End of training
Browse files- README.md +14 -2
- all_results.json +16 -0
- eval_results.json +10 -0
- train_results.json +9 -0
- trainer_state.json +2392 -0
README.md
CHANGED
|
@@ -2,11 +2,23 @@
|
|
| 2 |
library_name: transformers
|
| 3 |
tags:
|
| 4 |
- generated_from_trainer
|
|
|
|
|
|
|
| 5 |
metrics:
|
| 6 |
- accuracy
|
| 7 |
model-index:
|
| 8 |
- name: opt-babylm2-clean-spacy-32k_seed-42_3e-4
|
| 9 |
-
results:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 10 |
---
|
| 11 |
|
| 12 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
@@ -14,7 +26,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 14 |
|
| 15 |
# opt-babylm2-clean-spacy-32k_seed-42_3e-4
|
| 16 |
|
| 17 |
-
This model was trained from scratch on
|
| 18 |
It achieves the following results on the evaluation set:
|
| 19 |
- Loss: 3.0190
|
| 20 |
- Accuracy: 0.4234
|
|
|
|
| 2 |
library_name: transformers
|
| 3 |
tags:
|
| 4 |
- generated_from_trainer
|
| 5 |
+
datasets:
|
| 6 |
+
- kanishka/babylm2-clean-spacy
|
| 7 |
metrics:
|
| 8 |
- accuracy
|
| 9 |
model-index:
|
| 10 |
- name: opt-babylm2-clean-spacy-32k_seed-42_3e-4
|
| 11 |
+
results:
|
| 12 |
+
- task:
|
| 13 |
+
name: Causal Language Modeling
|
| 14 |
+
type: text-generation
|
| 15 |
+
dataset:
|
| 16 |
+
name: kanishka/babylm2-clean-spacy
|
| 17 |
+
type: kanishka/babylm2-clean-spacy
|
| 18 |
+
metrics:
|
| 19 |
+
- name: Accuracy
|
| 20 |
+
type: accuracy
|
| 21 |
+
value: 0.4234014597448438
|
| 22 |
---
|
| 23 |
|
| 24 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
|
|
| 26 |
|
| 27 |
# opt-babylm2-clean-spacy-32k_seed-42_3e-4
|
| 28 |
|
| 29 |
+
This model was trained from scratch on the kanishka/babylm2-clean-spacy dataset.
|
| 30 |
It achieves the following results on the evaluation set:
|
| 31 |
- Loss: 3.0190
|
| 32 |
- Accuracy: 0.4234
|
all_results.json
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 20.0,
|
| 3 |
+
"eval_accuracy": 0.4234014597448438,
|
| 4 |
+
"eval_loss": 3.0190186500549316,
|
| 5 |
+
"eval_runtime": 111.5107,
|
| 6 |
+
"eval_samples": 52440,
|
| 7 |
+
"eval_samples_per_second": 470.269,
|
| 8 |
+
"eval_steps_per_second": 7.354,
|
| 9 |
+
"perplexity": 20.47119242004321,
|
| 10 |
+
"total_flos": 1.29957250203648e+18,
|
| 11 |
+
"train_loss": 2.7038125842232534,
|
| 12 |
+
"train_runtime": 44110.0674,
|
| 13 |
+
"train_samples": 497364,
|
| 14 |
+
"train_samples_per_second": 225.51,
|
| 15 |
+
"train_steps_per_second": 7.047
|
| 16 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 20.0,
|
| 3 |
+
"eval_accuracy": 0.4234014597448438,
|
| 4 |
+
"eval_loss": 3.0190186500549316,
|
| 5 |
+
"eval_runtime": 111.5107,
|
| 6 |
+
"eval_samples": 52440,
|
| 7 |
+
"eval_samples_per_second": 470.269,
|
| 8 |
+
"eval_steps_per_second": 7.354,
|
| 9 |
+
"perplexity": 20.47119242004321
|
| 10 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 20.0,
|
| 3 |
+
"total_flos": 1.29957250203648e+18,
|
| 4 |
+
"train_loss": 2.7038125842232534,
|
| 5 |
+
"train_runtime": 44110.0674,
|
| 6 |
+
"train_samples": 497364,
|
| 7 |
+
"train_samples_per_second": 225.51,
|
| 8 |
+
"train_steps_per_second": 7.047
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2392 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 20.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 310860,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.06433764395547835,
|
| 13 |
+
"grad_norm": 0.9469536542892456,
|
| 14 |
+
"learning_rate": 9.375e-06,
|
| 15 |
+
"loss": 7.0597,
|
| 16 |
+
"step": 1000
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.1286752879109567,
|
| 20 |
+
"grad_norm": 0.9877486824989319,
|
| 21 |
+
"learning_rate": 1.875e-05,
|
| 22 |
+
"loss": 4.8574,
|
| 23 |
+
"step": 2000
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.19301293186643506,
|
| 27 |
+
"grad_norm": 1.1811100244522095,
|
| 28 |
+
"learning_rate": 2.8125e-05,
|
| 29 |
+
"loss": 4.5464,
|
| 30 |
+
"step": 3000
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.2573505758219134,
|
| 34 |
+
"grad_norm": 1.155553936958313,
|
| 35 |
+
"learning_rate": 3.75e-05,
|
| 36 |
+
"loss": 4.3086,
|
| 37 |
+
"step": 4000
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.32168821977739176,
|
| 41 |
+
"grad_norm": 1.0062898397445679,
|
| 42 |
+
"learning_rate": 4.6874999999999994e-05,
|
| 43 |
+
"loss": 4.1307,
|
| 44 |
+
"step": 5000
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.3860258637328701,
|
| 48 |
+
"grad_norm": 0.9749443531036377,
|
| 49 |
+
"learning_rate": 5.625e-05,
|
| 50 |
+
"loss": 3.986,
|
| 51 |
+
"step": 6000
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.45036350768834843,
|
| 55 |
+
"grad_norm": 0.9870838522911072,
|
| 56 |
+
"learning_rate": 6.5625e-05,
|
| 57 |
+
"loss": 3.8708,
|
| 58 |
+
"step": 7000
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.5147011516438268,
|
| 62 |
+
"grad_norm": 1.0740858316421509,
|
| 63 |
+
"learning_rate": 7.5e-05,
|
| 64 |
+
"loss": 3.781,
|
| 65 |
+
"step": 8000
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.5790387955993052,
|
| 69 |
+
"grad_norm": 0.969571053981781,
|
| 70 |
+
"learning_rate": 8.437499999999999e-05,
|
| 71 |
+
"loss": 3.6942,
|
| 72 |
+
"step": 9000
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.6433764395547835,
|
| 76 |
+
"grad_norm": 0.923062801361084,
|
| 77 |
+
"learning_rate": 9.374999999999999e-05,
|
| 78 |
+
"loss": 3.6225,
|
| 79 |
+
"step": 10000
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.7077140835102619,
|
| 83 |
+
"grad_norm": 0.87486732006073,
|
| 84 |
+
"learning_rate": 0.00010312499999999999,
|
| 85 |
+
"loss": 3.5667,
|
| 86 |
+
"step": 11000
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.7720517274657402,
|
| 90 |
+
"grad_norm": 0.8343172073364258,
|
| 91 |
+
"learning_rate": 0.000112490625,
|
| 92 |
+
"loss": 3.5107,
|
| 93 |
+
"step": 12000
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.8363893714212186,
|
| 97 |
+
"grad_norm": 0.8089198470115662,
|
| 98 |
+
"learning_rate": 0.000121865625,
|
| 99 |
+
"loss": 3.4681,
|
| 100 |
+
"step": 13000
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.9007270153766969,
|
| 104 |
+
"grad_norm": 0.8141182661056519,
|
| 105 |
+
"learning_rate": 0.00013123125,
|
| 106 |
+
"loss": 3.4337,
|
| 107 |
+
"step": 14000
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.9650646593321752,
|
| 111 |
+
"grad_norm": 0.7596079707145691,
|
| 112 |
+
"learning_rate": 0.00014060625,
|
| 113 |
+
"loss": 3.3944,
|
| 114 |
+
"step": 15000
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 1.0,
|
| 118 |
+
"eval_accuracy": 0.37339323372369543,
|
| 119 |
+
"eval_loss": 3.4211647510528564,
|
| 120 |
+
"eval_runtime": 112.5567,
|
| 121 |
+
"eval_samples_per_second": 465.898,
|
| 122 |
+
"eval_steps_per_second": 7.285,
|
| 123 |
+
"step": 15543
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 1.0294023032876536,
|
| 127 |
+
"grad_norm": 0.7583508491516113,
|
| 128 |
+
"learning_rate": 0.000149971875,
|
| 129 |
+
"loss": 3.345,
|
| 130 |
+
"step": 16000
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 1.093739947243132,
|
| 134 |
+
"grad_norm": 0.7395954728126526,
|
| 135 |
+
"learning_rate": 0.00015933749999999996,
|
| 136 |
+
"loss": 3.3182,
|
| 137 |
+
"step": 17000
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 1.1580775911986103,
|
| 141 |
+
"grad_norm": 0.7119142413139343,
|
| 142 |
+
"learning_rate": 0.00016871249999999996,
|
| 143 |
+
"loss": 3.304,
|
| 144 |
+
"step": 18000
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 1.2224152351540887,
|
| 148 |
+
"grad_norm": 0.7133814692497253,
|
| 149 |
+
"learning_rate": 0.00017808749999999999,
|
| 150 |
+
"loss": 3.2808,
|
| 151 |
+
"step": 19000
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 1.286752879109567,
|
| 155 |
+
"grad_norm": 0.6662284731864929,
|
| 156 |
+
"learning_rate": 0.00018745312499999998,
|
| 157 |
+
"loss": 3.2624,
|
| 158 |
+
"step": 20000
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 1.3510905230650454,
|
| 162 |
+
"grad_norm": 0.6821054816246033,
|
| 163 |
+
"learning_rate": 0.00019682812499999998,
|
| 164 |
+
"loss": 3.2468,
|
| 165 |
+
"step": 21000
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 1.4154281670205238,
|
| 169 |
+
"grad_norm": 0.6423399448394775,
|
| 170 |
+
"learning_rate": 0.00020619374999999998,
|
| 171 |
+
"loss": 3.2323,
|
| 172 |
+
"step": 22000
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 1.4797658109760021,
|
| 176 |
+
"grad_norm": 0.6489351987838745,
|
| 177 |
+
"learning_rate": 0.00021556874999999998,
|
| 178 |
+
"loss": 3.218,
|
| 179 |
+
"step": 23000
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 1.5441034549314803,
|
| 183 |
+
"grad_norm": 0.6388360261917114,
|
| 184 |
+
"learning_rate": 0.00022493437499999998,
|
| 185 |
+
"loss": 3.2063,
|
| 186 |
+
"step": 24000
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 1.6084410988869586,
|
| 190 |
+
"grad_norm": 0.6035541296005249,
|
| 191 |
+
"learning_rate": 0.00023430937499999997,
|
| 192 |
+
"loss": 3.1971,
|
| 193 |
+
"step": 25000
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 1.672778742842437,
|
| 197 |
+
"grad_norm": 0.5949345231056213,
|
| 198 |
+
"learning_rate": 0.00024367499999999997,
|
| 199 |
+
"loss": 3.1683,
|
| 200 |
+
"step": 26000
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 1.7371163867979154,
|
| 204 |
+
"grad_norm": 0.5953760147094727,
|
| 205 |
+
"learning_rate": 0.00025305,
|
| 206 |
+
"loss": 3.1728,
|
| 207 |
+
"step": 27000
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 1.8014540307533937,
|
| 211 |
+
"grad_norm": 0.5276063680648804,
|
| 212 |
+
"learning_rate": 0.000262415625,
|
| 213 |
+
"loss": 3.1607,
|
| 214 |
+
"step": 28000
|
| 215 |
+
},
|
| 216 |
+
{
|
| 217 |
+
"epoch": 1.865791674708872,
|
| 218 |
+
"grad_norm": 0.5257272124290466,
|
| 219 |
+
"learning_rate": 0.000271790625,
|
| 220 |
+
"loss": 3.1472,
|
| 221 |
+
"step": 29000
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 1.9301293186643504,
|
| 225 |
+
"grad_norm": 0.49043259024620056,
|
| 226 |
+
"learning_rate": 0.000281165625,
|
| 227 |
+
"loss": 3.1367,
|
| 228 |
+
"step": 30000
