End of training
Browse files- README.md +3 -4
- all_results.json +13 -0
- eval_results.json +8 -0
- train_results.json +8 -0
- trainer_state.json +2178 -0
README.md
CHANGED
|
@@ -1,6 +1,5 @@
|
|
| 1 |
---
|
| 2 |
library_name: transformers
|
| 3 |
-
license: other
|
| 4 |
base_model: google/mobilenet_v2_1.0_224
|
| 5 |
tags:
|
| 6 |
- generated_from_trainer
|
|
@@ -23,7 +22,7 @@ model-index:
|
|
| 23 |
metrics:
|
| 24 |
- name: Accuracy
|
| 25 |
type: accuracy
|
| 26 |
-
value: 0.
|
| 27 |
---
|
| 28 |
|
| 29 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
@@ -33,8 +32,8 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 33 |
|
| 34 |
This model is a fine-tuned version of [google/mobilenet_v2_1.0_224](https://huggingface.co/google/mobilenet_v2_1.0_224) on the imagefolder dataset.
|
| 35 |
It achieves the following results on the evaluation set:
|
| 36 |
-
- Loss: 0.
|
| 37 |
-
- Accuracy: 0.
|
| 38 |
|
| 39 |
## Model description
|
| 40 |
|
|
|
|
| 1 |
---
|
| 2 |
library_name: transformers
|
|
|
|
| 3 |
base_model: google/mobilenet_v2_1.0_224
|
| 4 |
tags:
|
| 5 |
- generated_from_trainer
|
|
|
|
| 22 |
metrics:
|
| 23 |
- name: Accuracy
|
| 24 |
type: accuracy
|
| 25 |
+
value: 0.8676470588235294
|
| 26 |
---
|
| 27 |
|
| 28 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
|
|
| 32 |
|
| 33 |
This model is a fine-tuned version of [google/mobilenet_v2_1.0_224](https://huggingface.co/google/mobilenet_v2_1.0_224) on the imagefolder dataset.
|
| 34 |
It achieves the following results on the evaluation set:
|
| 35 |
+
- Loss: 0.4698
|
| 36 |
+
- Accuracy: 0.8676
|
| 37 |
|
| 38 |
## Model description
|
| 39 |
|
all_results.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 59.6078431372549,
|
| 3 |
+
"eval_accuracy": 0.8676470588235294,
|
| 4 |
+
"eval_loss": 0.46983805298805237,
|
| 5 |
+
"eval_runtime": 21.608,
|
| 6 |
+
"eval_samples_per_second": 6.294,
|
| 7 |
+
"eval_steps_per_second": 0.787,
|
| 8 |
+
"total_flos": 1.9206376588836864e+17,
|
| 9 |
+
"train_loss": 0.5395637326073228,
|
| 10 |
+
"train_runtime": 14275.0415,
|
| 11 |
+
"train_samples_per_second": 5.145,
|
| 12 |
+
"train_steps_per_second": 0.16
|
| 13 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 59.6078431372549,
|
| 3 |
+
"eval_accuracy": 0.8676470588235294,
|
| 4 |
+
"eval_loss": 0.46983805298805237,
|
| 5 |
+
"eval_runtime": 21.608,
|
| 6 |
+
"eval_samples_per_second": 6.294,
|
| 7 |
+
"eval_steps_per_second": 0.787
|
| 8 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 59.6078431372549,
|
| 3 |
+
"total_flos": 1.9206376588836864e+17,
|
| 4 |
+
"train_loss": 0.5395637326073228,
|
| 5 |
+
"train_runtime": 14275.0415,
|
| 6 |
+
"train_samples_per_second": 5.145,
|
| 7 |
+
"train_steps_per_second": 0.16
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2178 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.8676470588235294,
|
| 3 |
+
"best_model_checkpoint": "mobilenet_v2_1.0_224-finetuned-papsmear/checkpoint-1797",
|
| 4 |
+
"epoch": 59.6078431372549,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 2280,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.26143790849673204,
|
| 13 |
+
"grad_norm": 9.104169845581055,
|
| 14 |
+
"learning_rate": 2.1929824561403507e-06,
|
| 15 |
+
"loss": 1.8301,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.5228758169934641,
|
| 20 |
+
"grad_norm": 8.698704719543457,
|
| 21 |
+
"learning_rate": 4.3859649122807014e-06,
|
| 22 |
+
"loss": 1.8234,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.7843137254901961,
|
| 27 |
+
"grad_norm": 9.768064498901367,
|
| 28 |
+
"learning_rate": 6.578947368421053e-06,
|
| 29 |
+
"loss": 1.7932,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.9934640522875817,
|
| 34 |
+
"eval_accuracy": 0.25,
|
| 35 |
+
"eval_loss": 1.7606836557388306,
|
| 36 |
+
"eval_runtime": 20.4249,
|
| 37 |
+
"eval_samples_per_second": 6.659,
|
| 38 |
+
"eval_steps_per_second": 0.832,
|
| 39 |
+
"step": 38
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 1.0457516339869282,
|
| 43 |
+
"grad_norm": 8.835221290588379,
|
| 44 |
+
"learning_rate": 8.771929824561403e-06,
|
| 45 |
+
"loss": 1.7895,
|
| 46 |
+
"step": 40
|
| 47 |
+
},
|
| 48 |
+
{
|
| 49 |
+
"epoch": 1.3071895424836601,
|
| 50 |
+
"grad_norm": 9.015777587890625,
|
| 51 |
+
"learning_rate": 1.0964912280701754e-05,
|
| 52 |
+
"loss": 1.7622,
|
| 53 |
+
"step": 50
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 1.5686274509803921,
|
| 57 |
+
"grad_norm": 8.489930152893066,
|
| 58 |
+
"learning_rate": 1.3157894736842106e-05,
|
| 59 |
+
"loss": 1.7277,
|
| 60 |
+
"step": 60
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"epoch": 1.8300653594771243,
|
| 64 |
+
"grad_norm": 8.624068260192871,
|
| 65 |
+
"learning_rate": 1.5350877192982457e-05,
|
| 66 |
+
"loss": 1.6542,
|
| 67 |
+
"step": 70
|
| 68 |
+
},
|
| 69 |
+
{
|
| 70 |
+
"epoch": 1.9869281045751634,
|
| 71 |
+
"eval_accuracy": 0.39705882352941174,
|
| 72 |
+
"eval_loss": 1.5735763311386108,
|
| 73 |
+
"eval_runtime": 19.6256,
|
| 74 |
+
"eval_samples_per_second": 6.93,
|
| 75 |
+
"eval_steps_per_second": 0.866,
|
| 76 |
+
"step": 76
|
| 77 |
+
},
|
| 78 |
+
{
|
| 79 |
+
"epoch": 2.0915032679738563,
|
| 80 |
+
"grad_norm": 8.306758880615234,
|
| 81 |
+
"learning_rate": 1.7543859649122806e-05,
|
| 82 |
+
"loss": 1.6402,
|
| 83 |
+
"step": 80
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 2.3529411764705883,
|
| 87 |
+
"grad_norm": 7.969709873199463,
|
| 88 |
+
"learning_rate": 1.9736842105263158e-05,
|
| 89 |
+
"loss": 1.6017,
|
| 90 |
+
"step": 90
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 2.6143790849673203,
|
| 94 |
+
"grad_norm": 8.150823593139648,
|
| 95 |
+
"learning_rate": 2.1929824561403507e-05,
|
| 96 |
+
"loss": 1.5358,
|
| 97 |
+
"step": 100
|
| 98 |
+
},
|
| 99 |
+
{
|
| 100 |
+
"epoch": 2.8758169934640523,
|
| 101 |
+
"grad_norm": 7.387766361236572,
|
| 102 |
+
"learning_rate": 2.412280701754386e-05,
|
| 103 |
+
"loss": 1.4692,
|
| 104 |
+
"step": 110
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 2.980392156862745,
|
| 108 |
+
"eval_accuracy": 0.36764705882352944,
|
| 109 |
+
"eval_loss": 1.4805188179016113,
|
| 110 |
+
"eval_runtime": 20.2426,
|
| 111 |
+
"eval_samples_per_second": 6.719,
|
| 112 |
+
"eval_steps_per_second": 0.84,
|
| 113 |
+
"step": 114
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 3.1372549019607843,
|
| 117 |
+
"grad_norm": 8.2266845703125,
|
| 118 |
+
"learning_rate": 2.6315789473684212e-05,
|
| 119 |
+
"loss": 1.439,
|
| 120 |
+
"step": 120
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 3.3986928104575163,
|
| 124 |
+
"grad_norm": 8.400202751159668,
|
| 125 |
+
"learning_rate": 2.850877192982456e-05,
|
| 126 |
+
"loss": 1.399,
|
| 127 |
+
"step": 130
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 3.6601307189542482,
|
| 131 |
+
"grad_norm": 8.897980690002441,
|
| 132 |
+
"learning_rate": 3.0701754385964913e-05,
|
| 133 |
+
"loss": 1.3496,
|
| 134 |
+
"step": 140
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 3.9215686274509802,
|
| 138 |
+
"grad_norm": 8.579508781433105,
|
| 139 |
+
"learning_rate": 3.289473684210527e-05,
|
| 140 |
+
"loss": 1.2759,
|
| 141 |
+
"step": 150
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 4.0,
|
| 145 |
+
"eval_accuracy": 0.5808823529411765,
|
| 146 |
+
"eval_loss": 1.2177482843399048,
|
| 147 |
+
"eval_runtime": 19.8987,
|
| 148 |
+
"eval_samples_per_second": 6.835,
|
| 149 |
+
"eval_steps_per_second": 0.854,
|
| 150 |
+
"step": 153
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 4.183006535947713,
|
| 154 |
+
"grad_norm": 7.850776195526123,
|
| 155 |
+
"learning_rate": 3.508771929824561e-05,
|
| 156 |
+
"loss": 1.2353,
|
| 157 |
+
"step": 160
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 4.444444444444445,
|
| 161 |
+
"grad_norm": 7.207574367523193,
|
| 162 |
+
"learning_rate": 3.728070175438597e-05,
|
| 163 |
+
"loss": 1.1957,
|
| 164 |
+
"step": 170
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 4.705882352941177,
|
| 168 |
+
"grad_norm": 7.722698211669922,
|
| 169 |
+
"learning_rate": 3.9473684210526316e-05,
|
| 170 |
+
"loss": 1.1894,
|
| 171 |
+
"step": 180
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 4.967320261437909,
|
| 175 |
+
"grad_norm": 7.626311779022217,
|
| 176 |
+
"learning_rate": 4.166666666666667e-05,
|
| 177 |
+
"loss": 1.1521,
|
| 178 |
+
"step": 190
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 4.993464052287582,
|
| 182 |
+
"eval_accuracy": 0.6470588235294118,
|
| 183 |
+
"eval_loss": 1.0726604461669922,
|
| 184 |
+
"eval_runtime": 19.9729,
|
| 185 |
+
"eval_samples_per_second": 6.809,
|
| 186 |
+
"eval_steps_per_second": 0.851,
|
| 187 |
+
"step": 191
|
| 188 |
+
},
|
| 189 |
+
{
|
| 190 |
+
"epoch": 5.228758169934641,
|
| 191 |
+
"grad_norm": 6.824063301086426,
|
| 192 |
+
"learning_rate": 4.3859649122807014e-05,
|
| 193 |
+
"loss": 1.1157,
|
| 194 |
+
"step": 200
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 5.490196078431373,
|
| 198 |
+
"grad_norm": 9.015517234802246,
|
| 199 |
+
"learning_rate": 4.605263157894737e-05,
|
| 200 |
+
"loss": 1.0953,
|
| 201 |
+
"step": 210
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 5.751633986928105,
|
| 205 |
+
"grad_norm": 7.684938430786133,
|
