frostsolutions/frost-vision-v2-google_vit-base-patch16-224-v2024-11-14
Browse files- README.md +9 -9
- all_results.json +16 -0
- eval_results.json +11 -0
- runs/Nov14_23-04-08_ba4b501b14a9/events.out.tfevents.1731626158.ba4b501b14a9.833.1 +3 -0
- train_results.json +8 -0
- trainer_state.json +2031 -0
README.md
CHANGED
|
@@ -26,16 +26,16 @@ model-index:
|
|
| 26 |
metrics:
|
| 27 |
- name: Accuracy
|
| 28 |
type: accuracy
|
| 29 |
-
value: 0.
|
| 30 |
- name: F1
|
| 31 |
type: f1
|
| 32 |
-
value: 0.
|
| 33 |
- name: Precision
|
| 34 |
type: precision
|
| 35 |
-
value: 0.
|
| 36 |
- name: Recall
|
| 37 |
type: recall
|
| 38 |
-
value: 0.
|
| 39 |
---
|
| 40 |
|
| 41 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
@@ -45,11 +45,11 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 45 |
|
| 46 |
This model is a fine-tuned version of [google/vit-base-patch16-224](https://huggingface.co/google/vit-base-patch16-224) on the webdataset dataset.
|
| 47 |
It achieves the following results on the evaluation set:
|
| 48 |
-
- Loss: 0.
|
| 49 |
-
- Accuracy: 0.
|
| 50 |
-
- F1: 0.
|
| 51 |
-
- Precision: 0.
|
| 52 |
-
- Recall: 0.
|
| 53 |
|
| 54 |
## Model description
|
| 55 |
|
|
|
|
| 26 |
metrics:
|
| 27 |
- name: Accuracy
|
| 28 |
type: accuracy
|
| 29 |
+
value: 0.924074074074074
|
| 30 |
- name: F1
|
| 31 |
type: f1
|
| 32 |
+
value: 0.8069073783359497
|
| 33 |
- name: Precision
|
| 34 |
type: precision
|
| 35 |
+
value: 0.8524046434494196
|
| 36 |
- name: Recall
|
| 37 |
type: recall
|
| 38 |
+
value: 0.7660208643815202
|
| 39 |
---
|
| 40 |
|
| 41 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
|
|
| 45 |
|
| 46 |
This model is a fine-tuned version of [google/vit-base-patch16-224](https://huggingface.co/google/vit-base-patch16-224) on the webdataset dataset.
|
| 47 |
It achieves the following results on the evaluation set:
|
| 48 |
+
- Loss: 0.1858
|
| 49 |
+
- Accuracy: 0.9241
|
| 50 |
+
- F1: 0.8069
|
| 51 |
+
- Precision: 0.8524
|
| 52 |
+
- Recall: 0.7660
|
| 53 |
|
| 54 |
## Model description
|
| 55 |
|
all_results.json
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 30.0,
|
| 3 |
+
"eval_accuracy": 0.924074074074074,
|
| 4 |
+
"eval_f1": 0.8069073783359497,
|
| 5 |
+
"eval_loss": 0.18580639362335205,
|
| 6 |
+
"eval_precision": 0.8524046434494196,
|
| 7 |
+
"eval_recall": 0.7660208643815202,
|
| 8 |
+
"eval_runtime": 3.2436,
|
| 9 |
+
"eval_samples_per_second": 99.889,
|
| 10 |
+
"eval_steps_per_second": 12.64,
|
| 11 |
+
"total_flos": 3.008454731998249e+18,
|
| 12 |
+
"train_loss": 0.08984789192185971,
|
| 13 |
+
"train_runtime": 670.5653,
|
| 14 |
+
"train_samples_per_second": 57.891,
|
| 15 |
+
"train_steps_per_second": 3.624
|
| 16 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 30.0,
|
| 3 |
+
"eval_accuracy": 0.924074074074074,
|
| 4 |
+
"eval_f1": 0.8069073783359497,
|
| 5 |
+
"eval_loss": 0.18580639362335205,
|
| 6 |
+
"eval_precision": 0.8524046434494196,
|
| 7 |
+
"eval_recall": 0.7660208643815202,
|
| 8 |
+
"eval_runtime": 3.2436,
|
| 9 |
+
"eval_samples_per_second": 99.889,
|
| 10 |
+
"eval_steps_per_second": 12.64
|
| 11 |
+
}
|
runs/Nov14_23-04-08_ba4b501b14a9/events.out.tfevents.1731626158.ba4b501b14a9.833.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fc344eddf54d845eeeeeb0cd2353e359c7b18313b46347e8bd5fb12ed9bb41dd
|
| 3 |
+
size 560
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 30.0,
|
| 3 |
+
"total_flos": 3.008454731998249e+18,
|
| 4 |
+
"train_loss": 0.08984789192185971,
|
| 5 |
+
"train_runtime": 670.5653,
|
| 6 |
+
"train_samples_per_second": 57.891,
|
| 7 |
+
"train_steps_per_second": 3.624
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2031 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.18580639362335205,
|
| 3 |
+
"best_model_checkpoint": "frostsolutions/frost-vision-v2-google_vit-base-patch16-224-v2024-11-14/checkpoint-500",
|
| 4 |
+
"epoch": 30.0,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 2430,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.12345679012345678,
|
| 13 |
+
"grad_norm": 1.8872668743133545,
|
| 14 |
+
"learning_rate": 8.23045267489712e-06,
|
| 15 |
+
"loss": 0.7179,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.24691358024691357,
|
| 20 |
+
"grad_norm": 1.4793915748596191,
|
| 21 |
+
"learning_rate": 1.646090534979424e-05,
|
| 22 |
+
"loss": 0.632,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.37037037037037035,
|
| 27 |
+
"grad_norm": 1.0972111225128174,
|
| 28 |
+
"learning_rate": 2.4691358024691357e-05,
|
| 29 |
+
"loss": 0.5057,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.49382716049382713,
|
| 34 |
+
"grad_norm": 1.1177233457565308,
|
| 35 |
+
"learning_rate": 3.292181069958848e-05,
|
| 36 |
+
"loss": 0.4289,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.6172839506172839,
|
| 41 |
+
"grad_norm": 1.0670571327209473,
|
| 42 |
+
"learning_rate": 4.11522633744856e-05,
|
| 43 |
+
"loss": 0.3715,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.7407407407407407,
|
| 48 |
+
"grad_norm": 1.502959132194519,
|
| 49 |
+
"learning_rate": 4.938271604938271e-05,
|
| 50 |
+
"loss": 0.3202,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.8641975308641975,
|
| 55 |
+
"grad_norm": 1.198879361152649,
|
| 56 |
+
"learning_rate": 5.761316872427984e-05,
|
| 57 |
+
"loss": 0.3032,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.9876543209876543,
|
| 62 |
+
"grad_norm": 1.1655505895614624,
|
| 63 |
+
"learning_rate": 6.584362139917696e-05,
|
| 64 |
+
"loss": 0.2917,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 1.1111111111111112,
|
| 69 |
+
"grad_norm": 1.1428577899932861,
|
| 70 |
+
"learning_rate": 7.407407407407407e-05,
|
| 71 |
+
"loss": 0.2521,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 1.2345679012345678,
|
| 76 |
+
"grad_norm": 0.6600608825683594,
|
| 77 |
+
"learning_rate": 8.23045267489712e-05,
|
| 78 |
+
"loss": 0.2436,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 1.2345679012345678,
|
| 83 |
+
"eval_accuracy": 0.9012345679012346,
|
| 84 |
+
"eval_f1": 0.744408945686901,
|
| 85 |
+
"eval_loss": 0.24551859498023987,
|
| 86 |
+
"eval_precision": 0.802065404475043,
|
| 87 |
+
"eval_recall": 0.6944858420268256,
|
| 88 |
+
"eval_runtime": 3.0062,
|
| 89 |
+
"eval_samples_per_second": 107.778,
|
| 90 |
+
"eval_steps_per_second": 13.639,
|
| 91 |
+
"step": 100
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.3580246913580247,
|
| 95 |
+
"grad_norm": 0.9017321467399597,
|
| 96 |
+
"learning_rate": 9.053497942386831e-05,
|
| 97 |
+
"loss": 0.2218,
|
| 98 |
+
"step": 110
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.4814814814814814,
|
| 102 |
+
"grad_norm": 0.7535004615783691,
|
| 103 |
+
"learning_rate": 9.876543209876543e-05,
|
| 104 |
+
"loss": 0.2229,
|
| 105 |
+
"step": 120
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.6049382716049383,
|
| 109 |
+
"grad_norm": 0.7849411964416504,
|
| 110 |
+
"learning_rate": 0.00010699588477366255,
|
| 111 |
+
"loss": 0.2478,
|
| 112 |
+
"step": 130
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.7283950617283952,
|
| 116 |
+
"grad_norm": 1.008712649345398,
|
| 117 |
+
"learning_rate": 0.00011522633744855968,
|
| 118 |
+
"loss": 0.237,
|
| 119 |
+
"step": 140
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.8518518518518519,
|
| 123 |
+
"grad_norm": 0.7667405605316162,
|
| 124 |
+
"learning_rate": 0.0001234567901234568,
|
| 125 |
+
"loss": 0.2143,
|
| 126 |
+
"step": 150
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 1.9753086419753085,
|
| 130 |
+
"grad_norm": 1.147663950920105,
|
| 131 |
+
"learning_rate": 0.00013168724279835392,
|
| 132 |
+
"loss": 0.2208,
|
| 133 |
+
"step": 160
|
| 134 |
+
},
|
| 135 |
+
{
|
| 136 |
+
"epoch": 2.0987654320987654,
|
| 137 |
+
"grad_norm": 0.746328592300415,
|
| 138 |
+
"learning_rate": 0.00013991769547325105,
|
| 139 |
+
"loss": 0.1848,
|
| 140 |
+
"step": 170
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 2.2222222222222223,
|
| 144 |
+
"grad_norm": 1.367189884185791,
|
| 145 |
+
"learning_rate": 0.00014814814814814815,
|
| 146 |
+
"loss": 0.1901,
|
| 147 |
+
"step": 180
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 2.3456790123456788,
|
| 151 |
+
"grad_norm": 0.8272459506988525,
|
| 152 |
+
"learning_rate": 0.00015637860082304527,
|
| 153 |
+
"loss": 0.1811,
|
| 154 |
+
"step": 190
|
| 155 |
+
},
|
| 156 |
+
{
|
| 157 |
+
"epoch": 2.4691358024691357,
|
| 158 |
+
"grad_norm": 0.7679163217544556,
|
| 159 |
+
"learning_rate": 0.0001646090534979424,
|
| 160 |
+
"loss": 0.1639,
|
| 161 |
+
"step": 200
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 2.4691358024691357,
|
| 165 |
+
"eval_accuracy": 0.917283950617284,
|
| 166 |
+
"eval_f1": 0.7981927710843374,
|
| 167 |
+
"eval_loss": 0.2006644308567047,
|
| 168 |
+
"eval_precision": 0.806697108066971,
|
