Model save
Browse files- README.md +2 -4
- adapter/adapter_config.json +4 -4
- adapter/adapter_model.safetensors +1 -1
- adapter/training_args.bin +1 -1
- all_results.json +3 -3
- train_results.json +3 -3
- trainer_state.json +1908 -8
README.md
CHANGED
|
@@ -1,19 +1,17 @@
|
|
| 1 |
---
|
| 2 |
base_model: deepseek-ai/DeepSeek-R1-Distill-Qwen-7B
|
| 3 |
-
datasets: knoveleng/open-rs
|
| 4 |
library_name: transformers
|
| 5 |
model_name: DAPO-7B
|
| 6 |
tags:
|
| 7 |
- generated_from_trainer
|
| 8 |
-
- open-r1
|
| 9 |
-
- dapo
|
| 10 |
- trl
|
|
|
|
| 11 |
licence: license
|
| 12 |
---
|
| 13 |
|
| 14 |
# Model Card for DAPO-7B
|
| 15 |
|
| 16 |
-
This model is a fine-tuned version of [deepseek-ai/DeepSeek-R1-Distill-Qwen-7B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-7B)
|
| 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-7B
|
|
|
|
| 3 |
library_name: transformers
|
| 4 |
model_name: DAPO-7B
|
| 5 |
tags:
|
| 6 |
- generated_from_trainer
|
|
|
|
|
|
|
| 7 |
- trl
|
| 8 |
+
- dapo
|
| 9 |
licence: license
|
| 10 |
---
|
| 11 |
|
| 12 |
# Model Card for DAPO-7B
|
| 13 |
|
| 14 |
+
This model is a fine-tuned version of [deepseek-ai/DeepSeek-R1-Distill-Qwen-7B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-7B).
|
| 15 |
It has been trained using [TRL](https://github.com/huggingface/trl).
|
| 16 |
|
| 17 |
## Quick start
|
adapter/adapter_config.json
CHANGED
|
@@ -29,13 +29,13 @@
|
|
| 29 |
"rank_pattern": {},
|
| 30 |
"revision": null,
|
| 31 |
"target_modules": [
|
| 32 |
-
"q_proj",
|
| 33 |
-
"gate_proj",
|
| 34 |
"v_proj",
|
| 35 |
-
"
|
| 36 |
"o_proj",
|
| 37 |
"up_proj",
|
| 38 |
-
"
|
|
|
|
|
|
|
| 39 |
],
|
| 40 |
"target_parameters": null,
|
| 41 |
"task_type": "CAUSAL_LM",
|
|
|
|
| 29 |
"rank_pattern": {},
|
| 30 |
"revision": null,
|
| 31 |
"target_modules": [
|
|
|
|
|
|
|
| 32 |
"v_proj",
|
| 33 |
+
"down_proj",
|
| 34 |
"o_proj",
|
| 35 |
"up_proj",
|
| 36 |
+
"q_proj",
|
| 37 |
+
"k_proj",
|
| 38 |
+
"gate_proj"
|
| 39 |
],
|
| 40 |
"target_parameters": null,
|
| 41 |
"task_type": "CAUSAL_LM",
|
adapter/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 323014560
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47969ecdc8316c40e6f959f6369d54cb33fa5660e1e4073f9f17e03ac0e208bd
|
| 3 |
size 323014560
|
adapter/training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 8760
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5d047cbc55cda44b752f416d36e3a69e56abc0bd3f8f2902b298452f531a4525
|
| 3 |
size 8760
|
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": 0.001
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"total_flos": 0.0,
|
| 3 |
+
"train_loss": 0.01698429927288089,
|
| 4 |
+
"train_runtime": 137940.7556,
|
| 5 |
"train_samples": 7000,
|
| 6 |
+
"train_samples_per_second": 0.07,
|
| 7 |
"train_steps_per_second": 0.001
|
| 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": 0.001
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"total_flos": 0.0,
|
| 3 |
+
"train_loss": 0.01698429927288089,
|
| 4 |
+
"train_runtime": 137940.7556,
|
| 5 |
"train_samples": 7000,
|
| 6 |
+
"train_samples_per_second": 0.07,
|
| 7 |
"train_steps_per_second": 0.001
|
| 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": 0.001
|
| 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": 2261.625045776367,
|
| 1915 |
+
"dapo/avg_reward_std": 0.2656887276419278,
|
| 1916 |
+
"dapo/filter_reward_index": 0.0,
|
| 1917 |
+
"dapo/kept_prompts_ratio": 0.3563218414783478,
|
| 1918 |
+
"dapo/num_sampling_attempts": 3.625,
|
| 1919 |
+
"dapo/sampling_efficiency": 30.952380952380942,
|
| 1920 |
+
"dapo/total_prompts_processed": 21.75,
|
| 1921 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 1922 |
+
"epoch": 0.11542857142857142,
|
| 1923 |
+
"grad_norm": 0.012256976217031479,
|
| 1924 |
+
"kl": 0.0002308487892150879,
|
| 1925 |
+
"learning_rate": 1e-07,
|
| 1926 |
+
"loss": 0.0255,
|
| 1927 |
+
"reward": 0.6794679276645184,
|
| 1928 |
+
"reward_std": 0.936141237616539,
|
| 1929 |
+
"step": 101
|
| 1930 |
+
},
|
| 1931 |
+
{
|
| 1932 |
+
"clip_fraction": 0.0,
|
| 1933 |
+
"completion_length": 2019.2500381469727,
|
| 1934 |
+
"dapo/avg_reward_std": 0.27603574914316975,
|
| 1935 |
+
"dapo/filter_reward_index": 0.0,
|
| 1936 |
+
"dapo/kept_prompts_ratio": 0.40322581414253483,
|
| 1937 |
+
"dapo/num_sampling_attempts": 3.875,
|
| 1938 |
+
"dapo/sampling_efficiency": 31.14583333333333,
|
| 1939 |
+
"dapo/total_prompts_processed": 23.25,
|
| 1940 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 1941 |
+
"epoch": 0.11657142857142858,
|
| 1942 |
+
"grad_norm": 0.011883130297064781,
|
| 1943 |
+
"kl": 0.00018447637557983398,
|
| 1944 |
+
"learning_rate": 6.203955092681039e-07,
|
| 1945 |
+
"loss": 0.0566,
|
| 1946 |
+
"reward": 0.9531724825501442,
|
| 1947 |
+
"reward_std": 0.9424103274941444,
|
| 1948 |
+
"step": 102
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"clip_fraction": 0.0,
|
| 1952 |
+
"completion_length": 2447.142379760742,
|
| 1953 |
+
"dapo/avg_reward_std": 0.2595460871855418,
|
| 1954 |
+
"dapo/filter_reward_index": 0.0,
|
| 1955 |
+
"dapo/kept_prompts_ratio": 0.3833333447575569,
|
| 1956 |
+
"dapo/num_sampling_attempts": 3.75,
|
| 1957 |
+
"dapo/sampling_efficiency": 28.958333333333332,
|
| 1958 |
+
"dapo/total_prompts_processed": 22.5,
|
| 1959 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 1960 |