| 206 |
+
"learning_rate": 4.824561403508772e-05,
|
| 207 |
+
"loss": 1.078,
|
| 208 |
+
"step": 220
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 5.9869281045751634,
|
| 212 |
+
"eval_accuracy": 0.6176470588235294,
|
| 213 |
+
"eval_loss": 0.9995782375335693,
|
| 214 |
+
"eval_runtime": 19.7917,
|
| 215 |
+
"eval_samples_per_second": 6.872,
|
| 216 |
+
"eval_steps_per_second": 0.859,
|
| 217 |
+
"step": 229
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 6.0130718954248366,
|
| 221 |
+
"grad_norm": 8.750350952148438,
|
| 222 |
+
"learning_rate": 4.9951267056530214e-05,
|
| 223 |
+
"loss": 1.0635,
|
| 224 |
+
"step": 230
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 6.2745098039215685,
|
| 228 |
+
"grad_norm": 6.832897186279297,
|
| 229 |
+
"learning_rate": 4.970760233918128e-05,
|
| 230 |
+
"loss": 0.9083,
|
| 231 |
+
"step": 240
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 6.5359477124183005,
|
| 235 |
+
"grad_norm": 6.695949554443359,
|
| 236 |
+
"learning_rate": 4.9463937621832366e-05,
|
| 237 |
+
"loss": 0.9306,
|
| 238 |
+
"step": 250
|
| 239 |
+
},
|
| 240 |
+
{
|
| 241 |
+
"epoch": 6.7973856209150325,
|
| 242 |
+
"grad_norm": 7.3760271072387695,
|
| 243 |
+
"learning_rate": 4.9220272904483435e-05,
|
| 244 |
+
"loss": 1.0235,
|
| 245 |
+
"step": 260
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 6.980392156862745,
|
| 249 |
+
"eval_accuracy": 0.7058823529411765,
|
| 250 |
+
"eval_loss": 0.8680174946784973,
|
| 251 |
+
"eval_runtime": 19.5178,
|
| 252 |
+
"eval_samples_per_second": 6.968,
|
| 253 |
+
"eval_steps_per_second": 0.871,
|
| 254 |
+
"step": 267
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 7.0588235294117645,
|
| 258 |
+
"grad_norm": 6.625696659088135,
|
| 259 |
+
"learning_rate": 4.8976608187134504e-05,
|
| 260 |
+
"loss": 0.885,
|
| 261 |
+
"step": 270
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 7.3202614379084965,
|
| 265 |
+
"grad_norm": 6.442111015319824,
|
| 266 |
+
"learning_rate": 4.8732943469785574e-05,
|
| 267 |
+
"loss": 0.8222,
|
| 268 |
+
"step": 280
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 7.5816993464052285,
|
| 272 |
+
"grad_norm": 8.278558731079102,
|
| 273 |
+
"learning_rate": 4.848927875243665e-05,
|
| 274 |
+
"loss": 0.8869,
|
| 275 |
+
"step": 290
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 7.8431372549019605,
|
| 279 |
+
"grad_norm": 7.1332106590271,
|
| 280 |
+
"learning_rate": 4.824561403508772e-05,
|
| 281 |
+
"loss": 0.9554,
|
| 282 |
+
"step": 300
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 8.0,
|
| 286 |
+
"eval_accuracy": 0.6397058823529411,
|
| 287 |
+
"eval_loss": 0.9272549748420715,
|
| 288 |
+
"eval_runtime": 19.5476,
|
| 289 |
+
"eval_samples_per_second": 6.957,
|
| 290 |
+
"eval_steps_per_second": 0.87,
|
| 291 |
+
"step": 306
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 8.104575163398692,
|
| 295 |
+
"grad_norm": 8.287984848022461,
|
| 296 |
+
"learning_rate": 4.8001949317738795e-05,
|
| 297 |
+
"loss": 0.9148,
|
| 298 |
+
"step": 310
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 8.366013071895425,
|
| 302 |
+
"grad_norm": 7.765283107757568,
|
| 303 |
+
"learning_rate": 4.7758284600389865e-05,
|
| 304 |
+
"loss": 0.7935,
|
| 305 |
+
"step": 320
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 8.627450980392156,
|
| 309 |
+
"grad_norm": 7.960378646850586,
|
| 310 |
+
"learning_rate": 4.751461988304094e-05,
|
| 311 |
+
"loss": 0.9253,
|
| 312 |
+
"step": 330
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 8.88888888888889,
|
| 316 |
+
"grad_norm": 7.798432350158691,
|
| 317 |
+
"learning_rate": 4.727095516569201e-05,
|
| 318 |
+
"loss": 0.7437,
|
| 319 |
+
"step": 340
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 8.993464052287582,
|
| 323 |
+
"eval_accuracy": 0.7058823529411765,
|
| 324 |
+
"eval_loss": 0.7389305233955383,
|
| 325 |
+
"eval_runtime": 20.3392,
|
| 326 |
+
"eval_samples_per_second": 6.687,
|
| 327 |
+
"eval_steps_per_second": 0.836,
|
| 328 |
+
"step": 344
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 9.15032679738562,
|
| 332 |
+
"grad_norm": 7.6827921867370605,
|
| 333 |
+
"learning_rate": 4.702729044834308e-05,
|
| 334 |
+
"loss": 0.8238,
|
| 335 |
+
"step": 350
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 9.411764705882353,
|
| 339 |
+
"grad_norm": 7.531745433807373,
|
| 340 |
+
"learning_rate": 4.678362573099415e-05,
|
| 341 |
+
"loss": 0.7379,
|
| 342 |
+
"step": 360
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"epoch": 9.673202614379084,
|
| 346 |
+
"grad_norm": 7.212475299835205,
|
| 347 |
+
"learning_rate": 4.6539961013645225e-05,
|
| 348 |
+
"loss": 0.7684,
|
| 349 |
+
"step": 370
|
| 350 |
+
},
|
| 351 |
+
{
|
| 352 |
+
"epoch": 9.934640522875817,
|
| 353 |
+
"grad_norm": 8.958950996398926,
|
| 354 |
+
"learning_rate": 4.62962962962963e-05,
|
| 355 |
+
"loss": 0.7876,
|
| 356 |
+
"step": 380
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 9.986928104575163,
|
| 360 |
+
"eval_accuracy": 0.7426470588235294,
|
| 361 |
+
"eval_loss": 0.6773847937583923,
|
| 362 |
+
"eval_runtime": 22.3142,
|
| 363 |
+
"eval_samples_per_second": 6.095,
|
| 364 |
+
"eval_steps_per_second": 0.762,
|
| 365 |
+
"step": 382
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 10.196078431372548,
|
| 369 |
+
"grad_norm": 8.386847496032715,
|
| 370 |
+
"learning_rate": 4.605263157894737e-05,
|
| 371 |
+
"loss": 0.7287,
|
| 372 |
+
"step": 390
|
| 373 |
+
},
|
| 374 |
+
{
|
| 375 |
+
"epoch": 10.457516339869281,
|
| 376 |
+
"grad_norm": 8.313080787658691,
|
| 377 |
+
"learning_rate": 4.580896686159844e-05,
|
| 378 |
+
"loss": 0.6713,
|
| 379 |
+
"step": 400
|
| 380 |
+
},
|
| 381 |
+
{
|
| 382 |
+
"epoch": 10.718954248366012,
|
| 383 |
+
"grad_norm": 7.874463081359863,
|
| 384 |
+
"learning_rate": 4.5565302144249516e-05,
|
| 385 |
+
"loss": 0.7802,
|
| 386 |
+
"step": 410
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 10.980392156862745,
|
| 390 |
+
"grad_norm": 7.682034492492676,
|
| 391 |
+
"learning_rate": 4.5321637426900585e-05,
|
| 392 |
+
"loss": 0.7698,
|
| 393 |
+
"step": 420
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 10.980392156862745,
|
| 397 |
+
"eval_accuracy": 0.7205882352941176,
|
| 398 |
+
"eval_loss": 0.6569061875343323,
|
| 399 |
+
"eval_runtime": 20.7833,
|
| 400 |
+
"eval_samples_per_second": 6.544,
|
| 401 |
+
"eval_steps_per_second": 0.818,
|
| 402 |
+
"step": 420
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 11.241830065359476,
|
| 406 |
+
"grad_norm": 8.04808521270752,
|
| 407 |
+
"learning_rate": 4.507797270955166e-05,
|
| 408 |
+
"loss": 0.7152,
|
| 409 |
+
"step": 430
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 11.50326797385621,
|
| 413 |
+
"grad_norm": 6.100086212158203,
|
| 414 |
+
"learning_rate": 4.483430799220273e-05,
|
| 415 |
+
"loss": 0.6517,
|
| 416 |
+
"step": 440
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 11.764705882352942,
|
| 420 |
+
"grad_norm": 10.316337585449219,
|
| 421 |
+
"learning_rate": 4.4590643274853806e-05,
|
| 422 |
+
"loss": 0.7597,
|
| 423 |
+
"step": 450
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 12.0,
|
| 427 |
+
"eval_accuracy": 0.7573529411764706,
|
| 428 |
+
"eval_loss": 0.6757553219795227,
|
| 429 |
+
"eval_runtime": 21.1029,
|
| 430 |
+
"eval_samples_per_second": 6.445,
|
| 431 |
+
"eval_steps_per_second": 0.806,
|
| 432 |
+
"step": 459
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 12.026143790849673,
|
| 436 |
+
"grad_norm": 8.15910816192627,
|
| 437 |
+
"learning_rate": 4.4346978557504876e-05,
|
| 438 |
+
"loss": 0.7,
|
| 439 |
+
"step": 460
|
| 440 |
+
},
|
| 441 |
+
{
|
| 442 |
+
"epoch": 12.287581699346406,
|
| 443 |
+
"grad_norm": 6.629967212677002,
|
| 444 |
+
"learning_rate": 4.4103313840155945e-05,
|
| 445 |
+
"loss": 0.6214,
|
| 446 |
+
"step": 470
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 12.549019607843137,
|
| 450 |
+
"grad_norm": 7.6857123374938965,
|
| 451 |
+
"learning_rate": 4.3859649122807014e-05,
|
| 452 |
+
"loss": 0.6272,
|
| 453 |
+
"step": 480
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 12.81045751633987,
|
| 457 |
+
"grad_norm": 5.317193031311035,
|
| 458 |
+
"learning_rate": 4.361598440545809e-05,
|
| 459 |
+
"loss": 0.6114,
|
| 460 |
+
"step": 490
|
| 461 |
+
},
|
| 462 |
+
{
|
| 463 |
+
"epoch": 12.993464052287582,
|
| 464 |
+
"eval_accuracy": 0.7132352941176471,
|
| 465 |
+
"eval_loss": 0.8278929591178894,
|
| 466 |
+
"eval_runtime": 24.3179,
|
| 467 |
+
"eval_samples_per_second": 5.593,
|
| 468 |
+
"eval_steps_per_second": 0.699,
|
| 469 |
+
"step": 497
|
| 470 |
+
},
|
| 471 |
+
{
|
| 472 |
+
"epoch": 13.071895424836601,
|
| 473 |
+
"grad_norm": 7.161328315734863,
|
| 474 |
+
"learning_rate": 4.3372319688109166e-05,
|
| 475 |
+
"loss": 0.66,
|
| 476 |
+
"step": 500
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 13.333333333333334,
|
| 480 |
+
"grad_norm": 7.760936260223389,
|
| 481 |
+
"learning_rate": 4.3128654970760236e-05,
|
| 482 |
+
"loss": 0.6246,
|
| 483 |
+
"step": 510
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 13.594771241830065,
|
| 487 |
+
"grad_norm": 7.746845245361328,
|
| 488 |
+
"learning_rate": 4.2884990253411305e-05,
|
| 489 |
+
"loss": 0.6391,
|
| 490 |
+
"step": 520
