Training in progress, step 690, checkpoint
Browse files
last-checkpoint/adapter_model.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 4731640
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:fb3724c874ea65782d8dcd4cb5c3209c89ff98f654f1190a20ab5fa9de8602ff
|
3 |
size 4731640
|
last-checkpoint/optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 2505850
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:bbcbb0ae8f5791ad43b6585f30b8675ce7b11665e8aa85d4cd2b157c82c0a59d
|
3 |
size 2505850
|
last-checkpoint/rng_state.pth
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 14244
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:eccaa3eca2e911dfe6c95134e6dab45713a2d23ff2880dd37ecc335fd49e8a8f
|
3 |
size 14244
|
last-checkpoint/scheduler.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 1064
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7e4c2fbcfa9fc7eb2f52e6e150e8b4f82c141787d028ff20f699dcb142cbe5fb
|
3 |
size 1064
|
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": 230,
|
6 |
-
"global_step":
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
@@ -3251,6 +3251,1624 @@
|
|
3251 |
"eval_samples_per_second": 139.095,
|
3252 |
"eval_steps_per_second": 69.727,
|
3253 |
"step": 460
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
3254 |
}
|
3255 |
],
|
3256 |
"logging_steps": 1,
|
@@ -3270,7 +4888,7 @@
|
|
3270 |
"attributes": {}
|
3271 |
}
|
3272 |
},
|
3273 |
-
"total_flos":
|
3274 |
"train_batch_size": 2,
|
3275 |
"trial_name": null,
|
3276 |
"trial_params": null
|
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
+
"epoch": 0.7516339869281046,
|
5 |
"eval_steps": 230,
|
6 |
+
"global_step": 690,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
|
|
3251 |
"eval_samples_per_second": 139.095,
|
3252 |
"eval_steps_per_second": 69.727,
|
3253 |
"step": 460
|
3254 |
+
},
|
3255 |
+
{
|
3256 |
+
"epoch": 0.5021786492374728,
|
3257 |
+
"grad_norm": 1697.1995849609375,
|
3258 |
+
"learning_rate": 0.00010103795250292778,
|
3259 |
+
"loss": 15.0099,
|
3260 |
+
"step": 461
|
3261 |
+
},
|
3262 |
+
{
|
3263 |
+
"epoch": 0.5032679738562091,
|
3264 |
+
"grad_norm": 1438.0911865234375,
|
3265 |
+
"learning_rate": 0.00010069197523827833,
|
3266 |
+
"loss": 16.3135,
|
3267 |
+
"step": 462
|
3268 |
+
},
|
3269 |
+
{
|
3270 |
+
"epoch": 0.5043572984749455,
|
3271 |
+
"grad_norm": 1578.0306396484375,
|
3272 |
+
"learning_rate": 0.00010034598969004705,
|
3273 |
+
"loss": 14.3354,
|
3274 |
+
"step": 463
|
3275 |
+
},
|
3276 |
+
{
|
3277 |
+
"epoch": 0.5054466230936819,
|
3278 |
+
"grad_norm": 1145.7254638671875,
|
3279 |
+
"learning_rate": 0.0001,
|
3280 |
+
"loss": 15.1201,
|
3281 |
+
"step": 464
|
3282 |
+
},
|
3283 |
+
{
|
3284 |
+
"epoch": 0.5065359477124183,
|
3285 |
+
"grad_norm": 1281.2977294921875,
|
3286 |
+
"learning_rate": 9.965401030995301e-05,
|
3287 |
+
"loss": 15.0423,
|
3288 |
+
"step": 465
|
3289 |
+
},
|
3290 |
+
{
|
3291 |
+
"epoch": 0.5076252723311547,
|
3292 |
+
"grad_norm": 2497.565673828125,
|
3293 |
+
"learning_rate": 9.930802476172169e-05,
|
3294 |
+
"loss": 12.9232,
|
3295 |
+
"step": 466
|
3296 |
+
},
|
3297 |
+
{
|
3298 |
+
"epoch": 0.5087145969498911,
|
3299 |
+
"grad_norm": 1575.5599365234375,
|
3300 |
+
"learning_rate": 9.896204749707228e-05,
|
3301 |
+
"loss": 14.6681,
|
3302 |
+
"step": 467
|
3303 |
+
},
|
3304 |
+
{
|
3305 |
+
"epoch": 0.5098039215686274,
|
3306 |
+
"grad_norm": 941.0279541015625,
|
3307 |
+
"learning_rate": 9.861608265767167e-05,
|
3308 |
+
"loss": 14.7297,
|
3309 |
+
"step": 468
|
3310 |
+
},
|
3311 |
+
{
|
3312 |
+
"epoch": 0.5108932461873639,
|
3313 |
+
"grad_norm": 1556.48388671875,
|
3314 |
+
"learning_rate": 9.827013438503822e-05,
|
3315 |
+
"loss": 14.6033,
|
3316 |
+
"step": 469
|
3317 |
+
},
|
3318 |
+
{
|
3319 |
+
"epoch": 0.5119825708061002,
|
3320 |
+
"grad_norm": 1751.578125,
|
3321 |
+
"learning_rate": 9.792420682049174e-05,
|
3322 |
+
"loss": 15.0415,
|
3323 |
+
"step": 470
|
3324 |
+
},
|
3325 |
+
{
|
3326 |
+
"epoch": 0.5130718954248366,
|
3327 |
+
"grad_norm": 1077.553955078125,
|
3328 |
+
"learning_rate": 9.757830410510433e-05,
|
3329 |
+
"loss": 13.5249,
|
3330 |
+
"step": 471
|
3331 |
+
},
|
3332 |
+
{
|
3333 |
+
"epoch": 0.514161220043573,
|
3334 |
+
"grad_norm": 1023.6539916992188,
|
3335 |
+
"learning_rate": 9.723243037965056e-05,
|
3336 |
+
"loss": 13.8337,
|
3337 |
+
"step": 472
|
3338 |
+
},
|
3339 |
+
{
|
3340 |
+
"epoch": 0.5152505446623094,
|
3341 |
+
"grad_norm": 1367.58984375,
|
3342 |
+
"learning_rate": 9.688658978455784e-05,
|
3343 |
+
"loss": 15.3791,
|
3344 |
+
"step": 473
|
3345 |
+
},
|
3346 |
+
{
|
3347 |
+
"epoch": 0.5163398692810458,
|
3348 |
+
"grad_norm": 1138.333740234375,
|
3349 |
+
"learning_rate": 9.654078645985722e-05,
|
3350 |
+
"loss": 14.343,
|
3351 |
+
"step": 474
|
3352 |
+
},
|
3353 |
+
{
|
3354 |
+
"epoch": 0.5174291938997821,
|
3355 |
+
"grad_norm": 958.8590698242188,
|
3356 |
+
"learning_rate": 9.619502454513338e-05,
|
3357 |
+
"loss": 14.1471,
|
3358 |
+
"step": 475
|
3359 |
+
},
|
3360 |
+
{
|
3361 |
+
"epoch": 0.5185185185185185,
|
3362 |
+
"grad_norm": 1362.31689453125,
|
3363 |
+
"learning_rate": 9.584930817947544e-05,
|
3364 |
+
"loss": 14.2489,
|
3365 |
+
"step": 476
|
3366 |
+
},
|
3367 |
+
{
|
3368 |
+
"epoch": 0.5196078431372549,
|
3369 |
+
"grad_norm": 1034.6534423828125,
|
3370 |
+
"learning_rate": 9.550364150142713e-05,
|
3371 |
+
"loss": 13.4652,
|
3372 |
+
"step": 477
|
3373 |
+
},
|
3374 |
+
{
|
3375 |
+
"epoch": 0.5206971677559913,
|
3376 |
+
"grad_norm": 1429.5125732421875,
|
3377 |
+
"learning_rate": 9.515802864893739e-05,
|
3378 |
+
"loss": 14.1628,
|
3379 |
+
"step": 478
|
3380 |
+
},
|
3381 |
+
{
|
3382 |
+
"epoch": 0.5217864923747276,
|
3383 |
+
"grad_norm": 1044.7430419921875,
|
3384 |
+
"learning_rate": 9.481247375931094e-05,
|
3385 |
+
"loss": 13.1049,
|
3386 |
+
"step": 479
|
3387 |
+
},
|
3388 |
+
{
|
3389 |
+
"epoch": 0.5228758169934641,
|
3390 |
+
"grad_norm": 1164.5643310546875,
|
3391 |
+
"learning_rate": 9.446698096915847e-05,
|
3392 |
+
"loss": 13.8229,
|
3393 |
+
"step": 480
|
3394 |
+
},
|
3395 |
+
{
|
3396 |
+
"epoch": 0.5239651416122004,
|
3397 |
+
"grad_norm": 1188.510986328125,
|
3398 |
+
"learning_rate": 9.412155441434741e-05,
|
3399 |
+
"loss": 13.4517,
|
3400 |
+
"step": 481
|
3401 |
+
},
|
3402 |
+
{
|
3403 |
+
"epoch": 0.5250544662309368,
|
3404 |
+
"grad_norm": 1218.689208984375,
