Training in progress, step 492, 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 83945296
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0bd01a843d72a649467ca0ef8aaef50f27c47ac052da47acba539e63005add66
|
3 |
size 83945296
|
last-checkpoint/optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:edd8c13f8967274a96ca7b3ebe0650a28c9586ed56d88d318df8624de39af605
|
3 |
+
size 43123028
|
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 sha256:08416b69f3ea1ef1c206c452662a86dc8094e9ca4c60db81379bdb5238fe1a17
|
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 sha256:7a02f5253505f86b36ab797e2d0183c254e06141b3cd157163c8863bd1f151e5
|
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": 500,
|
6 |
-
"global_step":
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
@@ -1729,6 +1729,1728 @@
|
|
1729 |
"learning_rate": 8.575059338986633e-05,
|
1730 |
"loss": 2.2807,
|
1731 |
"step": 246
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1732 |
}
|
1733 |
],
|
1734 |
"logging_steps": 1,
|
@@ -1748,7 +3470,7 @@
|
|
1748 |
"attributes": {}
|
1749 |
}
|
1750 |
},
|
1751 |
-
"total_flos":
|
1752 |
"train_batch_size": 4,
|
1753 |
"trial_name": null,
|
1754 |
"trial_params": null
|
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
+
"epoch": 0.5008908119114278,
|
5 |
"eval_steps": 500,
|
6 |
+
"global_step": 492,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
|
|
1729 |
"learning_rate": 8.575059338986633e-05,
|
1730 |
"loss": 2.2807,
|
1731 |
"step": 246
|
1732 |
+
},
|
1733 |
+
{
|
1734 |
+
"epoch": 0.25146347671163144,
|
1735 |
+
"grad_norm": 9.670249938964844,
|
1736 |
+
"learning_rate": 8.563812252348411e-05,
|
1737 |
+
"loss": 2.2475,
|
1738 |
+
"step": 247
|
1739 |
+
},
|
1740 |
+
{
|
1741 |
+
"epoch": 0.252481547467549,
|
1742 |
+
"grad_norm": 9.350224494934082,
|
1743 |
+
"learning_rate": 8.552528392082147e-05,
|
1744 |
+
"loss": 2.5073,
|
1745 |
+
"step": 248
|
1746 |
+
},
|
1747 |
+
{
|
1748 |
+
"epoch": 0.2534996182234665,
|
1749 |
+
"grad_norm": 9.628990173339844,
|
1750 |
+
"learning_rate": 8.541207874621718e-05,
|
1751 |
+
"loss": 2.0659,
|
1752 |
+
"step": 249
|
1753 |
+
},
|
1754 |
+
{
|
1755 |
+
"epoch": 0.2545176889793841,
|
1756 |
+
"grad_norm": 9.84993839263916,
|
1757 |
+
"learning_rate": 8.529850816779251e-05,
|
1758 |
+
"loss": 2.2365,
|
1759 |
+
"step": 250
|
1760 |
+
},
|
1761 |
+
{
|
1762 |
+
"epoch": 0.2555357597353016,
|
1763 |
+
"grad_norm": 8.845046997070312,
|
1764 |
+
"learning_rate": 8.518457335743926e-05,
|
1765 |
+
"loss": 3.1796,
|
1766 |
+
"step": 251
|
1767 |
+
},
|
1768 |
+
{
|
1769 |
+
"epoch": 0.25655383049121916,
|
1770 |
+
"grad_norm": 10.79970932006836,
|
1771 |
+
"learning_rate": 8.507027549080753e-05,
|
1772 |
+
"loss": 3.8036,
|
1773 |
+
"step": 252
|
1774 |
+
},
|
1775 |
+
{
|
1776 |
+
"epoch": 0.2575719012471367,
|
1777 |
+
"grad_norm": 8.892024993896484,
|
1778 |
+
"learning_rate": 8.495561574729369e-05,
|
1779 |
+
"loss": 2.9368,
|
1780 |
+
"step": 253
|
1781 |
+
},
|
1782 |
+
{
|
1783 |
+
"epoch": 0.2585899720030542,
|
1784 |
+
"grad_norm": 11.658991813659668,
|
1785 |
+
"learning_rate": 8.484059531002821e-05,
|
1786 |
+
"loss": 3.7456,
|
1787 |
+
"step": 254
|
1788 |
+
},
|
1789 |
+
{
|
1790 |
+
"epoch": 0.25960804275897176,
|
1791 |
+
"grad_norm": 11.338571548461914,
|
1792 |
+
"learning_rate": 8.472521536586335e-05,
|
1793 |
+
"loss": 3.9418,
|
1794 |
+
"step": 255
|
1795 |
+
},
|
1796 |
+
{
|
1797 |
+
"epoch": 0.2606261135148893,
|
1798 |
+
"grad_norm": 14.362560272216797,
|
1799 |
+
"learning_rate": 8.460947710536107e-05,
|
1800 |
+
"loss": 4.6011,
|
1801 |
+
"step": 256
|
1802 |
+
},
|
1803 |
+
{
|
1804 |
+
"epoch": 0.2616441842708068,
|
1805 |
+
"grad_norm": 13.662555694580078,
|
1806 |
+
"learning_rate": 8.449338172278059e-05,
|
1807 |
+
"loss": 5.049,
|
1808 |
+
"step": 257
|
1809 |
+
},
|
1810 |
+
{
|
1811 |
+
"epoch": 0.26266225502672436,
|
1812 |
+
"grad_norm": 15.036532402038574,
|
1813 |
+
"learning_rate": 8.437693041606618e-05,
|
1814 |
+
"loss": 4.0385,
|
1815 |
+
"step": 258
|
1816 |
+
},
|
1817 |
+
{
|
1818 |
+
"epoch": 0.2636803257826419,
|
1819 |
+
"grad_norm": 12.57422161102295,
|
1820 |
+
"learning_rate": 8.426012438683473e-05,
|
1821 |
+
"loss": 3.1101,
|
1822 |
+
"step": 259
|
1823 |
+
},
|
1824 |
+
{
|
1825 |
+
"epoch": 0.2646983965385594,
|
1826 |
+
"grad_norm": 8.874217987060547,
|
1827 |
+
"learning_rate": 8.414296484036339e-05,
|
1828 |
+
"loss": 2.3986,
|
1829 |
+
"step": 260
|
1830 |
+
},
|
1831 |
+
{
|
1832 |
+
"epoch": 0.26571646729447695,
|
1833 |
+
"grad_norm": 5.300018787384033,
|
1834 |
+
"learning_rate": 8.402545298557712e-05,
|
1835 |
+
"loss": 0.9408,
|
1836 |
+
"step": 261
|
1837 |
+
},
|
1838 |
+
{
|
1839 |
+
"epoch": 0.2667345380503945,
|
1840 |
+
"grad_norm": 6.752171039581299,
|
1841 |
+
"learning_rate": 8.390759003503623e-05,
|
1842 |
+
"loss": 1.8722,
|
1843 |
+
"step": 262
|
1844 |
+
},
|
1845 |
+
{
|
1846 |
+
"epoch": 0.267752608806312,
|
1847 |
+
"grad_norm": 7.508800029754639,
|
1848 |
+
"learning_rate": 8.378937720492384e-05,
|
1849 |
+
"loss": 1.7145,
|
1850 |
+
"step": 263
|
1851 |
+
},
|
1852 |
+
{
|
1853 |
+
"epoch": 0.26877067956222955,
|
1854 |
+
"grad_norm": 6.305329322814941,
|
1855 |
+
"learning_rate": 8.367081571503332e-05,
|
1856 |
+
"loss": 1.6567,
|
1857 |
+
"step": 264
|
1858 |
+
},
|
1859 |
+
{
|
1860 |
+
"epoch": 0.2697887503181471,
|
1861 |
+
"grad_norm": 9.371475219726562,
|
1862 |
+
"learning_rate": 8.355190678875578e-05,
|
1863 |
+
"loss": 2.4242,
|
1864 |
+
"step": 265
|
1865 |
+
},
|
1866 |
+
{
|
1867 |
+
"epoch": 0.2708068210740647,
|
1868 |
+
"grad_norm": 5.403337001800537,
|
1869 |
+
"learning_rate": 8.343265165306735e-05,
|
1870 |
+
"loss": 1.3716,
|
1871 |
+
"step": 266
|
1872 |
+
},
|
1873 |
+
{
|
1874 |
+
"epoch": 0.2718248918299822,
|
1875 |
+
"grad_norm": 8.240038871765137,
|
1876 |
+
"learning_rate": 8.331305153851658e-05,
|
1877 |
+
"loss": 2.2134,
|
1878 |
+
"step": 267
|
1879 |
+
},
|
1880 |
+
{
|
1881 |
+
"epoch": 0.27284296258589974,
|
1882 |
+
"grad_norm": 5.857060432434082,
|
1883 |
+
"learning_rate": 8.319310767921174e-05,
|
1884 |
+
"loss": 1.2823,
|
1885 |
+
"step": 268
|
1886 |
+
},
|
1887 |
+
{
|
1888 |
+
"epoch": 0.2738610333418173,
|
1889 |
+
"grad_norm": 6.11975622177124,
|
1890 |
+
"learning_rate": 8.307282131280804e-05,
|
1891 |
+
"loss": 1.7163,
|
1892 |
+
"step": 269
|
1893 |
+
},
|
1894 |
+
{
|
1895 |
+
"epoch": 0.2748791040977348,
|
1896 |
+
"grad_norm": 10.722909927368164,
|
1897 |
+
"learning_rate": 8.295219368049494e-05,
|
1898 |
+
"loss": 2.2343,
|
1899 |
+
"step": 270
|
1900 |
+
},
|
1901 |
+
{
|
1902 |
+
"epoch": 0.27589717485365234,
|
1903 |
+
"grad_norm": 6.3688507080078125,
|
1904 |
+
"learning_rate": 8.283122602698323e-05,