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 13.856209150326798,
|
| 494 |
+
"grad_norm": 7.603119373321533,
|
| 495 |
+
"learning_rate": 4.264132553606238e-05,
|
| 496 |
+
"loss": 0.6847,
|
| 497 |
+
"step": 530
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 13.986928104575163,
|
| 501 |
+
"eval_accuracy": 0.7132352941176471,
|
| 502 |
+
"eval_loss": 0.7504714131355286,
|
| 503 |
+
"eval_runtime": 24.2998,
|
| 504 |
+
"eval_samples_per_second": 5.597,
|
| 505 |
+
"eval_steps_per_second": 0.7,
|
| 506 |
+
"step": 535
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 14.117647058823529,
|
| 510 |
+
"grad_norm": 6.282960414886475,
|
| 511 |
+
"learning_rate": 4.239766081871345e-05,
|
| 512 |
+
"loss": 0.6034,
|
| 513 |
+
"step": 540
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 14.379084967320262,
|
| 517 |
+
"grad_norm": 9.423322677612305,
|
| 518 |
+
"learning_rate": 4.215399610136453e-05,
|
| 519 |
+
"loss": 0.6312,
|
| 520 |
+
"step": 550
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 14.640522875816993,
|
| 524 |
+
"grad_norm": 7.931453704833984,
|
| 525 |
+
"learning_rate": 4.1910331384015596e-05,
|
| 526 |
+
"loss": 0.6247,
|
| 527 |
+
"step": 560
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 14.901960784313726,
|
| 531 |
+
"grad_norm": 8.18138313293457,
|
| 532 |
+
"learning_rate": 4.166666666666667e-05,
|
| 533 |
+
"loss": 0.5902,
|
| 534 |
+
"step": 570
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 14.980392156862745,
|
| 538 |
+
"eval_accuracy": 0.6691176470588235,
|
| 539 |
+
"eval_loss": 0.7918919920921326,
|
| 540 |
+
"eval_runtime": 22.015,
|
| 541 |
+
"eval_samples_per_second": 6.178,
|
| 542 |
+
"eval_steps_per_second": 0.772,
|
| 543 |
+
"step": 573
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 15.163398692810457,
|
| 547 |
+
"grad_norm": 9.473512649536133,
|
| 548 |
+
"learning_rate": 4.142300194931774e-05,
|
| 549 |
+
"loss": 0.6104,
|
| 550 |
+
"step": 580
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 15.42483660130719,
|
| 554 |
+
"grad_norm": 8.635931015014648,
|
| 555 |
+
"learning_rate": 4.117933723196881e-05,
|
| 556 |
+
"loss": 0.6392,
|
| 557 |
+
"step": 590
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 15.686274509803921,
|
| 561 |
+
"grad_norm": 8.554533004760742,
|
| 562 |
+
"learning_rate": 4.093567251461988e-05,
|
| 563 |
+
"loss": 0.5461,
|
| 564 |
+
"step": 600
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 15.947712418300654,
|
| 568 |
+
"grad_norm": 6.450700283050537,
|
| 569 |
+
"learning_rate": 4.0692007797270956e-05,
|
| 570 |
+
"loss": 0.629,
|
| 571 |
+
"step": 610
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 16.0,
|
| 575 |
+
"eval_accuracy": 0.7867647058823529,
|
| 576 |
+
"eval_loss": 0.611703097820282,
|
| 577 |
+
"eval_runtime": 21.7065,
|
| 578 |
+
"eval_samples_per_second": 6.265,
|
| 579 |
+
"eval_steps_per_second": 0.783,
|
| 580 |
+
"step": 612
|
| 581 |
+
},
|
| 582 |
+
{
|
| 583 |
+
"epoch": 16.209150326797385,
|
| 584 |
+
"grad_norm": 7.3682026863098145,
|
| 585 |
+
"learning_rate": 4.044834307992203e-05,
|
| 586 |
+
"loss": 0.5648,
|
| 587 |
+
"step": 620
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 16.470588235294116,
|
| 591 |
+
"grad_norm": 7.454318523406982,
|
| 592 |
+
"learning_rate": 4.02046783625731e-05,
|
| 593 |
+
"loss": 0.5916,
|
| 594 |
+
"step": 630
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 16.73202614379085,
|
| 598 |
+
"grad_norm": 7.530863285064697,
|
| 599 |
+
"learning_rate": 3.996101364522417e-05,
|
| 600 |
+
"loss": 0.5476,
|
| 601 |
+
"step": 640
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 16.99346405228758,
|
| 605 |
+
"grad_norm": 6.975796699523926,
|
| 606 |
+
"learning_rate": 3.971734892787525e-05,
|
| 607 |
+
"loss": 0.5071,
|
| 608 |
+
"step": 650
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 16.99346405228758,
|
| 612 |
+
"eval_accuracy": 0.7352941176470589,
|
| 613 |
+
"eval_loss": 0.6048275828361511,
|
| 614 |
+
"eval_runtime": 22.49,
|
| 615 |
+
"eval_samples_per_second": 6.047,
|
| 616 |
+
"eval_steps_per_second": 0.756,
|
| 617 |
+
"step": 650
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 17.254901960784313,
|
| 621 |
+
"grad_norm": 7.351856708526611,
|
| 622 |
+
"learning_rate": 3.9473684210526316e-05,
|
| 623 |
+
"loss": 0.5372,
|
| 624 |
+
"step": 660
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 17.516339869281047,
|
| 628 |
+
"grad_norm": 8.070556640625,
|
| 629 |
+
"learning_rate": 3.923001949317739e-05,
|
| 630 |
+
"loss": 0.5503,
|
| 631 |
+
"step": 670
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"epoch": 17.77777777777778,
|
| 635 |
+
"grad_norm": 8.659597396850586,
|
| 636 |
+
"learning_rate": 3.898635477582846e-05,
|
| 637 |
+
"loss": 0.5453,
|
| 638 |
+
"step": 680
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 17.986928104575163,
|
| 642 |
+
"eval_accuracy": 0.7279411764705882,
|
| 643 |
+
"eval_loss": 0.8086075782775879,
|
| 644 |
+
"eval_runtime": 20.3748,
|
| 645 |
+
"eval_samples_per_second": 6.675,
|
| 646 |
+
"eval_steps_per_second": 0.834,
|
| 647 |
+
"step": 688
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 18.03921568627451,
|
| 651 |
+
"grad_norm": 8.510287284851074,
|
| 652 |
+
"learning_rate": 3.874269005847954e-05,
|
| 653 |
+
"loss": 0.5091,
|
| 654 |
+
"step": 690
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 18.30065359477124,
|
| 658 |
+
"grad_norm": 9.453730583190918,
|
| 659 |
+
"learning_rate": 3.849902534113061e-05,
|
| 660 |
+
"loss": 0.5115,
|
| 661 |
+
"step": 700
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 18.562091503267975,
|
| 665 |
+
"grad_norm": 6.948453903198242,
|
| 666 |
+
"learning_rate": 3.8255360623781676e-05,
|
| 667 |
+
"loss": 0.5383,
|
| 668 |
+
"step": 710
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 18.823529411764707,
|
| 672 |
+
"grad_norm": 7.508301734924316,
|
| 673 |
+
"learning_rate": 3.8011695906432746e-05,
|
| 674 |
+
"loss": 0.5071,
|
| 675 |
+
"step": 720
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 18.980392156862745,
|
| 679 |
+
"eval_accuracy": 0.7058823529411765,
|
| 680 |
+
"eval_loss": 0.783478856086731,
|
| 681 |
+
"eval_runtime": 21.0622,
|
| 682 |
+
"eval_samples_per_second": 6.457,
|
| 683 |
+
"eval_steps_per_second": 0.807,
|
| 684 |
+
"step": 726
|
| 685 |
+
},
|
| 686 |
+
{
|
| 687 |
+
"epoch": 19.084967320261438,
|
| 688 |
+
"grad_norm": 8.099154472351074,
|
| 689 |
+
"learning_rate": 3.776803118908382e-05,
|
| 690 |
+
"loss": 0.4653,
|
| 691 |
+
"step": 730
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 19.34640522875817,
|
| 695 |
+
"grad_norm": 6.376596927642822,
|
| 696 |
+
"learning_rate": 3.75243664717349e-05,
|
| 697 |
+
"loss": 0.426,
|
| 698 |
+
"step": 740
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 19.607843137254903,
|
| 702 |
+
"grad_norm": 9.521799087524414,
|
| 703 |
+
"learning_rate": 3.728070175438597e-05,
|
| 704 |
+
"loss": 0.4674,
|
| 705 |
+
"step": 750
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 19.869281045751634,
|
| 709 |
+
"grad_norm": 7.47611141204834,
|
| 710 |
+
"learning_rate": 3.7037037037037037e-05,
|
| 711 |
+
"loss": 0.5328,
|
| 712 |
+
"step": 760
|
| 713 |
+
},
|
| 714 |
+
{
|
| 715 |
+
"epoch": 20.0,
|
| 716 |
+
"eval_accuracy": 0.75,
|
| 717 |
+
"eval_loss": 0.6138941645622253,
|
| 718 |
+
"eval_runtime": 20.7772,
|
| 719 |
+
"eval_samples_per_second": 6.546,
|
| 720 |
+
"eval_steps_per_second": 0.818,
|
| 721 |
+
"step": 765
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 20.130718954248366,
|
| 725 |
+
"grad_norm": 8.478204727172852,
|
| 726 |
+
"learning_rate": 3.679337231968811e-05,
|
| 727 |
+
"loss": 0.5362,
|
| 728 |
+
"step": 770
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 20.392156862745097,
|
| 732 |
+
"grad_norm": 3.9403858184814453,
|
| 733 |
+
"learning_rate": 3.654970760233918e-05,
|
| 734 |
+
"loss": 0.4867,
|
| 735 |
+
"step": 780
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 20.65359477124183,
|
| 739 |
+
"grad_norm": 8.257362365722656,
|
| 740 |
+
"learning_rate": 3.630604288499025e-05,
|
| 741 |
+
"loss": 0.4665,
|
| 742 |
+
"step": 790
|
| 743 |
+
},
|
| 744 |
+
{
|
| 745 |
+
"epoch": 20.915032679738562,
|
| 746 |
+
"grad_norm": 7.73619270324707,
|
| 747 |
+
"learning_rate": 3.606237816764133e-05,
|
| 748 |
+
"loss": 0.5053,
|
| 749 |
+
"step": 800
|
| 750 |
+
},
|
| 751 |
+
{
|
| 752 |
+
"epoch": 20.99346405228758,
|
| 753 |
+
"eval_accuracy": 0.7867647058823529,
|
| 754 |
+
"eval_loss": 0.5980984568595886,
|
| 755 |
+
"eval_runtime": 20.8979,
|
| 756 |
+
"eval_samples_per_second": 6.508,
|
| 757 |
+
"eval_steps_per_second": 0.813,
|
| 758 |
+
"step": 803
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 21.176470588235293,
|
| 762 |
+
"grad_norm": 9.795605659484863,
|
| 763 |
+
"learning_rate": 3.5818713450292403e-05,
|
| 764 |
+
"loss": 0.4638,
|
| 765 |
+
"step": 810
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 21.437908496732025,
|
| 769 |
+
"grad_norm": 7.380087375640869,
|
| 770 |
+
"learning_rate": 3.557504873294347e-05,
|
| 771 |
+
"loss": 0.453,
|
| 772 |
+
"step": 820
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 21.69934640522876,
|
| 776 |
+
"grad_norm": 9.243858337402344,