|
| 229 |
+
},
|
| 230 |
+
{
|
| 231 |
+
"epoch": 1.9944669626198288,
|
| 232 |
+
"grad_norm": 0.5030378699302673,
|
| 233 |
+
"learning_rate": 0.000290521875,
|
| 234 |
+
"loss": 3.1245,
|
| 235 |
+
"step": 31000
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 2.0,
|
| 239 |
+
"eval_accuracy": 0.3939620256950988,
|
| 240 |
+
"eval_loss": 3.2037432193756104,
|
| 241 |
+
"eval_runtime": 113.0293,
|
| 242 |
+
"eval_samples_per_second": 463.951,
|
| 243 |
+
"eval_steps_per_second": 7.255,
|
| 244 |
+
"step": 31086
|
| 245 |
+
},
|
| 246 |
+
{
|
| 247 |
+
"epoch": 2.058804606575307,
|
| 248 |
+
"grad_norm": 0.5003546476364136,
|
| 249 |
+
"learning_rate": 0.000299896875,
|
| 250 |
+
"loss": 3.0828,
|
| 251 |
+
"step": 32000
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 2.1231422505307855,
|
| 255 |
+
"grad_norm": 0.48286330699920654,
|
| 256 |
+
"learning_rate": 0.00029893602524564295,
|
| 257 |
+
"loss": 3.08,
|
| 258 |
+
"step": 33000
|
| 259 |
+
},
|
| 260 |
+
{
|
| 261 |
+
"epoch": 2.187479894486264,
|
| 262 |
+
"grad_norm": 0.4852472245693207,
|
| 263 |
+
"learning_rate": 0.0002978602165961414,
|
| 264 |
+
"loss": 3.0633,
|
| 265 |
+
"step": 34000
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 2.2518175384417423,
|
| 269 |
+
"grad_norm": 0.4629572927951813,
|
| 270 |
+
"learning_rate": 0.00029678548375528934,
|
| 271 |
+
"loss": 3.063,
|
| 272 |
+
"step": 35000
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 2.3161551823972206,
|
| 276 |
+
"grad_norm": 0.4571368992328644,
|
| 277 |
+
"learning_rate": 0.0002957096751057878,
|
| 278 |
+
"loss": 3.0453,
|
| 279 |
+
"step": 36000
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 2.380492826352699,
|
| 283 |
+
"grad_norm": 0.44331055879592896,
|
| 284 |
+
"learning_rate": 0.0002946349422649358,
|
| 285 |
+
"loss": 3.0408,
|
| 286 |
+
"step": 37000
|
| 287 |
+
},
|
| 288 |
+
{
|
| 289 |
+
"epoch": 2.4448304703081774,
|
| 290 |
+
"grad_norm": 0.4230923354625702,
|
| 291 |
+
"learning_rate": 0.00029355913361543424,
|
| 292 |
+
"loss": 3.0359,
|
| 293 |
+
"step": 38000
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 2.5091681142636557,
|
| 297 |
+
"grad_norm": 0.4260108768939972,
|
| 298 |
+
"learning_rate": 0.0002924833249659327,
|
| 299 |
+
"loss": 3.0316,
|
| 300 |
+
"step": 39000
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"epoch": 2.573505758219134,
|
| 304 |
+
"grad_norm": 0.41887935996055603,
|
| 305 |
+
"learning_rate": 0.0002914085921250807,
|
| 306 |
+
"loss": 3.0299,
|
| 307 |
+
"step": 40000
|
| 308 |
+
},
|
| 309 |
+
{
|
| 310 |
+
"epoch": 2.6378434021746124,
|
| 311 |
+
"grad_norm": 0.41068920493125916,
|
| 312 |
+
"learning_rate": 0.00029033278347557914,
|
| 313 |
+
"loss": 3.0138,
|
| 314 |
+
"step": 41000
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 2.702181046130091,
|
| 318 |
+
"grad_norm": 0.39430394768714905,
|
| 319 |
+
"learning_rate": 0.0002892591264433766,
|
| 320 |
+
"loss": 3.0038,
|
| 321 |
+
"step": 42000
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 2.766518690085569,
|
| 325 |
+
"grad_norm": 0.4100017547607422,
|
| 326 |
+
"learning_rate": 0.00028818331779387505,
|
| 327 |
+
"loss": 3.0088,
|
| 328 |
+
"step": 43000
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 2.8308563340410475,
|
| 332 |
+
"grad_norm": 0.4101816415786743,
|
| 333 |
+
"learning_rate": 0.0002871075091443735,
|
| 334 |
+
"loss": 2.9937,
|
| 335 |
+
"step": 44000
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 2.895193977996526,
|
| 339 |
+
"grad_norm": 0.38294607400894165,
|
| 340 |
+
"learning_rate": 0.000286031700494872,
|
| 341 |
+
"loss": 2.9898,
|
| 342 |
+
"step": 45000
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"epoch": 2.9595316219520043,
|
| 346 |
+
"grad_norm": 0.37260037660598755,
|
| 347 |
+
"learning_rate": 0.00028495589184537043,
|
| 348 |
+
"loss": 2.9807,
|
| 349 |
+
"step": 46000
|
| 350 |
+
},
|
| 351 |
+
{
|
| 352 |
+
"epoch": 3.0,
|
| 353 |
+
"eval_accuracy": 0.40728960081362825,
|
| 354 |
+
"eval_loss": 3.079435110092163,
|
| 355 |
+
"eval_runtime": 112.8117,
|
| 356 |
+
"eval_samples_per_second": 464.845,
|
| 357 |
+
"eval_steps_per_second": 7.269,
|
| 358 |
+
"step": 46629
|
| 359 |
+
},
|
| 360 |
+
{
|
| 361 |
+
"epoch": 3.0238692659074826,
|
| 362 |
+
"grad_norm": 0.40783312916755676,
|
| 363 |
+
"learning_rate": 0.00028388115900451836,
|
| 364 |
+
"loss": 2.9534,
|
| 365 |
+
"step": 47000
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 3.088206909862961,
|
| 369 |
+
"grad_norm": 0.38361623883247375,
|
| 370 |
+
"learning_rate": 0.0002828053503550168,
|
| 371 |
+
"loss": 2.9122,
|
| 372 |
+
"step": 48000
|
| 373 |
+
},
|
| 374 |
+
{
|
| 375 |
+
"epoch": 3.1525445538184393,
|
| 376 |
+
"grad_norm": 0.3766690790653229,
|
| 377 |
+
"learning_rate": 0.0002817306175141648,
|
| 378 |
+
"loss": 2.9119,
|
| 379 |
+
"step": 49000
|
| 380 |
+
},
|
| 381 |
+
{
|
| 382 |
+
"epoch": 3.2168821977739177,
|
| 383 |
+
"grad_norm": 0.38536399602890015,
|
| 384 |
+
"learning_rate": 0.00028065480886466326,
|
| 385 |
+
"loss": 2.9135,
|
| 386 |
+
"step": 50000
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 3.2812198417293956,
|
| 390 |
+
"grad_norm": 0.38374361395835876,
|
| 391 |
+
"learning_rate": 0.0002795800760238112,
|
| 392 |
+
"loss": 2.9094,
|
| 393 |
+
"step": 51000
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 3.345557485684874,
|
| 397 |
+
"grad_norm": 0.3872029185295105,
|
| 398 |
+
"learning_rate": 0.00027850426737430965,
|
| 399 |
+
"loss": 2.9069,
|
| 400 |
+
"step": 52000
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 3.4098951296403524,
|
| 404 |
+
"grad_norm": 0.37401601672172546,
|
| 405 |
+
"learning_rate": 0.0002774284587248081,
|
| 406 |
+
"loss": 2.9092,
|
| 407 |
+
"step": 53000
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 3.4742327735958307,
|
| 411 |
+
"grad_norm": 0.34951257705688477,
|
| 412 |
+
"learning_rate": 0.0002763537258839561,
|
| 413 |
+
"loss": 2.9124,
|
| 414 |
+
"step": 54000
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 3.538570417551309,
|
| 418 |
+
"grad_norm": 0.36252182722091675,
|
| 419 |
+
"learning_rate": 0.000275278993043104,
|
| 420 |
+
"loss": 2.9086,
|
| 421 |
+
"step": 55000
|
| 422 |
+
},
|
| 423 |
+
{
|
| 424 |
+
"epoch": 3.6029080615067874,
|
| 425 |
+
"grad_norm": 0.3610841631889343,
|
| 426 |
+
"learning_rate": 0.0002742031843936025,
|
| 427 |
+
"loss": 2.9044,
|
| 428 |
+
"step": 56000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 3.667245705462266,
|
| 432 |
+
"grad_norm": 0.356315016746521,
|
| 433 |
+
"learning_rate": 0.00027312737574410094,
|
| 434 |
+
"loss": 2.8984,
|
| 435 |
+
"step": 57000
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 3.731583349417744,
|
| 439 |
+
"grad_norm": 0.3501368761062622,
|
| 440 |
+
"learning_rate": 0.0002720515670945994,
|
| 441 |
+
"loss": 2.8991,
|
| 442 |
+
"step": 58000
|
| 443 |
+
},
|
| 444 |
+
{
|
| 445 |
+
"epoch": 3.7959209933732225,
|
| 446 |
+
"grad_norm": 0.3654986619949341,
|
| 447 |
+
"learning_rate": 0.00027097575844509786,
|
| 448 |
+
"loss": 2.8946,
|
| 449 |
+
"step": 59000
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 3.860258637328701,
|
| 453 |
+
"grad_norm": 0.34233444929122925,
|
| 454 |
+
"learning_rate": 0.00026990102560424585,
|
| 455 |
+
"loss": 2.8981,
|
| 456 |
+
"step": 60000
|
| 457 |
+
},
|
| 458 |
+
{
|
| 459 |
+
"epoch": 3.9245962812841793,
|
| 460 |
+
"grad_norm": 0.35118335485458374,
|
| 461 |
+
"learning_rate": 0.0002688252169547443,
|
| 462 |
+
"loss": 2.8932,
|
| 463 |
+
"step": 61000
|
| 464 |
+
},
|
| 465 |
+
{
|
| 466 |
+
"epoch": 3.9889339252396576,
|
| 467 |
+
"grad_norm": 0.3542274236679077,
|
| 468 |
+
"learning_rate": 0.00026775048411389223,
|
| 469 |
+
"loss": 2.8872,
|
| 470 |
+
"step": 62000
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 4.0,
|
| 474 |
+
"eval_accuracy": 0.4139775803532702,
|
| 475 |
+
"eval_loss": 3.0204551219940186,
|
| 476 |
+
"eval_runtime": 113.294,
|
| 477 |
+
"eval_samples_per_second": 462.866,
|
| 478 |
+
"eval_steps_per_second": 7.238,
|
| 479 |
+
"step": 62172
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 4.053271569195136,
|
| 483 |
+
"grad_norm": 0.359250545501709,
|
| 484 |
+
"learning_rate": 0.0002666746754643907,
|
| 485 |
+
"loss": 2.8242,
|
| 486 |
+
"step": 63000
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 4.117609213150614,
|
| 490 |
+
"grad_norm": 0.34917619824409485,
|
| 491 |
+
"learning_rate": 0.00026559886681488915,
|
| 492 |
+
"loss": 2.8144,
|
| 493 |
+
"step": 64000
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 4.181946857106093,
|
| 497 |
+
"grad_norm": 0.351457417011261,
|
| 498 |
+
"learning_rate": 0.00026452413397403714,
|
| 499 |
+
"loss": 2.8132,
|
| 500 |
+
"step": 65000
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 4.246284501061571,
|
| 504 |
+
"grad_norm": 0.35231146216392517,
|
| 505 |
+
"learning_rate": 0.0002634483253245356,
|
| 506 |
+
"loss": 2.8203,
|
| 507 |
+
"step": 66000
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 4.310622145017049,
|
| 511 |
+
"grad_norm": 0.354030579328537,
|
| 512 |
+
"learning_rate": 0.0002623735924836836,
|
| 513 |
+
"loss": 2.8273,
|
| 514 |
+
"step": 67000
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 4.374959788972528,
|
| 518 |
+
"grad_norm": 0.3434860408306122,
|
| 519 |
+
"learning_rate": 0.00026129778383418204,
|
| 520 |
+
"loss": 2.8221,
|
| 521 |
+
"step": 68000
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 4.439297432928006,
|
| 525 |
+
"grad_norm": 0.35598379373550415,
|
| 526 |
+
"learning_rate": 0.0002602219751846805,
|
| 527 |
+
"loss": 2.8283,
|
| 528 |
+
"step": 69000
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 4.5036350768834845,
|
| 532 |
+
"grad_norm": 0.350340873003006,
|
| 533 |
+
"learning_rate": 0.00025914616653517896,
|
| 534 |
+
"loss": 2.8242,
|
| 535 |
+
"step": 70000
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 4.567972720838963,
|
| 539 |
+
"grad_norm": 0.34078752994537354,
|
| 540 |
+
"learning_rate": 0.0002580714336943269,
|
| 541 |
+
"loss": 2.8309,
|
| 542 |
+
"step": 71000