|
1905 |
+
"loss": 1.253,
|
1906 |
+
"step": 271
|
1907 |
+
},
|
1908 |
+
{
|
1909 |
+
"epoch": 0.27691524560956987,
|
1910 |
+
"grad_norm": 8.231119155883789,
|
1911 |
+
"learning_rate": 8.27099196004923e-05,
|
1912 |
+
"loss": 2.4005,
|
1913 |
+
"step": 272
|
1914 |
+
},
|
1915 |
+
{
|
1916 |
+
"epoch": 0.2779333163654874,
|
1917 |
+
"grad_norm": 6.623697757720947,
|
1918 |
+
"learning_rate": 8.258827565273718e-05,
|
1919 |
+
"loss": 1.5276,
|
1920 |
+
"step": 273
|
1921 |
+
},
|
1922 |
+
{
|
1923 |
+
"epoch": 0.27895138712140494,
|
1924 |
+
"grad_norm": 8.357768058776855,
|
1925 |
+
"learning_rate": 8.246629543891569e-05,
|
1926 |
+
"loss": 2.5312,
|
1927 |
+
"step": 274
|
1928 |
+
},
|
1929 |
+
{
|
1930 |
+
"epoch": 0.27996945787732247,
|
1931 |
+
"grad_norm": 7.037582874298096,
|
1932 |
+
"learning_rate": 8.23439802176954e-05,
|
1933 |
+
"loss": 2.6555,
|
1934 |
+
"step": 275
|
1935 |
+
},
|
1936 |
+
{
|
1937 |
+
"epoch": 0.28098752863324,
|
1938 |
+
"grad_norm": 8.760072708129883,
|
1939 |
+
"learning_rate": 8.222133125120076e-05,
|
1940 |
+
"loss": 2.051,
|
1941 |
+
"step": 276
|
1942 |
+
},
|
1943 |
+
{
|
1944 |
+
"epoch": 0.28200559938915754,
|
1945 |
+
"grad_norm": 10.407620429992676,
|
1946 |
+
"learning_rate": 8.209834980499995e-05,
|
1947 |
+
"loss": 2.7866,
|
1948 |
+
"step": 277
|
1949 |
+
},
|
1950 |
+
{
|
1951 |
+
"epoch": 0.28302367014507507,
|
1952 |
+
"grad_norm": 10.065641403198242,
|
1953 |
+
"learning_rate": 8.197503714809191e-05,
|
1954 |
+
"loss": 2.7393,
|
1955 |
+
"step": 278
|
1956 |
+
},
|
1957 |
+
{
|
1958 |
+
"epoch": 0.2840417409009926,
|
1959 |
+
"grad_norm": 8.072295188903809,
|
1960 |
+
"learning_rate": 8.185139455289322e-05,
|
1961 |
+
"loss": 2.1416,
|
1962 |
+
"step": 279
|
1963 |
+
},
|
1964 |
+
{
|
1965 |
+
"epoch": 0.28505981165691013,
|
1966 |
+
"grad_norm": 10.837465286254883,
|
1967 |
+
"learning_rate": 8.172742329522493e-05,
|
1968 |
+
"loss": 3.0516,
|
1969 |
+
"step": 280
|
1970 |
+
},
|
1971 |
+
{
|
1972 |
+
"epoch": 0.28607788241282767,
|
1973 |
+
"grad_norm": 8.73945426940918,
|
1974 |
+
"learning_rate": 8.160312465429952e-05,
|
1975 |
+
"loss": 2.6516,
|
1976 |
+
"step": 281
|
1977 |
+
},
|
1978 |
+
{
|
1979 |
+
"epoch": 0.28709595316874525,
|
1980 |
+
"grad_norm": 9.408519744873047,
|
1981 |
+
"learning_rate": 8.147849991270752e-05,
|
1982 |
+
"loss": 2.4367,
|
1983 |
+
"step": 282
|
1984 |
+
},
|
1985 |
+
{
|
1986 |
+
"epoch": 0.2881140239246628,
|
1987 |
+
"grad_norm": 8.221115112304688,
|
1988 |
+
"learning_rate": 8.135355035640444e-05,
|
1989 |
+
"loss": 2.4484,
|
1990 |
+
"step": 283
|
1991 |
+
},
|
1992 |
+
{
|
1993 |
+
"epoch": 0.2891320946805803,
|
1994 |
+
"grad_norm": 11.46916389465332,
|
1995 |
+
"learning_rate": 8.122827727469737e-05,
|
1996 |
+
"loss": 3.5208,
|
1997 |
+
"step": 284
|
1998 |
+
},
|
1999 |
+
{
|
2000 |
+
"epoch": 0.29015016543649785,
|
2001 |
+
"grad_norm": 8.721375465393066,
|
2002 |
+
"learning_rate": 8.110268196023179e-05,
|
2003 |
+
"loss": 2.3896,
|
2004 |
+
"step": 285
|
2005 |
+
},
|
2006 |
+
{
|
2007 |
+
"epoch": 0.2911682361924154,
|
2008 |
+
"grad_norm": 9.857149124145508,
|
2009 |
+
"learning_rate": 8.097676570897814e-05,
|
2010 |
+
"loss": 2.8248,
|
2011 |
+
"step": 286
|
2012 |
+
},
|
2013 |
+
{
|
2014 |
+
"epoch": 0.2921863069483329,
|
2015 |
+
"grad_norm": 7.732857704162598,
|
2016 |
+
"learning_rate": 8.085052982021847e-05,
|
2017 |
+
"loss": 2.0455,
|
2018 |
+
"step": 287
|
2019 |
+
},
|
2020 |
+
{
|
2021 |
+
"epoch": 0.29320437770425045,
|
2022 |
+
"grad_norm": 9.654264450073242,
|
2023 |
+
"learning_rate": 8.072397559653313e-05,
|
2024 |
+
"loss": 2.3959,
|
2025 |
+
"step": 288
|
2026 |
+
},
|
2027 |
+
{
|
2028 |
+
"epoch": 0.294222448460168,
|
2029 |
+
"grad_norm": 10.697869300842285,
|
2030 |
+
"learning_rate": 8.059710434378715e-05,
|
2031 |
+
"loss": 3.314,
|
2032 |
+
"step": 289
|
2033 |
+
},
|
2034 |
+
{
|
2035 |
+
"epoch": 0.2952405192160855,
|
2036 |
+
"grad_norm": 8.84398078918457,
|
2037 |
+
"learning_rate": 8.046991737111696e-05,
|
2038 |
+
"loss": 2.0514,
|
2039 |
+
"step": 290
|
2040 |
+
},
|
2041 |
+
{
|
2042 |
+
"epoch": 0.29625858997200305,
|
2043 |
+
"grad_norm": 7.673434257507324,
|
2044 |
+
"learning_rate": 8.034241599091665e-05,
|
2045 |
+
"loss": 2.165,
|
2046 |
+
"step": 291
|
2047 |
+
},
|
2048 |
+
{
|
2049 |
+
"epoch": 0.2972766607279206,
|
2050 |
+
"grad_norm": 10.299299240112305,
|
2051 |
+
"learning_rate": 8.021460151882471e-05,
|
2052 |
+
"loss": 3.0283,
|
2053 |
+
"step": 292
|
2054 |
+
},
|
2055 |
+
{
|
2056 |
+
"epoch": 0.2982947314838381,
|
2057 |
+
"grad_norm": 6.935961723327637,
|
2058 |
+
"learning_rate": 8.008647527371023e-05,
|
2059 |
+
"loss": 1.9187,
|
2060 |
+
"step": 293
|
2061 |
+
},
|
2062 |
+
{
|
2063 |
+
"epoch": 0.29931280223975565,
|
2064 |
+
"grad_norm": 9.410109519958496,
|
2065 |
+
"learning_rate": 7.995803857765933e-05,
|
2066 |
+
"loss": 2.4798,
|
2067 |
+
"step": 294
|
2068 |
+
},
|
2069 |
+
{
|
2070 |
+
"epoch": 0.3003308729956732,
|
2071 |
+
"grad_norm": 9.035164833068848,
|
2072 |
+
"learning_rate": 7.982929275596166e-05,
|
2073 |
+
"loss": 2.8312,
|
2074 |
+
"step": 295
|
2075 |
+
},
|
2076 |
+
{
|
2077 |
+
"epoch": 0.3013489437515907,
|
2078 |
+
"grad_norm": 8.214160919189453,
|
2079 |
+
"learning_rate": 7.970023913709652e-05,
|
2080 |
+
"loss": 2.4572,
|
2081 |
+
"step": 296
|
2082 |
+
},
|
2083 |
+
{
|
2084 |
+
"epoch": 0.30236701450750825,
|
2085 |
+
"grad_norm": 9.095396041870117,
|
2086 |
+
"learning_rate": 7.957087905271934e-05,
|
2087 |
+
"loss": 2.4834,
|
2088 |
+
"step": 297
|
2089 |
+
},
|
2090 |
+
{
|
2091 |
+
"epoch": 0.30338508526342584,
|
2092 |
+
"grad_norm": 9.806940078735352,
|
2093 |
+
"learning_rate": 7.944121383764776e-05,
|
2094 |
+
"loss": 2.6364,
|
2095 |
+
"step": 298
|
2096 |
+
},
|
2097 |
+
{
|
2098 |
+
"epoch": 0.30440315601934337,
|
2099 |
+
"grad_norm": 10.004327774047852,
|
2100 |
+
"learning_rate": 7.931124482984802e-05,
|
2101 |
+
"loss": 2.4236,
|
2102 |
+
"step": 299
|
2103 |
+
},
|
2104 |
+
{
|
2105 |
+
"epoch": 0.3054212267752609,
|
2106 |
+
"grad_norm": 12.964902877807617,
|
2107 |
+
"learning_rate": 7.918097337042105e-05,
|
2108 |
+
"loss": 2.542,
|
2109 |
+
"step": 300
|
2110 |
+
},
|
2111 |
+
{
|
2112 |
+
"epoch": 0.30643929753117843,
|
2113 |
+
"grad_norm": 7.418375015258789,
|
2114 |
+
"learning_rate": 7.905040080358868e-05,
|
2115 |
+
"loss": 2.7417,
|
2116 |
+
"step": 301
|
2117 |
+
},
|
2118 |
+
{
|
2119 |
+
"epoch": 0.30745736828709597,
|
2120 |
+
"grad_norm": 8.499213218688965,
|
2121 |
+
"learning_rate": 7.891952847667973e-05,
|
2122 |
+
"loss": 3.6269,
|
2123 |
+
"step": 302
|
2124 |
+
},
|
2125 |
+
{
|
2126 |
+
"epoch": 0.3084754390430135,
|
2127 |
+
"grad_norm": 8.255233764648438,