|
3405 |
+
"learning_rate": 9.377619822995219e-05,
|
3406 |
+
"loss": 15.032,
|
3407 |
+
"step": 482
|
3408 |
+
},
|
3409 |
+
{
|
3410 |
+
"epoch": 0.5261437908496732,
|
3411 |
+
"grad_norm": 802.5537109375,
|
3412 |
+
"learning_rate": 9.343091655020495e-05,
|
3413 |
+
"loss": 13.9797,
|
3414 |
+
"step": 483
|
3415 |
+
},
|
3416 |
+
{
|
3417 |
+
"epoch": 0.5272331154684096,
|
3418 |
+
"grad_norm": 938.1593627929688,
|
3419 |
+
"learning_rate": 9.308571350844584e-05,
|
3420 |
+
"loss": 13.8941,
|
3421 |
+
"step": 484
|
3422 |
+
},
|
3423 |
+
{
|
3424 |
+
"epoch": 0.528322440087146,
|
3425 |
+
"grad_norm": 1472.3194580078125,
|
3426 |
+
"learning_rate": 9.274059323707366e-05,
|
3427 |
+
"loss": 13.2638,
|
3428 |
+
"step": 485
|
3429 |
+
},
|
3430 |
+
{
|
3431 |
+
"epoch": 0.5294117647058824,
|
3432 |
+
"grad_norm": 1048.2354736328125,
|
3433 |
+
"learning_rate": 9.239555986749645e-05,
|
3434 |
+
"loss": 14.3531,
|
3435 |
+
"step": 486
|
3436 |
+
},
|
3437 |
+
{
|
3438 |
+
"epoch": 0.5305010893246187,
|
3439 |
+
"grad_norm": 1296.9527587890625,
|
3440 |
+
"learning_rate": 9.205061753008183e-05,
|
3441 |
+
"loss": 14.203,
|
3442 |
+
"step": 487
|
3443 |
+
},
|
3444 |
+
{
|
3445 |
+
"epoch": 0.5315904139433552,
|
3446 |
+
"grad_norm": 1657.743896484375,
|
3447 |
+
"learning_rate": 9.170577035410783e-05,
|
3448 |
+
"loss": 13.7112,
|
3449 |
+
"step": 488
|
3450 |
+
},
|
3451 |
+
{
|
3452 |
+
"epoch": 0.5326797385620915,
|
3453 |
+
"grad_norm": 1212.4053955078125,
|
3454 |
+
"learning_rate": 9.136102246771314e-05,
|
3455 |
+
"loss": 13.2756,
|
3456 |
+
"step": 489
|
3457 |
+
},
|
3458 |
+
{
|
3459 |
+
"epoch": 0.5337690631808278,
|
3460 |
+
"grad_norm": 1304.51220703125,
|
3461 |
+
"learning_rate": 9.101637799784804e-05,
|
3462 |
+
"loss": 14.4237,
|
3463 |
+
"step": 490
|
3464 |
+
},
|
3465 |
+
{
|
3466 |
+
"epoch": 0.5348583877995643,
|
3467 |
+
"grad_norm": 1542.832275390625,
|
3468 |
+
"learning_rate": 9.06718410702247e-05,
|
3469 |
+
"loss": 14.4378,
|
3470 |
+
"step": 491
|
3471 |
+
},
|
3472 |
+
{
|
3473 |
+
"epoch": 0.5359477124183006,
|
3474 |
+
"grad_norm": 1302.9571533203125,
|
3475 |
+
"learning_rate": 9.032741580926787e-05,
|
3476 |
+
"loss": 13.1055,
|
3477 |
+
"step": 492
|
3478 |
+
},
|
3479 |
+
{
|
3480 |
+
"epoch": 0.5370370370370371,
|
3481 |
+
"grad_norm": 1939.5728759765625,
|
3482 |
+
"learning_rate": 8.998310633806571e-05,
|
3483 |
+
"loss": 12.4175,
|
3484 |
+
"step": 493
|
3485 |
+
},
|
3486 |
+
{
|
3487 |
+
"epoch": 0.5381263616557734,
|
3488 |
+
"grad_norm": 1355.798095703125,
|
3489 |
+
"learning_rate": 8.963891677832011e-05,
|
3490 |
+
"loss": 12.9592,
|
3491 |
+
"step": 494
|
3492 |
+
},
|
3493 |
+
{
|
3494 |
+
"epoch": 0.5392156862745098,
|
3495 |
+
"grad_norm": 2074.481201171875,
|
3496 |
+
"learning_rate": 8.929485125029766e-05,
|
3497 |
+
"loss": 13.6742,
|
3498 |
+
"step": 495
|
3499 |
+
},
|
3500 |
+
{
|
3501 |
+
"epoch": 0.5403050108932462,
|
3502 |
+
"grad_norm": 1087.583740234375,
|
3503 |
+
"learning_rate": 8.895091387277999e-05,
|
3504 |
+
"loss": 14.9366,
|
3505 |
+
"step": 496
|
3506 |
+
},
|
3507 |
+
{
|
3508 |
+
"epoch": 0.5413943355119826,
|
3509 |
+
"grad_norm": 1182.4432373046875,
|
3510 |
+
"learning_rate": 8.860710876301484e-05,
|
3511 |
+
"loss": 12.0956,
|
3512 |
+
"step": 497
|
3513 |
+
},
|
3514 |
+
{
|
3515 |
+
"epoch": 0.5424836601307189,
|
3516 |
+
"grad_norm": 1062.96923828125,
|
3517 |
+
"learning_rate": 8.826344003666647e-05,
|
3518 |
+
"loss": 13.0663,
|
3519 |
+
"step": 498
|
3520 |
+
},
|
3521 |
+
{
|
3522 |
+
"epoch": 0.5435729847494554,
|
3523 |
+
"grad_norm": 1257.4434814453125,
|
3524 |
+
"learning_rate": 8.791991180776648e-05,
|
3525 |
+
"loss": 14.0272,
|
3526 |
+
"step": 499
|
3527 |
+
},
|
3528 |
+
{
|
3529 |
+
"epoch": 0.5446623093681917,
|
3530 |
+
"grad_norm": 1445.0548095703125,
|
3531 |
+
"learning_rate": 8.757652818866471e-05,
|
3532 |
+
"loss": 12.8504,
|
3533 |
+
"step": 500
|
3534 |
+
},
|
3535 |
+
{
|
3536 |
+
"epoch": 0.545751633986928,
|
3537 |
+
"grad_norm": 1665.5859375,
|
3538 |
+
"learning_rate": 8.723329328997973e-05,
|
3539 |
+
"loss": 13.8356,
|
3540 |
+
"step": 501
|
3541 |
+
},
|
3542 |
+
{
|
3543 |
+
"epoch": 0.5468409586056645,
|
3544 |
+
"grad_norm": 995.88427734375,
|
3545 |
+
"learning_rate": 8.689021122054996e-05,
|
3546 |
+
"loss": 12.8165,
|
3547 |
+
"step": 502
|
3548 |
+
},
|
3549 |
+
{
|
3550 |
+
"epoch": 0.5479302832244008,
|
3551 |
+
"grad_norm": 1658.434326171875,
|
3552 |
+
"learning_rate": 8.654728608738418e-05,
|
3553 |
+
"loss": 14.196,
|
3554 |
+
"step": 503
|
3555 |
+
},
|
3556 |
+
{
|
3557 |
+
"epoch": 0.5490196078431373,
|
3558 |
+
"grad_norm": 984.2662963867188,
|
3559 |
+
"learning_rate": 8.620452199561254e-05,
|
3560 |
+
"loss": 13.6334,
|
3561 |
+
"step": 504
|
3562 |
+
},
|
3563 |
+
{
|
3564 |
+
"epoch": 0.5501089324618736,
|
3565 |
+
"grad_norm": 1411.3946533203125,
|
3566 |
+
"learning_rate": 8.58619230484374e-05,
|
3567 |
+
"loss": 13.7294,
|
3568 |
+
"step": 505
|
3569 |
+
},
|
3570 |
+
{
|
3571 |
+
"epoch": 0.55119825708061,
|
3572 |
+
"grad_norm": 1083.3192138671875,
|
3573 |
+
"learning_rate": 8.551949334708415e-05,
|
3574 |
+
"loss": 12.9704,
|
3575 |
+
"step": 506
|
3576 |
+
},
|
3577 |
+
{
|
3578 |
+
"epoch": 0.5522875816993464,
|
3579 |
+
"grad_norm": 1257.832275390625,
|
3580 |
+
"learning_rate": 8.51772369907522e-05,
|
3581 |
+
"loss": 12.5718,
|
3582 |
+
"step": 507
|
3583 |
+
},
|
3584 |
+
{
|
3585 |
+
"epoch": 0.5533769063180828,
|
3586 |
+
"grad_norm": 871.9485473632812,
|
3587 |
+
"learning_rate": 8.483515807656576e-05,
|
3588 |
+
"loss": 13.1949,
|
3589 |
+
"step": 508
|
3590 |
+
},
|
3591 |
+
{
|
3592 |
+
"epoch": 0.5544662309368191,
|
3593 |
+
"grad_norm": 1094.6038818359375,
|
3594 |
+
"learning_rate": 8.449326069952506e-05,
|
3595 |
+
"loss": 14.2022,
|
3596 |
+
"step": 509
|
3597 |
+
},
|
3598 |
+
{
|
3599 |
+
"epoch": 0.5555555555555556,
|
3600 |
+
"grad_norm": 1254.0933837890625,
|
3601 |
+
"learning_rate": 8.415154895245697e-05,
|
3602 |
+
"loss": 13.1929,
|
3603 |
+
"step": 510
|
3604 |
+
},
|
3605 |
+
{
|
3606 |
+
"epoch": 0.5566448801742919,
|
3607 |
+
"grad_norm": 819.5022583007812,
|
3608 |
+
"learning_rate": 8.381002692596635e-05,
|
3609 |
+
"loss": 13.739,
|
3610 |
+
"step": 511
|
3611 |
+
},
|
3612 |
+
{
|
3613 |
+
"epoch": 0.5577342047930284,
|
3614 |
+
"grad_norm": 1080.34521484375,
|
3615 |
+