| 169 |
+
"eval_recall": 0.789865871833085,
|
| 170 |
+
"eval_runtime": 2.6836,
|
| 171 |
+
"eval_samples_per_second": 120.731,
|
| 172 |
+
"eval_steps_per_second": 15.278,
|
| 173 |
+
"step": 200
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 2.5925925925925926,
|
| 177 |
+
"grad_norm": 1.0779305696487427,
|
| 178 |
+
"learning_rate": 0.0001728395061728395,
|
| 179 |
+
"loss": 0.1977,
|
| 180 |
+
"step": 210
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"epoch": 2.7160493827160495,
|
| 184 |
+
"grad_norm": 1.054652452468872,
|
| 185 |
+
"learning_rate": 0.00018106995884773663,
|
| 186 |
+
"loss": 0.2195,
|
| 187 |
+
"step": 220
|
| 188 |
+
},
|
| 189 |
+
{
|
| 190 |
+
"epoch": 2.8395061728395063,
|
| 191 |
+
"grad_norm": 0.7771775722503662,
|
| 192 |
+
"learning_rate": 0.00018930041152263375,
|
| 193 |
+
"loss": 0.1997,
|
| 194 |
+
"step": 230
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 2.962962962962963,
|
| 198 |
+
"grad_norm": 1.1053166389465332,
|
| 199 |
+
"learning_rate": 0.00019753086419753085,
|
| 200 |
+
"loss": 0.2222,
|
| 201 |
+
"step": 240
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 3.0864197530864197,
|
| 205 |
+
"grad_norm": 0.7908689975738525,
|
| 206 |
+
"learning_rate": 0.00019935985368084132,
|
| 207 |
+
"loss": 0.2012,
|
| 208 |
+
"step": 250
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 3.2098765432098766,
|
| 212 |
+
"grad_norm": 1.0789804458618164,
|
| 213 |
+
"learning_rate": 0.0001984453589391861,
|
| 214 |
+
"loss": 0.1826,
|
| 215 |
+
"step": 260
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 3.3333333333333335,
|
| 219 |
+
"grad_norm": 0.8242624998092651,
|
| 220 |
+
"learning_rate": 0.00019753086419753085,
|
| 221 |
+
"loss": 0.2012,
|
| 222 |
+
"step": 270
|
| 223 |
+
},
|
| 224 |
+
{
|
| 225 |
+
"epoch": 3.45679012345679,
|
| 226 |
+
"grad_norm": 1.4627937078475952,
|
| 227 |
+
"learning_rate": 0.00019661636945587566,
|
| 228 |
+
"loss": 0.2098,
|
| 229 |
+
"step": 280
|
| 230 |
+
},
|
| 231 |
+
{
|
| 232 |
+
"epoch": 3.580246913580247,
|
| 233 |
+
"grad_norm": 1.009103536605835,
|
| 234 |
+
"learning_rate": 0.0001957018747142204,
|
| 235 |
+
"loss": 0.1744,
|
| 236 |
+
"step": 290
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 3.7037037037037037,
|
| 240 |
+
"grad_norm": 0.6586538553237915,
|
| 241 |
+
"learning_rate": 0.0001947873799725652,
|
| 242 |
+
"loss": 0.188,
|
| 243 |
+
"step": 300
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 3.7037037037037037,
|
| 247 |
+
"eval_accuracy": 0.9219135802469136,
|
| 248 |
+
"eval_f1": 0.8064269319051263,
|
| 249 |
+
"eval_loss": 0.19134539365768433,
|
| 250 |
+
"eval_precision": 0.8286163522012578,
|
| 251 |
+
"eval_recall": 0.7853949329359166,
|
| 252 |
+
"eval_runtime": 2.376,
|
| 253 |
+
"eval_samples_per_second": 136.362,
|
| 254 |
+
"eval_steps_per_second": 17.256,
|
| 255 |
+
"step": 300
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 3.8271604938271606,
|
| 259 |
+
"grad_norm": 0.6922177076339722,
|
| 260 |
+
"learning_rate": 0.00019387288523090994,
|
| 261 |
+
"loss": 0.1915,
|
| 262 |
+
"step": 310
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 3.950617283950617,
|
| 266 |
+
"grad_norm": 0.9769504070281982,
|
| 267 |
+
"learning_rate": 0.0001929583904892547,
|
| 268 |
+
"loss": 0.1663,
|
| 269 |
+
"step": 320
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 4.074074074074074,
|
| 273 |
+
"grad_norm": 0.6337828636169434,
|
| 274 |
+
"learning_rate": 0.00019204389574759947,
|
| 275 |
+
"loss": 0.159,
|
| 276 |
+
"step": 330
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 4.197530864197531,
|
| 280 |
+
"grad_norm": 0.7267969846725464,
|
| 281 |
+
"learning_rate": 0.00019112940100594422,
|
| 282 |
+
"loss": 0.1347,
|
| 283 |
+
"step": 340
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 4.320987654320987,
|
| 287 |
+
"grad_norm": 1.197481632232666,
|
| 288 |
+
"learning_rate": 0.000190214906264289,
|
| 289 |
+
"loss": 0.1392,
|
| 290 |
+
"step": 350
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 4.444444444444445,
|
| 294 |
+
"grad_norm": 1.4104204177856445,
|
| 295 |
+
"learning_rate": 0.00018930041152263375,
|
| 296 |
+
"loss": 0.1655,
|
| 297 |
+
"step": 360
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 4.567901234567901,
|
| 301 |
+
"grad_norm": 1.115108609199524,
|
| 302 |
+
"learning_rate": 0.00018838591678097853,
|
| 303 |
+
"loss": 0.1792,
|
| 304 |
+
"step": 370
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 4.6913580246913575,
|
| 308 |
+
"grad_norm": 1.7156143188476562,
|
| 309 |
+
"learning_rate": 0.00018747142203932328,
|
| 310 |
+
"loss": 0.1826,
|
| 311 |
+
"step": 380
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 4.814814814814815,
|
| 315 |
+
"grad_norm": 0.7819073796272278,
|
| 316 |
+
"learning_rate": 0.00018655692729766806,
|
| 317 |
+
"loss": 0.2012,
|
| 318 |
+
"step": 390
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 4.938271604938271,
|
| 322 |
+
"grad_norm": 0.9515486359596252,
|
| 323 |
+
"learning_rate": 0.00018564243255601281,
|
| 324 |
+
"loss": 0.1695,
|
| 325 |
+
"step": 400
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 4.938271604938271,
|
| 329 |
+
"eval_accuracy": 0.9243827160493827,
|
| 330 |
+
"eval_f1": 0.8167539267015707,
|
| 331 |
+
"eval_loss": 0.1824890375137329,
|
| 332 |
+
"eval_precision": 0.8198198198198198,
|
| 333 |
+
"eval_recall": 0.8137108792846498,
|
| 334 |
+
"eval_runtime": 2.4365,
|
| 335 |
+
"eval_samples_per_second": 132.979,
|
| 336 |
+
"eval_steps_per_second": 16.828,
|
| 337 |
+
"step": 400
|
| 338 |
+
},
|
| 339 |
+
{
|
| 340 |
+
"epoch": 5.061728395061729,
|
| 341 |
+
"grad_norm": 0.7488430738449097,
|
| 342 |
+
"learning_rate": 0.00018472793781435757,
|
| 343 |
+
"loss": 0.1436,
|
| 344 |
+
"step": 410
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 5.185185185185185,
|
| 348 |
+
"grad_norm": 0.7038572430610657,
|
| 349 |
+
"learning_rate": 0.00018381344307270234,
|
| 350 |
+
"loss": 0.132,
|
| 351 |
+
"step": 420
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 5.308641975308642,
|
| 355 |
+
"grad_norm": 0.8166447877883911,
|
| 356 |
+
"learning_rate": 0.0001828989483310471,
|
| 357 |
+
"loss": 0.1565,
|
| 358 |
+
"step": 430
|
| 359 |
+
},
|
| 360 |
+
{
|
| 361 |
+
"epoch": 5.432098765432099,
|
| 362 |
+
"grad_norm": 1.1276097297668457,
|
| 363 |
+
"learning_rate": 0.00018198445358939187,
|
| 364 |
+
"loss": 0.1642,
|
| 365 |
+
"step": 440
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 5.555555555555555,
|
| 369 |
+
"grad_norm": 0.6070886850357056,
|
| 370 |
+
"learning_rate": 0.00018106995884773663,
|
| 371 |
+
"loss": 0.1353,
|
| 372 |
+
"step": 450
|
| 373 |
+
},
|
| 374 |
+
{
|
| 375 |
+
"epoch": 5.679012345679013,
|
| 376 |
+
"grad_norm": 0.8393697738647461,
|
| 377 |
+
"learning_rate": 0.0001801554641060814,
|
| 378 |
+
"loss": 0.1169,
|
| 379 |
+
"step": 460
|
| 380 |
+
},
|
| 381 |
+
{
|
| 382 |
+
"epoch": 5.802469135802469,
|
| 383 |
+
"grad_norm": 1.5119699239730835,
|
| 384 |
+
"learning_rate": 0.00017924096936442616,
|
| 385 |
+
"loss": 0.1269,
|
| 386 |
+
"step": 470
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 5.925925925925926,
|
| 390 |
+
"grad_norm": 0.9335212707519531,
|
| 391 |
+
"learning_rate": 0.00017832647462277094,
|
| 392 |
+
"loss": 0.1603,
|
| 393 |
+
"step": 480
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 6.049382716049383,
|
| 397 |
+
"grad_norm": 0.6569867730140686,
|
| 398 |
+
"learning_rate": 0.0001774119798811157,
|
| 399 |
+
"loss": 0.123,
|
| 400 |
+
"step": 490
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 6.172839506172839,
|
| 404 |
+
"grad_norm": 0.5314657092094421,
|
| 405 |
+
"learning_rate": 0.00017649748513946044,
|
| 406 |
+
"loss": 0.1204,
|
| 407 |
+
"step": 500
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 6.172839506172839,
|
| 411 |
+
"eval_accuracy": 0.924074074074074,
|
| 412 |
+
"eval_f1": 0.8069073783359497,
|
| 413 |
+
"eval_loss": 0.18580639362335205,
|
| 414 |
+
"eval_precision": 0.8524046434494196,
|
| 415 |
+
"eval_recall": 0.7660208643815202,
|
| 416 |
+
"eval_runtime": 3.2628,
|
| 417 |
+
"eval_samples_per_second": 99.302,
|
| 418 |
+
"eval_steps_per_second": 12.566,
|
| 419 |
+
"step": 500
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.296296296296296,
|
| 423 |
+
"grad_norm": 0.8376681804656982,
|
| 424 |
+
"learning_rate": 0.00017558299039780522,
|
| 425 |
+
"loss": 0.1327,
|
| 426 |
+
"step": 510
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.419753086419753,
|
| 430 |
+
"grad_norm": 0.7419152855873108,
|
| 431 |
+
"learning_rate": 0.00017466849565614997,
|
| 432 |
+
"loss": 0.1355,
|
| 433 |
+
"step": 520
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 6.54320987654321,
|
| 437 |
+
"grad_norm": 1.5098567008972168,
|
| 438 |
+