|
2128 |
+
"learning_rate": 7.878835774011615e-05,
|
2129 |
+
"loss": 2.7048,
|
2130 |
+
"step": 303
|
2131 |
+
},
|
2132 |
+
{
|
2133 |
+
"epoch": 0.30949350979893103,
|
2134 |
+
"grad_norm": 9.611371040344238,
|
2135 |
+
"learning_rate": 7.865688994739907e-05,
|
2136 |
+
"loss": 3.9859,
|
2137 |
+
"step": 304
|
2138 |
+
},
|
2139 |
+
{
|
2140 |
+
"epoch": 0.31051158055484857,
|
2141 |
+
"grad_norm": 9.825111389160156,
|
2142 |
+
"learning_rate": 7.85251264550948e-05,
|
2143 |
+
"loss": 3.2304,
|
2144 |
+
"step": 305
|
2145 |
+
},
|
2146 |
+
{
|
2147 |
+
"epoch": 0.3115296513107661,
|
2148 |
+
"grad_norm": 13.28979778289795,
|
2149 |
+
"learning_rate": 7.839306862282089e-05,
|
2150 |
+
"loss": 3.8321,
|
2151 |
+
"step": 306
|
2152 |
+
},
|
2153 |
+
{
|
2154 |
+
"epoch": 0.31254772206668363,
|
2155 |
+
"grad_norm": 16.412532806396484,
|
2156 |
+
"learning_rate": 7.826071781323207e-05,
|
2157 |
+
"loss": 5.4082,
|
2158 |
+
"step": 307
|
2159 |
+
},
|
2160 |
+
{
|
2161 |
+
"epoch": 0.31356579282260116,
|
2162 |
+
"grad_norm": 9.727624893188477,
|
2163 |
+
"learning_rate": 7.812807539200622e-05,
|
2164 |
+
"loss": 2.5533,
|
2165 |
+
"step": 308
|
2166 |
+
},
|
2167 |
+
{
|
2168 |
+
"epoch": 0.3145838635785187,
|
2169 |
+
"grad_norm": 8.238031387329102,
|
2170 |
+
"learning_rate": 7.799514272783014e-05,
|
2171 |
+
"loss": 2.2857,
|
2172 |
+
"step": 309
|
2173 |
+
},
|
2174 |
+
{
|
2175 |
+
"epoch": 0.31560193433443623,
|
2176 |
+
"grad_norm": 6.882323741912842,
|
2177 |
+
"learning_rate": 7.786192119238567e-05,
|
2178 |
+
"loss": 2.1371,
|
2179 |
+
"step": 310
|
2180 |
+
},
|
2181 |
+
{
|
2182 |
+
"epoch": 0.31662000509035376,
|
2183 |
+
"grad_norm": 5.3293280601501465,
|
2184 |
+
"learning_rate": 7.772841216033533e-05,
|
2185 |
+
"loss": 1.1834,
|
2186 |
+
"step": 311
|
2187 |
+
},
|
2188 |
+
{
|
2189 |
+
"epoch": 0.3176380758462713,
|
2190 |
+
"grad_norm": 5.593384265899658,
|
2191 |
+
"learning_rate": 7.759461700930823e-05,
|
2192 |
+
"loss": 1.4746,
|
2193 |
+
"step": 312
|
2194 |
+
},
|
2195 |
+
{
|
2196 |
+
"epoch": 0.31865614660218883,
|
2197 |
+
"grad_norm": 5.281317234039307,
|
2198 |
+
"learning_rate": 7.746053711988583e-05,
|
2199 |
+
"loss": 1.1387,
|
2200 |
+
"step": 313
|
2201 |
+
},
|
2202 |
+
{
|
2203 |
+
"epoch": 0.3196742173581064,
|
2204 |
+
"grad_norm": 6.735507965087891,
|
2205 |
+
"learning_rate": 7.73261738755877e-05,
|
2206 |
+
"loss": 1.9801,
|
2207 |
+
"step": 314
|
2208 |
+
},
|
2209 |
+
{
|
2210 |
+
"epoch": 0.32069228811402395,
|
2211 |
+
"grad_norm": 6.708221912384033,
|
2212 |
+
"learning_rate": 7.719152866285721e-05,
|
2213 |
+
"loss": 1.863,
|
2214 |
+
"step": 315
|
2215 |
+
},
|
2216 |
+
{
|
2217 |
+
"epoch": 0.3217103588699415,
|
2218 |
+
"grad_norm": 8.238001823425293,
|
2219 |
+
"learning_rate": 7.70566028710473e-05,
|
2220 |
+
"loss": 2.16,
|
2221 |
+
"step": 316
|
2222 |
+
},
|
2223 |
+
{
|
2224 |
+
"epoch": 0.322728429625859,
|
2225 |
+
"grad_norm": 6.396310329437256,
|
2226 |
+
"learning_rate": 7.692139789240611e-05,
|
2227 |
+
"loss": 2.1722,
|
2228 |
+
"step": 317
|
2229 |
+
},
|
2230 |
+
{
|
2231 |
+
"epoch": 0.32374650038177655,
|
2232 |
+
"grad_norm": 8.0552396774292,
|
2233 |
+
"learning_rate": 7.678591512206255e-05,
|
2234 |
+
"loss": 2.2866,
|
2235 |
+
"step": 318
|
2236 |
+
},
|
2237 |
+
{
|
2238 |
+
"epoch": 0.3247645711376941,
|
2239 |
+
"grad_norm": 5.051511287689209,
|
2240 |
+
"learning_rate": 7.665015595801197e-05,
|
2241 |
+
"loss": 1.0846,
|
2242 |
+
"step": 319
|
2243 |
+
},
|
2244 |
+
{
|
2245 |
+
"epoch": 0.3257826418936116,
|
2246 |
+
"grad_norm": 7.378931045532227,
|
2247 |
+
"learning_rate": 7.651412180110176e-05,
|
2248 |
+
"loss": 1.7923,
|
2249 |
+
"step": 320
|
2250 |
+
},
|
2251 |
+
{
|
2252 |
+
"epoch": 0.32680071264952915,
|
2253 |
+
"grad_norm": 7.6839118003845215,
|
2254 |
+
"learning_rate": 7.637781405501681e-05,
|
2255 |
+
"loss": 1.3504,
|
2256 |
+
"step": 321
|
2257 |
+
},
|
2258 |
+
{
|
2259 |
+
"epoch": 0.3278187834054467,
|
2260 |
+
"grad_norm": 7.594010829925537,
|
2261 |
+
"learning_rate": 7.624123412626512e-05,
|
2262 |
+
"loss": 2.5312,
|
2263 |
+
"step": 322
|
2264 |
+
},
|
2265 |
+
{
|
2266 |
+
"epoch": 0.3288368541613642,
|
2267 |
+
"grad_norm": 7.310485363006592,
|
2268 |
+
"learning_rate": 7.610438342416319e-05,
|
2269 |
+
"loss": 1.8773,
|
2270 |
+
"step": 323
|
2271 |
+
},
|
2272 |
+
{
|
2273 |
+
"epoch": 0.32985492491728174,
|
2274 |
+
"grad_norm": 10.614038467407227,
|
2275 |
+
"learning_rate": 7.596726336082158e-05,
|
2276 |
+
"loss": 2.8128,
|
2277 |
+
"step": 324
|
2278 |
+
},
|
2279 |
+
{
|
2280 |
+
"epoch": 0.3308729956731993,
|
2281 |
+
"grad_norm": 7.768847465515137,
|
2282 |
+
"learning_rate": 7.582987535113023e-05,
|
2283 |
+
"loss": 2.2407,
|
2284 |
+
"step": 325
|
2285 |
+
},
|
2286 |
+
{
|
2287 |
+
"epoch": 0.3318910664291168,
|
2288 |
+
"grad_norm": 9.793490409851074,
|
2289 |
+
"learning_rate": 7.569222081274395e-05,
|
2290 |
+
"loss": 2.3751,
|
2291 |
+
"step": 326
|
2292 |
+
},
|
2293 |
+
{
|
2294 |
+
"epoch": 0.33290913718503434,
|
2295 |
+
"grad_norm": 9.077165603637695,
|
2296 |
+
"learning_rate": 7.555430116606778e-05,
|
2297 |
+
"loss": 2.3227,
|
2298 |
+
"step": 327
|
2299 |
+
},
|
2300 |
+
{
|
2301 |
+
"epoch": 0.3339272079409519,
|
2302 |
+
"grad_norm": 8.083662986755371,
|
2303 |
+
"learning_rate": 7.541611783424225e-05,
|
2304 |
+
"loss": 2.4783,
|
2305 |
+
"step": 328
|
2306 |
+
},
|
2307 |
+
{
|
2308 |
+
"epoch": 0.3349452786968694,
|
2309 |
+
"grad_norm": 8.170014381408691,
|
2310 |
+
"learning_rate": 7.527767224312883e-05,
|
2311 |
+
"loss": 2.5388,
|
2312 |
+
"step": 329
|
2313 |
+
},
|
2314 |
+
{
|
2315 |
+
"epoch": 0.33596334945278694,
|
2316 |
+
"grad_norm": 7.519252777099609,
|
2317 |
+
"learning_rate": 7.513896582129508e-05,
|
2318 |
+
"loss": 2.384,
|
2319 |
+
"step": 330
|
2320 |
+
},
|
2321 |
+
{
|
2322 |
+
"epoch": 0.33698142020870453,
|
2323 |
+
"grad_norm": 9.94649887084961,
|
2324 |
+
"learning_rate": 7.500000000000001e-05,
|
2325 |
+
"loss": 2.7797,
|
2326 |
+
"step": 331
|
2327 |
+
},
|
2328 |
+
{
|
2329 |
+
"epoch": 0.33799949096462206,
|
2330 |
+
"grad_norm": 8.595416069030762,
|
2331 |
+
"learning_rate": 7.486077621317926e-05,
|
2332 |
+
"loss": 2.7035,
|
2333 |
+
"step": 332
|
2334 |
+
},
|
2335 |
+
{
|
2336 |
+
"epoch": 0.3390175617205396,
|
2337 |
+
"grad_norm": 7.753424644470215,
|
2338 |
+
"learning_rate": 7.472129589743033e-05,
|
2339 |
+
"loss": 2.0287,
|
2340 |
+
"step": 333
|
2341 |
+
},
|
2342 |
+
{
|
2343 |
+
"epoch": 0.34003563247645713,
|
2344 |
+
"grad_norm": 9.271249771118164,
|
2345 |
+
"learning_rate": 7.458156049199775e-05,
|
2346 |
+
"loss": 2.1601,
|
2347 |
+
"step": 334
|
2348 |
+
},
|
2349 |
+
{
|
2350 |
+
"epoch": 0.34105370323237466,