"learning_rate": 8.346869870838685e-05,
|
3616 |
+
"loss": 12.4172,
|
3617 |
+
"step": 512
|
3618 |
+
},
|
3619 |
+
{
|
3620 |
+
"epoch": 0.5588235294117647,
|
3621 |
+
"grad_norm": 844.1549072265625,
|
3622 |
+
"learning_rate": 8.312756838573208e-05,
|
3623 |
+
"loss": 14.1843,
|
3624 |
+
"step": 513
|
3625 |
+
},
|
3626 |
+
{
|
3627 |
+
"epoch": 0.5599128540305011,
|
3628 |
+
"grad_norm": 1664.24462890625,
|
3629 |
+
"learning_rate": 8.278664004164665e-05,
|
3630 |
+
"loss": 13.9817,
|
3631 |
+
"step": 514
|
3632 |
+
},
|
3633 |
+
{
|
3634 |
+
"epoch": 0.5610021786492375,
|
3635 |
+
"grad_norm": 1837.83544921875,
|
3636 |
+
"learning_rate": 8.244591775735732e-05,
|
3637 |
+
"loss": 14.2748,
|
3638 |
+
"step": 515
|
3639 |
+
},
|
3640 |
+
{
|
3641 |
+
"epoch": 0.5620915032679739,
|
3642 |
+
"grad_norm": 794.2677612304688,
|
3643 |
+
"learning_rate": 8.210540561162412e-05,
|
3644 |
+
"loss": 13.3806,
|
3645 |
+
"step": 516
|
3646 |
+
},
|
3647 |
+
{
|
3648 |
+
"epoch": 0.5631808278867102,
|
3649 |
+
"grad_norm": 1501.6959228515625,
|
3650 |
+
"learning_rate": 8.176510768069147e-05,
|
3651 |
+
"loss": 13.5997,
|
3652 |
+
"step": 517
|
3653 |
+
},
|
3654 |
+
{
|
3655 |
+
"epoch": 0.5642701525054467,
|
3656 |
+
"grad_norm": 1058.7418212890625,
|
3657 |
+
"learning_rate": 8.142502803823955e-05,
|
3658 |
+
"loss": 12.9171,
|
3659 |
+
"step": 518
|
3660 |
+
},
|
3661 |
+
{
|
3662 |
+
"epoch": 0.565359477124183,
|
3663 |
+
"grad_norm": 1096.179443359375,
|
3664 |
+
"learning_rate": 8.108517075533531e-05,
|
3665 |
+
"loss": 13.1182,
|
3666 |
+
"step": 519
|
3667 |
+
},
|
3668 |
+
{
|
3669 |
+
"epoch": 0.5664488017429193,
|
3670 |
+
"grad_norm": 1459.3277587890625,
|
3671 |
+
"learning_rate": 8.074553990038395e-05,
|
3672 |
+
"loss": 14.0618,
|
3673 |
+
"step": 520
|
3674 |
+
},
|
3675 |
+
{
|
3676 |
+
"epoch": 0.5675381263616558,
|
3677 |
+
"grad_norm": 1117.51318359375,
|
3678 |
+
"learning_rate": 8.040613953908005e-05,
|
3679 |
+
"loss": 12.8453,
|
3680 |
+
"step": 521
|
3681 |
+
},
|
3682 |
+
{
|
3683 |
+
"epoch": 0.5686274509803921,
|
3684 |
+
"grad_norm": 1164.1474609375,
|
3685 |
+
"learning_rate": 8.0066973734359e-05,
|
3686 |
+
"loss": 13.228,
|
3687 |
+
"step": 522
|
3688 |
+
},
|
3689 |
+
{
|
3690 |
+
"epoch": 0.5697167755991286,
|
3691 |
+
"grad_norm": 1310.04296875,
|
3692 |
+
"learning_rate": 7.972804654634834e-05,
|
3693 |
+
"loss": 13.3479,
|
3694 |
+
"step": 523
|
3695 |
+
},
|
3696 |
+
{
|
3697 |
+
"epoch": 0.5708061002178649,
|
3698 |
+
"grad_norm": 1271.54833984375,
|
3699 |
+
"learning_rate": 7.938936203231912e-05,
|
3700 |
+
"loss": 12.7188,
|
3701 |
+
"step": 524
|
3702 |
+
},
|
3703 |
+
{
|
3704 |
+
"epoch": 0.5718954248366013,
|
3705 |
+
"grad_norm": 1231.713623046875,
|
3706 |
+
"learning_rate": 7.905092424663735e-05,
|
3707 |
+
"loss": 14.596,
|
3708 |
+
"step": 525
|
3709 |
+
},
|
3710 |
+
{
|
3711 |
+
"epoch": 0.5729847494553377,
|
3712 |
+
"grad_norm": 2087.1767578125,
|
3713 |
+
"learning_rate": 7.871273724071553e-05,
|
3714 |
+
"loss": 12.5966,
|
3715 |
+
"step": 526
|
3716 |
+
},
|
3717 |
+
{
|
3718 |
+
"epoch": 0.5740740740740741,
|
3719 |
+
"grad_norm": 1349.419189453125,
|
3720 |
+
"learning_rate": 7.837480506296404e-05,
|
3721 |
+
"loss": 12.99,
|
3722 |
+
"step": 527
|
3723 |
+
},
|
3724 |
+
{
|
3725 |
+
"epoch": 0.5751633986928104,
|
3726 |
+
"grad_norm": 1333.2598876953125,
|
3727 |
+
"learning_rate": 7.803713175874275e-05,
|
3728 |
+
"loss": 12.8456,
|
3729 |
+
"step": 528
|
3730 |
+
},
|
3731 |
+
{
|
3732 |
+
"epoch": 0.5762527233115469,
|
3733 |
+
"grad_norm": 1339.005615234375,
|
3734 |
+
"learning_rate": 7.769972137031262e-05,
|
3735 |
+
"loss": 14.4523,
|
3736 |
+
"step": 529
|
3737 |
+
},
|
3738 |
+
{
|
3739 |
+
"epoch": 0.5773420479302832,
|
3740 |
+
"grad_norm": 1520.0718994140625,
|
3741 |
+
"learning_rate": 7.736257793678714e-05,
|
3742 |
+
"loss": 13.4148,
|
3743 |
+
"step": 530
|
3744 |
+
},
|
3745 |
+
{
|
3746 |
+
"epoch": 0.5784313725490197,
|
3747 |
+
"grad_norm": 1333.5904541015625,
|
3748 |
+
"learning_rate": 7.702570549408428e-05,
|
3749 |
+
"loss": 13.4515,
|
3750 |
+
"step": 531
|
3751 |
+
},
|
3752 |
+
{
|
3753 |
+
"epoch": 0.579520697167756,
|
3754 |
+
"grad_norm": 1032.6591796875,
|
3755 |
+
"learning_rate": 7.668910807487783e-05,
|
3756 |
+
"loss": 13.9835,
|
3757 |
+
"step": 532
|
3758 |
+
},
|
3759 |
+
{
|
3760 |
+
"epoch": 0.5806100217864923,
|
3761 |
+
"grad_norm": 1167.9801025390625,
|
3762 |
+
"learning_rate": 7.635278970854943e-05,
|
3763 |
+
"loss": 13.1408,
|
3764 |
+
"step": 533
|
3765 |
+
},
|
3766 |
+
{
|
3767 |
+
"epoch": 0.5816993464052288,
|
3768 |
+
"grad_norm": 1407.249267578125,
|
3769 |
+
"learning_rate": 7.601675442114009e-05,
|
3770 |
+
"loss": 13.9407,
|
3771 |
+
"step": 534
|
3772 |
+
},
|
3773 |
+
{
|
3774 |
+
"epoch": 0.5827886710239651,
|
3775 |
+
"grad_norm": 2209.746826171875,
|
3776 |
+
"learning_rate": 7.568100623530217e-05,
|
3777 |
+
"loss": 12.4755,
|
3778 |
+
"step": 535
|
3779 |
+
},
|
3780 |
+
{
|
3781 |
+
"epoch": 0.5838779956427015,
|
3782 |
+
"grad_norm": 1797.746826171875,
|
3783 |
+
"learning_rate": 7.534554917025119e-05,
|
3784 |
+
"loss": 12.8268,
|
3785 |
+
"step": 536
|
3786 |
+
},
|
3787 |
+
{
|
3788 |
+
"epoch": 0.5849673202614379,
|
3789 |
+
"grad_norm": 1302.199462890625,
|
3790 |
+
"learning_rate": 7.501038724171756e-05,
|
3791 |
+
"loss": 13.197,
|
3792 |
+
"step": 537
|
3793 |
+
},
|
3794 |
+
{
|
3795 |
+
"epoch": 0.5860566448801743,
|
3796 |
+
"grad_norm": 2067.767822265625,
|
3797 |
+
"learning_rate": 7.46755244618988e-05,
|
3798 |
+
"loss": 12.8685,
|
3799 |
+
"step": 538
|
3800 |
+
},
|
3801 |
+
{
|
3802 |
+
"epoch": 0.5871459694989106,
|
3803 |
+
"grad_norm": 1580.0186767578125,
|
3804 |
+
"learning_rate": 7.434096483941115e-05,
|
3805 |
+
"loss": 13.4972,
|
3806 |
+
"step": 539
|
3807 |
+
},
|
3808 |
+
{
|
3809 |
+
"epoch": 0.5882352941176471,
|
3810 |
+
"grad_norm": 894.298828125,
|
3811 |
+
"learning_rate": 7.400671237924202e-05,
|
3812 |
+
"loss": 13.1393,
|
3813 |
+
"step": 540
|
3814 |
+
},
|
3815 |
+
{
|
3816 |
+
"epoch": 0.5893246187363834,
|
3817 |
+
"grad_norm": 1165.6744384765625,
|
3818 |
+
"learning_rate": 7.367277108270156e-05,
|
3819 |
+
"loss": 13.9111,
|
3820 |
+
"step": 541
|
3821 |
+
},
|
3822 |
+
{
|
3823 |
+
"epoch": 0.5904139433551199,
|
3824 |
+
"grad_norm": 1790.1239013671875,
|
3825 |
+
"learning_rate": 7.333914494737514e-05,
|
