End of training
Browse files- README.md +14 -2
- all_results.json +16 -0
- eval_results.json +10 -0
- tokenizer.json +1 -6
- train_results.json +9 -0
- trainer_state.json +945 -0
README.md
CHANGED
|
@@ -3,11 +3,23 @@ license: other
|
|
| 3 |
base_model: Qwen/Qwen1.5-4B
|
| 4 |
tags:
|
| 5 |
- generated_from_trainer
|
|
|
|
|
|
|
| 6 |
metrics:
|
| 7 |
- accuracy
|
| 8 |
model-index:
|
| 9 |
- name: lmind_nq_train6000_eval6489_v1_recite_qa_v3_Qwen_Qwen1.5-4B_3e-4_lora2
|
| 10 |
-
results:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 11 |
library_name: peft
|
| 12 |
---
|
| 13 |
|
|
@@ -16,7 +28,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 16 |
|
| 17 |
# lmind_nq_train6000_eval6489_v1_recite_qa_v3_Qwen_Qwen1.5-4B_3e-4_lora2
|
| 18 |
|
| 19 |
-
This model is a fine-tuned version of [Qwen/Qwen1.5-4B](https://huggingface.co/Qwen/Qwen1.5-4B) on
|
| 20 |
It achieves the following results on the evaluation set:
|
| 21 |
- Loss: 0.4311
|
| 22 |
- Accuracy: 0.7981
|
|
|
|
| 3 |
base_model: Qwen/Qwen1.5-4B
|
| 4 |
tags:
|
| 5 |
- generated_from_trainer
|
| 6 |
+
datasets:
|
| 7 |
+
- tyzhu/lmind_nq_train6000_eval6489_v1_recite_qa_v3
|
| 8 |
metrics:
|
| 9 |
- accuracy
|
| 10 |
model-index:
|
| 11 |
- name: lmind_nq_train6000_eval6489_v1_recite_qa_v3_Qwen_Qwen1.5-4B_3e-4_lora2
|
| 12 |
+
results:
|
| 13 |
+
- task:
|
| 14 |
+
name: Causal Language Modeling
|
| 15 |
+
type: text-generation
|
| 16 |
+
dataset:
|
| 17 |
+
name: tyzhu/lmind_nq_train6000_eval6489_v1_recite_qa_v3
|
| 18 |
+
type: tyzhu/lmind_nq_train6000_eval6489_v1_recite_qa_v3
|
| 19 |
+
metrics:
|
| 20 |
+
- name: Accuracy
|
| 21 |
+
type: accuracy
|
| 22 |
+
value: 0.7981434977578475
|
| 23 |
library_name: peft
|
| 24 |
---
|
| 25 |
|
|
|
|
| 28 |
|
| 29 |
# lmind_nq_train6000_eval6489_v1_recite_qa_v3_Qwen_Qwen1.5-4B_3e-4_lora2
|
| 30 |
|
| 31 |
+
This model is a fine-tuned version of [Qwen/Qwen1.5-4B](https://huggingface.co/Qwen/Qwen1.5-4B) on the tyzhu/lmind_nq_train6000_eval6489_v1_recite_qa_v3 dataset.
|
| 32 |
It achieves the following results on the evaluation set:
|
| 33 |
- Loss: 0.4311
|
| 34 |
- Accuracy: 0.7981
|
all_results.json
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 20.0,
|
| 3 |
+
"eval_accuracy": 0.7981434977578475,
|
| 4 |
+
"eval_loss": 0.43110987544059753,
|
| 5 |
+
"eval_runtime": 7.0185,
|
| 6 |
+
"eval_samples": 500,
|
| 7 |
+
"eval_samples_per_second": 71.24,
|
| 8 |
+
"eval_steps_per_second": 8.976,
|
| 9 |
+
"perplexity": 1.5389646350848876,
|
| 10 |
+
"total_flos": 9.64242245391745e+17,
|
| 11 |
+
"train_loss": 0.3966417311720677,
|
| 12 |
+
"train_runtime": 24179.107,
|
| 13 |
+
"train_samples": 16925,
|
| 14 |
+
"train_samples_per_second": 14.0,
|
| 15 |
+
"train_steps_per_second": 0.438
|
| 16 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 20.0,
|
| 3 |
+
"eval_accuracy": 0.7981434977578475,
|
| 4 |
+
"eval_loss": 0.43110987544059753,
|
| 5 |
+
"eval_runtime": 7.0185,
|
| 6 |
+
"eval_samples": 500,
|
| 7 |
+
"eval_samples_per_second": 71.24,
|
| 8 |
+
"eval_steps_per_second": 8.976,
|
| 9 |
+
"perplexity": 1.5389646350848876
|
| 10 |
+
}
|
tokenizer.json
CHANGED
|
@@ -1,11 +1,6 @@
|
|
| 1 |
{
|
| 2 |
"version": "1.0",
|
| 3 |
-
"truncation":
|
| 4 |
-
"direction": "Right",
|
| 5 |
-
"max_length": 1024,
|
| 6 |
-
"strategy": "LongestFirst",
|
| 7 |
-
"stride": 0
|
| 8 |
-
},
|
| 9 |
"padding": null,
|
| 10 |
"added_tokens": [
|
| 11 |
{
|
|
|
|
| 1 |
{
|
| 2 |
"version": "1.0",
|
| 3 |
+
"truncation": null,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4 |
"padding": null,
|
| 5 |
"added_tokens": [
|
| 6 |
{
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 20.0,
|
| 3 |
+
"total_flos": 9.64242245391745e+17,
|
| 4 |
+
"train_loss": 0.3966417311720677,
|
| 5 |
+
"train_runtime": 24179.107,
|
| 6 |
+
"train_samples": 16925,
|
| 7 |
+
"train_samples_per_second": 14.0,
|
| 8 |
+
"train_steps_per_second": 0.438
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,945 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 20.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 10580,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.1890359168241966,