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 4.632310364794441,
|
| 546 |
+
"grad_norm": 0.3571733832359314,
|
| 547 |
+
"learning_rate": 0.00025699670085347487,
|
| 548 |
+
"loss": 2.8248,
|
| 549 |
+
"step": 72000
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 4.69664800874992,
|
| 553 |
+
"grad_norm": 0.35940021276474,
|
| 554 |
+
"learning_rate": 0.00025592089220397333,
|
| 555 |
+
"loss": 2.8334,
|
| 556 |
+
"step": 73000
|
| 557 |
+
},
|
| 558 |
+
{
|
| 559 |
+
"epoch": 4.760985652705398,
|
| 560 |
+
"grad_norm": 0.3354775607585907,
|
| 561 |
+
"learning_rate": 0.0002548450835544718,
|
| 562 |
+
"loss": 2.8263,
|
| 563 |
+
"step": 74000
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 4.825323296660876,
|
| 567 |
+
"grad_norm": 0.330805242061615,
|
| 568 |
+
"learning_rate": 0.0002537703507136197,
|
| 569 |
+
"loss": 2.8296,
|
| 570 |
+
"step": 75000
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 4.889660940616355,
|
| 574 |
+
"grad_norm": 0.32566189765930176,
|
| 575 |
+
"learning_rate": 0.0002526945420641182,
|
| 576 |
+
"loss": 2.8208,
|
| 577 |
+
"step": 76000
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 4.953998584571833,
|
| 581 |
+
"grad_norm": 0.32299116253852844,
|
| 582 |
+
"learning_rate": 0.00025161980922326616,
|
| 583 |
+
"loss": 2.8286,
|
| 584 |
+
"step": 77000
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 5.0,
|
| 588 |
+
"eval_accuracy": 0.417990981289541,
|
| 589 |
+
"eval_loss": 2.988518238067627,
|
| 590 |
+
"eval_runtime": 113.0062,
|
| 591 |
+
"eval_samples_per_second": 464.045,
|
| 592 |
+
"eval_steps_per_second": 7.256,
|
| 593 |
+
"step": 77715
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 5.018336228527311,
|
| 597 |
+
"grad_norm": 0.332711786031723,
|
| 598 |
+
"learning_rate": 0.00025054400057376457,
|
| 599 |
+
"loss": 2.797,
|
| 600 |
+
"step": 78000
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"epoch": 5.08267387248279,
|
| 604 |
+
"grad_norm": 0.3597155809402466,
|
| 605 |
+
"learning_rate": 0.000249468191924263,
|
| 606 |
+
"loss": 2.7461,
|
| 607 |
+
"step": 79000
|
| 608 |
+
},
|
| 609 |
+
{
|
| 610 |
+
"epoch": 5.147011516438268,
|
| 611 |
+
"grad_norm": 0.3411096930503845,
|
| 612 |
+
"learning_rate": 0.000248393459083411,
|
| 613 |
+
"loss": 2.7493,
|
| 614 |
+
"step": 80000
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 5.2113491603937465,
|
| 618 |
+
"grad_norm": 0.35248109698295593,
|
| 619 |
+
"learning_rate": 0.00024731765043390947,
|
| 620 |
+
"loss": 2.7584,
|
| 621 |
+
"step": 81000
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 5.275686804349225,
|
| 625 |
+
"grad_norm": 0.3520190417766571,
|
| 626 |
+
"learning_rate": 0.00024624184178440793,
|
| 627 |
+
"loss": 2.755,
|
| 628 |
+
"step": 82000
|
| 629 |
+
},
|
| 630 |
+
{
|
| 631 |
+
"epoch": 5.340024448304703,
|
| 632 |
+
"grad_norm": 0.34867680072784424,
|
| 633 |
+
"learning_rate": 0.00024516710894355586,
|
| 634 |
+
"loss": 2.7649,
|
| 635 |
+
"step": 83000
|
| 636 |
+
},
|
| 637 |
+
{
|
| 638 |
+
"epoch": 5.404362092260182,
|
| 639 |
+
"grad_norm": 0.3400154709815979,
|
| 640 |
+
"learning_rate": 0.00024409130029405434,
|
| 641 |
+
"loss": 2.7586,
|
| 642 |
+
"step": 84000
|
| 643 |
+
},
|
| 644 |
+
{
|
| 645 |
+
"epoch": 5.46869973621566,
|
| 646 |
+
"grad_norm": 0.3640024662017822,
|
| 647 |
+
"learning_rate": 0.0002430154916445528,
|
| 648 |
+
"loss": 2.7606,
|
| 649 |
+
"step": 85000
|
| 650 |
+
},
|
| 651 |
+
{
|
| 652 |
+
"epoch": 5.533037380171138,
|
| 653 |
+
"grad_norm": 0.3456322252750397,
|
| 654 |
+
"learning_rate": 0.00024193968299505126,
|
| 655 |
+
"loss": 2.767,
|
| 656 |
+
"step": 86000
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 5.597375024126617,
|
| 660 |
+
"grad_norm": 0.3284786343574524,
|
| 661 |
+
"learning_rate": 0.00024086495015419922,
|
| 662 |
+
"loss": 2.7687,
|
| 663 |
+
"step": 87000
|
| 664 |
+
},
|
| 665 |
+
{
|
| 666 |
+
"epoch": 5.661712668082095,
|
| 667 |
+
"grad_norm": 0.3351786732673645,
|
| 668 |
+
"learning_rate": 0.00023978914150469768,
|
| 669 |
+
"loss": 2.7705,
|
| 670 |
+
"step": 88000
|
| 671 |
+
},
|
| 672 |
+
{
|
| 673 |
+
"epoch": 5.726050312037573,
|
| 674 |
+
"grad_norm": 0.3189627528190613,
|
| 675 |
+
"learning_rate": 0.00023871440866384563,
|
| 676 |
+
"loss": 2.7743,
|
| 677 |
+
"step": 89000
|
| 678 |
+
},
|
| 679 |
+
{
|
| 680 |
+
"epoch": 5.790387955993052,
|
| 681 |
+
"grad_norm": 0.3447468876838684,
|
| 682 |
+
"learning_rate": 0.0002376386000143441,
|
| 683 |
+
"loss": 2.7712,
|
| 684 |
+
"step": 90000
|
| 685 |
+
},
|
| 686 |
+
{
|
| 687 |
+
"epoch": 5.85472559994853,
|
| 688 |
+
"grad_norm": 0.3212040364742279,
|
| 689 |
+
"learning_rate": 0.00023656386717349205,
|
| 690 |
+
"loss": 2.7741,
|
| 691 |
+
"step": 91000
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 5.9190632439040085,
|
| 695 |
+
"grad_norm": 0.3384701609611511,
|
| 696 |
+
"learning_rate": 0.0002354880585239905,
|
| 697 |
+
"loss": 2.7747,
|
| 698 |
+
"step": 92000
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 5.983400887859487,
|
| 702 |
+
"grad_norm": 0.33266380429267883,
|
| 703 |
+
"learning_rate": 0.00023441224987448894,
|
| 704 |
+
"loss": 2.779,
|
| 705 |
+
"step": 93000
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 6.0,
|
| 709 |
+
"eval_accuracy": 0.4206485843765424,
|
| 710 |
+
"eval_loss": 2.969926595687866,
|
| 711 |
+
"eval_runtime": 113.3066,
|
| 712 |
+
"eval_samples_per_second": 462.815,
|
| 713 |
+
"eval_steps_per_second": 7.237,
|
| 714 |
+
"step": 93258
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 6.047738531814965,
|
| 718 |
+
"grad_norm": 0.35645657777786255,
|
| 719 |
+
"learning_rate": 0.0002333364412249874,
|
| 720 |
+
"loss": 2.7059,
|
| 721 |
+
"step": 94000
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 6.112076175770444,
|
| 725 |
+
"grad_norm": 0.35733386874198914,
|
| 726 |
+
"learning_rate": 0.0002322617083841354,
|
| 727 |
+
"loss": 2.6966,
|
| 728 |
+
"step": 95000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 6.176413819725922,
|
| 732 |
+
"grad_norm": 0.3435540199279785,
|
| 733 |
+
"learning_rate": 0.00023118589973463387,
|
| 734 |
+
"loss": 2.6986,
|
| 735 |
+
"step": 96000
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 6.2407514636814,
|
| 739 |
+
"grad_norm": 0.3479596972465515,
|
| 740 |
+
"learning_rate": 0.0002301100910851323,
|
| 741 |
+
"loss": 2.7068,
|
| 742 |
+
"step": 97000
|
| 743 |
+
},
|
| 744 |
+
{
|
| 745 |
+
"epoch": 6.305089107636879,
|
| 746 |
+
"grad_norm": 0.3150424659252167,
|
| 747 |
+
"learning_rate": 0.00022903428243563077,
|
| 748 |
+
"loss": 2.7074,
|
| 749 |
+
"step": 98000
|
| 750 |
+
},
|
| 751 |
+
{
|
| 752 |
+
"epoch": 6.369426751592357,
|
| 753 |
+
"grad_norm": 0.34055858850479126,
|
| 754 |
+
"learning_rate": 0.00022795954959477872,
|
| 755 |
+
"loss": 2.7065,
|
| 756 |
+
"step": 99000
|
| 757 |
+
},
|
| 758 |
+
{
|
| 759 |
+
"epoch": 6.433764395547835,
|
| 760 |
+
"grad_norm": 0.3491341769695282,
|
| 761 |
+
"learning_rate": 0.0002268848167539267,
|
| 762 |
+
"loss": 2.7156,
|
| 763 |
+
"step": 100000
|
| 764 |
+
},
|
| 765 |
+
{
|
| 766 |
+
"epoch": 6.498102039503314,
|
| 767 |
+
"grad_norm": 0.3347100019454956,
|
| 768 |
+
"learning_rate": 0.00022580900810442514,
|
| 769 |
+
"loss": 2.714,
|
| 770 |
+
"step": 101000
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 6.562439683458791,
|
| 774 |
+
"grad_norm": 0.35210439562797546,
|
| 775 |
+
"learning_rate": 0.00022473427526357312,
|
| 776 |
+
"loss": 2.7194,
|
| 777 |
+
"step": 102000
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 6.6267773274142705,
|
| 781 |
+
"grad_norm": 0.3326897919178009,
|
| 782 |
+
"learning_rate": 0.00022365846661407155,
|
| 783 |
+
"loss": 2.727,
|
| 784 |
+
"step": 103000
|
| 785 |
+
},
|
| 786 |
+
{
|
| 787 |
+
"epoch": 6.691114971369748,
|
| 788 |
+
"grad_norm": 0.3269229531288147,
|
| 789 |
+
"learning_rate": 0.00022258265796457,
|
| 790 |
+
"loss": 2.7203,
|
| 791 |
+
"step": 104000
|
| 792 |
+
},
|
| 793 |
+
{
|
| 794 |
+
"epoch": 6.755452615325227,
|
| 795 |
+
"grad_norm": 0.34183254837989807,
|
| 796 |
+
"learning_rate": 0.00022150684931506847,
|
| 797 |
+
"loss": 2.7328,
|
| 798 |
+
"step": 105000
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 6.819790259280705,
|
| 802 |
+
"grad_norm": 0.33449244499206543,
|
| 803 |
+
"learning_rate": 0.00022043211647421643,
|
| 804 |
+
"loss": 2.7291,
|
| 805 |
+
"step": 106000
|
| 806 |
+
},
|
| 807 |
+
{
|
| 808 |
+
"epoch": 6.884127903236184,
|
| 809 |
+
"grad_norm": 0.33734798431396484,
|
| 810 |
+
"learning_rate": 0.0002193563078247149,
|
| 811 |
+
"loss": 2.7277,
|
| 812 |
+
"step": 107000
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 6.948465547191661,
|
| 816 |
+
"grad_norm": 0.34088870882987976,
|
| 817 |
+
"learning_rate": 0.00021828157498386284,
|
| 818 |
+
"loss": 2.7316,
|
| 819 |
+
"step": 108000
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 7.0,
|
| 823 |
+
"eval_accuracy": 0.42224443247932275,
|
| 824 |
+
"eval_loss": 2.958820104598999,
|
| 825 |
+
"eval_runtime": 113.0743,
|
| 826 |
+
"eval_samples_per_second": 463.766,
|
| 827 |
+
"eval_steps_per_second": 7.252,
|
| 828 |
+
"step": 108801
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 7.01280319114714,
|
| 832 |
+
"grad_norm": 0.3517482876777649,
|
| 833 |
+
"learning_rate": 0.0002172057663343613,
|
| 834 |
+
"loss": 2.7116,
|
| 835 |
+
"step": 109000
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 7.077140835102618,
|
| 839 |
+
"grad_norm": 0.3411085903644562,
|
| 840 |
+
"learning_rate": 0.00021613103349350926,
|
| 841 |
+
"loss": 2.6469,
|
| 842 |
+
"step": 110000
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 7.1414784790580965,
|
| 846 |
+
"grad_norm": 0.3486618399620056,
|
| 847 |
+
"learning_rate": 0.00021505522484400772,
|
| 848 |
+
"loss": 2.6546,
|
| 849 |
+
"step": 111000
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 7.205816123013575,
|
| 853 |
+
"grad_norm": 0.35618531703948975,
|
| 854 |
+
"learning_rate": 0.00021397941619450618,
|
| 855 |
+
"loss": 2.6603,
|
| 856 |
+
"step": 112000