"learning_rate": 0.00017375400091449475,
|
| 439 |
+
"loss": 0.1223,
|
| 440 |
+
"step": 530
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 6.666666666666667,
|
| 444 |
+
"grad_norm": 0.4890337884426117,
|
| 445 |
+
"learning_rate": 0.0001728395061728395,
|
| 446 |
+
"loss": 0.1307,
|
| 447 |
+
"step": 540
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 6.790123456790123,
|
| 451 |
+
"grad_norm": 0.9887601137161255,
|
| 452 |
+
"learning_rate": 0.00017192501143118428,
|
| 453 |
+
"loss": 0.1118,
|
| 454 |
+
"step": 550
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"epoch": 6.91358024691358,
|
| 458 |
+
"grad_norm": 0.7607052326202393,
|
| 459 |
+
"learning_rate": 0.00017101051668952906,
|
| 460 |
+
"loss": 0.1177,
|
| 461 |
+
"step": 560
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 7.037037037037037,
|
| 465 |
+
"grad_norm": 0.6813036799430847,
|
| 466 |
+
"learning_rate": 0.0001700960219478738,
|
| 467 |
+
"loss": 0.1367,
|
| 468 |
+
"step": 570
|
| 469 |
+
},
|
| 470 |
+
{
|
| 471 |
+
"epoch": 7.160493827160494,
|
| 472 |
+
"grad_norm": 0.5105597972869873,
|
| 473 |
+
"learning_rate": 0.0001691815272062186,
|
| 474 |
+
"loss": 0.1151,
|
| 475 |
+
"step": 580
|
| 476 |
+
},
|
| 477 |
+
{
|
| 478 |
+
"epoch": 7.283950617283951,
|
| 479 |
+
"grad_norm": 0.7654514312744141,
|
| 480 |
+
"learning_rate": 0.00016826703246456334,
|
| 481 |
+
"loss": 0.1483,
|
| 482 |
+
"step": 590
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 7.407407407407407,
|
| 486 |
+
"grad_norm": 0.48647335171699524,
|
| 487 |
+
"learning_rate": 0.00016735253772290812,
|
| 488 |
+
"loss": 0.1268,
|
| 489 |
+
"step": 600
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 7.407407407407407,
|
| 493 |
+
"eval_accuracy": 0.9222222222222223,
|
| 494 |
+
"eval_f1": 0.8061538461538461,
|
| 495 |
+
"eval_loss": 0.18393439054489136,
|
| 496 |
+
"eval_precision": 0.8330683624801272,
|
| 497 |
+
"eval_recall": 0.7809239940387481,
|
| 498 |
+
"eval_runtime": 2.3802,
|
| 499 |
+
"eval_samples_per_second": 136.124,
|
| 500 |
+
"eval_steps_per_second": 17.226,
|
| 501 |
+
"step": 600
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 7.530864197530864,
|
| 505 |
+
"grad_norm": 0.6816988587379456,
|
| 506 |
+
"learning_rate": 0.00016643804298125287,
|
| 507 |
+
"loss": 0.1027,
|
| 508 |
+
"step": 610
|
| 509 |
+
},
|
| 510 |
+
{
|
| 511 |
+
"epoch": 7.654320987654321,
|
| 512 |
+
"grad_norm": 0.8071882128715515,
|
| 513 |
+
"learning_rate": 0.00016552354823959765,
|
| 514 |
+
"loss": 0.1145,
|
| 515 |
+
"step": 620
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 7.777777777777778,
|
| 519 |
+
"grad_norm": 0.532025933265686,
|
| 520 |
+
"learning_rate": 0.0001646090534979424,
|
| 521 |
+
"loss": 0.1179,
|
| 522 |
+
"step": 630
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 7.901234567901234,
|
| 526 |
+
"grad_norm": 0.7129656076431274,
|
| 527 |
+
"learning_rate": 0.00016369455875628715,
|
| 528 |
+
"loss": 0.1304,
|
| 529 |
+
"step": 640
|
| 530 |
+
},
|
| 531 |
+
{
|
| 532 |
+
"epoch": 8.024691358024691,
|
| 533 |
+
"grad_norm": 1.0503504276275635,
|
| 534 |
+
"learning_rate": 0.00016278006401463193,
|
| 535 |
+
"loss": 0.1319,
|
| 536 |
+
"step": 650
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 8.148148148148149,
|
| 540 |
+
"grad_norm": 0.8545478582382202,
|
| 541 |
+
"learning_rate": 0.00016186556927297668,
|
| 542 |
+
"loss": 0.1009,
|
| 543 |
+
"step": 660
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 8.271604938271604,
|
| 547 |
+
"grad_norm": 0.39897826313972473,
|
| 548 |
+
"learning_rate": 0.00016095107453132146,
|
| 549 |
+
"loss": 0.0829,
|
| 550 |
+
"step": 670
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 8.395061728395062,
|
| 554 |
+
"grad_norm": 0.5581258535385132,
|
| 555 |
+
"learning_rate": 0.0001600365797896662,
|
| 556 |
+
"loss": 0.1089,
|
| 557 |
+
"step": 680
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 8.518518518518519,
|
| 561 |
+
"grad_norm": 0.7780981659889221,
|
| 562 |
+
"learning_rate": 0.000159122085048011,
|
| 563 |
+
"loss": 0.116,
|
| 564 |
+
"step": 690
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 8.641975308641975,
|
| 568 |
+
"grad_norm": 0.9378405213356018,
|
| 569 |
+
"learning_rate": 0.00015820759030635574,
|
| 570 |
+
"loss": 0.1089,
|
| 571 |
+
"step": 700
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 8.641975308641975,
|
| 575 |
+
"eval_accuracy": 0.9271604938271605,
|
| 576 |
+
"eval_f1": 0.8214826021180031,
|
| 577 |
+
"eval_loss": 0.18152755498886108,
|
| 578 |
+
"eval_precision": 0.8341013824884793,
|
| 579 |
+
"eval_recall": 0.8092399403874814,
|
| 580 |
+
"eval_runtime": 3.9346,
|
| 581 |
+
"eval_samples_per_second": 82.345,
|
| 582 |
+
"eval_steps_per_second": 10.42,
|
| 583 |
+
"step": 700
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 8.765432098765432,
|
| 587 |
+
"grad_norm": 0.8875961899757385,
|
| 588 |
+
"learning_rate": 0.00015729309556470052,
|
| 589 |
+
"loss": 0.0944,
|
| 590 |
+
"step": 710
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 8.88888888888889,
|
| 594 |
+
"grad_norm": 0.5635790824890137,
|
| 595 |
+
"learning_rate": 0.00015637860082304527,
|
| 596 |
+
"loss": 0.1047,
|
| 597 |
+
"step": 720
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 9.012345679012345,
|
| 601 |
+
"grad_norm": 0.5476753115653992,
|
| 602 |
+
"learning_rate": 0.00015546410608139002,
|
| 603 |
+
"loss": 0.0961,
|
| 604 |
+
"step": 730
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 9.135802469135802,
|
| 608 |
+
"grad_norm": 0.895350456237793,
|
| 609 |
+
"learning_rate": 0.0001545496113397348,
|
| 610 |
+
"loss": 0.097,
|
| 611 |
+
"step": 740
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 9.25925925925926,
|
| 615 |
+
"grad_norm": 0.5731819868087769,
|
| 616 |
+
"learning_rate": 0.00015363511659807956,
|
| 617 |
+
"loss": 0.093,
|
| 618 |
+
"step": 750
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 9.382716049382717,
|
| 622 |
+
"grad_norm": 0.7692080140113831,
|
| 623 |
+
"learning_rate": 0.00015272062185642433,
|
| 624 |
+
"loss": 0.1054,
|
| 625 |
+
"step": 760
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 9.506172839506172,
|
| 629 |
+
"grad_norm": 0.5154545903205872,
|
| 630 |
+
"learning_rate": 0.00015180612711476909,
|
| 631 |
+
"loss": 0.083,
|
| 632 |
+
"step": 770
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 9.62962962962963,
|
| 636 |
+
"grad_norm": 0.9912018179893494,
|
| 637 |
+
"learning_rate": 0.00015089163237311386,
|
| 638 |
+
"loss": 0.079,
|
| 639 |
+
"step": 780
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 9.753086419753087,
|
| 643 |
+
"grad_norm": 0.6497861742973328,
|
| 644 |
+
"learning_rate": 0.00014997713763145862,
|
| 645 |
+
"loss": 0.0813,
|
| 646 |
+
"step": 790
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 9.876543209876543,
|
| 650 |
+
"grad_norm": 0.36326050758361816,
|
| 651 |
+
"learning_rate": 0.0001490626428898034,
|
| 652 |
+
"loss": 0.0863,
|
| 653 |
+
"step": 800
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 9.876543209876543,
|
| 657 |
+
"eval_accuracy": 0.924074074074074,
|
| 658 |
+
"eval_f1": 0.8095975232198143,
|
| 659 |
+
"eval_loss": 0.20340517163276672,
|
| 660 |
+
"eval_precision": 0.8421900161030595,
|
| 661 |
+
"eval_recall": 0.7794336810730254,
|
| 662 |
+
"eval_runtime": 2.5219,
|
| 663 |
+
"eval_samples_per_second": 128.472,
|
| 664 |
+
"eval_steps_per_second": 16.257,
|
| 665 |
+
"step": 800
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"epoch": 10.0,
|
| 669 |
+
"grad_norm": 1.1620311737060547,
|
| 670 |
+
"learning_rate": 0.00014814814814814815,
|
| 671 |
+
"loss": 0.1242,
|
| 672 |
+
"step": 810
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 10.123456790123457,
|
| 676 |
+
"grad_norm": 0.5626382827758789,
|
| 677 |
+
"learning_rate": 0.0001472336534064929,
|
| 678 |
+
"loss": 0.0817,
|
| 679 |
+
"step": 820
|
| 680 |
+
},
|
| 681 |
+
{
|
| 682 |
+
"epoch": 10.246913580246913,
|
| 683 |
+
"grad_norm": 0.4754534363746643,
|
| 684 |
+
"learning_rate": 0.00014631915866483768,
|
| 685 |
+
"loss": 0.0885,
|
| 686 |
+
"step": 830
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 10.37037037037037,
|
| 690 |
+
"grad_norm": 0.8991501927375793,
|
| 691 |
+
"learning_rate": 0.00014540466392318243,
|
| 692 |
+
"loss": 0.1016,
|
| 693 |
+
"step": 840
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 10.493827160493828,
|
| 697 |
+
"grad_norm": 0.4675813317298889,
|
| 698 |
+
"learning_rate": 0.00014449016918152723,
|
| 699 |
+
"loss": 0.0988,
|
| 700 |
+
"step": 850
|
| 701 |
+
},
|
| 702 |
+
{
|
| 703 |
+
"epoch": 10.617283950617283,
|
| 704 |
+
"grad_norm": 0.8737104535102844,
|
| 705 |
+
"learning_rate": 0.00014357567443987199,
|
| 706 |
+
"loss": 0.0838,
|
| 707 |
+
"step": 860
|
| 708 |
+
},
|