|
| 13 |
+
"grad_norm": 0.2088678479194641,
|
| 14 |
+
"learning_rate": 0.0003,
|
| 15 |
+
"loss": 1.909,
|
| 16 |
+
"step": 100
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.3780718336483932,
|
| 20 |
+
"grad_norm": 0.24642816185951233,
|
| 21 |
+
"learning_rate": 0.0003,
|
| 22 |
+
"loss": 1.8584,
|
| 23 |
+
"step": 200
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.5671077504725898,
|
| 27 |
+
"grad_norm": 0.27153250575065613,
|
| 28 |
+
"learning_rate": 0.0003,
|
| 29 |
+
"loss": 1.8175,
|
| 30 |
+
"step": 300
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.7561436672967864,
|
| 34 |
+
"grad_norm": 0.3125714957714081,
|
| 35 |
+
"learning_rate": 0.0003,
|
| 36 |
+
"loss": 1.8035,
|
| 37 |
+
"step": 400
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.945179584120983,
|
| 41 |
+
"grad_norm": 0.29630693793296814,
|
| 42 |
+
"learning_rate": 0.0003,
|
| 43 |
+
"loss": 1.7637,
|
| 44 |
+
"step": 500
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 1.0,
|
| 48 |
+
"eval_accuracy": 0.6288430493273542,
|
| 49 |
+
"eval_loss": 1.4994953870773315,
|
| 50 |
+
"eval_runtime": 6.5869,
|
| 51 |
+
"eval_samples_per_second": 75.908,
|
| 52 |
+
"eval_steps_per_second": 9.564,
|
| 53 |
+
"step": 529
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 1.1342155009451795,
|
| 57 |
+
"grad_norm": 0.42783793807029724,
|
| 58 |
+
"learning_rate": 0.0003,
|
| 59 |
+
"loss": 1.5915,
|
| 60 |
+
"step": 600
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"epoch": 1.3232514177693762,
|
| 64 |
+
"grad_norm": 0.5046432614326477,
|
| 65 |
+
"learning_rate": 0.0003,
|
| 66 |
+
"loss": 1.4725,
|
| 67 |
+
"step": 700
|
| 68 |
+
},
|
| 69 |
+
{
|
| 70 |
+
"epoch": 1.5122873345935728,
|
| 71 |
+
"grad_norm": 0.5139931440353394,
|
| 72 |
+
"learning_rate": 0.0003,
|
| 73 |
+
"loss": 1.4387,
|
| 74 |
+
"step": 800
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 1.7013232514177694,
|
| 78 |
+
"grad_norm": 0.5190715789794922,
|
| 79 |
+
"learning_rate": 0.0003,
|
| 80 |
+
"loss": 1.4124,
|
| 81 |
+
"step": 900
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 1.8903591682419658,
|
| 85 |
+
"grad_norm": 0.5980258584022522,
|
| 86 |
+
"learning_rate": 0.0003,
|
| 87 |
+
"loss": 1.3986,
|
| 88 |
+
"step": 1000
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"epoch": 2.0,
|
| 92 |
+
"eval_accuracy": 0.6719730941704036,
|
| 93 |
+
"eval_loss": 1.171139121055603,
|
| 94 |
+
"eval_runtime": 6.5804,
|
| 95 |
+
"eval_samples_per_second": 75.983,
|
| 96 |
+
"eval_steps_per_second": 9.574,
|
| 97 |
+
"step": 1058
|
| 98 |
+
},
|
| 99 |
+
{
|
| 100 |
+
"epoch": 2.0793950850661624,
|
| 101 |
+
"grad_norm": 0.614499568939209,
|
| 102 |
+
"learning_rate": 0.0003,
|
| 103 |
+
"loss": 1.207,
|
| 104 |
+
"step": 1100
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 2.268431001890359,
|
| 108 |
+
"grad_norm": 0.812962532043457,
|
| 109 |
+
"learning_rate": 0.0003,
|
| 110 |
+
"loss": 0.9878,
|
| 111 |
+
"step": 1200
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 2.4574669187145557,
|
| 115 |
+
"grad_norm": 0.7435634732246399,
|
| 116 |
+
"learning_rate": 0.0003,
|
| 117 |
+
"loss": 0.9801,
|
| 118 |
+
"step": 1300
|
| 119 |
+
},
|
| 120 |
+
{
|
| 121 |
+
"epoch": 2.6465028355387523,
|
| 122 |
+
"grad_norm": 0.778424859046936,
|
| 123 |
+
"learning_rate": 0.0003,
|
| 124 |
+
"loss": 0.972,
|
| 125 |
+
"step": 1400
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 2.835538752362949,
|
| 129 |
+
"grad_norm": 0.7896220684051514,
|
| 130 |
+
"learning_rate": 0.0003,
|
| 131 |
+
"loss": 0.9515,
|
| 132 |
+
"step": 1500
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"epoch": 3.0,
|
| 136 |
+
"eval_accuracy": 0.7147713004484305,
|
| 137 |
+
"eval_loss": 0.8765971660614014,
|
| 138 |
+
"eval_runtime": 6.6432,
|
| 139 |
+
"eval_samples_per_second": 75.265,
|
| 140 |
+
"eval_steps_per_second": 9.483,
|
| 141 |
+
"step": 1587
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 3.0245746691871456,
|
| 145 |
+