+
"epoch": 0.11771428571428572,
|
| 1961 |
+
"grad_norm": 0.010165790095925331,
|
| 1962 |
+
"kl": 0.00018906593322753906,
|
| 1963 |
+
"learning_rate": 6.126278954320294e-07,
|
| 1964 |
+
"loss": 0.0361,
|
| 1965 |
+
"reward": 0.8079591542482376,
|
| 1966 |
+
"reward_std": 0.9323313534259796,
|
| 1967 |
+
"step": 103
|
| 1968 |
+
},
|
| 1969 |
+
{
|
| 1970 |
+
"clip_fraction": 0.0,
|
| 1971 |
+
"completion_length": 2414.4305725097656,
|
| 1972 |
+
"dapo/avg_reward_std": 0.2675211922875766,
|
| 1973 |
+
"dapo/filter_reward_index": 0.0,
|
| 1974 |
+
"dapo/kept_prompts_ratio": 0.3505747174394542,
|
| 1975 |
+
"dapo/num_sampling_attempts": 3.625,
|
| 1976 |
+
"dapo/sampling_efficiency": 42.013888888888886,
|
| 1977 |
+
"dapo/total_prompts_processed": 21.75,
|
| 1978 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 1979 |
+
"epoch": 0.11885714285714286,
|
| 1980 |
+
"grad_norm": 0.009563453495502472,
|
| 1981 |
+
"kl": 0.0002244710922241211,
|
| 1982 |
+
"learning_rate": 6.048412045323164e-07,
|
| 1983 |
+
"loss": 0.0367,
|
| 1984 |
+
"reward": 0.6746065132319927,
|
| 1985 |
+
"reward_std": 0.9439321234822273,
|
| 1986 |
+
"step": 104
|
| 1987 |
+
},
|
| 1988 |
+
{
|
| 1989 |
+
"clip_fraction": 0.0,
|
| 1990 |
+
"completion_length": 2350.4653396606445,
|
| 1991 |
+
"dapo/avg_reward_std": 0.2709802109183687,
|
| 1992 |
+
"dapo/filter_reward_index": 0.0,
|
| 1993 |
+
"dapo/kept_prompts_ratio": 0.33838384621071094,
|
| 1994 |
+
"dapo/num_sampling_attempts": 4.125,
|
| 1995 |
+
"dapo/sampling_efficiency": 39.58333333333333,
|
| 1996 |
+
"dapo/total_prompts_processed": 24.75,
|
| 1997 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 1998 |
+
"epoch": 0.12,
|
| 1999 |
+
"grad_norm": 0.011944189667701721,
|
| 2000 |
+
"kl": 0.00023399293422698975,
|
| 2001 |
+
"learning_rate": 5.97037808470444e-07,
|
| 2002 |
+
"loss": 0.0133,
|
| 2003 |
+
"reward": 0.7501634955406189,
|
| 2004 |
+
"reward_std": 0.9493465423583984,
|
| 2005 |
+
"step": 105
|
| 2006 |
+
},
|
| 2007 |
+
{
|
| 2008 |
+
"clip_fraction": 0.0,
|
| 2009 |
+
"completion_length": 2232.5590209960938,
|
| 2010 |
+
"dapo/avg_reward_std": 0.2304972934311834,
|
| 2011 |
+
"dapo/filter_reward_index": 0.0,
|
| 2012 |
+
"dapo/kept_prompts_ratio": 0.32183908234382497,
|
| 2013 |
+
"dapo/num_sampling_attempts": 3.625,
|
| 2014 |
+
"dapo/sampling_efficiency": 46.05654761904762,
|
| 2015 |
+
"dapo/total_prompts_processed": 21.75,
|
| 2016 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2017 |
+
"epoch": 0.12114285714285715,
|
| 2018 |
+
"grad_norm": 0.011308341287076473,
|
| 2019 |
+
"kl": 0.0002868175506591797,
|
| 2020 |
+
"learning_rate": 5.892200842364462e-07,
|
| 2021 |
+
"loss": 0.017,
|
| 2022 |
+
"reward": 0.8449488952755928,
|
| 2023 |
+
"reward_std": 0.9235394075512886,
|
| 2024 |
+
"step": 106
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"clip_fraction": 0.0,
|
| 2028 |
+
"completion_length": 2245.4306259155273,
|
| 2029 |
+
"dapo/avg_reward_std": 0.32372228088586225,
|
| 2030 |
+
"dapo/filter_reward_index": 0.0,
|
| 2031 |
+
"dapo/kept_prompts_ratio": 0.37681160478488257,
|
| 2032 |
+
"dapo/num_sampling_attempts": 2.875,
|
| 2033 |
+
"dapo/sampling_efficiency": 45.53571428571428,
|
| 2034 |
+
"dapo/total_prompts_processed": 17.25,
|
| 2035 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2036 |
+
"epoch": 0.12228571428571429,
|
| 2037 |
+
"grad_norm": 0.01079760491847992,
|
| 2038 |
+
"kl": 0.00018143653869628906,
|
| 2039 |
+
"learning_rate": 5.813904131848564e-07,
|
| 2040 |
+
"loss": 0.0407,
|
| 2041 |
+
"reward": 0.876940418034792,
|
| 2042 |
+
"reward_std": 0.945194236934185,
|
| 2043 |
+
"step": 107
|
| 2044 |
+
},
|
| 2045 |
+
{
|
| 2046 |
+
"clip_fraction": 0.0,
|
| 2047 |
+
"completion_length": 2877.4410095214844,
|
| 2048 |
+
"dapo/avg_reward_std": 0.31851592376118615,
|
| 2049 |
+
"dapo/filter_reward_index": 0.0,
|
| 2050 |
+
"dapo/kept_prompts_ratio": 0.44444445201328825,
|
| 2051 |
+
"dapo/num_sampling_attempts": 2.625,
|
| 2052 |
+
"dapo/sampling_efficiency": 47.916666666666664,
|
| 2053 |
+
"dapo/total_prompts_processed": 15.75,
|
| 2054 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2055 |
+
"epoch": 0.12342857142857143,
|
| 2056 |
+
"grad_norm": 0.008532079868018627,
|
| 2057 |
+
"kl": 0.00026416778564453125,
|
| 2058 |
+
"learning_rate": 5.735511803093248e-07,
|
| 2059 |
+
"loss": 0.0189,
|
| 2060 |
+
"reward": 0.5354619715362787,
|
| 2061 |
+
"reward_std": 0.9343887642025948,
|
| 2062 |
+
"step": 108
|
| 2063 |
+
},
|
| 2064 |
+
{
|
| 2065 |
+
"clip_fraction": 0.0,
|
| 2066 |
+
"completion_length": 2287.3403396606445,
|
| 2067 |
+
"dapo/avg_reward_std": 0.2637126021660291,
|
| 2068 |
+
"dapo/filter_reward_index": 0.0,
|
| 2069 |
+
"dapo/kept_prompts_ratio": 0.41025641560554504,
|
| 2070 |
+
"dapo/num_sampling_attempts": 3.25,
|
| 2071 |
+
"dapo/sampling_efficiency": 44.6875,
|
| 2072 |
+
"dapo/total_prompts_processed": 19.5,
|
| 2073 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2074 |
+
"epoch": 0.12457142857142857,
|
| 2075 |
+
"grad_norm": 0.010662744753062725,
|
| 2076 |
+
"kl": 0.00025856494903564453,
|
| 2077 |
+
"learning_rate": 5.657047735161255e-07,
|
| 2078 |
+
"loss": 0.0139,
|
| 2079 |
+
"reward": 0.6945868469774723,
|
| 2080 |
+
"reward_std": 0.945196196436882,
|
| 2081 |
+
"step": 109