| 709 |
+
{
|
| 710 |
+
"epoch": 10.74074074074074,
|
| 711 |
+
"grad_norm": 0.7189881205558777,
|
| 712 |
+
"learning_rate": 0.00014266117969821676,
|
| 713 |
+
"loss": 0.0733,
|
| 714 |
+
"step": 870
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 10.864197530864198,
|
| 718 |
+
"grad_norm": 1.0447670221328735,
|
| 719 |
+
"learning_rate": 0.00014174668495656152,
|
| 720 |
+
"loss": 0.1154,
|
| 721 |
+
"step": 880
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 10.987654320987655,
|
| 725 |
+
"grad_norm": 0.6584729552268982,
|
| 726 |
+
"learning_rate": 0.00014083219021490627,
|
| 727 |
+
"loss": 0.0786,
|
| 728 |
+
"step": 890
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 11.11111111111111,
|
| 732 |
+
"grad_norm": 0.4896683096885681,
|
| 733 |
+
"learning_rate": 0.00013991769547325105,
|
| 734 |
+
"loss": 0.0758,
|
| 735 |
+
"step": 900
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 11.11111111111111,
|
| 739 |
+
"eval_accuracy": 0.929320987654321,
|
| 740 |
+
"eval_f1": 0.8287210172026926,
|
| 741 |
+
"eval_loss": 0.193252295255661,
|
| 742 |
+
"eval_precision": 0.8318318318318318,
|
| 743 |
+
"eval_recall": 0.8256333830104322,
|
| 744 |
+
"eval_runtime": 2.417,
|
| 745 |
+
"eval_samples_per_second": 134.049,
|
| 746 |
+
"eval_steps_per_second": 16.963,
|
| 747 |
+
"step": 900
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 11.234567901234568,
|
| 751 |
+
"grad_norm": 0.4879882037639618,
|
| 752 |
+
"learning_rate": 0.0001390032007315958,
|
| 753 |
+
"loss": 0.0758,
|
| 754 |
+
"step": 910
|
| 755 |
+
},
|
| 756 |
+
{
|
| 757 |
+
"epoch": 11.358024691358025,
|
| 758 |
+
"grad_norm": 0.40832993388175964,
|
| 759 |
+
"learning_rate": 0.00013808870598994058,
|
| 760 |
+
"loss": 0.0659,
|
| 761 |
+
"step": 920
|
| 762 |
+
},
|
| 763 |
+
{
|
| 764 |
+
"epoch": 11.481481481481481,
|
| 765 |
+
"grad_norm": 0.36086103320121765,
|
| 766 |
+
"learning_rate": 0.00013717421124828533,
|
| 767 |
+
"loss": 0.0564,
|
| 768 |
+
"step": 930
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"epoch": 11.604938271604938,
|
| 772 |
+
"grad_norm": 0.4690726101398468,
|
| 773 |
+
"learning_rate": 0.0001362597165066301,
|
| 774 |
+
"loss": 0.0853,
|
| 775 |
+
"step": 940
|
| 776 |
+
},
|
| 777 |
+
{
|
| 778 |
+
"epoch": 11.728395061728396,
|
| 779 |
+
"grad_norm": 0.4532856047153473,
|
| 780 |
+
"learning_rate": 0.00013534522176497486,
|
| 781 |
+
"loss": 0.082,
|
| 782 |
+
"step": 950
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 11.851851851851851,
|
| 786 |
+
"grad_norm": 0.4193461835384369,
|
| 787 |
+
"learning_rate": 0.00013443072702331964,
|
| 788 |
+
"loss": 0.0692,
|
| 789 |
+
"step": 960
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 11.975308641975309,
|
| 793 |
+
"grad_norm": 0.7810946702957153,
|
| 794 |
+
"learning_rate": 0.0001335162322816644,
|
| 795 |
+
"loss": 0.0724,
|
| 796 |
+
"step": 970
|
| 797 |
+
},
|
| 798 |
+
{
|
| 799 |
+
"epoch": 12.098765432098766,
|
| 800 |
+
"grad_norm": 0.49589404463768005,
|
| 801 |
+
"learning_rate": 0.00013260173754000914,
|
| 802 |
+
"loss": 0.058,
|
| 803 |
+
"step": 980
|
| 804 |
+
},
|
| 805 |
+
{
|
| 806 |
+
"epoch": 12.222222222222221,
|
| 807 |
+
"grad_norm": 0.43445077538490295,
|
| 808 |
+
"learning_rate": 0.00013168724279835392,
|
| 809 |
+
"loss": 0.0657,
|
| 810 |
+
"step": 990
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"epoch": 12.345679012345679,
|
| 814 |
+
"grad_norm": 0.4995817542076111,
|
| 815 |
+
"learning_rate": 0.00013077274805669867,
|
| 816 |
+
"loss": 0.0605,
|
| 817 |
+
"step": 1000
|
| 818 |
+
},
|
| 819 |
+
{
|
| 820 |
+
"epoch": 12.345679012345679,
|
| 821 |
+
"eval_accuracy": 0.9302469135802469,
|
| 822 |
+
"eval_f1": 0.8293051359516617,
|
| 823 |
+
"eval_loss": 0.1941601037979126,
|
| 824 |
+
"eval_precision": 0.8407350689127105,
|
| 825 |
+
"eval_recall": 0.8181818181818182,
|
| 826 |
+
"eval_runtime": 2.5849,
|
| 827 |
+
"eval_samples_per_second": 125.345,
|
| 828 |
+
"eval_steps_per_second": 15.862,
|
| 829 |
+
"step": 1000
|
| 830 |
+
},
|
| 831 |
+
{
|
| 832 |
+
"epoch": 12.469135802469136,
|
| 833 |
+
"grad_norm": 0.6132990121841431,
|
| 834 |
+
"learning_rate": 0.00012985825331504345,
|
| 835 |
+
"loss": 0.0618,
|
| 836 |
+
"step": 1010
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 12.592592592592592,
|
| 840 |
+
"grad_norm": 1.0372841358184814,
|
| 841 |
+
"learning_rate": 0.0001289437585733882,
|
| 842 |
+
"loss": 0.0839,
|
| 843 |
+
"step": 1020
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 12.716049382716049,
|
| 847 |
+
"grad_norm": 0.7716336846351624,
|
| 848 |
+
"learning_rate": 0.00012802926383173298,
|
| 849 |
+
"loss": 0.0868,
|
| 850 |
+
"step": 1030
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 12.839506172839506,
|
| 854 |
+
"grad_norm": 0.8065176010131836,
|
| 855 |
+
"learning_rate": 0.00012711476909007773,
|
| 856 |
+
"loss": 0.082,
|
| 857 |
+
"step": 1040
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 12.962962962962964,
|
| 861 |
+
"grad_norm": 0.5325743556022644,
|
| 862 |
+
"learning_rate": 0.0001262002743484225,
|
| 863 |
+
"loss": 0.0621,
|
| 864 |
+
"step": 1050
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 13.08641975308642,
|
| 868 |
+
"grad_norm": 0.595810055732727,
|
| 869 |
+
"learning_rate": 0.00012528577960676726,
|
| 870 |
+
"loss": 0.0704,
|
| 871 |
+
"step": 1060
|
| 872 |
+
},
|
| 873 |
+
{
|
| 874 |
+
"epoch": 13.209876543209877,
|
| 875 |
+
"grad_norm": 0.43285462260246277,
|
| 876 |
+
"learning_rate": 0.00012437128486511201,
|
| 877 |
+
"loss": 0.0731,
|
| 878 |
+
"step": 1070
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 13.333333333333334,
|
| 882 |
+
"grad_norm": 0.6610433459281921,
|
| 883 |
+
"learning_rate": 0.0001234567901234568,
|
| 884 |
+
"loss": 0.0883,
|
| 885 |
+
"step": 1080
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 13.45679012345679,
|
| 889 |
+
"grad_norm": 0.4961569011211395,
|
| 890 |
+
"learning_rate": 0.00012254229538180154,
|
| 891 |
+
"loss": 0.0628,
|
| 892 |
+
"step": 1090
|
| 893 |
+
},
|
| 894 |
+
{
|
| 895 |
+
"epoch": 13.580246913580247,
|
| 896 |
+
"grad_norm": 1.3226776123046875,
|
| 897 |
+
"learning_rate": 0.00012162780064014631,
|
| 898 |
+
"loss": 0.0726,
|
| 899 |
+
"step": 1100
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 13.580246913580247,
|
| 903 |
+
"eval_accuracy": 0.9271604938271605,
|
| 904 |
+
"eval_f1": 0.8238805970149253,
|
| 905 |
+
"eval_loss": 0.19768880307674408,
|
| 906 |
+
"eval_precision": 0.8251121076233184,
|
| 907 |
+
"eval_recall": 0.8226527570789866,
|
| 908 |
+
"eval_runtime": 2.5461,
|
| 909 |
+
"eval_samples_per_second": 127.251,
|
| 910 |
+
"eval_steps_per_second": 16.103,
|
| 911 |
+
"step": 1100
|
| 912 |
+
},
|
| 913 |
+
{
|
| 914 |
+
"epoch": 13.703703703703704,
|
| 915 |
+
"grad_norm": 0.36743324995040894,
|
| 916 |
+
"learning_rate": 0.00012071330589849108,
|
| 917 |
+
"loss": 0.0654,
|
| 918 |
+
"step": 1110
|
| 919 |
+
},
|
| 920 |
+
{
|
| 921 |
+
"epoch": 13.82716049382716,
|
| 922 |
+
"grad_norm": 0.251310259103775,
|
| 923 |
+
"learning_rate": 0.00011979881115683584,
|
| 924 |
+
"loss": 0.061,
|
| 925 |
+
"step": 1120
|
| 926 |
+
},
|
| 927 |
+
{
|
| 928 |
+
"epoch": 13.950617283950617,
|
| 929 |
+
"grad_norm": 0.29951387643814087,
|
| 930 |
+
"learning_rate": 0.0001188843164151806,
|
| 931 |
+
"loss": 0.0623,
|
| 932 |
+
"step": 1130
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 14.074074074074074,
|
| 936 |
+
"grad_norm": 0.18914303183555603,
|
| 937 |
+
"learning_rate": 0.0001179698216735254,
|
| 938 |
+
"loss": 0.0481,
|
| 939 |
+
"step": 1140
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 14.197530864197532,
|
| 943 |
+
"grad_norm": 0.5847985744476318,
|
| 944 |
+
"learning_rate": 0.00011705532693187016,
|
| 945 |
+
"loss": 0.0665,
|
| 946 |
+
"step": 1150
|
| 947 |
+
},
|
| 948 |
+
{
|
| 949 |
+
"epoch": 14.320987654320987,
|
| 950 |
+
"grad_norm": 0.8701656460762024,
|
| 951 |
+
"learning_rate": 0.00011614083219021491,
|
| 952 |
+
"loss": 0.0717,
|
| 953 |
+
"step": 1160
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"epoch": 14.444444444444445,
|
| 957 |
+
"grad_norm": 0.7420238256454468,
|
| 958 |
+
"learning_rate": 0.00011522633744855968,
|
| 959 |
+
"loss": 0.075,
|
| 960 |
+
"step": 1170
|
| 961 |
+
},
|
| 962 |
+
{
|
| 963 |
+
"epoch": 14.567901234567902,
|
| 964 |
+
"grad_norm": 0.6891819834709167,
|
| 965 |
+
"learning_rate": 0.00011431184270690445,
|
| 966 |
+
"loss": 0.0625,
|
| 967 |
+
"step": 1180
|
| 968 |
+
},
|
| 969 |
+
{
|
| 970 |
+
"epoch": 14.691358024691358,
|
| 971 |
+
"grad_norm": 0.6125108003616333,
|
| 972 |
+
"learning_rate": 0.00011339734796524921,
|
| 973 |
+
"loss": 0.0635,
|
| 974 |
+
"step": 1190
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 14.814814814814815,
|
| 978 |
+
