Model save
Browse files- README.md +2 -4
- all_results.json +4 -4
- train_results.json +4 -4
- trainer_state.json +1909 -9
README.md
CHANGED
|
@@ -1,19 +1,17 @@
|
|
| 1 |
---
|
| 2 |
base_model: deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B
|
| 3 |
-
datasets: knoveleng/open-rs
|
| 4 |
library_name: transformers
|
| 5 |
model_name: DAPO
|
| 6 |
tags:
|
| 7 |
- generated_from_trainer
|
| 8 |
-
- open-r1
|
| 9 |
-
- dapo
|
| 10 |
- trl
|
|
|
|
| 11 |
licence: license
|
| 12 |
---
|
| 13 |
|
| 14 |
# Model Card for DAPO
|
| 15 |
|
| 16 |
-
This model is a fine-tuned version of [deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B)
|
| 17 |
It has been trained using [TRL](https://github.com/huggingface/trl).
|
| 18 |
|
| 19 |
## Quick start
|
|
|
|
| 1 |
---
|
| 2 |
base_model: deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B
|
|
|
|
| 3 |
library_name: transformers
|
| 4 |
model_name: DAPO
|
| 5 |
tags:
|
| 6 |
- generated_from_trainer
|
|
|
|
|
|
|
| 7 |
- trl
|
| 8 |
+
- dapo
|
| 9 |
licence: license
|
| 10 |
---
|
| 11 |
|
| 12 |
# Model Card for DAPO
|
| 13 |
|
| 14 |
+
This model is a fine-tuned version of [deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B).
|
| 15 |
It has been trained using [TRL](https://github.com/huggingface/trl).
|
| 16 |
|
| 17 |
## Quick start
|
all_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
"total_flos": 0.0,
|
| 3 |
-
"train_loss": 0.
|
| 4 |
-
"train_runtime":
|
| 5 |
"train_samples": 7000,
|
| 6 |
-
"train_samples_per_second": 0.
|
| 7 |
-
"train_steps_per_second": 0.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"total_flos": 0.0,
|
| 3 |
+
"train_loss": 0.02940896774176508,
|
| 4 |
+
"train_runtime": 83918.4654,
|
| 5 |
"train_samples": 7000,
|
| 6 |
+
"train_samples_per_second": 0.114,
|
| 7 |
+
"train_steps_per_second": 0.002
|
| 8 |
}
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
"total_flos": 0.0,
|
| 3 |
-
"train_loss": 0.
|
| 4 |
-
"train_runtime":
|
| 5 |
"train_samples": 7000,
|
| 6 |
-
"train_samples_per_second": 0.
|
| 7 |
-
"train_steps_per_second": 0.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"total_flos": 0.0,
|
| 3 |
+
"train_loss": 0.02940896774176508,
|
| 4 |
+
"train_runtime": 83918.4654,
|
| 5 |
"train_samples": 7000,
|
| 6 |
+
"train_samples_per_second": 0.114,
|
| 7 |
+
"train_steps_per_second": 0.002
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -2,9 +2,9 @@
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
-
"epoch": 0.
|
| 6 |
"eval_steps": 500,
|
| 7 |
-
"global_step":
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
@@ -1910,17 +1910,1917 @@
|
|
| 1910 |
"step": 100
|
| 1911 |
},
|
| 1912 |
{
|
| 1913 |
-
"
|
| 1914 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1915 |
"total_flos": 0.0,
|
| 1916 |
-
"train_loss": 0.
|
| 1917 |
-
"train_runtime":
|
| 1918 |
-
"train_samples_per_second": 0.
|
| 1919 |
-
"train_steps_per_second": 0.
|
| 1920 |
}
|
| 1921 |
],
|
| 1922 |
"logging_steps": 1,
|
| 1923 |
-
"max_steps":
|
| 1924 |
"num_input_tokens_seen": 0,
|
| 1925 |
"num_train_epochs": 1,
|
| 1926 |
"save_steps": 10,
|
|
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.22857142857142856,
|
| 6 |
"eval_steps": 500,
|
| 7 |
+
"global_step": 200,
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
|
|
| 1910 |
"step": 100
|
| 1911 |
},
|
| 1912 |
{
|
| 1913 |
+
"clip_fraction": 0.0,
|
| 1914 |
+
"completion_length": 3318.513916015625,
|
| 1915 |
+
"dapo/avg_reward_std": 0.22042016812733242,
|
| 1916 |
+
"dapo/filter_reward_index": 0.0,
|
| 1917 |
+
"dapo/kept_prompts_ratio": 0.29523810063089645,
|
| 1918 |
+
"dapo/num_sampling_attempts": 4.375,
|
| 1919 |
+
"dapo/sampling_efficiency": 28.645833333333332,
|
| 1920 |
+
"dapo/total_prompts_processed": 26.25,
|
| 1921 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 1922 |
+
"epoch": 0.11542857142857142,
|
| 1923 |
+
"grad_norm": 0.22150926291942596,
|
| 1924 |
+
"kl": 0.011791229248046875,
|
| 1925 |
+
"learning_rate": 1e-07,
|
| 1926 |
+
"loss": 0.0631,
|
| 1927 |
+
"reward": 0.46524661034345627,
|
| 1928 |
+
"reward_std": 0.9665903598070145,
|
| 1929 |
+
"step": 101
|
| 1930 |
+
},
|
| 1931 |
+
{
|
| 1932 |
+
"clip_fraction": 0.0,
|
| 1933 |
+
"completion_length": 3083.875,
|
| 1934 |
+
"dapo/avg_reward_std": 0.21663353669232335,
|
| 1935 |
+
"dapo/filter_reward_index": 0.0,
|
| 1936 |
+
"dapo/kept_prompts_ratio": 0.3390804637095024,
|
| 1937 |
+
"dapo/num_sampling_attempts": 3.625,
|
| 1938 |
+
"dapo/sampling_efficiency": 39.93055555555555,
|
| 1939 |
+
"dapo/total_prompts_processed": 21.75,
|
| 1940 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 1941 |
+
"epoch": 0.11657142857142858,
|
| 1942 |
+
"grad_norm": 0.16289636492729187,
|
| 1943 |
+
"kl": 0.008695602416992188,
|
| 1944 |
+
"learning_rate": 6.203955092681039e-07,
|
| 1945 |
+
"loss": 0.098,
|
| 1946 |
+
"reward": 0.8642945289611816,
|
| 1947 |
+
"reward_std": 1.031830094754696,
|
| 1948 |
+
"step": 102
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"clip_fraction": 0.0,
|
| 1952 |
+
"completion_length": 3364.701446533203,
|
| 1953 |
+
"dapo/avg_reward_std": 0.24887267331923207,
|
| 1954 |
+
"dapo/filter_reward_index": 0.0,
|
| 1955 |
+
"dapo/kept_prompts_ratio": 0.3172043090866458,
|
| 1956 |
+
"dapo/num_sampling_attempts": 3.875,
|
| 1957 |
+
"dapo/sampling_efficiency": 31.69642857142857,
|
| 1958 |
+
"dapo/total_prompts_processed": 23.25,
|
| 1959 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 1960 |
+
"epoch": 0.11771428571428572,
|
| 1961 |
+
"grad_norm": 0.08825232833623886,
|
| 1962 |
+
"kl": 0.009820938110351562,
|
| 1963 |
+
"learning_rate": 6.126278954320294e-07,
|
| 1964 |
+
"loss": 0.0178,
|
| 1965 |
+
"reward": 0.3627179069444537,
|
| 1966 |
+
"reward_std": 0.8941863179206848,
|
| 1967 |
+
"step": 103
|
| 1968 |
+
},
|
| 1969 |
+
{
|
| 1970 |
+
"clip_fraction": 0.0,
|
| 1971 |
+
"completion_length": 3255.3055725097656,
|
| 1972 |
+
"dapo/avg_reward_std": 0.24808817549988074,
|
| 1973 |
+
"dapo/filter_reward_index": 0.0,
|
| 1974 |
+
"dapo/kept_prompts_ratio": 0.33950618074999916,
|
| 1975 |
+
"dapo/num_sampling_attempts": 3.375,
|
| 1976 |
+
"dapo/sampling_efficiency": 38.95833333333333,
|
| 1977 |
+
"dapo/total_prompts_processed": 20.25,
|
| 1978 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 1979 |
+
"epoch": 0.11885714285714286,
|
| 1980 |
+
"grad_norm": 0.13638561964035034,
|
| 1981 |
+
"kl": 0.011318206787109375,
|
| 1982 |
+
"learning_rate": 6.048412045323164e-07,
|
| 1983 |
+
"loss": 0.0643,
|
| 1984 |
+
"reward": 0.5508436523377895,
|
| 1985 |
+
"reward_std": 0.9409585371613503,
|
| 1986 |
+
"step": 104
|
| 1987 |
+
},
|
| 1988 |
+
{
|
| 1989 |
+
"clip_fraction": 0.0,
|
| 1990 |
+
"completion_length": 3270.4930419921875,
|
| 1991 |
+
"dapo/avg_reward_std": 0.23700118958950042,
|
| 1992 |
+
"dapo/filter_reward_index": 0.0,
|
| 1993 |
+
"dapo/kept_prompts_ratio": 0.3166666706403097,
|
| 1994 |
+
"dapo/num_sampling_attempts": 3.75,
|
| 1995 |
+
"dapo/sampling_efficiency": 61.07142857142857,
|
| 1996 |
+
"dapo/total_prompts_processed": 22.5,
|
| 1997 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 1998 |
+
"epoch": 0.12,
|
| 1999 |
+
"grad_norm": 0.10357476025819778,
|
| 2000 |
+
"kl": 0.0117034912109375,
|
| 2001 |
+
"learning_rate": 5.97037808470444e-07,
|
| 2002 |
+
"loss": 0.0278,
|
| 2003 |
+
"reward": 0.4137148158624768,
|
| 2004 |
+
"reward_std": 0.9205853268504143,
|
| 2005 |
+
"step": 105
|
| 2006 |
+
},
|
| 2007 |
+
{
|
| 2008 |
+
"clip_fraction": 0.0,
|
| 2009 |
+
"completion_length": 3118.9584045410156,
|
| 2010 |
+
"dapo/avg_reward_std": 0.22452521603554487,
|
| 2011 |
+
"dapo/filter_reward_index": 0.0,
|
| 2012 |
+
"dapo/kept_prompts_ratio": 0.3333333395421505,
|
| 2013 |
+
"dapo/num_sampling_attempts": 4.0,
|
| 2014 |
+
"dapo/sampling_efficiency": 28.869047619047613,
|
| 2015 |
+
"dapo/total_prompts_processed": 24.0,
|
| 2016 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2017 |
+
"epoch": 0.12114285714285715,
|
| 2018 |
+
"grad_norm": 0.11885393410921097,
|
| 2019 |
+
"kl": 0.011783599853515625,
|
| 2020 |
+
"learning_rate": 5.892200842364462e-07,
|
| 2021 |
+
"loss": 0.0786,
|
| 2022 |
+
"reward": 0.673494272865355,
|
| 2023 |
+
"reward_std": 0.9388571679592133,
|
| 2024 |
+
"step": 106
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"clip_fraction": 0.0,
|
| 2028 |
+
"completion_length": 3183.666717529297,
|
| 2029 |
+
"dapo/avg_reward_std": 0.23609773551716523,
|
| 2030 |
+
"dapo/filter_reward_index": 0.0,
|
| 2031 |
+
"dapo/kept_prompts_ratio": 0.30882353467099805,
|
| 2032 |
+
"dapo/num_sampling_attempts": 4.25,
|
| 2033 |
+
"dapo/sampling_efficiency": 37.74305555555556,
|
| 2034 |
+
"dapo/total_prompts_processed": 25.5,
|
| 2035 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2036 |
+
"epoch": 0.12228571428571429,
|
| 2037 |
+
"grad_norm": 0.13629400730133057,
|
| 2038 |
+
"kl": 0.0092010498046875,
|
| 2039 |
+
"learning_rate": 5.813904131848564e-07,
|
| 2040 |
+
"loss": 0.0615,
|
| 2041 |
+
"reward": 0.5680118557065725,
|
| 2042 |
+
"reward_std": 0.8982010260224342,
|
| 2043 |
+
"step": 107
|
| 2044 |
+
},
|
| 2045 |
+
{
|
| 2046 |
+
"clip_fraction": 0.0,
|
| 2047 |
+
"completion_length": 3170.263916015625,
|
| 2048 |
+
"dapo/avg_reward_std": 0.21017570431168014,
|
| 2049 |
+
"dapo/filter_reward_index": 0.0,
|
| 2050 |
+
"dapo/kept_prompts_ratio": 0.3018018079770578,
|
| 2051 |
+
"dapo/num_sampling_attempts": 4.625,
|
| 2052 |
+
"dapo/sampling_efficiency": 30.625,
|
| 2053 |
+
"dapo/total_prompts_processed": 27.75,
|
| 2054 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2055 |
+
"epoch": 0.12342857142857143,
|
| 2056 |
+
"grad_norm": 0.1134539544582367,
|
| 2057 |
+
"kl": 0.010692596435546875,
|
| 2058 |
+
"learning_rate": 5.735511803093248e-07,
|
| 2059 |
+
"loss": 0.0433,
|
| 2060 |
+
"reward": 0.6368884779512882,
|
| 2061 |
+
"reward_std": 0.9655679985880852,
|
| 2062 |
+
"step": 108
|
| 2063 |
+
},
|
| 2064 |
+
{
|
| 2065 |
+
"clip_fraction": 0.0,
|
| 2066 |
+
"completion_length": 2938.5243530273438,
|
| 2067 |
+
"dapo/avg_reward_std": 0.30796096875117374,
|
| 2068 |
+
"dapo/filter_reward_index": 0.0,
|
| 2069 |
+
"dapo/kept_prompts_ratio": 0.3974359052685591,
|
| 2070 |
+
"dapo/num_sampling_attempts": 3.25,
|
| 2071 |
+
"dapo/sampling_efficiency": 38.95833333333333,
|
| 2072 |
+
"dapo/total_prompts_processed": 19.5,
|
| 2073 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2074 |
+
"epoch": 0.12457142857142857,
|
| 2075 |
+
"grad_norm": 0.16064728796482086,
|
| 2076 |
+
"kl": 0.014812469482421875,
|
| 2077 |
+
"learning_rate": 5.657047735161255e-07,
|
| 2078 |
+
"loss": 0.0874,
|
| 2079 |
+
"reward": 0.4405923653393984,
|
| 2080 |
+
"reward_std": 0.899710550904274,
|
| 2081 |
+
"step": 109
