Upload folder using huggingface_hub
Browse files- adapter_model.safetensors +1 -1
- optimizer.pt +1 -1
- rng_state.pth +1 -1
- scheduler.pt +1 -1
- trainer_state.json +1053 -3
adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 3705288
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:afbb9fafdaa3ca57948b5758a6f7f91aff0d8bb88753df2cf01d7363ea7ed873
|
| 3 |
size 3705288
|
optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 2213690
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4ef832f29c75c0a4f145b612ac65d501410d3c9e4ad8c11582372091e773bf49
|
| 3 |
size 2213690
|
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 sha256:7fca791d6b819bad1dbc285bf8bd7345964c2aeaf16b7d702b5c0c5380f2a057
|
| 3 |
size 14244
|
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 sha256:6b04bc150d39b084b09acd98d05c3563c323fc35277059ff5584f9f2d3fde608
|
| 3 |
size 1064
|
trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch":
|
| 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,
|
|
@@ -2107,6 +2107,1056 @@
|
|
| 2107 |
"learning_rate": 0.00014187192118226603,
|
| 2108 |
"loss": 0.444,
|
| 2109 |
"step": 6000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2110 |
}
|
| 2111 |
],
|
| 2112 |
"logging_steps": 20,
|
|
@@ -2126,7 +3176,7 @@
|
|
| 2126 |
"attributes": {}
|
| 2127 |
}
|
| 2128 |
},
|
| 2129 |
-
"total_flos": 1.
|
| 2130 |
"train_batch_size": 1,
|
| 2131 |
"trial_name": null,
|
| 2132 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 43.92922513727883,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 9000,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 2107 |
"learning_rate": 0.00014187192118226603,
|
| 2108 |
"loss": 0.444,
|
| 2109 |
"step": 6000
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 29.383770591824284,
|
| 2113 |
+
"grad_norm": 4.88236141204834,
|
| 2114 |
+
"learning_rate": 0.00014167487684729063,
|
| 2115 |
+
"loss": 0.4639,
|
| 2116 |
+
"step": 6020
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 29.481391092129346,
|
| 2120 |
+
"grad_norm": 3.7870137691497803,
|
| 2121 |
+
"learning_rate": 0.00014147783251231528,
|
| 2122 |
+
"loss": 0.4873,
|
| 2123 |
+
"step": 6040
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 29.579011592434412,
|
| 2127 |
+
"grad_norm": 3.091411590576172,
|
| 2128 |
+
"learning_rate": 0.0001412807881773399,
|
| 2129 |
+
"loss": 0.4834,
|
| 2130 |
+
"step": 6060
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 29.676632092739474,
|
| 2134 |
+
"grad_norm": 2.7498538494110107,
|
| 2135 |
+
"learning_rate": 0.00014108374384236454,
|
| 2136 |
+
"loss": 0.4846,
|
| 2137 |
+
"step": 6080
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 29.77425259304454,
|
| 2141 |
+
"grad_norm": 3.2043850421905518,
|
| 2142 |
+
"learning_rate": 0.00014088669950738917,
|
| 2143 |
+
"loss": 0.4983,
|
| 2144 |
+
"step": 6100
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 29.871873093349603,
|
| 2148 |
+
"grad_norm": 3.270357847213745,
|
| 2149 |
+
"learning_rate": 0.0001406896551724138,
|
| 2150 |
+
"loss": 0.4803,
|
| 2151 |
+
"step": 6120
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 29.96949359365467,
|
| 2155 |
+
"grad_norm": 3.031405210494995,
|
| 2156 |
+
"learning_rate": 0.00014049261083743842,
|
| 2157 |
+
"loss": 0.5287,
|
| 2158 |
+
"step": 6140
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 30.06711409395973,
|
| 2162 |
+
"grad_norm": 3.390765905380249,
|
| 2163 |
+
"learning_rate": 0.00014029556650246307,
|
| 2164 |
+
"loss": 0.4619,
|
| 2165 |
+
"step": 6160
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 30.164734594264797,
|
| 2169 |
+
"grad_norm": 3.2783963680267334,
|
| 2170 |
+
"learning_rate": 0.0001400985221674877,
|
| 2171 |
+
"loss": 0.4328,
|
| 2172 |
+
"step": 6180
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 30.26235509456986,
|
| 2176 |
+
"grad_norm": 3.6925759315490723,
|
| 2177 |
+
"learning_rate": 0.00013990147783251233,
|
| 2178 |
+
"loss": 0.487,
|
| 2179 |
+
"step": 6200
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 30.359975594874925,
|
| 2183 |
+
"grad_norm": 3.0115065574645996,
|
| 2184 |
+
"learning_rate": 0.00013970443349753696,
|
| 2185 |
+
"loss": 0.467,
|
| 2186 |
+
"step": 6220
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 30.457596095179987,
|
| 2190 |
+
"grad_norm": 4.561310291290283,
|
| 2191 |
+
"learning_rate": 0.00013950738916256158,
|
| 2192 |
+
"loss": 0.4801,
|
| 2193 |
+
"step": 6240
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 30.555216595485053,