"grad_norm": 0.6992429494857788,
|
| 979 |
+
"learning_rate": 0.00011248285322359398,
|
| 980 |
+
"loss": 0.0637,
|
| 981 |
+
"step": 1200
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 14.814814814814815,
|
| 985 |
+
"eval_accuracy": 0.9311728395061728,
|
| 986 |
+
"eval_f1": 0.8324567993989481,
|
| 987 |
+
"eval_loss": 0.2040136605501175,
|
| 988 |
+
"eval_precision": 0.8393939393939394,
|
| 989 |
+
"eval_recall": 0.8256333830104322,
|
| 990 |
+
"eval_runtime": 2.5198,
|
| 991 |
+
"eval_samples_per_second": 128.58,
|
| 992 |
+
"eval_steps_per_second": 16.271,
|
| 993 |
+
"step": 1200
|
| 994 |
+
},
|
| 995 |
+
{
|
| 996 |
+
"epoch": 14.938271604938272,
|
| 997 |
+
"grad_norm": 0.37562182545661926,
|
| 998 |
+
"learning_rate": 0.00011156835848193874,
|
| 999 |
+
"loss": 0.0494,
|
| 1000 |
+
"step": 1210
|
| 1001 |
+
},
|
| 1002 |
+
{
|
| 1003 |
+
"epoch": 15.061728395061728,
|
| 1004 |
+
"grad_norm": 0.6514123678207397,
|
| 1005 |
+
"learning_rate": 0.0001106538637402835,
|
| 1006 |
+
"loss": 0.0513,
|
| 1007 |
+
"step": 1220
|
| 1008 |
+
},
|
| 1009 |
+
{
|
| 1010 |
+
"epoch": 15.185185185185185,
|
| 1011 |
+
"grad_norm": 0.3561560809612274,
|
| 1012 |
+
"learning_rate": 0.00010973936899862827,
|
| 1013 |
+
"loss": 0.069,
|
| 1014 |
+
"step": 1230
|
| 1015 |
+
},
|
| 1016 |
+
{
|
| 1017 |
+
"epoch": 15.308641975308642,
|
| 1018 |
+
"grad_norm": 1.3583102226257324,
|
| 1019 |
+
"learning_rate": 0.00010882487425697304,
|
| 1020 |
+
"loss": 0.0714,
|
| 1021 |
+
"step": 1240
|
| 1022 |
+
},
|
| 1023 |
+
{
|
| 1024 |
+
"epoch": 15.432098765432098,
|
| 1025 |
+
"grad_norm": 0.3257612884044647,
|
| 1026 |
+
"learning_rate": 0.00010791037951531779,
|
| 1027 |
+
"loss": 0.0494,
|
| 1028 |
+
"step": 1250
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 15.555555555555555,
|
| 1032 |
+
"grad_norm": 0.4650997519493103,
|
| 1033 |
+
"learning_rate": 0.00010699588477366255,
|
| 1034 |
+
"loss": 0.0464,
|
| 1035 |
+
"step": 1260
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 15.679012345679013,
|
| 1039 |
+
"grad_norm": 0.7076647281646729,
|
| 1040 |
+
"learning_rate": 0.00010608139003200732,
|
| 1041 |
+
"loss": 0.0575,
|
| 1042 |
+
"step": 1270
|
| 1043 |
+
},
|
| 1044 |
+
{
|
| 1045 |
+
"epoch": 15.802469135802468,
|
| 1046 |
+
"grad_norm": 0.3548375070095062,
|
| 1047 |
+
"learning_rate": 0.00010516689529035208,
|
| 1048 |
+
"loss": 0.0605,
|
| 1049 |
+
"step": 1280
|
| 1050 |
+
},
|
| 1051 |
+
{
|
| 1052 |
+
"epoch": 15.925925925925926,
|
| 1053 |
+
"grad_norm": 0.3160780966281891,
|
| 1054 |
+
"learning_rate": 0.00010425240054869685,
|
| 1055 |
+
"loss": 0.0695,
|
| 1056 |
+
"step": 1290
|
| 1057 |
+
},
|
| 1058 |
+
{
|
| 1059 |
+
"epoch": 16.049382716049383,
|
| 1060 |
+
"grad_norm": 0.43579068779945374,
|
| 1061 |
+
"learning_rate": 0.00010333790580704161,
|
| 1062 |
+
"loss": 0.0549,
|
| 1063 |
+
"step": 1300
|
| 1064 |
+
},
|
| 1065 |
+
{
|
| 1066 |
+
"epoch": 16.049382716049383,
|
| 1067 |
+
"eval_accuracy": 0.9308641975308642,
|
| 1068 |
+
"eval_f1": 0.8325859491778774,
|
| 1069 |
+
"eval_loss": 0.21728986501693726,
|
| 1070 |
+
"eval_precision": 0.8350824587706147,
|
| 1071 |
+
"eval_recall": 0.8301043219076006,
|
| 1072 |
+
"eval_runtime": 3.2646,
|
| 1073 |
+
"eval_samples_per_second": 99.247,
|
| 1074 |
+
"eval_steps_per_second": 12.559,
|
| 1075 |
+
"step": 1300
|
| 1076 |
+
},
|
| 1077 |
+
{
|
| 1078 |
+
"epoch": 16.17283950617284,
|
| 1079 |
+
"grad_norm": 0.6600210070610046,
|
| 1080 |
+
"learning_rate": 0.00010242341106538638,
|
| 1081 |
+
"loss": 0.0348,
|
| 1082 |
+
"step": 1310
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 16.296296296296298,
|
| 1086 |
+
"grad_norm": 0.45880362391471863,
|
| 1087 |
+
"learning_rate": 0.00010150891632373114,
|
| 1088 |
+
"loss": 0.053,
|
| 1089 |
+
"step": 1320
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 16.419753086419753,
|
| 1093 |
+
"grad_norm": 0.25807011127471924,
|
| 1094 |
+
"learning_rate": 0.00010059442158207591,
|
| 1095 |
+
"loss": 0.0534,
|
| 1096 |
+
"step": 1330
|
| 1097 |
+
},
|
| 1098 |
+
{
|
| 1099 |
+
"epoch": 16.54320987654321,
|
| 1100 |
+
"grad_norm": 0.7079629898071289,
|
| 1101 |
+
"learning_rate": 9.967992684042066e-05,
|
| 1102 |
+
"loss": 0.0519,
|
| 1103 |
+
"step": 1340
|
| 1104 |
+
},
|
| 1105 |
+
{
|
| 1106 |
+
"epoch": 16.666666666666668,
|
| 1107 |
+
"grad_norm": 0.5823590159416199,
|
| 1108 |
+
"learning_rate": 9.876543209876543e-05,
|
| 1109 |
+
"loss": 0.0532,
|
| 1110 |
+
"step": 1350
|
| 1111 |
+
},
|
| 1112 |
+
{
|
| 1113 |
+
"epoch": 16.790123456790123,
|
| 1114 |
+
"grad_norm": 0.8557542562484741,
|
| 1115 |
+
"learning_rate": 9.78509373571102e-05,
|
| 1116 |
+
"loss": 0.0554,
|
| 1117 |
+
"step": 1360
|
| 1118 |
+
},
|
| 1119 |
+
{
|
| 1120 |
+
"epoch": 16.91358024691358,
|
| 1121 |
+
"grad_norm": 0.35440272092819214,
|
| 1122 |
+
"learning_rate": 9.693644261545497e-05,
|
| 1123 |
+
"loss": 0.0447,
|
| 1124 |
+
"step": 1370
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 17.037037037037038,
|
| 1128 |
+
"grad_norm": 0.7405946850776672,
|
| 1129 |
+
"learning_rate": 9.602194787379974e-05,
|
| 1130 |
+
"loss": 0.0502,
|
| 1131 |
+
"step": 1380
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 17.160493827160494,
|
| 1135 |
+
"grad_norm": 0.350297212600708,
|
| 1136 |
+
"learning_rate": 9.51074531321445e-05,
|
| 1137 |
+
"loss": 0.039,
|
| 1138 |
+
"step": 1390
|
| 1139 |
+
},
|
| 1140 |
+
{
|
| 1141 |
+
"epoch": 17.28395061728395,
|
| 1142 |
+
"grad_norm": 0.6312181353569031,
|
| 1143 |
+
"learning_rate": 9.419295839048927e-05,
|
| 1144 |
+
"loss": 0.0566,
|
| 1145 |
+
"step": 1400
|
| 1146 |
+
},
|
| 1147 |
+
{
|
| 1148 |
+
"epoch": 17.28395061728395,
|
| 1149 |
+
"eval_accuracy": 0.9314814814814815,
|
| 1150 |
+
"eval_f1": 0.8353115727002968,
|
| 1151 |
+
"eval_loss": 0.20715734362602234,
|
| 1152 |
+
"eval_precision": 0.8316100443131462,
|
| 1153 |
+
"eval_recall": 0.8390461997019374,
|
| 1154 |
+
"eval_runtime": 2.7588,
|
| 1155 |
+
"eval_samples_per_second": 117.44,
|
| 1156 |
+
"eval_steps_per_second": 14.861,
|
| 1157 |
+
"step": 1400
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 17.40740740740741,
|
| 1161 |
+
"grad_norm": 0.17453396320343018,
|
| 1162 |
+
"learning_rate": 9.327846364883403e-05,
|
| 1163 |
+
"loss": 0.0498,
|
| 1164 |
+
"step": 1410
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 17.530864197530864,
|
| 1168 |
+
"grad_norm": 0.36732375621795654,
|
| 1169 |
+
"learning_rate": 9.236396890717878e-05,
|
| 1170 |
+
"loss": 0.0549,
|
| 1171 |
+
"step": 1420
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 17.65432098765432,
|
| 1175 |
+
"grad_norm": 0.7602197527885437,
|
| 1176 |
+
"learning_rate": 9.144947416552355e-05,
|
| 1177 |
+
"loss": 0.0447,
|
| 1178 |
+
"step": 1430
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 17.77777777777778,
|
| 1182 |
+
"grad_norm": 0.8259577751159668,
|
| 1183 |
+
"learning_rate": 9.053497942386831e-05,
|
| 1184 |
+
"loss": 0.0584,
|
| 1185 |
+
"step": 1440
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 17.901234567901234,
|
| 1189 |
+
"grad_norm": 0.43447449803352356,
|
| 1190 |
+
"learning_rate": 8.962048468221308e-05,
|
| 1191 |
+
"loss": 0.0359,
|
| 1192 |
+
"step": 1450
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 18.02469135802469,
|
| 1196 |
+
"grad_norm": 0.6444792151451111,
|
| 1197 |
+
"learning_rate": 8.870598994055784e-05,
|
| 1198 |
+
"loss": 0.0437,
|
| 1199 |
+
"step": 1460
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 18.14814814814815,
|
| 1203 |
+
"grad_norm": 0.9808084964752197,
|
| 1204 |
+
"learning_rate": 8.779149519890261e-05,
|
| 1205 |
+
"loss": 0.0428,
|
| 1206 |
+
"step": 1470
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 18.271604938271604,
|
| 1210 |
+
"grad_norm": 0.4970827102661133,
|
| 1211 |
+
"learning_rate": 8.687700045724737e-05,
|
| 1212 |
+
"loss": 0.0325,
|
| 1213 |
+
"step": 1480
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 18.395061728395063,
|
| 1217 |
+
"grad_norm": 0.3154457211494446,
|
| 1218 |
+
"learning_rate": 8.596250571559214e-05,
|
| 1219 |
+
"loss": 0.0443,
|
| 1220 |
+
"step": 1490
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 18.51851851851852,
|
| 1224 |
+
"grad_norm": 0.38939017057418823,
|
| 1225 |
+
"learning_rate": 8.50480109739369e-05,
|
| 1226 |
+
"loss": 0.0491,
|
| 1227 |
+
"step": 1500
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 18.51851851851852,
|
| 1231 |
+
"eval_accuracy": 0.928395061728395,
|
| 1232 |
+
"eval_f1": 0.824773413897281,
|
| 1233 |
+
"eval_loss": 0.2156645506620407,
|
| 1234 |
+
"eval_precision": 0.8361408882082695,
|
| 1235 |
+
"eval_recall": 0.8137108792846498,
|
| 1236 |
+
"eval_runtime": 2.3248,
|
| 1237 |
+
"eval_samples_per_second": 139.37,
|
| 1238 |
+
"eval_steps_per_second": 17.636,
|
| 1239 |
+