"grad_norm": 0.8365605473518372,
|
| 146 |
+
"learning_rate": 0.0003,
|
| 147 |
+
"loss": 0.8826,
|
| 148 |
+
"step": 1600
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 3.213610586011342,
|
| 152 |
+
"grad_norm": 0.8311352729797363,
|
| 153 |
+
"learning_rate": 0.0003,
|
| 154 |
+
"loss": 0.5808,
|
| 155 |
+
"step": 1700
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"epoch": 3.402646502835539,
|
| 159 |
+
"grad_norm": 0.7168679237365723,
|
| 160 |
+
"learning_rate": 0.0003,
|
| 161 |
+
"loss": 0.611,
|
| 162 |
+
"step": 1800
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"epoch": 3.5916824196597354,
|
| 166 |
+
"grad_norm": 0.9022857546806335,
|
| 167 |
+
"learning_rate": 0.0003,
|
| 168 |
+
"loss": 0.6249,
|
| 169 |
+
"step": 1900
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 3.780718336483932,
|
| 173 |
+
"grad_norm": 1.0664108991622925,
|
| 174 |
+
"learning_rate": 0.0003,
|
| 175 |
+
"loss": 0.6378,
|
| 176 |
+
"step": 2000
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 3.9697542533081287,
|
| 180 |
+
"grad_norm": 0.8847127556800842,
|
| 181 |
+
"learning_rate": 0.0003,
|
| 182 |
+
"loss": 0.642,
|
| 183 |
+
"step": 2100
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 4.0,
|
| 187 |
+
"eval_accuracy": 0.7477668161434977,
|
| 188 |
+
"eval_loss": 0.6719540953636169,
|
| 189 |
+
"eval_runtime": 7.1688,
|
| 190 |
+
"eval_samples_per_second": 69.747,
|
| 191 |
+
"eval_steps_per_second": 8.788,
|
| 192 |
+
"step": 2116
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 4.158790170132325,
|
| 196 |
+
"grad_norm": 0.765385627746582,
|
| 197 |
+
"learning_rate": 0.0003,
|
| 198 |
+
"loss": 0.4204,
|
| 199 |
+
"step": 2200
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 4.3478260869565215,
|
| 203 |
+
"grad_norm": 0.8121228218078613,
|
| 204 |
+
"learning_rate": 0.0003,
|
| 205 |
+
"loss": 0.3987,
|
| 206 |
+
"step": 2300
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 4.536862003780718,
|
| 210 |
+
"grad_norm": 0.8688237071037292,
|
| 211 |
+
"learning_rate": 0.0003,
|
| 212 |
+
"loss": 0.4055,
|
| 213 |
+
"step": 2400
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 4.725897920604915,
|
| 217 |
+
"grad_norm": 0.8493298292160034,
|
| 218 |
+
"learning_rate": 0.0003,
|
| 219 |
+
"loss": 0.4276,
|
| 220 |
+
"step": 2500
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 4.914933837429111,
|
| 224 |
+
"grad_norm": 0.8626968264579773,
|
| 225 |
+
"learning_rate": 0.0003,
|
| 226 |
+
"loss": 0.4362,
|
| 227 |
+
"step": 2600
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 5.0,
|
| 231 |
+
"eval_accuracy": 0.7697219730941705,
|
| 232 |
+
"eval_loss": 0.5458089709281921,
|
| 233 |
+
"eval_runtime": 6.5529,
|
| 234 |
+
"eval_samples_per_second": 76.302,
|
| 235 |
+
"eval_steps_per_second": 9.614,
|
| 236 |
+
"step": 2645
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 5.103969754253308,
|
| 240 |
+
"grad_norm": 0.8268809914588928,
|
| 241 |
+
"learning_rate": 0.0003,
|
| 242 |
+
"loss": 0.3507,
|
| 243 |
+
"step": 2700
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 5.293005671077505,
|
| 247 |
+
"grad_norm": 0.7695613503456116,
|
| 248 |
+
"learning_rate": 0.0003,
|
| 249 |
+
"loss": 0.2829,
|
| 250 |
+
"step": 2800
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 5.482041587901701,
|
| 254 |
+
"grad_norm": 0.8355693221092224,
|
| 255 |
+
"learning_rate": 0.0003,
|
| 256 |
+
"loss": 0.2987,
|
| 257 |
+
"step": 2900
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 5.671077504725898,
|
| 261 |
+
"grad_norm": 0.8978683948516846,
|
| 262 |
+
"learning_rate": 0.0003,
|
| 263 |
+
"loss": 0.3112,
|
| 264 |
+
"step": 3000
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 5.8601134215500945,
|
| 268 |
+
"grad_norm": 0.9739505648612976,
|
| 269 |
+
"learning_rate": 0.0003,
|
| 270 |
+
"loss": 0.3201,
|
| 271 |
+
"step": 3100
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 6.0,
|
| 275 |
+
"eval_accuracy": 0.7822780269058296,
|
| 276 |
+
"eval_loss": 0.4750530421733856,
|
| 277 |
+
"eval_runtime": 6.4801,
|
| 278 |
+
"eval_samples_per_second": 77.159,