3826 |
+
"loss": 13.351,
|
3827 |
+
"step": 542
|
3828 |
+
},
|
3829 |
+
{
|
3830 |
+
"epoch": 0.5915032679738562,
|
3831 |
+
"grad_norm": 984.1708374023438,
|
3832 |
+
"learning_rate": 7.300583796707539e-05,
|
3833 |
+
"loss": 14.5447,
|
3834 |
+
"step": 543
|
3835 |
+
},
|
3836 |
+
{
|
3837 |
+
"epoch": 0.5925925925925926,
|
3838 |
+
"grad_norm": 1341.8519287109375,
|
3839 |
+
"learning_rate": 7.267285413179421e-05,
|
3840 |
+
"loss": 14.1327,
|
3841 |
+
"step": 544
|
3842 |
+
},
|
3843 |
+
{
|
3844 |
+
"epoch": 0.593681917211329,
|
3845 |
+
"grad_norm": 1000.2693481445312,
|
3846 |
+
"learning_rate": 7.234019742765532e-05,
|
3847 |
+
"loss": 13.3989,
|
3848 |
+
"step": 545
|
3849 |
+
},
|
3850 |
+
{
|
3851 |
+
"epoch": 0.5947712418300654,
|
3852 |
+
"grad_norm": 1023.481689453125,
|
3853 |
+
"learning_rate": 7.200787183686625e-05,
|
3854 |
+
"loss": 13.069,
|
3855 |
+
"step": 546
|
3856 |
+
},
|
3857 |
+
{
|
3858 |
+
"epoch": 0.5958605664488017,
|
3859 |
+
"grad_norm": 1108.57470703125,
|
3860 |
+
"learning_rate": 7.167588133767091e-05,
|
3861 |
+
"loss": 12.2698,
|
3862 |
+
"step": 547
|
3863 |
+
},
|
3864 |
+
{
|
3865 |
+
"epoch": 0.5969498910675382,
|
3866 |
+
"grad_norm": 964.7139282226562,
|
3867 |
+
"learning_rate": 7.134422990430176e-05,
|
3868 |
+
"loss": 13.6585,
|
3869 |
+
"step": 548
|
3870 |
+
},
|
3871 |
+
{
|
3872 |
+
"epoch": 0.5980392156862745,
|
3873 |
+
"grad_norm": 1239.9150390625,
|
3874 |
+
"learning_rate": 7.101292150693241e-05,
|
3875 |
+
"loss": 13.0273,
|
3876 |
+
"step": 549
|
3877 |
+
},
|
3878 |
+
{
|
3879 |
+
"epoch": 0.599128540305011,
|
3880 |
+
"grad_norm": 846.2174072265625,
|
3881 |
+
"learning_rate": 7.068196011162994e-05,
|
3882 |
+
"loss": 13.0756,
|
3883 |
+
"step": 550
|
3884 |
+
},
|
3885 |
+
{
|
3886 |
+
"epoch": 0.6002178649237473,
|
3887 |
+
"grad_norm": 987.310546875,
|
3888 |
+
"learning_rate": 7.03513496803075e-05,
|
3889 |
+
"loss": 14.1704,
|
3890 |
+
"step": 551
|
3891 |
+
},
|
3892 |
+
{
|
3893 |
+
"epoch": 0.6013071895424836,
|
3894 |
+
"grad_norm": 5245.39697265625,
|
3895 |
+
"learning_rate": 7.002109417067697e-05,
|
3896 |
+
"loss": 14.4467,
|
3897 |
+
"step": 552
|
3898 |
+
},
|
3899 |
+
{
|
3900 |
+
"epoch": 0.6023965141612201,
|
3901 |
+
"grad_norm": 1111.834228515625,
|
3902 |
+
"learning_rate": 6.969119753620135e-05,
|
3903 |
+
"loss": 13.517,
|
3904 |
+
"step": 553
|
3905 |
+
},
|
3906 |
+
{
|
3907 |
+
"epoch": 0.6034858387799564,
|
3908 |
+
"grad_norm": 996.4967041015625,
|
3909 |
+
"learning_rate": 6.936166372604773e-05,
|
3910 |
+
"loss": 13.5025,
|
3911 |
+
"step": 554
|
3912 |
+
},
|
3913 |
+
{
|
3914 |
+
"epoch": 0.6045751633986928,
|
3915 |
+
"grad_norm": 867.002685546875,
|
3916 |
+
"learning_rate": 6.903249668503972e-05,
|
3917 |
+
"loss": 12.8567,
|
3918 |
+
"step": 555
|
3919 |
+
},
|
3920 |
+
{
|
3921 |
+
"epoch": 0.6056644880174292,
|
3922 |
+
"grad_norm": 1155.432373046875,
|
3923 |
+
"learning_rate": 6.87037003536104e-05,
|
3924 |
+
"loss": 13.0582,
|
3925 |
+
"step": 556
|
3926 |
+
},
|
3927 |
+
{
|
3928 |
+
"epoch": 0.6067538126361656,
|
3929 |
+
"grad_norm": 1074.625732421875,
|
3930 |
+
"learning_rate": 6.837527866775522e-05,
|
3931 |
+
"loss": 13.5709,
|
3932 |
+
"step": 557
|
3933 |
+
},
|
3934 |
+
{
|
3935 |
+
"epoch": 0.6078431372549019,
|
3936 |
+
"grad_norm": 1409.87841796875,
|
3937 |
+
"learning_rate": 6.804723555898458e-05,
|
3938 |
+
"loss": 13.7728,
|
3939 |
+
"step": 558
|
3940 |
+
},
|
3941 |
+
{
|
3942 |
+
"epoch": 0.6089324618736384,
|
3943 |
+
"grad_norm": 987.9152221679688,
|
3944 |
+
"learning_rate": 6.771957495427716e-05,
|
3945 |
+
"loss": 13.0499,
|
3946 |
+
"step": 559
|
3947 |
+
},
|
3948 |
+
{
|
3949 |
+
"epoch": 0.6100217864923747,
|
3950 |
+
"grad_norm": 1056.71240234375,
|
3951 |
+
"learning_rate": 6.739230077603259e-05,
|
3952 |
+
"loss": 13.4185,
|
3953 |
+
"step": 560
|
3954 |
+
},
|
3955 |
+
{
|
3956 |
+
"epoch": 0.6111111111111112,
|
3957 |
+
"grad_norm": 965.7537841796875,
|
3958 |
+
"learning_rate": 6.706541694202471e-05,
|
3959 |
+
"loss": 13.3033,
|
3960 |
+
"step": 561
|
3961 |
+
},
|
3962 |
+
{
|
3963 |
+
"epoch": 0.6122004357298475,
|
3964 |
+
"grad_norm": 950.843505859375,
|
3965 |
+
"learning_rate": 6.673892736535448e-05,
|
3966 |
+
"loss": 13.2638,
|
3967 |
+
"step": 562
|
3968 |
+
},
|
3969 |
+
{
|
3970 |
+
"epoch": 0.6132897603485838,
|
3971 |
+
"grad_norm": 888.200439453125,
|
3972 |
+
"learning_rate": 6.641283595440323e-05,
|
3973 |
+
"loss": 14.2555,
|
3974 |
+
"step": 563
|
3975 |
+
},
|
3976 |
+
{
|
3977 |
+
"epoch": 0.6143790849673203,
|
3978 |
+
"grad_norm": 1082.6995849609375,
|
3979 |
+
"learning_rate": 6.608714661278606e-05,
|
3980 |
+
"loss": 13.0727,
|
3981 |
+
"step": 564
|
3982 |
+
},
|
3983 |
+
{
|
3984 |
+
"epoch": 0.6154684095860566,
|
3985 |
+
"grad_norm": 1750.720947265625,
|
3986 |
+
"learning_rate": 6.576186323930466e-05,
|
3987 |
+
"loss": 12.9111,
|
3988 |
+
"step": 565
|
3989 |
+
},
|
3990 |
+
{
|
3991 |
+
"epoch": 0.616557734204793,
|
3992 |
+
"grad_norm": 1437.4244384765625,
|
3993 |
+
"learning_rate": 6.543698972790117e-05,
|
3994 |
+
"loss": 14.7961,
|
3995 |
+
"step": 566
|
3996 |
+
},
|
3997 |
+
{
|
3998 |
+
"epoch": 0.6176470588235294,
|
3999 |
+
"grad_norm": 1005.5140380859375,
|
4000 |
+
"learning_rate": 6.51125299676111e-05,
|
4001 |
+
"loss": 14.3845,
|
4002 |
+
"step": 567
|
4003 |
+
},
|
4004 |
+
{
|
4005 |
+
"epoch": 0.6187363834422658,
|
4006 |
+
"grad_norm": 802.7247924804688,
|
4007 |
+
"learning_rate": 6.478848784251713e-05,
|
4008 |
+
"loss": 14.1054,
|
4009 |
+
"step": 568
|
4010 |
+
},
|
4011 |
+
{
|
4012 |
+
"epoch": 0.6198257080610022,
|
4013 |
+
"grad_norm": 780.930908203125,
|
4014 |
+
"learning_rate": 6.446486723170236e-05,
|
4015 |
+
"loss": 13.4999,
|
4016 |
+
"step": 569
|
4017 |
+
},
|
4018 |
+
{
|
4019 |
+
"epoch": 0.6209150326797386,
|
4020 |
+
"grad_norm": 1568.1295166015625,
|
4021 |
+
"learning_rate": 6.414167200920391e-05,
|
4022 |
+
"loss": 14.3808,
|
4023 |
+
"step": 570
|
4024 |
+
},
|
4025 |
+
{
|
4026 |
+
"epoch": 0.6220043572984749,
|
4027 |
+
"grad_norm": 649.1087036132812,
|
4028 |
+
"learning_rate": 6.381890604396687e-05,
|
4029 |
+
"loss": 14.4266,
|
4030 |
+
"step": 571
|
4031 |
+
},
|
4032 |
+
{
|
4033 |
+
"epoch": 0.6230936819172114,
|
4034 |
+
"grad_norm": 1113.2542724609375,
|
4035 |
+
"learning_rate": 6.349657319979742e-05,
|
4036 |
+
"loss": 14.2873,
|
4037 |