|
| 777 |
+
"learning_rate": 3.533138401559454e-05,
|
| 778 |
+
"loss": 0.4768,
|
| 779 |
+
"step": 830
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 21.96078431372549,
|
| 783 |
+
"grad_norm": 9.421792030334473,
|
| 784 |
+
"learning_rate": 3.508771929824561e-05,
|
| 785 |
+
"loss": 0.4436,
|
| 786 |
+
"step": 840
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 21.986928104575163,
|
| 790 |
+
"eval_accuracy": 0.8014705882352942,
|
| 791 |
+
"eval_loss": 0.5218998193740845,
|
| 792 |
+
"eval_runtime": 20.8439,
|
| 793 |
+
"eval_samples_per_second": 6.525,
|
| 794 |
+
"eval_steps_per_second": 0.816,
|
| 795 |
+
"step": 841
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 22.22222222222222,
|
| 799 |
+
"grad_norm": 9.261748313903809,
|
| 800 |
+
"learning_rate": 3.484405458089669e-05,
|
| 801 |
+
"loss": 0.4918,
|
| 802 |
+
"step": 850
|
| 803 |
+
},
|
| 804 |
+
{
|
| 805 |
+
"epoch": 22.483660130718953,
|
| 806 |
+
"grad_norm": 7.44433069229126,
|
| 807 |
+
"learning_rate": 3.4600389863547764e-05,
|
| 808 |
+
"loss": 0.4141,
|
| 809 |
+
"step": 860
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 22.745098039215687,
|
| 813 |
+
"grad_norm": 7.531349182128906,
|
| 814 |
+
"learning_rate": 3.435672514619883e-05,
|
| 815 |
+
"loss": 0.5025,
|
| 816 |
+
"step": 870
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"epoch": 22.980392156862745,
|
| 820 |
+
"eval_accuracy": 0.8088235294117647,
|
| 821 |
+
"eval_loss": 0.4959449768066406,
|
| 822 |
+
"eval_runtime": 21.8633,
|
| 823 |
+
"eval_samples_per_second": 6.22,
|
| 824 |
+
"eval_steps_per_second": 0.778,
|
| 825 |
+
"step": 879
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 23.00653594771242,
|
| 829 |
+
"grad_norm": 7.8833794593811035,
|
| 830 |
+
"learning_rate": 3.41130604288499e-05,
|
| 831 |
+
"loss": 0.4356,
|
| 832 |
+
"step": 880
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 23.26797385620915,
|
| 836 |
+
"grad_norm": 8.349276542663574,
|
| 837 |
+
"learning_rate": 3.386939571150098e-05,
|
| 838 |
+
"loss": 0.421,
|
| 839 |
+
"step": 890
|
| 840 |
+
},
|
| 841 |
+
{
|
| 842 |
+
"epoch": 23.529411764705884,
|
| 843 |
+
"grad_norm": 7.9358086585998535,
|
| 844 |
+
"learning_rate": 3.362573099415205e-05,
|
| 845 |
+
"loss": 0.4414,
|
| 846 |
+
"step": 900
|
| 847 |
+
},
|
| 848 |
+
{
|
| 849 |
+
"epoch": 23.790849673202615,
|
| 850 |
+
"grad_norm": 7.962532043457031,
|
| 851 |
+
"learning_rate": 3.338206627680312e-05,
|
| 852 |
+
"loss": 0.4984,
|
| 853 |
+
"step": 910
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 24.0,
|
| 857 |
+
"eval_accuracy": 0.7794117647058824,
|
| 858 |
+
"eval_loss": 0.5701327323913574,
|
| 859 |
+
"eval_runtime": 20.7703,
|
| 860 |
+
"eval_samples_per_second": 6.548,
|
| 861 |
+
"eval_steps_per_second": 0.818,
|
| 862 |
+
"step": 918
|
| 863 |
+
},
|
| 864 |
+
{
|
| 865 |
+
"epoch": 24.052287581699346,
|
| 866 |
+
"grad_norm": 7.35078239440918,
|
| 867 |
+
"learning_rate": 3.313840155945419e-05,
|
| 868 |
+
"loss": 0.4853,
|
| 869 |
+
"step": 920
|
| 870 |
+
},
|
| 871 |
+
{
|
| 872 |
+
"epoch": 24.313725490196077,
|
| 873 |
+
"grad_norm": 10.068733215332031,
|
| 874 |
+
"learning_rate": 3.289473684210527e-05,
|
| 875 |
+
"loss": 0.4273,
|
| 876 |
+
"step": 930
|
| 877 |
+
},
|
| 878 |
+
{
|
| 879 |
+
"epoch": 24.575163398692812,
|
| 880 |
+
"grad_norm": 8.412373542785645,
|
| 881 |
+
"learning_rate": 3.265107212475634e-05,
|
| 882 |
+
"loss": 0.5412,
|
| 883 |
+
"step": 940
|
| 884 |
+
},
|
| 885 |
+
{
|
| 886 |
+
"epoch": 24.836601307189543,
|
| 887 |
+
"grad_norm": 6.781473636627197,
|
| 888 |
+
"learning_rate": 3.240740740740741e-05,
|
| 889 |
+
"loss": 0.4655,
|
| 890 |
+
"step": 950
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 24.99346405228758,
|
| 894 |
+
"eval_accuracy": 0.7205882352941176,
|
| 895 |
+
"eval_loss": 0.7179181575775146,
|
| 896 |
+
"eval_runtime": 21.5807,
|
| 897 |
+
"eval_samples_per_second": 6.302,
|
| 898 |
+
"eval_steps_per_second": 0.788,
|
| 899 |
+
"step": 956
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 25.098039215686274,
|
| 903 |
+
"grad_norm": 8.029053688049316,
|
| 904 |
+
"learning_rate": 3.216374269005848e-05,
|
| 905 |
+
"loss": 0.476,
|
| 906 |
+
"step": 960
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"epoch": 25.359477124183005,
|
| 910 |
+
"grad_norm": 6.743289470672607,
|
| 911 |
+
"learning_rate": 3.192007797270955e-05,
|
| 912 |
+
"loss": 0.4561,
|
| 913 |
+
"step": 970
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 25.62091503267974,
|
| 917 |
+
"grad_norm": 8.96541690826416,
|
| 918 |
+
"learning_rate": 3.167641325536063e-05,
|
| 919 |
+
"loss": 0.4343,
|
| 920 |
+
"step": 980
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 25.88235294117647,
|
| 924 |
+
"grad_norm": 6.772902965545654,
|
| 925 |
+
"learning_rate": 3.14327485380117e-05,
|
| 926 |
+
"loss": 0.3848,
|
| 927 |
+
"step": 990
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 25.986928104575163,
|
| 931 |
+
"eval_accuracy": 0.8088235294117647,
|
| 932 |
+
"eval_loss": 0.5074525475502014,
|
| 933 |
+
"eval_runtime": 19.6885,
|
| 934 |
+
"eval_samples_per_second": 6.908,
|
| 935 |
+
"eval_steps_per_second": 0.863,
|
| 936 |
+
"step": 994
|
| 937 |
+
},
|
| 938 |
+
{
|
| 939 |
+
"epoch": 26.143790849673202,
|
| 940 |
+
"grad_norm": 8.97864055633545,
|
| 941 |
+
"learning_rate": 3.118908382066277e-05,
|
| 942 |
+
"loss": 0.4338,
|
| 943 |
+
"step": 1000
|
| 944 |
+
},
|
| 945 |
+
{
|
| 946 |
+
"epoch": 26.405228758169933,
|
| 947 |
+
"grad_norm": 7.333850383758545,
|
| 948 |
+
"learning_rate": 3.0945419103313844e-05,
|
| 949 |
+
"loss": 0.4137,
|
| 950 |
+
"step": 1010
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 26.666666666666668,
|
| 954 |
+
"grad_norm": 7.425909519195557,
|
| 955 |
+
"learning_rate": 3.0701754385964913e-05,
|
| 956 |
+
"loss": 0.4776,
|
| 957 |
+
"step": 1020
|
| 958 |
+
},
|
| 959 |
+
{
|
| 960 |
+
"epoch": 26.9281045751634,
|
| 961 |
+
"grad_norm": 6.110713481903076,
|
| 962 |
+
"learning_rate": 3.0458089668615986e-05,
|
| 963 |
+
"loss": 0.3824,
|
| 964 |
+
"step": 1030
|
| 965 |
+
},
|
| 966 |
+
{
|
| 967 |
+
"epoch": 26.980392156862745,
|
| 968 |
+
"eval_accuracy": 0.7426470588235294,
|
| 969 |
+
"eval_loss": 0.6644823551177979,
|
| 970 |
+
"eval_runtime": 22.0924,
|
| 971 |
+
"eval_samples_per_second": 6.156,
|
| 972 |
+
"eval_steps_per_second": 0.769,
|
| 973 |
+
"step": 1032
|
| 974 |
+
},
|
| 975 |
+
{
|
| 976 |
+
"epoch": 27.18954248366013,
|
| 977 |
+
"grad_norm": 7.540946006774902,
|
| 978 |
+
"learning_rate": 3.0214424951267055e-05,
|
| 979 |
+
"loss": 0.3977,
|
| 980 |
+
"step": 1040
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 27.45098039215686,
|
| 984 |
+
"grad_norm": 8.191174507141113,
|
| 985 |
+
"learning_rate": 2.997076023391813e-05,
|
| 986 |
+
"loss": 0.4704,
|
| 987 |
+
"step": 1050
|
| 988 |
+
},
|
| 989 |
+
{
|
| 990 |
+
"epoch": 27.712418300653596,
|
| 991 |
+
"grad_norm": 10.757161140441895,
|
| 992 |
+
"learning_rate": 2.9727095516569204e-05,
|
| 993 |
+
"loss": 0.4258,
|
| 994 |
+
"step": 1060
|
| 995 |
+
},
|
| 996 |
+
{
|
| 997 |
+
"epoch": 27.973856209150327,
|
| 998 |
+
"grad_norm": 7.706909656524658,
|
| 999 |
+
"learning_rate": 2.9483430799220274e-05,
|
| 1000 |
+
"loss": 0.4901,
|
| 1001 |
+
"step": 1070
|
| 1002 |
+
},
|
| 1003 |
+
{
|
| 1004 |
+
"epoch": 28.0,
|
| 1005 |
+
"eval_accuracy": 0.6985294117647058,
|
| 1006 |
+
"eval_loss": 0.7287850379943848,
|
| 1007 |
+
"eval_runtime": 23.7514,
|
| 1008 |
+
"eval_samples_per_second": 5.726,
|
| 1009 |
+
"eval_steps_per_second": 0.716,
|
| 1010 |
+
"step": 1071
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 28.235294117647058,
|
| 1014 |
+
"grad_norm": 5.9652886390686035,
|
| 1015 |
+
"learning_rate": 2.9239766081871346e-05,
|
| 1016 |
+
"loss": 0.4311,
|
| 1017 |
+
"step": 1080
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 28.49673202614379,
|
| 1021 |
+
"grad_norm": 8.528486251831055,
|
| 1022 |
+
"learning_rate": 2.8996101364522422e-05,
|
| 1023 |
+
"loss": 0.3898,
|
| 1024 |
+
"step": 1090
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 28.758169934640524,
|
| 1028 |
+
"grad_norm": 8.398402214050293,
|
| 1029 |
+
"learning_rate": 2.875243664717349e-05,
|
| 1030 |
+
"loss": 0.397,
|
| 1031 |
+
"step": 1100
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 28.99346405228758,
|
| 1035 |
+
"eval_accuracy": 0.7279411764705882,
|
| 1036 |
+
"eval_loss": 0.7250666618347168,
|
| 1037 |
+
"eval_runtime": 21.8786,
|
| 1038 |
+
"eval_samples_per_second": 6.216,
|
| 1039 |
+
"eval_steps_per_second": 0.777,
|
| 1040 |
+
"step": 1109
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 29.019607843137255,
|
| 1044 |
+
"grad_norm": 5.904468536376953,
|
| 1045 |
+
"learning_rate": 2.850877192982456e-05,
|
| 1046 |
+
"loss": 0.3617,
|
| 1047 |
+
"step": 1110
|
| 1048 |
+
},
|
| 1049 |
+
{
|
| 1050 |
+
"epoch": 29.281045751633986,
|
| 1051 |
+
"grad_norm": 5.135966777801514,
|
| 1052 |
+
"learning_rate": 2.8265107212475634e-05,
|
| 1053 |
+
"loss": 0.4141,
|
| 1054 |
+
"step": 1120
|
| 1055 |
+
},
|
| 1056 |
+
{
|
| 1057 |
+
"epoch": 29.54248366013072,
|
| 1058 |
+
"grad_norm": 7.617613792419434,
|
| 1059 |
+
"learning_rate": 2.802144249512671e-05,
|
| 1060 |
+