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"clip_fraction": 0.0,
|
| 2085 |
+
"completion_length": 3333.5556030273438,
|
| 2086 |
+
"dapo/avg_reward_std": 0.17683410130698105,
|
| 2087 |
+
"dapo/filter_reward_index": 0.0,
|
| 2088 |
+
"dapo/kept_prompts_ratio": 0.28735632475080164,
|
| 2089 |
+
"dapo/num_sampling_attempts": 3.625,
|
| 2090 |
+
"dapo/sampling_efficiency": 40.104166666666664,
|
| 2091 |
+
"dapo/total_prompts_processed": 21.75,
|
| 2092 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2093 |
+
"epoch": 0.12571428571428572,
|
| 2094 |
+
"grad_norm": 0.1374766230583191,
|
| 2095 |
+
"kl": 0.00823211669921875,
|
| 2096 |
+
"learning_rate": 5.578535828967777e-07,
|
| 2097 |
+
"loss": 0.0525,
|
| 2098 |
+
"reward": 0.6373127717524767,
|
| 2099 |
+
"reward_std": 0.949370414018631,
|
| 2100 |
+
"step": 110
|
| 2101 |
+
},
|
| 2102 |
+
{
|
| 2103 |
+
"clip_fraction": 0.0,
|
| 2104 |
+
"completion_length": 3404.166717529297,
|
| 2105 |
+
"dapo/avg_reward_std": 0.2707539377734065,
|
| 2106 |
+
"dapo/filter_reward_index": 0.0,
|
| 2107 |
+
"dapo/kept_prompts_ratio": 0.3437500074505806,
|
| 2108 |
+
"dapo/num_sampling_attempts": 4.0,
|
| 2109 |
+
"dapo/sampling_efficiency": 28.124999999999996,
|
| 2110 |
+
"dapo/total_prompts_processed": 24.0,
|
| 2111 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2112 |
+
"epoch": 0.12685714285714286,
|
| 2113 |
+
"grad_norm": 0.09096160531044006,
|
| 2114 |
+
"kl": 0.0152435302734375,
|
| 2115 |
+
"learning_rate": 5.5e-07,
|
| 2116 |
+
"loss": 0.0286,
|
| 2117 |
+
"reward": 0.4166172882542014,
|
| 2118 |
+
"reward_std": 0.9417606145143509,
|
| 2119 |
+
"step": 111
|
| 2120 |
+
},
|
| 2121 |
+
{
|
| 2122 |
+
"clip_fraction": 0.0,
|
| 2123 |
+
"completion_length": 3306.263946533203,
|
| 2124 |
+
"dapo/avg_reward_std": 0.17227381931410896,
|
| 2125 |
+
"dapo/filter_reward_index": 0.0,
|
| 2126 |
+
"dapo/kept_prompts_ratio": 0.21481482055452134,
|
| 2127 |
+
"dapo/num_sampling_attempts": 5.625,
|
| 2128 |
+
"dapo/sampling_efficiency": 27.395833333333332,
|
| 2129 |
+
"dapo/total_prompts_processed": 33.75,
|
| 2130 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2131 |
+
"epoch": 0.128,
|
| 2132 |
+
"grad_norm": 0.11950567364692688,
|
| 2133 |
+
"kl": 0.01320648193359375,
|
| 2134 |
+
"learning_rate": 5.421464171032224e-07,
|
| 2135 |
+
"loss": 0.0449,
|
| 2136 |
+
"reward": 0.4937558462843299,
|
| 2137 |
+
"reward_std": 0.9720155894756317,
|
| 2138 |
+
"step": 112
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"clip_fraction": 0.0,
|
| 2142 |
+
"completion_length": 3117.1979064941406,
|
| 2143 |
+
"dapo/avg_reward_std": 0.30339551545106447,
|
| 2144 |
+
"dapo/filter_reward_index": 0.0,
|
| 2145 |
+
"dapo/kept_prompts_ratio": 0.3846153886272357,
|
| 2146 |
+
"dapo/num_sampling_attempts": 3.25,
|
| 2147 |
+
"dapo/sampling_efficiency": 38.95833333333333,
|
| 2148 |
+
"dapo/total_prompts_processed": 19.5,
|
| 2149 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2150 |
+
"epoch": 0.12914285714285714,
|
| 2151 |
+
"grad_norm": 0.15823398530483246,
|
| 2152 |
+
"kl": 0.01418304443359375,
|
| 2153 |
+
"learning_rate": 5.342952264838747e-07,
|
| 2154 |
+
"loss": 0.0743,
|
| 2155 |
+
"reward": 0.5596551271155477,
|
| 2156 |
+
"reward_std": 0.8979872986674309,
|
| 2157 |
+
"step": 113
|
| 2158 |
+
},
|
| 2159 |
+
{
|
| 2160 |
+
"clip_fraction": 0.0,
|
| 2161 |
+
"completion_length": 3239.031280517578,
|
| 2162 |
+
"dapo/avg_reward_std": 0.24120492219924927,
|
| 2163 |
+
"dapo/filter_reward_index": 0.0,
|
| 2164 |
+
"dapo/kept_prompts_ratio": 0.34000000298023225,
|
| 2165 |
+
"dapo/num_sampling_attempts": 3.125,
|
| 2166 |
+
"dapo/sampling_efficiency": 56.770833333333336,
|
| 2167 |
+
"dapo/total_prompts_processed": 18.75,
|
| 2168 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2169 |
+
"epoch": 0.13028571428571428,
|
| 2170 |
+
"grad_norm": 0.20106364786624908,
|
| 2171 |
+
"kl": 0.01206207275390625,
|
| 2172 |
+
"learning_rate": 5.264488196906752e-07,
|
| 2173 |
+
"loss": 0.0817,
|
| 2174 |
+
"reward": 0.697497084736824,
|
| 2175 |
+
"reward_std": 0.9489930346608162,
|
| 2176 |
+
"step": 114
|
| 2177 |
+
},
|
| 2178 |
+
{
|
| 2179 |
+
"clip_fraction": 0.0,
|
| 2180 |
+
"completion_length": 3197.2430725097656,
|
| 2181 |
+
"dapo/avg_reward_std": 0.20663932577157632,
|
| 2182 |
+
"dapo/filter_reward_index": 0.0,
|
| 2183 |
+
"dapo/kept_prompts_ratio": 0.26495727056112045,
|
| 2184 |
+
"dapo/num_sampling_attempts": 4.875,
|
| 2185 |
+
"dapo/sampling_efficiency": 38.4375,
|
| 2186 |
+
"dapo/total_prompts_processed": 29.25,
|
| 2187 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2188 |
+
"epoch": 0.13142857142857142,
|
| 2189 |
+
"grad_norm": 0.15399962663650513,
|
| 2190 |
+
"kl": 0.015567779541015625,
|
| 2191 |
+
"learning_rate": 5.186095868151436e-07,
|
| 2192 |
+
"loss": 0.0667,
|
| 2193 |
+
"reward": 0.5802914081141353,
|
| 2194 |
+
"reward_std": 0.9295158162713051,
|
| 2195 |
+
"step": 115
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"clip_fraction": 0.0,
|
| 2199 |
+
"completion_length": 3272.6007080078125,
|
| 2200 |
+
"dapo/avg_reward_std": 0.22710687816143035,
|
| 2201 |
+
"dapo/filter_reward_index": 0.0,
|
| 2202 |
+
"dapo/kept_prompts_ratio": 0.3166666701436043,
|
| 2203 |
+
"dapo/num_sampling_attempts": 3.75,
|
| 2204 |
+
"dapo/sampling_efficiency": 37.61904761904762,
|
| 2205 |
+
"dapo/total_prompts_processed": 22.5,
|
| 2206 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2207 |
+
"epoch": 0.13257142857142856,
|
| 2208 |
+
"grad_norm": 0.140142023563385,
|
| 2209 |
+
"kl": 0.01934814453125,
|
| 2210 |
+
"learning_rate": 5.107799157635538e-07,
|
| 2211 |
+
"loss": 0.0611,
|
| 2212 |
+
"reward": 0.6176847349852324,
|
| 2213 |
+
"reward_std": 0.944318100810051,
|
| 2214 |
+
"step": 116
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"clip_fraction": 0.0,
|
| 2218 |
+
"completion_length": 3268.4305725097656,
|
| 2219 |
+
"dapo/avg_reward_std": 0.23266587586238466,
|
| 2220 |
+
"dapo/filter_reward_index": 0.0,
|
| 2221 |
+
"dapo/kept_prompts_ratio": 0.344827591345228,
|
| 2222 |
+
"dapo/num_sampling_attempts": 3.625,
|
| 2223 |
+
"dapo/sampling_efficiency": 38.125,
|
| 2224 |
+
"dapo/total_prompts_processed": 21.75,
|
| 2225 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2226 |
+
"epoch": 0.1337142857142857,
|
| 2227 |
+
"grad_norm": 0.1582440286874771,
|
| 2228 |
+
"kl": 0.01198577880859375,
|
| 2229 |
+
"learning_rate": 5.02962191529556e-07,
|
| 2230 |
+
"loss": 0.0556,
|
| 2231 |
+
"reward": 0.5785031230188906,
|
| 2232 |
+
"reward_std": 0.954645112156868,
|
| 2233 |
+
"step": 117
|
| 2234 |
+
},
|
| 2235 |
+
{
|
| 2236 |
+
"clip_fraction": 0.0,
|
| 2237 |
+
"completion_length": 2941.9722595214844,
|
| 2238 |
+
"dapo/avg_reward_std": 0.24969401342027328,
|
| 2239 |
+
"dapo/filter_reward_index": 0.0,
|
| 2240 |
+
"dapo/kept_prompts_ratio": 0.3284313814604984,
|
| 2241 |
+
"dapo/num_sampling_attempts": 4.25,
|
| 2242 |
+
"dapo/sampling_efficiency": 27.20238095238095,
|
| 2243 |
+
"dapo/total_prompts_processed": 25.5,
|
| 2244 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2245 |
+
"epoch": 0.13485714285714287,
|
| 2246 |
+
"grad_norm": 0.1869765818119049,
|
| 2247 |
+
"kl": 0.01676177978515625,
|
| 2248 |
+
"learning_rate": 4.951587954676837e-07,
|
| 2249 |
+
"loss": 0.1063,
|
| 2250 |
+
"reward": 0.6486848145723343,
|
| 2251 |
+
"reward_std": 0.9332743212580681,
|
| 2252 |
+
"step": 118
|
| 2253 |
+
},
|
| 2254 |
+
{
|
| 2255 |
+
"clip_fraction": 0.0,
|
| 2256 |
+
"completion_length": 3206.982635498047,
|
| 2257 |
+
"dapo/avg_reward_std": 0.20580977627209254,
|
| 2258 |
+
"dapo/filter_reward_index": 0.0,
|
| 2259 |
+
"dapo/kept_prompts_ratio": 0.26666667333671024,
|
| 2260 |
+
"dapo/num_sampling_attempts": 4.375,
|
| 2261 |
+
"dapo/sampling_efficiency": 41.28472222222222,
|
| 2262 |
+
"dapo/total_prompts_processed": 26.25,
|
| 2263 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2264 |
+
"epoch": 0.136,
|
| 2265 |
+
"grad_norm": 0.13004696369171143,
|
| 2266 |
+
"kl": 0.015842437744140625,
|
| 2267 |
+
"learning_rate": 4.873721045679706e-07,
|
| 2268 |
+
"loss": 0.0453,
|
| 2269 |
+
"reward": 0.4798949249088764,
|
| 2270 |
+
"reward_std": 0.9390313774347305,
|
| 2271 |
+
"step": 119
|
| 2272 |
+
},
|
| 2273 |
+
{
|
| 2274 |
+
"clip_fraction": 0.0,
|
| 2275 |
+
"completion_length": 3015.545135498047,
|
| 2276 |
+
"dapo/avg_reward_std": 0.22217401381461852,
|
| 2277 |
+
"dapo/filter_reward_index": 0.0,
|
| 2278 |
+
"dapo/kept_prompts_ratio": 0.3548387149649282,
|
| 2279 |
+
"dapo/num_sampling_attempts": 3.875,
|
| 2280 |
+
"dapo/sampling_efficiency": 28.95833333333333,
|
| 2281 |
+
"dapo/total_prompts_processed": 23.25,
|
| 2282 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2283 |
+
"epoch": 0.13714285714285715,
|
| 2284 |
+
"grad_norm": 0.229897141456604,
|
| 2285 |
+
"kl": 0.02198028564453125,
|
| 2286 |
+
"learning_rate": 4.79604490731896e-07,
|
| 2287 |
+
"loss": 0.0749,
|
| 2288 |
+
"reward": 0.7311479561030865,
|
| 2289 |
+
"reward_std": 0.9607837572693825,
|
| 2290 |
+
"step": 120
|
| 2291 |
+
},
|
| 2292 |
+
{
|
| 2293 |
+
"clip_fraction": 0.0,
|
| 2294 |
+
"completion_length": 3098.656280517578,
|
| 2295 |
+
"dapo/avg_reward_std": 0.22588159143924713,
|
| 2296 |
+
"dapo/filter_reward_index": 0.0,
|
| 2297 |
+
"dapo/kept_prompts_ratio": 0.32777778506278993,
|
| 2298 |
+
"dapo/num_sampling_attempts": 3.75,
|
| 2299 |
+
"dapo/sampling_efficiency": 44.613095238095234,
|
| 2300 |
+
"dapo/total_prompts_processed": 22.5,
|
| 2301 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2302 |
+
"epoch": 0.1382857142857143,
|
| 2303 |
+
"grad_norm": 0.13800247013568878,
|
| 2304 |
+
"kl": 0.014202117919921875,
|
| 2305 |
+
"learning_rate": 4.7185832004988133e-07,
|
| 2306 |
+
"loss": 0.0814,
|
| 2307 |
+
"reward": 0.8461479842662811,
|
| 2308 |
+
"reward_std": 0.9660850539803505,
|
| 2309 |
+
"step": 121
|
| 2310 |
+
},
|
| 2311 |
+
{
|
| 2312 |
+
"clip_fraction": 0.0,
|
| 2313 |
+
"completion_length": 3064.3924255371094,
|
| 2314 |
+
"dapo/avg_reward_std": 0.16500467896461488,
|
| 2315 |
+
"dapo/filter_reward_index": 0.0,
|
| 2316 |
+
"dapo/kept_prompts_ratio": 0.19666667193174361,
|
| 2317 |
+
"dapo/num_sampling_attempts": 6.25,
|
| 2318 |
+
"dapo/sampling_efficiency": 21.07142857142857,
|
| 2319 |
+
"dapo/total_prompts_processed": 37.5,
|
| 2320 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2321 |
+
"epoch": 0.13942857142857143,
|
| 2322 |
+
"grad_norm": 0.1680934727191925,
|
| 2323 |
+
"kl": 0.01361083984375,
|
| 2324 |
+
"learning_rate": 4.641359520805548e-07,
|
| 2325 |
+
"loss": 0.066,
|
| 2326 |
+
"reward": 0.7812346797436476,
|
| 2327 |
+
"reward_std": 0.9529108256101608,
|
| 2328 |
+
"step": 122
|
| 2329 |
+
},
|
| 2330 |
+
{
|
| 2331 |
+
