5387 |
+
{
|
| 5388 |
+
"epoch": 0.8796226060698006,
|
| 5389 |
+
"grad_norm": 140.75,
|
| 5390 |
+
"learning_rate": 1.3360739979445017e-06,
|
| 5391 |
+
"loss": 83.0828,
|
| 5392 |
+
"step": 3805
|
| 5393 |
+
},
|
| 5394 |
+
{
|
| 5395 |
+
"epoch": 0.8807784833445309,
|
| 5396 |
+
"grad_norm": 135.375,
|
| 5397 |
+
"learning_rate": 1.3232271325796505e-06,
|
| 5398 |
+
"loss": 82.2537,
|
| 5399 |
+
"step": 3810
|
| 5400 |
+
},
|
| 5401 |
+
{
|
| 5402 |
+
"epoch": 0.8819343606192612,
|
| 5403 |
+
"grad_norm": 143.75,
|
| 5404 |
+
"learning_rate": 1.3103802672147998e-06,
|
| 5405 |
+
"loss": 82.8528,
|
| 5406 |
+
"step": 3815
|
| 5407 |
+
},
|
| 5408 |
+
{
|
| 5409 |
+
"epoch": 0.8830902378939917,
|
| 5410 |
+
"grad_norm": 140.0,
|
| 5411 |
+
"learning_rate": 1.2975334018499486e-06,
|
| 5412 |
+
"loss": 83.4548,
|
| 5413 |
+
"step": 3820
|
| 5414 |
+
},
|
| 5415 |
+
{
|
| 5416 |
+
"epoch": 0.884246115168722,
|
| 5417 |
+
"grad_norm": 137.5,
|
| 5418 |
+
"learning_rate": 1.2846865364850977e-06,
|
| 5419 |
+
"loss": 82.9603,
|
| 5420 |
+
"step": 3825
|
| 5421 |
+
},
|
| 5422 |
+
{
|
| 5423 |
+
"epoch": 0.8854019924434523,
|
| 5424 |
+
"grad_norm": 143.75,
|
| 5425 |
+
"learning_rate": 1.2718396711202468e-06,
|
| 5426 |
+
"loss": 82.612,
|
| 5427 |
+
"step": 3830
|
| 5428 |
+
},
|
| 5429 |
+
{
|
| 5430 |
+
"epoch": 0.8865578697181826,
|
| 5431 |
+
"grad_norm": 134.75,
|
| 5432 |
+
"learning_rate": 1.2589928057553958e-06,
|
| 5433 |
+
"loss": 81.9109,
|
| 5434 |
+
"step": 3835
|
| 5435 |
+
},
|
| 5436 |
+
{
|
| 5437 |
+
"epoch": 0.887713746992913,
|
| 5438 |
+
"grad_norm": 149.875,
|
| 5439 |
+
"learning_rate": 1.2461459403905449e-06,
|
| 5440 |
+
"loss": 83.5588,
|
| 5441 |
+
"step": 3840
|
| 5442 |
+
},
|
| 5443 |
+
{
|
| 5444 |
+
"epoch": 0.8888696242676434,
|
| 5445 |
+
"grad_norm": 140.125,
|
| 5446 |
+
"learning_rate": 1.233299075025694e-06,
|
| 5447 |
+
"loss": 81.9114,
|
| 5448 |
+
"step": 3845
|
| 5449 |
+
},
|
| 5450 |
+
{
|
| 5451 |
+
"epoch": 0.8900255015423737,
|
| 5452 |
+
"grad_norm": 140.875,
|
| 5453 |
+
"learning_rate": 1.2204522096608428e-06,
|
| 5454 |
+
"loss": 82.8397,
|
| 5455 |
+
"step": 3850
|
| 5456 |
+
},
|
| 5457 |
+
{
|
| 5458 |
+
"epoch": 0.8911813788171041,
|
| 5459 |
+
"grad_norm": 133.75,
|
| 5460 |
+
"learning_rate": 1.2076053442959918e-06,
|
| 5461 |
+
"loss": 83.9099,
|
| 5462 |
+
"step": 3855
|
| 5463 |
+
},
|
| 5464 |
+
{
|
| 5465 |
+
"epoch": 0.8923372560918345,
|
| 5466 |
+
"grad_norm": 142.75,
|
| 5467 |
+
"learning_rate": 1.194758478931141e-06,
|
| 5468 |
+
"loss": 83.4947,
|
| 5469 |
+
"step": 3860
|
| 5470 |
+
},
|
| 5471 |
+
{
|
| 5472 |
+
"epoch": 0.8934931333665648,
|
| 5473 |
+
"grad_norm": 144.25,
|
| 5474 |
+
"learning_rate": 1.18191161356629e-06,
|
| 5475 |
+
"loss": 83.0275,
|
| 5476 |
+
"step": 3865
|
| 5477 |
+
},
|
| 5478 |
+
{
|
| 5479 |
+
"epoch": 0.8946490106412952,
|
| 5480 |
+
"grad_norm": 157.0,
|
| 5481 |
+
"learning_rate": 1.1690647482014388e-06,
|
| 5482 |
+