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 7.270153766969053,
|
| 860 |
+
"grad_norm": 0.34740447998046875,
|
| 861 |
+
"learning_rate": 0.00021290468335365413,
|
| 862 |
+
"loss": 2.6632,
|
| 863 |
+
"step": 113000
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 7.334491410924532,
|
| 867 |
+
"grad_norm": 0.339108407497406,
|
| 868 |
+
"learning_rate": 0.0002118288747041526,
|
| 869 |
+
"loss": 2.6682,
|
| 870 |
+
"step": 114000
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 7.39882905488001,
|
| 874 |
+
"grad_norm": 0.36686399579048157,
|
| 875 |
+
"learning_rate": 0.00021075306605465105,
|
| 876 |
+
"loss": 2.6718,
|
| 877 |
+
"step": 115000
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 7.463166698835488,
|
| 881 |
+
"grad_norm": 0.3336213529109955,
|
| 882 |
+
"learning_rate": 0.000209678333213799,
|
| 883 |
+
"loss": 2.6806,
|
| 884 |
+
"step": 116000
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 7.527504342790967,
|
| 888 |
+
"grad_norm": 0.34256553649902344,
|
| 889 |
+
"learning_rate": 0.00020860252456429747,
|
| 890 |
+
"loss": 2.6772,
|
| 891 |
+
"step": 117000
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 7.591841986746445,
|
| 895 |
+
"grad_norm": 0.3527204096317291,
|
| 896 |
+
"learning_rate": 0.00020752671591479593,
|
| 897 |
+
"loss": 2.6786,
|
| 898 |
+
"step": 118000
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 7.656179630701923,
|
| 902 |
+
"grad_norm": 0.34285178780555725,
|
| 903 |
+
"learning_rate": 0.0002064509072652944,
|
| 904 |
+
"loss": 2.6816,
|
| 905 |
+
"step": 119000
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 7.720517274657402,
|
| 909 |
+
"grad_norm": 0.3418208658695221,
|
| 910 |
+
"learning_rate": 0.00020537617442444234,
|
| 911 |
+
"loss": 2.6893,
|
| 912 |
+
"step": 120000
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 7.78485491861288,
|
| 916 |
+
"grad_norm": 0.34486138820648193,
|
| 917 |
+
"learning_rate": 0.0002043003657749408,
|
| 918 |
+
"loss": 2.6847,
|
| 919 |
+
"step": 121000
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 7.8491925625683585,
|
| 923 |
+
"grad_norm": 0.348530650138855,
|
| 924 |
+
"learning_rate": 0.00020322563293408876,
|
| 925 |
+
"loss": 2.6826,
|
| 926 |
+
"step": 122000
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 7.913530206523837,
|
| 930 |
+
"grad_norm": 0.33808425068855286,
|
| 931 |
+
"learning_rate": 0.00020215090009323674,
|
| 932 |
+
"loss": 2.6905,
|
| 933 |
+
"step": 123000
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 7.977867850479315,
|
| 937 |
+
"grad_norm": 0.3486366868019104,
|
| 938 |
+
"learning_rate": 0.0002010750914437352,
|
| 939 |
+
"loss": 2.6909,
|
| 940 |
+
"step": 124000
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 8.0,
|
| 944 |
+
"eval_accuracy": 0.4232837528604119,
|
| 945 |
+
"eval_loss": 2.9554243087768555,
|
| 946 |
+
"eval_runtime": 111.9904,
|
| 947 |
+
"eval_samples_per_second": 468.254,
|
| 948 |
+
"eval_steps_per_second": 7.322,
|
| 949 |
+
"step": 124344
|
| 950 |
+
},
|
| 951 |
+
{
|
| 952 |
+
"epoch": 8.042205494434794,
|
| 953 |
+
"grad_norm": 0.35380104184150696,
|
| 954 |
+
"learning_rate": 0.00020000035860288316,
|
| 955 |
+
"loss": 2.6303,
|
| 956 |
+
"step": 125000
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 8.106543138390272,
|
| 960 |
+
"grad_norm": 0.3654320240020752,
|
| 961 |
+
"learning_rate": 0.00019892454995338162,
|
| 962 |
+
"loss": 2.6128,
|
| 963 |
+
"step": 126000
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 8.170880782345751,
|
| 967 |
+
"grad_norm": 0.3670574724674225,
|
| 968 |
+
"learning_rate": 0.00019784874130388008,
|
| 969 |
+
"loss": 2.617,
|
| 970 |
+
"step": 127000
|
| 971 |
+
},
|
| 972 |
+
{
|
| 973 |
+
"epoch": 8.235218426301229,
|
| 974 |
+
"grad_norm": 0.38059455156326294,
|
| 975 |
+
"learning_rate": 0.00019677400846302803,
|
| 976 |
+
"loss": 2.6274,
|
| 977 |
+
"step": 128000
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"epoch": 8.299556070256708,
|
| 981 |
+
"grad_norm": 0.3698261082172394,
|
| 982 |
+
"learning_rate": 0.00019569927562217599,
|
| 983 |
+
"loss": 2.6309,
|
| 984 |
+
"step": 129000
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 8.363893714212185,
|
| 988 |
+
"grad_norm": 0.3583601117134094,
|
| 989 |
+
"learning_rate": 0.00019462346697267445,
|
| 990 |
+
"loss": 2.6312,
|
| 991 |
+
"step": 130000
|
| 992 |
+
},
|
| 993 |
+
{
|
| 994 |
+
"epoch": 8.428231358167665,
|
| 995 |
+
"grad_norm": 0.3602234721183777,
|
| 996 |
+
"learning_rate": 0.0001935476583231729,
|
| 997 |
+
"loss": 2.6368,
|
| 998 |
+
"step": 131000
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 8.492569002123142,
|
| 1002 |
+
"grad_norm": 0.3441711664199829,
|
| 1003 |
+
"learning_rate": 0.00019247184967367137,
|
| 1004 |
+
"loss": 2.6372,
|
| 1005 |
+
"step": 132000
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 8.556906646078621,
|
| 1009 |
+
"grad_norm": 0.3533187508583069,
|
| 1010 |
+
"learning_rate": 0.00019139604102416983,
|
| 1011 |
+
"loss": 2.6443,
|
| 1012 |
+
"step": 133000
|
| 1013 |
+
},
|
| 1014 |
+
{
|
| 1015 |
+
"epoch": 8.621244290034099,
|
| 1016 |
+
"grad_norm": 0.3579193651676178,
|
| 1017 |
+
"learning_rate": 0.00019032130818331778,
|
| 1018 |
+
"loss": 2.6481,
|
| 1019 |
+
"step": 134000
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"epoch": 8.685581933989578,
|
| 1023 |
+
"grad_norm": 0.3524502217769623,
|
| 1024 |
+
"learning_rate": 0.00018924549953381624,
|
| 1025 |
+
"loss": 2.6509,
|
| 1026 |
+
"step": 135000
|
| 1027 |
+
},
|
| 1028 |
+
{
|
| 1029 |
+
"epoch": 8.749919577945056,
|
| 1030 |
+
"grad_norm": 0.36159747838974,
|
| 1031 |
+
"learning_rate": 0.0001881707666929642,
|
| 1032 |
+
"loss": 2.6456,
|
| 1033 |
+
"step": 136000
|
| 1034 |
+
},
|
| 1035 |
+
{
|
| 1036 |
+
"epoch": 8.814257221900533,
|
| 1037 |
+
"grad_norm": 0.34249147772789,
|
| 1038 |
+
"learning_rate": 0.00018709495804346266,
|
| 1039 |
+
"loss": 2.6538,
|
| 1040 |
+
"step": 137000
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 8.878594865856012,
|
| 1044 |
+
"grad_norm": 0.34867429733276367,
|
| 1045 |
+
"learning_rate": 0.0001860202252026106,
|
| 1046 |
+
"loss": 2.6558,
|
| 1047 |
+
"step": 138000
|
| 1048 |
+
},
|
| 1049 |
+
{
|
| 1050 |
+
"epoch": 8.942932509811492,
|
| 1051 |
+
"grad_norm": 0.3351230025291443,
|
| 1052 |
+
"learning_rate": 0.00018494441655310907,
|
| 1053 |
+
"loss": 2.6504,
|
| 1054 |
+
"step": 139000
|
| 1055 |
+
},
|
| 1056 |
+
{
|
| 1057 |
+
"epoch": 9.0,
|
| 1058 |
+
"eval_accuracy": 0.4238085730096768,
|
| 1059 |
+
"eval_loss": 2.9544410705566406,
|
| 1060 |
+
"eval_runtime": 112.4181,
|
| 1061 |
+
"eval_samples_per_second": 466.473,
|
| 1062 |
+
"eval_steps_per_second": 7.294,
|
| 1063 |
+
"step": 139887
|
| 1064 |
+
},
|
| 1065 |
+
{
|
| 1066 |
+
"epoch": 9.007270153766969,
|
| 1067 |
+
"grad_norm": 0.36276528239250183,
|
| 1068 |
+
"learning_rate": 0.00018386968371225703,
|
| 1069 |
+
"loss": 2.6469,
|
| 1070 |
+
"step": 140000
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 9.071607797722447,
|
| 1074 |
+
"grad_norm": 0.36368831992149353,
|
| 1075 |
+
"learning_rate": 0.0001827938750627555,
|
| 1076 |
+
"loss": 2.5666,
|
| 1077 |
+
"step": 141000
|
| 1078 |
+
},
|
| 1079 |
+
{
|
| 1080 |
+
"epoch": 9.135945441677926,
|
| 1081 |
+
"grad_norm": 0.36417004466056824,
|
| 1082 |
+
"learning_rate": 0.00018171806641325395,
|
| 1083 |
+
"loss": 2.5832,
|
| 1084 |
+
"step": 142000
|
| 1085 |
+
},
|
| 1086 |
+
{
|
| 1087 |
+
"epoch": 9.200283085633403,
|
| 1088 |
+
"grad_norm": 0.3550620973110199,
|
| 1089 |
+
"learning_rate": 0.0001806422577637524,
|
| 1090 |
+
"loss": 2.5888,
|
| 1091 |
+
"step": 143000
|
| 1092 |
+
},
|
| 1093 |
+
{
|
| 1094 |
+
"epoch": 9.264620729588882,
|
| 1095 |
+
"grad_norm": 0.3513035178184509,
|
| 1096 |
+
"learning_rate": 0.00017956644911425084,
|
| 1097 |
+
"loss": 2.5872,
|
| 1098 |
+
"step": 144000
|
| 1099 |
+
},
|
| 1100 |
+
{
|
| 1101 |
+
"epoch": 9.32895837354436,
|
| 1102 |
+
"grad_norm": 0.3576969802379608,
|
| 1103 |
+
"learning_rate": 0.00017849279208204832,
|
| 1104 |
+
"loss": 2.599,
|
| 1105 |
+
"step": 145000
|
| 1106 |
+
},
|
| 1107 |
+
{
|
| 1108 |
+
"epoch": 9.39329601749984,
|
| 1109 |
+
"grad_norm": 0.3496710956096649,
|
| 1110 |
+
"learning_rate": 0.00017741698343254678,
|
| 1111 |
+
"loss": 2.6042,
|
| 1112 |
+
"step": 146000
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 9.457633661455317,
|
| 1116 |
+
"grad_norm": 0.3502206802368164,
|
| 1117 |
+
"learning_rate": 0.00017634225059169476,
|
| 1118 |
+
"loss": 2.6069,
|
| 1119 |
+
"step": 147000
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 9.521971305410796,
|
| 1123 |
+
"grad_norm": 0.3516786992549896,
|
| 1124 |
+
"learning_rate": 0.00017526644194219322,
|
| 1125 |
+
"loss": 2.606,
|
| 1126 |
+
"step": 148000
|
| 1127 |
+
},
|
| 1128 |
+
{
|
| 1129 |
+
"epoch": 9.586308949366273,
|
| 1130 |
+
"grad_norm": 0.3671824336051941,
|
| 1131 |
+
"learning_rate": 0.00017419063329269168,
|
| 1132 |
+
"loss": 2.6151,
|
| 1133 |
+
"step": 149000
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"epoch": 9.650646593321753,
|
| 1137 |
+
"grad_norm": 0.36615684628486633,
|
| 1138 |
+
"learning_rate": 0.00017311590045183964,
|
| 1139 |
+
"loss": 2.6174,
|
| 1140 |
+
"step": 150000
|
| 1141 |
+
},
|
| 1142 |
+
{
|
| 1143 |
+
"epoch": 9.71498423727723,
|
| 1144 |
+
"grad_norm": 0.369759202003479,
|
| 1145 |
+
"learning_rate": 0.0001720400918023381,
|
| 1146 |
+
"loss": 2.6162,
|
| 1147 |
+
"step": 151000
|
| 1148 |
+
},
|
| 1149 |
+
{
|
| 1150 |
+
"epoch": 9.77932188123271,
|
| 1151 |
+
"grad_norm": 0.3495037257671356,
|
| 1152 |
+
"learning_rate": 0.00017096428315283656,
|
| 1153 |
+
"loss": 2.6186,
|
| 1154 |
+
"step": 152000
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 9.843659525188187,
|
| 1158 |
+
"grad_norm": 0.3635868728160858,
|
| 1159 |
+
"learning_rate": 0.0001698895503119845,
|
| 1160 |
+
"loss": 2.616,
|
| 1161 |
+
"step": 153000
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 9.907997169143666,
|
| 1165 |
+
"grad_norm": 0.352250337600708,
|
| 1166 |
+
"learning_rate": 0.00016881374166248297,
|