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"clip_fraction": 0.0,
|
| 2085 |
+
"completion_length": 2099.197898864746,
|
| 2086 |
+
"dapo/avg_reward_std": 0.2950383967586926,
|
| 2087 |
+
"dapo/filter_reward_index": 0.0,
|
| 2088 |
+
"dapo/kept_prompts_ratio": 0.380952388048172,
|
| 2089 |
+
"dapo/num_sampling_attempts": 3.5,
|
| 2090 |
+
"dapo/sampling_efficiency": 38.541666666666664,
|
| 2091 |
+
"dapo/total_prompts_processed": 21.0,
|
| 2092 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2093 |
+
"epoch": 0.12571428571428572,
|
| 2094 |
+
"grad_norm": 0.011256680823862553,
|
| 2095 |
+
"kl": 0.0002060532569885254,
|
| 2096 |
+
"learning_rate": 5.578535828967777e-07,
|
| 2097 |
+
"loss": 0.0106,
|
| 2098 |
+
"reward": 0.6000825632363558,
|
| 2099 |
+
"reward_std": 0.9193084537982941,
|
| 2100 |
+
"step": 110
|
| 2101 |
+
},
|
| 2102 |
+
{
|
| 2103 |
+
"clip_fraction": 0.0,
|
| 2104 |
+
"completion_length": 2716.079864501953,
|
| 2105 |
+
"dapo/avg_reward_std": 0.2741352463590688,
|
| 2106 |
+
"dapo/filter_reward_index": 0.0,
|
| 2107 |
+
"dapo/kept_prompts_ratio": 0.3620689691140734,
|
| 2108 |
+
"dapo/num_sampling_attempts": 3.625,
|
| 2109 |
+
"dapo/sampling_efficiency": 41.488095238095234,
|
| 2110 |
+
"dapo/total_prompts_processed": 21.75,
|
| 2111 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2112 |
+
"epoch": 0.12685714285714286,
|
| 2113 |
+
"grad_norm": 0.009062284603714943,
|
| 2114 |
+
"kl": 0.0002288222312927246,
|
| 2115 |
+
"learning_rate": 5.5e-07,
|
| 2116 |
+
"loss": 0.0461,
|
| 2117 |
+
"reward": 0.6751261968165636,
|
| 2118 |
+
"reward_std": 0.9856812655925751,
|
| 2119 |
+
"step": 111
|
| 2120 |
+
},
|
| 2121 |
+
{
|
| 2122 |
+
"clip_fraction": 0.0,
|
| 2123 |
+
"completion_length": 3045.125,
|
| 2124 |
+
"dapo/avg_reward_std": 0.36701818108558654,
|
| 2125 |
+
"dapo/filter_reward_index": 0.0,
|
| 2126 |
+
"dapo/kept_prompts_ratio": 0.5888888945182165,
|
| 2127 |
+
"dapo/num_sampling_attempts": 1.875,
|
| 2128 |
+
"dapo/sampling_efficiency": 65.625,
|
| 2129 |
+
"dapo/total_prompts_processed": 11.25,
|
| 2130 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2131 |
+
"epoch": 0.128,
|
| 2132 |
+
"grad_norm": 0.013615131378173828,
|
| 2133 |
+
"kl": 0.0003104209899902344,
|
| 2134 |
+
"learning_rate": 5.421464171032224e-07,
|
| 2135 |
+
"loss": 0.0541,
|
| 2136 |
+
"reward": 0.6107649356126785,
|
| 2137 |
+
"reward_std": 0.9386496767401695,
|
| 2138 |
+
"step": 112
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"clip_fraction": 0.0,
|
| 2142 |
+
"completion_length": 2261.1597442626953,
|
| 2143 |
+
"dapo/avg_reward_std": 0.2829060518741608,
|
| 2144 |
+
"dapo/filter_reward_index": 0.0,
|
| 2145 |
+
"dapo/kept_prompts_ratio": 0.413333340883255,
|
| 2146 |
+
"dapo/num_sampling_attempts": 3.125,
|
| 2147 |
+
"dapo/sampling_efficiency": 41.041666666666664,
|
| 2148 |
+
"dapo/total_prompts_processed": 18.75,
|
| 2149 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2150 |
+
"epoch": 0.12914285714285714,
|
| 2151 |
+
"grad_norm": 0.01245199330151081,
|
| 2152 |
+
"kl": 0.0002949833869934082,
|
| 2153 |
+
"learning_rate": 5.342952264838747e-07,
|
| 2154 |
+
"loss": 0.0273,
|
| 2155 |
+
"reward": 0.7544166818261147,
|
| 2156 |
+
"reward_std": 0.9633913785219193,
|
| 2157 |
+
"step": 113
|
| 2158 |
+
},
|
| 2159 |
+
{
|
| 2160 |
+
"clip_fraction": 0.0,
|
| 2161 |
+
"completion_length": 2030.1041946411133,
|
| 2162 |
+
"dapo/avg_reward_std": 0.2660287490912846,
|
| 2163 |
+
"dapo/filter_reward_index": 0.0,
|
| 2164 |
+
"dapo/kept_prompts_ratio": 0.32738095788019045,
|
| 2165 |
+
"dapo/num_sampling_attempts": 3.5,
|
| 2166 |
+
"dapo/sampling_efficiency": 45.535714285714285,
|
| 2167 |
+
"dapo/total_prompts_processed": 21.0,
|
| 2168 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2169 |
+
"epoch": 0.13028571428571428,
|
| 2170 |
+
"grad_norm": 0.011100664734840393,
|
| 2171 |
+
"kl": 0.00016885995864868164,
|
| 2172 |
+
"learning_rate": 5.264488196906752e-07,
|
| 2173 |
+
"loss": 0.0649,
|
| 2174 |
+
"reward": 0.5986752398312092,
|
| 2175 |
+
"reward_std": 0.9739916548132896,
|
| 2176 |
+
"step": 114
|
| 2177 |
+
},
|
| 2178 |
+
{
|
| 2179 |
+
"clip_fraction": 0.0,
|
| 2180 |
+
"completion_length": 2791.465301513672,
|
| 2181 |
+
"dapo/avg_reward_std": 0.297807412147522,
|
| 2182 |
+
"dapo/filter_reward_index": 0.0,
|
| 2183 |
+
"dapo/kept_prompts_ratio": 0.44000000655651095,
|
| 2184 |
+
"dapo/num_sampling_attempts": 3.125,
|
| 2185 |
+
"dapo/sampling_efficiency": 42.08333333333333,
|
| 2186 |
+
"dapo/total_prompts_processed": 18.75,
|
| 2187 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2188 |
+
"epoch": 0.13142857142857142,
|
| 2189 |
+
"grad_norm": 0.011278674006462097,
|
| 2190 |
+
"kl": 0.0002925395965576172,
|
| 2191 |
+
"learning_rate": 5.186095868151436e-07,
|
| 2192 |
+
"loss": 0.0586,
|
| 2193 |
+
"reward": 0.6219565980136395,
|
| 2194 |
+
"reward_std": 0.9591977074742317,
|
| 2195 |
+
"step": 115
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"clip_fraction": 0.0,
|
| 2199 |
+
"completion_length": 2804.6493606567383,
|
| 2200 |
+
"dapo/avg_reward_std": 0.35703572371731634,
|
| 2201 |
+
"dapo/filter_reward_index": 0.0,
|
| 2202 |
+
"dapo/kept_prompts_ratio": 0.42028986371081806,
|
| 2203 |
+
"dapo/num_sampling_attempts": 2.875,
|
| 2204 |
+
"dapo/sampling_efficiency": 43.12499999999999,
|
| 2205 |
+