|
2351 |
+
"grad_norm": 8.564419746398926,
|
2352 |
+
"learning_rate": 7.44415714387582e-05,
|
2353 |
+
"loss": 2.9013,
|
2354 |
+
"step": 335
|
2355 |
+
},
|
2356 |
+
{
|
2357 |
+
"epoch": 0.3420717739882922,
|
2358 |
+
"grad_norm": 8.431052207946777,
|
2359 |
+
"learning_rate": 7.430133018220567e-05,
|
2360 |
+
"loss": 2.4643,
|
2361 |
+
"step": 336
|
2362 |
+
},
|
2363 |
+
{
|
2364 |
+
"epoch": 0.3430898447442097,
|
2365 |
+
"grad_norm": 6.7154436111450195,
|
2366 |
+
"learning_rate": 7.416083816943653e-05,
|
2367 |
+
"loss": 2.271,
|
2368 |
+
"step": 337
|
2369 |
+
},
|
2370 |
+
{
|
2371 |
+
"epoch": 0.34410791550012726,
|
2372 |
+
"grad_norm": 9.483381271362305,
|
2373 |
+
"learning_rate": 7.402009685013463e-05,
|
2374 |
+
"loss": 2.489,
|
2375 |
+
"step": 338
|
2376 |
+
},
|
2377 |
+
{
|
2378 |
+
"epoch": 0.3451259862560448,
|
2379 |
+
"grad_norm": 7.885382175445557,
|
2380 |
+
"learning_rate": 7.38791076765563e-05,
|
2381 |
+
"loss": 3.1875,
|
2382 |
+
"step": 339
|
2383 |
+
},
|
2384 |
+
{
|
2385 |
+
"epoch": 0.3461440570119623,
|
2386 |
+
"grad_norm": 7.622438430786133,
|
2387 |
+
"learning_rate": 7.373787210351541e-05,
|
2388 |
+
"loss": 2.0865,
|
2389 |
+
"step": 340
|
2390 |
+
},
|
2391 |
+
{
|
2392 |
+
"epoch": 0.34716212776787986,
|
2393 |
+
"grad_norm": 7.785037517547607,
|
2394 |
+
"learning_rate": 7.359639158836828e-05,
|
2395 |
+
"loss": 2.0806,
|
2396 |
+
"step": 341
|
2397 |
+
},
|
2398 |
+
{
|
2399 |
+
"epoch": 0.3481801985237974,
|
2400 |
+
"grad_norm": 7.861757755279541,
|
2401 |
+
"learning_rate": 7.345466759099875e-05,
|
2402 |
+
"loss": 2.4029,
|
2403 |
+
"step": 342
|
2404 |
+
},
|
2405 |
+
{
|
2406 |
+
"epoch": 0.3491982692797149,
|
2407 |
+
"grad_norm": 8.165399551391602,
|
2408 |
+
"learning_rate": 7.331270157380303e-05,
|
2409 |
+
"loss": 2.1438,
|
2410 |
+
"step": 343
|
2411 |
+
},
|
2412 |
+
{
|
2413 |
+
"epoch": 0.35021634003563246,
|
2414 |
+
"grad_norm": 8.010607719421387,
|
2415 |
+
"learning_rate": 7.317049500167465e-05,
|
2416 |
+
"loss": 1.8633,
|
2417 |
+
"step": 344
|
2418 |
+
},
|
2419 |
+
{
|
2420 |
+
"epoch": 0.35123441079155,
|
2421 |
+
"grad_norm": 8.535947799682617,
|
2422 |
+
"learning_rate": 7.302804934198936e-05,
|
2423 |
+
"loss": 2.2122,
|
2424 |
+
"step": 345
|
2425 |
+
},
|
2426 |
+
{
|
2427 |
+
"epoch": 0.3522524815474675,
|
2428 |
+
"grad_norm": 8.71279239654541,
|
2429 |
+
"learning_rate": 7.28853660645899e-05,
|
2430 |
+
"loss": 2.4223,
|
2431 |
+
"step": 346
|
2432 |
+
},
|
2433 |
+
{
|
2434 |
+
"epoch": 0.3532705523033851,
|
2435 |
+
"grad_norm": 9.44794750213623,
|
2436 |
+
"learning_rate": 7.274244664177097e-05,
|
2437 |
+
"loss": 2.5881,
|
2438 |
+
"step": 347
|
2439 |
+
},
|
2440 |
+
{
|
2441 |
+
"epoch": 0.35428862305930264,
|
2442 |
+
"grad_norm": 10.474303245544434,
|
2443 |
+
"learning_rate": 7.259929254826392e-05,
|
2444 |
+
"loss": 1.9121,
|
2445 |
+
"step": 348
|
2446 |
+
},
|
2447 |
+
{
|
2448 |
+
"epoch": 0.3553066938152202,
|
2449 |
+
"grad_norm": 9.976726531982422,
|
2450 |
+
"learning_rate": 7.245590526122159e-05,
|
2451 |
+
"loss": 2.3289,
|
2452 |
+
"step": 349
|
2453 |
+
},
|
2454 |
+
{
|
2455 |
+
"epoch": 0.3563247645711377,
|
2456 |
+
"grad_norm": 10.359407424926758,
|
2457 |
+
"learning_rate": 7.231228626020304e-05,
|
2458 |
+
"loss": 2.0781,
|
2459 |
+
"step": 350
|
2460 |
+
},
|
2461 |
+
{
|
2462 |
+
"epoch": 0.35734283532705524,
|
2463 |
+
"grad_norm": 10.392163276672363,
|
2464 |
+
"learning_rate": 7.216843702715831e-05,
|
2465 |
+
"loss": 4.3803,
|
2466 |
+
"step": 351
|
2467 |
+
},
|
2468 |
+
{
|
2469 |
+
"epoch": 0.3583609060829728,
|
2470 |
+
"grad_norm": 8.603669166564941,
|
2471 |
+
"learning_rate": 7.202435904641315e-05,
|
2472 |
+
"loss": 3.3045,
|
2473 |
+
"step": 352
|
2474 |
+
},
|
2475 |
+
{
|
2476 |
+
"epoch": 0.3593789768388903,
|
2477 |
+
"grad_norm": 11.44635009765625,
|
2478 |
+
"learning_rate": 7.188005380465364e-05,
|
2479 |
+
"loss": 5.3026,
|
2480 |
+
"step": 353
|
2481 |
+
},
|
2482 |
+
{
|
2483 |
+
"epoch": 0.36039704759480784,
|
2484 |
+
"grad_norm": 11.560089111328125,
|
2485 |
+
"learning_rate": 7.173552279091087e-05,
|
2486 |
+
"loss": 4.0946,
|
2487 |
+
"step": 354
|
2488 |
+
},
|
2489 |
+
{
|
2490 |
+
"epoch": 0.3614151183507254,
|
2491 |
+
"grad_norm": 12.211684226989746,
|
2492 |
+
"learning_rate": 7.159076749654559e-05,
|
2493 |
+
"loss": 3.7029,
|
2494 |
+
"step": 355
|
2495 |
+
},
|
2496 |
+
{
|
2497 |
+
"epoch": 0.3624331891066429,
|
2498 |
+
"grad_norm": 13.772133827209473,
|
2499 |
+
"learning_rate": 7.144578941523284e-05,
|
2500 |
+
"loss": 4.3069,
|
2501 |
+
"step": 356
|
2502 |
+
},
|
2503 |
+
{
|
2504 |
+
"epoch": 0.36345125986256044,
|
2505 |
+
"grad_norm": 11.917257308959961,
|
2506 |
+
"learning_rate": 7.130059004294647e-05,
|
2507 |
+
"loss": 3.3222,
|
2508 |
+
"step": 357
|
2509 |
+
},
|
2510 |
+
{
|
2511 |
+
"epoch": 0.36446933061847797,
|
2512 |
+
"grad_norm": 15.86817741394043,
|
2513 |
+
"learning_rate": 7.115517087794381e-05,
|
2514 |
+
"loss": 3.5182,
|
2515 |
+
"step": 358
|
2516 |
+
},
|
2517 |
+
{
|
2518 |
+
"epoch": 0.3654874013743955,
|
2519 |
+
"grad_norm": 6.566526412963867,
|
2520 |
+
"learning_rate": 7.10095334207501e-05,
|
2521 |
+
"loss": 1.7594,
|
2522 |
+
"step": 359
|
2523 |
+
},
|
2524 |
+
{
|
2525 |
+
"epoch": 0.36650547213031304,
|
2526 |
+
"grad_norm": 6.9844560623168945,
|
2527 |
+
"learning_rate": 7.086367917414306e-05,
|
2528 |
+
"loss": 1.6552,
|
2529 |
+
"step": 360
|
2530 |
+
},
|
2531 |
+
{
|
2532 |
+
"epoch": 0.36752354288623057,
|
2533 |
+
"grad_norm": 9.259458541870117,
|
2534 |
+
"learning_rate": 7.07176096431374e-05,
|
2535 |
+
"loss": 2.2072,
|
2536 |
+
"step": 361
|
2537 |
+
},
|
2538 |
+
{
|
2539 |
+
"epoch": 0.3685416136421481,
|
2540 |
+
"grad_norm": 8.759178161621094,
|
2541 |
+
"learning_rate": 7.057132633496923e-05,
|
2542 |
+
"loss": 1.7327,
|
2543 |
+
"step": 362
|
2544 |
+
},
|
2545 |
+
{
|
2546 |
+
"epoch": 0.3695596843980657,
|
2547 |
+
"grad_norm": 7.46987247467041,
|
2548 |
+
"learning_rate": 7.042483075908062e-05,
|
2549 |
+
"loss": 1.6727,
|
2550 |
+
"step": 363
|
2551 |
+
},
|
2552 |
+
{
|
2553 |
+
"epoch": 0.3705777551539832,
|
2554 |
+
"grad_norm": 7.011016368865967,
|
2555 |
+
"learning_rate": 7.027812442710385e-05,
|
2556 |
+
"loss": 1.643,
|
2557 |
+
"step": 364
|
2558 |
+
},
|
2559 |
+
{
|
2560 |
+
"epoch": 0.37159582590990076,
|
2561 |
+
"grad_norm": 8.475665092468262,
|
2562 |
+
"learning_rate": 7.013120885284598e-05,
|
2563 |
+
"loss": 2.4295,
|
2564 |
+
"step": 365
|
2565 |
+
},
|
2566 |
+
{
|
2567 |
+
"epoch": 0.3726138966658183,
|
2568 |
+
"grad_norm": 5.769803047180176,
|
2569 |
+
"learning_rate": 6.998408555227314e-05,
|
2570 |
+
"loss": 1.3708,
|
2571 |
+
"step": 366
|
2572 |
+