| 1167 |
+
"loss": 2.626,
|
| 1168 |
+
"step": 154000
|
| 1169 |
+
},
|
| 1170 |
+
{
|
| 1171 |
+
"epoch": 9.972334813099144,
|
| 1172 |
+
"grad_norm": 0.3688776195049286,
|
| 1173 |
+
"learning_rate": 0.00016773900882163093,
|
| 1174 |
+
"loss": 2.6246,
|
| 1175 |
+
"step": 155000
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"epoch": 10.0,
|
| 1179 |
+
"eval_accuracy": 0.424411016885778,
|
| 1180 |
+
"eval_loss": 2.9523308277130127,
|
| 1181 |
+
"eval_runtime": 112.0636,
|
| 1182 |
+
"eval_samples_per_second": 467.948,
|
| 1183 |
+
"eval_steps_per_second": 7.317,
|
| 1184 |
+
"step": 155430
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 10.036672457054623,
|
| 1188 |
+
"grad_norm": 0.3961314558982849,
|
| 1189 |
+
"learning_rate": 0.0001666632001721294,
|
| 1190 |
+
"loss": 2.5827,
|
| 1191 |
+
"step": 156000
|
| 1192 |
+
},
|
| 1193 |
+
{
|
| 1194 |
+
"epoch": 10.1010101010101,
|
| 1195 |
+
"grad_norm": 0.3705954849720001,
|
| 1196 |
+
"learning_rate": 0.00016558739152262782,
|
| 1197 |
+
"loss": 2.5413,
|
| 1198 |
+
"step": 157000
|
| 1199 |
+
},
|
| 1200 |
+
{
|
| 1201 |
+
"epoch": 10.16534774496558,
|
| 1202 |
+
"grad_norm": 0.37091416120529175,
|
| 1203 |
+
"learning_rate": 0.00016451158287312628,
|
| 1204 |
+
"loss": 2.5502,
|
| 1205 |
+
"step": 158000
|
| 1206 |
+
},
|
| 1207 |
+
{
|
| 1208 |
+
"epoch": 10.229685388921057,
|
| 1209 |
+
"grad_norm": 0.38428565859794617,
|
| 1210 |
+
"learning_rate": 0.00016343685003227424,
|
| 1211 |
+
"loss": 2.5592,
|
| 1212 |
+
"step": 159000
|
| 1213 |
+
},
|
| 1214 |
+
{
|
| 1215 |
+
"epoch": 10.294023032876536,
|
| 1216 |
+
"grad_norm": 0.3688577115535736,
|
| 1217 |
+
"learning_rate": 0.0001623610413827727,
|
| 1218 |
+
"loss": 2.5673,
|
| 1219 |
+
"step": 160000
|
| 1220 |
+
},
|
| 1221 |
+
{
|
| 1222 |
+
"epoch": 10.358360676832014,
|
| 1223 |
+
"grad_norm": 0.38183775544166565,
|
| 1224 |
+
"learning_rate": 0.00016128630854192065,
|
| 1225 |
+
"loss": 2.5697,
|
| 1226 |
+
"step": 161000
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 10.422698320787493,
|
| 1230 |
+
"grad_norm": 0.37677517533302307,
|
| 1231 |
+
"learning_rate": 0.0001602104998924191,
|
| 1232 |
+
"loss": 2.5713,
|
| 1233 |
+
"step": 162000
|
| 1234 |
+
},
|
| 1235 |
+
{
|
| 1236 |
+
"epoch": 10.48703596474297,
|
| 1237 |
+
"grad_norm": 0.3694332540035248,
|
| 1238 |
+
"learning_rate": 0.00015913576705156707,
|
| 1239 |
+
"loss": 2.5751,
|
| 1240 |
+
"step": 163000
|
| 1241 |
+
},
|
| 1242 |
+
{
|
| 1243 |
+
"epoch": 10.55137360869845,
|
| 1244 |
+
"grad_norm": 0.3814958333969116,
|
| 1245 |
+
"learning_rate": 0.00015806103421071502,
|
| 1246 |
+
"loss": 2.5792,
|
| 1247 |
+
"step": 164000
|
| 1248 |
+
},
|
| 1249 |
+
{
|
| 1250 |
+
"epoch": 10.615711252653927,
|
| 1251 |
+
"grad_norm": 0.38280004262924194,
|
| 1252 |
+
"learning_rate": 0.00015698522556121348,
|
| 1253 |
+
"loss": 2.5782,
|
| 1254 |
+
"step": 165000
|
| 1255 |
+
},
|
| 1256 |
+
{
|
| 1257 |
+
"epoch": 10.680048896609406,
|
| 1258 |
+
"grad_norm": 0.3659280240535736,
|
| 1259 |
+
"learning_rate": 0.00015590941691171194,
|
| 1260 |
+
"loss": 2.5862,
|
| 1261 |
+
"step": 166000
|
| 1262 |
+
},
|
| 1263 |
+
{
|
| 1264 |
+
"epoch": 10.744386540564884,
|
| 1265 |
+
"grad_norm": 0.34562841057777405,
|
| 1266 |
+
"learning_rate": 0.0001548336082622104,
|
| 1267 |
+
"loss": 2.5869,
|
| 1268 |
+
"step": 167000
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 10.808724184520363,
|
| 1272 |
+
"grad_norm": 0.3570345938205719,
|
| 1273 |
+
"learning_rate": 0.00015375887542135836,
|
| 1274 |
+
"loss": 2.59,
|
| 1275 |
+
"step": 168000
|
| 1276 |
+
},
|
| 1277 |
+
{
|
| 1278 |
+
"epoch": 10.87306182847584,
|
| 1279 |
+
"grad_norm": 0.360215961933136,
|
| 1280 |
+
"learning_rate": 0.00015268306677185682,
|
| 1281 |
+
"loss": 2.5979,
|
| 1282 |
+
"step": 169000
|
| 1283 |
+
},
|
| 1284 |
+
{
|
| 1285 |
+
"epoch": 10.93739947243132,
|
| 1286 |
+
"grad_norm": 0.370670884847641,
|
| 1287 |
+
"learning_rate": 0.00015160725812235528,
|
| 1288 |
+
"loss": 2.5988,
|
| 1289 |
+
"step": 170000
|
| 1290 |
+
},
|
| 1291 |
+
{
|
| 1292 |
+
"epoch": 11.0,
|
| 1293 |
+
"eval_accuracy": 0.4248191770987571,
|
| 1294 |
+
"eval_loss": 2.9567785263061523,
|
| 1295 |
+
"eval_runtime": 112.9375,
|
| 1296 |
+
"eval_samples_per_second": 464.328,
|
| 1297 |
+
"eval_steps_per_second": 7.261,
|
| 1298 |
+
"step": 170973
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 11.001737116386797,
|
| 1302 |
+
"grad_norm": 0.38218948245048523,
|
| 1303 |
+
"learning_rate": 0.00015053252528150323,
|
| 1304 |
+
"loss": 2.5933,
|
| 1305 |
+
"step": 171000
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 11.066074760342277,
|
| 1309 |
+
"grad_norm": 0.396331787109375,
|
| 1310 |
+
"learning_rate": 0.00014945671663200172,
|
| 1311 |
+
"loss": 2.5023,
|
| 1312 |
+
"step": 172000
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 11.130412404297754,
|
| 1316 |
+
"grad_norm": 0.3751789927482605,
|
| 1317 |
+
"learning_rate": 0.00014838090798250018,
|
| 1318 |
+
"loss": 2.5227,
|
| 1319 |
+
"step": 173000
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 11.194750048253233,
|
| 1323 |
+
"grad_norm": 0.37265828251838684,
|
| 1324 |
+
"learning_rate": 0.00014730509933299864,
|
| 1325 |
+
"loss": 2.5299,
|
| 1326 |
+
"step": 174000
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 11.25908769220871,
|
| 1330 |
+
"grad_norm": 0.37080228328704834,
|
| 1331 |
+
"learning_rate": 0.0001462303664921466,
|
| 1332 |
+
"loss": 2.5333,
|
| 1333 |
+
"step": 175000
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 11.32342533616419,
|
| 1337 |
+
"grad_norm": 0.3808966875076294,
|
| 1338 |
+
"learning_rate": 0.00014515563365129455,
|
| 1339 |
+
"loss": 2.5376,
|
| 1340 |
+
"step": 176000
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 11.387762980119668,
|
| 1344 |
+
"grad_norm": 0.38901346921920776,
|
| 1345 |
+
"learning_rate": 0.000144079825001793,
|
| 1346 |
+
"loss": 2.5422,
|
| 1347 |
+
"step": 177000
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 11.452100624075147,
|
| 1351 |
+
"grad_norm": 0.380100816488266,
|
| 1352 |
+
"learning_rate": 0.00014300401635229144,
|
| 1353 |
+
"loss": 2.5533,
|
| 1354 |
+
"step": 178000
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 11.516438268030624,
|
| 1358 |
+
"grad_norm": 0.39306920766830444,
|
| 1359 |
+
"learning_rate": 0.0001419282077027899,
|
| 1360 |
+
"loss": 2.5507,
|
| 1361 |
+
"step": 179000
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 11.580775911986104,
|
| 1365 |
+
"grad_norm": 0.3917422890663147,
|
| 1366 |
+
"learning_rate": 0.00014085239905328836,
|
| 1367 |
+
"loss": 2.5579,
|
| 1368 |
+
"step": 180000
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 11.645113555941581,
|
| 1372 |
+
"grad_norm": 0.38742849230766296,
|
| 1373 |
+
"learning_rate": 0.00013977766621243632,
|
| 1374 |
+
"loss": 2.5531,
|
| 1375 |
+
"step": 181000
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 11.70945119989706,
|
| 1379 |
+
"grad_norm": 0.3767852187156677,
|
| 1380 |
+
"learning_rate": 0.00013870185756293478,
|
| 1381 |
+
"loss": 2.5633,
|
| 1382 |
+
"step": 182000
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 11.773788843852538,
|
| 1386 |
+
"grad_norm": 0.39576900005340576,
|
| 1387 |
+
"learning_rate": 0.00013762604891343324,
|
| 1388 |
+
"loss": 2.5648,
|
| 1389 |
+
"step": 183000
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 11.838126487808017,
|
| 1393 |
+
"grad_norm": 0.37659791111946106,
|
| 1394 |
+
"learning_rate": 0.00013655131607258122,
|
| 1395 |
+
"loss": 2.5631,
|
| 1396 |
+
"step": 184000
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 11.902464131763494,
|
| 1400 |
+
"grad_norm": 0.38377416133880615,
|
| 1401 |
+
"learning_rate": 0.00013547658323172918,
|
| 1402 |
+
"loss": 2.5631,
|
| 1403 |
+
"step": 185000
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 11.966801775718974,
|
| 1407 |
+
"grad_norm": 0.37857234477996826,
|
| 1408 |
+
"learning_rate": 0.00013440077458222764,
|
| 1409 |
+
"loss": 2.5639,
|
| 1410 |
+
"step": 186000
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 12.0,
|
| 1414 |
+
"eval_accuracy": 0.4247610714766456,
|
| 1415 |
+
"eval_loss": 2.9595353603363037,
|
| 1416 |
+
"eval_runtime": 112.0415,
|
| 1417 |
+
"eval_samples_per_second": 468.041,
|
| 1418 |
+
"eval_steps_per_second": 7.319,
|
| 1419 |
+
"step": 186516
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"epoch": 12.031139419674451,
|
| 1423 |
+
"grad_norm": 0.4024442136287689,
|
| 1424 |
+
"learning_rate": 0.0001333249659327261,
|
| 1425 |
+
"loss": 2.5273,
|
| 1426 |
+
"step": 187000
|
| 1427 |
+
},
|
| 1428 |
+
{
|
| 1429 |
+
"epoch": 12.09547706362993,
|
| 1430 |
+
"grad_norm": 0.4137458801269531,
|
| 1431 |
+
"learning_rate": 0.00013225023309187405,
|
| 1432 |
+
"loss": 2.4933,
|
| 1433 |
+
"step": 188000
|
| 1434 |
+
},
|
| 1435 |
+
{
|
| 1436 |
+
"epoch": 12.159814707585408,
|
| 1437 |
+
"grad_norm": 0.409184992313385,
|
| 1438 |
+
"learning_rate": 0.0001311744244423725,
|
| 1439 |
+
"loss": 2.4967,
|
| 1440 |
+
"step": 189000
|
| 1441 |
+
},
|
| 1442 |
+
{
|
| 1443 |
+
"epoch": 12.224152351540887,
|
| 1444 |
+
"grad_norm": 0.41316309571266174,
|
| 1445 |
+
"learning_rate": 0.00013009861579287097,
|
| 1446 |
+
"loss": 2.5063,
|
| 1447 |
+
"step": 190000
|
| 1448 |
+
},
|
| 1449 |
+
{
|
| 1450 |
+
"epoch": 12.288489995496365,
|
| 1451 |
+
"grad_norm": 0.3909110724925995,
|
| 1452 |
+
"learning_rate": 0.00012902280714336943,
|
| 1453 |
+
"loss": 2.5153,
|
| 1454 |
+
"step": 191000
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 12.352827639451844,
|
| 1458 |
+
"grad_norm": 0.39046111702919006,
|
| 1459 |
+
"learning_rate": 0.0001279469984938679,
|
| 1460 |
+
"loss": 2.5115,
|
| 1461 |
+
"step": 192000
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 12.417165283407321,
|
| 1465 |
+
"grad_norm": 0.40070855617523193,
|
| 1466 |
+
"learning_rate": 0.00012687226565301585,
|
| 1467 |
+
"loss": 2.5157,
|
| 1468 |
+
"step": 193000
|
| 1469 |
+
},
|
| 1470 |
+
{
|
| 1471 |
+
"epoch": 12.4815029273628,
|
| 1472 |
+
"grad_norm": 0.3970703184604645,