|
| 279 |
+
"eval_steps_per_second": 9.722,
|
| 280 |
+
"step": 3174
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 6.049149338374291,
|
| 284 |
+
"grad_norm": 0.8119204044342041,
|
| 285 |
+
"learning_rate": 0.0003,
|
| 286 |
+
"loss": 0.3033,
|
| 287 |
+
"step": 3200
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 6.238185255198488,
|
| 291 |
+
"grad_norm": 0.7731099128723145,
|
| 292 |
+
"learning_rate": 0.0003,
|
| 293 |
+
"loss": 0.2191,
|
| 294 |
+
"step": 3300
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 6.427221172022684,
|
| 298 |
+
"grad_norm": 0.7242079377174377,
|
| 299 |
+
"learning_rate": 0.0003,
|
| 300 |
+
"loss": 0.2313,
|
| 301 |
+
"step": 3400
|
| 302 |
+
},
|
| 303 |
+
{
|
| 304 |
+
"epoch": 6.616257088846881,
|
| 305 |
+
"grad_norm": 0.8276740908622742,
|
| 306 |
+
"learning_rate": 0.0003,
|
| 307 |
+
"loss": 0.2433,
|
| 308 |
+
"step": 3500
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 6.805293005671078,
|
| 312 |
+
"grad_norm": 0.8737579584121704,
|
| 313 |
+
"learning_rate": 0.0003,
|
| 314 |
+
"loss": 0.2545,
|
| 315 |
+
"step": 3600
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 6.994328922495274,
|
| 319 |
+
"grad_norm": 0.8738126158714294,
|
| 320 |
+
"learning_rate": 0.0003,
|
| 321 |
+
"loss": 0.2652,
|
| 322 |
+
"step": 3700
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"epoch": 7.0,
|
| 326 |
+
"eval_accuracy": 0.7886816143497758,
|
| 327 |
+
"eval_loss": 0.4510054588317871,
|
| 328 |
+
"eval_runtime": 7.2639,
|
| 329 |
+
"eval_samples_per_second": 68.834,
|
| 330 |
+
"eval_steps_per_second": 8.673,
|
| 331 |
+
"step": 3703
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 7.183364839319471,
|
| 335 |
+
"grad_norm": 0.8802294731140137,
|
| 336 |
+
"learning_rate": 0.0003,
|
| 337 |
+
"loss": 0.1865,
|
| 338 |
+
"step": 3800
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 7.3724007561436675,
|
| 342 |
+
"grad_norm": 0.740033745765686,
|
| 343 |
+
"learning_rate": 0.0003,
|
| 344 |
+
"loss": 0.1971,
|
| 345 |
+
"step": 3900
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 7.561436672967864,
|
| 349 |
+
"grad_norm": 0.8324846029281616,
|
| 350 |
+
"learning_rate": 0.0003,
|
| 351 |
+
"loss": 0.2043,
|
| 352 |
+
"step": 4000
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 7.750472589792061,
|
| 356 |
+
"grad_norm": 0.8003520369529724,
|
| 357 |
+
"learning_rate": 0.0003,
|
| 358 |
+
"loss": 0.2136,
|
| 359 |
+
"step": 4100
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 7.939508506616257,
|
| 363 |
+
"grad_norm": 0.9331985116004944,
|
| 364 |
+
"learning_rate": 0.0003,
|
| 365 |
+
"loss": 0.2263,
|
| 366 |
+
"step": 4200
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 8.0,
|
| 370 |
+
"eval_accuracy": 0.791372197309417,
|
| 371 |
+
"eval_loss": 0.4372054934501648,
|
| 372 |
+
"eval_runtime": 6.6756,
|
| 373 |
+
"eval_samples_per_second": 74.9,
|
| 374 |
+
"eval_steps_per_second": 9.437,
|
| 375 |
+
"step": 4232
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 8.128544423440454,
|
| 379 |
+
"grad_norm": 0.668910801410675,
|
| 380 |
+
"learning_rate": 0.0003,
|
| 381 |
+
"loss": 0.1888,
|
| 382 |
+
"step": 4300
|
| 383 |
+
},
|
| 384 |
+
{
|
| 385 |
+
"epoch": 8.31758034026465,
|
| 386 |
+
"grad_norm": 0.7265163660049438,
|
| 387 |
+
"learning_rate": 0.0003,
|
| 388 |
+
"loss": 0.172,
|
| 389 |
+
"step": 4400
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 8.506616257088847,
|
| 393 |
+
"grad_norm": 0.6801394820213318,
|
| 394 |
+
"learning_rate": 0.0003,
|
| 395 |
+
"loss": 0.1861,
|
| 396 |
+
"step": 4500
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 8.695652173913043,
|
| 400 |
+
"grad_norm": 0.709992527961731,
|
| 401 |
+
"learning_rate": 0.0003,
|
| 402 |
+
"loss": 0.1923,
|
| 403 |
+
"step": 4600
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"epoch": 8.88468809073724,
|
| 407 |
+
"grad_norm": 0.7939429879188538,
|
| 408 |
+
"learning_rate": 0.0003,
|
| 409 |
+
"loss": 0.2035,
|
| 410 |
+
"step": 4700
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 9.0,