"loss": 81.9333,
|
| 5483 |
+
"step": 3870
|
| 5484 |
+
},
|
| 5485 |
+
{
|
| 5486 |
+
"epoch": 0.8958048879160255,
|
| 5487 |
+
"grad_norm": 136.5,
|
| 5488 |
+
"learning_rate": 1.156217882836588e-06,
|
| 5489 |
+
"loss": 81.9801,
|
| 5490 |
+
"step": 3875
|
| 5491 |
+
},
|
| 5492 |
+
{
|
| 5493 |
+
"epoch": 0.8969607651907558,
|
| 5494 |
+
"grad_norm": 141.375,
|
| 5495 |
+
"learning_rate": 1.1433710174717371e-06,
|
| 5496 |
+
"loss": 82.9318,
|
| 5497 |
+
"step": 3880
|
| 5498 |
+
},
|
| 5499 |
+
{
|
| 5500 |
+
"epoch": 0.8981166424654863,
|
| 5501 |
+
"grad_norm": 140.625,
|
| 5502 |
+
"learning_rate": 1.130524152106886e-06,
|
| 5503 |
+
"loss": 82.2715,
|
| 5504 |
+
"step": 3885
|
| 5505 |
+
},
|
| 5506 |
+
{
|
| 5507 |
+
"epoch": 0.8992725197402166,
|
| 5508 |
+
"grad_norm": 139.25,
|
| 5509 |
+
"learning_rate": 1.117677286742035e-06,
|
| 5510 |
+
"loss": 82.4903,
|
| 5511 |
+
"step": 3890
|
| 5512 |
+
},
|
| 5513 |
+
{
|
| 5514 |
+
"epoch": 0.9004283970149469,
|
| 5515 |
+
"grad_norm": 140.0,
|
| 5516 |
+
"learning_rate": 1.1048304213771841e-06,
|
| 5517 |
+
"loss": 82.3743,
|
| 5518 |
+
"step": 3895
|
| 5519 |
+
},
|
| 5520 |
+
{
|
| 5521 |
+
"epoch": 0.9015842742896772,
|
| 5522 |
+
"grad_norm": 141.875,
|
| 5523 |
+
"learning_rate": 1.091983556012333e-06,
|
| 5524 |
+
"loss": 79.7922,
|
| 5525 |
+
"step": 3900
|
| 5526 |
+
},
|
| 5527 |
+
{
|
| 5528 |
+
"epoch": 0.9027401515644077,
|
| 5529 |
+
"grad_norm": 140.125,
|
| 5530 |
+
"learning_rate": 1.079136690647482e-06,
|
| 5531 |
+
"loss": 80.9088,
|
| 5532 |
+
"step": 3905
|
| 5533 |
+
},
|
| 5534 |
+
{
|
| 5535 |
+
"epoch": 0.903896028839138,
|
| 5536 |
+
"grad_norm": 139.5,
|
| 5537 |
+
"learning_rate": 1.066289825282631e-06,
|
| 5538 |
+
"loss": 82.9754,
|
| 5539 |
+
"step": 3910
|
| 5540 |
+
},
|
| 5541 |
+
{
|
| 5542 |
+
"epoch": 0.9050519061138683,
|
| 5543 |
+
"grad_norm": 135.25,
|
| 5544 |
+
"learning_rate": 1.0534429599177801e-06,
|
| 5545 |
+
"loss": 82.7032,
|
| 5546 |
+
"step": 3915
|
| 5547 |
+
},
|
| 5548 |
+
{
|
| 5549 |
+
"epoch": 0.9062077833885988,
|
| 5550 |
+
"grad_norm": 138.5,
|
| 5551 |
+
"learning_rate": 1.0405960945529292e-06,
|
| 5552 |
+
"loss": 80.908,
|
| 5553 |
+
"step": 3920
|
| 5554 |
+
},
|
| 5555 |
+
{
|
| 5556 |
+
"epoch": 0.9073636606633291,
|
| 5557 |
+
"grad_norm": 151.25,
|
| 5558 |
+
"learning_rate": 1.0277492291880783e-06,
|
| 5559 |
+
"loss": 82.3395,
|
| 5560 |
+
"step": 3925
|
| 5561 |
+
},
|
| 5562 |
+
{
|
| 5563 |
+
"epoch": 0.9085195379380594,
|
| 5564 |
+
"grad_norm": 141.625,
|
| 5565 |
+
"learning_rate": 1.0149023638232273e-06,
|
| 5566 |
+
"loss": 81.7353,
|
| 5567 |
+
"step": 3930
|
| 5568 |
+
},
|
| 5569 |
+
{
|
| 5570 |
+
"epoch": 0.9096754152127898,
|
| 5571 |
+
"grad_norm": 154.375,
|
| 5572 |
+
"learning_rate": 1.0020554984583762e-06,
|
| 5573 |
+
"loss": 81.4314,
|
| 5574 |
+
"step": 3935
|
| 5575 |
+
},
|
| 5576 |
+
{
|
| 5577 |
+
"epoch": 0.9108312924875201,
|
| 5578 |
+
"grad_norm": 137.625,
|
| 5579 |
+