"dapo/total_prompts_processed": 17.25,
|
| 2206 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2207 |
+
"epoch": 0.13257142857142856,
|
| 2208 |
+
"grad_norm": 0.01122019812464714,
|
| 2209 |
+
"kl": 0.00034046173095703125,
|
| 2210 |
+
"learning_rate": 5.107799157635538e-07,
|
| 2211 |
+
"loss": 0.0233,
|
| 2212 |
+
"reward": 0.469740716740489,
|
| 2213 |
+
"reward_std": 0.9214994236826897,
|
| 2214 |
+
"step": 116
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"clip_fraction": 0.0,
|
| 2218 |
+
"completion_length": 2037.885456085205,
|
| 2219 |
+
"dapo/avg_reward_std": 0.30805256009101867,
|
| 2220 |
+
"dapo/filter_reward_index": 0.0,
|
| 2221 |
+
"dapo/kept_prompts_ratio": 0.4666666799783707,
|
| 2222 |
+
"dapo/num_sampling_attempts": 3.125,
|
| 2223 |
+
"dapo/sampling_efficiency": 44.27083333333333,
|
| 2224 |
+
"dapo/total_prompts_processed": 18.75,
|
| 2225 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2226 |
+
"epoch": 0.1337142857142857,
|
| 2227 |
+
"grad_norm": 0.014056873507797718,
|
| 2228 |
+
"kl": 0.0002486705780029297,
|
| 2229 |
+
"learning_rate": 5.02962191529556e-07,
|
| 2230 |
+
"loss": 0.038,
|
| 2231 |
+
"reward": 0.9076524265110493,
|
| 2232 |
+
"reward_std": 0.9655390456318855,
|
| 2233 |
+
"step": 117
|
| 2234 |
+
},
|
| 2235 |
+
{
|
| 2236 |
+
"clip_fraction": 0.0,
|
| 2237 |
+
"completion_length": 2517.215316772461,
|
| 2238 |
+
"dapo/avg_reward_std": 0.23199922059263503,
|
| 2239 |
+
"dapo/filter_reward_index": 0.0,
|
| 2240 |
+
"dapo/kept_prompts_ratio": 0.3571428635290691,
|
| 2241 |
+
"dapo/num_sampling_attempts": 3.5,
|
| 2242 |
+
"dapo/sampling_efficiency": 45.535714285714285,
|
| 2243 |
+
"dapo/total_prompts_processed": 21.0,
|
| 2244 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2245 |
+
"epoch": 0.13485714285714287,
|
| 2246 |
+
"grad_norm": 0.011827170848846436,
|
| 2247 |
+
"kl": 0.00034999847412109375,
|
| 2248 |
+
"learning_rate": 4.951587954676837e-07,
|
| 2249 |
+
"loss": 0.023,
|
| 2250 |
+
"reward": 0.5725362580269575,
|
| 2251 |
+
"reward_std": 0.9489376917481422,
|
| 2252 |
+
"step": 118
|
| 2253 |
+
},
|
| 2254 |
+
{
|
| 2255 |
+
"clip_fraction": 0.0,
|
| 2256 |
+
"completion_length": 2309.763916015625,
|
| 2257 |
+
"dapo/avg_reward_std": 0.33521059803340747,
|
| 2258 |
+
"dapo/filter_reward_index": 0.0,
|
| 2259 |
+
"dapo/kept_prompts_ratio": 0.48550726084605506,
|
| 2260 |
+
"dapo/num_sampling_attempts": 2.875,
|
| 2261 |
+
"dapo/sampling_efficiency": 44.166666666666664,
|
| 2262 |
+
"dapo/total_prompts_processed": 17.25,
|
| 2263 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2264 |
+
"epoch": 0.136,
|
| 2265 |
+
"grad_norm": 0.014920210465788841,
|
| 2266 |
+
"kl": 0.0003477334976196289,
|
| 2267 |
+
"learning_rate": 4.873721045679706e-07,
|
| 2268 |
+
"loss": 0.0967,
|
| 2269 |
+
"reward": 0.7152486853301525,
|
| 2270 |
+
"reward_std": 0.9450967088341713,
|
| 2271 |
+
"step": 119
|
| 2272 |
+
},
|
| 2273 |
+
{
|
| 2274 |
+
"clip_fraction": 0.0,
|
| 2275 |
+
"completion_length": 2588.423629760742,
|
| 2276 |
+
"dapo/avg_reward_std": 0.322092001636823,
|
| 2277 |
+
"dapo/filter_reward_index": 0.0,
|
| 2278 |
+
"dapo/kept_prompts_ratio": 0.4236111156642437,
|
| 2279 |
+
"dapo/num_sampling_attempts": 3.0,
|
| 2280 |
+
"dapo/sampling_efficiency": 40.62499999999999,
|
| 2281 |
+
"dapo/total_prompts_processed": 18.0,
|
| 2282 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2283 |
+
"epoch": 0.13714285714285715,
|
| 2284 |
+
"grad_norm": 0.009259553626179695,
|
| 2285 |
+
"kl": 0.0002923011779785156,
|
| 2286 |
+
"learning_rate": 4.79604490731896e-07,
|
| 2287 |
+
"loss": 0.0204,
|
| 2288 |
+
"reward": 0.5492150112986565,
|
| 2289 |
+
"reward_std": 0.9336576908826828,
|
| 2290 |
+
"step": 120
|
| 2291 |
+
},
|
| 2292 |
+
{
|
| 2293 |
+
"clip_fraction": 0.0,
|
| 2294 |
+
"completion_length": 2224.4583282470703,
|
| 2295 |
+
"dapo/avg_reward_std": 0.2846992796375638,
|
| 2296 |
+
"dapo/filter_reward_index": 0.0,
|
| 2297 |
+
"dapo/kept_prompts_ratio": 0.4603174633923031,
|
| 2298 |
+
"dapo/num_sampling_attempts": 2.625,
|
| 2299 |
+
"dapo/sampling_efficiency": 51.45833333333333,
|
| 2300 |
+
"dapo/total_prompts_processed": 15.75,
|
| 2301 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2302 |
+
"epoch": 0.1382857142857143,
|
| 2303 |
+
"grad_norm": 0.012681272812187672,
|
| 2304 |
+
"kl": 0.0002828836441040039,
|
| 2305 |
+
"learning_rate": 4.7185832004988133e-07,
|
| 2306 |
+
"loss": 0.084,
|
| 2307 |
+
"reward": 0.8260641098022461,
|
| 2308 |
+
"reward_std": 0.9569381102919579,
|
| 2309 |
+
"step": 121
|
| 2310 |
+
},
|
| 2311 |
+
{
|
| 2312 |
+
"clip_fraction": 0.0,
|
| 2313 |
+
"completion_length": 2042.4618453979492,
|
| 2314 |
+
"dapo/avg_reward_std": 0.21980997684754824,
|
| 2315 |
+
"dapo/filter_reward_index": 0.0,
|
| 2316 |
+
"dapo/kept_prompts_ratio": 0.2500000039213582,
|
| 2317 |
+
"dapo/num_sampling_attempts": 4.75,
|
| 2318 |
+
"dapo/sampling_efficiency": 32.916666666666664,
|
| 2319 |
+
"dapo/total_prompts_processed": 28.5,
|
| 2320 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2321 |
+
"epoch": 0.13942857142857143,
|
| 2322 |
+
"grad_norm": 0.014447196386754513,
|
| 2323 |
+
"kl": 0.0002307891845703125,
|
| 2324 |
+
"learning_rate": 4.641359520805548e-07,
|
| 2325 |
+
"loss": 0.0797,
|
| 2326 |
+
"reward": 0.5401283344253898,
|
| 2327 |
+