|
| 414 |
+
"eval_accuracy": 0.7940269058295965,
|
| 415 |
+
"eval_loss": 0.43346819281578064,
|
| 416 |
+
"eval_runtime": 6.6609,
|
| 417 |
+
"eval_samples_per_second": 75.064,
|
| 418 |
+
"eval_steps_per_second": 9.458,
|
| 419 |
+
"step": 4761
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 9.073724007561436,
|
| 423 |
+
"grad_norm": 0.5764950513839722,
|
| 424 |
+
"learning_rate": 0.0003,
|
| 425 |
+
"loss": 0.1902,
|
| 426 |
+
"step": 4800
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 9.262759924385634,
|
| 430 |
+
"grad_norm": 0.7292913198471069,
|
| 431 |
+
"learning_rate": 0.0003,
|
| 432 |
+
"loss": 0.162,
|
| 433 |
+
"step": 4900
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 9.45179584120983,
|
| 437 |
+
"grad_norm": 0.8819996118545532,
|
| 438 |
+
"learning_rate": 0.0003,
|
| 439 |
+
"loss": 0.168,
|
| 440 |
+
"step": 5000
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 9.640831758034027,
|
| 444 |
+
"grad_norm": 0.8124964237213135,
|
| 445 |
+
"learning_rate": 0.0003,
|
| 446 |
+
"loss": 0.1814,
|
| 447 |
+
"step": 5100
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 9.829867674858223,
|
| 451 |
+
"grad_norm": 0.9616438150405884,
|
| 452 |
+
"learning_rate": 0.0003,
|
| 453 |
+
"loss": 0.1913,
|
| 454 |
+
"step": 5200
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"epoch": 10.0,
|
| 458 |
+
"eval_accuracy": 0.7949955156950672,
|
| 459 |
+
"eval_loss": 0.43220242857933044,
|
| 460 |
+
"eval_runtime": 6.518,
|
| 461 |
+
"eval_samples_per_second": 76.71,
|
| 462 |
+
"eval_steps_per_second": 9.666,
|
| 463 |
+
"step": 5290
|
| 464 |
+
},
|
| 465 |
+
{
|
| 466 |
+
"epoch": 10.01890359168242,
|
| 467 |
+
"grad_norm": 0.693879246711731,
|
| 468 |
+
"learning_rate": 0.0003,
|
| 469 |
+
"loss": 0.1943,
|
| 470 |
+
"step": 5300
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 10.207939508506616,
|
| 474 |
+
"grad_norm": 0.7181903719902039,
|
| 475 |
+
"learning_rate": 0.0003,
|
| 476 |
+
"loss": 0.1527,
|
| 477 |
+
"step": 5400
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 10.396975425330814,
|
| 481 |
+
"grad_norm": 0.7881558537483215,
|
| 482 |
+
"learning_rate": 0.0003,
|
| 483 |
+
"loss": 0.1599,
|
| 484 |
+
"step": 5500
|
| 485 |
+
},
|
| 486 |
+
{
|
| 487 |
+
"epoch": 10.58601134215501,
|
| 488 |
+
"grad_norm": 0.8658528327941895,
|
| 489 |
+
"learning_rate": 0.0003,
|
| 490 |
+
"loss": 0.1683,
|
| 491 |
+
"step": 5600
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 10.775047258979207,
|
| 495 |
+
"grad_norm": 0.8280355334281921,
|
| 496 |
+
"learning_rate": 0.0003,
|
| 497 |
+
"loss": 0.1794,
|
| 498 |
+
"step": 5700
|
| 499 |
+
},
|
| 500 |
+
{
|
| 501 |
+
"epoch": 10.964083175803403,
|
| 502 |
+
"grad_norm": 0.8619106411933899,
|
| 503 |
+
"learning_rate": 0.0003,
|
| 504 |
+
"loss": 0.188,
|
| 505 |
+
"step": 5800
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 11.0,
|
| 509 |
+
"eval_accuracy": 0.7945470852017937,
|
| 510 |
+
"eval_loss": 0.43794524669647217,
|
| 511 |
+
"eval_runtime": 6.1793,
|
| 512 |
+
"eval_samples_per_second": 80.915,
|
| 513 |
+
"eval_steps_per_second": 10.195,
|
| 514 |
+
"step": 5819
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 11.1531190926276,
|
| 518 |
+
"grad_norm": 0.6532451510429382,
|
| 519 |
+
"learning_rate": 0.0003,
|
| 520 |
+
"loss": 0.1553,
|
| 521 |
+
"step": 5900
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 11.342155009451796,
|
| 525 |
+
"grad_norm": 0.6573035717010498,
|
| 526 |
+
"learning_rate": 0.0003,
|
| 527 |
+
"loss": 0.152,
|
| 528 |
+
"step": 6000
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 11.531190926275993,
|
| 532 |
+
"grad_norm": 0.7298913598060608,
|
| 533 |
+
"learning_rate": 0.0003,
|
| 534 |
+
"loss": 0.1611,
|
| 535 |
+
"step": 6100
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 11.720226843100189,
|
| 539 |
+
"grad_norm": 0.7905253767967224,
|
| 540 |
+
"learning_rate": 0.0003,
|
| 541 |
+
"loss": 0.1699,
|
| 542 |
+
"step": 6200
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 11.909262759924385,
|
| 546 |
+
"grad_norm": 0.5873724222183228,