"clip_fraction": 0.0,
|
| 2332 |
+
"completion_length": 3097.4861755371094,
|
| 2333 |
+
"dapo/avg_reward_std": 0.22939075000824466,
|
| 2334 |
+
"dapo/filter_reward_index": 0.0,
|
| 2335 |
+
"dapo/kept_prompts_ratio": 0.33333334038334506,
|
| 2336 |
+
"dapo/num_sampling_attempts": 3.875,
|
| 2337 |
+
"dapo/sampling_efficiency": 33.75,
|
| 2338 |
+
"dapo/total_prompts_processed": 23.25,
|
| 2339 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2340 |
+
"epoch": 0.14057142857142857,
|
| 2341 |
+
"grad_norm": 0.18081900477409363,
|
| 2342 |
+
"kl": 0.014842987060546875,
|
| 2343 |
+
"learning_rate": 4.5643973913200837e-07,
|
| 2344 |
+
"loss": 0.0877,
|
| 2345 |
+
"reward": 0.7531900368630886,
|
| 2346 |
+
"reward_std": 0.9868133068084717,
|
| 2347 |
+
"step": 123
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"clip_fraction": 0.0,
|
| 2351 |
+
"completion_length": 3203.888885498047,
|
| 2352 |
+
"dapo/avg_reward_std": 0.24352495979379724,
|
| 2353 |
+
"dapo/filter_reward_index": 0.0,
|
| 2354 |
+
"dapo/kept_prompts_ratio": 0.35185185737080044,
|
| 2355 |
+
"dapo/num_sampling_attempts": 3.375,
|
| 2356 |
+
"dapo/sampling_efficiency": 43.05555555555556,
|
| 2357 |
+
"dapo/total_prompts_processed": 20.25,
|
| 2358 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2359 |
+
"epoch": 0.1417142857142857,
|
| 2360 |
+
"grad_norm": 0.16807734966278076,
|
| 2361 |
+
"kl": 0.0139007568359375,
|
| 2362 |
+
"learning_rate": 4.4877202554526084e-07,
|
| 2363 |
+
"loss": 0.0612,
|
| 2364 |
+
"reward": 0.715996683575213,
|
| 2365 |
+
"reward_std": 0.9595553278923035,
|
| 2366 |
+
"step": 124
|
| 2367 |
+
},
|
| 2368 |
+
{
|
| 2369 |
+
"clip_fraction": 0.0,
|
| 2370 |
+
"completion_length": 2885.5625610351562,
|
| 2371 |
+
"dapo/avg_reward_std": 0.2548297820612788,
|
| 2372 |
+
"dapo/filter_reward_index": 0.0,
|
| 2373 |
+
"dapo/kept_prompts_ratio": 0.31770833814516664,
|
| 2374 |
+
"dapo/num_sampling_attempts": 4.0,
|
| 2375 |
+
"dapo/sampling_efficiency": 27.20238095238095,
|
| 2376 |
+
"dapo/total_prompts_processed": 24.0,
|
| 2377 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2378 |
+
"epoch": 0.14285714285714285,
|
| 2379 |
+
"grad_norm": 0.16355834901332855,
|
| 2380 |
+
"kl": 0.02027130126953125,
|
| 2381 |
+
"learning_rate": 4.4113514698014953e-07,
|
| 2382 |
+
"loss": 0.0597,
|
| 2383 |
+
"reward": 0.8311022147536278,
|
| 2384 |
+
"reward_std": 0.9600836709141731,
|
| 2385 |
+
"step": 125
|
| 2386 |
+
},
|
| 2387 |
+
{
|
| 2388 |
+
"clip_fraction": 0.0,
|
| 2389 |
+
"completion_length": 3250.843780517578,
|
| 2390 |
+
"dapo/avg_reward_std": 0.2203440727858708,
|
| 2391 |
+
"dapo/filter_reward_index": 0.0,
|
| 2392 |
+
"dapo/kept_prompts_ratio": 0.32758621152105005,
|
| 2393 |
+
"dapo/num_sampling_attempts": 3.625,
|
| 2394 |
+
"dapo/sampling_efficiency": 46.770833333333336,
|
| 2395 |
+
"dapo/total_prompts_processed": 21.75,
|
| 2396 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2397 |
+
"epoch": 0.144,
|
| 2398 |
+
"grad_norm": 0.18190248310565948,
|
| 2399 |
+
"kl": 0.0158843994140625,
|
| 2400 |
+
"learning_rate": 4.3353142970386557e-07,
|
| 2401 |
+
"loss": 0.068,
|
| 2402 |
+
"reward": 0.7400151332840323,
|
| 2403 |
+
"reward_std": 0.9569809287786484,
|
| 2404 |
+
"step": 126
|
| 2405 |
+
},
|
| 2406 |
+
{
|
| 2407 |
+
"clip_fraction": 0.0,
|
| 2408 |
+
"completion_length": 3264.420166015625,
|
| 2409 |
+
"dapo/avg_reward_std": 0.25137073759521755,
|
| 2410 |
+
"dapo/filter_reward_index": 0.0,
|
| 2411 |
+
"dapo/kept_prompts_ratio": 0.41666667429464205,
|
| 2412 |
+
"dapo/num_sampling_attempts": 3.5,
|
| 2413 |
+
"dapo/sampling_efficiency": 40.11904761904761,
|
| 2414 |
+
"dapo/total_prompts_processed": 21.0,
|
| 2415 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2416 |
+
"epoch": 0.14514285714285713,
|
| 2417 |
+
"grad_norm": 0.17950685322284698,
|
| 2418 |
+
"kl": 0.0223236083984375,
|
| 2419 |
+
"learning_rate": 4.2596318988235037e-07,
|
| 2420 |
+
"loss": 0.0528,
|
| 2421 |
+
"reward": 0.5194851458072662,
|
| 2422 |
+
"reward_std": 0.9414050430059433,
|
| 2423 |
+
"step": 127
|
| 2424 |
+
},
|
| 2425 |
+
{
|
| 2426 |
+
"clip_fraction": 0.0,
|
| 2427 |
+
"completion_length": 2892.9132690429688,
|
| 2428 |
+
"dapo/avg_reward_std": 0.2416491061449051,
|
| 2429 |
+
"dapo/filter_reward_index": 0.0,
|
| 2430 |
+
"dapo/kept_prompts_ratio": 0.2631579002267436,
|
| 2431 |
+
"dapo/num_sampling_attempts": 4.75,
|
| 2432 |
+
"dapo/sampling_efficiency": 26.9047619047619,
|
| 2433 |
+
"dapo/total_prompts_processed": 28.5,
|
| 2434 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2435 |
+
"epoch": 0.1462857142857143,
|
| 2436 |
+
"grad_norm": 0.25602471828460693,
|
| 2437 |
+
"kl": 0.02016448974609375,
|
| 2438 |
+
"learning_rate": 4.1843273287476854e-07,
|
| 2439 |
+
"loss": 0.0933,
|
| 2440 |
+
"reward": 0.8592288717627525,
|
| 2441 |
+
"reward_std": 0.9212958365678787,
|
| 2442 |
+
"step": 128
|
| 2443 |
+
},
|
| 2444 |
+
{
|
| 2445 |
+
"clip_fraction": 0.0,
|
| 2446 |
+
"completion_length": 3146.6944580078125,
|
| 2447 |
+
"dapo/avg_reward_std": 0.22558308675371366,
|
| 2448 |
+
"dapo/filter_reward_index": 0.0,
|
| 2449 |
+
"dapo/kept_prompts_ratio": 0.3218390854268238,
|
| 2450 |
+
"dapo/num_sampling_attempts": 3.625,
|
| 2451 |
+
"dapo/sampling_efficiency": 54.07738095238095,
|
| 2452 |
+
"dapo/total_prompts_processed": 21.75,
|
| 2453 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2454 |
+
"epoch": 0.14742857142857144,
|
| 2455 |
+
"grad_norm": 0.21352027356624603,
|
| 2456 |
+
"kl": 0.0198211669921875,
|
| 2457 |
+
"learning_rate": 4.1094235253127374e-07,
|
| 2458 |
+
"loss": 0.0679,
|
| 2459 |
+
"reward": 0.5732525363564491,
|
| 2460 |
+
"reward_std": 0.9645283669233322,
|
| 2461 |
+
"step": 129
|
| 2462 |
+
},
|
| 2463 |
+
{
|
| 2464 |
+
"clip_fraction": 0.0,
|
| 2465 |
+
"completion_length": 3248.4236450195312,
|
| 2466 |
+
"dapo/avg_reward_std": 0.35807471639580196,
|
| 2467 |
+
"dapo/filter_reward_index": 0.0,
|
| 2468 |
+
"dapo/kept_prompts_ratio": 0.5000000066227384,
|
| 2469 |
+
"dapo/num_sampling_attempts": 2.25,
|
| 2470 |
+
"dapo/sampling_efficiency": 51.041666666666664,
|
| 2471 |
+
"dapo/total_prompts_processed": 13.5,
|
| 2472 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2473 |
+
"epoch": 0.14857142857142858,
|
| 2474 |
+
"grad_norm": 0.1599435657262802,
|
| 2475 |
+
"kl": 0.0216827392578125,
|
| 2476 |
+
"learning_rate": 4.034943304942796e-07,
|
| 2477 |
+
"loss": 0.0443,
|
| 2478 |
+
"reward": 0.5955070666968822,
|
| 2479 |
+
"reward_std": 0.9924386888742447,
|
| 2480 |
+
"step": 130
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"clip_fraction": 0.0,
|
| 2484 |
+
"completion_length": 2958.5347595214844,
|
| 2485 |
+
"dapo/avg_reward_std": 0.18185590389298228,
|
| 2486 |
+
"dapo/filter_reward_index": 0.0,
|
| 2487 |
+
"dapo/kept_prompts_ratio": 0.23170731998071437,
|
| 2488 |
+
"dapo/num_sampling_attempts": 5.125,
|
| 2489 |
+
"dapo/sampling_efficiency": 24.945436507936506,
|
| 2490 |
+
"dapo/total_prompts_processed": 30.75,
|
| 2491 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2492 |
+
"epoch": 0.14971428571428572,
|
| 2493 |
+
"grad_norm": 0.21188445389270782,
|
| 2494 |
+
"kl": 0.02074432373046875,
|
| 2495 |
+
"learning_rate": 3.9609093550344907e-07,
|
| 2496 |
+
"loss": 0.0628,
|
| 2497 |
+
"reward": 0.8608505353331566,
|
| 2498 |
+
"reward_std": 0.9059992283582687,
|
| 2499 |
+
"step": 131
|
| 2500 |
+
},
|
| 2501 |
+
{
|
| 2502 |
+
"clip_fraction": 0.0,
|
| 2503 |
+
"completion_length": 3019.888931274414,
|
| 2504 |
+
"dapo/avg_reward_std": 0.3038036392794715,
|
| 2505 |
+
"dapo/filter_reward_index": 0.0,
|
| 2506 |
+
"dapo/kept_prompts_ratio": 0.36419753785486575,
|
| 2507 |
+
"dapo/num_sampling_attempts": 3.375,
|
| 2508 |
+
"dapo/sampling_efficiency": 38.33333333333333,
|
| 2509 |
+
"dapo/total_prompts_processed": 20.25,
|
| 2510 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2511 |
+
"epoch": 0.15085714285714286,
|
| 2512 |
+
"grad_norm": 0.19752100110054016,
|
| 2513 |
+
"kl": 0.024078369140625,
|
| 2514 |
+
"learning_rate": 3.8873442270461485e-07,
|
| 2515 |
+
"loss": 0.0698,
|
| 2516 |
+
"reward": 0.7191393785178661,
|
| 2517 |
+
"reward_std": 0.9548436179757118,
|
| 2518 |
+
"step": 132
|
| 2519 |
+
},
|
| 2520 |
+
{
|
| 2521 |
+
"clip_fraction": 0.0,
|
| 2522 |
+
"completion_length": 3251.6909790039062,
|
| 2523 |
+
"dapo/avg_reward_std": 0.17617152915114448,
|
| 2524 |
+
"dapo/filter_reward_index": 0.0,
|
| 2525 |
+
"dapo/kept_prompts_ratio": 0.22222222494227545,
|
| 2526 |
+
"dapo/num_sampling_attempts": 5.25,
|
| 2527 |
+
"dapo/sampling_efficiency": 31.369047619047613,
|
| 2528 |
+
"dapo/total_prompts_processed": 31.5,
|
| 2529 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2530 |
+
"epoch": 0.152,
|
| 2531 |
+
"grad_norm": 0.1220565065741539,
|
| 2532 |
+
"kl": 0.01824951171875,
|
| 2533 |
+
"learning_rate": 3.8142703296283953e-07,
|
| 2534 |
+
"loss": 0.0249,
|
| 2535 |
+
"reward": 0.3546891317819245,
|
| 2536 |
+
"reward_std": 0.9377138167619705,
|
| 2537 |
+
"step": 133
|
| 2538 |
+
},
|
| 2539 |
+
{
|
| 2540 |
+
"clip_fraction": 0.0,
|
| 2541 |
+
"completion_length": 3146.545196533203,
|
| 2542 |
+
"dapo/avg_reward_std": 0.2565364229679108,
|
| 2543 |
+
"dapo/filter_reward_index": 0.0,
|
| 2544 |
+
"dapo/kept_prompts_ratio": 0.32666667103767394,
|
| 2545 |
+
"dapo/num_sampling_attempts": 3.125,
|
| 2546 |
+
"dapo/sampling_efficiency": 47.08333333333333,
|
| 2547 |
+
"dapo/total_prompts_processed": 18.75,
|
| 2548 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2549 |
+
"epoch": 0.15314285714285714,
|
| 2550 |
+
"grad_norm": 0.15810362994670868,
|
| 2551 |
+
"kl": 0.03081512451171875,
|
| 2552 |
+
"learning_rate": 3.7417099217982686e-07,
|
| 2553 |
+
"loss": 0.0306,
|
| 2554 |
+
"reward": 0.5206232005730271,
|
| 2555 |
+
"reward_std": 0.9619846642017365,
|
| 2556 |
+
"step": 134
|
| 2557 |
+
},
|
| 2558 |
+
{
|
| 2559 |
+
"clip_fraction": 0.0,
|
| 2560 |
+
"completion_length": 3085.5972900390625,
|
| 2561 |
+
"dapo/avg_reward_std": 0.30491976333515985,
|
| 2562 |
+
"dapo/filter_reward_index": 0.0,
|
| 2563 |
+
"dapo/kept_prompts_ratio": 0.40476191469601225,
|
| 2564 |
+
"dapo/num_sampling_attempts": 3.5,
|
| 2565 |
+
"dapo/sampling_efficiency": 31.666666666666664,
|
| 2566 |
+
"dapo/total_prompts_processed": 21.0,
|
| 2567 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2568 |
+
"epoch": 0.15428571428571428,
|
| 2569 |
+
"grad_norm": 0.2133372277021408,
|
| 2570 |
+
"kl": 0.0204620361328125,
|
| 2571 |
+
"learning_rate": 3.6696851061588994e-07,
|
| 2572 |
+
"loss": 0.0681,
|
| 2573 |
+
"reward": 0.7713347226381302,
|
| 2574 |
+
"reward_std": 0.9403144493699074,
|
| 2575 |
+
"step": 135
|
| 2576 |
+
},
|
| 2577 |
+
{
|
| 2578 |
+
"clip_fraction": 0.0,