"step": 1500
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 18.641975308641975,
|
| 1243 |
+
"grad_norm": 0.7808548808097839,
|
| 1244 |
+
"learning_rate": 8.413351623228167e-05,
|
| 1245 |
+
"loss": 0.0342,
|
| 1246 |
+
"step": 1510
|
| 1247 |
+
},
|
| 1248 |
+
{
|
| 1249 |
+
"epoch": 18.765432098765434,
|
| 1250 |
+
"grad_norm": 0.11940323561429977,
|
| 1251 |
+
"learning_rate": 8.321902149062643e-05,
|
| 1252 |
+
"loss": 0.0528,
|
| 1253 |
+
"step": 1520
|
| 1254 |
+
},
|
| 1255 |
+
{
|
| 1256 |
+
"epoch": 18.88888888888889,
|
| 1257 |
+
"grad_norm": 0.6686835885047913,
|
| 1258 |
+
"learning_rate": 8.23045267489712e-05,
|
| 1259 |
+
"loss": 0.037,
|
| 1260 |
+
"step": 1530
|
| 1261 |
+
},
|
| 1262 |
+
{
|
| 1263 |
+
"epoch": 19.012345679012345,
|
| 1264 |
+
"grad_norm": 0.21024620532989502,
|
| 1265 |
+
"learning_rate": 8.139003200731597e-05,
|
| 1266 |
+
"loss": 0.0505,
|
| 1267 |
+
"step": 1540
|
| 1268 |
+
},
|
| 1269 |
+
{
|
| 1270 |
+
"epoch": 19.135802469135804,
|
| 1271 |
+
"grad_norm": 0.5132074952125549,
|
| 1272 |
+
"learning_rate": 8.047553726566073e-05,
|
| 1273 |
+
"loss": 0.0404,
|
| 1274 |
+
"step": 1550
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 19.25925925925926,
|
| 1278 |
+
"grad_norm": 0.32788872718811035,
|
| 1279 |
+
"learning_rate": 7.95610425240055e-05,
|
| 1280 |
+
"loss": 0.0513,
|
| 1281 |
+
"step": 1560
|
| 1282 |
+
},
|
| 1283 |
+
{
|
| 1284 |
+
"epoch": 19.382716049382715,
|
| 1285 |
+
"grad_norm": 0.31631699204444885,
|
| 1286 |
+
"learning_rate": 7.864654778235026e-05,
|
| 1287 |
+
"loss": 0.0343,
|
| 1288 |
+
"step": 1570
|
| 1289 |
+
},
|
| 1290 |
+
{
|
| 1291 |
+
"epoch": 19.506172839506174,
|
| 1292 |
+
"grad_norm": 0.6206240057945251,
|
| 1293 |
+
"learning_rate": 7.773205304069501e-05,
|
| 1294 |
+
"loss": 0.044,
|
| 1295 |
+
"step": 1580
|
| 1296 |
+
},
|
| 1297 |
+
{
|
| 1298 |
+
"epoch": 19.62962962962963,
|
| 1299 |
+
"grad_norm": 0.46290040016174316,
|
| 1300 |
+
"learning_rate": 7.681755829903978e-05,
|
| 1301 |
+
"loss": 0.0456,
|
| 1302 |
+
"step": 1590
|
| 1303 |
+
},
|
| 1304 |
+
{
|
| 1305 |
+
"epoch": 19.753086419753085,
|
| 1306 |
+
"grad_norm": 0.5506296753883362,
|
| 1307 |
+
"learning_rate": 7.590306355738454e-05,
|
| 1308 |
+
"loss": 0.0496,
|
| 1309 |
+
"step": 1600
|
| 1310 |
+
},
|
| 1311 |
+
{
|
| 1312 |
+
"epoch": 19.753086419753085,
|
| 1313 |
+
"eval_accuracy": 0.9351851851851852,
|
| 1314 |
+
"eval_f1": 0.84375,
|
| 1315 |
+
"eval_loss": 0.19679902493953705,
|
| 1316 |
+
"eval_precision": 0.8424962852897474,
|
| 1317 |
+
"eval_recall": 0.8450074515648286,
|
| 1318 |
+
"eval_runtime": 3.3713,
|
| 1319 |
+
"eval_samples_per_second": 96.105,
|
| 1320 |
+
"eval_steps_per_second": 12.161,
|
| 1321 |
+
"step": 1600
|
| 1322 |
+
},
|
| 1323 |
+
{
|
| 1324 |
+
"epoch": 19.876543209876544,
|
| 1325 |
+
"grad_norm": 0.48592284321784973,
|
| 1326 |
+
"learning_rate": 7.498856881572931e-05,
|
| 1327 |
+
"loss": 0.042,
|
| 1328 |
+
"step": 1610
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 20.0,
|
| 1332 |
+
"grad_norm": 0.3968632221221924,
|
| 1333 |
+
"learning_rate": 7.407407407407407e-05,
|
| 1334 |
+
"loss": 0.0434,
|
| 1335 |
+
"step": 1620
|
| 1336 |
+
},
|
| 1337 |
+
{
|
| 1338 |
+
"epoch": 20.123456790123456,
|
| 1339 |
+
"grad_norm": 0.16500796377658844,
|
| 1340 |
+
"learning_rate": 7.315957933241884e-05,
|
| 1341 |
+
"loss": 0.0479,
|
| 1342 |
+
"step": 1630
|
| 1343 |
+
},
|
| 1344 |
+
{
|
| 1345 |
+
"epoch": 20.246913580246915,
|
| 1346 |
+
"grad_norm": 0.12196658551692963,
|
| 1347 |
+
"learning_rate": 7.224508459076362e-05,
|
| 1348 |
+
"loss": 0.0309,
|
| 1349 |
+
"step": 1640
|
| 1350 |
+
},
|
| 1351 |
+
{
|
| 1352 |
+
"epoch": 20.37037037037037,
|
| 1353 |
+
"grad_norm": 0.46755895018577576,
|
| 1354 |
+
"learning_rate": 7.133058984910838e-05,
|
| 1355 |
+
"loss": 0.0448,
|
| 1356 |
+
"step": 1650
|
| 1357 |
+
},
|
| 1358 |
+
{
|
| 1359 |
+
"epoch": 20.493827160493826,
|
| 1360 |
+
"grad_norm": 0.26269474625587463,
|
| 1361 |
+
"learning_rate": 7.041609510745313e-05,
|
| 1362 |
+
"loss": 0.0281,
|
| 1363 |
+
"step": 1660
|
| 1364 |
+
},
|
| 1365 |
+
{
|
| 1366 |
+
"epoch": 20.617283950617285,
|
| 1367 |
+
"grad_norm": 0.28748786449432373,
|
| 1368 |
+
"learning_rate": 6.95016003657979e-05,
|
| 1369 |
+
"loss": 0.0419,
|
| 1370 |
+
"step": 1670
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 20.74074074074074,
|
| 1374 |
+
"grad_norm": 0.48799601197242737,
|
| 1375 |
+
"learning_rate": 6.858710562414266e-05,
|
| 1376 |
+
"loss": 0.0408,
|
| 1377 |
+
"step": 1680
|
| 1378 |
+
},
|
| 1379 |
+
{
|
| 1380 |
+
"epoch": 20.864197530864196,
|
| 1381 |
+
"grad_norm": 0.259390264749527,
|
| 1382 |
+
"learning_rate": 6.767261088248743e-05,
|
| 1383 |
+
"loss": 0.0285,
|
| 1384 |
+
"step": 1690
|
| 1385 |
+
},
|
| 1386 |
+
{
|
| 1387 |
+
"epoch": 20.987654320987655,
|
| 1388 |
+
"grad_norm": 0.29502254724502563,
|
| 1389 |
+
"learning_rate": 6.67581161408322e-05,
|
| 1390 |
+
"loss": 0.0317,
|
| 1391 |
+
"step": 1700
|
| 1392 |
+
},
|
| 1393 |
+
{
|
| 1394 |
+
"epoch": 20.987654320987655,
|
| 1395 |
+
"eval_accuracy": 0.9361111111111111,
|
| 1396 |
+
"eval_f1": 0.8456375838926175,
|
| 1397 |
+
"eval_loss": 0.2053374946117401,
|
| 1398 |
+
"eval_precision": 0.8462686567164179,
|
| 1399 |
+
"eval_recall": 0.8450074515648286,
|
| 1400 |
+
"eval_runtime": 2.8712,
|
| 1401 |
+
"eval_samples_per_second": 112.844,
|
| 1402 |
+
"eval_steps_per_second": 14.28,
|
| 1403 |
+
"step": 1700
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 21.11111111111111,
|
| 1407 |
+
"grad_norm": 0.5067708492279053,
|
| 1408 |
+
"learning_rate": 6.584362139917696e-05,
|
| 1409 |
+
"loss": 0.0398,
|
| 1410 |
+
"step": 1710
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 21.234567901234566,
|
| 1414 |
+
"grad_norm": 0.3558381199836731,
|
| 1415 |
+
"learning_rate": 6.492912665752173e-05,
|
| 1416 |
+
"loss": 0.0396,
|
| 1417 |
+
"step": 1720
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 21.358024691358025,
|
| 1421 |
+
"grad_norm": 0.1958431601524353,
|
| 1422 |
+
"learning_rate": 6.401463191586649e-05,
|
| 1423 |
+
"loss": 0.0284,
|
| 1424 |
+
"step": 1730
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 21.48148148148148,
|
| 1428 |
+
"grad_norm": 0.12868855893611908,
|
| 1429 |
+
"learning_rate": 6.310013717421126e-05,
|
| 1430 |
+
"loss": 0.0249,
|
| 1431 |
+
"step": 1740
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 21.604938271604937,
|
| 1435 |
+
"grad_norm": 0.1328994780778885,
|
| 1436 |
+
"learning_rate": 6.218564243255601e-05,
|
| 1437 |
+
"loss": 0.0336,
|
| 1438 |
+
"step": 1750
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 21.728395061728396,
|
| 1442 |
+
"grad_norm": 0.2639734745025635,
|
| 1443 |
+
"learning_rate": 6.127114769090077e-05,
|
| 1444 |
+
"loss": 0.0281,
|
| 1445 |
+
"step": 1760
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 21.85185185185185,
|
| 1449 |
+
"grad_norm": 0.20830489695072174,
|
| 1450 |
+
"learning_rate": 6.035665294924554e-05,
|
| 1451 |
+
"loss": 0.0371,
|
| 1452 |
+
"step": 1770
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 21.97530864197531,
|
| 1456 |
+
"grad_norm": 0.3154972195625305,
|
| 1457 |
+
"learning_rate": 5.94421582075903e-05,
|
| 1458 |
+
"loss": 0.0449,
|
| 1459 |
+
"step": 1780
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 22.098765432098766,
|
| 1463 |
+
"grad_norm": 0.15233543515205383,
|
| 1464 |
+
"learning_rate": 5.852766346593508e-05,
|
| 1465 |
+
"loss": 0.033,
|
| 1466 |
+
"step": 1790
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 22.22222222222222,
|
| 1470 |
+
"grad_norm": 0.2932221293449402,
|
| 1471 |
+
"learning_rate": 5.761316872427984e-05,
|
| 1472 |
+
"loss": 0.0326,
|
| 1473 |
+
"step": 1800
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 22.22222222222222,
|
| 1477 |
+
"eval_accuracy": 0.9305555555555556,
|
| 1478 |
+
"eval_f1": 0.8288973384030418,
|
| 1479 |
+
"eval_loss": 0.21322904527187347,
|
| 1480 |
+
"eval_precision": 0.8462732919254659,
|
| 1481 |
+
"eval_recall": 0.812220566318927,
|
| 1482 |
+
"eval_runtime": 2.3717,
|
| 1483 |
+
"eval_samples_per_second": 136.612,
|
| 1484 |
+
"eval_steps_per_second": 17.287,
|
| 1485 |
+
"step": 1800
|
| 1486 |
+
},
|
| 1487 |
+
{
|
| 1488 |
+
"epoch": 22.34567901234568,
|
| 1489 |
+
"grad_norm": 0.2980441153049469,
|
| 1490 |
+
"learning_rate": 5.6698673982624605e-05,
|
| 1491 |
+
"loss": 0.0293,
|
| 1492 |
+
"step": 1810
|
| 1493 |
+
},
|
| 1494 |
+
{
|
| 1495 |
+
"epoch": 22.469135802469136,
|
| 1496 |
+
"grad_norm": 0.28522634506225586,
|
| 1497 |
+
"learning_rate": 5.578417924096937e-05,
|
| 1498 |
+
"loss": 0.0334,
|
| 1499 |
+
"step": 1820
|
| 1500 |
+
},
|
| 1501 |
+
{
|
| 1502 |
+
"epoch": 22.59259259259259,
|
| 1503 |
+
"grad_norm": 0.40435144305229187,
|
| 1504 |
+