+
"step": 572
|
4038 |
+
},
|
4039 |
+
{
|
4040 |
+
"epoch": 0.6241830065359477,
|
4041 |
+
"grad_norm": 735.5198974609375,
|
4042 |
+
"learning_rate": 6.317467733531712e-05,
|
4043 |
+
"loss": 14.1431,
|
4044 |
+
"step": 573
|
4045 |
+
},
|
4046 |
+
{
|
4047 |
+
"epoch": 0.6252723311546841,
|
4048 |
+
"grad_norm": 889.42822265625,
|
4049 |
+
"learning_rate": 6.28532223039163e-05,
|
4050 |
+
"loss": 14.5709,
|
4051 |
+
"step": 574
|
4052 |
+
},
|
4053 |
+
{
|
4054 |
+
"epoch": 0.6263616557734205,
|
4055 |
+
"grad_norm": 1000.8912963867188,
|
4056 |
+
"learning_rate": 6.253221195370826e-05,
|
4057 |
+
"loss": 14.9789,
|
4058 |
+
"step": 575
|
4059 |
+
},
|
4060 |
+
{
|
4061 |
+
"epoch": 0.6274509803921569,
|
4062 |
+
"grad_norm": 1058.2774658203125,
|
4063 |
+
"learning_rate": 6.221165012748297e-05,
|
4064 |
+
"loss": 13.2536,
|
4065 |
+
"step": 576
|
4066 |
+
},
|
4067 |
+
{
|
4068 |
+
"epoch": 0.6285403050108932,
|
4069 |
+
"grad_norm": 752.534912109375,
|
4070 |
+
"learning_rate": 6.189154066266112e-05,
|
4071 |
+
"loss": 14.1625,
|
4072 |
+
"step": 577
|
4073 |
+
},
|
4074 |
+
{
|
4075 |
+
"epoch": 0.6296296296296297,
|
4076 |
+
"grad_norm": 756.6798095703125,
|
4077 |
+
"learning_rate": 6.157188739124834e-05,
|
4078 |
+
"loss": 13.6298,
|
4079 |
+
"step": 578
|
4080 |
+
},
|
4081 |
+
{
|
4082 |
+
"epoch": 0.630718954248366,
|
4083 |
+
"grad_norm": 1065.2354736328125,
|
4084 |
+
"learning_rate": 6.125269413978907e-05,
|
4085 |
+
"loss": 13.4351,
|
4086 |
+
"step": 579
|
4087 |
+
},
|
4088 |
+
{
|
4089 |
+
"epoch": 0.6318082788671024,
|
4090 |
+
"grad_norm": 1031.8475341796875,
|
4091 |
+
"learning_rate": 6.093396472932103e-05,
|
4092 |
+
"loss": 12.9427,
|
4093 |
+
"step": 580
|
4094 |
+
},
|
4095 |
+
{
|
4096 |
+
"epoch": 0.6328976034858388,
|
4097 |
+
"grad_norm": 545.2572631835938,
|
4098 |
+
"learning_rate": 6.0615702975329194e-05,
|
4099 |
+
"loss": 13.4271,
|
4100 |
+
"step": 581
|
4101 |
+
},
|
4102 |
+
{
|
4103 |
+
"epoch": 0.6339869281045751,
|
4104 |
+
"grad_norm": 629.153076171875,
|
4105 |
+
"learning_rate": 6.029791268770029e-05,
|
4106 |
+
"loss": 13.7342,
|
4107 |
+
"step": 582
|
4108 |
+
},
|
4109 |
+
{
|
4110 |
+
"epoch": 0.6350762527233116,
|
4111 |
+
"grad_norm": 808.3855590820312,
|
4112 |
+
"learning_rate": 5.998059767067728e-05,
|
4113 |
+
"loss": 12.8523,
|
4114 |
+
"step": 583
|
4115 |
+
},
|
4116 |
+
{
|
4117 |
+
"epoch": 0.6361655773420479,
|
4118 |
+
"grad_norm": 966.9733276367188,
|
4119 |
+
"learning_rate": 5.9663761722813495e-05,
|
4120 |
+
"loss": 13.8446,
|
4121 |
+
"step": 584
|
4122 |
+
},
|
4123 |
+
{
|
4124 |
+
"epoch": 0.6372549019607843,
|
4125 |
+
"grad_norm": 832.753173828125,
|
4126 |
+
"learning_rate": 5.934740863692759e-05,
|
4127 |
+
"loss": 14.0291,
|
4128 |
+
"step": 585
|
4129 |
+
},
|
4130 |
+
{
|
4131 |
+
"epoch": 0.6383442265795207,
|
4132 |
+
"grad_norm": 870.2808837890625,
|
4133 |
+
"learning_rate": 5.903154220005771e-05,
|
4134 |
+
"loss": 14.7135,
|
4135 |
+
"step": 586
|
4136 |
+
},
|
4137 |
+
{
|
4138 |
+
"epoch": 0.6394335511982571,
|
4139 |
+
"grad_norm": 888.448974609375,
|
4140 |
+
"learning_rate": 5.871616619341653e-05,
|
4141 |
+
"loss": 13.6237,
|
4142 |
+
"step": 587
|
4143 |
+
},
|
4144 |
+
{
|
4145 |
+
"epoch": 0.6405228758169934,
|
4146 |
+
"grad_norm": 604.5468139648438,
|
4147 |
+
"learning_rate": 5.840128439234571e-05,
|
4148 |
+
"loss": 13.1112,
|
4149 |
+
"step": 588
|
4150 |
+
},
|
4151 |
+
{
|
4152 |
+
"epoch": 0.6416122004357299,
|
4153 |
+
"grad_norm": 740.3869018554688,
|
4154 |
+
"learning_rate": 5.80869005662708e-05,
|
4155 |
+
"loss": 13.2927,
|
4156 |
+
"step": 589
|
4157 |
+
},
|
4158 |
+
{
|
4159 |
+
"epoch": 0.6427015250544662,
|
4160 |
+
"grad_norm": 729.2908935546875,
|
4161 |
+
"learning_rate": 5.777301847865629e-05,
|
4162 |
+
"loss": 13.1883,
|
4163 |
+
"step": 590
|
4164 |
+
},
|
4165 |
+
{
|
4166 |
+
"epoch": 0.6437908496732027,
|
4167 |
+
"grad_norm": 679.76171875,
|
4168 |
+
"learning_rate": 5.7459641886960244e-05,
|
4169 |
+
"loss": 12.7278,
|
4170 |
+
"step": 591
|
4171 |
+
},
|
4172 |
+
{
|
4173 |
+
"epoch": 0.644880174291939,
|
4174 |
+
"grad_norm": 771.1044921875,
|
4175 |
+
"learning_rate": 5.714677454258947e-05,
|
4176 |
+
"loss": 14.2043,
|
4177 |
+
"step": 592
|
4178 |
+
},
|
4179 |
+
{
|
4180 |
+
"epoch": 0.6459694989106753,
|
4181 |
+
"grad_norm": 525.4684448242188,
|
4182 |
+
"learning_rate": 5.6834420190854745e-05,
|
4183 |
+
"loss": 13.7949,
|
4184 |
+
"step": 593
|
4185 |
+
},
|
4186 |
+
{
|
4187 |
+
"epoch": 0.6470588235294118,
|
4188 |
+
"grad_norm": 1367.8231201171875,
|
4189 |
+
"learning_rate": 5.652258257092569e-05,
|
4190 |
+
"loss": 13.5713,
|
4191 |
+
"step": 594
|
4192 |
+
},
|
4193 |
+
{
|
4194 |
+
"epoch": 0.6481481481481481,
|
4195 |
+
"grad_norm": 1199.23876953125,
|
4196 |
+
"learning_rate": 5.621126541578632e-05,
|
4197 |
+
"loss": 13.6268,
|
4198 |
+
"step": 595
|
4199 |
+
},
|
4200 |
+
{
|
4201 |
+
"epoch": 0.6492374727668845,
|
4202 |
+
"grad_norm": 995.0260009765625,
|
4203 |
+
"learning_rate": 5.590047245219009e-05,
|
4204 |
+
"loss": 14.3565,
|
4205 |
+
"step": 596
|
4206 |
+
},
|
4207 |
+
{
|
4208 |
+
"epoch": 0.6503267973856209,
|
4209 |
+
"grad_norm": 805.2388916015625,
|
4210 |
+
"learning_rate": 5.559020740061549e-05,
|
4211 |
+
"loss": 14.2207,
|
4212 |
+
"step": 597
|
4213 |
+
},
|
4214 |
+
{
|
4215 |
+
"epoch": 0.6514161220043573,
|
4216 |
+
"grad_norm": 779.6170654296875,
|
4217 |
+
"learning_rate": 5.528047397522133e-05,
|
4218 |
+
"loss": 13.4574,
|
4219 |
+
"step": 598
|
4220 |
+
},
|
4221 |
+
{
|
4222 |
+
"epoch": 0.6525054466230937,
|
4223 |
+
"grad_norm": 974.6607055664062,
|
4224 |
+
"learning_rate": 5.497127588380244e-05,
|
4225 |
+
"loss": 13.5537,
|
4226 |
+
"step": 599
|
4227 |
+
},
|
4228 |
+
{
|
4229 |
+
"epoch": 0.6535947712418301,
|
4230 |
+
"grad_norm": 957.2470092773438,
|
4231 |
+
"learning_rate": 5.4662616827745185e-05,
|
4232 |
+
"loss": 13.6918,
|
4233 |
+
"step": 600
|
4234 |
+
},
|
4235 |
+
{
|
4236 |
+
"epoch": 0.6546840958605664,
|
4237 |
+
"grad_norm": 564.2131958007812,
|
4238 |
+
"learning_rate": 5.4354500501983074e-05,
|
4239 |
+
"loss": 14.5732,
|
4240 |
+
"step": 601
|
4241 |
+
},
|
4242 |
+
{
|
4243 |
+
"epoch": 0.6557734204793029,
|
4244 |
+
"grad_norm": 970.4561157226562,
|
4245 |
+
"learning_rate": 5.404693059495285e-05,
|
4246 |
+
"loss": 14.259,
|
4247 |
+
"step": 602
|
4248 |