|
| 2579 |
+
"completion_length": 3326.295196533203,
|
| 2580 |
+
"dapo/avg_reward_std": 0.22884555886953306,
|
| 2581 |
+
"dapo/filter_reward_index": 0.0,
|
| 2582 |
+
"dapo/kept_prompts_ratio": 0.24358974817471626,
|
| 2583 |
+
"dapo/num_sampling_attempts": 4.875,
|
| 2584 |
+
"dapo/sampling_efficiency": 25.868055555555557,
|
| 2585 |
+
"dapo/total_prompts_processed": 29.25,
|
| 2586 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2587 |
+
"epoch": 0.15542857142857142,
|
| 2588 |
+
"grad_norm": 0.18792302906513214,
|
| 2589 |
+
"kl": 0.029754638671875,
|
| 2590 |
+
"learning_rate": 3.5982178221668533e-07,
|
| 2591 |
+
"loss": 0.0468,
|
| 2592 |
+
"reward": 0.5651950668543577,
|
| 2593 |
+
"reward_std": 0.9934203922748566,
|
| 2594 |
+
"step": 136
|
| 2595 |
+
},
|
| 2596 |
+
{
|
| 2597 |
+
"clip_fraction": 0.0,
|
| 2598 |
+
"completion_length": 3265.2882080078125,
|
| 2599 |
+
"dapo/avg_reward_std": 0.304972759137551,
|
| 2600 |
+
"dapo/filter_reward_index": 0.0,
|
| 2601 |
+
"dapo/kept_prompts_ratio": 0.43055556155741215,
|
| 2602 |
+
"dapo/num_sampling_attempts": 3.0,
|
| 2603 |
+
"dapo/sampling_efficiency": 54.375,
|
| 2604 |
+
"dapo/total_prompts_processed": 18.0,
|
| 2605 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2606 |
+
"epoch": 0.15657142857142858,
|
| 2607 |
+
"grad_norm": 0.13081717491149902,
|
| 2608 |
+
"kl": 0.0223846435546875,
|
| 2609 |
+
"learning_rate": 3.5273298394491515e-07,
|
| 2610 |
+
"loss": 0.0443,
|
| 2611 |
+
"reward": 0.5535581167787313,
|
| 2612 |
+
"reward_std": 0.9467164501547813,
|
| 2613 |
+
"step": 137
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"clip_fraction": 0.0,
|
| 2617 |
+
"completion_length": 2895.8646545410156,
|
| 2618 |
+
"dapo/avg_reward_std": 0.2690910736719767,
|
| 2619 |
+
"dapo/filter_reward_index": 0.0,
|
| 2620 |
+
"dapo/kept_prompts_ratio": 0.3333333387970924,
|
| 2621 |
+
"dapo/num_sampling_attempts": 3.75,
|
| 2622 |
+
"dapo/sampling_efficiency": 32.82738095238095,
|
| 2623 |
+
"dapo/total_prompts_processed": 22.5,
|
| 2624 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2625 |
+
"epoch": 0.15771428571428572,
|
| 2626 |
+
"grad_norm": 0.18165208399295807,
|
| 2627 |
+
"kl": 0.032073974609375,
|
| 2628 |
+
"learning_rate": 3.45704275117204e-07,
|
| 2629 |
+
"loss": 0.0288,
|
| 2630 |
+
"reward": 0.5253790076822042,
|
| 2631 |
+
"reward_std": 0.9247673749923706,
|
| 2632 |
+
"step": 138
|
| 2633 |
+
},
|
| 2634 |
+
{
|
| 2635 |
+
"clip_fraction": 0.0,
|
| 2636 |
+
"completion_length": 3049.8507080078125,
|
| 2637 |
+
"dapo/avg_reward_std": 0.2440622321196965,
|
| 2638 |
+
"dapo/filter_reward_index": 0.0,
|
| 2639 |
+
"dapo/kept_prompts_ratio": 0.33928572067192625,
|
| 2640 |
+
"dapo/num_sampling_attempts": 3.5,
|
| 2641 |
+
"dapo/sampling_efficiency": 40.11904761904761,
|
| 2642 |
+
"dapo/total_prompts_processed": 21.0,
|
| 2643 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2644 |
+
"epoch": 0.15885714285714286,
|
| 2645 |
+
"grad_norm": 0.19676071405410767,
|
| 2646 |
+
"kl": 0.03052520751953125,
|
| 2647 |
+
"learning_rate": 3.387377967463493e-07,
|
| 2648 |
+
"loss": 0.0477,
|
| 2649 |
+
"reward": 0.6778539270162582,
|
| 2650 |
+
"reward_std": 0.9344745948910713,
|
| 2651 |
+
"step": 139
|
| 2652 |
+
},
|
| 2653 |
+
{
|
| 2654 |
+
"clip_fraction": 0.0,
|
| 2655 |
+
"completion_length": 3029.0486450195312,
|
| 2656 |
+
"dapo/avg_reward_std": 0.3111469969153404,
|
| 2657 |
+
"dapo/filter_reward_index": 0.0,
|
| 2658 |
+
"dapo/kept_prompts_ratio": 0.4916666768491268,
|
| 2659 |
+
"dapo/num_sampling_attempts": 2.5,
|
| 2660 |
+
"dapo/sampling_efficiency": 41.666666666666664,
|
| 2661 |
+
"dapo/total_prompts_processed": 15.0,
|
| 2662 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2663 |
+
"epoch": 0.16,
|
| 2664 |
+
"grad_norm": 0.18594416975975037,
|
| 2665 |
+
"kl": 0.0277557373046875,
|
| 2666 |
+
"learning_rate": 3.3183567088914833e-07,
|
| 2667 |
+
"loss": 0.0431,
|
| 2668 |
+
"reward": 0.5210836753249168,
|
| 2669 |
+
"reward_std": 0.9851464107632637,
|
| 2670 |
+
"step": 140
|
| 2671 |
+
},
|
| 2672 |
+
{
|
| 2673 |
+
"clip_fraction": 0.0,
|
| 2674 |
+
"completion_length": 3151.5486755371094,
|
| 2675 |
+
"dapo/avg_reward_std": 0.23511080997330802,
|
| 2676 |
+
"dapo/filter_reward_index": 0.0,
|
| 2677 |
+
"dapo/kept_prompts_ratio": 0.3095238127878734,
|
| 2678 |
+
"dapo/num_sampling_attempts": 4.375,
|
| 2679 |
+
"dapo/sampling_efficiency": 26.18055555555555,
|
| 2680 |
+
"dapo/total_prompts_processed": 26.25,
|
| 2681 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2682 |
+
"epoch": 0.16114285714285714,
|
| 2683 |
+
"grad_norm": 0.17807213962078094,
|
| 2684 |
+
"kl": 0.0266265869140625,
|
| 2685 |
+
"learning_rate": 3.250000000000001e-07,
|
| 2686 |
+
"loss": 0.0498,
|
| 2687 |
+
"reward": 0.5591800361871719,
|
| 2688 |
+
"reward_std": 0.9730060175061226,
|
| 2689 |
+
"step": 141
|
| 2690 |
+
},
|
| 2691 |
+
{
|
| 2692 |
+
"clip_fraction": 0.0,
|
| 2693 |
+
"completion_length": 2963.59033203125,
|
| 2694 |
+
"dapo/avg_reward_std": 0.19928012508898973,
|
| 2695 |
+
"dapo/filter_reward_index": 0.0,
|
| 2696 |
+
"dapo/kept_prompts_ratio": 0.2812500069849193,
|
| 2697 |
+
"dapo/num_sampling_attempts": 4.0,
|
| 2698 |
+
"dapo/sampling_efficiency": 38.02083333333333,
|
| 2699 |
+
"dapo/total_prompts_processed": 24.0,
|
| 2700 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2701 |
+
"epoch": 0.16228571428571428,
|
| 2702 |
+
"grad_norm": 0.24388359487056732,
|
| 2703 |
+
"kl": 0.0318603515625,
|
| 2704 |
+
"learning_rate": 3.182328662904756e-07,
|
| 2705 |
+
"loss": 0.0567,
|
| 2706 |
+
"reward": 0.7148469444364309,
|
| 2707 |
+
"reward_std": 0.9495278596878052,
|
| 2708 |
+
"step": 142
|
| 2709 |
+
},
|
| 2710 |
+
{
|
| 2711 |
+
"clip_fraction": 0.0,
|
| 2712 |
+
"completion_length": 3157.791717529297,
|
| 2713 |
+
"dapo/avg_reward_std": 0.23966079843895777,
|
| 2714 |
+
"dapo/filter_reward_index": 0.0,
|
| 2715 |
+
"dapo/kept_prompts_ratio": 0.3214285767504147,
|
| 2716 |
+
"dapo/num_sampling_attempts": 3.5,
|
| 2717 |
+
"dapo/sampling_efficiency": 39.166666666666664,
|
| 2718 |
+
"dapo/total_prompts_processed": 21.0,
|
| 2719 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2720 |
+
"epoch": 0.16342857142857142,
|
| 2721 |
+
"grad_norm": 0.20528583228588104,
|
| 2722 |
+
"kl": 0.041290283203125,
|
| 2723 |
+
"learning_rate": 3.115363310950578e-07,
|
| 2724 |
+
"loss": 0.0443,
|
| 2725 |
+
"reward": 0.5249591246247292,
|
| 2726 |
+
"reward_std": 0.9509934857487679,
|
| 2727 |
+
"step": 143
|
| 2728 |
+
},
|
| 2729 |
+
{
|
| 2730 |
+
"clip_fraction": 0.0,
|
| 2731 |
+
"completion_length": 3030.187530517578,
|
| 2732 |
+
"dapo/avg_reward_std": 0.30880050485332805,
|
| 2733 |
+
"dapo/filter_reward_index": 0.0,
|
| 2734 |
+
"dapo/kept_prompts_ratio": 0.4375000099341075,
|
| 2735 |
+
"dapo/num_sampling_attempts": 3.0,
|
| 2736 |
+
"dapo/sampling_efficiency": 41.04166666666666,
|
| 2737 |
+
"dapo/total_prompts_processed": 18.0,
|
| 2738 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2739 |
+
"epoch": 0.16457142857142856,
|
| 2740 |
+
"grad_norm": 0.15082307159900665,
|
| 2741 |
+
"kl": 0.02729034423828125,
|
| 2742 |
+
"learning_rate": 3.0491243424323783e-07,
|
| 2743 |
+
"loss": 0.0511,
|
| 2744 |
+
"reward": 0.5894143544137478,
|
| 2745 |
+
"reward_std": 0.954010546207428,
|
| 2746 |
+
"step": 144
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"clip_fraction": 0.0,
|
| 2750 |
+
"completion_length": 2973.3993225097656,
|
| 2751 |
+
"dapo/avg_reward_std": 0.32683228328824043,
|
| 2752 |
+
"dapo/filter_reward_index": 0.0,
|
| 2753 |
+
"dapo/kept_prompts_ratio": 0.4236111181477706,
|
| 2754 |
+
"dapo/num_sampling_attempts": 3.0,
|
| 2755 |
+
"dapo/sampling_efficiency": 48.66071428571428,
|
| 2756 |
+
"dapo/total_prompts_processed": 18.0,
|
| 2757 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2758 |
+
"epoch": 0.1657142857142857,
|
| 2759 |
+
"grad_norm": 0.2588576078414917,
|
| 2760 |
+
"kl": 0.038238525390625,
|
| 2761 |
+
"learning_rate": 2.9836319343816397e-07,
|
| 2762 |
+
"loss": 0.0611,
|
| 2763 |
+
"reward": 0.6702784113585949,
|
| 2764 |
+
"reward_std": 0.9678368121385574,
|
| 2765 |
+
"step": 145
|
| 2766 |
+
},
|
| 2767 |
+
{
|
| 2768 |
+
"clip_fraction": 0.0,
|
| 2769 |
+
"completion_length": 3289.8368530273438,
|
| 2770 |
+
"dapo/avg_reward_std": 0.29686578666722335,
|
| 2771 |
+
"dapo/filter_reward_index": 0.0,
|
| 2772 |
+
"dapo/kept_prompts_ratio": 0.34567901823255753,
|
| 2773 |
+
"dapo/num_sampling_attempts": 3.375,
|
| 2774 |
+
"dapo/sampling_efficiency": 51.57738095238095,
|
| 2775 |
+
"dapo/total_prompts_processed": 20.25,
|
| 2776 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2777 |
+
"epoch": 0.16685714285714287,
|
| 2778 |
+
"grad_norm": 0.2035798877477646,
|
| 2779 |
+
"kl": 0.0394744873046875,
|
| 2780 |
+
"learning_rate": 2.918906036420294e-07,
|
| 2781 |
+
"loss": 0.0576,
|
| 2782 |
+
"reward": 0.4602743685245514,
|
| 2783 |
+
"reward_std": 0.9194413796067238,
|
| 2784 |
+
"step": 146
|
| 2785 |
+
},
|
| 2786 |
+
{
|
| 2787 |
+
"clip_fraction": 0.0,
|
| 2788 |
+
"completion_length": 3068.7604064941406,
|
| 2789 |
+
"dapo/avg_reward_std": 0.27814541943371296,
|
| 2790 |
+
"dapo/filter_reward_index": 0.0,
|
| 2791 |
+
"dapo/kept_prompts_ratio": 0.3437500069849193,
|
| 2792 |
+
"dapo/num_sampling_attempts": 4.0,
|
| 2793 |
+
"dapo/sampling_efficiency": 36.666666666666664,
|
| 2794 |
+
"dapo/total_prompts_processed": 24.0,
|
| 2795 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2796 |
+
"epoch": 0.168,
|
| 2797 |
+
"grad_norm": 0.22469140589237213,
|
| 2798 |
+
"kl": 0.030426025390625,
|
| 2799 |
+
"learning_rate": 2.854966364683872e-07,
|
| 2800 |
+
"loss": 0.0696,
|
| 2801 |
+
"reward": 0.6243265215307474,
|
| 2802 |
+
"reward_std": 0.9174878597259521,
|
| 2803 |
+
"step": 147
|
| 2804 |
+
},
|
| 2805 |
+
{
|
| 2806 |
+
"clip_fraction": 0.0,
|
| 2807 |
+
"completion_length": 3041.357635498047,
|
| 2808 |
+
"dapo/avg_reward_std": 0.2907161459326744,
|
| 2809 |
+
"dapo/filter_reward_index": 0.0,
|
| 2810 |
+
"dapo/kept_prompts_ratio": 0.458333346247673,
|
| 2811 |
+
"dapo/num_sampling_attempts": 2.5,
|
| 2812 |
+
"dapo/sampling_efficiency": 57.70833333333333,
|
| 2813 |
+
"dapo/total_prompts_processed": 15.0,
|
| 2814 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2815 |
+
"epoch": 0.16914285714285715,
|
| 2816 |
+
"grad_norm": 0.3123789429664612,
|
| 2817 |
+
"kl": 0.0328521728515625,
|
| 2818 |
+
"learning_rate": 2.791832395815782e-07,
|
| 2819 |
+
"loss": 0.0819,
|
| 2820 |
+
"reward": 0.8250775411725044,
|
| 2821 |
+
"reward_std": 0.9233218431472778,
|
| 2822 |
+
"step": 148
|
| 2823 |
+
},
|
| 2824 |
+
{
|
| 2825 |
+
"clip_fraction": 0.0,
|
| 2826 |
+
"completion_length": 2433.0694732666016,