|
| 1473 |
+
"learning_rate": 0.00012579645700351428,
|
| 1474 |
+
"loss": 2.5198,
|
| 1475 |
+
"step": 194000
|
| 1476 |
+
},
|
| 1477 |
+
{
|
| 1478 |
+
"epoch": 12.545840571318278,
|
| 1479 |
+
"grad_norm": 0.40202242136001587,
|
| 1480 |
+
"learning_rate": 0.00012472064835401274,
|
| 1481 |
+
"loss": 2.526,
|
| 1482 |
+
"step": 195000
|
| 1483 |
+
},
|
| 1484 |
+
{
|
| 1485 |
+
"epoch": 12.610178215273757,
|
| 1486 |
+
"grad_norm": 0.3841732144355774,
|
| 1487 |
+
"learning_rate": 0.0001236459155131607,
|
| 1488 |
+
"loss": 2.5295,
|
| 1489 |
+
"step": 196000
|
| 1490 |
+
},
|
| 1491 |
+
{
|
| 1492 |
+
"epoch": 12.674515859229235,
|
| 1493 |
+
"grad_norm": 0.40759024024009705,
|
| 1494 |
+
"learning_rate": 0.00012257010686365916,
|
| 1495 |
+
"loss": 2.5307,
|
| 1496 |
+
"step": 197000
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 12.738853503184714,
|
| 1500 |
+
"grad_norm": 0.3963831663131714,
|
| 1501 |
+
"learning_rate": 0.00012149429821415763,
|
| 1502 |
+
"loss": 2.534,
|
| 1503 |
+
"step": 198000
|
| 1504 |
+
},
|
| 1505 |
+
{
|
| 1506 |
+
"epoch": 12.803191147140192,
|
| 1507 |
+
"grad_norm": 0.37255486845970154,
|
| 1508 |
+
"learning_rate": 0.0001204195653733056,
|
| 1509 |
+
"loss": 2.5354,
|
| 1510 |
+
"step": 199000
|
| 1511 |
+
},
|
| 1512 |
+
{
|
| 1513 |
+
"epoch": 12.86752879109567,
|
| 1514 |
+
"grad_norm": 0.397368460893631,
|
| 1515 |
+
"learning_rate": 0.00011934375672380406,
|
| 1516 |
+
"loss": 2.5352,
|
| 1517 |
+
"step": 200000
|
| 1518 |
+
},
|
| 1519 |
+
{
|
| 1520 |
+
"epoch": 12.931866435051148,
|
| 1521 |
+
"grad_norm": 0.379574716091156,
|
| 1522 |
+
"learning_rate": 0.00011826902388295201,
|
| 1523 |
+
"loss": 2.5397,
|
| 1524 |
+
"step": 201000
|
| 1525 |
+
},
|
| 1526 |
+
{
|
| 1527 |
+
"epoch": 12.996204079006628,
|
| 1528 |
+
"grad_norm": 0.3803842067718506,
|
| 1529 |
+
"learning_rate": 0.00011719321523345048,
|
| 1530 |
+
"loss": 2.5361,
|
| 1531 |
+
"step": 202000
|
| 1532 |
+
},
|
| 1533 |
+
{
|
| 1534 |
+
"epoch": 13.0,
|
| 1535 |
+
"eval_accuracy": 0.42475613586395655,
|
| 1536 |
+
"eval_loss": 2.9698119163513184,
|
| 1537 |
+
"eval_runtime": 112.2708,
|
| 1538 |
+
"eval_samples_per_second": 467.085,
|
| 1539 |
+
"eval_steps_per_second": 7.304,
|
| 1540 |
+
"step": 202059
|
| 1541 |
+
},
|
| 1542 |
+
{
|
| 1543 |
+
"epoch": 13.060541722962105,
|
| 1544 |
+
"grad_norm": 0.40816885232925415,
|
| 1545 |
+
"learning_rate": 0.00011611740658394894,
|
| 1546 |
+
"loss": 2.4669,
|
| 1547 |
+
"step": 203000
|
| 1548 |
+
},
|
| 1549 |
+
{
|
| 1550 |
+
"epoch": 13.124879366917584,
|
| 1551 |
+
"grad_norm": 0.42818671464920044,
|
| 1552 |
+
"learning_rate": 0.00011504159793444738,
|
| 1553 |
+
"loss": 2.467,
|
| 1554 |
+
"step": 204000
|
| 1555 |
+
},
|
| 1556 |
+
{
|
| 1557 |
+
"epoch": 13.189217010873062,
|
| 1558 |
+
"grad_norm": 0.40255987644195557,
|
| 1559 |
+
"learning_rate": 0.00011396686509359535,
|
| 1560 |
+
"loss": 2.4753,
|
| 1561 |
+
"step": 205000
|
| 1562 |
+
},
|
| 1563 |
+
{
|
| 1564 |
+
"epoch": 13.253554654828541,
|
| 1565 |
+
"grad_norm": 0.4254453778266907,
|
| 1566 |
+
"learning_rate": 0.0001128921322527433,
|
| 1567 |
+
"loss": 2.4808,
|
| 1568 |
+
"step": 206000
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 13.317892298784018,
|
| 1572 |
+
"grad_norm": 0.4060657322406769,
|
| 1573 |
+
"learning_rate": 0.00011181632360324175,
|
| 1574 |
+
"loss": 2.4932,
|
| 1575 |
+
"step": 207000
|
| 1576 |
+
},
|
| 1577 |
+
{
|
| 1578 |
+
"epoch": 13.382229942739498,
|
| 1579 |
+
"grad_norm": 0.4138365387916565,
|
| 1580 |
+
"learning_rate": 0.00011074051495374021,
|
| 1581 |
+
"loss": 2.4922,
|
| 1582 |
+
"step": 208000
|
| 1583 |
+
},
|
| 1584 |
+
{
|
| 1585 |
+
"epoch": 13.446567586694975,
|
| 1586 |
+
"grad_norm": 0.4098254442214966,
|
| 1587 |
+
"learning_rate": 0.00010966578211288817,
|
| 1588 |
+
"loss": 2.4948,
|
| 1589 |
+
"step": 209000
|
| 1590 |
+
},
|
| 1591 |
+
{
|
| 1592 |
+
"epoch": 13.510905230650454,
|
| 1593 |
+
"grad_norm": 0.4242159128189087,
|
| 1594 |
+
"learning_rate": 0.00010858997346338663,
|
| 1595 |
+
"loss": 2.5012,
|
| 1596 |
+
"step": 210000
|
| 1597 |
+
},
|
| 1598 |
+
{
|
| 1599 |
+
"epoch": 13.575242874605932,
|
| 1600 |
+
"grad_norm": 0.42177829146385193,
|
| 1601 |
+
"learning_rate": 0.00010751416481388509,
|
| 1602 |
+
"loss": 2.4998,
|
| 1603 |
+
"step": 211000
|
| 1604 |
+
},
|
| 1605 |
+
{
|
| 1606 |
+
"epoch": 13.63958051856141,
|
| 1607 |
+
"grad_norm": 0.4196189045906067,
|
| 1608 |
+
"learning_rate": 0.00010643943197303304,
|
| 1609 |
+
"loss": 2.5048,
|
| 1610 |
+
"step": 212000
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 13.703918162516889,
|
| 1614 |
+
"grad_norm": 0.3965640366077423,
|
| 1615 |
+
"learning_rate": 0.0001053636233235315,
|
| 1616 |
+
"loss": 2.5092,
|
| 1617 |
+
"step": 213000
|
| 1618 |
+
},
|
| 1619 |
+
{
|
| 1620 |
+
"epoch": 13.768255806472368,
|
| 1621 |
+
"grad_norm": 0.39778339862823486,
|
| 1622 |
+
"learning_rate": 0.00010428781467402996,
|
| 1623 |
+
"loss": 2.5121,
|
| 1624 |
+
"step": 214000
|
| 1625 |
+
},
|
| 1626 |
+
{
|
| 1627 |
+
"epoch": 13.832593450427845,
|
| 1628 |
+
"grad_norm": 0.40292391180992126,
|
| 1629 |
+
"learning_rate": 0.00010321308183317793,
|
| 1630 |
+
"loss": 2.5119,
|
| 1631 |
+
"step": 215000
|
| 1632 |
+
},
|
| 1633 |
+
{
|
| 1634 |
+
"epoch": 13.896931094383323,
|
| 1635 |
+
"grad_norm": 0.41673198342323303,
|
| 1636 |
+
"learning_rate": 0.00010213727318367639,
|
| 1637 |
+
"loss": 2.5112,
|
| 1638 |
+
"step": 216000
|
| 1639 |
+
},
|
| 1640 |
+
{
|
| 1641 |
+
"epoch": 13.961268738338802,
|
| 1642 |
+
"grad_norm": 0.40400612354278564,
|
| 1643 |
+
"learning_rate": 0.00010106254034282435,
|
| 1644 |
+
"loss": 2.5098,
|
| 1645 |
+
"step": 217000
|
| 1646 |
+
},
|
| 1647 |
+
{
|
| 1648 |
+
"epoch": 14.0,
|
| 1649 |
+
"eval_accuracy": 0.424743796832234,
|
| 1650 |
+
"eval_loss": 2.9747180938720703,
|
| 1651 |
+
"eval_runtime": 112.0802,
|
| 1652 |
+
"eval_samples_per_second": 467.879,
|
| 1653 |
+
"eval_steps_per_second": 7.316,
|
| 1654 |
+
"step": 217602
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"epoch": 14.02560638229428,
|
| 1658 |
+
"grad_norm": 0.40745100378990173,
|
| 1659 |
+
"learning_rate": 9.998673169332281e-05,
|
| 1660 |
+
"loss": 2.4894,
|
| 1661 |
+
"step": 218000
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 14.089944026249759,
|
| 1665 |
+
"grad_norm": 0.42399463057518005,
|
| 1666 |
+
"learning_rate": 9.891092304382127e-05,
|
| 1667 |
+
"loss": 2.449,
|
| 1668 |
+
"step": 219000
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 14.154281670205236,
|
| 1672 |
+
"grad_norm": 0.4149724841117859,
|
| 1673 |
+
"learning_rate": 9.783511439431973e-05,
|
| 1674 |
+
"loss": 2.4534,
|
| 1675 |
+
"step": 220000
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 14.218619314160716,
|
| 1679 |
+
"grad_norm": 0.40756285190582275,
|
| 1680 |
+
"learning_rate": 9.676145736211718e-05,
|
| 1681 |
+
"loss": 2.4576,
|
| 1682 |
+
"step": 221000
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 14.282956958116193,
|
| 1686 |
+
"grad_norm": 0.4224795997142792,
|
| 1687 |
+
"learning_rate": 9.568564871261564e-05,
|
| 1688 |
+
"loss": 2.4584,
|
| 1689 |
+
"step": 222000
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 14.347294602071672,
|
| 1693 |
+
"grad_norm": 0.41213053464889526,
|
| 1694 |
+
"learning_rate": 9.461091587176359e-05,
|
| 1695 |
+
"loss": 2.4707,
|
| 1696 |
+
"step": 223000
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 14.41163224602715,
|
| 1700 |
+
"grad_norm": 0.4161031246185303,
|
| 1701 |
+
"learning_rate": 9.353510722226205e-05,
|
| 1702 |
+
"loss": 2.4701,
|
| 1703 |
+
"step": 224000
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 14.475969889982629,
|
| 1707 |
+
"grad_norm": 0.42417025566101074,
|
| 1708 |
+
"learning_rate": 9.245929857276051e-05,
|
| 1709 |
+
"loss": 2.4706,
|
| 1710 |
+
"step": 225000
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 14.540307533938106,
|
| 1714 |
+
"grad_norm": 0.4227360785007477,
|
| 1715 |
+
"learning_rate": 9.138348992325897e-05,
|
| 1716 |
+
"loss": 2.4678,
|
| 1717 |
+
"step": 226000
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 14.604645177893586,
|
| 1721 |
+
"grad_norm": 0.3956305682659149,
|
| 1722 |
+
"learning_rate": 9.030768127375742e-05,
|
| 1723 |
+
"loss": 2.4816,
|
| 1724 |
+
"step": 227000
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 14.668982821849063,
|
| 1728 |
+
"grad_norm": 0.42013561725616455,
|
| 1729 |
+
"learning_rate": 8.92329484329054e-05,
|
| 1730 |
+
"loss": 2.4791,
|
| 1731 |
+
"step": 228000
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 14.733320465804542,
|
| 1735 |
+
"grad_norm": 0.41232335567474365,
|
| 1736 |
+
"learning_rate": 8.815713978340386e-05,
|
| 1737 |
+
"loss": 2.4861,
|
| 1738 |
+
"step": 229000
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 14.79765810976002,
|
| 1742 |
+
"grad_norm": 0.398253858089447,
|
| 1743 |
+
"learning_rate": 8.708240694255182e-05,
|
| 1744 |
+
"loss": 2.4857,
|
| 1745 |
+
"step": 230000
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 14.8619957537155,
|
| 1749 |
+
"grad_norm": 0.41056615114212036,
|
| 1750 |
+
"learning_rate": 8.600659829305028e-05,
|
| 1751 |
+
"loss": 2.4826,
|
| 1752 |
+
"step": 231000
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 14.926333397670977,
|
| 1756 |
+
"grad_norm": 0.4065124988555908,
|
| 1757 |
+
"learning_rate": 8.493186545219823e-05,
|
| 1758 |
+
"loss": 2.4791,
|
| 1759 |
+
"step": 232000
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 14.990671041626456,
|
| 1763 |
+
"grad_norm": 0.42194780707359314,
|
| 1764 |
+
"learning_rate": 8.385605680269669e-05,
|
| 1765 |
+
"loss": 2.4899,
|
| 1766 |
+
"step": 233000
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 15.0,
|
| 1770 |
+
"eval_accuracy": 0.4246625087868862,
|
| 1771 |
+
"eval_loss": 2.9792003631591797,
|
| 1772 |
+
"eval_runtime": 112.3403,
|
| 1773 |
+
"eval_samples_per_second": 466.796,
|
| 1774 |
+
"eval_steps_per_second": 7.299,
|
| 1775 |
+
"step": 233145
|
| 1776 |
+
},
|
| 1777 |
+
{
|
| 1778 |
+
"epoch": 15.055008685581933,