|
| 2197 |
+
"grad_norm": 3.2879674434661865,
|
| 2198 |
+
"learning_rate": 0.0001393103448275862,
|
| 2199 |
+
"loss": 0.4638,
|
| 2200 |
+
"step": 6260
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 30.652837095790115,
|
| 2204 |
+
"grad_norm": 2.793945789337158,
|
| 2205 |
+
"learning_rate": 0.00013911330049261084,
|
| 2206 |
+
"loss": 0.463,
|
| 2207 |
+
"step": 6280
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 30.75045759609518,
|
| 2211 |
+
"grad_norm": 3.615793466567993,
|
| 2212 |
+
"learning_rate": 0.0001389162561576355,
|
| 2213 |
+
"loss": 0.4907,
|
| 2214 |
+
"step": 6300
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 30.848078096400243,
|
| 2218 |
+
"grad_norm": 3.160133123397827,
|
| 2219 |
+
"learning_rate": 0.0001387192118226601,
|
| 2220 |
+
"loss": 0.477,
|
| 2221 |
+
"step": 6320
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 30.94569859670531,
|
| 2225 |
+
"grad_norm": 3.62670636177063,
|
| 2226 |
+
"learning_rate": 0.00013852216748768475,
|
| 2227 |
+
"loss": 0.4945,
|
| 2228 |
+
"step": 6340
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 31.04331909701037,
|
| 2232 |
+
"grad_norm": 3.346158981323242,
|
| 2233 |
+
"learning_rate": 0.00013832512315270935,
|
| 2234 |
+
"loss": 0.4543,
|
| 2235 |
+
"step": 6360
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 31.140939597315437,
|
| 2239 |
+
"grad_norm": 2.8707423210144043,
|
| 2240 |
+
"learning_rate": 0.000138128078817734,
|
| 2241 |
+
"loss": 0.4352,
|
| 2242 |
+
"step": 6380
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 31.2385600976205,
|
| 2246 |
+
"grad_norm": 2.5617620944976807,
|
| 2247 |
+
"learning_rate": 0.00013793103448275863,
|
| 2248 |
+
"loss": 0.4611,
|
| 2249 |
+
"step": 6400
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 31.336180597925566,
|
| 2253 |
+
"grad_norm": 3.2273828983306885,
|
| 2254 |
+
"learning_rate": 0.00013773399014778325,
|
| 2255 |
+
"loss": 0.4593,
|
| 2256 |
+
"step": 6420
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 31.433801098230628,
|
| 2260 |
+
"grad_norm": 3.502797842025757,
|
| 2261 |
+
"learning_rate": 0.00013753694581280788,
|
| 2262 |
+
"loss": 0.4717,
|
| 2263 |
+
"step": 6440
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 31.531421598535694,
|
| 2267 |
+
"grad_norm": 3.9278218746185303,
|
| 2268 |
+
"learning_rate": 0.0001373399014778325,
|
| 2269 |
+
"loss": 0.4813,
|
| 2270 |
+
"step": 6460
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 31.629042098840756,
|
| 2274 |
+
"grad_norm": 3.013709545135498,
|
| 2275 |
+
"learning_rate": 0.00013714285714285716,
|
| 2276 |
+
"loss": 0.4305,
|
| 2277 |
+
"step": 6480
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 31.726662599145822,
|
| 2281 |
+
"grad_norm": 2.661198377609253,
|
| 2282 |
+
"learning_rate": 0.0001369458128078818,
|
| 2283 |
+
"loss": 0.4495,
|
| 2284 |
+
"step": 6500
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 31.824283099450884,
|
| 2288 |
+
"grad_norm": 2.6343297958374023,
|
| 2289 |
+
"learning_rate": 0.00013674876847290642,
|
| 2290 |
+
"loss": 0.4809,
|
| 2291 |
+
"step": 6520
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 31.92190359975595,
|
| 2295 |
+
"grad_norm": 6.334170818328857,
|
| 2296 |
+
"learning_rate": 0.00013655172413793104,
|
| 2297 |
+
"loss": 0.4576,
|
| 2298 |
+
"step": 6540
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 32.01952410006101,
|
| 2302 |
+
"grad_norm": 3.728727102279663,
|
| 2303 |
+
"learning_rate": 0.00013635467980295567,
|
| 2304 |
+
"loss": 0.5034,
|
| 2305 |
+
"step": 6560
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 32.117144600366075,
|
| 2309 |
+
"grad_norm": 2.0572702884674072,
|
| 2310 |
+
"learning_rate": 0.0001361576354679803,
|
| 2311 |
+
"loss": 0.4161,
|
| 2312 |
+
"step": 6580
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 32.214765100671144,
|
| 2316 |
+
"grad_norm": 2.7006356716156006,
|
| 2317 |
+
"learning_rate": 0.00013596059113300492,
|
| 2318 |
+
"loss": 0.4357,
|
| 2319 |
+
"step": 6600
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 32.31238560097621,
|
| 2323 |
+
"grad_norm": 3.526782989501953,
|
| 2324 |
+
"learning_rate": 0.00013576354679802955,
|
| 2325 |
+
"loss": 0.4367,
|
| 2326 |
+
"step": 6620
|
| 2327 |
+
},
|
| 2328 |
+
{
|
| 2329 |
+
"epoch": 32.41000610128127,
|
| 2330 |
+
"grad_norm": 3.240647792816162,
|
| 2331 |
+
"learning_rate": 0.0001355665024630542,
|
| 2332 |
+
"loss": 0.4416,
|
| 2333 |
+
"step": 6640