|
| 2827 |
+
"dapo/avg_reward_std": 0.22243764168686336,
|
| 2828 |
+
"dapo/filter_reward_index": 0.0,
|
| 2829 |
+
"dapo/kept_prompts_ratio": 0.2777777839865949,
|
| 2830 |
+
"dapo/num_sampling_attempts": 4.5,
|
| 2831 |
+
"dapo/sampling_efficiency": 35.75892857142857,
|
| 2832 |
+
"dapo/total_prompts_processed": 27.0,
|
| 2833 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2834 |
+
"epoch": 0.1702857142857143,
|
| 2835 |
+
"grad_norm": 0.2827485203742981,
|
| 2836 |
+
"kl": 0.0386505126953125,
|
| 2837 |
+
"learning_rate": 2.729523361034538e-07,
|
| 2838 |
+
"loss": 0.0784,
|
| 2839 |
+
"reward": 0.6995697831735015,
|
| 2840 |
+
"reward_std": 0.9434132054448128,
|
| 2841 |
+
"step": 149
|
| 2842 |
+
},
|
| 2843 |
+
{
|
| 2844 |
+
"clip_fraction": 0.0,
|
| 2845 |
+
"completion_length": 3096.59033203125,
|
| 2846 |
+
"dapo/avg_reward_std": 0.347408726811409,
|
| 2847 |
+
"dapo/filter_reward_index": 0.0,
|
| 2848 |
+
"dapo/kept_prompts_ratio": 0.541666672565043,
|
| 2849 |
+
"dapo/num_sampling_attempts": 2.0,
|
| 2850 |
+
"dapo/sampling_efficiency": 63.541666666666664,
|
| 2851 |
+
"dapo/total_prompts_processed": 12.0,
|
| 2852 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2853 |
+
"epoch": 0.17142857142857143,
|
| 2854 |
+
"grad_norm": 0.30529579520225525,
|
| 2855 |
+
"kl": 0.03045654296875,
|
| 2856 |
+
"learning_rate": 2.6680582402757324e-07,
|
| 2857 |
+
"loss": 0.0868,
|
| 2858 |
+
"reward": 0.7112221932038665,
|
| 2859 |
+
"reward_std": 0.9602288007736206,
|
| 2860 |
+
"step": 150
|
| 2861 |
+
},
|
| 2862 |
+
{
|
| 2863 |
+
"clip_fraction": 0.0,
|
| 2864 |
+
"completion_length": 3184.611083984375,
|
| 2865 |
+
"dapo/avg_reward_std": 0.1674806038115887,
|
| 2866 |
+
"dapo/filter_reward_index": 0.0,
|
| 2867 |
+
"dapo/kept_prompts_ratio": 0.20212766528129578,
|
| 2868 |
+
"dapo/num_sampling_attempts": 5.875,
|
| 2869 |
+
"dapo/sampling_efficiency": 23.749999999999996,
|
| 2870 |
+
"dapo/total_prompts_processed": 35.25,
|
| 2871 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2872 |
+
"epoch": 0.17257142857142857,
|
| 2873 |
+
"grad_norm": 0.19142813980579376,
|
| 2874 |
+
"kl": 0.037353515625,
|
| 2875 |
+
"learning_rate": 2.6074557564105724e-07,
|
| 2876 |
+
"loss": 0.045,
|
| 2877 |
+
"reward": 0.41017685225233436,
|
| 2878 |
+
"reward_std": 0.9152907580137253,
|
| 2879 |
+
"step": 151
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"clip_fraction": 0.0,
|
| 2883 |
+
"completion_length": 3437.3541564941406,
|
| 2884 |
+
"dapo/avg_reward_std": 0.208841644014631,
|
| 2885 |
+
"dapo/filter_reward_index": 0.0,
|
| 2886 |
+
"dapo/kept_prompts_ratio": 0.2571428622518267,
|
| 2887 |
+
"dapo/num_sampling_attempts": 4.375,
|
| 2888 |
+
"dapo/sampling_efficiency": 40.416666666666664,
|
| 2889 |
+
"dapo/total_prompts_processed": 26.25,
|
| 2890 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2891 |
+
"epoch": 0.1737142857142857,
|
| 2892 |
+
"grad_norm": 0.15321692824363708,
|
| 2893 |
+
"kl": 0.03997802734375,
|
| 2894 |
+
"learning_rate": 2.547734369542718e-07,
|
| 2895 |
+
"loss": 0.0346,
|
| 2896 |
+
"reward": 0.34562894329428673,
|
| 2897 |
+
"reward_std": 0.856454074382782,
|
| 2898 |
+
"step": 152
|
| 2899 |
+
},
|
| 2900 |
+
{
|
| 2901 |
+
"clip_fraction": 0.0,
|
| 2902 |
+
"completion_length": 3008.1285095214844,
|
| 2903 |
+
"dapo/avg_reward_std": 0.3009934023022652,
|
| 2904 |
+
"dapo/filter_reward_index": 0.0,
|
| 2905 |
+
"dapo/kept_prompts_ratio": 0.5000000096857548,
|
| 2906 |
+
"dapo/num_sampling_attempts": 2.5,
|
| 2907 |
+
"dapo/sampling_efficiency": 43.75,
|
| 2908 |
+
"dapo/total_prompts_processed": 15.0,
|
| 2909 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2910 |
+
"epoch": 0.17485714285714285,
|
| 2911 |
+
"grad_norm": 0.20332548022270203,
|
| 2912 |
+
"kl": 0.0509033203125,
|
| 2913 |
+
"learning_rate": 2.488912271385139e-07,
|
| 2914 |
+
"loss": 0.0536,
|
| 2915 |
+
"reward": 0.7641689777374268,
|
| 2916 |
+
"reward_std": 0.95648343116045,
|
| 2917 |
+
"step": 153
|
| 2918 |
+
},
|
| 2919 |
+
{
|
| 2920 |
+
"clip_fraction": 0.0,
|
| 2921 |
+
"completion_length": 3165.52783203125,
|
| 2922 |
+
"dapo/avg_reward_std": 0.2268627045246271,
|
| 2923 |
+
"dapo/filter_reward_index": 0.0,
|
| 2924 |
+
"dapo/kept_prompts_ratio": 0.35256410905948055,
|
| 2925 |
+
"dapo/num_sampling_attempts": 3.25,
|
| 2926 |
+
"dapo/sampling_efficiency": 40.625,
|
| 2927 |
+
"dapo/total_prompts_processed": 19.5,
|
| 2928 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2929 |
+
"epoch": 0.176,
|
| 2930 |
+
"grad_norm": 0.2415708601474762,
|
| 2931 |
+
"kl": 0.032623291015625,
|
| 2932 |
+
"learning_rate": 2.4310073797187573e-07,
|
| 2933 |
+
"loss": 0.0658,
|
| 2934 |
+
"reward": 0.6375892572104931,
|
| 2935 |
+
"reward_std": 0.9544621706008911,
|
| 2936 |
+
"step": 154
|
| 2937 |
+
},
|
| 2938 |
+
{
|
| 2939 |
+
"clip_fraction": 0.0,
|
| 2940 |
+
"completion_length": 3226.4652709960938,
|
| 2941 |
+
"dapo/avg_reward_std": 0.2563069482644399,
|
| 2942 |
+
"dapo/filter_reward_index": 0.0,
|
| 2943 |
+
"dapo/kept_prompts_ratio": 0.38333334078391396,
|
| 2944 |
+
"dapo/num_sampling_attempts": 3.75,
|
| 2945 |
+
"dapo/sampling_efficiency": 31.249999999999996,
|
| 2946 |
+
"dapo/total_prompts_processed": 22.5,
|
| 2947 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2948 |
+
"epoch": 0.17714285714285713,
|
| 2949 |
+
"grad_norm": 0.2137623131275177,
|
| 2950 |
+
"kl": 0.0427093505859375,
|
| 2951 |
+
"learning_rate": 2.374037332934512e-07,
|
| 2952 |
+
"loss": 0.0533,
|
| 2953 |
+
"reward": 0.537381574511528,
|
| 2954 |
+
"reward_std": 0.9281218275427818,
|
| 2955 |
+
"step": 155
|
| 2956 |
+
},
|
| 2957 |
+
{
|
| 2958 |
+
"clip_fraction": 0.0,
|
| 2959 |
+
"completion_length": 2680.3090209960938,
|
| 2960 |
+
"dapo/avg_reward_std": 0.22888225678241614,
|
| 2961 |
+
"dapo/filter_reward_index": 0.0,
|
| 2962 |
+
"dapo/kept_prompts_ratio": 0.3181818226973216,
|
| 2963 |
+
"dapo/num_sampling_attempts": 4.125,
|
| 2964 |
+
"dapo/sampling_efficiency": 31.29960317460317,
|
| 2965 |
+
"dapo/total_prompts_processed": 24.75,
|
| 2966 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2967 |
+
"epoch": 0.1782857142857143,
|
| 2968 |
+
"grad_norm": 0.3409210443496704,
|
| 2969 |
+
"kl": 0.03851318359375,
|
| 2970 |
+
"learning_rate": 2.3180194846605364e-07,
|
| 2971 |
+
"loss": 0.0962,
|
| 2972 |
+
"reward": 0.8820424377918243,
|
| 2973 |
+
"reward_std": 0.9246840327978134,
|
| 2974 |
+
"step": 156
|
| 2975 |
+
},
|
| 2976 |
+
{
|
| 2977 |
+
"clip_fraction": 0.0,
|
| 2978 |
+
"completion_length": 3045.3299255371094,
|
| 2979 |
+
"dapo/avg_reward_std": 0.2491180575810946,
|
| 2980 |
+
"dapo/filter_reward_index": 0.0,
|
| 2981 |
+
"dapo/kept_prompts_ratio": 0.3653846222620744,
|
| 2982 |
+
"dapo/num_sampling_attempts": 3.25,
|
| 2983 |
+
"dapo/sampling_efficiency": 45.83333333333332,
|
| 2984 |
+
"dapo/total_prompts_processed": 19.5,
|
| 2985 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2986 |
+
"epoch": 0.17942857142857144,
|
| 2987 |
+
"grad_norm": 0.23701035976409912,
|
| 2988 |
+
"kl": 0.0436248779296875,
|
| 2989 |
+
"learning_rate": 2.2629708984760706e-07,
|
| 2990 |
+
"loss": 0.0414,
|
| 2991 |
+
"reward": 0.6551959328353405,
|
| 2992 |
+
"reward_std": 0.9744707196950912,
|
| 2993 |
+
"step": 157
|
| 2994 |
+
},
|
| 2995 |
+
{
|
| 2996 |
+
"clip_fraction": 0.0,
|
| 2997 |
+
"completion_length": 2918.892364501953,
|
| 2998 |
+
"dapo/avg_reward_std": 0.22537656256130764,
|
| 2999 |
+
"dapo/filter_reward_index": 0.0,
|
| 3000 |
+
"dapo/kept_prompts_ratio": 0.33333333730697634,
|
| 3001 |
+
"dapo/num_sampling_attempts": 4.375,
|
| 3002 |
+
"dapo/sampling_efficiency": 39.93055555555556,
|
| 3003 |
+
"dapo/total_prompts_processed": 26.25,
|
| 3004 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3005 |
+
"epoch": 0.18057142857142858,
|
| 3006 |
+
"grad_norm": 0.3551786541938782,
|
| 3007 |
+
"kl": 0.0572357177734375,
|
| 3008 |
+
"learning_rate": 2.2089083427137329e-07,
|
| 3009 |
+
"loss": 0.0732,
|
| 3010 |
+
"reward": 0.5248121619224548,
|
| 3011 |
+
"reward_std": 0.9334831684827805,
|
| 3012 |
+
"step": 158
|
| 3013 |
+
},
|
| 3014 |
+
{
|
| 3015 |
+
"clip_fraction": 0.0,
|
| 3016 |
+
"completion_length": 2874.0729446411133,
|
| 3017 |
+
"dapo/avg_reward_std": 0.18832522351294756,
|
| 3018 |
+
"dapo/filter_reward_index": 0.0,
|
| 3019 |
+
"dapo/kept_prompts_ratio": 0.2812500046566129,
|
| 3020 |
+
"dapo/num_sampling_attempts": 4.0,
|
| 3021 |
+
"dapo/sampling_efficiency": 38.69047619047618,
|
| 3022 |
+
"dapo/total_prompts_processed": 24.0,
|
| 3023 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3024 |
+
"epoch": 0.18171428571428572,
|
| 3025 |
+
"grad_norm": 0.25500980019569397,
|
| 3026 |
+
"kl": 0.03741455078125,
|
| 3027 |
+
"learning_rate": 2.1558482853517253e-07,
|
| 3028 |
+
"loss": 0.0537,
|
| 3029 |
+
"reward": 0.7963100634515285,
|
| 3030 |
+
"reward_std": 0.987776905298233,
|
| 3031 |
+
"step": 159
|
| 3032 |
+
},
|
| 3033 |
+
{
|
| 3034 |
+
"clip_fraction": 0.0,
|
| 3035 |
+
"completion_length": 2940.701385498047,
|
| 3036 |
+
"dapo/avg_reward_std": 0.16297742784023284,
|
| 3037 |
+
"dapo/filter_reward_index": 0.0,
|
| 3038 |
+
"dapo/kept_prompts_ratio": 0.20000000536441803,
|
| 3039 |
+
"dapo/num_sampling_attempts": 6.25,
|
| 3040 |
+
"dapo/sampling_efficiency": 18.368055555555557,
|
| 3041 |
+
"dapo/total_prompts_processed": 37.5,
|
| 3042 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3043 |
+
"epoch": 0.18285714285714286,
|
| 3044 |
+
"grad_norm": 0.2898014187812805,
|
| 3045 |
+
"kl": 0.058013916015625,
|
| 3046 |
+
"learning_rate": 2.1038068889975259e-07,
|
| 3047 |
+
"loss": 0.037,
|
| 3048 |
+
"reward": 0.5323189618065953,
|
| 3049 |
+
"reward_std": 0.9483579620718956,
|
| 3050 |
+
"step": 160
|
| 3051 |
+
},
|
| 3052 |
+
{
|
| 3053 |
+
"clip_fraction": 0.0,
|
| 3054 |
+
"completion_length": 3090.7882385253906,
|
| 3055 |
+
"dapo/avg_reward_std": 0.3046227526664734,
|
| 3056 |
+
"dapo/filter_reward_index": 0.0,
|
| 3057 |
+
"dapo/kept_prompts_ratio": 0.3733333414793015,
|
| 3058 |
+
"dapo/num_sampling_attempts": 3.125,
|
| 3059 |
+
"dapo/sampling_efficiency": 43.45238095238095,
|
| 3060 |
+
"dapo/total_prompts_processed": 18.75,
|
| 3061 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3062 |
+
"epoch": 0.184,
|
| 3063 |
+
"grad_norm": 0.28573325276374817,
|
| 3064 |
+
"kl": 0.040771484375,
|
| 3065 |
+
"learning_rate": 2.0528000059645995e-07,
|
| 3066 |
+
"loss": 0.0511,
|
| 3067 |
+
"reward": 0.6970310118049383,
|
| 3068 |
+
"reward_std": 0.9432796016335487,
|
| 3069 |
+
"step": 161
|
| 3070 |
+
},
|
| 3071 |
+
{
|
| 3072 |
+
"clip_fraction": 0.0,
|
| 3073 |
+
"completion_length": 3205.4270629882812,
|
| 3074 |
+
"dapo/avg_reward_std": 0.36972329020500183,
|
| 3075 |
+