},
|
2573 |
+
{
|
2574 |
+
"epoch": 0.3736319674217358,
|
2575 |
+
"grad_norm": 6.653828144073486,
|
2576 |
+
"learning_rate": 6.983675604349493e-05,
|
2577 |
+
"loss": 2.111,
|
2578 |
+
"step": 367
|
2579 |
+
},
|
2580 |
+
{
|
2581 |
+
"epoch": 0.37465003817765336,
|
2582 |
+
"grad_norm": 8.172953605651855,
|
2583 |
+
"learning_rate": 6.968922184674867e-05,
|
2584 |
+
"loss": 2.3177,
|
2585 |
+
"step": 368
|
2586 |
+
},
|
2587 |
+
{
|
2588 |
+
"epoch": 0.3756681089335709,
|
2589 |
+
"grad_norm": 6.391868591308594,
|
2590 |
+
"learning_rate": 6.954148448438389e-05,
|
2591 |
+
"loss": 1.2711,
|
2592 |
+
"step": 369
|
2593 |
+
},
|
2594 |
+
{
|
2595 |
+
"epoch": 0.3766861796894884,
|
2596 |
+
"grad_norm": 8.03226375579834,
|
2597 |
+
"learning_rate": 6.93935454808464e-05,
|
2598 |
+
"loss": 1.7306,
|
2599 |
+
"step": 370
|
2600 |
+
},
|
2601 |
+
{
|
2602 |
+
"epoch": 0.37770425044540595,
|
2603 |
+
"grad_norm": 5.273244857788086,
|
2604 |
+
"learning_rate": 6.924540636266272e-05,
|
2605 |
+
"loss": 1.3542,
|
2606 |
+
"step": 371
|
2607 |
+
},
|
2608 |
+
{
|
2609 |
+
"epoch": 0.3787223212013235,
|
2610 |
+
"grad_norm": 9.628256797790527,
|
2611 |
+
"learning_rate": 6.909706865842429e-05,
|
2612 |
+
"loss": 2.7357,
|
2613 |
+
"step": 372
|
2614 |
+
},
|
2615 |
+
{
|
2616 |
+
"epoch": 0.379740391957241,
|
2617 |
+
"grad_norm": 11.36279582977295,
|
2618 |
+
"learning_rate": 6.894853389877163e-05,
|
2619 |
+
"loss": 2.2367,
|
2620 |
+
"step": 373
|
2621 |
+
},
|
2622 |
+
{
|
2623 |
+
"epoch": 0.38075846271315855,
|
2624 |
+
"grad_norm": 8.72659969329834,
|
2625 |
+
"learning_rate": 6.879980361637866e-05,
|
2626 |
+
"loss": 2.2005,
|
2627 |
+
"step": 374
|
2628 |
+
},
|
2629 |
+
{
|
2630 |
+
"epoch": 0.3817765334690761,
|
2631 |
+
"grad_norm": 7.1913042068481445,
|
2632 |
+
"learning_rate": 6.86508793459368e-05,
|
2633 |
+
"loss": 2.1034,
|
2634 |
+
"step": 375
|
2635 |
+
},
|
2636 |
+
{
|
2637 |
+
"epoch": 0.3827946042249936,
|
2638 |
+
"grad_norm": 8.96323299407959,
|
2639 |
+
"learning_rate": 6.850176262413912e-05,
|
2640 |
+
"loss": 2.8465,
|
2641 |
+
"step": 376
|
2642 |
+
},
|
2643 |
+
{
|
2644 |
+
"epoch": 0.38381267498091115,
|
2645 |
+
"grad_norm": 6.918330192565918,
|
2646 |
+
"learning_rate": 6.835245498966461e-05,
|
2647 |
+
"loss": 1.8181,
|
2648 |
+
"step": 377
|
2649 |
+
},
|
2650 |
+
{
|
2651 |
+
"epoch": 0.3848307457368287,
|
2652 |
+
"grad_norm": 9.063780784606934,
|
2653 |
+
"learning_rate": 6.820295798316214e-05,
|
2654 |
+
"loss": 2.4932,
|
2655 |
+
"step": 378
|
2656 |
+
},
|
2657 |
+
{
|
2658 |
+
"epoch": 0.38584881649274627,
|
2659 |
+
"grad_norm": 13.343623161315918,
|
2660 |
+
"learning_rate": 6.805327314723468e-05,
|
2661 |
+
"loss": 3.0215,
|
2662 |
+
"step": 379
|
2663 |
+
},
|
2664 |
+
{
|
2665 |
+
"epoch": 0.3868668872486638,
|
2666 |
+
"grad_norm": 7.741687774658203,
|
2667 |
+
"learning_rate": 6.790340202642332e-05,
|
2668 |
+
"loss": 2.1827,
|
2669 |
+
"step": 380
|
2670 |
+
},
|
2671 |
+
{
|
2672 |
+
"epoch": 0.38788495800458134,
|
2673 |
+
"grad_norm": 5.362514495849609,
|
2674 |
+
"learning_rate": 6.775334616719136e-05,
|
2675 |
+
"loss": 1.7059,
|
2676 |
+
"step": 381
|
2677 |
+
},
|
2678 |
+
{
|
2679 |
+
"epoch": 0.38890302876049887,
|
2680 |
+
"grad_norm": 10.59506893157959,
|
2681 |
+
"learning_rate": 6.760310711790832e-05,
|
2682 |
+
"loss": 2.6664,
|
2683 |
+
"step": 382
|
2684 |
+
},
|
2685 |
+
{
|
2686 |
+
"epoch": 0.3899210995164164,
|
2687 |
+
"grad_norm": 9.036832809448242,
|
2688 |
+
"learning_rate": 6.745268642883404e-05,
|
2689 |
+
"loss": 2.5482,
|
2690 |
+
"step": 383
|
2691 |
+
},
|
2692 |
+
{
|
2693 |
+
"epoch": 0.39093917027233394,
|
2694 |
+
"grad_norm": 7.859615802764893,
|
2695 |
+
"learning_rate": 6.73020856521026e-05,
|
2696 |
+
"loss": 2.1062,
|
2697 |
+
"step": 384
|
2698 |
+
},
|
2699 |
+
{
|
2700 |
+
"epoch": 0.39195724102825147,
|
2701 |
+
"grad_norm": 9.639580726623535,
|
2702 |
+
"learning_rate": 6.715130634170635e-05,
|
2703 |
+
"loss": 3.0521,
|
2704 |
+
"step": 385
|
2705 |
+
},
|
2706 |
+
{
|
2707 |
+
"epoch": 0.392975311784169,
|
2708 |
+
"grad_norm": 8.098716735839844,
|
2709 |
+
"learning_rate": 6.700035005347983e-05,
|
2710 |
+
"loss": 2.6295,
|
2711 |
+
"step": 386
|
2712 |
+
},
|
2713 |
+
{
|
2714 |
+
"epoch": 0.39399338254008653,
|
2715 |
+
"grad_norm": 11.05691146850586,
|
2716 |
+
"learning_rate": 6.684921834508379e-05,
|
2717 |
+
"loss": 2.6122,
|
2718 |
+
"step": 387
|
2719 |
+
},
|
2720 |
+
{
|
2721 |
+
"epoch": 0.39501145329600407,
|
2722 |
+
"grad_norm": 9.071178436279297,
|
2723 |
+
"learning_rate": 6.669791277598904e-05,
|
2724 |
+
"loss": 2.3797,
|
2725 |
+
"step": 388
|
2726 |
+
},
|
2727 |
+
{
|
2728 |
+
"epoch": 0.3960295240519216,
|
2729 |
+
"grad_norm": 9.826159477233887,
|
2730 |
+
"learning_rate": 6.654643490746042e-05,
|
2731 |
+
"loss": 2.4635,
|
2732 |
+
"step": 389
|
2733 |
+
},
|
2734 |
+
{
|
2735 |
+
"epoch": 0.39704759480783913,
|
2736 |
+
"grad_norm": 7.310181140899658,
|
2737 |
+
"learning_rate": 6.639478630254064e-05,
|
2738 |
+
"loss": 2.06,
|
2739 |
+
"step": 390
|
2740 |
+
},
|
2741 |
+
{
|
2742 |
+
"epoch": 0.39806566556375667,
|
2743 |
+
"grad_norm": 9.507575035095215,
|
2744 |
+
"learning_rate": 6.624296852603419e-05,
|
2745 |
+
"loss": 2.9877,
|
2746 |
+
"step": 391
|
2747 |
+
},
|
2748 |
+
{
|
2749 |
+
"epoch": 0.3990837363196742,
|
2750 |
+
"grad_norm": 7.882664680480957,
|
2751 |
+
"learning_rate": 6.609098314449116e-05,
|
2752 |
+
"loss": 2.1182,
|
2753 |
+
"step": 392
|
2754 |
+
},
|
2755 |
+
{
|
2756 |
+
"epoch": 0.40010180707559173,
|
2757 |
+
"grad_norm": 9.889808654785156,
|
2758 |
+
"learning_rate": 6.593883172619111e-05,
|
2759 |
+
"loss": 3.1559,
|
2760 |
+
"step": 393
|
2761 |
+
},
|
2762 |
+
{
|
2763 |
+
"epoch": 0.40111987783150926,
|
2764 |
+
"grad_norm": 9.180678367614746,
|
2765 |
+
"learning_rate": 6.578651584112686e-05,
|
2766 |
+
"loss": 2.6704,
|
2767 |
+
"step": 394
|
2768 |
+
},
|
2769 |
+
{
|
2770 |
+
"epoch": 0.40213794858742685,
|
2771 |
+
"grad_norm": 8.127158164978027,
|
2772 |
+
"learning_rate": 6.563403706098833e-05,
|
2773 |
+
"loss": 2.1218,
|
2774 |
+
"step": 395
|
2775 |
+
},
|
2776 |
+
{
|
2777 |
+
"epoch": 0.4031560193433444,
|
2778 |
+
"grad_norm": 9.298659324645996,
|
2779 |
+
"learning_rate": 6.548139695914622e-05,
|
2780 |
+
"loss": 3.3196,
|
2781 |
+
"step": 396
|
2782 |
+
},
|
2783 |
+
{
|
2784 |
+
"epoch": 0.4041740900992619,
|
2785 |
+
"grad_norm": 7.540867328643799,
|
2786 |
+
"learning_rate": 6.532859711063594e-05,
|
2787 |
+
"loss": 1.9205,
|
2788 |
+
"step": 397
|
2789 |
+
},
|
2790 |
+
{
|
2791 |
+
"epoch": 0.40519216085517945,
|
2792 |
+
"grad_norm": 8.0338716506958,
|
2793 |
+
"learning_rate": 6.51756390921412e-05,
|
2794 |
+
"loss": 2.2858,