|
| 547 |
+
"learning_rate": 0.0003,
|
| 548 |
+
"loss": 0.1777,
|
| 549 |
+
"step": 6300
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 12.0,
|
| 553 |
+
"eval_accuracy": 0.795695067264574,
|
| 554 |
+
"eval_loss": 0.4279196858406067,
|
| 555 |
+
"eval_runtime": 7.2474,
|
| 556 |
+
"eval_samples_per_second": 68.99,
|
| 557 |
+
"eval_steps_per_second": 8.693,
|
| 558 |
+
"step": 6348
|
| 559 |
+
},
|
| 560 |
+
{
|
| 561 |
+
"epoch": 12.098298676748582,
|
| 562 |
+
"grad_norm": 0.6252820491790771,
|
| 563 |
+
"learning_rate": 0.0003,
|
| 564 |
+
"loss": 0.1608,
|
| 565 |
+
"step": 6400
|
| 566 |
+
},
|
| 567 |
+
{
|
| 568 |
+
"epoch": 12.287334593572778,
|
| 569 |
+
"grad_norm": 0.7915599346160889,
|
| 570 |
+
"learning_rate": 0.0003,
|
| 571 |
+
"loss": 0.1467,
|
| 572 |
+
"step": 6500
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 12.476370510396976,
|
| 576 |
+
"grad_norm": 0.7576895356178284,
|
| 577 |
+
"learning_rate": 0.0003,
|
| 578 |
+
"loss": 0.1539,
|
| 579 |
+
"step": 6600
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 12.665406427221171,
|
| 583 |
+
"grad_norm": 0.7499473690986633,
|
| 584 |
+
"learning_rate": 0.0003,
|
| 585 |
+
"loss": 0.1648,
|
| 586 |
+
"step": 6700
|
| 587 |
+
},
|
| 588 |
+
{
|
| 589 |
+
"epoch": 12.854442344045369,
|
| 590 |
+
"grad_norm": 0.8016745448112488,
|
| 591 |
+
"learning_rate": 0.0003,
|
| 592 |
+
"loss": 0.1723,
|
| 593 |
+
"step": 6800
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 13.0,
|
| 597 |
+
"eval_accuracy": 0.7956322869955157,
|
| 598 |
+
"eval_loss": 0.43264538049697876,
|
| 599 |
+
"eval_runtime": 6.6858,
|
| 600 |
+
"eval_samples_per_second": 74.785,
|
| 601 |
+
"eval_steps_per_second": 9.423,
|
| 602 |
+
"step": 6877
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 13.043478260869565,
|
| 606 |
+
"grad_norm": 0.6766420006752014,
|
| 607 |
+
"learning_rate": 0.0003,
|
| 608 |
+
"loss": 0.1715,
|
| 609 |
+
"step": 6900
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 13.232514177693762,
|
| 613 |
+
"grad_norm": 0.6546654105186462,
|
| 614 |
+
"learning_rate": 0.0003,
|
| 615 |
+
"loss": 0.1461,
|
| 616 |
+
"step": 7000
|
| 617 |
+
},
|
| 618 |
+
{
|
| 619 |
+
"epoch": 13.421550094517958,
|
| 620 |
+
"grad_norm": 0.7697082757949829,
|
| 621 |
+
"learning_rate": 0.0003,
|
| 622 |
+
"loss": 0.1498,
|
| 623 |
+
"step": 7100
|
| 624 |
+
},
|
| 625 |
+
{
|
| 626 |
+
"epoch": 13.610586011342155,
|
| 627 |
+
"grad_norm": 0.6504223346710205,
|
| 628 |
+
"learning_rate": 0.0003,
|
| 629 |
+
"loss": 0.1589,
|
| 630 |
+
"step": 7200
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 13.799621928166351,
|
| 634 |
+
"grad_norm": 0.834709882736206,
|
| 635 |
+
"learning_rate": 0.0003,
|
| 636 |
+
"loss": 0.1664,
|
| 637 |
+
"step": 7300
|
| 638 |
+
},
|
| 639 |
+
{
|
| 640 |
+
"epoch": 13.988657844990549,
|
| 641 |
+
"grad_norm": 0.7842487692832947,
|
| 642 |
+
"learning_rate": 0.0003,
|
| 643 |
+
"loss": 0.1767,
|
| 644 |
+
"step": 7400
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 14.0,
|
| 648 |
+
"eval_accuracy": 0.7966636771300448,
|
| 649 |
+
"eval_loss": 0.43290552496910095,
|
| 650 |
+
"eval_runtime": 6.5452,
|
| 651 |
+
"eval_samples_per_second": 76.392,
|
| 652 |
+
"eval_steps_per_second": 9.625,
|
| 653 |
+
"step": 7406
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 14.177693761814744,
|
| 657 |
+
"grad_norm": 0.762667715549469,
|
| 658 |
+
"learning_rate": 0.0003,
|
| 659 |
+
"loss": 0.1438,
|
| 660 |
+
"step": 7500
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 14.366729678638942,
|
| 664 |
+
"grad_norm": 0.6443786025047302,
|
| 665 |
+
"learning_rate": 0.0003,
|
| 666 |
+
"loss": 0.1467,
|
| 667 |
+
"step": 7600
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 14.555765595463138,
|
| 671 |
+
"grad_norm": 0.7382057309150696,
|
| 672 |
+
"learning_rate": 0.0003,
|
| 673 |
+
"loss": 0.1525,
|
| 674 |
+
"step": 7700
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 14.744801512287335,
|
| 678 |
+
"grad_norm": 0.8144531846046448,
|
| 679 |
+
"learning_rate": 0.0003,
|