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 32.50762660158633,
|
| 2337 |
+
"grad_norm": 2.965851306915283,
|
| 2338 |
+
"learning_rate": 0.0001353694581280788,
|
| 2339 |
+
"loss": 0.4649,
|
| 2340 |
+
"step": 6660
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 32.6052471018914,
|
| 2344 |
+
"grad_norm": 3.028812885284424,
|
| 2345 |
+
"learning_rate": 0.00013517241379310346,
|
| 2346 |
+
"loss": 0.4381,
|
| 2347 |
+
"step": 6680
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 32.70286760219646,
|
| 2351 |
+
"grad_norm": 4.041370391845703,
|
| 2352 |
+
"learning_rate": 0.0001349753694581281,
|
| 2353 |
+
"loss": 0.4671,
|
| 2354 |
+
"step": 6700
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 32.800488102501525,
|
| 2358 |
+
"grad_norm": 5.677656650543213,
|
| 2359 |
+
"learning_rate": 0.00013477832512315271,
|
| 2360 |
+
"loss": 0.4718,
|
| 2361 |
+
"step": 6720
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 32.89810860280659,
|
| 2365 |
+
"grad_norm": 3.1538727283477783,
|
| 2366 |
+
"learning_rate": 0.00013458128078817737,
|
| 2367 |
+
"loss": 0.4705,
|
| 2368 |
+
"step": 6740
|
| 2369 |
+
},
|
| 2370 |
+
{
|
| 2371 |
+
"epoch": 32.99572910311166,
|
| 2372 |
+
"grad_norm": 3.8186867237091064,
|
| 2373 |
+
"learning_rate": 0.00013438423645320197,
|
| 2374 |
+
"loss": 0.4724,
|
| 2375 |
+
"step": 6760
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 33.09334960341672,
|
| 2379 |
+
"grad_norm": 2.8248584270477295,
|
| 2380 |
+
"learning_rate": 0.00013418719211822662,
|
| 2381 |
+
"loss": 0.4399,
|
| 2382 |
+
"step": 6780
|
| 2383 |
+
},
|
| 2384 |
+
{
|
| 2385 |
+
"epoch": 33.19097010372178,
|
| 2386 |
+
"grad_norm": 2.2694895267486572,
|
| 2387 |
+
"learning_rate": 0.00013399014778325122,
|
| 2388 |
+
"loss": 0.4147,
|
| 2389 |
+
"step": 6800
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 33.288590604026844,
|
| 2393 |
+
"grad_norm": 3.305610418319702,
|
| 2394 |
+
"learning_rate": 0.00013379310344827588,
|
| 2395 |
+
"loss": 0.4028,
|
| 2396 |
+
"step": 6820
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 33.38621110433191,
|
| 2400 |
+
"grad_norm": 3.610136032104492,
|
| 2401 |
+
"learning_rate": 0.0001335960591133005,
|
| 2402 |
+
"loss": 0.4319,
|
| 2403 |
+
"step": 6840
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 33.483831604636975,
|
| 2407 |
+
"grad_norm": 3.4783689975738525,
|
| 2408 |
+
"learning_rate": 0.00013339901477832513,
|
| 2409 |
+
"loss": 0.4361,
|
| 2410 |
+
"step": 6860
|
| 2411 |
+
},
|
| 2412 |
+
{
|
| 2413 |
+
"epoch": 33.58145210494204,
|
| 2414 |
+
"grad_norm": 3.0984203815460205,
|
| 2415 |
+
"learning_rate": 0.00013320197044334976,
|
| 2416 |
+
"loss": 0.4488,
|
| 2417 |
+
"step": 6880
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 33.6790726052471,
|
| 2421 |
+
"grad_norm": 3.1558122634887695,
|
| 2422 |
+
"learning_rate": 0.00013300492610837438,
|
| 2423 |
+
"loss": 0.4262,
|
| 2424 |
+
"step": 6900
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 33.77669310555217,
|
| 2428 |
+
"grad_norm": 4.813379764556885,
|
| 2429 |
+
"learning_rate": 0.000132807881773399,
|
| 2430 |
+
"loss": 0.452,
|
| 2431 |
+
"step": 6920
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 33.87431360585723,
|
| 2435 |
+
"grad_norm": 3.047551393508911,
|
| 2436 |
+
"learning_rate": 0.00013261083743842364,
|
| 2437 |
+
"loss": 0.4517,
|
| 2438 |
+
"step": 6940
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 33.971934106162294,
|
| 2442 |
+
"grad_norm": 3.0880701541900635,
|
| 2443 |
+
"learning_rate": 0.0001324137931034483,
|
| 2444 |
+
"loss": 0.5147,
|
| 2445 |
+
"step": 6960
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 34.06955460646736,
|
| 2449 |
+
"grad_norm": 2.824169874191284,
|
| 2450 |
+
"learning_rate": 0.00013221674876847292,
|
| 2451 |
+
"loss": 0.4017,
|
| 2452 |
+
"step": 6980
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 34.16717510677242,
|
| 2456 |
+
"grad_norm": 3.1136012077331543,
|
| 2457 |
+
"learning_rate": 0.00013201970443349755,
|
| 2458 |
+
"loss": 0.4291,
|
| 2459 |
+
"step": 7000
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 34.26479560707749,
|
| 2463 |
+
"grad_norm": 4.246958255767822,
|
| 2464 |
+
"learning_rate": 0.00013182266009852217,
|
| 2465 |
+
"loss": 0.4318,
|
| 2466 |
+
"step": 7020
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 34.36241610738255,
|
| 2470 |
+
"grad_norm": 2.4655661582946777,
|
| 2471 |
+
"learning_rate": 0.0001316256157635468,
|
| 2472 |
+
"loss": 0.4283,
|