|
2795 |
+
"step": 398
|
2796 |
+
},
|
2797 |
+
{
|
2798 |
+
"epoch": 0.406210231611097,
|
2799 |
+
"grad_norm": 8.012909889221191,
|
2800 |
+
"learning_rate": 6.502252448197782e-05,
|
2801 |
+
"loss": 2.3761,
|
2802 |
+
"step": 399
|
2803 |
+
},
|
2804 |
+
{
|
2805 |
+
"epoch": 0.4072283023670145,
|
2806 |
+
"grad_norm": 7.803510665893555,
|
2807 |
+
"learning_rate": 6.486925486007742e-05,
|
2808 |
+
"loss": 2.0418,
|
2809 |
+
"step": 400
|
2810 |
+
},
|
2811 |
+
{
|
2812 |
+
"epoch": 0.40824637312293205,
|
2813 |
+
"grad_norm": 6.121700286865234,
|
2814 |
+
"learning_rate": 6.471583180797121e-05,
|
2815 |
+
"loss": 2.1481,
|
2816 |
+
"step": 401
|
2817 |
+
},
|
2818 |
+
{
|
2819 |
+
"epoch": 0.4092644438788496,
|
2820 |
+
"grad_norm": 8.690316200256348,
|
2821 |
+
"learning_rate": 6.456225690877344e-05,
|
2822 |
+
"loss": 3.0496,
|
2823 |
+
"step": 402
|
2824 |
+
},
|
2825 |
+
{
|
2826 |
+
"epoch": 0.4102825146347671,
|
2827 |
+
"grad_norm": 8.961786270141602,
|
2828 |
+
"learning_rate": 6.440853174716534e-05,
|
2829 |
+
"loss": 3.4188,
|
2830 |
+
"step": 403
|
2831 |
+
},
|
2832 |
+
{
|
2833 |
+
"epoch": 0.41130058539068465,
|
2834 |
+
"grad_norm": 9.848220825195312,
|
2835 |
+
"learning_rate": 6.425465790937861e-05,
|
2836 |
+
"loss": 3.948,
|
2837 |
+
"step": 404
|
2838 |
+
},
|
2839 |
+
{
|
2840 |
+
"epoch": 0.4123186561466022,
|
2841 |
+
"grad_norm": 11.758772850036621,
|
2842 |
+
"learning_rate": 6.410063698317901e-05,
|
2843 |
+
"loss": 4.4288,
|
2844 |
+
"step": 405
|
2845 |
+
},
|
2846 |
+
{
|
2847 |
+
"epoch": 0.4133367269025197,
|
2848 |
+
"grad_norm": 10.132964134216309,
|
2849 |
+
"learning_rate": 6.394647055785017e-05,
|
2850 |
+
"loss": 3.4126,
|
2851 |
+
"step": 406
|
2852 |
+
},
|
2853 |
+
{
|
2854 |
+
"epoch": 0.41435479765843725,
|
2855 |
+
"grad_norm": 9.949785232543945,
|
2856 |
+
"learning_rate": 6.379216022417696e-05,
|
2857 |
+
"loss": 2.584,
|
2858 |
+
"step": 407
|
2859 |
+
},
|
2860 |
+
{
|
2861 |
+
"epoch": 0.4153728684143548,
|
2862 |
+
"grad_norm": 13.347235679626465,
|
2863 |
+
"learning_rate": 6.363770757442927e-05,
|
2864 |
+
"loss": 2.8766,
|
2865 |
+
"step": 408
|
2866 |
+
},
|
2867 |
+
{
|
2868 |
+
"epoch": 0.4163909391702723,
|
2869 |
+
"grad_norm": 7.714400768280029,
|
2870 |
+
"learning_rate": 6.348311420234542e-05,
|
2871 |
+
"loss": 2.4595,
|
2872 |
+
"step": 409
|
2873 |
+
},
|
2874 |
+
{
|
2875 |
+
"epoch": 0.41740900992618984,
|
2876 |
+
"grad_norm": 5.146259307861328,
|
2877 |
+
"learning_rate": 6.332838170311585e-05,
|
2878 |
+
"loss": 1.4562,
|
2879 |
+
"step": 410
|
2880 |
+
},
|
2881 |
+
{
|
2882 |
+
"epoch": 0.41842708068210743,
|
2883 |
+
"grad_norm": 5.76894474029541,
|
2884 |
+
"learning_rate": 6.31735116733666e-05,
|
2885 |
+
"loss": 1.3719,
|
2886 |
+
"step": 411
|
2887 |
+
},
|
2888 |
+
{
|
2889 |
+
"epoch": 0.41944515143802497,
|
2890 |
+
"grad_norm": 5.680314540863037,
|
2891 |
+
"learning_rate": 6.301850571114281e-05,
|
2892 |
+
"loss": 1.173,
|
2893 |
+
"step": 412
|
2894 |
+
},
|
2895 |
+
{
|
2896 |
+
"epoch": 0.4204632221939425,
|
2897 |
+
"grad_norm": 8.60592269897461,
|
2898 |
+
"learning_rate": 6.286336541589224e-05,
|
2899 |
+
"loss": 2.2157,
|
2900 |
+
"step": 413
|
2901 |
+
},
|
2902 |
+
{
|
2903 |
+
"epoch": 0.42148129294986003,
|
2904 |
+
"grad_norm": 6.760054588317871,
|
2905 |
+
"learning_rate": 6.27080923884488e-05,
|
2906 |
+
"loss": 1.6651,
|
2907 |
+
"step": 414
|
2908 |
+
},
|
2909 |
+
{
|
2910 |
+
"epoch": 0.42249936370577756,
|
2911 |
+
"grad_norm": 6.813849925994873,
|
2912 |
+
"learning_rate": 6.255268823101605e-05,
|
2913 |
+
"loss": 1.6109,
|
2914 |
+
"step": 415
|
2915 |
+
},
|
2916 |
+
{
|
2917 |
+
"epoch": 0.4235174344616951,
|
2918 |
+
"grad_norm": 7.729984760284424,
|
2919 |
+
"learning_rate": 6.239715454715054e-05,
|
2920 |
+
"loss": 2.0043,
|
2921 |
+
"step": 416
|
2922 |
+
},
|
2923 |
+
{
|
2924 |
+
"epoch": 0.42453550521761263,
|
2925 |
+
"grad_norm": 6.723374366760254,
|
2926 |
+
"learning_rate": 6.224149294174548e-05,
|
2927 |
+
"loss": 1.7516,
|
2928 |
+
"step": 417
|
2929 |
+
},
|
2930 |
+
{
|
2931 |
+
"epoch": 0.42555357597353016,
|
2932 |
+
"grad_norm": 7.92853307723999,
|
2933 |
+
"learning_rate": 6.208570502101393e-05,
|
2934 |
+
"loss": 2.0667,
|
2935 |
+
"step": 418
|
2936 |
+
},
|
2937 |
+
{
|
2938 |
+
"epoch": 0.4265716467294477,
|
2939 |
+
"grad_norm": 8.561442375183105,
|
2940 |
+
"learning_rate": 6.192979239247243e-05,
|
2941 |
+
"loss": 2.3514,
|
2942 |
+
"step": 419
|
2943 |
+
},
|
2944 |
+
{
|
2945 |
+
"epoch": 0.42758971748536523,
|
2946 |
+
"grad_norm": 6.065650463104248,
|
2947 |
+
"learning_rate": 6.177375666492431e-05,
|
2948 |
+
"loss": 1.6079,
|
2949 |
+
"step": 420
|
2950 |
+
},
|
2951 |
+
{
|
2952 |
+
"epoch": 0.42860778824128276,
|
2953 |
+
"grad_norm": 7.547060489654541,
|
2954 |
+
"learning_rate": 6.161759944844308e-05,
|
2955 |
+
"loss": 1.6165,
|
2956 |
+
"step": 421
|
2957 |
+
},
|
2958 |
+
{
|
2959 |
+
"epoch": 0.4296258589972003,
|
2960 |
+
"grad_norm": 7.750328540802002,
|
2961 |
+
"learning_rate": 6.146132235435591e-05,
|
2962 |
+
"loss": 2.6017,
|
2963 |
+
"step": 422
|
2964 |
+
},
|
2965 |
+
{
|
2966 |
+
"epoch": 0.4306439297531178,
|
2967 |
+
"grad_norm": 8.826605796813965,
|
2968 |
+
"learning_rate": 6.13049269952269e-05,
|
2969 |
+
"loss": 2.8712,
|
2970 |
+
"step": 423
|
2971 |
+
},
|
2972 |
+
{
|
2973 |
+
"epoch": 0.43166200050903536,
|
2974 |
+
"grad_norm": 9.169108390808105,
|
2975 |
+
"learning_rate": 6.114841498484048e-05,
|
2976 |
+
"loss": 3.1703,
|
2977 |
+
"step": 424
|
2978 |
+
},
|
2979 |
+
{
|
2980 |
+
"epoch": 0.4326800712649529,
|
2981 |
+
"grad_norm": 10.452327728271484,
|
2982 |
+
"learning_rate": 6.0991787938184784e-05,
|
2983 |
+
"loss": 2.8737,
|
2984 |
+
"step": 425
|
2985 |
+
},
|
2986 |
+
{
|
2987 |
+
"epoch": 0.4336981420208704,
|
2988 |
+
"grad_norm": 6.921334743499756,
|
2989 |
+
"learning_rate": 6.0835047471434955e-05,
|
2990 |
+
"loss": 2.1521,
|
2991 |
+
"step": 426
|
2992 |
+
},
|
2993 |
+
{
|
2994 |
+
"epoch": 0.434716212776788,
|
2995 |
+
"grad_norm": 7.088666915893555,
|
2996 |
+
"learning_rate": 6.067819520193645e-05,
|
2997 |
+
"loss": 1.833,
|
2998 |
+
"step": 427
|
2999 |
+
},
|
3000 |
+
{
|
3001 |
+
"epoch": 0.43573428353270555,
|
3002 |
+
"grad_norm": 7.354036331176758,
|
3003 |
+
"learning_rate": 6.052123274818842e-05,
|
3004 |
+
"loss": 2.0779,
|
3005 |
+
"step": 428
|
3006 |
+
},
|
3007 |
+
{
|
3008 |
+
"epoch": 0.4367523542886231,
|
3009 |
+
"grad_norm": 6.732209205627441,
|
3010 |
+
"learning_rate": 6.0364161729826905e-05,
|
3011 |
+
"loss": 1.8935,
|
3012 |
+
"step": 429
|
3013 |
+
},
|
3014 |
+
{
|
3015 |
+
"epoch": 0.4377704250445406,
|
3016 |
+
"grad_norm": 7.724286079406738,
|
3017 |
+