"reward_std": 0.8589324243366718,
|
| 2328 |
+
"step": 122
|
| 2329 |
+
},
|
| 2330 |
+
{
|
| 2331 |
+
"clip_fraction": 0.0,
|
| 2332 |
+
"completion_length": 1821.270851135254,
|
| 2333 |
+
"dapo/avg_reward_std": 0.30661167701085407,
|
| 2334 |
+
"dapo/filter_reward_index": 0.0,
|
| 2335 |
+
"dapo/kept_prompts_ratio": 0.4652777910232544,
|
| 2336 |
+
"dapo/num_sampling_attempts": 3.0,
|
| 2337 |
+
"dapo/sampling_efficiency": 51.979166666666664,
|
| 2338 |
+
"dapo/total_prompts_processed": 18.0,
|
| 2339 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2340 |
+
"epoch": 0.14057142857142857,
|
| 2341 |
+
"grad_norm": 0.012464089319109917,
|
| 2342 |
+
"kl": 0.00021943449974060059,
|
| 2343 |
+
"learning_rate": 4.5643973913200837e-07,
|
| 2344 |
+
"loss": 0.0524,
|
| 2345 |
+
"reward": 0.7340994998812675,
|
| 2346 |
+
"reward_std": 0.948041707277298,
|
| 2347 |
+
"step": 123
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"clip_fraction": 0.0,
|
| 2351 |
+
"completion_length": 1917.8576431274414,
|
| 2352 |
+
"dapo/avg_reward_std": 0.26710175829274313,
|
| 2353 |
+
"dapo/filter_reward_index": 0.0,
|
| 2354 |
+
"dapo/kept_prompts_ratio": 0.3452381023338863,
|
| 2355 |
+
"dapo/num_sampling_attempts": 3.5,
|
| 2356 |
+
"dapo/sampling_efficiency": 35.83333333333333,
|
| 2357 |
+
"dapo/total_prompts_processed": 21.0,
|
| 2358 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2359 |
+
"epoch": 0.1417142857142857,
|
| 2360 |
+
"grad_norm": 0.01295761950314045,
|
| 2361 |
+
"kl": 0.00027441978454589844,
|
| 2362 |
+
"learning_rate": 4.4877202554526084e-07,
|
| 2363 |
+
"loss": 0.0395,
|
| 2364 |
+
"reward": 0.44990649446845055,
|
| 2365 |
+
"reward_std": 0.9298848733305931,
|
| 2366 |
+
"step": 124
|
| 2367 |
+
},
|
| 2368 |
+
{
|
| 2369 |
+
"clip_fraction": 0.0,
|
| 2370 |
+
"completion_length": 2151.381965637207,
|
| 2371 |
+
"dapo/avg_reward_std": 0.2770674500776374,
|
| 2372 |
+
"dapo/filter_reward_index": 0.0,
|
| 2373 |
+
"dapo/kept_prompts_ratio": 0.4347826164701711,
|
| 2374 |
+
"dapo/num_sampling_attempts": 2.875,
|
| 2375 |
+
"dapo/sampling_efficiency": 55.729166666666664,
|
| 2376 |
+
"dapo/total_prompts_processed": 17.25,
|
| 2377 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2378 |
+
"epoch": 0.14285714285714285,
|
| 2379 |
+
"grad_norm": 0.011758905835449696,
|
| 2380 |
+
"kl": 0.00027173757553100586,
|
| 2381 |
+
"learning_rate": 4.4113514698014953e-07,
|
| 2382 |
+
"loss": -0.0284,
|
| 2383 |
+
"reward": 0.5582777298986912,
|
| 2384 |
+
"reward_std": 0.9428363367915154,
|
| 2385 |
+
"step": 125
|
| 2386 |
+
},
|
| 2387 |
+
{
|
| 2388 |
+
"clip_fraction": 0.0,
|
| 2389 |
+
"completion_length": 1810.0104522705078,
|
| 2390 |
+
"dapo/avg_reward_std": 0.21576001878940698,
|
| 2391 |
+
"dapo/filter_reward_index": 0.0,
|
| 2392 |
+
"dapo/kept_prompts_ratio": 0.2777777860562007,
|
| 2393 |
+
"dapo/num_sampling_attempts": 4.125,
|
| 2394 |
+
"dapo/sampling_efficiency": 36.45833333333333,
|
| 2395 |
+
"dapo/total_prompts_processed": 24.75,
|
| 2396 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2397 |
+
"epoch": 0.144,
|
| 2398 |
+
"grad_norm": 0.011465526185929775,
|
| 2399 |
+
"kl": 0.00021535158157348633,
|
| 2400 |
+
"learning_rate": 4.3353142970386557e-07,
|
| 2401 |
+
"loss": -0.0108,
|
| 2402 |
+
"reward": 0.6622855560854077,
|
| 2403 |
+
"reward_std": 0.9075748100876808,
|
| 2404 |
+
"step": 126
|
| 2405 |
+
},
|
| 2406 |
+
{
|
| 2407 |
+
"clip_fraction": 0.0,
|
| 2408 |
+
"completion_length": 2243.732635498047,
|
| 2409 |
+
"dapo/avg_reward_std": 0.2920382275031163,
|
| 2410 |
+
"dapo/filter_reward_index": 0.0,
|
| 2411 |
+
"dapo/kept_prompts_ratio": 0.3782051377571546,
|
| 2412 |
+
"dapo/num_sampling_attempts": 3.25,
|
| 2413 |
+
"dapo/sampling_efficiency": 32.291666666666664,
|
| 2414 |
+
"dapo/total_prompts_processed": 19.5,
|
| 2415 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2416 |
+
"epoch": 0.14514285714285713,
|
| 2417 |
+
"grad_norm": 0.01050955057144165,
|
| 2418 |
+
"kl": 0.00027239322662353516,
|
| 2419 |
+
"learning_rate": 4.2596318988235037e-07,
|
| 2420 |
+
"loss": 0.0464,
|
| 2421 |
+
"reward": 0.533456489443779,
|
| 2422 |
+
"reward_std": 0.9191579967737198,
|
| 2423 |
+
"step": 127
|
| 2424 |
+
},
|
| 2425 |
+
{
|
| 2426 |
+
"clip_fraction": 0.0,
|
| 2427 |
+
"completion_length": 2544.6875,
|
| 2428 |
+
"dapo/avg_reward_std": 0.27494730835869197,
|
| 2429 |
+
"dapo/filter_reward_index": 0.0,
|
| 2430 |
+
"dapo/kept_prompts_ratio": 0.46031746906893595,
|
| 2431 |
+
"dapo/num_sampling_attempts": 2.625,
|
| 2432 |
+
"dapo/sampling_efficiency": 46.24999999999999,
|
| 2433 |
+
"dapo/total_prompts_processed": 15.75,
|
| 2434 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2435 |
+
"epoch": 0.1462857142857143,
|
| 2436 |
+
"grad_norm": 0.017111552879214287,
|
| 2437 |
+
"kl": 0.0003796815872192383,
|
| 2438 |
+
"learning_rate": 4.1843273287476854e-07,
|
| 2439 |
+
"loss": 0.0784,
|
| 2440 |
+
"reward": 0.7016365043818951,
|
| 2441 |
+
"reward_std": 0.986565351486206,
|
| 2442 |
+
"step": 128
|
| 2443 |
+
},
|
| 2444 |
+
{
|
| 2445 |
+
"clip_fraction": 0.0,
|
| 2446 |
+
"completion_length": 2367.3646240234375,
|
| 2447 |
+
"dapo/avg_reward_std": 0.23454804884062874,
|
| 2448 |
+
"dapo/filter_reward_index": 0.0,
|
| 2449 |
+
"dapo/kept_prompts_ratio": 0.34567901988824207,
|
| 2450 |
+
"dapo/num_sampling_attempts": 3.375,