"loss": 0.4057,
|
| 1061 |
+
"step": 1130
|
| 1062 |
+
},
|
| 1063 |
+
{
|
| 1064 |
+
"epoch": 29.80392156862745,
|
| 1065 |
+
"grad_norm": 9.047426223754883,
|
| 1066 |
+
"learning_rate": 2.777777777777778e-05,
|
| 1067 |
+
"loss": 0.3818,
|
| 1068 |
+
"step": 1140
|
| 1069 |
+
},
|
| 1070 |
+
{
|
| 1071 |
+
"epoch": 29.986928104575163,
|
| 1072 |
+
"eval_accuracy": 0.7941176470588235,
|
| 1073 |
+
"eval_loss": 0.6250298023223877,
|
| 1074 |
+
"eval_runtime": 21.1786,
|
| 1075 |
+
"eval_samples_per_second": 6.422,
|
| 1076 |
+
"eval_steps_per_second": 0.803,
|
| 1077 |
+
"step": 1147
|
| 1078 |
+
},
|
| 1079 |
+
{
|
| 1080 |
+
"epoch": 30.065359477124183,
|
| 1081 |
+
"grad_norm": 8.303121566772461,
|
| 1082 |
+
"learning_rate": 2.7534113060428852e-05,
|
| 1083 |
+
"loss": 0.3813,
|
| 1084 |
+
"step": 1150
|
| 1085 |
+
},
|
| 1086 |
+
{
|
| 1087 |
+
"epoch": 30.326797385620914,
|
| 1088 |
+
"grad_norm": 6.311716079711914,
|
| 1089 |
+
"learning_rate": 2.729044834307992e-05,
|
| 1090 |
+
"loss": 0.3967,
|
| 1091 |
+
"step": 1160
|
| 1092 |
+
},
|
| 1093 |
+
{
|
| 1094 |
+
"epoch": 30.58823529411765,
|
| 1095 |
+
"grad_norm": 6.965037822723389,
|
| 1096 |
+
"learning_rate": 2.7046783625730997e-05,
|
| 1097 |
+
"loss": 0.4462,
|
| 1098 |
+
"step": 1170
|
| 1099 |
+
},
|
| 1100 |
+
{
|
| 1101 |
+
"epoch": 30.84967320261438,
|
| 1102 |
+
"grad_norm": 9.514204978942871,
|
| 1103 |
+
"learning_rate": 2.680311890838207e-05,
|
| 1104 |
+
"loss": 0.3412,
|
| 1105 |
+
"step": 1180
|
| 1106 |
+
},
|
| 1107 |
+
{
|
| 1108 |
+
"epoch": 30.980392156862745,
|
| 1109 |
+
"eval_accuracy": 0.7279411764705882,
|
| 1110 |
+
"eval_loss": 0.7065072059631348,
|
| 1111 |
+
"eval_runtime": 21.6767,
|
| 1112 |
+
"eval_samples_per_second": 6.274,
|
| 1113 |
+
"eval_steps_per_second": 0.784,
|
| 1114 |
+
"step": 1185
|
| 1115 |
+
},
|
| 1116 |
+
{
|
| 1117 |
+
"epoch": 31.11111111111111,
|
| 1118 |
+
"grad_norm": 9.511887550354004,
|
| 1119 |
+
"learning_rate": 2.655945419103314e-05,
|
| 1120 |
+
"loss": 0.4346,
|
| 1121 |
+
"step": 1190
|
| 1122 |
+
},
|
| 1123 |
+
{
|
| 1124 |
+
"epoch": 31.372549019607842,
|
| 1125 |
+
"grad_norm": 8.251455307006836,
|
| 1126 |
+
"learning_rate": 2.6315789473684212e-05,
|
| 1127 |
+
"loss": 0.3567,
|
| 1128 |
+
"step": 1200
|
| 1129 |
+
},
|
| 1130 |
+
{
|
| 1131 |
+
"epoch": 31.633986928104576,
|
| 1132 |
+
"grad_norm": 8.693254470825195,
|
| 1133 |
+
"learning_rate": 2.6072124756335288e-05,
|
| 1134 |
+
"loss": 0.4056,
|
| 1135 |
+
"step": 1210
|
| 1136 |
+
},
|
| 1137 |
+
{
|
| 1138 |
+
"epoch": 31.895424836601308,
|
| 1139 |
+
"grad_norm": 6.94719123840332,
|
| 1140 |
+
"learning_rate": 2.5828460038986357e-05,
|
| 1141 |
+
"loss": 0.3627,
|
| 1142 |
+
"step": 1220
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 32.0,
|
| 1146 |
+
"eval_accuracy": 0.7426470588235294,
|
| 1147 |
+
"eval_loss": 0.6876980662345886,
|
| 1148 |
+
"eval_runtime": 21.8581,
|
| 1149 |
+
"eval_samples_per_second": 6.222,
|
| 1150 |
+
"eval_steps_per_second": 0.778,
|
| 1151 |
+
"step": 1224
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 32.15686274509804,
|
| 1155 |
+
"grad_norm": 6.152271270751953,
|
| 1156 |
+
"learning_rate": 2.5584795321637427e-05,
|
| 1157 |
+
"loss": 0.294,
|
| 1158 |
+
"step": 1230
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"epoch": 32.41830065359477,
|
| 1162 |
+
"grad_norm": 9.349310874938965,
|
| 1163 |
+
"learning_rate": 2.53411306042885e-05,
|
| 1164 |
+
"loss": 0.361,
|
| 1165 |
+
"step": 1240
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 32.6797385620915,
|
| 1169 |
+
"grad_norm": 8.874456405639648,
|
| 1170 |
+
"learning_rate": 2.5097465886939576e-05,
|
| 1171 |
+
"loss": 0.3795,
|
| 1172 |
+
"step": 1250
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 32.94117647058823,
|
| 1176 |
+
"grad_norm": 6.504201412200928,
|
| 1177 |
+
"learning_rate": 2.485380116959064e-05,
|
| 1178 |
+
"loss": 0.3557,
|
| 1179 |
+
"step": 1260
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 32.99346405228758,
|
| 1183 |
+
"eval_accuracy": 0.8529411764705882,
|
| 1184 |
+
"eval_loss": 0.4244842231273651,
|
| 1185 |
+
"eval_runtime": 21.3842,
|
| 1186 |
+
"eval_samples_per_second": 6.36,
|
| 1187 |
+
"eval_steps_per_second": 0.795,
|
| 1188 |
+
"step": 1262
|
| 1189 |
+
},
|
| 1190 |
+
{
|
| 1191 |
+
"epoch": 33.20261437908497,
|
| 1192 |
+
"grad_norm": 6.195323467254639,
|
| 1193 |
+
"learning_rate": 2.4610136452241718e-05,
|
| 1194 |
+
"loss": 0.3335,
|
| 1195 |
+
"step": 1270
|
| 1196 |
+
},
|
| 1197 |
+
{
|
| 1198 |
+
"epoch": 33.4640522875817,
|
| 1199 |
+
"grad_norm": 9.20754623413086,
|
| 1200 |
+
"learning_rate": 2.4366471734892787e-05,
|
| 1201 |
+
"loss": 0.3705,
|
| 1202 |
+
"step": 1280
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 33.72549019607843,
|
| 1206 |
+
"grad_norm": 7.307880878448486,
|
| 1207 |
+
"learning_rate": 2.412280701754386e-05,
|
| 1208 |
+
"loss": 0.3626,
|
| 1209 |
+
"step": 1290
|
| 1210 |
+
},
|
| 1211 |
+
{
|
| 1212 |
+
"epoch": 33.98692810457516,
|
| 1213 |
+
"grad_norm": 11.704693794250488,
|
| 1214 |
+
"learning_rate": 2.3879142300194932e-05,
|
| 1215 |
+
"loss": 0.441,
|
| 1216 |
+
"step": 1300
|
| 1217 |
+
},
|
| 1218 |
+
{
|
| 1219 |
+
"epoch": 33.98692810457516,
|
| 1220 |
+
"eval_accuracy": 0.75,
|
| 1221 |
+
"eval_loss": 0.6974249482154846,
|
| 1222 |
+
"eval_runtime": 21.7309,
|
| 1223 |
+
"eval_samples_per_second": 6.258,
|
| 1224 |
+
"eval_steps_per_second": 0.782,
|
| 1225 |
+
"step": 1300
|
| 1226 |
+
},
|
| 1227 |
+
{
|
| 1228 |
+
"epoch": 34.248366013071895,
|
| 1229 |
+
"grad_norm": 7.3577094078063965,
|
| 1230 |
+
"learning_rate": 2.3635477582846005e-05,
|
| 1231 |
+
"loss": 0.4363,
|
| 1232 |
+
"step": 1310
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 34.509803921568626,
|
| 1236 |
+
"grad_norm": 4.33771276473999,
|
| 1237 |
+
"learning_rate": 2.3391812865497074e-05,
|
| 1238 |
+
"loss": 0.3408,
|
| 1239 |
+
"step": 1320
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 34.77124183006536,
|
| 1243 |
+
"grad_norm": 7.208470821380615,
|
| 1244 |
+
"learning_rate": 2.314814814814815e-05,
|
| 1245 |
+
"loss": 0.3036,
|
| 1246 |
+
"step": 1330
|
| 1247 |
+
},
|
| 1248 |
+
{
|
| 1249 |
+
"epoch": 34.98039215686274,
|
| 1250 |
+
"eval_accuracy": 0.7426470588235294,
|
| 1251 |
+
"eval_loss": 0.6457906365394592,
|
| 1252 |
+
"eval_runtime": 22.2732,
|
| 1253 |
+
"eval_samples_per_second": 6.106,
|
| 1254 |
+
"eval_steps_per_second": 0.763,
|
| 1255 |
+
"step": 1338
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 35.032679738562095,
|
| 1259 |
+
"grad_norm": 7.787372589111328,
|
| 1260 |
+
"learning_rate": 2.290448343079922e-05,
|
| 1261 |
+
"loss": 0.3159,
|
| 1262 |
+
"step": 1340
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 35.294117647058826,
|
| 1266 |
+
"grad_norm": 6.915111064910889,
|
| 1267 |
+
"learning_rate": 2.2660818713450292e-05,
|
| 1268 |
+
"loss": 0.3424,
|
| 1269 |
+
"step": 1350
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 35.55555555555556,
|
| 1273 |
+
"grad_norm": 8.578511238098145,
|
| 1274 |
+
"learning_rate": 2.2417153996101365e-05,
|
| 1275 |
+
"loss": 0.3452,
|
| 1276 |
+
"step": 1360
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 35.81699346405229,
|
| 1280 |
+
"grad_norm": 8.574066162109375,
|
| 1281 |
+
"learning_rate": 2.2173489278752438e-05,
|
| 1282 |
+
"loss": 0.3213,
|
| 1283 |
+
"step": 1370
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 36.0,
|
| 1287 |
+
"eval_accuracy": 0.7941176470588235,
|
| 1288 |
+
"eval_loss": 0.5578574538230896,
|
| 1289 |
+
"eval_runtime": 21.3424,
|
| 1290 |
+
"eval_samples_per_second": 6.372,
|
| 1291 |
+
"eval_steps_per_second": 0.797,
|
| 1292 |
+
"step": 1377
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 36.07843137254902,
|
| 1296 |
+
"grad_norm": 7.610787391662598,
|
| 1297 |
+
"learning_rate": 2.1929824561403507e-05,
|
| 1298 |
+
"loss": 0.4071,
|
| 1299 |
+
"step": 1380
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 36.33986928104575,
|
| 1303 |
+
"grad_norm": 8.612502098083496,
|
| 1304 |
+
"learning_rate": 2.1686159844054583e-05,
|
| 1305 |
+
"loss": 0.2995,
|
| 1306 |
+
"step": 1390
|
| 1307 |
+
},
|
| 1308 |
+
{
|
| 1309 |
+
"epoch": 36.60130718954248,
|
| 1310 |
+
"grad_norm": 8.465425491333008,
|
| 1311 |
+
"learning_rate": 2.1442495126705653e-05,
|
| 1312 |
+
"loss": 0.3456,
|
| 1313 |
+
"step": 1400
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"epoch": 36.86274509803921,
|
| 1317 |
+
"grad_norm": 7.883861541748047,
|
| 1318 |
+
"learning_rate": 2.1198830409356725e-05,
|
| 1319 |
+
"loss": 0.402,
|
| 1320 |
+
"step": 1410
|
| 1321 |
+
},
|
| 1322 |
+
{
|
| 1323 |
+
"epoch": 36.99346405228758,
|
| 1324 |
+
"eval_accuracy": 0.8382352941176471,
|
| 1325 |
+
"eval_loss": 0.4578179121017456,
|
| 1326 |
+
"eval_runtime": 20.8376,
|
| 1327 |
+
"eval_samples_per_second": 6.527,
|
| 1328 |
+
"eval_steps_per_second": 0.816,
|
| 1329 |
+
"step": 1415
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 37.12418300653595,
|
| 1333 |
+
"grad_norm": 8.734749794006348,
|
| 1334 |
+
"learning_rate": 2.0955165692007798e-05,
|
| 1335 |
+
"loss": 0.3742,
|
| 1336 |
+
"step": 1420
|
| 1337 |
+
},
|
| 1338 |
+
{
|
| 1339 |
+
"epoch": 37.38562091503268,
|
| 1340 |
+