"learning_rate": 9.892086330935252e-07,
|
| 5580 |
+
"loss": 82.1363,
|
| 5581 |
+
"step": 3940
|
| 5582 |
+
},
|
| 5583 |
+
{
|
| 5584 |
+
"epoch": 0.9119871697622505,
|
| 5585 |
+
"grad_norm": 141.0,
|
| 5586 |
+
"learning_rate": 9.763617677286743e-07,
|
| 5587 |
+
"loss": 83.2808,
|
| 5588 |
+
"step": 3945
|
| 5589 |
+
},
|
| 5590 |
+
{
|
| 5591 |
+
"epoch": 0.9131430470369809,
|
| 5592 |
+
"grad_norm": 140.25,
|
| 5593 |
+
"learning_rate": 9.635149023638233e-07,
|
| 5594 |
+
"loss": 81.7309,
|
| 5595 |
+
"step": 3950
|
| 5596 |
+
},
|
| 5597 |
+
{
|
| 5598 |
+
"epoch": 0.9142989243117112,
|
| 5599 |
+
"grad_norm": 136.125,
|
| 5600 |
+
"learning_rate": 9.506680369989724e-07,
|
| 5601 |
+
"loss": 80.9205,
|
| 5602 |
+
"step": 3955
|
| 5603 |
+
},
|
| 5604 |
+
{
|
| 5605 |
+
"epoch": 0.9154548015864415,
|
| 5606 |
+
"grad_norm": 140.25,
|
| 5607 |
+
"learning_rate": 9.378211716341213e-07,
|
| 5608 |
+
"loss": 83.6719,
|
| 5609 |
+
"step": 3960
|
| 5610 |
+
},
|
| 5611 |
+
{
|
| 5612 |
+
"epoch": 0.9166106788611719,
|
| 5613 |
+
"grad_norm": 142.5,
|
| 5614 |
+
"learning_rate": 9.249743062692704e-07,
|
| 5615 |
+
"loss": 82.9615,
|
| 5616 |
+
"step": 3965
|
| 5617 |
+
},
|
| 5618 |
+
{
|
| 5619 |
+
"epoch": 0.9177665561359023,
|
| 5620 |
+
"grad_norm": 137.125,
|
| 5621 |
+
"learning_rate": 9.121274409044194e-07,
|
| 5622 |
+
"loss": 83.3444,
|
| 5623 |
+
"step": 3970
|
| 5624 |
+
},
|
| 5625 |
+
{
|
| 5626 |
+
"epoch": 0.9189224334106326,
|
| 5627 |
+
"grad_norm": 137.75,
|
| 5628 |
+
"learning_rate": 8.992805755395684e-07,
|
| 5629 |
+
"loss": 82.6694,
|
| 5630 |
+
"step": 3975
|
| 5631 |
+
},
|
| 5632 |
+
{
|
| 5633 |
+
"epoch": 0.9200783106853629,
|
| 5634 |
+
"grad_norm": 142.0,
|
| 5635 |
+
"learning_rate": 8.864337101747174e-07,
|
| 5636 |
+
"loss": 81.7997,
|
| 5637 |
+
"step": 3980
|
| 5638 |
+
},
|
| 5639 |
+
{
|
| 5640 |
+
"epoch": 0.9212341879600934,
|
| 5641 |
+
"grad_norm": 138.875,
|
| 5642 |
+
"learning_rate": 8.735868448098665e-07,
|
| 5643 |
+
"loss": 81.687,
|
| 5644 |
+
"step": 3985
|
| 5645 |
+
},
|
| 5646 |
+
{
|
| 5647 |
+
"epoch": 0.9223900652348237,
|
| 5648 |
+
"grad_norm": 135.0,
|
| 5649 |
+
"learning_rate": 8.607399794450155e-07,
|
| 5650 |
+
"loss": 81.6286,
|
| 5651 |
+
"step": 3990
|
| 5652 |
+
},
|
| 5653 |
+
{
|
| 5654 |
+
"epoch": 0.923545942509554,
|
| 5655 |
+
"grad_norm": 140.625,
|
| 5656 |
+
"learning_rate": 8.478931140801645e-07,
|
| 5657 |
+
"loss": 80.7711,
|
| 5658 |
+
"step": 3995
|
| 5659 |
+
},
|
| 5660 |
+
{
|
| 5661 |
+
"epoch": 0.9247018197842845,
|
| 5662 |
+
"grad_norm": 149.375,
|
| 5663 |
+
"learning_rate": 8.350462487153135e-07,
|
| 5664 |
+
"loss": 83.3313,
|
| 5665 |
+
"step": 4000
|
| 5666 |
+
},
|
| 5667 |
+
{
|
| 5668 |
+
"epoch": 0.9247018197842845,
|
| 5669 |
+
"eval_loss": NaN,
|
| 5670 |
+
"eval_runtime": 381.0966,
|
| 5671 |
+
"eval_samples_per_second": 611.745,
|
| 5672 |
+
"eval_steps_per_second": 38.234,
|
| 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
|