| 2473 |
+
"step": 7040
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 34.46003660768761,
|
| 2477 |
+
"grad_norm": 4.322596549987793,
|
| 2478 |
+
"learning_rate": 0.00013142857142857143,
|
| 2479 |
+
"loss": 0.4323,
|
| 2480 |
+
"step": 7060
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 34.557657107992675,
|
| 2484 |
+
"grad_norm": 4.425800800323486,
|
| 2485 |
+
"learning_rate": 0.00013123152709359608,
|
| 2486 |
+
"loss": 0.4376,
|
| 2487 |
+
"step": 7080
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 34.655277608297745,
|
| 2491 |
+
"grad_norm": 3.796889305114746,
|
| 2492 |
+
"learning_rate": 0.00013103448275862068,
|
| 2493 |
+
"loss": 0.4276,
|
| 2494 |
+
"step": 7100
|
| 2495 |
+
},
|
| 2496 |
+
{
|
| 2497 |
+
"epoch": 34.75289810860281,
|
| 2498 |
+
"grad_norm": 3.9222586154937744,
|
| 2499 |
+
"learning_rate": 0.00013083743842364534,
|
| 2500 |
+
"loss": 0.4658,
|
| 2501 |
+
"step": 7120
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 34.85051860890787,
|
| 2505 |
+
"grad_norm": 4.5007548332214355,
|
| 2506 |
+
"learning_rate": 0.00013064039408866994,
|
| 2507 |
+
"loss": 0.4293,
|
| 2508 |
+
"step": 7140
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 34.94813910921293,
|
| 2512 |
+
"grad_norm": 3.0858423709869385,
|
| 2513 |
+
"learning_rate": 0.0001304433497536946,
|
| 2514 |
+
"loss": 0.4214,
|
| 2515 |
+
"step": 7160
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 35.045759609518,
|
| 2519 |
+
"grad_norm": 3.586949586868286,
|
| 2520 |
+
"learning_rate": 0.00013024630541871922,
|
| 2521 |
+
"loss": 0.4199,
|
| 2522 |
+
"step": 7180
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 35.14338010982306,
|
| 2526 |
+
"grad_norm": 2.916937828063965,
|
| 2527 |
+
"learning_rate": 0.00013004926108374385,
|
| 2528 |
+
"loss": 0.4071,
|
| 2529 |
+
"step": 7200
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 35.241000610128125,
|
| 2533 |
+
"grad_norm": 3.1324169635772705,
|
| 2534 |
+
"learning_rate": 0.00012985221674876847,
|
| 2535 |
+
"loss": 0.4151,
|
| 2536 |
+
"step": 7220
|
| 2537 |
+
},
|
| 2538 |
+
{
|
| 2539 |
+
"epoch": 35.33862111043319,
|
| 2540 |
+
"grad_norm": 2.8730344772338867,
|
| 2541 |
+
"learning_rate": 0.0001296551724137931,
|
| 2542 |
+
"loss": 0.3984,
|
| 2543 |
+
"step": 7240
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 35.43624161073826,
|
| 2547 |
+
"grad_norm": 3.0865273475646973,
|
| 2548 |
+
"learning_rate": 0.00012945812807881775,
|
| 2549 |
+
"loss": 0.4273,
|
| 2550 |
+
"step": 7260
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 35.53386211104332,
|
| 2554 |
+
"grad_norm": 4.397771835327148,
|
| 2555 |
+
"learning_rate": 0.00012926108374384238,
|
| 2556 |
+
"loss": 0.4232,
|
| 2557 |
+
"step": 7280
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 35.63148261134838,
|
| 2561 |
+
"grad_norm": 2.4203243255615234,
|
| 2562 |
+
"learning_rate": 0.000129064039408867,
|
| 2563 |
+
"loss": 0.4035,
|
| 2564 |
+
"step": 7300
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 35.729103111653444,
|
| 2568 |
+
"grad_norm": 2.94404673576355,
|
| 2569 |
+
"learning_rate": 0.00012886699507389164,
|
| 2570 |
+
"loss": 0.4332,
|
| 2571 |
+
"step": 7320
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 35.82672361195851,
|
| 2575 |
+
"grad_norm": 3.4141249656677246,
|
| 2576 |
+
"learning_rate": 0.00012866995073891626,
|
| 2577 |
+
"loss": 0.4484,
|
| 2578 |
+
"step": 7340
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 35.924344112263576,
|
| 2582 |
+
"grad_norm": 2.8227927684783936,
|
| 2583 |
+
"learning_rate": 0.0001284729064039409,
|
| 2584 |
+
"loss": 0.4509,
|
| 2585 |
+
"step": 7360
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 36.02196461256864,
|
| 2589 |
+
"grad_norm": 2.768937110900879,
|
| 2590 |
+
"learning_rate": 0.00012827586206896552,
|
| 2591 |
+
"loss": 0.4391,
|
| 2592 |
+
"step": 7380
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 36.1195851128737,
|
| 2596 |
+
"grad_norm": 4.155871391296387,
|
| 2597 |
+
"learning_rate": 0.00012807881773399014,
|
| 2598 |
+
"loss": 0.3954,
|
| 2599 |
+
"step": 7400
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 36.21720561317877,
|
| 2603 |
+
"grad_norm": 2.484731912612915,
|
| 2604 |
+
"learning_rate": 0.0001278817733990148,
|
| 2605 |
+
"loss": 0.4363,
|
| 2606 |
+
"step": 7420
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 36.31482611348383,
|
| 2610 |
+
"grad_norm": 2.7758595943450928,
|
| 2611 |
+
"learning_rate": 0.0001276847290640394,