|
| 1779 |
+
"grad_norm": 0.444181889295578,
|
| 1780 |
+
"learning_rate": 8.278024815319514e-05,
|
| 1781 |
+
"loss": 2.4309,
|
| 1782 |
+
"step": 234000
|
| 1783 |
+
},
|
| 1784 |
+
{
|
| 1785 |
+
"epoch": 15.119346329537413,
|
| 1786 |
+
"grad_norm": 0.4177301526069641,
|
| 1787 |
+
"learning_rate": 8.17044395036936e-05,
|
| 1788 |
+
"loss": 2.4254,
|
| 1789 |
+
"step": 235000
|
| 1790 |
+
},
|
| 1791 |
+
{
|
| 1792 |
+
"epoch": 15.18368397349289,
|
| 1793 |
+
"grad_norm": 0.43864157795906067,
|
| 1794 |
+
"learning_rate": 8.062970666284155e-05,
|
| 1795 |
+
"loss": 2.432,
|
| 1796 |
+
"step": 236000
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 15.24802161744837,
|
| 1800 |
+
"grad_norm": 0.43071264028549194,
|
| 1801 |
+
"learning_rate": 7.955497382198951e-05,
|
| 1802 |
+
"loss": 2.4372,
|
| 1803 |
+
"step": 237000
|
| 1804 |
+
},
|
| 1805 |
+
{
|
| 1806 |
+
"epoch": 15.312359261403847,
|
| 1807 |
+
"grad_norm": 0.44551989436149597,
|
| 1808 |
+
"learning_rate": 7.847916517248797e-05,
|
| 1809 |
+
"loss": 2.4441,
|
| 1810 |
+
"step": 238000
|
| 1811 |
+
},
|
| 1812 |
+
{
|
| 1813 |
+
"epoch": 15.376696905359326,
|
| 1814 |
+
"grad_norm": 0.42598387598991394,
|
| 1815 |
+
"learning_rate": 7.740335652298643e-05,
|
| 1816 |
+
"loss": 2.4448,
|
| 1817 |
+
"step": 239000
|
| 1818 |
+
},
|
| 1819 |
+
{
|
| 1820 |
+
"epoch": 15.441034549314804,
|
| 1821 |
+
"grad_norm": 0.4412069618701935,
|
| 1822 |
+
"learning_rate": 7.632754787348489e-05,
|
| 1823 |
+
"loss": 2.4481,
|
| 1824 |
+
"step": 240000
|
| 1825 |
+
},
|
| 1826 |
+
{
|
| 1827 |
+
"epoch": 15.505372193270283,
|
| 1828 |
+
"grad_norm": 0.4257245361804962,
|
| 1829 |
+
"learning_rate": 7.525173922398335e-05,
|
| 1830 |
+
"loss": 2.4496,
|
| 1831 |
+
"step": 241000
|
| 1832 |
+
},
|
| 1833 |
+
{
|
| 1834 |
+
"epoch": 15.56970983722576,
|
| 1835 |
+
"grad_norm": 0.4463740885257721,
|
| 1836 |
+
"learning_rate": 7.417593057448181e-05,
|
| 1837 |
+
"loss": 2.4583,
|
| 1838 |
+
"step": 242000
|
| 1839 |
+
},
|
| 1840 |
+
{
|
| 1841 |
+
"epoch": 15.63404748118124,
|
| 1842 |
+
"grad_norm": 0.40843266248703003,
|
| 1843 |
+
"learning_rate": 7.310119773362977e-05,
|
| 1844 |
+
"loss": 2.4549,
|
| 1845 |
+
"step": 243000
|
| 1846 |
+
},
|
| 1847 |
+
{
|
| 1848 |
+
"epoch": 15.698385125136717,
|
| 1849 |
+
"grad_norm": 0.43823161721229553,
|
| 1850 |
+
"learning_rate": 7.202538908412823e-05,
|
| 1851 |
+
"loss": 2.4565,
|
| 1852 |
+
"step": 244000
|
| 1853 |
+
},
|
| 1854 |
+
{
|
| 1855 |
+
"epoch": 15.762722769092196,
|
| 1856 |
+
"grad_norm": 0.4224304258823395,
|
| 1857 |
+
"learning_rate": 7.09506562432762e-05,
|
| 1858 |
+
"loss": 2.4664,
|
| 1859 |
+
"step": 245000
|
| 1860 |
+
},
|
| 1861 |
+
{
|
| 1862 |
+
"epoch": 15.827060413047674,
|
| 1863 |
+
"grad_norm": 0.42779698967933655,
|
| 1864 |
+
"learning_rate": 6.987484759377464e-05,
|
| 1865 |
+
"loss": 2.4607,
|
| 1866 |
+
"step": 246000
|
| 1867 |
+
},
|
| 1868 |
+
{
|
| 1869 |
+
"epoch": 15.891398057003153,
|
| 1870 |
+
"grad_norm": 0.41904374957084656,
|
| 1871 |
+
"learning_rate": 6.880011475292261e-05,
|
| 1872 |
+
"loss": 2.463,
|
| 1873 |
+
"step": 247000
|
| 1874 |
+
},
|
| 1875 |
+
{
|
| 1876 |
+
"epoch": 15.95573570095863,
|
| 1877 |
+
"grad_norm": 0.4636126458644867,
|
| 1878 |
+
"learning_rate": 6.772430610342107e-05,
|
| 1879 |
+
"loss": 2.4626,
|
| 1880 |
+
"step": 248000
|
| 1881 |
+
},
|
| 1882 |
+
{
|
| 1883 |
+
"epoch": 16.0,
|
| 1884 |
+
"eval_accuracy": 0.4244173733566653,
|
| 1885 |
+
"eval_loss": 2.9882283210754395,
|
| 1886 |
+
"eval_runtime": 112.1119,
|
| 1887 |
+
"eval_samples_per_second": 467.747,
|
| 1888 |
+
"eval_steps_per_second": 7.314,
|
| 1889 |
+
"step": 248688
|
| 1890 |
+
},
|
| 1891 |
+
{
|
| 1892 |
+
"epoch": 16.02007334491411,
|
| 1893 |
+
"grad_norm": 0.44689562916755676,
|
| 1894 |
+
"learning_rate": 6.664849745391953e-05,
|
| 1895 |
+
"loss": 2.4432,
|
| 1896 |
+
"step": 249000
|
| 1897 |
+
},
|
| 1898 |
+
{
|
| 1899 |
+
"epoch": 16.08441098886959,
|
| 1900 |
+
"grad_norm": 0.45889049768447876,
|
| 1901 |
+
"learning_rate": 6.557376461306749e-05,
|
| 1902 |
+
"loss": 2.4048,
|
| 1903 |
+
"step": 250000
|
| 1904 |
+
},
|
| 1905 |
+
{
|
| 1906 |
+
"epoch": 16.148748632825065,
|
| 1907 |
+
"grad_norm": 0.4538269639015198,
|
| 1908 |
+
"learning_rate": 6.449795596356593e-05,
|
| 1909 |
+
"loss": 2.4123,
|
| 1910 |
+
"step": 251000
|
| 1911 |
+
},
|
| 1912 |
+
{
|
| 1913 |
+
"epoch": 16.213086276780544,
|
| 1914 |
+
"grad_norm": 0.44775742292404175,
|
| 1915 |
+
"learning_rate": 6.342214731406439e-05,
|
| 1916 |
+
"loss": 2.4137,
|
| 1917 |
+
"step": 252000
|
| 1918 |
+
},
|
| 1919 |
+
{
|
| 1920 |
+
"epoch": 16.277423920736023,
|
| 1921 |
+
"grad_norm": 0.4506843090057373,
|
| 1922 |
+
"learning_rate": 6.234741447321236e-05,
|
| 1923 |
+
"loss": 2.4152,
|
| 1924 |
+
"step": 253000
|
| 1925 |
+
},
|
| 1926 |
+
{
|
| 1927 |
+
"epoch": 16.341761564691502,
|
| 1928 |
+
"grad_norm": 0.4564642310142517,
|
| 1929 |
+
"learning_rate": 6.127160582371082e-05,
|
| 1930 |
+
"loss": 2.4236,
|
| 1931 |
+
"step": 254000
|
| 1932 |
+
},
|
| 1933 |
+
{
|
| 1934 |
+
"epoch": 16.406099208646978,
|
| 1935 |
+
"grad_norm": 0.4492376744747162,
|
| 1936 |
+
"learning_rate": 6.0195797174209275e-05,
|
| 1937 |
+
"loss": 2.4222,
|
| 1938 |
+
"step": 255000
|
| 1939 |
+
},
|
| 1940 |
+
{
|
| 1941 |
+
"epoch": 16.470436852602457,
|
| 1942 |
+
"grad_norm": 0.44002753496170044,
|
| 1943 |
+
"learning_rate": 5.9119988524707736e-05,
|
| 1944 |
+
"loss": 2.4277,
|
| 1945 |
+
"step": 256000
|
| 1946 |
+
},
|
| 1947 |
+
{
|
| 1948 |
+
"epoch": 16.534774496557937,
|
| 1949 |
+
"grad_norm": 0.437580406665802,
|
| 1950 |
+
"learning_rate": 5.8044179875206196e-05,
|
| 1951 |
+
"loss": 2.4303,
|
| 1952 |
+
"step": 257000
|
| 1953 |
+
},
|
| 1954 |
+
{
|
| 1955 |
+
"epoch": 16.599112140513416,
|
| 1956 |
+
"grad_norm": 0.42502424120903015,
|
| 1957 |
+
"learning_rate": 5.697052284300365e-05,
|
| 1958 |
+
"loss": 2.4359,
|
| 1959 |
+
"step": 258000
|
| 1960 |
+
},
|
| 1961 |
+
{
|
| 1962 |
+
"epoch": 16.66344978446889,
|
| 1963 |
+
"grad_norm": 0.44441190361976624,
|
| 1964 |
+
"learning_rate": 5.5894714193502106e-05,
|
| 1965 |
+
"loss": 2.4306,
|
| 1966 |
+
"step": 259000
|
| 1967 |
+
},
|
| 1968 |
+
{
|
| 1969 |
+
"epoch": 16.72778742842437,
|
| 1970 |
+
"grad_norm": 0.4539526700973511,
|
| 1971 |
+
"learning_rate": 5.4818905544000566e-05,
|
| 1972 |
+
"loss": 2.4342,
|
| 1973 |
+
"step": 260000
|
| 1974 |
+
},
|
| 1975 |
+
{
|
| 1976 |
+
"epoch": 16.79212507237985,
|
| 1977 |
+
"grad_norm": 0.4554595947265625,
|
| 1978 |
+
"learning_rate": 5.374417270314853e-05,
|
| 1979 |
+
"loss": 2.4388,
|
| 1980 |
+
"step": 261000
|
| 1981 |
+
},
|
| 1982 |
+
{
|
| 1983 |
+
"epoch": 16.85646271633533,
|
| 1984 |
+
"grad_norm": 0.4573330283164978,
|
| 1985 |
+
"learning_rate": 5.266836405364699e-05,
|
| 1986 |
+
"loss": 2.441,
|
| 1987 |
+
"step": 262000
|
| 1988 |
+
},
|
| 1989 |
+
{
|
| 1990 |
+
"epoch": 16.920800360290805,
|
| 1991 |
+
"grad_norm": 0.449770450592041,
|
| 1992 |
+
"learning_rate": 5.159255540414545e-05,
|
| 1993 |
+
"loss": 2.4411,
|
| 1994 |
+
"step": 263000
|
| 1995 |
+
},
|
| 1996 |
+
{
|
| 1997 |
+
"epoch": 16.985138004246284,
|
| 1998 |
+
"grad_norm": 0.48139625787734985,
|
| 1999 |
+
"learning_rate": 5.05178225632934e-05,
|
| 2000 |
+
"loss": 2.4399,
|
| 2001 |
+
"step": 264000
|
| 2002 |
+
},
|
| 2003 |
+
{
|
| 2004 |
+
"epoch": 17.0,
|
| 2005 |
+
"eval_accuracy": 0.4242649676193895,
|
| 2006 |
+
"eval_loss": 2.9961202144622803,
|
| 2007 |
+
"eval_runtime": 112.3106,
|
| 2008 |
+
"eval_samples_per_second": 466.919,
|
| 2009 |
+
"eval_steps_per_second": 7.301,
|
| 2010 |
+
"step": 264231
|
| 2011 |
+
},
|
| 2012 |
+
{
|
| 2013 |
+
"epoch": 17.049475648201764,
|
| 2014 |
+
"grad_norm": 0.4543195366859436,
|
| 2015 |
+
"learning_rate": 4.9442013913791863e-05,
|
| 2016 |
+
"loss": 2.4013,
|
| 2017 |
+
"step": 265000
|
| 2018 |
+
},
|
| 2019 |
+
{
|
| 2020 |
+
"epoch": 17.113813292157243,
|
| 2021 |
+
"grad_norm": 0.4699794054031372,
|
| 2022 |
+
"learning_rate": 4.836620526429032e-05,
|
| 2023 |
+
"loss": 2.3928,
|
| 2024 |
+
"step": 266000
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 17.17815093611272,
|
| 2028 |
+
"grad_norm": 0.4636929929256439,
|
| 2029 |
+
"learning_rate": 4.7291472423438285e-05,
|
| 2030 |
+
"loss": 2.3989,
|
| 2031 |
+
"step": 267000
|
| 2032 |
+
},
|
| 2033 |
+
{
|
| 2034 |
+
"epoch": 17.242488580068198,
|
| 2035 |
+
"grad_norm": 0.4614698886871338,
|
| 2036 |
+
"learning_rate": 4.6215663773936746e-05,
|
| 2037 |
+
"loss": 2.4004,
|
| 2038 |
+
"step": 268000
|
| 2039 |
+
},
|
| 2040 |
+
{
|
| 2041 |
+
"epoch": 17.306826224023677,
|
| 2042 |
+
"grad_norm": 0.46002906560897827,
|
| 2043 |
+
"learning_rate": 4.513985512443519e-05,
|
| 2044 |
+
"loss": 2.3982,
|
| 2045 |
+
"step": 269000
|
| 2046 |
+
},
|
| 2047 |
+
{
|
| 2048 |
+
"epoch": 17.371163867979156,
|
| 2049 |
+
"grad_norm": 0.42619064450263977,
|
| 2050 |
+
"learning_rate": 4.4065122283583154e-05,
|
| 2051 |
+
"loss": 2.4053,
|
| 2052 |
+
"step": 270000
|
| 2053 |
+
},
|
| 2054 |
+
{
|
| 2055 |
+
"epoch": 17.435501511934632,
|
| 2056 |
+
"grad_norm": 0.45975300669670105,
|
| 2057 |
+
"learning_rate": 4.2989313634081614e-05,
|
| 2058 |
+
"loss": 2.4041,
|
| 2059 |
+
"step": 271000
|
| 2060 |
+
},
|
| 2061 |
+
{
|
| 2062 |
+
"epoch": 17.49983915589011,
|
| 2063 |
+
"grad_norm": 0.4545740485191345,
|
| 2064 |
+
"learning_rate": 4.1913504984580075e-05,
|
| 2065 |
+
"loss": 2.406,
|
| 2066 |
+
"step": 272000
|
| 2067 |
+
},
|
| 2068 |
+
{
|
| 2069 |
+
"epoch": 17.56417679984559,
|
| 2070 |
+
"grad_norm": 0.458011269569397,
|
| 2071 |
+
"learning_rate": 4.083769633507853e-05,
|
| 2072 |
+
"loss": 2.4168,
|
| 2073 |
+
"step": 273000
|
| 2074 |
+
},
|
| 2075 |
+
{
|
| 2076 |
+
"epoch": 17.62851444380107,
|
| 2077 |
+
"grad_norm": 0.4604107439517975,
|
| 2078 |
+
"learning_rate": 3.976296349422649e-05,
|
| 2079 |
+
"loss": 2.411,
|
| 2080 |
+
"step": 274000
|
| 2081 |
+
},
|
| 2082 |
+
{
|
| 2083 |
+
"epoch": 17.692852087756545,
|
| 2084 |
+
"grad_norm": 0.4420773684978485,
|
| 2085 |
+
"learning_rate": 3.8687154844724944e-05,