"grad_norm": 10.651437759399414,
|
| 1341 |
+
"learning_rate": 2.071150097465887e-05,
|
| 1342 |
+
"loss": 0.3225,
|
| 1343 |
+
"step": 1430
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 37.64705882352941,
|
| 1347 |
+
"grad_norm": 8.572120666503906,
|
| 1348 |
+
"learning_rate": 2.046783625730994e-05,
|
| 1349 |
+
"loss": 0.3638,
|
| 1350 |
+
"step": 1440
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 37.908496732026144,
|
| 1354 |
+
"grad_norm": 7.853600978851318,
|
| 1355 |
+
"learning_rate": 2.0224171539961016e-05,
|
| 1356 |
+
"loss": 0.2897,
|
| 1357 |
+
"step": 1450
|
| 1358 |
+
},
|
| 1359 |
+
{
|
| 1360 |
+
"epoch": 37.98692810457516,
|
| 1361 |
+
"eval_accuracy": 0.7867647058823529,
|
| 1362 |
+
"eval_loss": 0.5368921756744385,
|
| 1363 |
+
"eval_runtime": 21.9563,
|
| 1364 |
+
"eval_samples_per_second": 6.194,
|
| 1365 |
+
"eval_steps_per_second": 0.774,
|
| 1366 |
+
"step": 1453
|
| 1367 |
+
},
|
| 1368 |
+
{
|
| 1369 |
+
"epoch": 38.169934640522875,
|
| 1370 |
+
"grad_norm": 7.19242000579834,
|
| 1371 |
+
"learning_rate": 1.9980506822612085e-05,
|
| 1372 |
+
"loss": 0.3608,
|
| 1373 |
+
"step": 1460
|
| 1374 |
+
},
|
| 1375 |
+
{
|
| 1376 |
+
"epoch": 38.431372549019606,
|
| 1377 |
+
"grad_norm": 5.555720329284668,
|
| 1378 |
+
"learning_rate": 1.9736842105263158e-05,
|
| 1379 |
+
"loss": 0.3496,
|
| 1380 |
+
"step": 1470
|
| 1381 |
+
},
|
| 1382 |
+
{
|
| 1383 |
+
"epoch": 38.69281045751634,
|
| 1384 |
+
"grad_norm": 7.676553249359131,
|
| 1385 |
+
"learning_rate": 1.949317738791423e-05,
|
| 1386 |
+
"loss": 0.3087,
|
| 1387 |
+
"step": 1480
|
| 1388 |
+
},
|
| 1389 |
+
{
|
| 1390 |
+
"epoch": 38.95424836601307,
|
| 1391 |
+
"grad_norm": 9.612250328063965,
|
| 1392 |
+
"learning_rate": 1.9249512670565304e-05,
|
| 1393 |
+
"loss": 0.348,
|
| 1394 |
+
"step": 1490
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 38.98039215686274,
|
| 1398 |
+
"eval_accuracy": 0.7941176470588235,
|
| 1399 |
+
"eval_loss": 0.6819477081298828,
|
| 1400 |
+
"eval_runtime": 20.1264,
|
| 1401 |
+
"eval_samples_per_second": 6.757,
|
| 1402 |
+
"eval_steps_per_second": 0.845,
|
| 1403 |
+
"step": 1491
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 39.21568627450981,
|
| 1407 |
+
"grad_norm": 5.8754167556762695,
|
| 1408 |
+
"learning_rate": 1.9005847953216373e-05,
|
| 1409 |
+
"loss": 0.2784,
|
| 1410 |
+
"step": 1500
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 39.47712418300654,
|
| 1414 |
+
"grad_norm": 9.667128562927246,
|
| 1415 |
+
"learning_rate": 1.876218323586745e-05,
|
| 1416 |
+
"loss": 0.3254,
|
| 1417 |
+
"step": 1510
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 39.73856209150327,
|
| 1421 |
+
"grad_norm": 7.417483329772949,
|
| 1422 |
+
"learning_rate": 1.8518518518518518e-05,
|
| 1423 |
+
"loss": 0.3416,
|
| 1424 |
+
"step": 1520
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 40.0,
|
| 1428 |
+
"grad_norm": 10.744291305541992,
|
| 1429 |
+
"learning_rate": 1.827485380116959e-05,
|
| 1430 |
+
"loss": 0.3929,
|
| 1431 |
+
"step": 1530
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 40.0,
|
| 1435 |
+
"eval_accuracy": 0.7867647058823529,
|
| 1436 |
+
"eval_loss": 0.5809765458106995,
|
| 1437 |
+
"eval_runtime": 22.4901,
|
| 1438 |
+
"eval_samples_per_second": 6.047,
|
| 1439 |
+
"eval_steps_per_second": 0.756,
|
| 1440 |
+
"step": 1530
|
| 1441 |
+
},
|
| 1442 |
+
{
|
| 1443 |
+
"epoch": 40.26143790849673,
|
| 1444 |
+
"grad_norm": 8.990880012512207,
|
| 1445 |
+
"learning_rate": 1.8031189083820664e-05,
|
| 1446 |
+
"loss": 0.3546,
|
| 1447 |
+
"step": 1540
|
| 1448 |
+
},
|
| 1449 |
+
{
|
| 1450 |
+
"epoch": 40.52287581699346,
|
| 1451 |
+
"grad_norm": 9.770576477050781,
|
| 1452 |
+
"learning_rate": 1.7787524366471736e-05,
|
| 1453 |
+
"loss": 0.344,
|
| 1454 |
+
"step": 1550
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 40.78431372549019,
|
| 1458 |
+
"grad_norm": 7.653594017028809,
|
| 1459 |
+
"learning_rate": 1.7543859649122806e-05,
|
| 1460 |
+
"loss": 0.3173,
|
| 1461 |
+
"step": 1560
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 40.99346405228758,
|
| 1465 |
+
"eval_accuracy": 0.7426470588235294,
|
| 1466 |
+
"eval_loss": 0.7875370979309082,
|
| 1467 |
+
"eval_runtime": 20.0479,
|
| 1468 |
+
"eval_samples_per_second": 6.784,
|
| 1469 |
+
"eval_steps_per_second": 0.848,
|
| 1470 |
+
"step": 1568
|
| 1471 |
+
},
|
| 1472 |
+
{
|
| 1473 |
+
"epoch": 41.04575163398693,
|
| 1474 |
+
"grad_norm": 10.298673629760742,
|
| 1475 |
+
"learning_rate": 1.7300194931773882e-05,
|
| 1476 |
+
"loss": 0.3148,
|
| 1477 |
+
"step": 1570
|
| 1478 |
+
},
|
| 1479 |
+
{
|
| 1480 |
+
"epoch": 41.30718954248366,
|
| 1481 |
+
"grad_norm": 9.334877967834473,
|
| 1482 |
+
"learning_rate": 1.705653021442495e-05,
|
| 1483 |
+
"loss": 0.295,
|
| 1484 |
+
"step": 1580
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 41.568627450980394,
|
| 1488 |
+
"grad_norm": 6.963639736175537,
|
| 1489 |
+
"learning_rate": 1.6812865497076024e-05,
|
| 1490 |
+
"loss": 0.3274,
|
| 1491 |
+
"step": 1590
|
| 1492 |
+
},
|
| 1493 |
+
{
|
| 1494 |
+
"epoch": 41.830065359477125,
|
| 1495 |
+
"grad_norm": 9.359763145446777,
|
| 1496 |
+
"learning_rate": 1.6569200779727097e-05,
|
| 1497 |
+
"loss": 0.3499,
|
| 1498 |
+
"step": 1600
|
| 1499 |
+
},
|
| 1500 |
+
{
|
| 1501 |
+
"epoch": 41.98692810457516,
|
| 1502 |
+
"eval_accuracy": 0.8014705882352942,
|
| 1503 |
+
"eval_loss": 0.5050505995750427,
|
| 1504 |
+
"eval_runtime": 19.6525,
|
| 1505 |
+
"eval_samples_per_second": 6.92,
|
| 1506 |
+
"eval_steps_per_second": 0.865,
|
| 1507 |
+
"step": 1606
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 42.091503267973856,
|
| 1511 |
+
"grad_norm": 8.404752731323242,
|
| 1512 |
+
"learning_rate": 1.632553606237817e-05,
|
| 1513 |
+
"loss": 0.3297,
|
| 1514 |
+
"step": 1610
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 42.35294117647059,
|
| 1518 |
+
"grad_norm": 7.797074794769287,
|
| 1519 |
+
"learning_rate": 1.608187134502924e-05,
|
| 1520 |
+
"loss": 0.2784,
|
| 1521 |
+
"step": 1620
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 42.61437908496732,
|
| 1525 |
+
"grad_norm": 6.273430347442627,
|
| 1526 |
+
"learning_rate": 1.5838206627680315e-05,
|
| 1527 |
+
"loss": 0.3084,
|
| 1528 |
+
"step": 1630
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 42.87581699346405,
|
| 1532 |
+
"grad_norm": 8.596699714660645,
|
| 1533 |
+
"learning_rate": 1.5594541910331384e-05,
|
| 1534 |
+
"loss": 0.3053,
|
| 1535 |
+
"step": 1640
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 42.98039215686274,
|
| 1539 |
+
"eval_accuracy": 0.7426470588235294,
|
| 1540 |
+
"eval_loss": 0.751028299331665,
|
| 1541 |
+
"eval_runtime": 19.7977,
|
| 1542 |
+
"eval_samples_per_second": 6.869,
|
| 1543 |
+
"eval_steps_per_second": 0.859,
|
| 1544 |
+
"step": 1644
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 43.13725490196079,
|
| 1548 |
+
"grad_norm": 10.001852989196777,
|
| 1549 |
+
"learning_rate": 1.5350877192982457e-05,
|
| 1550 |
+
"loss": 0.3818,
|
| 1551 |
+
"step": 1650
|
| 1552 |
+
},
|
| 1553 |
+
{
|
| 1554 |
+
"epoch": 43.39869281045752,
|
| 1555 |
+
"grad_norm": 6.75031042098999,
|
| 1556 |
+
"learning_rate": 1.5107212475633528e-05,
|
| 1557 |
+
"loss": 0.2612,
|
| 1558 |
+
"step": 1660
|
| 1559 |
+
},
|
| 1560 |
+
{
|
| 1561 |
+
"epoch": 43.66013071895425,
|
| 1562 |
+
"grad_norm": 4.408693790435791,
|
| 1563 |
+
"learning_rate": 1.4863547758284602e-05,
|
| 1564 |
+
"loss": 0.2901,
|
| 1565 |
+
"step": 1670
|
| 1566 |
+
},
|
| 1567 |
+
{
|
| 1568 |
+
"epoch": 43.92156862745098,
|
| 1569 |
+
"grad_norm": 9.371530532836914,
|
| 1570 |
+
"learning_rate": 1.4619883040935673e-05,
|
| 1571 |
+
"loss": 0.4109,
|
| 1572 |
+
"step": 1680
|
| 1573 |
+
},
|
| 1574 |
+
{
|
| 1575 |
+
"epoch": 44.0,
|
| 1576 |
+
"eval_accuracy": 0.75,
|
| 1577 |
+
"eval_loss": 0.6528570055961609,
|
| 1578 |
+
"eval_runtime": 20.0083,
|
| 1579 |
+
"eval_samples_per_second": 6.797,
|
| 1580 |
+
"eval_steps_per_second": 0.85,
|
| 1581 |
+
"step": 1683
|
| 1582 |
+
},
|
| 1583 |
+
{
|
| 1584 |
+
"epoch": 44.18300653594771,
|
| 1585 |
+
"grad_norm": 7.803284168243408,
|
| 1586 |
+
"learning_rate": 1.4376218323586746e-05,
|
| 1587 |
+
"loss": 0.2882,
|
| 1588 |
+
"step": 1690
|
| 1589 |
+
},
|
| 1590 |
+
{
|
| 1591 |
+
"epoch": 44.44444444444444,
|
| 1592 |
+
"grad_norm": 6.260004043579102,
|
| 1593 |
+
"learning_rate": 1.4132553606237817e-05,
|
| 1594 |
+
"loss": 0.333,
|
| 1595 |
+
"step": 1700
|
| 1596 |
+
},
|
| 1597 |
+
{
|
| 1598 |
+
"epoch": 44.705882352941174,
|
| 1599 |
+
"grad_norm": 11.425740242004395,
|
| 1600 |
+
"learning_rate": 1.388888888888889e-05,
|
| 1601 |
+
"loss": 0.3003,
|
| 1602 |
+
"step": 1710
|
| 1603 |
+
},
|
| 1604 |
+
{
|
| 1605 |
+
"epoch": 44.967320261437905,
|
| 1606 |
+
"grad_norm": 4.394773960113525,
|
| 1607 |
+
"learning_rate": 1.364522417153996e-05,
|
| 1608 |
+
"loss": 0.3846,
|
| 1609 |
+
"step": 1720
|
| 1610 |
+
},
|
| 1611 |
+
{
|
| 1612 |
+
"epoch": 44.99346405228758,
|
| 1613 |
+
"eval_accuracy": 0.7132352941176471,
|
| 1614 |
+
"eval_loss": 0.961465060710907,
|
| 1615 |
+
"eval_runtime": 19.8391,
|
| 1616 |
+
"eval_samples_per_second": 6.855,
|
| 1617 |
+
"eval_steps_per_second": 0.857,
|
| 1618 |
+
"step": 1721
|
| 1619 |