"dapo/filter_reward_index": 0.0,
|
| 3076 |
+
"dapo/kept_prompts_ratio": 0.5438596621940011,
|
| 3077 |
+
"dapo/num_sampling_attempts": 2.375,
|
| 3078 |
+
"dapo/sampling_efficiency": 55.625,
|
| 3079 |
+
"dapo/total_prompts_processed": 14.25,
|
| 3080 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3081 |
+
"epoch": 0.18514285714285714,
|
| 3082 |
+
"grad_norm": 0.390523225069046,
|
| 3083 |
+
"kl": 0.052459716796875,
|
| 3084 |
+
"learning_rate": 2.0028431734436308e-07,
|
| 3085 |
+
"loss": 0.0818,
|
| 3086 |
+
"reward": 0.6346883065998554,
|
| 3087 |
+
"reward_std": 0.9713371768593788,
|
| 3088 |
+
"step": 162
|
| 3089 |
+
},
|
| 3090 |
+
{
|
| 3091 |
+
"clip_fraction": 0.0,
|
| 3092 |
+
"completion_length": 3082.107635498047,
|
| 3093 |
+
"dapo/avg_reward_std": 0.2315557522158469,
|
| 3094 |
+
"dapo/filter_reward_index": 0.0,
|
| 3095 |
+
"dapo/kept_prompts_ratio": 0.3440860264724301,
|
| 3096 |
+
"dapo/num_sampling_attempts": 3.875,
|
| 3097 |
+
"dapo/sampling_efficiency": 44.513888888888886,
|
| 3098 |
+
"dapo/total_prompts_processed": 23.25,
|
| 3099 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3100 |
+
"epoch": 0.18628571428571428,
|
| 3101 |
+
"grad_norm": 0.31898149847984314,
|
| 3102 |
+
"kl": 0.05328369140625,
|
| 3103 |
+
"learning_rate": 1.9539516087697517e-07,
|
| 3104 |
+
"loss": 0.0722,
|
| 3105 |
+
"reward": 0.6942785531282425,
|
| 3106 |
+
"reward_std": 0.9776681512594223,
|
| 3107 |
+
"step": 163
|
| 3108 |
+
},
|
| 3109 |
+
{
|
| 3110 |
+
"clip_fraction": 0.0,
|
| 3111 |
+
"completion_length": 3027.0243530273438,
|
| 3112 |
+
"dapo/avg_reward_std": 0.15836979811255997,
|
| 3113 |
+
"dapo/filter_reward_index": 0.0,
|
| 3114 |
+
"dapo/kept_prompts_ratio": 0.22972973214613424,
|
| 3115 |
+
"dapo/num_sampling_attempts": 4.625,
|
| 3116 |
+
"dapo/sampling_efficiency": 41.69642857142857,
|
| 3117 |
+
"dapo/total_prompts_processed": 27.75,
|
| 3118 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3119 |
+
"epoch": 0.18742857142857142,
|
| 3120 |
+
"grad_norm": 0.2931766211986542,
|
| 3121 |
+
"kl": 0.033111572265625,
|
| 3122 |
+
"learning_rate": 1.9061402047871833e-07,
|
| 3123 |
+
"loss": 0.0754,
|
| 3124 |
+
"reward": 0.944303285330534,
|
| 3125 |
+
"reward_std": 0.9451126903295517,
|
| 3126 |
+
"step": 164
|
| 3127 |
+
},
|
| 3128 |
+
{
|
| 3129 |
+
"clip_fraction": 0.0,
|
| 3130 |
+
"completion_length": 2894.260482788086,
|
| 3131 |
+
"dapo/avg_reward_std": 0.224585828371346,
|
| 3132 |
+
"dapo/filter_reward_index": 0.0,
|
| 3133 |
+
"dapo/kept_prompts_ratio": 0.2916666716337204,
|
| 3134 |
+
"dapo/num_sampling_attempts": 4.0,
|
| 3135 |
+
"dapo/sampling_efficiency": 37.5,
|
| 3136 |
+
"dapo/total_prompts_processed": 24.0,
|
| 3137 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3138 |
+
"epoch": 0.18857142857142858,
|
| 3139 |
+
"grad_norm": 0.24178634583950043,
|
| 3140 |
+
"kl": 0.0533447265625,
|
| 3141 |
+
"learning_rate": 1.8594235253127372e-07,
|
| 3142 |
+
"loss": 0.0505,
|
| 3143 |
+
"reward": 0.6519163623452187,
|
| 3144 |
+
"reward_std": 0.9615699052810669,
|
| 3145 |
+
"step": 165
|
| 3146 |
+
},
|
| 3147 |
+
{
|
| 3148 |
+
"clip_fraction": 0.0,
|
| 3149 |
+
"completion_length": 3002.7882385253906,
|
| 3150 |
+
"dapo/avg_reward_std": 0.29886600477942105,
|
| 3151 |
+
"dapo/filter_reward_index": 0.0,
|
| 3152 |
+
"dapo/kept_prompts_ratio": 0.3160919598464308,
|
| 3153 |
+
"dapo/num_sampling_attempts": 3.625,
|
| 3154 |
+
"dapo/sampling_efficiency": 35.416666666666664,
|
| 3155 |
+
"dapo/total_prompts_processed": 21.75,
|
| 3156 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3157 |
+
"epoch": 0.18971428571428572,
|
| 3158 |
+
"grad_norm": 0.31221655011177063,
|
| 3159 |
+
"kl": 0.047943115234375,
|
| 3160 |
+
"learning_rate": 1.8138158006995363e-07,
|
| 3161 |
+
"loss": 0.066,
|
| 3162 |
+
"reward": 0.6383479349315166,
|
| 3163 |
+
"reward_std": 0.9029820337891579,
|
| 3164 |
+
"step": 166
|
| 3165 |
+
},
|
| 3166 |
+
{
|
| 3167 |
+
"clip_fraction": 0.0,
|
| 3168 |
+
"completion_length": 2927.295150756836,
|
| 3169 |
+
"dapo/avg_reward_std": 0.34752671499001353,
|
| 3170 |
+
"dapo/filter_reward_index": 0.0,
|
| 3171 |
+
"dapo/kept_prompts_ratio": 0.5438596621940011,
|
| 3172 |
+
"dapo/num_sampling_attempts": 2.375,
|
| 3173 |
+
"dapo/sampling_efficiency": 48.95833333333333,
|
| 3174 |
+
"dapo/total_prompts_processed": 14.25,
|
| 3175 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3176 |
+
"epoch": 0.19085714285714286,
|
| 3177 |
+
"grad_norm": 0.2697528600692749,
|
| 3178 |
+
"kl": 0.045745849609375,
|
| 3179 |
+
"learning_rate": 1.7693309235023127e-07,
|
| 3180 |
+
"loss": 0.0483,
|
| 3181 |
+
"reward": 0.8266985702211969,
|
| 3182 |
+
"reward_std": 0.9544429406523705,
|
| 3183 |
+
"step": 167
|
| 3184 |
+
},
|
| 3185 |
+
{
|
| 3186 |
+
"clip_fraction": 0.0,
|
| 3187 |
+
"completion_length": 3212.857666015625,
|
| 3188 |
+
"dapo/avg_reward_std": 0.263968757220677,
|
| 3189 |
+
"dapo/filter_reward_index": 0.0,
|
| 3190 |
+
"dapo/kept_prompts_ratio": 0.3690476247242519,
|
| 3191 |
+
"dapo/num_sampling_attempts": 3.5,
|
| 3192 |
+
"dapo/sampling_efficiency": 41.388888888888886,
|
| 3193 |
+
"dapo/total_prompts_processed": 21.0,
|
| 3194 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3195 |
+
"epoch": 0.192,
|
| 3196 |
+
"grad_norm": 0.27940821647644043,
|
| 3197 |
+
"kl": 0.05059814453125,
|
| 3198 |
+
"learning_rate": 1.7259824442455923e-07,
|
| 3199 |
+
"loss": 0.0415,
|
| 3200 |
+
"reward": 0.7715255841612816,
|
| 3201 |
+
"reward_std": 0.95072440803051,
|
| 3202 |
+
"step": 168
|
| 3203 |
+
},
|
| 3204 |
+
{
|
| 3205 |
+
"clip_fraction": 0.0,
|
| 3206 |
+
"completion_length": 3112.5799255371094,
|
| 3207 |
+
"dapo/avg_reward_std": 0.22730760558231458,
|
| 3208 |
+
"dapo/filter_reward_index": 0.0,
|
| 3209 |
+
"dapo/kept_prompts_ratio": 0.3153153222960395,
|
| 3210 |
+
"dapo/num_sampling_attempts": 4.625,
|
| 3211 |
+
"dapo/sampling_efficiency": 26.249999999999996,
|
| 3212 |
+
"dapo/total_prompts_processed": 27.75,
|
| 3213 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3214 |
+
"epoch": 0.19314285714285714,
|
| 3215 |
+
"grad_norm": 0.4339730143547058,
|
| 3216 |
+
"kl": 0.06396484375,
|
| 3217 |
+
"learning_rate": 1.6837835672960831e-07,
|
| 3218 |
+
"loss": 0.0777,
|
| 3219 |
+
"reward": 0.5262689627707005,
|
| 3220 |
+
"reward_std": 0.9779800549149513,
|
| 3221 |
+
"step": 169
|
| 3222 |
+
},
|
| 3223 |
+
{
|
| 3224 |
+
"clip_fraction": 0.0,
|
| 3225 |
+
"completion_length": 3088.6632385253906,
|
| 3226 |
+
"dapo/avg_reward_std": 0.2333034286275506,
|
| 3227 |
+
"dapo/filter_reward_index": 0.0,
|
| 3228 |
+
"dapo/kept_prompts_ratio": 0.33333333721384406,
|
| 3229 |
+
"dapo/num_sampling_attempts": 4.0,
|
| 3230 |
+
"dapo/sampling_efficiency": 38.263888888888886,
|
| 3231 |
+
"dapo/total_prompts_processed": 24.0,
|
| 3232 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3233 |
+
"epoch": 0.19428571428571428,
|
| 3234 |
+
"grad_norm": 0.48384836316108704,
|
| 3235 |
+
"kl": 0.0555419921875,
|
| 3236 |
+
"learning_rate": 1.6427471468404952e-07,
|
| 3237 |
+
"loss": 0.0974,
|
| 3238 |
+
"reward": 0.7407102398574352,
|
| 3239 |
+
"reward_std": 0.9568767622113228,
|
| 3240 |
+
"step": 170
|
| 3241 |
+
},
|
| 3242 |
+
{
|
| 3243 |
+
"clip_fraction": 0.0,
|
| 3244 |
+
"completion_length": 3099.347198486328,
|
| 3245 |
+
"dapo/avg_reward_std": 0.17301563743282766,
|
| 3246 |
+
"dapo/filter_reward_index": 0.0,
|
| 3247 |
+
"dapo/kept_prompts_ratio": 0.27941177127992406,
|
| 3248 |
+
"dapo/num_sampling_attempts": 4.25,
|
| 3249 |
+
"dapo/sampling_efficiency": 31.874999999999996,
|
| 3250 |
+
"dapo/total_prompts_processed": 25.5,
|
| 3251 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3252 |
+
"epoch": 0.19542857142857142,
|
| 3253 |
+
"grad_norm": 0.42263394594192505,
|
| 3254 |
+
"kl": 0.0595703125,
|
| 3255 |
+
"learning_rate": 1.6028856829700258e-07,
|
| 3256 |
+
"loss": 0.0812,
|
| 3257 |
+
"reward": 0.4282900430262089,
|
| 3258 |
+
"reward_std": 0.914498083293438,
|
| 3259 |
+
"step": 171
|
| 3260 |
+
},
|
| 3261 |
+
{
|
| 3262 |
+
"clip_fraction": 0.0,
|
| 3263 |
+
"completion_length": 3111.232696533203,
|
| 3264 |
+
"dapo/avg_reward_std": 0.2433939976617694,
|
| 3265 |
+
"dapo/filter_reward_index": 0.0,
|
| 3266 |
+
"dapo/kept_prompts_ratio": 0.33333334093913436,
|
| 3267 |
+
"dapo/num_sampling_attempts": 4.0,
|
| 3268 |
+
"dapo/sampling_efficiency": 36.80555555555555,
|
| 3269 |
+
"dapo/total_prompts_processed": 24.0,
|
| 3270 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3271 |
+
"epoch": 0.19657142857142856,
|
| 3272 |
+
"grad_norm": 0.4814501404762268,
|
| 3273 |
+
"kl": 0.05926513671875,
|
| 3274 |
+
"learning_rate": 1.5642113178727193e-07,
|
| 3275 |
+
"loss": 0.0843,
|
| 3276 |
+
"reward": 0.6843680012971163,
|
| 3277 |
+
"reward_std": 0.8743765726685524,
|
| 3278 |
+
"step": 172
|
| 3279 |
+
},
|
| 3280 |
+
{
|
| 3281 |
+
"clip_fraction": 0.0,
|
| 3282 |
+
"completion_length": 3008.6563110351562,
|
| 3283 |
+
"dapo/avg_reward_std": 0.25363275137814606,
|
| 3284 |
+
"dapo/filter_reward_index": 0.0,
|
| 3285 |
+
"dapo/kept_prompts_ratio": 0.31818182811592566,
|
| 3286 |
+
"dapo/num_sampling_attempts": 4.125,
|
| 3287 |
+
"dapo/sampling_efficiency": 38.78472222222222,
|
| 3288 |
+
"dapo/total_prompts_processed": 24.75,
|
| 3289 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3290 |
+
"epoch": 0.1977142857142857,
|
| 3291 |
+
"grad_norm": 0.285697877407074,
|
| 3292 |
+
"kl": 0.05755615234375,
|
| 3293 |
+
"learning_rate": 1.5267358321348285e-07,
|
| 3294 |
+
"loss": 0.0456,
|
| 3295 |
+
"reward": 0.5798944532871246,
|
| 3296 |
+
"reward_std": 0.984041191637516,
|
| 3297 |
+
"step": 173
|
| 3298 |
+
},
|
| 3299 |
+
{
|
| 3300 |
+
"clip_fraction": 0.0,
|
| 3301 |
+
"completion_length": 3067.9791870117188,
|
| 3302 |
+
"dapo/avg_reward_std": 0.3438388824462891,
|
| 3303 |
+
"dapo/filter_reward_index": 0.0,
|
| 3304 |
+
"dapo/kept_prompts_ratio": 0.4500000044703484,
|
| 3305 |
+
"dapo/num_sampling_attempts": 2.5,
|
| 3306 |
+
"dapo/sampling_efficiency": 48.33333333333333,
|
| 3307 |
+
"dapo/total_prompts_processed": 15.0,
|
| 3308 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3309 |
+
"epoch": 0.19885714285714284,
|
| 3310 |
+
"grad_norm": 0.43520498275756836,
|
| 3311 |
+
"kl": 0.07098388671875,
|
| 3312 |
+
"learning_rate": 1.4904706411523448e-07,
|
| 3313 |
+
"loss": 0.0716,
|
| 3314 |
+
"reward": 0.5646946905180812,
|
| 3315 |
+
"reward_std": 0.9460153579711914,
|
| 3316 |
+
"step": 174
|
| 3317 |
+
},
|
| 3318 |
+
{
|
| 3319 |
+
"clip_fraction": 0.0,
|
| 3320 |
+
"completion_length": 3223.2916870117188,
|
| 3321 |
+
"dapo/avg_reward_std": 0.2690600073337555,
|
| 3322 |
+
"dapo/filter_reward_index": 0.0,
|
| 3323 |
+
"dapo/kept_prompts_ratio": 0.4133333420753479,