"learning_rate": 5.4869684499314136e-05,
|
| 1505 |
+
"loss": 0.0377,
|
| 1506 |
+
"step": 1830
|
| 1507 |
+
},
|
| 1508 |
+
{
|
| 1509 |
+
"epoch": 22.71604938271605,
|
| 1510 |
+
"grad_norm": 0.43217939138412476,
|
| 1511 |
+
"learning_rate": 5.3955189757658894e-05,
|
| 1512 |
+
"loss": 0.0288,
|
| 1513 |
+
"step": 1840
|
| 1514 |
+
},
|
| 1515 |
+
{
|
| 1516 |
+
"epoch": 22.839506172839506,
|
| 1517 |
+
"grad_norm": 0.30265820026397705,
|
| 1518 |
+
"learning_rate": 5.304069501600366e-05,
|
| 1519 |
+
"loss": 0.0359,
|
| 1520 |
+
"step": 1850
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 22.962962962962962,
|
| 1524 |
+
"grad_norm": 0.0610734224319458,
|
| 1525 |
+
"learning_rate": 5.2126200274348424e-05,
|
| 1526 |
+
"loss": 0.037,
|
| 1527 |
+
"step": 1860
|
| 1528 |
+
},
|
| 1529 |
+
{
|
| 1530 |
+
"epoch": 23.08641975308642,
|
| 1531 |
+
"grad_norm": 0.5076401233673096,
|
| 1532 |
+
"learning_rate": 5.121170553269319e-05,
|
| 1533 |
+
"loss": 0.0344,
|
| 1534 |
+
"step": 1870
|
| 1535 |
+
},
|
| 1536 |
+
{
|
| 1537 |
+
"epoch": 23.209876543209877,
|
| 1538 |
+
"grad_norm": 0.26529771089553833,
|
| 1539 |
+
"learning_rate": 5.0297210791037955e-05,
|
| 1540 |
+
"loss": 0.0321,
|
| 1541 |
+
"step": 1880
|
| 1542 |
+
},
|
| 1543 |
+
{
|
| 1544 |
+
"epoch": 23.333333333333332,
|
| 1545 |
+
"grad_norm": 0.46441182494163513,
|
| 1546 |
+
"learning_rate": 4.938271604938271e-05,
|
| 1547 |
+
"loss": 0.0255,
|
| 1548 |
+
"step": 1890
|
| 1549 |
+
},
|
| 1550 |
+
{
|
| 1551 |
+
"epoch": 23.45679012345679,
|
| 1552 |
+
"grad_norm": 0.32864171266555786,
|
| 1553 |
+
"learning_rate": 4.8468221307727485e-05,
|
| 1554 |
+
"loss": 0.0307,
|
| 1555 |
+
"step": 1900
|
| 1556 |
+
},
|
| 1557 |
+
{
|
| 1558 |
+
"epoch": 23.45679012345679,
|
| 1559 |
+
"eval_accuracy": 0.9308641975308642,
|
| 1560 |
+
"eval_f1": 0.8330849478390462,
|
| 1561 |
+
"eval_loss": 0.21188341081142426,
|
| 1562 |
+
"eval_precision": 0.8330849478390462,
|
| 1563 |
+
"eval_recall": 0.8330849478390462,
|
| 1564 |
+
"eval_runtime": 2.4849,
|
| 1565 |
+
"eval_samples_per_second": 130.385,
|
| 1566 |
+
"eval_steps_per_second": 16.499,
|
| 1567 |
+
"step": 1900
|
| 1568 |
+
},
|
| 1569 |
+
{
|
| 1570 |
+
"epoch": 23.580246913580247,
|
| 1571 |
+
"grad_norm": 0.06886278837919235,
|
| 1572 |
+
"learning_rate": 4.755372656607225e-05,
|
| 1573 |
+
"loss": 0.0283,
|
| 1574 |
+
"step": 1910
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"epoch": 23.703703703703702,
|
| 1578 |
+
"grad_norm": 0.15195724368095398,
|
| 1579 |
+
"learning_rate": 4.6639231824417016e-05,
|
| 1580 |
+
"loss": 0.0274,
|
| 1581 |
+
"step": 1920
|
| 1582 |
+
},
|
| 1583 |
+
{
|
| 1584 |
+
"epoch": 23.82716049382716,
|
| 1585 |
+
"grad_norm": 0.2511798143386841,
|
| 1586 |
+
"learning_rate": 4.5724737082761774e-05,
|
| 1587 |
+
"loss": 0.0254,
|
| 1588 |
+
"step": 1930
|
| 1589 |
+
},
|
| 1590 |
+
{
|
| 1591 |
+
"epoch": 23.950617283950617,
|
| 1592 |
+
"grad_norm": 0.41059139370918274,
|
| 1593 |
+
"learning_rate": 4.481024234110654e-05,
|
| 1594 |
+
"loss": 0.0322,
|
| 1595 |
+
"step": 1940
|
| 1596 |
+
},
|
| 1597 |
+
{
|
| 1598 |
+
"epoch": 24.074074074074073,
|
| 1599 |
+
"grad_norm": 0.22714923322200775,
|
| 1600 |
+
"learning_rate": 4.3895747599451304e-05,
|
| 1601 |
+
"loss": 0.0245,
|
| 1602 |
+
"step": 1950
|
| 1603 |
+
},
|
| 1604 |
+
{
|
| 1605 |
+
"epoch": 24.19753086419753,
|
| 1606 |
+
"grad_norm": 0.4791244566440582,
|
| 1607 |
+
"learning_rate": 4.298125285779607e-05,
|
| 1608 |
+
"loss": 0.037,
|
| 1609 |
+
"step": 1960
|
| 1610 |
+
},
|
| 1611 |
+
{
|
| 1612 |
+
"epoch": 24.320987654320987,
|
| 1613 |
+
"grad_norm": 0.24066051840782166,
|
| 1614 |
+
"learning_rate": 4.2066758116140835e-05,
|
| 1615 |
+
"loss": 0.0269,
|
| 1616 |
+
"step": 1970
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 24.444444444444443,
|
| 1620 |
+
"grad_norm": 0.43945303559303284,
|
| 1621 |
+
"learning_rate": 4.11522633744856e-05,
|
| 1622 |
+
"loss": 0.0355,
|
| 1623 |
+
"step": 1980
|
| 1624 |
+
},
|
| 1625 |
+
{
|
| 1626 |
+
"epoch": 24.567901234567902,
|
| 1627 |
+
"grad_norm": 0.21818186342716217,
|
| 1628 |
+
"learning_rate": 4.0237768632830365e-05,
|
| 1629 |
+
"loss": 0.0223,
|
| 1630 |
+
"step": 1990
|
| 1631 |
+
},
|
| 1632 |
+
{
|
| 1633 |
+
"epoch": 24.691358024691358,
|
| 1634 |
+
"grad_norm": 0.18842598795890808,
|
| 1635 |
+
"learning_rate": 3.932327389117513e-05,
|
| 1636 |
+
"loss": 0.0147,
|
| 1637 |
+
"step": 2000
|
| 1638 |
+
},
|
| 1639 |
+
{
|
| 1640 |
+
"epoch": 24.691358024691358,
|
| 1641 |
+
"eval_accuracy": 0.9358024691358025,
|
| 1642 |
+
"eval_f1": 0.844776119402985,
|
| 1643 |
+
"eval_loss": 0.2153269648551941,
|
| 1644 |
+
"eval_precision": 0.8460388639760837,
|
| 1645 |
+
"eval_recall": 0.8435171385991058,
|
| 1646 |
+
"eval_runtime": 3.6738,
|
| 1647 |
+
"eval_samples_per_second": 88.193,
|
| 1648 |
+
"eval_steps_per_second": 11.16,
|
| 1649 |
+
"step": 2000
|
| 1650 |
+
},
|
| 1651 |
+
{
|
| 1652 |
+
"epoch": 24.814814814814813,
|
| 1653 |
+
"grad_norm": 0.13068807125091553,
|
| 1654 |
+
"learning_rate": 3.840877914951989e-05,
|
| 1655 |
+
"loss": 0.0217,
|
| 1656 |
+
"step": 2010
|
| 1657 |
+
},
|
| 1658 |
+
{
|
| 1659 |
+
"epoch": 24.938271604938272,
|
| 1660 |
+
"grad_norm": 0.17009228467941284,
|
| 1661 |
+
"learning_rate": 3.7494284407864654e-05,
|
| 1662 |
+
"loss": 0.0217,
|
| 1663 |
+
"step": 2020
|
| 1664 |
+
},
|
| 1665 |
+
{
|
| 1666 |
+
"epoch": 25.061728395061728,
|
| 1667 |
+
"grad_norm": 0.2452479898929596,
|
| 1668 |
+
"learning_rate": 3.657978966620942e-05,
|
| 1669 |
+
"loss": 0.0235,
|
| 1670 |
+
"step": 2030
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"epoch": 25.185185185185187,
|
| 1674 |
+
"grad_norm": 0.09778738021850586,
|
| 1675 |
+
"learning_rate": 3.566529492455419e-05,
|
| 1676 |
+
"loss": 0.0222,
|
| 1677 |
+
"step": 2040
|
| 1678 |
+
},
|
| 1679 |
+
{
|
| 1680 |
+
"epoch": 25.308641975308642,
|
| 1681 |
+
"grad_norm": 0.3846670687198639,
|
| 1682 |
+
"learning_rate": 3.475080018289895e-05,
|
| 1683 |
+
"loss": 0.0311,
|
| 1684 |
+
"step": 2050
|
| 1685 |
+
},
|
| 1686 |
+
{
|
| 1687 |
+
"epoch": 25.432098765432098,
|
| 1688 |
+
"grad_norm": 0.3299199640750885,
|
| 1689 |
+
"learning_rate": 3.3836305441243715e-05,
|
| 1690 |
+
"loss": 0.0282,
|
| 1691 |
+
"step": 2060
|
| 1692 |
+
},
|
| 1693 |
+
{
|
| 1694 |
+
"epoch": 25.555555555555557,
|
| 1695 |
+
"grad_norm": 0.33067819476127625,
|
| 1696 |
+
"learning_rate": 3.292181069958848e-05,
|
| 1697 |
+
"loss": 0.0342,
|
| 1698 |
+
"step": 2070
|
| 1699 |
+
},
|
| 1700 |
+
{
|
| 1701 |
+
"epoch": 25.679012345679013,
|
| 1702 |
+
"grad_norm": 0.1282743513584137,
|
| 1703 |
+
"learning_rate": 3.2007315957933245e-05,
|
| 1704 |
+
"loss": 0.0228,
|
| 1705 |
+
"step": 2080
|
| 1706 |
+
},
|
| 1707 |
+
{
|
| 1708 |
+
"epoch": 25.80246913580247,
|
| 1709 |
+
"grad_norm": 0.1972442865371704,
|
| 1710 |
+
"learning_rate": 3.1092821216278004e-05,
|
| 1711 |
+
"loss": 0.0217,
|
| 1712 |
+
"step": 2090
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 25.925925925925927,
|
| 1716 |
+
"grad_norm": 0.6959353089332581,
|
| 1717 |
+
"learning_rate": 3.017832647462277e-05,
|
| 1718 |
+
"loss": 0.03,
|
| 1719 |
+
"step": 2100
|
| 1720 |
+
},
|
| 1721 |
+
{
|
| 1722 |
+
"epoch": 25.925925925925927,
|
| 1723 |
+
"eval_accuracy": 0.9342592592592592,
|
| 1724 |
+
"eval_f1": 0.8421052631578947,
|
| 1725 |
+
"eval_loss": 0.21478094160556793,
|
| 1726 |
+
"eval_precision": 0.8377581120943953,
|
| 1727 |
+
"eval_recall": 0.8464977645305514,
|
| 1728 |
+
"eval_runtime": 2.688,
|
| 1729 |
+
"eval_samples_per_second": 120.534,
|
| 1730 |
+
"eval_steps_per_second": 15.253,
|
| 1731 |
+
"step": 2100
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 26.049382716049383,
|
| 1735 |
+
"grad_norm": 0.18116125464439392,
|
| 1736 |
+
"learning_rate": 2.926383173296754e-05,
|
| 1737 |
+
"loss": 0.0392,
|
| 1738 |
+
"step": 2110
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 26.17283950617284,
|
| 1742 |
+
"grad_norm": 0.12619613111019135,
|
| 1743 |
+
"learning_rate": 2.8349336991312303e-05,
|
| 1744 |
+
"loss": 0.0189,
|
| 1745 |
+
"step": 2120
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 26.296296296296298,
|
| 1749 |
+
"grad_norm": 0.041821230202913284,
|
| 1750 |
+
"learning_rate": 2.7434842249657068e-05,
|
| 1751 |
+
"loss": 0.0334,
|
| 1752 |
+
"step": 2130
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 26.419753086419753,
|
| 1756 |
+
"grad_norm": 0.25579944252967834,
|
| 1757 |
+
"learning_rate": 2.652034750800183e-05,
|
| 1758 |
+
"loss": 0.0252,
|
| 1759 |
+
"step": 2140
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 26.54320987654321,
|
| 1763 |
+
"grad_norm": 0.5486050844192505,
|
| 1764 |
+
"learning_rate": 2.5605852766346595e-05,
|
| 1765 |
+
"loss": 0.022,
|
| 1766 |
+
"step": 2150