|
| 2451 |
+
"dapo/sampling_efficiency": 47.222222222222214,
|
| 2452 |
+
"dapo/total_prompts_processed": 20.25,
|
| 2453 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2454 |
+
"epoch": 0.14742857142857144,
|
| 2455 |
+
"grad_norm": 0.009941876865923405,
|
| 2456 |
+
"kl": 0.00034287571907043457,
|
| 2457 |
+
"learning_rate": 4.1094235253127374e-07,
|
| 2458 |
+
"loss": 0.0061,
|
| 2459 |
+
"reward": 0.7930427435785532,
|
| 2460 |
+
"reward_std": 0.9500019550323486,
|
| 2461 |
+
"step": 129
|
| 2462 |
+
},
|
| 2463 |
+
{
|
| 2464 |
+
"clip_fraction": 0.0,
|
| 2465 |
+
"completion_length": 2704.125045776367,
|
| 2466 |
+
"dapo/avg_reward_std": 0.33850957382292973,
|
| 2467 |
+
"dapo/filter_reward_index": 0.0,
|
| 2468 |
+
"dapo/kept_prompts_ratio": 0.5158730284089134,
|
| 2469 |
+
"dapo/num_sampling_attempts": 2.625,
|
| 2470 |
+
"dapo/sampling_efficiency": 47.291666666666664,
|
| 2471 |
+
"dapo/total_prompts_processed": 15.75,
|
| 2472 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2473 |
+
"epoch": 0.14857142857142858,
|
| 2474 |
+
"grad_norm": 0.012839604169130325,
|
| 2475 |
+
"kl": 0.0004292726516723633,
|
| 2476 |
+
"learning_rate": 4.034943304942796e-07,
|
| 2477 |
+
"loss": 0.0353,
|
| 2478 |
+
"reward": 0.6285950914025307,
|
| 2479 |
+
"reward_std": 0.9615181535482407,
|
| 2480 |
+
"step": 130
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"clip_fraction": 0.0,
|
| 2484 |
+
"completion_length": 2475.5104370117188,
|
| 2485 |
+
"dapo/avg_reward_std": 0.30972740189595654,
|
| 2486 |
+
"dapo/filter_reward_index": 0.0,
|
| 2487 |
+
"dapo/kept_prompts_ratio": 0.4696969762444496,
|
| 2488 |
+
"dapo/num_sampling_attempts": 2.75,
|
| 2489 |
+
"dapo/sampling_efficiency": 52.82738095238095,
|
| 2490 |
+
"dapo/total_prompts_processed": 16.5,
|
| 2491 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2492 |
+
"epoch": 0.14971428571428572,
|
| 2493 |
+
"grad_norm": 0.00980184692889452,
|
| 2494 |
+
"kl": 0.00036525726318359375,
|
| 2495 |
+
"learning_rate": 3.9609093550344907e-07,
|
| 2496 |
+
"loss": -0.0176,
|
| 2497 |
+
"reward": 0.7969067245721817,
|
| 2498 |
+
"reward_std": 0.9501392021775246,
|
| 2499 |
+
"step": 131
|
| 2500 |
+
},
|
| 2501 |
+
{
|
| 2502 |
+
"clip_fraction": 0.0,
|
| 2503 |
+
"completion_length": 2550.0034790039062,
|
| 2504 |
+
"dapo/avg_reward_std": 0.2723326214722225,
|
| 2505 |
+
"dapo/filter_reward_index": 0.0,
|
| 2506 |
+
"dapo/kept_prompts_ratio": 0.30952381661960054,
|
| 2507 |
+
"dapo/num_sampling_attempts": 3.5,
|
| 2508 |
+
"dapo/sampling_efficiency": 29.999999999999993,
|
| 2509 |
+
"dapo/total_prompts_processed": 21.0,
|
| 2510 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2511 |
+
"epoch": 0.15085714285714286,
|
| 2512 |
+
"grad_norm": 0.017491888254880905,
|
| 2513 |
+
"kl": 0.00042629241943359375,
|
| 2514 |
+
"learning_rate": 3.8873442270461485e-07,
|
| 2515 |
+
"loss": 0.0772,
|
| 2516 |
+
"reward": 0.6202478259801865,
|
| 2517 |
+
"reward_std": 0.9556004330515862,
|
| 2518 |
+
"step": 132
|
| 2519 |
+
},
|
| 2520 |
+
{
|
| 2521 |
+
"clip_fraction": 0.0,
|
| 2522 |
+
"completion_length": 2056.5000534057617,
|
| 2523 |
+
"dapo/avg_reward_std": 0.3334644228219986,
|
| 2524 |
+
"dapo/filter_reward_index": 0.0,
|
| 2525 |
+
"dapo/kept_prompts_ratio": 0.5333333447575569,
|
| 2526 |
+
"dapo/num_sampling_attempts": 2.5,
|
| 2527 |
+
"dapo/sampling_efficiency": 47.916666666666664,
|
| 2528 |
+
"dapo/total_prompts_processed": 15.0,
|
| 2529 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2530 |
+
"epoch": 0.152,
|
| 2531 |
+
"grad_norm": 0.012553170323371887,
|
| 2532 |
+
"kl": 0.0004407167434692383,
|
| 2533 |
+
"learning_rate": 3.8142703296283953e-07,
|
| 2534 |
+
"loss": -0.0185,
|
| 2535 |
+
"reward": 0.7429189011454582,
|
| 2536 |
+
"reward_std": 1.0187850967049599,
|
| 2537 |
+
"step": 133
|
| 2538 |
+
},
|
| 2539 |
+
{
|
| 2540 |
+
"clip_fraction": 0.0,
|
| 2541 |
+
"completion_length": 2446.9270935058594,
|
| 2542 |
+
"dapo/avg_reward_std": 0.28044995562783603,
|
| 2543 |
+
"dapo/filter_reward_index": 0.0,
|
| 2544 |
+
"dapo/kept_prompts_ratio": 0.32758621254871634,
|
| 2545 |
+
"dapo/num_sampling_attempts": 3.625,
|
| 2546 |
+
"dapo/sampling_efficiency": 32.41071428571428,
|
| 2547 |
+
"dapo/total_prompts_processed": 21.75,
|
| 2548 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2549 |
+
"epoch": 0.15314285714285714,
|
| 2550 |
+
"grad_norm": 0.010931877419352531,
|
| 2551 |
+
"kl": 0.00035947561264038086,
|
| 2552 |
+
"learning_rate": 3.7417099217982686e-07,
|
| 2553 |
+
"loss": 0.0372,
|
| 2554 |
+
"reward": 0.6385626457631588,
|
| 2555 |
+
"reward_std": 0.9372833296656609,
|
| 2556 |
+
"step": 134
|
| 2557 |
+
},
|
| 2558 |
+
{
|
| 2559 |
+
"clip_fraction": 0.0,
|
| 2560 |
+
"completion_length": 2216.4479370117188,
|
| 2561 |
+
"dapo/avg_reward_std": 0.3021530819435914,
|
| 2562 |
+
"dapo/filter_reward_index": 0.0,
|
| 2563 |
+
"dapo/kept_prompts_ratio": 0.40972223070760566,
|
| 2564 |
+
"dapo/num_sampling_attempts": 3.0,
|
| 2565 |
+
"dapo/sampling_efficiency": 47.291666666666664,
|
| 2566 |
+
"dapo/total_prompts_processed": 18.0,
|
| 2567 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2568 |
+
"epoch": 0.15428571428571428,
|
| 2569 |
+
"grad_norm": 0.014175104908645153,
|
| 2570 |
+
"kl": 0.00040656328201293945,
|
| 2571 |
+
"learning_rate": 3.6696851061588994e-07,