| 680 |
+
"loss": 0.1611,
|
| 681 |
+
"step": 7800
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 14.93383742911153,
|
| 685 |
+
"grad_norm": 0.7852641940116882,
|
| 686 |
+
"learning_rate": 0.0003,
|
| 687 |
+
"loss": 0.1666,
|
| 688 |
+
"step": 7900
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 15.0,
|
| 692 |
+
"eval_accuracy": 0.7961793721973094,
|
| 693 |
+
"eval_loss": 0.43958863615989685,
|
| 694 |
+
"eval_runtime": 6.6743,
|
| 695 |
+
"eval_samples_per_second": 74.914,
|
| 696 |
+
"eval_steps_per_second": 9.439,
|
| 697 |
+
"step": 7935
|
| 698 |
+
},
|
| 699 |
+
{
|
| 700 |
+
"epoch": 15.122873345935728,
|
| 701 |
+
"grad_norm": 0.5839329361915588,
|
| 702 |
+
"learning_rate": 0.0003,
|
| 703 |
+
"loss": 0.1497,
|
| 704 |
+
"step": 8000
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 15.311909262759924,
|
| 708 |
+
"grad_norm": 0.6967843174934387,
|
| 709 |
+
"learning_rate": 0.0003,
|
| 710 |
+
"loss": 0.1413,
|
| 711 |
+
"step": 8100
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 15.500945179584122,
|
| 715 |
+
"grad_norm": 0.7435296177864075,
|
| 716 |
+
"learning_rate": 0.0003,
|
| 717 |
+
"loss": 0.145,
|
| 718 |
+
"step": 8200
|
| 719 |
+
},
|
| 720 |
+
{
|
| 721 |
+
"epoch": 15.689981096408317,
|
| 722 |
+
"grad_norm": 0.7955066561698914,
|
| 723 |
+
"learning_rate": 0.0003,
|
| 724 |
+
"loss": 0.1559,
|
| 725 |
+
"step": 8300
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"epoch": 15.879017013232515,
|
| 729 |
+
"grad_norm": 0.6731191873550415,
|
| 730 |
+
"learning_rate": 0.0003,
|
| 731 |
+
"loss": 0.1642,
|
| 732 |
+
"step": 8400
|
| 733 |
+
},
|
| 734 |
+
{
|
| 735 |
+
"epoch": 16.0,
|
| 736 |
+
"eval_accuracy": 0.7965022421524663,
|
| 737 |
+
"eval_loss": 0.43914830684661865,
|
| 738 |
+
"eval_runtime": 6.6825,
|
| 739 |
+
"eval_samples_per_second": 74.822,
|
| 740 |
+
"eval_steps_per_second": 9.428,
|
| 741 |
+
"step": 8464
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 16.068052930056712,
|
| 745 |
+
"grad_norm": 0.616921603679657,
|
| 746 |
+
"learning_rate": 0.0003,
|
| 747 |
+
"loss": 0.1555,
|
| 748 |
+
"step": 8500
|
| 749 |
+
},
|
| 750 |
+
{
|
| 751 |
+
"epoch": 16.257088846880908,
|
| 752 |
+
"grad_norm": 0.6410200595855713,
|
| 753 |
+
"learning_rate": 0.0003,
|
| 754 |
+
"loss": 0.1379,
|
| 755 |
+
"step": 8600
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"epoch": 16.446124763705104,
|
| 759 |
+
"grad_norm": 0.7206813097000122,
|
| 760 |
+
"learning_rate": 0.0003,
|
| 761 |
+
"loss": 0.1433,
|
| 762 |
+
"step": 8700
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 16.6351606805293,
|
| 766 |
+
"grad_norm": 1.1165937185287476,
|
| 767 |
+
"learning_rate": 0.0003,
|
| 768 |
+
"loss": 0.1526,
|
| 769 |
+
"step": 8800
|
| 770 |
+
},
|
| 771 |
+
{
|
| 772 |
+
"epoch": 16.8241965973535,
|
| 773 |
+
"grad_norm": 0.7585830688476562,
|
| 774 |
+
"learning_rate": 0.0003,
|
| 775 |
+
"loss": 0.1575,
|
| 776 |
+
"step": 8900
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 17.0,
|
| 780 |
+
"eval_accuracy": 0.7966816143497758,
|
| 781 |
+
"eval_loss": 0.44049155712127686,
|
| 782 |
+
"eval_runtime": 6.6339,
|
| 783 |
+
"eval_samples_per_second": 75.37,
|
| 784 |
+
"eval_steps_per_second": 9.497,
|
| 785 |
+
"step": 8993
|
| 786 |
+
},
|
| 787 |
+
{
|
| 788 |
+
"epoch": 17.013232514177695,
|
| 789 |
+
"grad_norm": 0.655029833316803,
|
| 790 |
+
"learning_rate": 0.0003,
|
| 791 |
+
"loss": 0.1624,
|
| 792 |
+
"step": 9000
|
| 793 |
+
},
|
| 794 |
+
{
|
| 795 |
+
"epoch": 17.20226843100189,
|
| 796 |
+
"grad_norm": 0.7475863099098206,
|
| 797 |
+
"learning_rate": 0.0003,
|
| 798 |
+
"loss": 0.1356,
|
| 799 |
+
"step": 9100
|
| 800 |
+
},
|
| 801 |
+
{
|
| 802 |
+
"epoch": 17.391304347826086,
|
| 803 |
+
"grad_norm": 0.6971881985664368,
|
| 804 |
+
"learning_rate": 0.0003,
|
| 805 |
+
"loss": 0.1406,
|
| 806 |
+
"step": 9200
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 17.58034026465028,
|
| 810 |
+
"grad_norm": 0.6662493944168091,
|
| 811 |
+
"learning_rate": 0.0003,
|
| 812 |
+
"loss": 0.1472,
|
| 813 |
+
"step": 9300