+
},
|
| 1620 |
+
{
|
| 1621 |
+
"epoch": 45.22875816993464,
|
| 1622 |
+
"grad_norm": 11.91213321685791,
|
| 1623 |
+
"learning_rate": 1.3401559454191035e-05,
|
| 1624 |
+
"loss": 0.3757,
|
| 1625 |
+
"step": 1730
|
| 1626 |
+
},
|
| 1627 |
+
{
|
| 1628 |
+
"epoch": 45.490196078431374,
|
| 1629 |
+
"grad_norm": 7.775343894958496,
|
| 1630 |
+
"learning_rate": 1.3157894736842106e-05,
|
| 1631 |
+
"loss": 0.3259,
|
| 1632 |
+
"step": 1740
|
| 1633 |
+
},
|
| 1634 |
+
{
|
| 1635 |
+
"epoch": 45.751633986928105,
|
| 1636 |
+
"grad_norm": 4.623126029968262,
|
| 1637 |
+
"learning_rate": 1.2914230019493179e-05,
|
| 1638 |
+
"loss": 0.3222,
|
| 1639 |
+
"step": 1750
|
| 1640 |
+
},
|
| 1641 |
+
{
|
| 1642 |
+
"epoch": 45.98692810457516,
|
| 1643 |
+
"eval_accuracy": 0.6691176470588235,
|
| 1644 |
+
"eval_loss": 0.8888900279998779,
|
| 1645 |
+
"eval_runtime": 19.6212,
|
| 1646 |
+
"eval_samples_per_second": 6.931,
|
| 1647 |
+
"eval_steps_per_second": 0.866,
|
| 1648 |
+
"step": 1759
|
| 1649 |
+
},
|
| 1650 |
+
{
|
| 1651 |
+
"epoch": 46.01307189542484,
|
| 1652 |
+
"grad_norm": 7.839041233062744,
|
| 1653 |
+
"learning_rate": 1.267056530214425e-05,
|
| 1654 |
+
"loss": 0.2718,
|
| 1655 |
+
"step": 1760
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 46.27450980392157,
|
| 1659 |
+
"grad_norm": 10.323081016540527,
|
| 1660 |
+
"learning_rate": 1.242690058479532e-05,
|
| 1661 |
+
"loss": 0.2538,
|
| 1662 |
+
"step": 1770
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 46.5359477124183,
|
| 1666 |
+
"grad_norm": 8.446836471557617,
|
| 1667 |
+
"learning_rate": 1.2183235867446393e-05,
|
| 1668 |
+
"loss": 0.2974,
|
| 1669 |
+
"step": 1780
|
| 1670 |
+
},
|
| 1671 |
+
{
|
| 1672 |
+
"epoch": 46.79738562091503,
|
| 1673 |
+
"grad_norm": 8.584868431091309,
|
| 1674 |
+
"learning_rate": 1.1939571150097466e-05,
|
| 1675 |
+
"loss": 0.3293,
|
| 1676 |
+
"step": 1790
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 46.98039215686274,
|
| 1680 |
+
"eval_accuracy": 0.8676470588235294,
|
| 1681 |
+
"eval_loss": 0.46983805298805237,
|
| 1682 |
+
"eval_runtime": 20.3581,
|
| 1683 |
+
"eval_samples_per_second": 6.68,
|
| 1684 |
+
"eval_steps_per_second": 0.835,
|
| 1685 |
+
"step": 1797
|
| 1686 |
+
},
|
| 1687 |
+
{
|
| 1688 |
+
"epoch": 47.05882352941177,
|
| 1689 |
+
"grad_norm": 6.30557918548584,
|
| 1690 |
+
"learning_rate": 1.1695906432748537e-05,
|
| 1691 |
+
"loss": 0.2915,
|
| 1692 |
+
"step": 1800
|
| 1693 |
+
},
|
| 1694 |
+
{
|
| 1695 |
+
"epoch": 47.3202614379085,
|
| 1696 |
+
"grad_norm": 6.145007610321045,
|
| 1697 |
+
"learning_rate": 1.145224171539961e-05,
|
| 1698 |
+
"loss": 0.2948,
|
| 1699 |
+
"step": 1810
|
| 1700 |
+
},
|
| 1701 |
+
{
|
| 1702 |
+
"epoch": 47.58169934640523,
|
| 1703 |
+
"grad_norm": 8.976215362548828,
|
| 1704 |
+
"learning_rate": 1.1208576998050683e-05,
|
| 1705 |
+
"loss": 0.3189,
|
| 1706 |
+
"step": 1820
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 47.84313725490196,
|
| 1710 |
+
"grad_norm": 5.437533855438232,
|
| 1711 |
+
"learning_rate": 1.0964912280701754e-05,
|
| 1712 |
+
"loss": 0.293,
|
| 1713 |
+
"step": 1830
|
| 1714 |
+
},
|
| 1715 |
+
{
|
| 1716 |
+
"epoch": 48.0,
|
| 1717 |
+
"eval_accuracy": 0.8014705882352942,
|
| 1718 |
+
"eval_loss": 0.5995640754699707,
|
| 1719 |
+
"eval_runtime": 22.4921,
|
| 1720 |
+
"eval_samples_per_second": 6.047,
|
| 1721 |
+
"eval_steps_per_second": 0.756,
|
| 1722 |
+
"step": 1836
|
| 1723 |
+
},
|
| 1724 |
+
{
|
| 1725 |
+
"epoch": 48.10457516339869,
|
| 1726 |
+
"grad_norm": 7.454949855804443,
|
| 1727 |
+
"learning_rate": 1.0721247563352826e-05,
|
| 1728 |
+
"loss": 0.3486,
|
| 1729 |
+
"step": 1840
|
| 1730 |
+
},
|
| 1731 |
+
{
|
| 1732 |
+
"epoch": 48.36601307189542,
|
| 1733 |
+
"grad_norm": 5.587437152862549,
|
| 1734 |
+
"learning_rate": 1.0477582846003899e-05,
|
| 1735 |
+
"loss": 0.3269,
|
| 1736 |
+
"step": 1850
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 48.627450980392155,
|
| 1740 |
+
"grad_norm": 7.708277702331543,
|
| 1741 |
+
"learning_rate": 1.023391812865497e-05,
|
| 1742 |
+
"loss": 0.3436,
|
| 1743 |
+
"step": 1860
|
| 1744 |
+
},
|
| 1745 |
+
{
|
| 1746 |
+
"epoch": 48.888888888888886,
|
| 1747 |
+
"grad_norm": 9.139860153198242,
|
| 1748 |
+
"learning_rate": 9.990253411306043e-06,
|
| 1749 |
+
"loss": 0.2363,
|
| 1750 |
+
"step": 1870
|
| 1751 |
+
},
|
| 1752 |
+
{
|
| 1753 |
+
"epoch": 48.99346405228758,
|
| 1754 |
+
"eval_accuracy": 0.8308823529411765,
|
| 1755 |
+
"eval_loss": 0.5006946921348572,
|
| 1756 |
+
"eval_runtime": 22.6256,
|
| 1757 |
+
"eval_samples_per_second": 6.011,
|
| 1758 |
+
"eval_steps_per_second": 0.751,
|
| 1759 |
+
"step": 1874
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 49.150326797385624,
|
| 1763 |
+
"grad_norm": 10.596453666687012,
|
| 1764 |
+
"learning_rate": 9.746588693957115e-06,
|
| 1765 |
+
"loss": 0.3911,
|
| 1766 |
+
"step": 1880
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 49.411764705882355,
|
| 1770 |
+
"grad_norm": 7.040250301361084,
|
| 1771 |
+
"learning_rate": 9.502923976608186e-06,
|
| 1772 |
+
"loss": 0.3426,
|
| 1773 |
+
"step": 1890
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 49.673202614379086,
|
| 1777 |
+
"grad_norm": 8.111015319824219,
|
| 1778 |
+
"learning_rate": 9.259259259259259e-06,
|
| 1779 |
+
"loss": 0.3189,
|
| 1780 |
+
"step": 1900
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 49.93464052287582,
|
| 1784 |
+
"grad_norm": 4.160747528076172,
|
| 1785 |
+
"learning_rate": 9.015594541910332e-06,
|
| 1786 |
+
"loss": 0.2811,
|
| 1787 |
+
"step": 1910
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 49.98692810457516,
|
| 1791 |
+
"eval_accuracy": 0.7941176470588235,
|
| 1792 |
+
"eval_loss": 0.6748296022415161,
|
| 1793 |
+
"eval_runtime": 21.1042,
|
| 1794 |
+
"eval_samples_per_second": 6.444,
|
| 1795 |
+
"eval_steps_per_second": 0.806,
|
| 1796 |
+
"step": 1912
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 50.19607843137255,
|
| 1800 |
+
"grad_norm": 7.058052062988281,
|
| 1801 |
+
"learning_rate": 8.771929824561403e-06,
|
| 1802 |
+
"loss": 0.2881,
|
| 1803 |
+
"step": 1920
|
| 1804 |
+
},
|
| 1805 |
+
{
|
| 1806 |
+
"epoch": 50.45751633986928,
|
| 1807 |
+
"grad_norm": 5.950506210327148,
|
| 1808 |
+
"learning_rate": 8.528265107212476e-06,
|
| 1809 |
+
"loss": 0.2872,
|
| 1810 |
+
"step": 1930
|
| 1811 |
+
},
|
| 1812 |
+
{
|
| 1813 |
+
"epoch": 50.71895424836601,
|
| 1814 |
+
"grad_norm": 7.726902484893799,
|
| 1815 |
+
"learning_rate": 8.284600389863548e-06,
|
| 1816 |
+
"loss": 0.3346,
|
| 1817 |
+
"step": 1940
|
| 1818 |
+
},
|
| 1819 |
+
{
|
| 1820 |
+
"epoch": 50.98039215686274,
|
| 1821 |
+
"grad_norm": 7.297433376312256,
|
| 1822 |
+
"learning_rate": 8.04093567251462e-06,
|
| 1823 |
+
"loss": 0.2403,
|
| 1824 |
+
"step": 1950
|
| 1825 |
+
},
|
| 1826 |
+
{
|
| 1827 |
+
"epoch": 50.98039215686274,
|
| 1828 |
+
"eval_accuracy": 0.7941176470588235,
|
| 1829 |
+
"eval_loss": 0.6595063209533691,
|
| 1830 |
+
"eval_runtime": 21.555,
|
| 1831 |
+
"eval_samples_per_second": 6.309,
|
| 1832 |
+
"eval_steps_per_second": 0.789,
|
| 1833 |
+
"step": 1950
|
| 1834 |
+
},
|
| 1835 |
+
{
|
| 1836 |
+
"epoch": 51.24183006535948,
|
| 1837 |
+
"grad_norm": 7.367440700531006,
|
| 1838 |
+
"learning_rate": 7.797270955165692e-06,
|
| 1839 |
+
"loss": 0.2824,
|
| 1840 |
+
"step": 1960
|
| 1841 |
+
},
|
| 1842 |
+
{
|
| 1843 |
+
"epoch": 51.50326797385621,
|
| 1844 |
+
"grad_norm": 10.130813598632812,
|
| 1845 |
+
"learning_rate": 7.553606237816764e-06,
|
| 1846 |
+
"loss": 0.3367,
|
| 1847 |
+
"step": 1970
|
| 1848 |
+
},
|
| 1849 |
+
{
|
| 1850 |
+
"epoch": 51.76470588235294,
|
| 1851 |
+
"grad_norm": 7.926657199859619,
|
| 1852 |
+
"learning_rate": 7.3099415204678366e-06,
|
| 1853 |
+
"loss": 0.2553,
|
| 1854 |
+
"step": 1980
|
| 1855 |
+
},
|
| 1856 |
+
{
|
| 1857 |
+
"epoch": 52.0,
|
| 1858 |
+
"eval_accuracy": 0.7794117647058824,
|
| 1859 |
+
"eval_loss": 0.5987350344657898,
|
| 1860 |
+
"eval_runtime": 20.0722,
|
| 1861 |
+
"eval_samples_per_second": 6.776,
|
| 1862 |
+
"eval_steps_per_second": 0.847,
|
| 1863 |
+
"step": 1989
|
| 1864 |
+
},
|
| 1865 |
+
{
|
| 1866 |
+
"epoch": 52.02614379084967,
|
| 1867 |
+
"grad_norm": 8.90279769897461,
|
| 1868 |
+
"learning_rate": 7.066276803118908e-06,
|
| 1869 |
+
"loss": 0.3068,
|
| 1870 |
+
"step": 1990
|
| 1871 |
+
},
|
| 1872 |
+
{
|
| 1873 |
+
"epoch": 52.287581699346404,
|
| 1874 |
+
"grad_norm": 11.345956802368164,
|
| 1875 |
+
"learning_rate": 6.82261208576998e-06,
|
| 1876 |
+
"loss": 0.2884,
|
| 1877 |
+
"step": 2000
|
| 1878 |
+
},
|
| 1879 |
+
{
|
| 1880 |
+
"epoch": 52.549019607843135,
|
| 1881 |
+
"grad_norm": 8.919641494750977,
|
| 1882 |
+
"learning_rate": 6.578947368421053e-06,
|
| 1883 |
+
"loss": 0.295,
|
| 1884 |
+
"step": 2010
|
| 1885 |
+
},
|
| 1886 |
+
{
|
| 1887 |
+
"epoch": 52.810457516339866,
|
| 1888 |
+
"grad_norm": 7.802643775939941,
|
| 1889 |
+
"learning_rate": 6.335282651072125e-06,
|
| 1890 |
+
"loss": 0.2959,
|
| 1891 |
+
"step": 2020
|
| 1892 |
+
},
|
| 1893 |
+
{
|
| 1894 |
+
"epoch": 52.99346405228758,
|
| 1895 |
+
"eval_accuracy": 0.8235294117647058,
|
| 1896 |
+
"eval_loss": 0.5459432005882263,
|
| 1897 |
+
"eval_runtime": 20.7402,
|
| 1898 |
+
"eval_samples_per_second": 6.557,