|
| 2612 |
+
"loss": 0.4058,
|
| 2613 |
+
"step": 7440
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 36.412446613788894,
|
| 2617 |
+
"grad_norm": 3.9609923362731934,
|
| 2618 |
+
"learning_rate": 0.00012748768472906405,
|
| 2619 |
+
"loss": 0.3845,
|
| 2620 |
+
"step": 7460
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 36.51006711409396,
|
| 2624 |
+
"grad_norm": 3.963120222091675,
|
| 2625 |
+
"learning_rate": 0.00012729064039408868,
|
| 2626 |
+
"loss": 0.4301,
|
| 2627 |
+
"step": 7480
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 36.607687614399026,
|
| 2631 |
+
"grad_norm": 2.77718448638916,
|
| 2632 |
+
"learning_rate": 0.0001270935960591133,
|
| 2633 |
+
"loss": 0.4034,
|
| 2634 |
+
"step": 7500
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 36.70530811470409,
|
| 2638 |
+
"grad_norm": 3.6000113487243652,
|
| 2639 |
+
"learning_rate": 0.00012689655172413793,
|
| 2640 |
+
"loss": 0.4087,
|
| 2641 |
+
"step": 7520
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 36.80292861500915,
|
| 2645 |
+
"grad_norm": 3.4430975914001465,
|
| 2646 |
+
"learning_rate": 0.00012669950738916256,
|
| 2647 |
+
"loss": 0.4109,
|
| 2648 |
+
"step": 7540
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 36.90054911531421,
|
| 2652 |
+
"grad_norm": 3.3932645320892334,
|
| 2653 |
+
"learning_rate": 0.00012650246305418721,
|
| 2654 |
+
"loss": 0.4394,
|
| 2655 |
+
"step": 7560
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 36.99816961561928,
|
| 2659 |
+
"grad_norm": 4.054554462432861,
|
| 2660 |
+
"learning_rate": 0.00012630541871921181,
|
| 2661 |
+
"loss": 0.4203,
|
| 2662 |
+
"step": 7580
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 37.095790115924345,
|
| 2666 |
+
"grad_norm": 2.8766210079193115,
|
| 2667 |
+
"learning_rate": 0.00012610837438423647,
|
| 2668 |
+
"loss": 0.3861,
|
| 2669 |
+
"step": 7600
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 37.19341061622941,
|
| 2673 |
+
"grad_norm": 4.115131855010986,
|
| 2674 |
+
"learning_rate": 0.0001259113300492611,
|
| 2675 |
+
"loss": 0.4236,
|
| 2676 |
+
"step": 7620
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 37.29103111653447,
|
| 2680 |
+
"grad_norm": 2.776914358139038,
|
| 2681 |
+
"learning_rate": 0.00012571428571428572,
|
| 2682 |
+
"loss": 0.4244,
|
| 2683 |
+
"step": 7640
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 37.38865161683954,
|
| 2687 |
+
"grad_norm": 3.8428800106048584,
|
| 2688 |
+
"learning_rate": 0.00012551724137931035,
|
| 2689 |
+
"loss": 0.4028,
|
| 2690 |
+
"step": 7660
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 37.4862721171446,
|
| 2694 |
+
"grad_norm": 3.028683662414551,
|
| 2695 |
+
"learning_rate": 0.00012532019704433498,
|
| 2696 |
+
"loss": 0.4127,
|
| 2697 |
+
"step": 7680
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 37.58389261744966,
|
| 2701 |
+
"grad_norm": 2.678617477416992,
|
| 2702 |
+
"learning_rate": 0.0001251231527093596,
|
| 2703 |
+
"loss": 0.4251,
|
| 2704 |
+
"step": 7700
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 37.681513117754726,
|
| 2708 |
+
"grad_norm": 3.496917247772217,
|
| 2709 |
+
"learning_rate": 0.00012492610837438423,
|
| 2710 |
+
"loss": 0.404,
|
| 2711 |
+
"step": 7720
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 37.779133618059795,
|
| 2715 |
+
"grad_norm": 4.018653869628906,
|
| 2716 |
+
"learning_rate": 0.00012472906403940889,
|
| 2717 |
+
"loss": 0.4028,
|
| 2718 |
+
"step": 7740
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 37.87675411836486,
|
| 2722 |
+
"grad_norm": 3.317580223083496,
|
| 2723 |
+
"learning_rate": 0.0001245320197044335,
|
| 2724 |
+
"loss": 0.4032,
|
| 2725 |
+
"step": 7760
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 37.97437461866992,
|
| 2729 |
+
"grad_norm": 3.7693002223968506,
|
| 2730 |
+
"learning_rate": 0.00012433497536945814,
|
| 2731 |
+
"loss": 0.3935,
|
| 2732 |
+
"step": 7780
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 38.07199511897498,
|
| 2736 |
+
"grad_norm": 2.809558629989624,
|
| 2737 |
+
"learning_rate": 0.00012413793103448277,
|
| 2738 |
+
"loss": 0.4113,
|
| 2739 |
+
"step": 7800
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 38.16961561928005,
|
| 2743 |
+
"grad_norm": 3.2092092037200928,
|
| 2744 |
+
"learning_rate": 0.0001239408866995074,
|
| 2745 |
+
"loss": 0.4019,
|
| 2746 |
+
"step": 7820
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 38.267236119585114,
|
| 2750 |
+
"grad_norm": 3.3514404296875,
|