|
| 814 |
+
},
|
| 815 |
+
{
|
| 816 |
+
"epoch": 17.76937618147448,
|
| 817 |
+
"grad_norm": 0.7289919257164001,
|
| 818 |
+
"learning_rate": 0.0003,
|
| 819 |
+
"loss": 0.1578,
|
| 820 |
+
"step": 9400
|
| 821 |
+
},
|
| 822 |
+
{
|
| 823 |
+
"epoch": 17.958412098298677,
|
| 824 |
+
"grad_norm": 0.7780668139457703,
|
| 825 |
+
"learning_rate": 0.0003,
|
| 826 |
+
"loss": 0.1634,
|
| 827 |
+
"step": 9500
|
| 828 |
+
},
|
| 829 |
+
{
|
| 830 |
+
"epoch": 18.0,
|
| 831 |
+
"eval_accuracy": 0.7976233183856503,
|
| 832 |
+
"eval_loss": 0.42652928829193115,
|
| 833 |
+
"eval_runtime": 6.7415,
|
| 834 |
+
"eval_samples_per_second": 74.167,
|
| 835 |
+
"eval_steps_per_second": 9.345,
|
| 836 |
+
"step": 9522
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 18.147448015122873,
|
| 840 |
+
"grad_norm": 0.5619010925292969,
|
| 841 |
+
"learning_rate": 0.0003,
|
| 842 |
+
"loss": 0.1401,
|
| 843 |
+
"step": 9600
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 18.33648393194707,
|
| 847 |
+
"grad_norm": 0.7481386065483093,
|
| 848 |
+
"learning_rate": 0.0003,
|
| 849 |
+
"loss": 0.1378,
|
| 850 |
+
"step": 9700
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 18.525519848771268,
|
| 854 |
+
"grad_norm": 0.6953665614128113,
|
| 855 |
+
"learning_rate": 0.0003,
|
| 856 |
+
"loss": 0.144,
|
| 857 |
+
"step": 9800
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 18.714555765595463,
|
| 861 |
+
"grad_norm": 0.8559867143630981,
|
| 862 |
+
"learning_rate": 0.0003,
|
| 863 |
+
"loss": 0.1521,
|
| 864 |
+
"step": 9900
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 18.90359168241966,
|
| 868 |
+
"grad_norm": 0.7916247844696045,
|
| 869 |
+
"learning_rate": 0.0003,
|
| 870 |
+
"loss": 0.1593,
|
| 871 |
+
"step": 10000
|
| 872 |
+
},
|
| 873 |
+
{
|
| 874 |
+
"epoch": 19.0,
|
| 875 |
+
"eval_accuracy": 0.7978116591928252,
|
| 876 |
+
"eval_loss": 0.43227851390838623,
|
| 877 |
+
"eval_runtime": 6.4486,
|
| 878 |
+
"eval_samples_per_second": 77.537,
|
| 879 |
+
"eval_steps_per_second": 9.77,
|
| 880 |
+
"step": 10051
|
| 881 |
+
},
|
| 882 |
+
{
|
| 883 |
+
"epoch": 19.092627599243855,
|
| 884 |
+
"grad_norm": 0.638944685459137,
|
| 885 |
+
"learning_rate": 0.0003,
|
| 886 |
+
"loss": 0.1468,
|
| 887 |
+
"step": 10100
|
| 888 |
+
},
|
| 889 |
+
{
|
| 890 |
+
"epoch": 19.281663516068054,
|
| 891 |
+
"grad_norm": 0.7771846652030945,
|
| 892 |
+
"learning_rate": 0.0003,
|
| 893 |
+
"loss": 0.1345,
|
| 894 |
+
"step": 10200
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"epoch": 19.47069943289225,
|
| 898 |
+
"grad_norm": 0.5761379599571228,
|
| 899 |
+
"learning_rate": 0.0003,
|
| 900 |
+
"loss": 0.1404,
|
| 901 |
+
"step": 10300
|
| 902 |
+
},
|
| 903 |
+
{
|
| 904 |
+
"epoch": 19.659735349716446,
|
| 905 |
+
"grad_norm": 0.6479610204696655,
|
| 906 |
+
"learning_rate": 0.0003,
|
| 907 |
+
"loss": 0.1489,
|
| 908 |
+
"step": 10400
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 19.84877126654064,
|
| 912 |
+
"grad_norm": 0.6894132494926453,
|
| 913 |
+
"learning_rate": 0.0003,
|
| 914 |
+
"loss": 0.153,
|
| 915 |
+
"step": 10500
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 20.0,
|
| 919 |
+
"eval_accuracy": 0.7981434977578475,
|
| 920 |
+
"eval_loss": 0.43110987544059753,
|
| 921 |
+
"eval_runtime": 6.5758,
|
| 922 |
+
"eval_samples_per_second": 76.036,
|
| 923 |
+
"eval_steps_per_second": 9.581,
|
| 924 |
+
"step": 10580
|
| 925 |
+
},
|
| 926 |
+
{
|
| 927 |
+
"epoch": 20.0,
|
| 928 |
+
"step": 10580,
|
| 929 |
+
"total_flos": 9.64242245391745e+17,
|
| 930 |
+
"train_loss": 0.3966417311720677,
|
| 931 |
+
"train_runtime": 24179.107,
|
| 932 |
+
"train_samples_per_second": 14.0,
|
| 933 |
+
"train_steps_per_second": 0.438
|
| 934 |
+
}
|
| 935 |
+
],
|
| 936 |
+
"logging_steps": 100,
|
| 937 |
+
"max_steps": 10580,
|
| 938 |
+
"num_input_tokens_seen": 0,
|
| 939 |
+
"num_train_epochs": 20,
|
| 940 |
+
"save_steps": 500,
|
| 941 |
+
"total_flos": 9.64242245391745e+17,
|
| 942 |
+
"train_batch_size": 1,
|
| 943 |
+
"trial_name": null,
|
| 944 |
+
"trial_params": null
|
| 945 |
+
}
|