+
},
|
4249 |
+
{
|
4250 |
+
"epoch": 0.6568627450980392,
|
4251 |
+
"grad_norm": 1475.91943359375,
|
4252 |
+
"learning_rate": 5.373991078854992e-05,
|
4253 |
+
"loss": 13.7511,
|
4254 |
+
"step": 603
|
4255 |
+
},
|
4256 |
+
{
|
4257 |
+
"epoch": 0.6579520697167756,
|
4258 |
+
"grad_norm": 1206.491943359375,
|
4259 |
+
"learning_rate": 5.3433444758084604e-05,
|
4260 |
+
"loss": 13.9658,
|
4261 |
+
"step": 604
|
4262 |
+
},
|
4263 |
+
{
|
4264 |
+
"epoch": 0.659041394335512,
|
4265 |
+
"grad_norm": 729.8560791015625,
|
4266 |
+
"learning_rate": 5.312753617223794e-05,
|
4267 |
+
"loss": 12.3194,
|
4268 |
+
"step": 605
|
4269 |
+
},
|
4270 |
+
{
|
4271 |
+
"epoch": 0.6601307189542484,
|
4272 |
+
"grad_norm": 833.2518310546875,
|
4273 |
+
"learning_rate": 5.282218869301788e-05,
|
4274 |
+
"loss": 13.2262,
|
4275 |
+
"step": 606
|
4276 |
+
},
|
4277 |
+
{
|
4278 |
+
"epoch": 0.6612200435729847,
|
4279 |
+
"grad_norm": 1121.98388671875,
|
4280 |
+
"learning_rate": 5.251740597571542e-05,
|
4281 |
+
"loss": 13.1887,
|
4282 |
+
"step": 607
|
4283 |
+
},
|
4284 |
+
{
|
4285 |
+
"epoch": 0.6623093681917211,
|
4286 |
+
"grad_norm": 732.55810546875,
|
4287 |
+
"learning_rate": 5.221319166886073e-05,
|
4288 |
+
"loss": 11.9579,
|
4289 |
+
"step": 608
|
4290 |
+
},
|
4291 |
+
{
|
4292 |
+
"epoch": 0.6633986928104575,
|
4293 |
+
"grad_norm": 795.4283447265625,
|
4294 |
+
"learning_rate": 5.190954941417977e-05,
|
4295 |
+
"loss": 13.6673,
|
4296 |
+
"step": 609
|
4297 |
+
},
|
4298 |
+
{
|
4299 |
+
"epoch": 0.664488017429194,
|
4300 |
+
"grad_norm": 742.3599243164062,
|
4301 |
+
"learning_rate": 5.160648284655032e-05,
|
4302 |
+
"loss": 13.4396,
|
4303 |
+
"step": 610
|
4304 |
+
},
|
4305 |
+
{
|
4306 |
+
"epoch": 0.6655773420479303,
|
4307 |
+
"grad_norm": 872.8551635742188,
|
4308 |
+
"learning_rate": 5.1303995593958824e-05,
|
4309 |
+
"loss": 14.0764,
|
4310 |
+
"step": 611
|
4311 |
+
},
|
4312 |
+
{
|
4313 |
+
"epoch": 0.6666666666666666,
|
4314 |
+
"grad_norm": 865.8281860351562,
|
4315 |
+
"learning_rate": 5.100209127745661e-05,
|
4316 |
+
"loss": 13.2594,
|
4317 |
+
"step": 612
|
4318 |
+
},
|
4319 |
+
{
|
4320 |
+
"epoch": 0.6677559912854031,
|
4321 |
+
"grad_norm": 553.1524658203125,
|
4322 |
+
"learning_rate": 5.0700773511116906e-05,
|
4323 |
+
"loss": 13.4783,
|
4324 |
+
"step": 613
|
4325 |
+
},
|
4326 |
+
{
|
4327 |
+
"epoch": 0.6688453159041394,
|
4328 |
+
"grad_norm": 981.0257568359375,
|
4329 |
+
"learning_rate": 5.040004590199128e-05,
|
4330 |
+
"loss": 14.2846,
|
4331 |
+
"step": 614
|
4332 |
+
},
|
4333 |
+
{
|
4334 |
+
"epoch": 0.6699346405228758,
|
4335 |
+
"grad_norm": 817.2470703125,
|
4336 |
+
"learning_rate": 5.0099912050066556e-05,
|
4337 |
+
"loss": 12.6324,
|
4338 |
+
"step": 615
|
4339 |
+
},
|
4340 |
+
{
|
4341 |
+
"epoch": 0.6710239651416122,
|
4342 |
+
"grad_norm": 593.127197265625,
|
4343 |
+
"learning_rate": 4.9800375548221845e-05,
|
4344 |
+
"loss": 13.0678,
|
4345 |
+
"step": 616
|
4346 |
+
},
|
4347 |
+
{
|
4348 |
+
"epoch": 0.6721132897603486,
|
4349 |
+
"grad_norm": 940.2001953125,
|
4350 |
+
"learning_rate": 4.950143998218531e-05,
|
4351 |
+
"loss": 13.5529,
|
4352 |
+
"step": 617
|
4353 |
+
},
|
4354 |
+
{
|
4355 |
+
"epoch": 0.673202614379085,
|
4356 |
+
"grad_norm": 717.47021484375,
|
4357 |
+
"learning_rate": 4.920310893049146e-05,
|
4358 |
+
"loss": 12.7114,
|
4359 |
+
"step": 618
|
4360 |
+
},
|
4361 |
+
{
|
4362 |
+
"epoch": 0.6742919389978214,
|
4363 |
+
"grad_norm": 874.2285766601562,
|
4364 |
+
"learning_rate": 4.89053859644381e-05,
|
4365 |
+
"loss": 14.4422,
|
4366 |
+
"step": 619
|
4367 |
+
},
|
4368 |
+
{
|
4369 |
+
"epoch": 0.6753812636165577,
|
4370 |
+
"grad_norm": 847.6348876953125,
|
4371 |
+
"learning_rate": 4.860827464804383e-05,
|
4372 |
+
"loss": 12.8582,
|
4373 |
+
"step": 620
|
4374 |
+
},
|
4375 |
+
{
|
4376 |
+
"epoch": 0.6764705882352942,
|
4377 |
+
"grad_norm": 552.3037719726562,
|
4378 |
+
"learning_rate": 4.831177853800511e-05,
|
4379 |
+
"loss": 13.5564,
|
4380 |
+
"step": 621
|
4381 |
+
},
|
4382 |
+
{
|
4383 |
+
"epoch": 0.6775599128540305,
|
4384 |
+
"grad_norm": 704.1598510742188,
|
4385 |
+
"learning_rate": 4.801590118365383e-05,
|
4386 |
+
"loss": 14.2366,
|
4387 |
+
"step": 622
|
4388 |
+
},
|
4389 |
+
{
|
4390 |
+
"epoch": 0.6786492374727668,
|
4391 |
+
"grad_norm": 535.1793823242188,
|
4392 |
+
"learning_rate": 4.77206461269149e-05,
|
4393 |
+
"loss": 13.5502,
|
4394 |
+
"step": 623
|
4395 |
+
},
|
4396 |
+
{
|
4397 |
+
"epoch": 0.6797385620915033,
|
4398 |
+
"grad_norm": 1089.44775390625,
|
4399 |
+
"learning_rate": 4.7426016902263636e-05,
|
4400 |
+
"loss": 13.391,
|
4401 |
+
"step": 624
|
4402 |
+
},
|
4403 |
+
{
|
4404 |
+
"epoch": 0.6808278867102396,
|
4405 |
+
"grad_norm": 793.9139404296875,
|
4406 |
+
"learning_rate": 4.713201703668367e-05,
|
4407 |
+
"loss": 13.9273,
|
4408 |
+
"step": 625
|
4409 |
+
},
|
4410 |
+
{
|
4411 |
+
"epoch": 0.681917211328976,
|
4412 |
+
"grad_norm": 1263.728759765625,
|
4413 |
+
"learning_rate": 4.683865004962452e-05,
|
4414 |
+
"loss": 13.2224,
|
4415 |
+
"step": 626
|
4416 |
+
},
|
4417 |
+
{
|
4418 |
+
"epoch": 0.6830065359477124,
|
4419 |
+
"grad_norm": 583.6502685546875,
|
4420 |
+
"learning_rate": 4.654591945295969e-05,
|
4421 |
+
"loss": 13.1198,
|
4422 |
+
"step": 627
|
4423 |
+
},
|
4424 |
+
{
|
4425 |
+
"epoch": 0.6840958605664488,
|
4426 |
+
"grad_norm": 546.5349731445312,
|
4427 |
+
"learning_rate": 4.6253828750944375e-05,
|
4428 |
+
"loss": 14.2708,
|
4429 |
+
"step": 628
|
4430 |
+
},
|
4431 |
+
{
|
4432 |
+
"epoch": 0.6851851851851852,
|
4433 |
+
"grad_norm": 573.7505493164062,
|
4434 |
+
"learning_rate": 4.596238144017369e-05,
|
4435 |
+
"loss": 14.4553,
|
4436 |
+
"step": 629
|
4437 |
+
},
|
4438 |
+
{
|
4439 |
+
"epoch": 0.6862745098039216,
|
4440 |
+
"grad_norm": 485.56976318359375,
|
4441 |
+
"learning_rate": 4.567158100954083e-05,
|
4442 |
+
"loss": 13.9009,
|
4443 |
+
"step": 630
|
4444 |
+
},
|
4445 |
+
{
|
4446 |
+
"epoch": 0.6873638344226579,
|
4447 |
+
"grad_norm": 859.7692260742188,
|
4448 |
+
"learning_rate": 4.53814309401951e-05,
|
4449 |
+
"loss": 13.4996,
|
4450 |
+
"step": 631
|
4451 |
+
},
|
4452 |
+
{
|
4453 |
+
"epoch": 0.6884531590413944,
|
4454 |
+
"grad_norm": 806.7265014648438,
|
4455 |
+
"learning_rate": 4.509193470550056e-05,
|
4456 |
+
"loss": 13.8831,
|
4457 |
+
"step": 632
|
4458 |
+
},
|
4459 |
+
{
|
4460 |