| 2751 |
+
"learning_rate": 0.00012374384236453202,
|
| 2752 |
+
"loss": 0.4013,
|
| 2753 |
+
"step": 7840
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 38.364856619890176,
|
| 2757 |
+
"grad_norm": 3.9514451026916504,
|
| 2758 |
+
"learning_rate": 0.00012354679802955667,
|
| 2759 |
+
"loss": 0.3889,
|
| 2760 |
+
"step": 7860
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 38.46247712019524,
|
| 2764 |
+
"grad_norm": 2.7896828651428223,
|
| 2765 |
+
"learning_rate": 0.00012334975369458127,
|
| 2766 |
+
"loss": 0.377,
|
| 2767 |
+
"step": 7880
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 38.56009762050031,
|
| 2771 |
+
"grad_norm": 3.522840738296509,
|
| 2772 |
+
"learning_rate": 0.00012315270935960593,
|
| 2773 |
+
"loss": 0.4158,
|
| 2774 |
+
"step": 7900
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 38.65771812080537,
|
| 2778 |
+
"grad_norm": 3.422250270843506,
|
| 2779 |
+
"learning_rate": 0.00012295566502463053,
|
| 2780 |
+
"loss": 0.3837,
|
| 2781 |
+
"step": 7920
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 38.75533862111043,
|
| 2785 |
+
"grad_norm": 3.0469913482666016,
|
| 2786 |
+
"learning_rate": 0.00012275862068965518,
|
| 2787 |
+
"loss": 0.4036,
|
| 2788 |
+
"step": 7940
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 38.852959121415495,
|
| 2792 |
+
"grad_norm": 2.904141664505005,
|
| 2793 |
+
"learning_rate": 0.0001225615763546798,
|
| 2794 |
+
"loss": 0.3928,
|
| 2795 |
+
"step": 7960
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 38.950579621720564,
|
| 2799 |
+
"grad_norm": 3.7538552284240723,
|
| 2800 |
+
"learning_rate": 0.00012236453201970444,
|
| 2801 |
+
"loss": 0.4092,
|
| 2802 |
+
"step": 7980
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 39.04820012202563,
|
| 2806 |
+
"grad_norm": 3.562114715576172,
|
| 2807 |
+
"learning_rate": 0.00012216748768472906,
|
| 2808 |
+
"loss": 0.3982,
|
| 2809 |
+
"step": 8000
|
| 2810 |
+
},
|
| 2811 |
+
{
|
| 2812 |
+
"epoch": 39.14582062233069,
|
| 2813 |
+
"grad_norm": 2.4931962490081787,
|
| 2814 |
+
"learning_rate": 0.00012197044334975369,
|
| 2815 |
+
"loss": 0.3547,
|
| 2816 |
+
"step": 8020
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 39.24344112263575,
|
| 2820 |
+
"grad_norm": 2.461050271987915,
|
| 2821 |
+
"learning_rate": 0.00012177339901477833,
|
| 2822 |
+
"loss": 0.3762,
|
| 2823 |
+
"step": 8040
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 39.34106162294082,
|
| 2827 |
+
"grad_norm": 3.1320595741271973,
|
| 2828 |
+
"learning_rate": 0.00012157635467980295,
|
| 2829 |
+
"loss": 0.3907,
|
| 2830 |
+
"step": 8060
|
| 2831 |
+
},
|
| 2832 |
+
{
|
| 2833 |
+
"epoch": 39.43868212324588,
|
| 2834 |
+
"grad_norm": 3.044754981994629,
|
| 2835 |
+
"learning_rate": 0.00012137931034482759,
|
| 2836 |
+
"loss": 0.4068,
|
| 2837 |
+
"step": 8080
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 39.536302623550945,
|
| 2841 |
+
"grad_norm": 2.9243273735046387,
|
| 2842 |
+
"learning_rate": 0.00012118226600985223,
|
| 2843 |
+
"loss": 0.3903,
|
| 2844 |
+
"step": 8100
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 39.63392312385601,
|
| 2848 |
+
"grad_norm": 4.234837055206299,
|
| 2849 |
+
"learning_rate": 0.00012098522167487685,
|
| 2850 |
+
"loss": 0.3841,
|
| 2851 |
+
"step": 8120
|
| 2852 |
+
},
|
| 2853 |
+
{
|
| 2854 |
+
"epoch": 39.73154362416108,
|
| 2855 |
+
"grad_norm": 3.993495464324951,
|
| 2856 |
+
"learning_rate": 0.00012078817733990148,
|
| 2857 |
+
"loss": 0.4082,
|
| 2858 |
+
"step": 8140
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 39.82916412446614,
|
| 2862 |
+
"grad_norm": 3.8363142013549805,
|
| 2863 |
+
"learning_rate": 0.00012059113300492611,
|
| 2864 |
+
"loss": 0.3939,
|
| 2865 |
+
"step": 8160
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 39.9267846247712,
|
| 2869 |
+
"grad_norm": 4.398952007293701,
|
| 2870 |
+
"learning_rate": 0.00012039408866995075,
|
| 2871 |
+
"loss": 0.4145,
|
| 2872 |
+
"step": 8180
|
| 2873 |
+
},
|
| 2874 |
+
{
|
| 2875 |
+
"epoch": 40.024405125076264,
|
| 2876 |
+
"grad_norm": 2.7002291679382324,
|
| 2877 |
+
"learning_rate": 0.00012019704433497539,
|
| 2878 |
+
"loss": 0.386,
|
| 2879 |
+
"step": 8200
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 40.12202562538133,
|
| 2883 |
+
"grad_norm": 3.1867945194244385,
|
| 2884 |
+
"learning_rate": 0.00012,
|
| 2885 |
+
"loss": 0.3924,
|
| 2886 |
+
"step": 8220
|
| 2887 |
+
},
|
| 2888 |
+
{
|
| 2889 |
+
"epoch": 40.219646125686396,
|
| 2890 |
+