|
| 2572 |
+
"loss": 0.0637,
|
| 2573 |
+
"reward": 0.6612157337367535,
|
| 2574 |
+
"reward_std": 0.9335344135761261,
|
| 2575 |
+
"step": 135
|
| 2576 |
+
},
|
| 2577 |
+
{
|
| 2578 |
+
"clip_fraction": 0.0,
|
| 2579 |
+
"completion_length": 1879.3889083862305,
|
| 2580 |
+
"dapo/avg_reward_std": 0.24328701660550875,
|
| 2581 |
+
"dapo/filter_reward_index": 0.0,
|
| 2582 |
+
"dapo/kept_prompts_ratio": 0.36206897219707224,
|
| 2583 |
+
"dapo/num_sampling_attempts": 3.625,
|
| 2584 |
+
"dapo/sampling_efficiency": 31.77083333333333,
|
| 2585 |
+
"dapo/total_prompts_processed": 21.75,
|
| 2586 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2587 |
+
"epoch": 0.15542857142857142,
|
| 2588 |
+
"grad_norm": 0.011906570754945278,
|
| 2589 |
+
"kl": 0.0002865791320800781,
|
| 2590 |
+
"learning_rate": 3.5982178221668533e-07,
|
| 2591 |
+
"loss": 0.0254,
|
| 2592 |
+
"reward": 0.621966740116477,
|
| 2593 |
+
"reward_std": 0.9788949191570282,
|
| 2594 |
+
"step": 136
|
| 2595 |
+
},
|
| 2596 |
+
{
|
| 2597 |
+
"clip_fraction": 0.0,
|
| 2598 |
+
"completion_length": 2137.1180725097656,
|
| 2599 |
+
"dapo/avg_reward_std": 0.19872227481433324,
|
| 2600 |
+
"dapo/filter_reward_index": 0.0,
|
| 2601 |
+
"dapo/kept_prompts_ratio": 0.2857142903975078,
|
| 2602 |
+
"dapo/num_sampling_attempts": 4.375,
|
| 2603 |
+
"dapo/sampling_efficiency": 26.180555555555557,
|
| 2604 |
+
"dapo/total_prompts_processed": 26.25,
|
| 2605 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2606 |
+
"epoch": 0.15657142857142858,
|
| 2607 |
+
"grad_norm": 0.011921432800590992,
|
| 2608 |
+
"kl": 0.000335007905960083,
|
| 2609 |
+
"learning_rate": 3.5273298394491515e-07,
|
| 2610 |
+
"loss": 0.0425,
|
| 2611 |
+
"reward": 0.8858193010091782,
|
| 2612 |
+
"reward_std": 0.960886999964714,
|
| 2613 |
+
"step": 137
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"clip_fraction": 0.0,
|
| 2617 |
+
"completion_length": 2133.6875381469727,
|
| 2618 |
+
"dapo/avg_reward_std": 0.30985672700972783,
|
| 2619 |
+
"dapo/filter_reward_index": 0.0,
|
| 2620 |
+
"dapo/kept_prompts_ratio": 0.5079365216550373,
|
| 2621 |
+
"dapo/num_sampling_attempts": 2.625,
|
| 2622 |
+
"dapo/sampling_efficiency": 50.41666666666666,
|
| 2623 |
+
"dapo/total_prompts_processed": 15.75,
|
| 2624 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2625 |
+
"epoch": 0.15771428571428572,
|
| 2626 |
+
"grad_norm": 0.011959100142121315,
|
| 2627 |
+
"kl": 0.00031131505966186523,
|
| 2628 |
+
"learning_rate": 3.45704275117204e-07,
|
| 2629 |
+
"loss": 0.0473,
|
| 2630 |
+
"reward": 0.8114638328552246,
|
| 2631 |
+
"reward_std": 0.9208285436034203,
|
| 2632 |
+
"step": 138
|
| 2633 |
+
},
|
| 2634 |
+
{
|
| 2635 |
+
"clip_fraction": 0.0,
|
| 2636 |
+
"completion_length": 1887.6423797607422,
|
| 2637 |
+
"dapo/avg_reward_std": 0.30113077312707903,
|
| 2638 |
+
"dapo/filter_reward_index": 0.0,
|
| 2639 |
+
"dapo/kept_prompts_ratio": 0.45833334401249887,
|
| 2640 |
+
"dapo/num_sampling_attempts": 2.5,
|
| 2641 |
+
"dapo/sampling_efficiency": 47.916666666666664,
|
| 2642 |
+
"dapo/total_prompts_processed": 15.0,
|
| 2643 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2644 |
+
"epoch": 0.15885714285714286,
|
| 2645 |
+
"grad_norm": 0.01248843315988779,
|
| 2646 |
+
"kl": 0.0003123283386230469,
|
| 2647 |
+
"learning_rate": 3.387377967463493e-07,
|
| 2648 |
+
"loss": 0.0133,
|
| 2649 |
+
"reward": 0.4802711680531502,
|
| 2650 |
+
"reward_std": 0.9749159812927246,
|
| 2651 |
+
"step": 139
|
| 2652 |
+
},
|
| 2653 |
+
{
|
| 2654 |
+
"clip_fraction": 0.0,
|
| 2655 |
+
"completion_length": 1555.0173835754395,
|
| 2656 |
+
"dapo/avg_reward_std": 0.2354262595375379,
|
| 2657 |
+
"dapo/filter_reward_index": 0.0,
|
| 2658 |
+
"dapo/kept_prompts_ratio": 0.3750000074505806,
|
| 2659 |
+
"dapo/num_sampling_attempts": 3.0,
|
| 2660 |
+
"dapo/sampling_efficiency": 58.05555555555555,
|
| 2661 |
+
"dapo/total_prompts_processed": 18.0,
|
| 2662 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2663 |
+
"epoch": 0.16,
|
| 2664 |
+
"grad_norm": 0.01651233807206154,
|
| 2665 |
+
"kl": 0.0003075599670410156,
|
| 2666 |
+
"learning_rate": 3.3183567088914833e-07,
|
| 2667 |
+
"loss": -0.0302,
|
| 2668 |
+
"reward": 0.8893436007201672,
|
| 2669 |
+
"reward_std": 0.9632327631115913,
|
| 2670 |
+
"step": 140
|
| 2671 |
+
},
|
| 2672 |
+
{
|
| 2673 |
+
"clip_fraction": 0.0,
|
| 2674 |
+
"completion_length": 2886.878517150879,
|
| 2675 |
+
"dapo/avg_reward_std": 0.2881770460378556,
|
| 2676 |
+
"dapo/filter_reward_index": 0.0,
|
| 2677 |
+
"dapo/kept_prompts_ratio": 0.48412698933056425,
|
| 2678 |
+
"dapo/num_sampling_attempts": 2.625,
|
| 2679 |
+
"dapo/sampling_efficiency": 53.12499999999999,
|
| 2680 |
+
"dapo/total_prompts_processed": 15.75,
|
| 2681 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2682 |
+
"epoch": 0.16114285714285714,
|
| 2683 |
+
"grad_norm": 0.010870919562876225,
|
| 2684 |
+
"kl": 0.0004533529281616211,
|
| 2685 |
+
"learning_rate": 3.250000000000001e-07,
|
| 2686 |
+
"loss": 0.0545,
|
| 2687 |
+
"reward": 0.612054293975234,
|
| 2688 |
+
"reward_std": 0.9482586532831192,
|
| 2689 |
+
"step": 141
|
| 2690 |
+
},
|
| 2691 |
+
{
|
| 2692 |
+
"clip_fraction": 0.0,
|
| 2693 |
+
"completion_length": 1937.1458740234375,
|
| 2694 |
+
"dapo/avg_reward_std": 0.3629945723906807,
|
| 2695 |
+