+
"epoch": 0.6895424836601307,
|
4461 |
+
"grad_norm": 678.3446044921875,
|
4462 |
+
"learning_rate": 4.4803095770994106e-05,
|
4463 |
+
"loss": 12.5697,
|
4464 |
+
"step": 633
|
4465 |
+
},
|
4466 |
+
{
|
4467 |
+
"epoch": 0.690631808278867,
|
4468 |
+
"grad_norm": 760.7838745117188,
|
4469 |
+
"learning_rate": 4.4514917594344184e-05,
|
4470 |
+
"loss": 13.5064,
|
4471 |
+
"step": 634
|
4472 |
+
},
|
4473 |
+
{
|
4474 |
+
"epoch": 0.6917211328976035,
|
4475 |
+
"grad_norm": 1103.8131103515625,
|
4476 |
+
"learning_rate": 4.422740362530945e-05,
|
4477 |
+
"loss": 14.8133,
|
4478 |
+
"step": 635
|
4479 |
+
},
|
4480 |
+
{
|
4481 |
+
"epoch": 0.6928104575163399,
|
4482 |
+
"grad_norm": 737.2034912109375,
|
4483 |
+
"learning_rate": 4.3940557305697226e-05,
|
4484 |
+
"loss": 14.7038,
|
4485 |
+
"step": 636
|
4486 |
+
},
|
4487 |
+
{
|
4488 |
+
"epoch": 0.6938997821350763,
|
4489 |
+
"grad_norm": 579.3258056640625,
|
4490 |
+
"learning_rate": 4.3654382069322644e-05,
|
4491 |
+
"loss": 14.0494,
|
4492 |
+
"step": 637
|
4493 |
+
},
|
4494 |
+
{
|
4495 |
+
"epoch": 0.6949891067538126,
|
4496 |
+
"grad_norm": 680.4292602539062,
|
4497 |
+
"learning_rate": 4.3368881341967135e-05,
|
4498 |
+
"loss": 14.3115,
|
4499 |
+
"step": 638
|
4500 |
+
},
|
4501 |
+
{
|
4502 |
+
"epoch": 0.696078431372549,
|
4503 |
+
"grad_norm": 617.1260986328125,
|
4504 |
+
"learning_rate": 4.308405854133786e-05,
|
4505 |
+
"loss": 13.788,
|
4506 |
+
"step": 639
|
4507 |
+
},
|
4508 |
+
{
|
4509 |
+
"epoch": 0.6971677559912854,
|
4510 |
+
"grad_norm": 951.0760498046875,
|
4511 |
+
"learning_rate": 4.2799917077026394e-05,
|
4512 |
+
"loss": 13.1248,
|
4513 |
+
"step": 640
|
4514 |
+
},
|
4515 |
+
{
|
4516 |
+
"epoch": 0.6982570806100218,
|
4517 |
+
"grad_norm": 606.4050903320312,
|
4518 |
+
"learning_rate": 4.251646035046814e-05,
|
4519 |
+
"loss": 14.0699,
|
4520 |
+
"step": 641
|
4521 |
+
},
|
4522 |
+
{
|
4523 |
+
"epoch": 0.6993464052287581,
|
4524 |
+
"grad_norm": 596.975830078125,
|
4525 |
+
"learning_rate": 4.223369175490162e-05,
|
4526 |
+
"loss": 11.8161,
|
4527 |
+
"step": 642
|
4528 |
+
},
|
4529 |
+
{
|
4530 |
+
"epoch": 0.7004357298474946,
|
4531 |
+
"grad_norm": 867.7359619140625,
|
4532 |
+
"learning_rate": 4.195161467532769e-05,
|
4533 |
+
"loss": 13.4987,
|
4534 |
+
"step": 643
|
4535 |
+
},
|
4536 |
+
{
|
4537 |
+
"epoch": 0.7015250544662309,
|
4538 |
+
"grad_norm": 802.6876831054688,
|
4539 |
+
"learning_rate": 4.167023248846925e-05,
|
4540 |
+
"loss": 12.8087,
|
4541 |
+
"step": 644
|
4542 |
+
},
|
4543 |
+
{
|
4544 |
+
"epoch": 0.7026143790849673,
|
4545 |
+
"grad_norm": 687.2286987304688,
|
4546 |
+
"learning_rate": 4.138954856273054e-05,
|
4547 |
+
"loss": 12.1043,
|
4548 |
+
"step": 645
|
4549 |
+
},
|
4550 |
+
{
|
4551 |
+
"epoch": 0.7037037037037037,
|
4552 |
+
"grad_norm": 669.9190673828125,
|
4553 |
+
"learning_rate": 4.110956625815713e-05,
|
4554 |
+
"loss": 12.1478,
|
4555 |
+
"step": 646
|
4556 |
+
},
|
4557 |
+
{
|
4558 |
+
"epoch": 0.7047930283224401,
|
4559 |
+
"grad_norm": 841.2921752929688,
|
4560 |
+
"learning_rate": 4.083028892639541e-05,
|
4561 |
+
"loss": 12.806,
|
4562 |
+
"step": 647
|
4563 |
+
},
|
4564 |
+
{
|
4565 |
+
"epoch": 0.7058823529411765,
|
4566 |
+
"grad_norm": 808.080078125,
|
4567 |
+
"learning_rate": 4.055171991065262e-05,
|
4568 |
+
"loss": 13.3545,
|
4569 |
+
"step": 648
|
4570 |
+
},
|
4571 |
+
{
|
4572 |
+
"epoch": 0.7069716775599129,
|
4573 |
+
"grad_norm": 919.4234619140625,
|
4574 |
+
"learning_rate": 4.027386254565688e-05,
|
4575 |
+
"loss": 13.9735,
|
4576 |
+
"step": 649
|
4577 |
+
},
|
4578 |
+
{
|
4579 |
+
"epoch": 0.7080610021786492,
|
4580 |
+
"grad_norm": 769.6393432617188,
|
4581 |
+
"learning_rate": 3.9996720157617094e-05,
|
4582 |
+
"loss": 14.6133,
|
4583 |
+
"step": 650
|
4584 |
+
},
|
4585 |
+
{
|
4586 |
+
"epoch": 0.7091503267973857,
|
4587 |
+
"grad_norm": 880.525146484375,
|
4588 |
+
"learning_rate": 3.972029606418335e-05,
|
4589 |
+
"loss": 13.0994,
|
4590 |
+
"step": 651
|
4591 |
+
},
|
4592 |
+
{
|
4593 |
+
"epoch": 0.710239651416122,
|
4594 |
+
"grad_norm": 945.1436767578125,
|
4595 |
+
"learning_rate": 3.9444593574406915e-05,
|
4596 |
+
"loss": 14.2055,
|
4597 |
+
"step": 652
|
4598 |
+
},
|
4599 |
+
{
|
4600 |
+
"epoch": 0.7113289760348583,
|
4601 |
+
"grad_norm": 690.2129516601562,
|
4602 |
+
"learning_rate": 3.9169615988701e-05,
|
4603 |
+
"loss": 13.2737,
|
4604 |
+
"step": 653
|
4605 |
+
},
|
4606 |
+
{
|
4607 |
+
"epoch": 0.7124183006535948,
|
4608 |
+
"grad_norm": 563.0156860351562,
|
4609 |
+
"learning_rate": 3.8895366598800896e-05,
|
4610 |
+
"loss": 12.8401,
|
4611 |
+
"step": 654
|
4612 |
+
},
|
4613 |
+
{
|
4614 |
+
"epoch": 0.7135076252723311,
|
4615 |
+
"grad_norm": 814.678466796875,
|
4616 |
+
"learning_rate": 3.862184868772473e-05,
|
4617 |
+
"loss": 12.2818,
|
4618 |
+
"step": 655
|
4619 |
+
},
|
4620 |
+
{
|
4621 |
+
"epoch": 0.7145969498910676,
|
4622 |
+
"grad_norm": 712.4736938476562,
|
4623 |
+
"learning_rate": 3.834906552973424e-05,
|
4624 |
+
"loss": 12.5831,
|
4625 |
+
"step": 656
|
4626 |
+
},
|
4627 |
+
{
|
4628 |
+
"epoch": 0.7156862745098039,
|
4629 |
+
"grad_norm": 1027.3265380859375,
|
4630 |
+
"learning_rate": 3.807702039029539e-05,
|
4631 |
+
"loss": 13.827,
|
4632 |
+
"step": 657
|
4633 |
+
},
|
4634 |
+
{
|
4635 |
+
"epoch": 0.7167755991285403,
|
4636 |
+
"grad_norm": 1087.4530029296875,
|
4637 |
+
"learning_rate": 3.780571652603949e-05,
|
4638 |
+
"loss": 13.3193,
|
4639 |
+
"step": 658
|
4640 |
+
},
|
4641 |
+
{
|
4642 |
+
"epoch": 0.7178649237472767,
|
4643 |
+
"grad_norm": 926.59326171875,
|
4644 |
+
"learning_rate": 3.753515718472402e-05,
|
4645 |
+
"loss": 13.3761,
|
4646 |
+
"step": 659
|
4647 |
+
},
|
4648 |
+
{
|
4649 |
+
"epoch": 0.7189542483660131,
|
4650 |
+
"grad_norm": 758.7875366210938,
|
4651 |
+
"learning_rate": 3.726534560519381e-05,
|
4652 |
+
"loss": 12.2786,
|
4653 |
+
"step": 660
|
4654 |
+
},
|
4655 |
+
{
|
4656 |
+
"epoch": 0.7200435729847494,
|
4657 |
+
"grad_norm": 693.2098999023438,
|
4658 |
+
"learning_rate": 3.6996285017342406e-05,
|
4659 |
+
"loss": 12.9392,
|
4660 |
+
"step": 661
|
4661 |
+
},
|
4662 |
+
{
|
4663 |
+
"epoch": 0.7211328976034859,
|
4664 |
+
"grad_norm": 1132.9970703125,
|
4665 |
+
"learning_rate": 3.672797864207316e-05,
|
4666 |
+
"loss": 12.6866,
|
4667 |
+
"step": 662
|
4668 |
+
},
|
4669 |
+
{
|
4670 |
+
"epoch": 0.7222222222222222,