"learning_rate": 6.020698376760824e-05,
|
3018 |
+
"loss": 2.334,
|
3019 |
+
"step": 430
|
3020 |
+
},
|
3021 |
+
{
|
3022 |
+
"epoch": 0.43878849580045814,
|
3023 |
+
"grad_norm": 7.147849082946777,
|
3024 |
+
"learning_rate": 6.004970048339226e-05,
|
3025 |
+
"loss": 1.9002,
|
3026 |
+
"step": 431
|
3027 |
+
},
|
3028 |
+
{
|
3029 |
+
"epoch": 0.4398065665563757,
|
3030 |
+
"grad_norm": 10.327398300170898,
|
3031 |
+
"learning_rate": 5.989231350012554e-05,
|
3032 |
+
"loss": 3.3501,
|
3033 |
+
"step": 432
|
3034 |
+
},
|
3035 |
+
{
|
3036 |
+
"epoch": 0.4408246373122932,
|
3037 |
+
"grad_norm": 8.666500091552734,
|
3038 |
+
"learning_rate": 5.973482444182475e-05,
|
3039 |
+
"loss": 2.4769,
|
3040 |
+
"step": 433
|
3041 |
+
},
|
3042 |
+
{
|
3043 |
+
"epoch": 0.44184270806821074,
|
3044 |
+
"grad_norm": 10.457250595092773,
|
3045 |
+
"learning_rate": 5.9577234933559764e-05,
|
3046 |
+
"loss": 3.03,
|
3047 |
+
"step": 434
|
3048 |
+
},
|
3049 |
+
{
|
3050 |
+
"epoch": 0.4428607788241283,
|
3051 |
+
"grad_norm": 11.054841995239258,
|
3052 |
+
"learning_rate": 5.941954660143703e-05,
|
3053 |
+
"loss": 2.6938,
|
3054 |
+
"step": 435
|
3055 |
+
},
|
3056 |
+
{
|
3057 |
+
"epoch": 0.4438788495800458,
|
3058 |
+
"grad_norm": 8.820106506347656,
|
3059 |
+
"learning_rate": 5.9261761072582655e-05,
|
3060 |
+
"loss": 2.6282,
|
3061 |
+
"step": 436
|
3062 |
+
},
|
3063 |
+
{
|
3064 |
+
"epoch": 0.44489692033596334,
|
3065 |
+
"grad_norm": 9.201048851013184,
|
3066 |
+
"learning_rate": 5.910387997512573e-05,
|
3067 |
+
"loss": 3.0258,
|
3068 |
+
"step": 437
|
3069 |
+
},
|
3070 |
+
{
|
3071 |
+
"epoch": 0.4459149910918809,
|
3072 |
+
"grad_norm": 8.373756408691406,
|
3073 |
+
"learning_rate": 5.8945904938181484e-05,
|
3074 |
+
"loss": 2.0512,
|
3075 |
+
"step": 438
|
3076 |
+
},
|
3077 |
+
{
|
3078 |
+
"epoch": 0.4469330618477984,
|
3079 |
+
"grad_norm": 7.493674278259277,
|
3080 |
+
"learning_rate": 5.878783759183442e-05,
|
3081 |
+
"loss": 2.3767,
|
3082 |
+
"step": 439
|
3083 |
+
},
|
3084 |
+
{
|
3085 |
+
"epoch": 0.44795113260371594,
|
3086 |
+
"grad_norm": 8.552105903625488,
|
3087 |
+
"learning_rate": 5.86296795671216e-05,
|
3088 |
+
"loss": 2.4706,
|
3089 |
+
"step": 440
|
3090 |
+
},
|
3091 |
+
{
|
3092 |
+
"epoch": 0.4489692033596335,
|
3093 |
+
"grad_norm": 7.708653450012207,
|
3094 |
+
"learning_rate": 5.847143249601574e-05,
|
3095 |
+
"loss": 2.2514,
|
3096 |
+
"step": 441
|
3097 |
+
},
|
3098 |
+
{
|
3099 |
+
"epoch": 0.449987274115551,
|
3100 |
+
"grad_norm": 9.070602416992188,
|
3101 |
+
"learning_rate": 5.8313098011408406e-05,
|
3102 |
+
"loss": 2.322,
|
3103 |
+
"step": 442
|
3104 |
+
},
|
3105 |
+
{
|
3106 |
+
"epoch": 0.4510053448714686,
|
3107 |
+
"grad_norm": 7.668588638305664,
|
3108 |
+
"learning_rate": 5.8154677747093134e-05,
|
3109 |
+
"loss": 1.8555,
|
3110 |
+
"step": 443
|
3111 |
+
},
|
3112 |
+
{
|
3113 |
+
"epoch": 0.4520234156273861,
|
3114 |
+
"grad_norm": 9.315807342529297,
|
3115 |
+
"learning_rate": 5.7996173337748606e-05,
|
3116 |
+
"loss": 2.1384,
|
3117 |
+
"step": 444
|
3118 |
+
},
|
3119 |
+
{
|
3120 |
+
"epoch": 0.45304148638330366,
|
3121 |
+
"grad_norm": 10.211068153381348,
|
3122 |
+
"learning_rate": 5.783758641892172e-05,
|
3123 |
+
"loss": 2.8774,
|
3124 |
+
"step": 445
|
3125 |
+
},
|
3126 |
+
{
|
3127 |
+
"epoch": 0.4540595571392212,
|
3128 |
+
"grad_norm": 9.461262702941895,
|
3129 |
+
"learning_rate": 5.767891862701082e-05,
|
3130 |
+
"loss": 2.5156,
|
3131 |
+
"step": 446
|
3132 |
+
},
|
3133 |
+
{
|
3134 |
+
"epoch": 0.4550776278951387,
|
3135 |
+
"grad_norm": 9.720065116882324,
|
3136 |
+
"learning_rate": 5.7520171599248704e-05,
|
3137 |
+
"loss": 2.6157,
|
3138 |
+
"step": 447
|
3139 |
+
},
|
3140 |
+
{
|
3141 |
+
"epoch": 0.45609569865105626,
|
3142 |
+
"grad_norm": 8.700965881347656,
|
3143 |
+
"learning_rate": 5.7361346973685794e-05,
|
3144 |
+
"loss": 2.2904,
|
3145 |
+
"step": 448
|
3146 |
+
},
|
3147 |
+
{
|
3148 |
+
"epoch": 0.4571137694069738,
|
3149 |
+
"grad_norm": 7.846927642822266,
|
3150 |
+
"learning_rate": 5.7202446389173223e-05,
|
3151 |
+
"loss": 1.968,
|
3152 |
+
"step": 449
|
3153 |
+
},
|
3154 |
+
{
|
3155 |
+
"epoch": 0.4581318401628913,
|
3156 |
+
"grad_norm": 11.256888389587402,
|
3157 |
+
"learning_rate": 5.704347148534589e-05,
|
3158 |
+
"loss": 2.0931,
|
3159 |
+
"step": 450
|
3160 |
+
},
|
3161 |
+
{
|
3162 |
+
"epoch": 0.45914991091880886,
|
3163 |
+
"grad_norm": 8.149317741394043,
|
3164 |
+
"learning_rate": 5.688442390260559e-05,
|
3165 |
+
"loss": 3.8319,
|
3166 |
+
"step": 451
|
3167 |
+
},
|
3168 |
+
{
|
3169 |
+
"epoch": 0.4601679816747264,
|
3170 |
+
"grad_norm": 10.932345390319824,
|
3171 |
+
"learning_rate": 5.672530528210405e-05,
|
3172 |
+
"loss": 3.0198,
|
3173 |
+
"step": 452
|
3174 |
+
},
|
3175 |
+
{
|
3176 |
+
"epoch": 0.4611860524306439,
|
3177 |
+
"grad_norm": 9.099678993225098,
|
3178 |
+
"learning_rate": 5.6566117265726006e-05,
|
3179 |
+
"loss": 3.3294,
|
3180 |
+
"step": 453
|
3181 |
+
},
|
3182 |
+
{
|
3183 |
+
"epoch": 0.46220412318656146,
|
3184 |
+
"grad_norm": 11.72851848602295,
|
3185 |
+
"learning_rate": 5.640686149607228e-05,
|
3186 |
+
"loss": 4.2732,
|
3187 |
+
"step": 454
|
3188 |
+
},
|
3189 |
+
{
|
3190 |
+
"epoch": 0.463222193942479,
|
3191 |
+
"grad_norm": 11.902469635009766,
|
3192 |
+
"learning_rate": 5.624753961644281e-05,
|
3193 |
+
"loss": 3.6108,
|
3194 |
+
"step": 455
|
3195 |
+
},
|
3196 |
+
{
|
3197 |
+
"epoch": 0.4642402646983965,
|
3198 |
+
"grad_norm": 10.879993438720703,
|
3199 |
+
"learning_rate": 5.608815327081969e-05,
|
3200 |
+
"loss": 2.6874,
|
3201 |
+
"step": 456
|
3202 |
+
},
|
3203 |
+
{
|
3204 |
+
"epoch": 0.46525833545431405,
|
3205 |
+
"grad_norm": 14.943718910217285,
|
3206 |
+
"learning_rate": 5.5928704103850206e-05,
|
3207 |
+
"loss": 4.6303,
|
3208 |
+
"step": 457
|
3209 |
+
},
|
3210 |
+
{
|
3211 |
+
"epoch": 0.4662764062102316,
|
3212 |
+
"grad_norm": 10.216432571411133,
|
3213 |
+
"learning_rate": 5.57691937608299e-05,
|
3214 |
+
"loss": 2.4846,
|
3215 |
+
"step": 458
|
3216 |
+
},
|
3217 |
+
{
|
3218 |
+
"epoch": 0.4672944769661492,
|
3219 |
+
"grad_norm": 8.506292343139648,
|
3220 |
+
"learning_rate": 5.5609623887685535e-05,
|
3221 |
+
"loss": 2.476,
|
3222 |
+
"step": 459
|
3223 |
+
},
|
3224 |
+
{
|
3225 |
+
"epoch": 0.4683125477220667,
|
3226 |
+
"grad_norm": 8.514628410339355,
|
3227 |
+
"learning_rate": 5.544999613095818e-05,
|
3228 |
+
"loss": 2.377,
|
3229 |
+
"step": 460
|
3230 |
+
},
|
3231 |
+
{
|
3232 |
+
"epoch": 0.46933061847798424,
|
3233 |
+
"grad_norm": 5.901019096374512,
|
3234 |
+
"learning_rate": 5.5290312137786146e-05,
|
3235 |
+
"loss": 1.5461,
|
3236 |
+
"step": 461
|
3237 |
+
},
|
3238 |
+
{
|
3239 |
+
"epoch": 0.4703486892339018,
|
3240 |
+