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 26.666666666666668,
|
| 1770 |
+
"grad_norm": 0.204714834690094,
|
| 1771 |
+
"learning_rate": 2.4691358024691357e-05,
|
| 1772 |
+
"loss": 0.0242,
|
| 1773 |
+
"step": 2160
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 26.790123456790123,
|
| 1777 |
+
"grad_norm": 0.9011967182159424,
|
| 1778 |
+
"learning_rate": 2.3776863283036125e-05,
|
| 1779 |
+
"loss": 0.0416,
|
| 1780 |
+
"step": 2170
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 26.91358024691358,
|
| 1784 |
+
"grad_norm": 0.072830930352211,
|
| 1785 |
+
"learning_rate": 2.2862368541380887e-05,
|
| 1786 |
+
"loss": 0.0148,
|
| 1787 |
+
"step": 2180
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 27.037037037037038,
|
| 1791 |
+
"grad_norm": 0.13628140091896057,
|
| 1792 |
+
"learning_rate": 2.1947873799725652e-05,
|
| 1793 |
+
"loss": 0.0256,
|
| 1794 |
+
"step": 2190
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 27.160493827160494,
|
| 1798 |
+
"grad_norm": 0.48816823959350586,
|
| 1799 |
+
"learning_rate": 2.1033379058070417e-05,
|
| 1800 |
+
"loss": 0.0228,
|
| 1801 |
+
"step": 2200
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 27.160493827160494,
|
| 1805 |
+
"eval_accuracy": 0.9385802469135802,
|
| 1806 |
+
"eval_f1": 0.8511593118922962,
|
| 1807 |
+
"eval_loss": 0.20911623537540436,
|
| 1808 |
+
"eval_precision": 0.8543543543543544,
|
| 1809 |
+
"eval_recall": 0.8479880774962743,
|
| 1810 |
+
"eval_runtime": 3.05,
|
| 1811 |
+
"eval_samples_per_second": 106.231,
|
| 1812 |
+
"eval_steps_per_second": 13.443,
|
| 1813 |
+
"step": 2200
|
| 1814 |
+
},
|
| 1815 |
+
{
|
| 1816 |
+
"epoch": 27.28395061728395,
|
| 1817 |
+
"grad_norm": 0.17725728452205658,
|
| 1818 |
+
"learning_rate": 2.0118884316415183e-05,
|
| 1819 |
+
"loss": 0.0367,
|
| 1820 |
+
"step": 2210
|
| 1821 |
+
},
|
| 1822 |
+
{
|
| 1823 |
+
"epoch": 27.40740740740741,
|
| 1824 |
+
"grad_norm": 0.14292789995670319,
|
| 1825 |
+
"learning_rate": 1.9204389574759944e-05,
|
| 1826 |
+
"loss": 0.0179,
|
| 1827 |
+
"step": 2220
|
| 1828 |
+
},
|
| 1829 |
+
{
|
| 1830 |
+
"epoch": 27.530864197530864,
|
| 1831 |
+
"grad_norm": 0.5695340037345886,
|
| 1832 |
+
"learning_rate": 1.828989483310471e-05,
|
| 1833 |
+
"loss": 0.025,
|
| 1834 |
+
"step": 2230
|
| 1835 |
+
},
|
| 1836 |
+
{
|
| 1837 |
+
"epoch": 27.65432098765432,
|
| 1838 |
+
"grad_norm": 0.12956739962100983,
|
| 1839 |
+
"learning_rate": 1.7375400091449475e-05,
|
| 1840 |
+
"loss": 0.024,
|
| 1841 |
+
"step": 2240
|
| 1842 |
+
},
|
| 1843 |
+
{
|
| 1844 |
+
"epoch": 27.77777777777778,
|
| 1845 |
+
"grad_norm": 0.07494452595710754,
|
| 1846 |
+
"learning_rate": 1.646090534979424e-05,
|
| 1847 |
+
"loss": 0.0313,
|
| 1848 |
+
"step": 2250
|
| 1849 |
+
},
|
| 1850 |
+
{
|
| 1851 |
+
"epoch": 27.901234567901234,
|
| 1852 |
+
"grad_norm": 0.1356113702058792,
|
| 1853 |
+
"learning_rate": 1.5546410608139002e-05,
|
| 1854 |
+
"loss": 0.0146,
|
| 1855 |
+
"step": 2260
|
| 1856 |
+
},
|
| 1857 |
+
{
|
| 1858 |
+
"epoch": 28.02469135802469,
|
| 1859 |
+
"grad_norm": 0.08553437143564224,
|
| 1860 |
+
"learning_rate": 1.463191586648377e-05,
|
| 1861 |
+
"loss": 0.0184,
|
| 1862 |
+
"step": 2270
|
| 1863 |
+
},
|
| 1864 |
+
{
|
| 1865 |
+
"epoch": 28.14814814814815,
|
| 1866 |
+
"grad_norm": 0.6898983120918274,
|
| 1867 |
+
"learning_rate": 1.3717421124828534e-05,
|
| 1868 |
+
"loss": 0.0215,
|
| 1869 |
+
"step": 2280
|
| 1870 |
+
},
|
| 1871 |
+
{
|
| 1872 |
+
"epoch": 28.271604938271604,
|
| 1873 |
+
"grad_norm": 0.150452122092247,
|
| 1874 |
+
"learning_rate": 1.2802926383173297e-05,
|
| 1875 |
+
"loss": 0.0289,
|
| 1876 |
+
"step": 2290
|
| 1877 |
+
},
|
| 1878 |
+
{
|
| 1879 |
+
"epoch": 28.395061728395063,
|
| 1880 |
+
"grad_norm": 0.23294192552566528,
|
| 1881 |
+
"learning_rate": 1.1888431641518063e-05,
|
| 1882 |
+
"loss": 0.0167,
|
| 1883 |
+
"step": 2300
|
| 1884 |
+
},
|
| 1885 |
+
{
|
| 1886 |
+
"epoch": 28.395061728395063,
|
| 1887 |
+
"eval_accuracy": 0.9361111111111111,
|
| 1888 |
+
"eval_f1": 0.8447111777944486,
|
| 1889 |
+
"eval_loss": 0.21039289236068726,
|
| 1890 |
+
"eval_precision": 0.850453172205438,
|
| 1891 |
+
"eval_recall": 0.8390461997019374,
|
| 1892 |
+
"eval_runtime": 2.8019,
|
| 1893 |
+
"eval_samples_per_second": 115.635,
|
| 1894 |
+
"eval_steps_per_second": 14.633,
|
| 1895 |
+
"step": 2300
|
| 1896 |
+
},
|
| 1897 |
+
{
|
| 1898 |
+
"epoch": 28.51851851851852,
|
| 1899 |
+
"grad_norm": 0.04379408434033394,
|
| 1900 |
+
"learning_rate": 1.0973936899862826e-05,
|
| 1901 |
+
"loss": 0.0171,
|
| 1902 |
+
"step": 2310
|
| 1903 |
+
},
|
| 1904 |
+
{
|
| 1905 |
+
"epoch": 28.641975308641975,
|
| 1906 |
+
"grad_norm": 0.35769739747047424,
|
| 1907 |
+
"learning_rate": 1.0059442158207591e-05,
|
| 1908 |
+
"loss": 0.0185,
|
| 1909 |
+
"step": 2320
|
| 1910 |
+
},
|
| 1911 |
+
{
|
| 1912 |
+
"epoch": 28.765432098765434,
|
| 1913 |
+
"grad_norm": 0.1033988893032074,
|
| 1914 |
+
"learning_rate": 9.144947416552355e-06,
|
| 1915 |
+
"loss": 0.0225,
|
| 1916 |
+
"step": 2330
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"epoch": 28.88888888888889,
|
| 1920 |
+
"grad_norm": 0.26813939213752747,
|
| 1921 |
+
"learning_rate": 8.23045267489712e-06,
|
| 1922 |
+
"loss": 0.0187,
|
| 1923 |
+
"step": 2340
|
| 1924 |
+
},
|
| 1925 |
+
{
|
| 1926 |
+
"epoch": 29.012345679012345,
|
| 1927 |
+
"grad_norm": 0.23706835508346558,
|
| 1928 |
+
"learning_rate": 7.315957933241885e-06,
|
| 1929 |
+
"loss": 0.0169,
|
| 1930 |
+
"step": 2350
|
| 1931 |
+
},
|
| 1932 |
+
{
|
| 1933 |
+
"epoch": 29.135802469135804,
|
| 1934 |
+
"grad_norm": 0.2444809526205063,
|
| 1935 |
+
"learning_rate": 6.401463191586649e-06,
|
| 1936 |
+
"loss": 0.0125,
|
| 1937 |
+
"step": 2360
|
| 1938 |
+
},
|
| 1939 |
+
{
|
| 1940 |
+
"epoch": 29.25925925925926,
|
| 1941 |
+
"grad_norm": 0.22470110654830933,
|
| 1942 |
+
"learning_rate": 5.486968449931413e-06,
|
| 1943 |
+
"loss": 0.0291,
|
| 1944 |
+
"step": 2370
|
| 1945 |
+
},
|
| 1946 |
+
{
|
| 1947 |
+
"epoch": 29.382716049382715,
|
| 1948 |
+
"grad_norm": 0.10346284508705139,
|
| 1949 |
+
"learning_rate": 4.572473708276177e-06,
|
| 1950 |
+
"loss": 0.0234,
|
| 1951 |
+
"step": 2380
|
| 1952 |
+
},
|
| 1953 |
+
{
|
| 1954 |
+
"epoch": 29.506172839506174,
|
| 1955 |
+
"grad_norm": 0.3980329930782318,
|
| 1956 |
+
"learning_rate": 3.6579789666209426e-06,
|
| 1957 |
+
"loss": 0.0175,
|
| 1958 |
+
"step": 2390
|
| 1959 |
+
},
|
| 1960 |
+
{
|
| 1961 |
+
"epoch": 29.62962962962963,
|
| 1962 |
+
"grad_norm": 0.19435462355613708,
|
| 1963 |
+
"learning_rate": 2.7434842249657065e-06,
|
| 1964 |
+
"loss": 0.0273,
|
| 1965 |
+
"step": 2400
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 29.62962962962963,
|
| 1969 |
+
"eval_accuracy": 0.9364197530864198,
|
| 1970 |
+
"eval_f1": 0.8458083832335329,
|
| 1971 |
+
"eval_loss": 0.20891080796718597,
|
| 1972 |
+
"eval_precision": 0.849624060150376,
|
| 1973 |
+
"eval_recall": 0.842026825633383,
|
| 1974 |
+
"eval_runtime": 2.5442,
|
| 1975 |
+
"eval_samples_per_second": 127.348,
|
| 1976 |
+
"eval_steps_per_second": 16.115,
|
| 1977 |
+
"step": 2400
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 29.753086419753085,
|
| 1981 |
+
"grad_norm": 0.3585284352302551,
|
| 1982 |
+
"learning_rate": 1.8289894833104713e-06,
|
| 1983 |
+
"loss": 0.0206,
|
| 1984 |
+
"step": 2410
|
| 1985 |
+
},
|
| 1986 |
+
{
|
| 1987 |
+
"epoch": 29.876543209876544,
|
| 1988 |
+
"grad_norm": 0.11663785576820374,
|
| 1989 |
+
"learning_rate": 9.144947416552356e-07,
|
| 1990 |
+
"loss": 0.0169,
|
| 1991 |
+
"step": 2420
|
| 1992 |
+
},
|
| 1993 |
+
{
|
| 1994 |
+
"epoch": 30.0,
|
| 1995 |
+
"grad_norm": 0.14282989501953125,
|
| 1996 |
+
"learning_rate": 0.0,
|
| 1997 |
+
"loss": 0.0191,
|
| 1998 |
+
"step": 2430
|
| 1999 |
+
},
|
| 2000 |
+
{
|
| 2001 |
+
"epoch": 30.0,
|
| 2002 |
+
"step": 2430,
|
| 2003 |
+
"total_flos": 3.008454731998249e+18,
|
| 2004 |
+
"train_loss": 0.08984789192185971,
|
| 2005 |
+
"train_runtime": 670.5653,
|
| 2006 |
+
"train_samples_per_second": 57.891,
|
| 2007 |
+
"train_steps_per_second": 3.624
|
| 2008 |
+
}
|
| 2009 |
+
],
|
| 2010 |
+
"logging_steps": 10,
|
| 2011 |
+
"max_steps": 2430,
|
| 2012 |
+
"num_input_tokens_seen": 0,
|
| 2013 |
+
"num_train_epochs": 30,
|
| 2014 |
+
"save_steps": 500,
|
| 2015 |
+
"stateful_callbacks": {
|
| 2016 |
+
"TrainerControl": {
|
| 2017 |
+
"args": {
|
| 2018 |
+
"should_epoch_stop": false,
|
| 2019 |
+
"should_evaluate": false,
|
| 2020 |
+
"should_log": false,
|
| 2021 |
+
"should_save": true,
|
| 2022 |
+
"should_training_stop": true
|
| 2023 |
+
},
|
| 2024 |
+
"attributes": {}
|
| 2025 |
+
}
|
| 2026 |
+
},
|
| 2027 |
+
"total_flos": 3.008454731998249e+18,
|
| 2028 |
+
"train_batch_size": 16,
|
| 2029 |
+
"trial_name": null,
|
| 2030 |
+
"trial_params": null
|
| 2031 |
+
}
|