|
| 3324 |
+
"dapo/num_sampling_attempts": 3.125,
|
| 3325 |
+
"dapo/sampling_efficiency": 41.45833333333333,
|
| 3326 |
+
"dapo/total_prompts_processed": 18.75,
|
| 3327 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3328 |
+
"epoch": 0.2,
|
| 3329 |
+
"grad_norm": 0.35144945979118347,
|
| 3330 |
+
"kl": 0.06170654296875,
|
| 3331 |
+
"learning_rate": 1.4554267916537495e-07,
|
| 3332 |
+
"loss": 0.0348,
|
| 3333 |
+
"reward": 0.556399748660624,
|
| 3334 |
+
"reward_std": 0.9192204177379608,
|
| 3335 |
+
"step": 175
|
| 3336 |
+
},
|
| 3337 |
+
{
|
| 3338 |
+
"clip_fraction": 0.0,
|
| 3339 |
+
"completion_length": 2946.0799102783203,
|
| 3340 |
+
"dapo/avg_reward_std": 0.25316954652468365,
|
| 3341 |
+
"dapo/filter_reward_index": 0.0,
|
| 3342 |
+
"dapo/kept_prompts_ratio": 0.37500000558793545,
|
| 3343 |
+
"dapo/num_sampling_attempts": 3.0,
|
| 3344 |
+
"dapo/sampling_efficiency": 43.75,
|
| 3345 |
+
"dapo/total_prompts_processed": 18.0,
|
| 3346 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3347 |
+
"epoch": 0.20114285714285715,
|
| 3348 |
+
"grad_norm": 0.46807849407196045,
|
| 3349 |
+
"kl": 0.063018798828125,
|
| 3350 |
+
"learning_rate": 1.4216149583350755e-07,
|
| 3351 |
+
"loss": 0.0796,
|
| 3352 |
+
"reward": 0.6736351866275072,
|
| 3353 |
+
"reward_std": 0.9649264737963676,
|
| 3354 |
+
"step": 176
|
| 3355 |
+
},
|
| 3356 |
+
{
|
| 3357 |
+
"clip_fraction": 0.0,
|
| 3358 |
+
"completion_length": 3096.829864501953,
|
| 3359 |
+
"dapo/avg_reward_std": 0.31567848042437907,
|
| 3360 |
+
"dapo/filter_reward_index": 0.0,
|
| 3361 |
+
"dapo/kept_prompts_ratio": 0.482456142965116,
|
| 3362 |
+
"dapo/num_sampling_attempts": 2.375,
|
| 3363 |
+
"dapo/sampling_efficiency": 55.625,
|
| 3364 |
+
"dapo/total_prompts_processed": 14.25,
|
| 3365 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3366 |
+
"epoch": 0.2022857142857143,
|
| 3367 |
+
"grad_norm": 0.31731271743774414,
|
| 3368 |
+
"kl": 0.055938720703125,
|
| 3369 |
+
"learning_rate": 1.3890454406082956e-07,
|
| 3370 |
+
"loss": 0.0386,
|
| 3371 |
+
"reward": 0.681073285639286,
|
| 3372 |
+
"reward_std": 0.9661536440253258,
|
| 3373 |
+
"step": 177
|
| 3374 |
+
},
|
| 3375 |
+
{
|
| 3376 |
+
"clip_fraction": 0.0,
|
| 3377 |
+
"completion_length": 3235.8056030273438,
|
| 3378 |
+
"dapo/avg_reward_std": 0.24198689542967697,
|
| 3379 |
+
"dapo/filter_reward_index": 0.0,
|
| 3380 |
+
"dapo/kept_prompts_ratio": 0.3448275898037286,
|
| 3381 |
+
"dapo/num_sampling_attempts": 3.625,
|
| 3382 |
+
"dapo/sampling_efficiency": 47.08333333333333,
|
| 3383 |
+
"dapo/total_prompts_processed": 21.75,
|
| 3384 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3385 |
+
"epoch": 0.20342857142857143,
|
| 3386 |
+
"grad_norm": 0.4640950560569763,
|
| 3387 |
+
"kl": 0.072052001953125,
|
| 3388 |
+
"learning_rate": 1.3577281594640182e-07,
|
| 3389 |
+
"loss": 0.0702,
|
| 3390 |
+
"reward": 0.5520291309803724,
|
| 3391 |
+
"reward_std": 0.9967257082462311,
|
| 3392 |
+
"step": 178
|
| 3393 |
+
},
|
| 3394 |
+
{
|
| 3395 |
+
"clip_fraction": 0.0,
|
| 3396 |
+
"completion_length": 3237.77783203125,
|
| 3397 |
+
"dapo/avg_reward_std": 0.30828417566689575,
|
| 3398 |
+
"dapo/filter_reward_index": 0.0,
|
| 3399 |
+
"dapo/kept_prompts_ratio": 0.4242424341765317,
|
| 3400 |
+
"dapo/num_sampling_attempts": 2.75,
|
| 3401 |
+
"dapo/sampling_efficiency": 52.82738095238095,
|
| 3402 |
+
"dapo/total_prompts_processed": 16.5,
|
| 3403 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3404 |
+
"epoch": 0.20457142857142857,
|
| 3405 |
+
"grad_norm": 0.4502318203449249,
|
| 3406 |
+
"kl": 0.07550048828125,
|
| 3407 |
+
"learning_rate": 1.3276726544494571e-07,
|
| 3408 |
+
"loss": 0.0614,
|
| 3409 |
+
"reward": 0.6213867999613285,
|
| 3410 |
+
"reward_std": 0.9431608989834785,
|
| 3411 |
+
"step": 179
|
| 3412 |
+
},
|
| 3413 |
+
{
|
| 3414 |
+
"clip_fraction": 0.0,
|
| 3415 |
+
"completion_length": 2887.9236450195312,
|
| 3416 |
+
"dapo/avg_reward_std": 0.2488611958645008,
|
| 3417 |
+
"dapo/filter_reward_index": 0.0,
|
| 3418 |
+
"dapo/kept_prompts_ratio": 0.3518518612340645,
|
| 3419 |
+
"dapo/num_sampling_attempts": 3.375,
|
| 3420 |
+
"dapo/sampling_efficiency": 48.035714285714285,
|
| 3421 |
+
"dapo/total_prompts_processed": 20.25,
|
| 3422 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3423 |
+
"epoch": 0.2057142857142857,
|
| 3424 |
+
"grad_norm": 0.44646504521369934,
|
| 3425 |
+
"kl": 0.073760986328125,
|
| 3426 |
+
"learning_rate": 1.2988880807625927e-07,
|
| 3427 |
+
"loss": 0.0683,
|
| 3428 |
+
"reward": 0.5839751102030277,
|
| 3429 |
+
"reward_std": 0.9090578481554985,
|
| 3430 |
+
"step": 180
|
| 3431 |
+
},
|
| 3432 |
+
{
|
| 3433 |
+
"clip_fraction": 0.0,
|
| 3434 |
+
"completion_length": 3021.2916870117188,
|
| 3435 |
+
"dapo/avg_reward_std": 0.20883248069069602,
|
| 3436 |
+
"dapo/filter_reward_index": 0.0,
|
| 3437 |
+
"dapo/kept_prompts_ratio": 0.2878787942004926,
|
| 3438 |
+
"dapo/num_sampling_attempts": 4.125,
|
| 3439 |
+
"dapo/sampling_efficiency": 39.632936507936506,
|
| 3440 |
+
"dapo/total_prompts_processed": 24.75,
|
| 3441 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3442 |
+
"epoch": 0.20685714285714285,
|
| 3443 |
+
"grad_norm": 0.36042678356170654,
|
| 3444 |
+
"kl": 0.07421875,
|
| 3445 |
+
"learning_rate": 1.2713832064634125e-07,
|
| 3446 |
+
"loss": 0.054,
|
| 3447 |
+
"reward": 0.5517729418352246,
|
| 3448 |
+
"reward_std": 0.9483400657773018,
|
| 3449 |
+
"step": 181
|
| 3450 |
+
},
|
| 3451 |
+
{
|
| 3452 |
+
"clip_fraction": 0.0,
|
| 3453 |
+
"completion_length": 3249.2118530273438,
|
| 3454 |
+
"dapo/avg_reward_std": 0.2615335573043142,
|
| 3455 |
+
"dapo/filter_reward_index": 0.0,
|
| 3456 |
+
"dapo/kept_prompts_ratio": 0.33333333847778185,
|
| 3457 |
+
"dapo/num_sampling_attempts": 3.5,
|
| 3458 |
+
"dapo/sampling_efficiency": 46.785714285714285,
|
| 3459 |
+
"dapo/total_prompts_processed": 21.0,
|
| 3460 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3461 |
+
"epoch": 0.208,
|
| 3462 |
+
"grad_norm": 0.4518042504787445,
|
| 3463 |
+
"kl": 0.072021484375,
|
| 3464 |
+
"learning_rate": 1.2451664098030743e-07,
|
| 3465 |
+
"loss": 0.0654,
|
| 3466 |
+
"reward": 0.686168298125267,
|
| 3467 |
+
"reward_std": 0.9350233674049377,
|
| 3468 |
+
"step": 182
|
| 3469 |
+
},
|
| 3470 |
+
{
|
| 3471 |
+
"clip_fraction": 0.0,
|
| 3472 |
+
"completion_length": 3221.6631774902344,
|
| 3473 |
+
"dapo/avg_reward_std": 0.27866364789731574,
|
| 3474 |
+
"dapo/filter_reward_index": 0.0,
|
| 3475 |
+
"dapo/kept_prompts_ratio": 0.3686868738044392,
|
| 3476 |
+
"dapo/num_sampling_attempts": 4.125,
|
| 3477 |
+
"dapo/sampling_efficiency": 28.4375,
|
| 3478 |
+
"dapo/total_prompts_processed": 24.75,
|
| 3479 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3480 |
+
"epoch": 0.20914285714285713,
|
| 3481 |
+
"grad_norm": 0.32408109307289124,
|
| 3482 |
+
"kl": 0.062255859375,
|
| 3483 |
+
"learning_rate": 1.220245676671809e-07,
|
| 3484 |
+
"loss": 0.0384,
|
| 3485 |
+
"reward": 0.6384344138205051,
|
| 3486 |
+
"reward_std": 0.9783304929733276,
|
| 3487 |
+
"step": 183
|
| 3488 |
+
},
|
| 3489 |
+
{
|
| 3490 |
+
"clip_fraction": 0.0,
|
| 3491 |
+
"completion_length": 3199.1354370117188,
|
| 3492 |
+
"dapo/avg_reward_std": 0.2816663732131322,
|
| 3493 |
+
"dapo/filter_reward_index": 0.0,
|
| 3494 |
+
"dapo/kept_prompts_ratio": 0.316666671137015,
|
| 3495 |
+
"dapo/num_sampling_attempts": 3.75,
|
| 3496 |
+
"dapo/sampling_efficiency": 45.55555555555555,
|
| 3497 |
+
"dapo/total_prompts_processed": 22.5,
|
| 3498 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3499 |
+
"epoch": 0.2102857142857143,
|
| 3500 |
+
"grad_norm": 0.2197091430425644,
|
| 3501 |
+
"kl": 0.07550048828125,
|
| 3502 |
+
"learning_rate": 1.1966285981663407e-07,
|
| 3503 |
+
"loss": 0.0211,
|
| 3504 |
+
"reward": 0.45471471454948187,
|
| 3505 |
+
"reward_std": 0.9136239141225815,
|
| 3506 |
+
"step": 184
|
| 3507 |
+
},
|
| 3508 |
+
{
|
| 3509 |
+
"clip_fraction": 0.0,
|
| 3510 |
+
"completion_length": 3037.420166015625,
|
| 3511 |
+
"dapo/avg_reward_std": 0.17516983683044846,
|
| 3512 |
+
"dapo/filter_reward_index": 0.0,
|
| 3513 |
+
"dapo/kept_prompts_ratio": 0.2657657728807346,
|
| 3514 |
+
"dapo/num_sampling_attempts": 4.625,
|
| 3515 |
+
"dapo/sampling_efficiency": 25.729166666666664,
|
| 3516 |
+
"dapo/total_prompts_processed": 27.75,
|
| 3517 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3518 |
+
"epoch": 0.21142857142857144,
|
| 3519 |
+
"grad_norm": 0.4012245535850525,
|
| 3520 |
+
"kl": 0.091796875,
|
| 3521 |
+
"learning_rate": 1.1743223682775649e-07,
|
| 3522 |
+
"loss": 0.0442,
|
| 3523 |
+
"reward": 0.7168623730540276,
|
| 3524 |
+
"reward_std": 0.9515729621052742,
|
| 3525 |
+
"step": 185
|
| 3526 |
+
},
|
| 3527 |
+
{
|
| 3528 |
+
"clip_fraction": 0.0,
|
| 3529 |
+
"completion_length": 3222.767364501953,
|
| 3530 |
+
"dapo/avg_reward_std": 0.2550514280796051,
|
| 3531 |
+
"dapo/filter_reward_index": 0.0,
|
| 3532 |
+
"dapo/kept_prompts_ratio": 0.3444444512327512,
|
| 3533 |
+
"dapo/num_sampling_attempts": 3.75,
|
| 3534 |
+
"dapo/sampling_efficiency": 43.64583333333333,
|
| 3535 |
+
"dapo/total_prompts_processed": 22.5,
|
| 3536 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3537 |
+
"epoch": 0.21257142857142858,
|
| 3538 |
+
"grad_norm": 0.4945845305919647,
|
| 3539 |
+
"kl": 0.083465576171875,
|
| 3540 |
+
"learning_rate": 1.1533337816991931e-07,
|
| 3541 |
+
"loss": 0.0667,
|
| 3542 |
+
"reward": 0.5391142014414072,
|
| 3543 |
+
"reward_std": 0.9342528805136681,
|
| 3544 |
+
"step": 186
|
| 3545 |
+
},
|
| 3546 |
+
{
|
| 3547 |
+
"clip_fraction": 0.0,
|
| 3548 |
+
"completion_length": 2858.5659942626953,
|
| 3549 |
+
"dapo/avg_reward_std": 0.23423856112264818,
|
| 3550 |
+
"dapo/filter_reward_index": 0.0,
|
| 3551 |
+
"dapo/kept_prompts_ratio": 0.3118279609949358,
|
| 3552 |
+
"dapo/num_sampling_attempts": 3.875,
|
| 3553 |
+
"dapo/sampling_efficiency": 40.0297619047619,
|
| 3554 |
+
"dapo/total_prompts_processed": 23.25,
|
| 3555 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3556 |
+
"epoch": 0.21371428571428572,
|
| 3557 |
+
"grad_norm": 0.4291866421699524,
|
| 3558 |
+
"kl": 0.091796875,
|
| 3559 |
+
"learning_rate": 1.1336692317580158e-07,
|
| 3560 |
+
"loss": 0.0384,
|
| 3561 |
+
"reward": 0.7481220848858356,
|
| 3562 |
+
"reward_std": 0.9474795907735825,
|
| 3563 |
+
"step": 187
|
| 3564 |
+
},
|
| 3565 |
+
{
|
| 3566 |
+
"clip_fraction": 0.0,
|
| 3567 |
+
"completion_length": 3123.170166015625,
|
| 3568 |
+
"dapo/avg_reward_std": 0.1988734739857751,
|
| 3569 |
+
"dapo/filter_reward_index": 0.0,
|
| 3570 |
+
"dapo/kept_prompts_ratio": 0.25675675997862946,
|
| 3571 |
+
"dapo/num_sampling_attempts": 4.625,
|
| 3572 |
+
"dapo/sampling_efficiency": 31.875,
|
| 3573 |