"grad_norm": 5.75610876083374,
|
3241 |
+
"learning_rate": 5.513057355588804e-05,
|
3242 |
+
"loss": 1.4872,
|
3243 |
+
"step": 462
|
3244 |
+
},
|
3245 |
+
{
|
3246 |
+
"epoch": 0.4713667599898193,
|
3247 |
+
"grad_norm": 5.7381272315979,
|
3248 |
+
"learning_rate": 5.4970782033545774e-05,
|
3249 |
+
"loss": 1.4357,
|
3250 |
+
"step": 463
|
3251 |
+
},
|
3252 |
+
{
|
3253 |
+
"epoch": 0.47238483074573684,
|
3254 |
+
"grad_norm": 7.901957035064697,
|
3255 |
+
"learning_rate": 5.4810939219587485e-05,
|
3256 |
+
"loss": 1.8938,
|
3257 |
+
"step": 464
|
3258 |
+
},
|
3259 |
+
{
|
3260 |
+
"epoch": 0.47340290150165437,
|
3261 |
+
"grad_norm": 5.663755893707275,
|
3262 |
+
"learning_rate": 5.465104676337062e-05,
|
3263 |
+
"loss": 1.2289,
|
3264 |
+
"step": 465
|
3265 |
+
},
|
3266 |
+
{
|
3267 |
+
"epoch": 0.4744209722575719,
|
3268 |
+
"grad_norm": 7.933018684387207,
|
3269 |
+
"learning_rate": 5.44911063147648e-05,
|
3270 |
+
"loss": 2.2704,
|
3271 |
+
"step": 466
|
3272 |
+
},
|
3273 |
+
{
|
3274 |
+
"epoch": 0.47543904301348944,
|
3275 |
+
"grad_norm": 7.943240165710449,
|
3276 |
+
"learning_rate": 5.433111952413495e-05,
|
3277 |
+
"loss": 2.3279,
|
3278 |
+
"step": 467
|
3279 |
+
},
|
3280 |
+
{
|
3281 |
+
"epoch": 0.47645711376940697,
|
3282 |
+
"grad_norm": 6.650790691375732,
|
3283 |
+
"learning_rate": 5.417108804232409e-05,
|
3284 |
+
"loss": 2.2172,
|
3285 |
+
"step": 468
|
3286 |
+
},
|
3287 |
+
{
|
3288 |
+
"epoch": 0.4774751845253245,
|
3289 |
+
"grad_norm": 7.444812774658203,
|
3290 |
+
"learning_rate": 5.401101352063647e-05,
|
3291 |
+
"loss": 2.2142,
|
3292 |
+
"step": 469
|
3293 |
+
},
|
3294 |
+
{
|
3295 |
+
"epoch": 0.47849325528124204,
|
3296 |
+
"grad_norm": 7.4753642082214355,
|
3297 |
+
"learning_rate": 5.3850897610820396e-05,
|
3298 |
+
"loss": 1.757,
|
3299 |
+
"step": 470
|
3300 |
+
},
|
3301 |
+
{
|
3302 |
+
"epoch": 0.47951132603715957,
|
3303 |
+
"grad_norm": 5.919788360595703,
|
3304 |
+
"learning_rate": 5.369074196505125e-05,
|
3305 |
+
"loss": 1.6481,
|
3306 |
+
"step": 471
|
3307 |
+
},
|
3308 |
+
{
|
3309 |
+
"epoch": 0.4805293967930771,
|
3310 |
+
"grad_norm": 8.422210693359375,
|
3311 |
+
"learning_rate": 5.3530548235914454e-05,
|
3312 |
+
"loss": 2.157,
|
3313 |
+
"step": 472
|
3314 |
+
},
|
3315 |
+
{
|
3316 |
+
"epoch": 0.48154746754899463,
|
3317 |
+
"grad_norm": 6.410617351531982,
|
3318 |
+
"learning_rate": 5.33703180763884e-05,
|
3319 |
+
"loss": 1.8347,
|
3320 |
+
"step": 473
|
3321 |
+
},
|
3322 |
+
{
|
3323 |
+
"epoch": 0.48256553830491217,
|
3324 |
+
"grad_norm": 6.645679473876953,
|
3325 |
+
"learning_rate": 5.321005313982738e-05,
|
3326 |
+
"loss": 1.9199,
|
3327 |
+
"step": 474
|
3328 |
+
},
|
3329 |
+
{
|
3330 |
+
"epoch": 0.4835836090608297,
|
3331 |
+
"grad_norm": 8.815511703491211,
|
3332 |
+
"learning_rate": 5.3049755079944527e-05,
|
3333 |
+
"loss": 1.9625,
|
3334 |
+
"step": 475
|
3335 |
+
},
|
3336 |
+
{
|
3337 |
+
"epoch": 0.4846016798167473,
|
3338 |
+
"grad_norm": 6.998859882354736,
|
3339 |
+
"learning_rate": 5.288942555079479e-05,
|
3340 |
+
"loss": 1.9826,
|
3341 |
+
"step": 476
|
3342 |
+
},
|
3343 |
+
{
|
3344 |
+
"epoch": 0.4856197505726648,
|
3345 |
+
"grad_norm": 9.812437057495117,
|
3346 |
+
"learning_rate": 5.272906620675779e-05,
|
3347 |
+
"loss": 2.8627,
|
3348 |
+
"step": 477
|
3349 |
+
},
|
3350 |
+
{
|
3351 |
+
"epoch": 0.48663782132858235,
|
3352 |
+
"grad_norm": 6.514684677124023,
|
3353 |
+
"learning_rate": 5.256867870252087e-05,
|
3354 |
+
"loss": 1.8711,
|
3355 |
+
"step": 478
|
3356 |
+
},
|
3357 |
+
{
|
3358 |
+
"epoch": 0.4876558920844999,
|
3359 |
+
"grad_norm": 8.101947784423828,
|
3360 |
+
"learning_rate": 5.240826469306187e-05,
|
3361 |
+
"loss": 2.2444,
|
3362 |
+
"step": 479
|
3363 |
+
},
|
3364 |
+
{
|
3365 |
+
"epoch": 0.4886739628404174,
|
3366 |
+
"grad_norm": 7.826203346252441,
|
3367 |
+
"learning_rate": 5.224782583363215e-05,
|
3368 |
+
"loss": 2.3479,
|
3369 |
+
"step": 480
|
3370 |
+
},
|
3371 |
+
{
|
3372 |
+
"epoch": 0.48969203359633495,
|
3373 |
+
"grad_norm": 7.306244850158691,
|
3374 |
+
"learning_rate": 5.208736377973954e-05,
|
3375 |
+
"loss": 2.2877,
|
3376 |
+
"step": 481
|
3377 |
+
},
|
3378 |
+
{
|
3379 |
+
"epoch": 0.4907101043522525,
|
3380 |
+
"grad_norm": 10.923294067382812,
|
3381 |
+
"learning_rate": 5.192688018713113e-05,
|
3382 |
+
"loss": 3.4528,
|
3383 |
+
"step": 482
|
3384 |
+
},
|
3385 |
+
{
|
3386 |
+
"epoch": 0.49172817510817,
|
3387 |
+
"grad_norm": 7.382264614105225,
|
3388 |
+
"learning_rate": 5.176637671177631e-05,
|
3389 |
+
"loss": 2.1969,
|
3390 |
+
"step": 483
|
3391 |
+
},
|
3392 |
+
{
|
3393 |
+
"epoch": 0.49274624586408755,
|
3394 |
+
"grad_norm": 10.03930377960205,
|
3395 |
+
"learning_rate": 5.1605855009849614e-05,
|
3396 |
+
"loss": 3.5883,
|
3397 |
+
"step": 484
|
3398 |
+
},
|
3399 |
+
{
|
3400 |
+
"epoch": 0.4937643166200051,
|
3401 |
+
"grad_norm": 7.927920818328857,
|
3402 |
+
"learning_rate": 5.144531673771363e-05,
|
3403 |
+
"loss": 2.3655,
|
3404 |
+
"step": 485
|
3405 |
+
},
|
3406 |
+
{
|
3407 |
+
"epoch": 0.4947823873759226,
|
3408 |
+
"grad_norm": 8.499062538146973,
|
3409 |
+
"learning_rate": 5.1284763551901995e-05,
|
3410 |
+
"loss": 2.5173,
|
3411 |
+
"step": 486
|
3412 |
+
},
|
3413 |
+
{
|
3414 |
+
"epoch": 0.49580045813184015,
|
3415 |
+
"grad_norm": 6.909058094024658,
|
3416 |
+
"learning_rate": 5.112419710910213e-05,
|
3417 |
+
"loss": 2.0323,
|
3418 |
+
"step": 487
|
3419 |
+
},
|
3420 |
+
{
|
3421 |
+
"epoch": 0.4968185288877577,
|
3422 |
+
"grad_norm": 8.88456916809082,
|
3423 |
+
"learning_rate": 5.096361906613836e-05,
|
3424 |
+
"loss": 2.6987,
|
3425 |
+
"step": 488
|
3426 |
+
},
|
3427 |
+
{
|
3428 |
+
"epoch": 0.4978365996436752,
|
3429 |
+
"grad_norm": 7.5175580978393555,
|
3430 |
+
"learning_rate": 5.080303107995461e-05,
|
3431 |
+
"loss": 2.1691,
|
3432 |
+
"step": 489
|
3433 |
+
},
|
3434 |
+
{
|
3435 |
+
"epoch": 0.49885467039959275,
|
3436 |
+
"grad_norm": 7.615448951721191,
|
3437 |
+
"learning_rate": 5.064243480759748e-05,
|
3438 |
+
"loss": 2.0365,
|
3439 |
+
"step": 490
|
3440 |
+
},
|
3441 |
+
{
|
3442 |
+
"epoch": 0.4998727411555103,
|
3443 |
+
"grad_norm": 9.251805305480957,
|
3444 |
+
"learning_rate": 5.048183190619904e-05,
|
3445 |
+
"loss": 2.2146,
|
3446 |
+
"step": 491
|
3447 |
+
},
|
3448 |
+
{
|
3449 |
+
"epoch": 0.5008908119114278,
|
3450 |
+
"grad_norm": 7.876194477081299,
|
3451 |
+
"learning_rate": 5.032122403295977e-05,
|
3452 |
+
"loss": 2.3439,
|
3453 |
+
"step": 492
|
3454 |
}
|
3455 |
],
|
3456 |
"logging_steps": 1,
|
|
|
3470 |
"attributes": {}
|
3471 |
}
|
3472 |
},
|
3473 |
+
"total_flos": 4.032897809991598e+17,
|
3474 |
"train_batch_size": 4,
|
3475 |
"trial_name": null,
|
3476 |
"trial_params": null
|