|
| 1899 |
+
"eval_steps_per_second": 0.82,
|
| 1900 |
+
"step": 2027
|
| 1901 |
+
},
|
| 1902 |
+
{
|
| 1903 |
+
"epoch": 53.071895424836605,
|
| 1904 |
+
"grad_norm": 6.629459381103516,
|
| 1905 |
+
"learning_rate": 6.091617933723197e-06,
|
| 1906 |
+
"loss": 0.3203,
|
| 1907 |
+
"step": 2030
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"epoch": 53.333333333333336,
|
| 1911 |
+
"grad_norm": 7.91646671295166,
|
| 1912 |
+
"learning_rate": 5.8479532163742686e-06,
|
| 1913 |
+
"loss": 0.3267,
|
| 1914 |
+
"step": 2040
|
| 1915 |
+
},
|
| 1916 |
+
{
|
| 1917 |
+
"epoch": 53.59477124183007,
|
| 1918 |
+
"grad_norm": 5.941601276397705,
|
| 1919 |
+
"learning_rate": 5.604288499025341e-06,
|
| 1920 |
+
"loss": 0.2962,
|
| 1921 |
+
"step": 2050
|
| 1922 |
+
},
|
| 1923 |
+
{
|
| 1924 |
+
"epoch": 53.8562091503268,
|
| 1925 |
+
"grad_norm": 7.222002029418945,
|
| 1926 |
+
"learning_rate": 5.360623781676413e-06,
|
| 1927 |
+
"loss": 0.3066,
|
| 1928 |
+
"step": 2060
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 53.98692810457516,
|
| 1932 |
+
"eval_accuracy": 0.7867647058823529,
|
| 1933 |
+
"eval_loss": 0.6198421716690063,
|
| 1934 |
+
"eval_runtime": 20.555,
|
| 1935 |
+
"eval_samples_per_second": 6.616,
|
| 1936 |
+
"eval_steps_per_second": 0.827,
|
| 1937 |
+
"step": 2065
|
| 1938 |
+
},
|
| 1939 |
+
{
|
| 1940 |
+
"epoch": 54.11764705882353,
|
| 1941 |
+
"grad_norm": 6.425452709197998,
|
| 1942 |
+
"learning_rate": 5.116959064327485e-06,
|
| 1943 |
+
"loss": 0.2961,
|
| 1944 |
+
"step": 2070
|
| 1945 |
+
},
|
| 1946 |
+
{
|
| 1947 |
+
"epoch": 54.37908496732026,
|
| 1948 |
+
"grad_norm": 9.548799514770508,
|
| 1949 |
+
"learning_rate": 4.873294346978558e-06,
|
| 1950 |
+
"loss": 0.3363,
|
| 1951 |
+
"step": 2080
|
| 1952 |
+
},
|
| 1953 |
+
{
|
| 1954 |
+
"epoch": 54.64052287581699,
|
| 1955 |
+
"grad_norm": 7.3025970458984375,
|
| 1956 |
+
"learning_rate": 4.6296296296296296e-06,
|
| 1957 |
+
"loss": 0.3843,
|
| 1958 |
+
"step": 2090
|
| 1959 |
+
},
|
| 1960 |
+
{
|
| 1961 |
+
"epoch": 54.90196078431372,
|
| 1962 |
+
"grad_norm": 8.06686019897461,
|
| 1963 |
+
"learning_rate": 4.3859649122807014e-06,
|
| 1964 |
+
"loss": 0.2981,
|
| 1965 |
+
"step": 2100
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 54.98039215686274,
|
| 1969 |
+
"eval_accuracy": 0.8308823529411765,
|
| 1970 |
+
"eval_loss": 0.4885517656803131,
|
| 1971 |
+
"eval_runtime": 22.3067,
|
| 1972 |
+
"eval_samples_per_second": 6.097,
|
| 1973 |
+
"eval_steps_per_second": 0.762,
|
| 1974 |
+
"step": 2103
|
| 1975 |
+
},
|
| 1976 |
+
{
|
| 1977 |
+
"epoch": 55.16339869281046,
|
| 1978 |
+
"grad_norm": 7.39772891998291,
|
| 1979 |
+
"learning_rate": 4.142300194931774e-06,
|
| 1980 |
+
"loss": 0.315,
|
| 1981 |
+
"step": 2110
|
| 1982 |
+
},
|
| 1983 |
+
{
|
| 1984 |
+
"epoch": 55.42483660130719,
|
| 1985 |
+
"grad_norm": 6.316030502319336,
|
| 1986 |
+
"learning_rate": 3.898635477582846e-06,
|
| 1987 |
+
"loss": 0.2416,
|
| 1988 |
+
"step": 2120
|
| 1989 |
+
},
|
| 1990 |
+
{
|
| 1991 |
+
"epoch": 55.68627450980392,
|
| 1992 |
+
"grad_norm": 9.764727592468262,
|
| 1993 |
+
"learning_rate": 3.6549707602339183e-06,
|
| 1994 |
+
"loss": 0.3007,
|
| 1995 |
+
"step": 2130
|
| 1996 |
+
},
|
| 1997 |
+
{
|
| 1998 |
+
"epoch": 55.947712418300654,
|
| 1999 |
+
"grad_norm": 9.428828239440918,
|
| 2000 |
+
"learning_rate": 3.41130604288499e-06,
|
| 2001 |
+
"loss": 0.2658,
|
| 2002 |
+
"step": 2140
|
| 2003 |
+
},
|
| 2004 |
+
{
|
| 2005 |
+
"epoch": 56.0,
|
| 2006 |
+
"eval_accuracy": 0.7794117647058824,
|
| 2007 |
+
"eval_loss": 0.6422163844108582,
|
| 2008 |
+
"eval_runtime": 20.6286,
|
| 2009 |
+
"eval_samples_per_second": 6.593,
|
| 2010 |
+
"eval_steps_per_second": 0.824,
|
| 2011 |
+
"step": 2142
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 56.209150326797385,
|
| 2015 |
+
"grad_norm": 11.038329124450684,
|
| 2016 |
+
"learning_rate": 3.1676413255360624e-06,
|
| 2017 |
+
"loss": 0.3298,
|
| 2018 |
+
"step": 2150
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 56.470588235294116,
|
| 2022 |
+
"grad_norm": 7.585492134094238,
|
| 2023 |
+
"learning_rate": 2.9239766081871343e-06,
|
| 2024 |
+
"loss": 0.3365,
|
| 2025 |
+
"step": 2160
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 56.73202614379085,
|
| 2029 |
+
"grad_norm": 11.289450645446777,
|
| 2030 |
+
"learning_rate": 2.6803118908382066e-06,
|
| 2031 |
+
"loss": 0.3559,
|
| 2032 |
+
"step": 2170
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 56.99346405228758,
|
| 2036 |
+
"grad_norm": 8.215266227722168,
|
| 2037 |
+
"learning_rate": 2.436647173489279e-06,
|
| 2038 |
+
"loss": 0.2371,
|
| 2039 |
+
"step": 2180
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 56.99346405228758,
|
| 2043 |
+
"eval_accuracy": 0.8382352941176471,
|
| 2044 |
+
"eval_loss": 0.4999590516090393,
|
| 2045 |
+
"eval_runtime": 19.8968,
|
| 2046 |
+
"eval_samples_per_second": 6.835,
|
| 2047 |
+
"eval_steps_per_second": 0.854,
|
| 2048 |
+
"step": 2180
|
| 2049 |
+
},
|
| 2050 |
+
{
|
| 2051 |
+
"epoch": 57.254901960784316,
|
| 2052 |
+
"grad_norm": 6.436427116394043,
|
| 2053 |
+
"learning_rate": 2.1929824561403507e-06,
|
| 2054 |
+
"loss": 0.2826,
|
| 2055 |
+
"step": 2190
|
| 2056 |
+
},
|
| 2057 |
+
{
|
| 2058 |
+
"epoch": 57.51633986928105,
|
| 2059 |
+
"grad_norm": 10.186029434204102,
|
| 2060 |
+
"learning_rate": 1.949317738791423e-06,
|
| 2061 |
+
"loss": 0.3155,
|
| 2062 |
+
"step": 2200
|
| 2063 |
+
},
|
| 2064 |
+
{
|
| 2065 |
+
"epoch": 57.77777777777778,
|
| 2066 |
+
"grad_norm": 5.971257209777832,
|
| 2067 |
+
"learning_rate": 1.705653021442495e-06,
|
| 2068 |
+
"loss": 0.2331,
|
| 2069 |
+
"step": 2210
|
| 2070 |
+
},
|
| 2071 |
+
{
|
| 2072 |
+
"epoch": 57.98692810457516,
|
| 2073 |
+
"eval_accuracy": 0.7132352941176471,
|
| 2074 |
+
"eval_loss": 0.8853857517242432,
|
| 2075 |
+
"eval_runtime": 21.9358,
|
| 2076 |
+
"eval_samples_per_second": 6.2,
|
| 2077 |
+
"eval_steps_per_second": 0.775,
|
| 2078 |
+
"step": 2218
|
| 2079 |
+
},
|
| 2080 |
+
{
|
| 2081 |
+
"epoch": 58.03921568627451,
|
| 2082 |
+
"grad_norm": 5.772573947906494,
|
| 2083 |
+
"learning_rate": 1.4619883040935671e-06,
|
| 2084 |
+
"loss": 0.2802,
|
| 2085 |
+
"step": 2220
|
| 2086 |
+
},
|
| 2087 |
+
{
|
| 2088 |
+
"epoch": 58.30065359477124,
|
| 2089 |
+
"grad_norm": 8.661057472229004,
|
| 2090 |
+
"learning_rate": 1.2183235867446394e-06,
|
| 2091 |
+
"loss": 0.286,
|
| 2092 |
+
"step": 2230
|
| 2093 |
+
},
|
| 2094 |
+
{
|
| 2095 |
+
"epoch": 58.56209150326797,
|
| 2096 |
+
"grad_norm": 7.350020408630371,
|
| 2097 |
+
"learning_rate": 9.746588693957115e-07,
|
| 2098 |
+
"loss": 0.2523,
|
| 2099 |
+
"step": 2240
|
| 2100 |
+
},
|
| 2101 |
+
{
|
| 2102 |
+
"epoch": 58.8235294117647,
|
| 2103 |
+
"grad_norm": 10.219001770019531,
|
| 2104 |
+
"learning_rate": 7.309941520467836e-07,
|
| 2105 |
+
"loss": 0.2777,
|
| 2106 |
+
"step": 2250
|
| 2107 |
+
},
|
| 2108 |
+
{
|
| 2109 |
+
"epoch": 58.98039215686274,
|
| 2110 |
+
"eval_accuracy": 0.8014705882352942,
|
| 2111 |
+
"eval_loss": 0.6190231442451477,
|
| 2112 |
+
"eval_runtime": 19.9884,
|
| 2113 |
+
"eval_samples_per_second": 6.804,
|
| 2114 |
+
"eval_steps_per_second": 0.85,
|
| 2115 |
+
"step": 2256
|
| 2116 |
+
},
|
| 2117 |
+
{
|
| 2118 |
+
"epoch": 59.08496732026144,
|
| 2119 |
+
"grad_norm": 6.8238396644592285,
|
| 2120 |
+
"learning_rate": 4.873294346978557e-07,
|
| 2121 |
+
"loss": 0.2722,
|
| 2122 |
+
"step": 2260
|
| 2123 |
+
},
|
| 2124 |
+
{
|
| 2125 |
+
"epoch": 59.34640522875817,
|
| 2126 |
+
"grad_norm": 9.155200004577637,
|
| 2127 |
+
"learning_rate": 2.436647173489279e-07,
|
| 2128 |
+
"loss": 0.2591,
|
| 2129 |
+
"step": 2270
|
| 2130 |
+
},
|
| 2131 |
+
{
|
| 2132 |
+
"epoch": 59.6078431372549,
|
| 2133 |
+
"grad_norm": 7.341146469116211,
|
| 2134 |
+
"learning_rate": 0.0,
|
| 2135 |
+
"loss": 0.3047,
|
| 2136 |
+
"step": 2280
|
| 2137 |
+
},
|
| 2138 |
+
{
|
| 2139 |
+
"epoch": 59.6078431372549,
|
| 2140 |
+
"eval_accuracy": 0.7647058823529411,
|
| 2141 |
+
"eval_loss": 0.604828417301178,
|
| 2142 |
+
"eval_runtime": 20.7055,
|
| 2143 |
+
"eval_samples_per_second": 6.568,
|
| 2144 |
+
"eval_steps_per_second": 0.821,
|
| 2145 |
+
"step": 2280
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 59.6078431372549,
|
| 2149 |
+
"step": 2280,
|
| 2150 |
+
"total_flos": 1.9206376588836864e+17,
|
| 2151 |
+
"train_loss": 0.5395637326073228,
|
| 2152 |
+
"train_runtime": 14275.0415,
|
| 2153 |
+
"train_samples_per_second": 5.145,
|
| 2154 |
+
"train_steps_per_second": 0.16
|
| 2155 |
+
}
|
| 2156 |
+
],
|
| 2157 |
+
"logging_steps": 10,
|
| 2158 |
+
"max_steps": 2280,
|
| 2159 |
+
"num_input_tokens_seen": 0,
|
| 2160 |
+
"num_train_epochs": 60,
|
| 2161 |
+
"save_steps": 500,
|
| 2162 |
+
"stateful_callbacks": {
|
| 2163 |
+
"TrainerControl": {
|
| 2164 |
+
"args": {
|
| 2165 |
+
"should_epoch_stop": false,
|
| 2166 |
+
"should_evaluate": false,
|
| 2167 |
+
"should_log": false,
|
| 2168 |
+
"should_save": true,
|
| 2169 |
+
"should_training_stop": true
|
| 2170 |
+
},
|
| 2171 |
+
"attributes": {}
|
| 2172 |
+
}
|
| 2173 |
+
},
|
| 2174 |
+
"total_flos": 1.9206376588836864e+17,
|
| 2175 |
+
"train_batch_size": 8,
|
| 2176 |
+
"trial_name": null,
|
| 2177 |
+
"trial_params": null
|
| 2178 |
+
}
|