+
"dapo/total_prompts_processed": 27.75,
|
| 3574 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3575 |
+
"epoch": 0.21485714285714286,
|
| 3576 |
+
"grad_norm": 0.30453264713287354,
|
| 3577 |
+
"kl": 0.080657958984375,
|
| 3578 |
+
"learning_rate": 1.1153347084664419e-07,
|
| 3579 |
+
"loss": 0.0273,
|
| 3580 |
+
"reward": 0.6236942922696471,
|
| 3581 |
+
"reward_std": 0.9715093299746513,
|
| 3582 |
+
"step": 188
|
| 3583 |
+
},
|
| 3584 |
+
{
|
| 3585 |
+
"clip_fraction": 0.0,
|
| 3586 |
+
"completion_length": 2872.9618530273438,
|
| 3587 |
+
"dapo/avg_reward_std": 0.21385114904372923,
|
| 3588 |
+
"dapo/filter_reward_index": 0.0,
|
| 3589 |
+
"dapo/kept_prompts_ratio": 0.3333333365378841,
|
| 3590 |
+
"dapo/num_sampling_attempts": 3.875,
|
| 3591 |
+
"dapo/sampling_efficiency": 39.18154761904762,
|
| 3592 |
+
"dapo/total_prompts_processed": 23.25,
|
| 3593 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3594 |
+
"epoch": 0.216,
|
| 3595 |
+
"grad_norm": 0.5780288577079773,
|
| 3596 |
+
"kl": 0.08612060546875,
|
| 3597 |
+
"learning_rate": 1.0983357966978745e-07,
|
| 3598 |
+
"loss": 0.0607,
|
| 3599 |
+
"reward": 0.7514887787401676,
|
| 3600 |
+
"reward_std": 1.0098591819405556,
|
| 3601 |
+
"step": 189
|
| 3602 |
+
},
|
| 3603 |
+
{
|
| 3604 |
+
"clip_fraction": 0.0,
|
| 3605 |
+
"completion_length": 2937.093780517578,
|
| 3606 |
+
"dapo/avg_reward_std": 0.1677520631575117,
|
| 3607 |
+
"dapo/filter_reward_index": 0.0,
|
| 3608 |
+
"dapo/kept_prompts_ratio": 0.21895425284610076,
|
| 3609 |
+
"dapo/num_sampling_attempts": 6.375,
|
| 3610 |
+
"dapo/sampling_efficiency": 20.689484126984123,
|
| 3611 |
+
"dapo/total_prompts_processed": 38.25,
|
| 3612 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3613 |
+
"epoch": 0.21714285714285714,
|
| 3614 |
+
"grad_norm": 0.3947860896587372,
|
| 3615 |
+
"kl": 0.076263427734375,
|
| 3616 |
+
"learning_rate": 1.0826776744855121e-07,
|
| 3617 |
+
"loss": 0.0487,
|
| 3618 |
+
"reward": 0.6180934552103281,
|
| 3619 |
+
"reward_std": 0.9050487726926804,
|
| 3620 |
+
"step": 190
|
| 3621 |
+
},
|
| 3622 |
+
{
|
| 3623 |
+
"clip_fraction": 0.0,
|
| 3624 |
+
"completion_length": 3252.090301513672,
|
| 3625 |
+
"dapo/avg_reward_std": 0.24265852073828378,
|
| 3626 |
+
"dapo/filter_reward_index": 0.0,
|
| 3627 |
+
"dapo/kept_prompts_ratio": 0.3055555621782939,
|
| 3628 |
+
"dapo/num_sampling_attempts": 3.75,
|
| 3629 |
+
"dapo/sampling_efficiency": 38.020833333333336,
|
| 3630 |
+
"dapo/total_prompts_processed": 22.5,
|
| 3631 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3632 |
+
"epoch": 0.21828571428571428,
|
| 3633 |
+
"grad_norm": 0.48333072662353516,
|
| 3634 |
+
"kl": 0.09661865234375,
|
| 3635 |
+
"learning_rate": 1.068365111445064e-07,
|
| 3636 |
+
"loss": 0.0584,
|
| 3637 |
+
"reward": 0.4759152363985777,
|
| 3638 |
+
"reward_std": 0.9479196071624756,
|
| 3639 |
+
"step": 191
|
| 3640 |
+
},
|
| 3641 |
+
{
|
| 3642 |
+
"clip_fraction": 0.0,
|
| 3643 |
+
"completion_length": 3074.420166015625,
|
| 3644 |
+
"dapo/avg_reward_std": 0.2189681170315578,
|
| 3645 |
+
"dapo/filter_reward_index": 0.0,
|
| 3646 |
+
"dapo/kept_prompts_ratio": 0.3333333371014431,
|
| 3647 |
+
"dapo/num_sampling_attempts": 3.625,
|
| 3648 |
+
"dapo/sampling_efficiency": 46.45833333333333,
|
| 3649 |
+
"dapo/total_prompts_processed": 21.75,
|
| 3650 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3651 |
+
"epoch": 0.21942857142857142,
|
| 3652 |
+
"grad_norm": 0.5536202192306519,
|
| 3653 |
+
"kl": 0.09814453125,
|
| 3654 |
+
"learning_rate": 1.0554024673218806e-07,
|
| 3655 |
+
"loss": 0.0731,
|
| 3656 |
+
"reward": 0.48804986744653434,
|
| 3657 |
+
"reward_std": 0.9367131069302559,
|
| 3658 |
+
"step": 192
|
| 3659 |
+
},
|
| 3660 |
+
{
|
| 3661 |
+
"clip_fraction": 0.0,
|
| 3662 |
+
"completion_length": 3026.9097595214844,
|
| 3663 |
+
"dapo/avg_reward_std": 0.21337791310774312,
|
| 3664 |
+
"dapo/filter_reward_index": 0.0,
|
| 3665 |
+
"dapo/kept_prompts_ratio": 0.256756762395034,
|
| 3666 |
+
"dapo/num_sampling_attempts": 4.625,
|
| 3667 |
+
"dapo/sampling_efficiency": 30.3125,
|
| 3668 |
+
"dapo/total_prompts_processed": 27.75,
|
| 3669 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3670 |
+
"epoch": 0.22057142857142858,
|
| 3671 |
+
"grad_norm": 0.5239105224609375,
|
| 3672 |
+
"kl": 0.0985107421875,
|
| 3673 |
+
"learning_rate": 1.0437936906629334e-07,
|
| 3674 |
+
"loss": 0.0561,
|
| 3675 |
+
"reward": 0.45341441221535206,
|
| 3676 |
+
"reward_std": 0.8912393003702164,
|
| 3677 |
+
"step": 193
|
| 3678 |
+
},
|
| 3679 |
+
{
|
| 3680 |
+
"clip_fraction": 0.0,
|
| 3681 |
+
"completion_length": 2896.656280517578,
|
| 3682 |
+
"dapo/avg_reward_std": 0.31374274492263793,
|
| 3683 |
+
"dapo/filter_reward_index": 0.0,
|
| 3684 |
+
"dapo/kept_prompts_ratio": 0.4333333417773247,
|
| 3685 |
+
"dapo/num_sampling_attempts": 2.5,
|
| 3686 |
+
"dapo/sampling_efficiency": 46.875,
|
| 3687 |
+
"dapo/total_prompts_processed": 15.0,
|
| 3688 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3689 |
+
"epoch": 0.22171428571428572,
|
| 3690 |
+
"grad_norm": 0.6310634016990662,
|
| 3691 |
+
"kl": 0.108062744140625,
|
| 3692 |
+
"learning_rate": 1.0335423176140511e-07,
|
| 3693 |
+
"loss": 0.0809,
|
| 3694 |
+
"reward": 0.6844924800097942,
|
| 3695 |
+
"reward_std": 0.9649646729230881,
|
| 3696 |
+
"step": 194
|
| 3697 |
+
},
|
| 3698 |
+
{
|
| 3699 |
+
"clip_fraction": 0.0,
|
| 3700 |
+
"completion_length": 3319.7048950195312,
|
| 3701 |
+
"dapo/avg_reward_std": 0.21983732057340216,
|
| 3702 |
+
"dapo/filter_reward_index": 0.0,
|
| 3703 |
+
"dapo/kept_prompts_ratio": 0.30303031251286017,
|
| 3704 |
+
"dapo/num_sampling_attempts": 4.125,
|
| 3705 |
+
"dapo/sampling_efficiency": 29.479166666666664,
|
| 3706 |
+
"dapo/total_prompts_processed": 24.75,
|
| 3707 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3708 |
+
"epoch": 0.22285714285714286,
|
| 3709 |
+
"grad_norm": 0.47936248779296875,
|
| 3710 |
+
"kl": 0.0997314453125,
|
| 3711 |
+
"learning_rate": 1.0246514708427701e-07,
|
| 3712 |
+
"loss": 0.0479,
|
| 3713 |
+
"reward": 0.3993752491660416,
|
| 3714 |
+
"reward_std": 0.9481607303023338,
|
| 3715 |
+
"step": 195
|
| 3716 |
+
},
|
| 3717 |
+
{
|
| 3718 |
+
"clip_fraction": 0.0,
|
| 3719 |
+
"completion_length": 3298.1736450195312,
|
| 3720 |
+
"dapo/avg_reward_std": 0.2514548934996128,
|
| 3721 |
+
"dapo/filter_reward_index": 0.0,
|
| 3722 |
+
"dapo/kept_prompts_ratio": 0.28333333916962145,
|
| 3723 |
+
"dapo/num_sampling_attempts": 5.0,
|
| 3724 |
+
"dapo/sampling_efficiency": 33.13988095238095,
|
| 3725 |
+
"dapo/total_prompts_processed": 30.0,
|
| 3726 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3727 |
+
"epoch": 0.224,
|
| 3728 |
+
"grad_norm": 0.36350947618484497,
|
| 3729 |
+
"kl": 0.1043701171875,
|
| 3730 |
+
"learning_rate": 1.017123858587145e-07,
|
| 3731 |
+
"loss": 0.0389,
|
| 3732 |
+
"reward": 0.31427645590156317,
|
| 3733 |
+
"reward_std": 0.8980218172073364,
|
| 3734 |
+
"step": 196
|
| 3735 |
+
},
|
| 3736 |
+
{
|
| 3737 |
+
"clip_fraction": 0.0,
|
| 3738 |
+
"completion_length": 3260.4861450195312,
|
| 3739 |
+
"dapo/avg_reward_std": 0.1836753969009106,
|
| 3740 |
+
"dapo/filter_reward_index": 0.0,
|
| 3741 |
+
"dapo/kept_prompts_ratio": 0.24786325486806723,
|
| 3742 |
+
"dapo/num_sampling_attempts": 4.875,
|
| 3743 |
+
"dapo/sampling_efficiency": 28.154761904761905,
|
| 3744 |
+
"dapo/total_prompts_processed": 29.25,
|
| 3745 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3746 |
+
"epoch": 0.22514285714285714,
|
| 3747 |
+
"grad_norm": 0.3354601562023163,
|
| 3748 |
+
"kl": 0.0946044921875,
|
| 3749 |
+
"learning_rate": 1.0109617738307911e-07,
|
| 3750 |
+
"loss": 0.0301,
|
| 3751 |
+
"reward": 0.5015182960778475,
|
| 3752 |
+
"reward_std": 0.9334053322672844,
|
| 3753 |
+
"step": 197
|
| 3754 |
+
},
|
| 3755 |
+
{
|
| 3756 |
+
"clip_fraction": 0.0,
|
| 3757 |
+
"completion_length": 3031.3958129882812,
|
| 3758 |
+
"dapo/avg_reward_std": 0.3008538554696476,
|
| 3759 |
+
"dapo/filter_reward_index": 0.0,
|
| 3760 |
+
"dapo/kept_prompts_ratio": 0.5196078463512308,
|
| 3761 |
+
"dapo/num_sampling_attempts": 2.125,
|
| 3762 |
+
"dapo/sampling_efficiency": 76.5625,
|
| 3763 |
+
"dapo/total_prompts_processed": 12.75,
|
| 3764 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3765 |
+
"epoch": 0.22628571428571428,
|
| 3766 |
+
"grad_norm": 0.48223650455474854,
|
| 3767 |
+
"kl": 0.10247802734375,
|
| 3768 |
+
"learning_rate": 1.0061670936044178e-07,
|
| 3769 |
+
"loss": 0.0648,
|
| 3770 |
+
"reward": 0.573589576408267,
|
| 3771 |
+
"reward_std": 0.9578919112682343,
|
| 3772 |
+
"step": 198
|
| 3773 |
+
},
|
| 3774 |
+
{
|
| 3775 |
+
"clip_fraction": 0.0,
|
| 3776 |
+
"completion_length": 2948.3854064941406,
|
| 3777 |
+
"dapo/avg_reward_std": 0.43072181940078735,
|
| 3778 |
+
"dapo/filter_reward_index": 0.0,
|
| 3779 |
+
"dapo/kept_prompts_ratio": 0.8333333373069763,
|
| 3780 |
+
"dapo/num_sampling_attempts": 1.25,
|
| 3781 |
+
"dapo/sampling_efficiency": 87.5,
|
| 3782 |
+
"dapo/total_prompts_processed": 7.5,
|
| 3783 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3784 |
+
"epoch": 0.22742857142857142,
|
| 3785 |
+
"grad_norm": 0.6141620874404907,
|
| 3786 |
+
"kl": 0.09808349609375,
|
| 3787 |
+
"learning_rate": 1.002741278414069e-07,
|
| 3788 |
+
"loss": 0.0827,
|
| 3789 |
+
"reward": 0.7053878791630268,
|
| 3790 |
+
"reward_std": 0.9694960787892342,
|
| 3791 |
+
"step": 199
|
| 3792 |
+
},
|
| 3793 |
+
{
|
| 3794 |
+
"clip_fraction": 0.0,
|
| 3795 |
+
"completion_length": 2714.482666015625,
|
| 3796 |
+
"dapo/avg_reward_std": 0.26207208441149804,
|
| 3797 |
+
"dapo/filter_reward_index": 0.0,
|
| 3798 |
+
"dapo/kept_prompts_ratio": 0.37096774914572317,
|
| 3799 |
+
"dapo/num_sampling_attempts": 3.875,
|
| 3800 |
+
"dapo/sampling_efficiency": 30.624999999999993,
|
| 3801 |
+
"dapo/total_prompts_processed": 23.25,
|
| 3802 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3803 |
+
"epoch": 0.22857142857142856,
|
| 3804 |
+
"grad_norm": 0.2072688341140747,
|
| 3805 |
+
"kl": 0.1064453125,
|
| 3806 |
+
"learning_rate": 1.0006853717962393e-07,
|
| 3807 |
+
"loss": 0.0122,
|
| 3808 |
+
"reward": 0.5771910101175308,
|
| 3809 |
+
"reward_std": 0.9156405553221703,
|
| 3810 |
+
"step": 200
|
| 3811 |
+
},
|
| 3812 |
+
{
|
| 3813 |
+
"epoch": 0.22857142857142856,
|
| 3814 |
+
"step": 200,
|
| 3815 |
"total_flos": 0.0,
|
| 3816 |
+
"train_loss": 0.02940896774176508,
|
| 3817 |
+
"train_runtime": 83918.4654,
|
| 3818 |
+
"train_samples_per_second": 0.114,
|
| 3819 |
+
"train_steps_per_second": 0.002
|
| 3820 |
}
|
| 3821 |
],
|
| 3822 |
"logging_steps": 1,
|
| 3823 |
+
"max_steps": 200,
|
| 3824 |
"num_input_tokens_seen": 0,
|
| 3825 |
"num_train_epochs": 1,
|
| 3826 |
"save_steps": 10,
|