"grad_norm": 2.9179584980010986,
|
| 2891 |
+
"learning_rate": 0.00011980295566502464,
|
| 2892 |
+
"loss": 0.3741,
|
| 2893 |
+
"step": 8240
|
| 2894 |
+
},
|
| 2895 |
+
{
|
| 2896 |
+
"epoch": 40.31726662599146,
|
| 2897 |
+
"grad_norm": 5.108730316162109,
|
| 2898 |
+
"learning_rate": 0.00011960591133004926,
|
| 2899 |
+
"loss": 0.371,
|
| 2900 |
+
"step": 8260
|
| 2901 |
+
},
|
| 2902 |
+
{
|
| 2903 |
+
"epoch": 40.41488712629652,
|
| 2904 |
+
"grad_norm": 3.4418270587921143,
|
| 2905 |
+
"learning_rate": 0.0001194088669950739,
|
| 2906 |
+
"loss": 0.3845,
|
| 2907 |
+
"step": 8280
|
| 2908 |
+
},
|
| 2909 |
+
{
|
| 2910 |
+
"epoch": 40.51250762660159,
|
| 2911 |
+
"grad_norm": 3.245562791824341,
|
| 2912 |
+
"learning_rate": 0.00011921182266009854,
|
| 2913 |
+
"loss": 0.375,
|
| 2914 |
+
"step": 8300
|
| 2915 |
+
},
|
| 2916 |
+
{
|
| 2917 |
+
"epoch": 40.61012812690665,
|
| 2918 |
+
"grad_norm": 2.6644446849823,
|
| 2919 |
+
"learning_rate": 0.00011901477832512315,
|
| 2920 |
+
"loss": 0.3839,
|
| 2921 |
+
"step": 8320
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 40.707748627211714,
|
| 2925 |
+
"grad_norm": 4.975727558135986,
|
| 2926 |
+
"learning_rate": 0.00011881773399014779,
|
| 2927 |
+
"loss": 0.3889,
|
| 2928 |
+
"step": 8340
|
| 2929 |
+
},
|
| 2930 |
+
{
|
| 2931 |
+
"epoch": 40.80536912751678,
|
| 2932 |
+
"grad_norm": 3.6427066326141357,
|
| 2933 |
+
"learning_rate": 0.0001186206896551724,
|
| 2934 |
+
"loss": 0.393,
|
| 2935 |
+
"step": 8360
|
| 2936 |
+
},
|
| 2937 |
+
{
|
| 2938 |
+
"epoch": 40.902989627821846,
|
| 2939 |
+
"grad_norm": 3.7799060344696045,
|
| 2940 |
+
"learning_rate": 0.00011842364532019705,
|
| 2941 |
+
"loss": 0.3894,
|
| 2942 |
+
"step": 8380
|
| 2943 |
+
},
|
| 2944 |
+
{
|
| 2945 |
+
"epoch": 41.00061012812691,
|
| 2946 |
+
"grad_norm": 4.170138835906982,
|
| 2947 |
+
"learning_rate": 0.00011822660098522169,
|
| 2948 |
+
"loss": 0.3965,
|
| 2949 |
+
"step": 8400
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 41.09823062843197,
|
| 2953 |
+
"grad_norm": 2.660006523132324,
|
| 2954 |
+
"learning_rate": 0.00011802955665024631,
|
| 2955 |
+
"loss": 0.3412,
|
| 2956 |
+
"step": 8420
|
| 2957 |
+
},
|
| 2958 |
+
{
|
| 2959 |
+
"epoch": 41.19585112873703,
|
| 2960 |
+
"grad_norm": 3.9118030071258545,
|
| 2961 |
+
"learning_rate": 0.00011783251231527096,
|
| 2962 |
+
"loss": 0.3608,
|
| 2963 |
+
"step": 8440
|
| 2964 |
+
},
|
| 2965 |
+
{
|
| 2966 |
+
"epoch": 41.2934716290421,
|
| 2967 |
+
"grad_norm": 4.68622350692749,
|
| 2968 |
+
"learning_rate": 0.00011763546798029557,
|
| 2969 |
+
"loss": 0.3742,
|
| 2970 |
+
"step": 8460
|
| 2971 |
+
},
|
| 2972 |
+
{
|
| 2973 |
+
"epoch": 41.391092129347165,
|
| 2974 |
+
"grad_norm": 2.5423784255981445,
|
| 2975 |
+
"learning_rate": 0.00011743842364532021,
|
| 2976 |
+
"loss": 0.3901,
|
| 2977 |
+
"step": 8480
|
| 2978 |
+
},
|
| 2979 |
+
{
|
| 2980 |
+
"epoch": 41.48871262965223,
|
| 2981 |
+
"grad_norm": 3.6446280479431152,
|
| 2982 |
+
"learning_rate": 0.00011724137931034482,
|
| 2983 |
+
"loss": 0.3518,
|
| 2984 |
+
"step": 8500
|
| 2985 |
+
},
|
| 2986 |
+
{
|
| 2987 |
+
"epoch": 41.58633312995729,
|
| 2988 |
+
"grad_norm": 2.6701178550720215,
|
| 2989 |
+
"learning_rate": 0.00011704433497536946,
|
| 2990 |
+
"loss": 0.3809,
|
| 2991 |
+
"step": 8520
|
| 2992 |
+
},
|
| 2993 |
+
{
|
| 2994 |
+
"epoch": 41.68395363026236,
|
| 2995 |
+
"grad_norm": 3.226100206375122,
|
| 2996 |
+
"learning_rate": 0.0001168472906403941,
|
| 2997 |
+
"loss": 0.3834,
|
| 2998 |
+
"step": 8540
|
| 2999 |
+
},
|
| 3000 |
+
{
|
| 3001 |
+
"epoch": 41.78157413056742,
|
| 3002 |
+
"grad_norm": 3.4181952476501465,
|
| 3003 |
+
"learning_rate": 0.00011665024630541872,
|
| 3004 |
+
"loss": 0.4098,
|
| 3005 |
+
"step": 8560
|
| 3006 |
+
},
|
| 3007 |
+
{
|
| 3008 |
+
"epoch": 41.87919463087248,
|
| 3009 |
+
"grad_norm": 2.9190330505371094,
|
| 3010 |
+
"learning_rate": 0.00011645320197044336,
|
| 3011 |
+
"loss": 0.3838,
|
| 3012 |
+
"step": 8580
|
| 3013 |
+
},
|
| 3014 |
+
{
|
| 3015 |
+
"epoch": 41.976815131177545,
|
| 3016 |
+
"grad_norm": 4.082178115844727,
|
| 3017 |
+
"learning_rate": 0.00011625615763546797,
|
| 3018 |
+
"loss": 0.4109,
|
| 3019 |
+
"step": 8600
|
| 3020 |
+
},
|
| 3021 |
+
{
|
| 3022 |
+
"epoch": 42.074435631482615,
|
| 3023 |
+
"grad_norm": 2.899162530899048,
|
| 3024 |
+
"learning_rate": 0.00011605911330049261,
|
| 3025 |
+
"loss": 0.3624,
|
| 3026 |
+
"step": 8620
|
| 3027 |
+
},
|
| 3028 |
+
{
|
| 3029 |
+
"epoch": 42.17205613178768,