|
| 2086 |
+
"loss": 2.4144,
|
| 2087 |
+
"step": 275000
|
| 2088 |
+
},
|
| 2089 |
+
{
|
| 2090 |
+
"epoch": 17.757189731712025,
|
| 2091 |
+
"grad_norm": 0.45774900913238525,
|
| 2092 |
+
"learning_rate": 3.7611346195223404e-05,
|
| 2093 |
+
"loss": 2.412,
|
| 2094 |
+
"step": 276000
|
| 2095 |
+
},
|
| 2096 |
+
{
|
| 2097 |
+
"epoch": 17.821527375667504,
|
| 2098 |
+
"grad_norm": 0.4509606659412384,
|
| 2099 |
+
"learning_rate": 3.6536613354371366e-05,
|
| 2100 |
+
"loss": 2.4086,
|
| 2101 |
+
"step": 277000
|
| 2102 |
+
},
|
| 2103 |
+
{
|
| 2104 |
+
"epoch": 17.885865019622983,
|
| 2105 |
+
"grad_norm": 0.4442935883998871,
|
| 2106 |
+
"learning_rate": 3.5460804704869826e-05,
|
| 2107 |
+
"loss": 2.4134,
|
| 2108 |
+
"step": 278000
|
| 2109 |
+
},
|
| 2110 |
+
{
|
| 2111 |
+
"epoch": 17.95020266357846,
|
| 2112 |
+
"grad_norm": 0.42292436957359314,
|
| 2113 |
+
"learning_rate": 3.438607186401778e-05,
|
| 2114 |
+
"loss": 2.4186,
|
| 2115 |
+
"step": 279000
|
| 2116 |
+
},
|
| 2117 |
+
{
|
| 2118 |
+
"epoch": 18.0,
|
| 2119 |
+
"eval_accuracy": 0.42388941236296196,
|
| 2120 |
+
"eval_loss": 3.0051016807556152,
|
| 2121 |
+
"eval_runtime": 112.3705,
|
| 2122 |
+
"eval_samples_per_second": 466.67,
|
| 2123 |
+
"eval_steps_per_second": 7.297,
|
| 2124 |
+
"step": 279774
|
| 2125 |
+
},
|
| 2126 |
+
{
|
| 2127 |
+
"epoch": 18.014540307533938,
|
| 2128 |
+
"grad_norm": 0.48824623227119446,
|
| 2129 |
+
"learning_rate": 3.331026321451624e-05,
|
| 2130 |
+
"loss": 2.4055,
|
| 2131 |
+
"step": 280000
|
| 2132 |
+
},
|
| 2133 |
+
{
|
| 2134 |
+
"epoch": 18.078877951489417,
|
| 2135 |
+
"grad_norm": 0.46934977173805237,
|
| 2136 |
+
"learning_rate": 3.22355303736642e-05,
|
| 2137 |
+
"loss": 2.3736,
|
| 2138 |
+
"step": 281000
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 18.143215595444893,
|
| 2142 |
+
"grad_norm": 0.5045217275619507,
|
| 2143 |
+
"learning_rate": 3.115972172416266e-05,
|
| 2144 |
+
"loss": 2.382,
|
| 2145 |
+
"step": 282000
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 18.207553239400372,
|
| 2149 |
+
"grad_norm": 0.46461954712867737,
|
| 2150 |
+
"learning_rate": 3.008391307466112e-05,
|
| 2151 |
+
"loss": 2.3806,
|
| 2152 |
+
"step": 283000
|
| 2153 |
+
},
|
| 2154 |
+
{
|
| 2155 |
+
"epoch": 18.27189088335585,
|
| 2156 |
+
"grad_norm": 0.4565331041812897,
|
| 2157 |
+
"learning_rate": 2.9009180233809078e-05,
|
| 2158 |
+
"loss": 2.3813,
|
| 2159 |
+
"step": 284000
|
| 2160 |
+
},
|
| 2161 |
+
{
|
| 2162 |
+
"epoch": 18.33622852731133,
|
| 2163 |
+
"grad_norm": 0.4561784863471985,
|
| 2164 |
+
"learning_rate": 2.793337158430754e-05,
|
| 2165 |
+
"loss": 2.3863,
|
| 2166 |
+
"step": 285000
|
| 2167 |
+
},
|
| 2168 |
+
{
|
| 2169 |
+
"epoch": 18.400566171266806,
|
| 2170 |
+
"grad_norm": 0.4438989758491516,
|
| 2171 |
+
"learning_rate": 2.6858638743455493e-05,
|
| 2172 |
+
"loss": 2.3845,
|
| 2173 |
+
"step": 286000
|
| 2174 |
+
},
|
| 2175 |
+
{
|
| 2176 |
+
"epoch": 18.464903815222286,
|
| 2177 |
+
"grad_norm": 0.461086630821228,
|
| 2178 |
+
"learning_rate": 2.578283009395395e-05,
|
| 2179 |
+
"loss": 2.3833,
|
| 2180 |
+
"step": 287000
|
| 2181 |
+
},
|
| 2182 |
+
{
|
| 2183 |
+
"epoch": 18.529241459177765,
|
| 2184 |
+
"grad_norm": 0.4639764726161957,
|
| 2185 |
+
"learning_rate": 2.470702144445241e-05,
|
| 2186 |
+
"loss": 2.3918,
|
| 2187 |
+
"step": 288000
|
| 2188 |
+
},
|
| 2189 |
+
{
|
| 2190 |
+
"epoch": 18.593579103133244,
|
| 2191 |
+
"grad_norm": 0.4645422697067261,
|
| 2192 |
+
"learning_rate": 2.3631212794950868e-05,
|
| 2193 |
+
"loss": 2.3953,
|
| 2194 |
+
"step": 289000
|
| 2195 |
+
},
|
| 2196 |
+
{
|
| 2197 |
+
"epoch": 18.65791674708872,
|
| 2198 |
+
"grad_norm": 0.47392553091049194,
|
| 2199 |
+
"learning_rate": 2.2555404145449328e-05,
|
| 2200 |
+
"loss": 2.3829,
|
| 2201 |
+
"step": 290000
|
| 2202 |
+
},
|
| 2203 |
+
{
|
| 2204 |
+
"epoch": 18.7222543910442,
|
| 2205 |
+
"grad_norm": 0.4530762732028961,
|
| 2206 |
+
"learning_rate": 2.148174711324679e-05,
|
| 2207 |
+
"loss": 2.3904,
|
| 2208 |
+
"step": 291000
|
| 2209 |
+
},
|
| 2210 |
+
{
|
| 2211 |
+
"epoch": 18.78659203499968,
|
| 2212 |
+
"grad_norm": 0.47473639249801636,
|
| 2213 |
+
"learning_rate": 2.0405938463745248e-05,
|
| 2214 |
+
"loss": 2.3966,
|
| 2215 |
+
"step": 292000
|
| 2216 |
+
},
|
| 2217 |
+
{
|
| 2218 |
+
"epoch": 18.850929678955158,
|
| 2219 |
+
"grad_norm": 0.43500351905822754,
|
| 2220 |
+
"learning_rate": 1.9330129814243705e-05,
|
| 2221 |
+
"loss": 2.396,
|
| 2222 |
+
"step": 293000
|
| 2223 |
+
},
|
| 2224 |
+
{
|
| 2225 |
+
"epoch": 18.915267322910633,
|
| 2226 |
+
"grad_norm": 0.45157596468925476,
|
| 2227 |
+
"learning_rate": 1.8254321164742165e-05,
|
| 2228 |
+
"loss": 2.3959,
|
| 2229 |
+
"step": 294000
|
| 2230 |
+
},
|
| 2231 |
+
{
|
| 2232 |
+
"epoch": 18.979604966866113,
|
| 2233 |
+
"grad_norm": 0.4546051621437073,
|
| 2234 |
+
"learning_rate": 1.7180664132539624e-05,
|
| 2235 |
+
"loss": 2.3869,
|
| 2236 |
+
"step": 295000
|
| 2237 |
+
},
|
| 2238 |
+
{
|
| 2239 |
+
"epoch": 19.0,
|
| 2240 |
+
"eval_accuracy": 0.42373925008599933,
|
| 2241 |
+
"eval_loss": 3.011887311935425,
|
| 2242 |
+
"eval_runtime": 112.2632,
|
| 2243 |
+
"eval_samples_per_second": 467.116,
|
| 2244 |
+
"eval_steps_per_second": 7.304,
|
| 2245 |
+
"step": 295317
|
| 2246 |
+
},
|
| 2247 |
+
{
|
| 2248 |
+
"epoch": 19.043942610821592,
|
| 2249 |
+
"grad_norm": 0.46901893615722656,
|
| 2250 |
+
"learning_rate": 1.610485548303808e-05,
|
| 2251 |
+
"loss": 2.3726,
|
| 2252 |
+
"step": 296000
|
| 2253 |
+
},
|
| 2254 |
+
{
|
| 2255 |
+
"epoch": 19.10828025477707,
|
| 2256 |
+
"grad_norm": 0.43862438201904297,
|
| 2257 |
+
"learning_rate": 1.502904683353654e-05,
|
| 2258 |
+
"loss": 2.3688,
|
| 2259 |
+
"step": 297000
|
| 2260 |
+
},
|
| 2261 |
+
{
|
| 2262 |
+
"epoch": 19.172617898732547,
|
| 2263 |
+
"grad_norm": 0.4580424427986145,
|
| 2264 |
+
"learning_rate": 1.3954313992684501e-05,
|
| 2265 |
+
"loss": 2.3682,
|
| 2266 |
+
"step": 298000
|
| 2267 |
+
},
|
| 2268 |
+
{
|
| 2269 |
+
"epoch": 19.236955542688026,
|
| 2270 |
+
"grad_norm": 0.47557470202445984,
|
| 2271 |
+
"learning_rate": 1.2878505343182957e-05,
|
| 2272 |
+
"loss": 2.3687,
|
| 2273 |
+
"step": 299000
|
| 2274 |
+
},
|
| 2275 |
+
{
|
| 2276 |
+
"epoch": 19.301293186643505,
|
| 2277 |
+
"grad_norm": 0.48615992069244385,
|
| 2278 |
+
"learning_rate": 1.1802696693681415e-05,
|
| 2279 |
+
"loss": 2.3636,
|
| 2280 |
+
"step": 300000
|
| 2281 |
+
},
|
| 2282 |
+
{
|
| 2283 |
+
"epoch": 19.365630830598985,
|
| 2284 |
+
"grad_norm": 0.5019800662994385,
|
| 2285 |
+
"learning_rate": 1.0726888044179874e-05,
|
| 2286 |
+
"loss": 2.3668,
|
| 2287 |
+
"step": 301000
|
| 2288 |
+
},
|
| 2289 |
+
{
|
| 2290 |
+
"epoch": 19.42996847455446,
|
| 2291 |
+
"grad_norm": 0.4481401741504669,
|
| 2292 |
+
"learning_rate": 9.652155203327834e-06,
|
| 2293 |
+
"loss": 2.3721,
|
| 2294 |
+
"step": 302000
|
| 2295 |
+
},
|
| 2296 |
+
{
|
| 2297 |
+
"epoch": 19.49430611850994,
|
| 2298 |
+
"grad_norm": 0.4632056653499603,
|
| 2299 |
+
"learning_rate": 8.577422362475793e-06,
|
| 2300 |
+
"loss": 2.372,
|
| 2301 |
+
"step": 303000
|
| 2302 |
+
},
|
| 2303 |
+
{
|
| 2304 |
+
"epoch": 19.55864376246542,
|
| 2305 |
+
"grad_norm": 0.4590476453304291,
|
| 2306 |
+
"learning_rate": 7.5016137129742514e-06,
|
| 2307 |
+
"loss": 2.3725,
|
| 2308 |
+
"step": 304000
|
| 2309 |
+
},
|
| 2310 |
+
{
|
| 2311 |
+
"epoch": 19.622981406420898,
|
| 2312 |
+
"grad_norm": 0.4774569272994995,
|
| 2313 |
+
"learning_rate": 6.42580506347271e-06,
|
| 2314 |
+
"loss": 2.37,
|
| 2315 |
+
"step": 305000
|
| 2316 |
+
},
|
| 2317 |
+
{
|
| 2318 |
+
"epoch": 19.687319050376374,
|
| 2319 |
+
"grad_norm": 0.47048863768577576,
|
| 2320 |
+
"learning_rate": 5.351072222620669e-06,
|
| 2321 |
+
"loss": 2.371,
|
| 2322 |
+
"step": 306000
|
| 2323 |
+
},
|
| 2324 |
+
{
|
| 2325 |
+
"epoch": 19.751656694331853,
|
| 2326 |
+
"grad_norm": 0.4567144215106964,
|
| 2327 |
+
"learning_rate": 4.275263573119128e-06,
|
| 2328 |
+
"loss": 2.3706,
|
| 2329 |
+
"step": 307000
|
| 2330 |
+
},
|
| 2331 |
+
{
|
| 2332 |
+
"epoch": 19.815994338287332,
|
| 2333 |
+
"grad_norm": 0.4492277503013611,
|
| 2334 |
+
"learning_rate": 3.200530732267087e-06,
|
| 2335 |
+
"loss": 2.3714,
|
| 2336 |
+
"step": 308000
|
| 2337 |
+
},
|
| 2338 |
+
{
|
| 2339 |
+
"epoch": 19.88033198224281,
|
| 2340 |
+
"grad_norm": 0.44562822580337524,
|
| 2341 |
+
"learning_rate": 2.1247220827655454e-06,
|
| 2342 |
+
"loss": 2.3711,
|
| 2343 |
+
"step": 309000
|
| 2344 |
+
},
|
| 2345 |
+
{
|
| 2346 |
+
"epoch": 19.944669626198287,
|
| 2347 |
+
"grad_norm": 0.4758046269416809,
|
| 2348 |
+
"learning_rate": 1.0499892419135049e-06,
|
| 2349 |
+
"loss": 2.3686,
|
| 2350 |
+
"step": 310000
|
| 2351 |
+
},
|
| 2352 |
+
{
|
| 2353 |
+
"epoch": 20.0,
|
| 2354 |
+
"eval_accuracy": 0.4234014597448438,
|
| 2355 |
+
"eval_loss": 3.0190186500549316,
|
| 2356 |
+
"eval_runtime": 112.3566,
|
| 2357 |
+
"eval_samples_per_second": 466.728,
|
| 2358 |
+
"eval_steps_per_second": 7.298,
|
| 2359 |
+
"step": 310860
|
| 2360 |
+
},
|
| 2361 |
+
{
|
| 2362 |
+
"epoch": 20.0,
|
| 2363 |
+
"step": 310860,
|
| 2364 |
+
"total_flos": 1.29957250203648e+18,
|
| 2365 |
+
"train_loss": 2.7038125842232534,
|
| 2366 |
+
"train_runtime": 44110.0674,
|
| 2367 |
+
"train_samples_per_second": 225.51,
|
| 2368 |
+
"train_steps_per_second": 7.047
|
| 2369 |
+
}
|
| 2370 |
+
],
|
| 2371 |
+
"logging_steps": 1000,
|
| 2372 |
+
"max_steps": 310860,
|
| 2373 |
+
"num_input_tokens_seen": 0,
|
| 2374 |
+
"num_train_epochs": 20,
|
| 2375 |
+
"save_steps": 5000,
|
| 2376 |
+
"stateful_callbacks": {
|
| 2377 |
+
"TrainerControl": {
|
| 2378 |
+
"args": {
|
| 2379 |
+
"should_epoch_stop": false,
|
| 2380 |
+
"should_evaluate": false,
|
| 2381 |
+
"should_log": false,
|
| 2382 |
+
"should_save": true,
|
| 2383 |
+
"should_training_stop": true
|
| 2384 |
+
},
|
| 2385 |
+
"attributes": {}
|
| 2386 |
+
}
|
| 2387 |
+
},
|
| 2388 |
+
"total_flos": 1.29957250203648e+18,
|
| 2389 |
+
"train_batch_size": 32,
|
| 2390 |
+
"trial_name": null,
|
| 2391 |
+
"trial_params": null
|
| 2392 |
+
}
|