|
4671 |
+
"grad_norm": 868.7568969726562,
|
4672 |
+
"learning_rate": 3.646042969126093e-05,
|
4673 |
+
"loss": 12.7426,
|
4674 |
+
"step": 663
|
4675 |
+
},
|
4676 |
+
{
|
4677 |
+
"epoch": 0.7233115468409586,
|
4678 |
+
"grad_norm": 800.123291015625,
|
4679 |
+
"learning_rate": 3.619364136771337e-05,
|
4680 |
+
"loss": 12.4544,
|
4681 |
+
"step": 664
|
4682 |
+
},
|
4683 |
+
{
|
4684 |
+
"epoch": 0.724400871459695,
|
4685 |
+
"grad_norm": 509.1244201660156,
|
4686 |
+
"learning_rate": 3.5927616865132884e-05,
|
4687 |
+
"loss": 13.2459,
|
4688 |
+
"step": 665
|
4689 |
+
},
|
4690 |
+
{
|
4691 |
+
"epoch": 0.7254901960784313,
|
4692 |
+
"grad_norm": 568.7617797851562,
|
4693 |
+
"learning_rate": 3.566235936807808e-05,
|
4694 |
+
"loss": 13.3732,
|
4695 |
+
"step": 666
|
4696 |
+
},
|
4697 |
+
{
|
4698 |
+
"epoch": 0.7265795206971678,
|
4699 |
+
"grad_norm": 839.1197509765625,
|
4700 |
+
"learning_rate": 3.539787205192586e-05,
|
4701 |
+
"loss": 12.6018,
|
4702 |
+
"step": 667
|
4703 |
+
},
|
4704 |
+
{
|
4705 |
+
"epoch": 0.7276688453159041,
|
4706 |
+
"grad_norm": 750.1957397460938,
|
4707 |
+
"learning_rate": 3.513415808283341e-05,
|
4708 |
+
"loss": 13.1899,
|
4709 |
+
"step": 668
|
4710 |
+
},
|
4711 |
+
{
|
4712 |
+
"epoch": 0.7287581699346405,
|
4713 |
+
"grad_norm": 604.1763305664062,
|
4714 |
+
"learning_rate": 3.4871220617700126e-05,
|
4715 |
+
"loss": 13.0681,
|
4716 |
+
"step": 669
|
4717 |
+
},
|
4718 |
+
{
|
4719 |
+
"epoch": 0.7298474945533769,
|
4720 |
+
"grad_norm": 637.3112182617188,
|
4721 |
+
"learning_rate": 3.460906280413007e-05,
|
4722 |
+
"loss": 13.0931,
|
4723 |
+
"step": 670
|
4724 |
+
},
|
4725 |
+
{
|
4726 |
+
"epoch": 0.7309368191721133,
|
4727 |
+
"grad_norm": 778.7605590820312,
|
4728 |
+
"learning_rate": 3.4347687780394e-05,
|
4729 |
+
"loss": 13.7031,
|
4730 |
+
"step": 671
|
4731 |
+
},
|
4732 |
+
{
|
4733 |
+
"epoch": 0.7320261437908496,
|
4734 |
+
"grad_norm": 1044.1873779296875,
|
4735 |
+
"learning_rate": 3.4087098675392104e-05,
|
4736 |
+
"loss": 12.2163,
|
4737 |
+
"step": 672
|
4738 |
+
},
|
4739 |
+
{
|
4740 |
+
"epoch": 0.7331154684095861,
|
4741 |
+
"grad_norm": 833.3889770507812,
|
4742 |
+
"learning_rate": 3.382729860861632e-05,
|
4743 |
+
"loss": 13.2927,
|
4744 |
+
"step": 673
|
4745 |
+
},
|
4746 |
+
{
|
4747 |
+
"epoch": 0.7342047930283224,
|
4748 |
+
"grad_norm": 655.1200561523438,
|
4749 |
+
"learning_rate": 3.3568290690113034e-05,
|
4750 |
+
"loss": 12.0668,
|
4751 |
+
"step": 674
|
4752 |
+
},
|
4753 |
+
{
|
4754 |
+
"epoch": 0.7352941176470589,
|
4755 |
+
"grad_norm": 702.258056640625,
|
4756 |
+
"learning_rate": 3.331007802044601e-05,
|
4757 |
+
"loss": 11.1181,
|
4758 |
+
"step": 675
|
4759 |
+
},
|
4760 |
+
{
|
4761 |
+
"epoch": 0.7363834422657952,
|
4762 |
+
"grad_norm": 791.837890625,
|
4763 |
+
"learning_rate": 3.305266369065901e-05,
|
4764 |
+
"loss": 12.7031,
|
4765 |
+
"step": 676
|
4766 |
+
},
|
4767 |
+
{
|
4768 |
+
"epoch": 0.7374727668845316,
|
4769 |
+
"grad_norm": 636.5592041015625,
|
4770 |
+
"learning_rate": 3.279605078223906e-05,
|
4771 |
+
"loss": 14.1468,
|
4772 |
+
"step": 677
|
4773 |
+
},
|
4774 |
+
{
|
4775 |
+
"epoch": 0.738562091503268,
|
4776 |
+
"grad_norm": 1960.34228515625,
|
4777 |
+
"learning_rate": 3.25402423670793e-05,
|
4778 |
+
"loss": 12.7793,
|
4779 |
+
"step": 678
|
4780 |
+
},
|
4781 |
+
{
|
4782 |
+
"epoch": 0.7396514161220044,
|
4783 |
+
"grad_norm": 987.132080078125,
|
4784 |
+
"learning_rate": 3.228524150744249e-05,
|
4785 |
+
"loss": 13.8105,
|
4786 |
+
"step": 679
|
4787 |
+
},
|
4788 |
+
{
|
4789 |
+
"epoch": 0.7407407407407407,
|
4790 |
+
"grad_norm": 536.3349609375,
|
4791 |
+
"learning_rate": 3.2031051255924085e-05,
|
4792 |
+
"loss": 13.0451,
|
4793 |
+
"step": 680
|
4794 |
+
},
|
4795 |
+
{
|
4796 |
+
"epoch": 0.7418300653594772,
|
4797 |
+
"grad_norm": 829.6603393554688,
|
4798 |
+
"learning_rate": 3.1777674655415834e-05,
|
4799 |
+
"loss": 13.352,
|
4800 |
+
"step": 681
|
4801 |
+
},
|
4802 |
+
{
|
4803 |
+
"epoch": 0.7429193899782135,
|
4804 |
+
"grad_norm": 825.07275390625,
|
4805 |
+
"learning_rate": 3.1525114739069415e-05,
|
4806 |
+
"loss": 13.4622,
|
4807 |
+
"step": 682
|
4808 |
+
},
|
4809 |
+
{
|
4810 |
+
"epoch": 0.7440087145969498,
|
4811 |
+
"grad_norm": 1563.3502197265625,
|
4812 |
+
"learning_rate": 3.127337453025994e-05,
|
4813 |
+
"loss": 13.9679,
|
4814 |
+
"step": 683
|
4815 |
+
},
|
4816 |
+
{
|
4817 |
+
"epoch": 0.7450980392156863,
|
4818 |
+
"grad_norm": 955.0587158203125,
|
4819 |
+
"learning_rate": 3.102245704254995e-05,
|
4820 |
+
"loss": 12.3828,
|
4821 |
+
"step": 684
|
4822 |
+
},
|
4823 |
+
{
|
4824 |
+
"epoch": 0.7461873638344226,
|
4825 |
+
"grad_norm": 784.834716796875,
|
4826 |
+
"learning_rate": 3.077236527965318e-05,
|
4827 |
+
"loss": 13.1804,
|
4828 |
+
"step": 685
|
4829 |
+
},
|
4830 |
+
{
|
4831 |
+
"epoch": 0.7472766884531591,
|
4832 |
+
"grad_norm": 735.6173706054688,
|
4833 |
+
"learning_rate": 3.0523102235398714e-05,
|
4834 |
+
"loss": 13.0926,
|
4835 |
+
"step": 686
|
4836 |
+
},
|
4837 |
+
{
|
4838 |
+
"epoch": 0.7483660130718954,
|
4839 |
+
"grad_norm": 1419.035888671875,
|
4840 |
+
"learning_rate": 3.0274670893695147e-05,
|
4841 |
+
"loss": 13.365,
|
4842 |
+
"step": 687
|
4843 |
+
},
|
4844 |
+
{
|
4845 |
+
"epoch": 0.7494553376906318,
|
4846 |
+
"grad_norm": 1074.7969970703125,
|
4847 |
+
"learning_rate": 3.002707422849472e-05,
|
4848 |
+
"loss": 13.357,
|
4849 |
+
"step": 688
|
4850 |
+
},
|
4851 |
+
{
|
4852 |
+
"epoch": 0.7505446623093682,
|
4853 |
+
"grad_norm": 1391.6951904296875,
|
4854 |
+
"learning_rate": 2.978031520375798e-05,
|
4855 |
+
"loss": 13.65,
|
4856 |
+
"step": 689
|
4857 |
+
},
|
4858 |
+
{
|
4859 |
+
"epoch": 0.7516339869281046,
|
4860 |
+
"grad_norm": 789.4784545898438,
|
4861 |
+
"learning_rate": 2.9534396773417994e-05,
|
4862 |
+
"loss": 12.0752,
|
4863 |
+
"step": 690
|
4864 |
+
},
|
4865 |
+
{
|
4866 |
+
"epoch": 0.7516339869281046,
|
4867 |
+
"eval_loss": 3.2088863849639893,
|
4868 |
+
"eval_runtime": 2.6648,
|
4869 |
+
"eval_samples_per_second": 145.229,
|
4870 |
+
"eval_steps_per_second": 72.802,
|
4871 |
+
"step": 690
|
4872 |
}
|
4873 |
],
|
4874 |
"logging_steps": 1,
|
|
|
4888 |
"attributes": {}
|
4889 |
}
|
4890 |
},
|
4891 |
+
"total_flos": 2341481805250560.0,
|
4892 |
"train_batch_size": 2,
|
4893 |
"trial_name": null,
|
4894 |
"trial_params": null
|