|
| 3030 |
+
"grad_norm": 2.4065990447998047,
|
| 3031 |
+
"learning_rate": 0.00011586206896551725,
|
| 3032 |
+
"loss": 0.3573,
|
| 3033 |
+
"step": 8640
|
| 3034 |
+
},
|
| 3035 |
+
{
|
| 3036 |
+
"epoch": 42.26967663209274,
|
| 3037 |
+
"grad_norm": 2.818037509918213,
|
| 3038 |
+
"learning_rate": 0.00011566502463054188,
|
| 3039 |
+
"loss": 0.3699,
|
| 3040 |
+
"step": 8660
|
| 3041 |
+
},
|
| 3042 |
+
{
|
| 3043 |
+
"epoch": 42.3672971323978,
|
| 3044 |
+
"grad_norm": 2.8875226974487305,
|
| 3045 |
+
"learning_rate": 0.00011546798029556651,
|
| 3046 |
+
"loss": 0.3489,
|
| 3047 |
+
"step": 8680
|
| 3048 |
+
},
|
| 3049 |
+
{
|
| 3050 |
+
"epoch": 42.464917632702864,
|
| 3051 |
+
"grad_norm": 3.0840396881103516,
|
| 3052 |
+
"learning_rate": 0.00011527093596059113,
|
| 3053 |
+
"loss": 0.3733,
|
| 3054 |
+
"step": 8700
|
| 3055 |
+
},
|
| 3056 |
+
{
|
| 3057 |
+
"epoch": 42.56253813300793,
|
| 3058 |
+
"grad_norm": 2.6554925441741943,
|
| 3059 |
+
"learning_rate": 0.00011507389162561578,
|
| 3060 |
+
"loss": 0.3541,
|
| 3061 |
+
"step": 8720
|
| 3062 |
+
},
|
| 3063 |
+
{
|
| 3064 |
+
"epoch": 42.660158633312996,
|
| 3065 |
+
"grad_norm": 2.766045331954956,
|
| 3066 |
+
"learning_rate": 0.00011487684729064042,
|
| 3067 |
+
"loss": 0.3682,
|
| 3068 |
+
"step": 8740
|
| 3069 |
+
},
|
| 3070 |
+
{
|
| 3071 |
+
"epoch": 42.75777913361806,
|
| 3072 |
+
"grad_norm": 3.0672762393951416,
|
| 3073 |
+
"learning_rate": 0.00011467980295566503,
|
| 3074 |
+
"loss": 0.3943,
|
| 3075 |
+
"step": 8760
|
| 3076 |
+
},
|
| 3077 |
+
{
|
| 3078 |
+
"epoch": 42.85539963392312,
|
| 3079 |
+
"grad_norm": 2.898484468460083,
|
| 3080 |
+
"learning_rate": 0.00011448275862068967,
|
| 3081 |
+
"loss": 0.3702,
|
| 3082 |
+
"step": 8780
|
| 3083 |
+
},
|
| 3084 |
+
{
|
| 3085 |
+
"epoch": 42.95302013422819,
|
| 3086 |
+
"grad_norm": 2.7023797035217285,
|
| 3087 |
+
"learning_rate": 0.00011428571428571428,
|
| 3088 |
+
"loss": 0.388,
|
| 3089 |
+
"step": 8800
|
| 3090 |
+
},
|
| 3091 |
+
{
|
| 3092 |
+
"epoch": 43.05064063453325,
|
| 3093 |
+
"grad_norm": 2.4088499546051025,
|
| 3094 |
+
"learning_rate": 0.00011408866995073892,
|
| 3095 |
+
"loss": 0.3615,
|
| 3096 |
+
"step": 8820
|
| 3097 |
+
},
|
| 3098 |
+
{
|
| 3099 |
+
"epoch": 43.148261134838314,
|
| 3100 |
+
"grad_norm": 2.3739655017852783,
|
| 3101 |
+
"learning_rate": 0.00011389162561576354,
|
| 3102 |
+
"loss": 0.3703,
|
| 3103 |
+
"step": 8840
|
| 3104 |
+
},
|
| 3105 |
+
{
|
| 3106 |
+
"epoch": 43.24588163514338,
|
| 3107 |
+
"grad_norm": 3.2558271884918213,
|
| 3108 |
+
"learning_rate": 0.00011369458128078818,
|
| 3109 |
+
"loss": 0.3478,
|
| 3110 |
+
"step": 8860
|
| 3111 |
+
},
|
| 3112 |
+
{
|
| 3113 |
+
"epoch": 43.343502135448446,
|
| 3114 |
+
"grad_norm": 2.931380271911621,
|
| 3115 |
+
"learning_rate": 0.00011349753694581282,
|
| 3116 |
+
"loss": 0.3553,
|
| 3117 |
+
"step": 8880
|
| 3118 |
+
},
|
| 3119 |
+
{
|
| 3120 |
+
"epoch": 43.44112263575351,
|
| 3121 |
+
"grad_norm": 2.5165908336639404,
|
| 3122 |
+
"learning_rate": 0.00011330049261083743,
|
| 3123 |
+
"loss": 0.3495,
|
| 3124 |
+
"step": 8900
|
| 3125 |
+
},
|
| 3126 |
+
{
|
| 3127 |
+
"epoch": 43.53874313605857,
|
| 3128 |
+
"grad_norm": 3.5619068145751953,
|
| 3129 |
+
"learning_rate": 0.00011310344827586207,
|
| 3130 |
+
"loss": 0.3692,
|
| 3131 |
+
"step": 8920
|
| 3132 |
+
},
|
| 3133 |
+
{
|
| 3134 |
+
"epoch": 43.63636363636363,
|
| 3135 |
+
"grad_norm": 2.39534068107605,
|
| 3136 |
+
"learning_rate": 0.0001129064039408867,
|
| 3137 |
+
"loss": 0.3674,
|
| 3138 |
+
"step": 8940
|
| 3139 |
+
},
|
| 3140 |
+
{
|
| 3141 |
+
"epoch": 43.7339841366687,
|
| 3142 |
+
"grad_norm": 3.495316505432129,
|
| 3143 |
+
"learning_rate": 0.00011270935960591134,
|
| 3144 |
+
"loss": 0.367,
|
| 3145 |
+
"step": 8960
|
| 3146 |
+
},
|
| 3147 |
+
{
|
| 3148 |
+
"epoch": 43.831604636973765,
|
| 3149 |
+
"grad_norm": 2.8195016384124756,
|
| 3150 |
+
"learning_rate": 0.00011251231527093598,
|
| 3151 |
+
"loss": 0.411,
|
| 3152 |
+
"step": 8980
|
| 3153 |
+
},
|
| 3154 |
+
{
|
| 3155 |
+
"epoch": 43.92922513727883,
|
| 3156 |
+
"grad_norm": 3.446014165878296,
|
| 3157 |
+
"learning_rate": 0.0001123152709359606,
|
| 3158 |
+
"loss": 0.3774,
|
| 3159 |
+
"step": 9000
|
| 3160 |
}
|
| 3161 |
],
|
| 3162 |
"logging_steps": 20,
|
|
|
|
| 3176 |
"attributes": {}
|
| 3177 |
}
|
| 3178 |
},
|
| 3179 |
+
"total_flos": 1.79330460401664e+17,
|
| 3180 |
"train_batch_size": 1,
|
| 3181 |
"trial_name": null,
|
| 3182 |
"trial_params": null
|