"dapo/filter_reward_index": 0.0,
|
| 2696 |
+
"dapo/kept_prompts_ratio": 0.47826088057911914,
|
| 2697 |
+
"dapo/num_sampling_attempts": 2.875,
|
| 2698 |
+
"dapo/sampling_efficiency": 47.39583333333333,
|
| 2699 |
+
"dapo/total_prompts_processed": 17.25,
|
| 2700 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2701 |
+
"epoch": 0.16228571428571428,
|
| 2702 |
+
"grad_norm": 0.011180016212165356,
|
| 2703 |
+
"kl": 0.00029921531677246094,
|
| 2704 |
+
"learning_rate": 3.182328662904756e-07,
|
| 2705 |
+
"loss": -0.0113,
|
| 2706 |
+
"reward": 0.6175431702286005,
|
| 2707 |
+
"reward_std": 0.9589766189455986,
|
| 2708 |
+
"step": 142
|
| 2709 |
+
},
|
| 2710 |
+
{
|
| 2711 |
+
"clip_fraction": 0.0,
|
| 2712 |
+
"completion_length": 2465.3159942626953,
|
| 2713 |
+
"dapo/avg_reward_std": 0.3803708272821763,
|
| 2714 |
+
"dapo/filter_reward_index": 0.0,
|
| 2715 |
+
"dapo/kept_prompts_ratio": 0.5588235381771537,
|
| 2716 |
+
"dapo/num_sampling_attempts": 2.125,
|
| 2717 |
+
"dapo/sampling_efficiency": 60.41666666666666,
|
| 2718 |
+
"dapo/total_prompts_processed": 12.75,
|
| 2719 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2720 |
+
"epoch": 0.16342857142857142,
|
| 2721 |
+
"grad_norm": 0.012431374751031399,
|
| 2722 |
+
"kl": 0.00048232078552246094,
|
| 2723 |
+
"learning_rate": 3.115363310950578e-07,
|
| 2724 |
+
"loss": 0.0679,
|
| 2725 |
+
"reward": 0.7579541122540832,
|
| 2726 |
+
"reward_std": 0.9723308756947517,
|
| 2727 |
+
"step": 143
|
| 2728 |
+
},
|
| 2729 |
+
{
|
| 2730 |
+
"clip_fraction": 0.0,
|
| 2731 |
+
"completion_length": 2344.1180419921875,
|
| 2732 |
+
"dapo/avg_reward_std": 0.21175828889796608,
|
| 2733 |
+
"dapo/filter_reward_index": 0.0,
|
| 2734 |
+
"dapo/kept_prompts_ratio": 0.26315790179528686,
|
| 2735 |
+
"dapo/num_sampling_attempts": 4.75,
|
| 2736 |
+
"dapo/sampling_efficiency": 25.347222222222218,
|
| 2737 |
+
"dapo/total_prompts_processed": 28.5,
|
| 2738 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2739 |
+
"epoch": 0.16457142857142856,
|
| 2740 |
+
"grad_norm": 0.010860033333301544,
|
| 2741 |
+
"kl": 0.0004448890686035156,
|
| 2742 |
+
"learning_rate": 3.0491243424323783e-07,
|
| 2743 |
+
"loss": -0.0005,
|
| 2744 |
+
"reward": 0.5643926626071334,
|
| 2745 |
+
"reward_std": 0.9328553825616837,
|
| 2746 |
+
"step": 144
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"clip_fraction": 0.0,
|
| 2750 |
+
"completion_length": 2482.1389389038086,
|
| 2751 |
+
"dapo/avg_reward_std": 0.37865253537893295,
|
| 2752 |
+
"dapo/filter_reward_index": 0.0,
|
| 2753 |
+
"dapo/kept_prompts_ratio": 0.6145833432674408,
|
| 2754 |
+
"dapo/num_sampling_attempts": 2.0,
|
| 2755 |
+
"dapo/sampling_efficiency": 58.33333333333333,
|
| 2756 |
+
"dapo/total_prompts_processed": 12.0,
|
| 2757 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2758 |
+
"epoch": 0.1657142857142857,
|
| 2759 |
+
"grad_norm": 0.011065399274230003,
|
| 2760 |
+
"kl": 0.0004658699035644531,
|
| 2761 |
+
"learning_rate": 2.9836319343816397e-07,
|
| 2762 |
+
"loss": 0.0412,
|
| 2763 |
+
"reward": 0.8742740526795387,
|
| 2764 |
+
"reward_std": 0.9688765779137611,
|
| 2765 |
+
"step": 145
|
| 2766 |
+
},
|
| 2767 |
+
{
|
| 2768 |
+
"clip_fraction": 0.0,
|
| 2769 |
+
"completion_length": 2297.6806030273438,
|
| 2770 |
+
"dapo/avg_reward_std": 0.40159281912971945,
|
| 2771 |
+
"dapo/filter_reward_index": 0.0,
|
| 2772 |
+
"dapo/kept_prompts_ratio": 0.627450992955881,
|
| 2773 |
+
"dapo/num_sampling_attempts": 2.125,
|
| 2774 |
+
"dapo/sampling_efficiency": 56.24999999999999,
|
| 2775 |
+
"dapo/total_prompts_processed": 12.75,
|
| 2776 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2777 |
+
"epoch": 0.16685714285714287,
|
| 2778 |
+
"grad_norm": 0.014852684922516346,
|
| 2779 |
+
"kl": 0.00038546323776245117,
|
| 2780 |
+
"learning_rate": 2.918906036420294e-07,
|
| 2781 |
+
"loss": 0.1043,
|
| 2782 |
+
"reward": 0.7259054481983185,
|
| 2783 |
+
"reward_std": 0.9452414810657501,
|
| 2784 |
+
"step": 146
|
| 2785 |
+
},
|
| 2786 |
+
{
|
| 2787 |
+
"clip_fraction": 0.0,
|
| 2788 |
+
"completion_length": 2485.2639389038086,
|
| 2789 |
+
"dapo/avg_reward_std": 0.2594580222731051,
|
| 2790 |
+
"dapo/filter_reward_index": 0.0,
|
| 2791 |
+
"dapo/kept_prompts_ratio": 0.42028986435869464,
|
| 2792 |
+
"dapo/num_sampling_attempts": 2.875,
|
| 2793 |
+
"dapo/sampling_efficiency": 43.125,
|
| 2794 |
+
"dapo/total_prompts_processed": 17.25,
|
| 2795 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2796 |
+
"epoch": 0.168,
|
| 2797 |
+
"grad_norm": 0.011770485900342464,
|
| 2798 |
+
"kl": 0.00037994980812072754,
|
| 2799 |
+
"learning_rate": 2.854966364683872e-07,
|
| 2800 |
+
"loss": 0.0414,
|
| 2801 |
+
"reward": 0.596230074763298,
|
| 2802 |
+
"reward_std": 0.944911852478981,
|
| 2803 |
+
"step": 147
|
| 2804 |
+
},
|
| 2805 |
+
{
|
| 2806 |
+
"clip_fraction": 0.0,
|
| 2807 |
+
"completion_length": 2030.6180877685547,
|
| 2808 |
+
"dapo/avg_reward_std": 0.28245899453759193,
|
| 2809 |
+
"dapo/filter_reward_index": 0.0,
|
| 2810 |
+
"dapo/kept_prompts_ratio": 0.43055556528270245,
|
| 2811 |
+
"dapo/num_sampling_attempts": 3.0,
|
| 2812 |
+
"dapo/sampling_efficiency": 47.70833333333333,
|
| 2813 |
+
"dapo/total_prompts_processed": 18.0,
|
| 2814 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2815 |
+
"epoch": 0.16914285714285715,
|
| 2816 |
+
"grad_norm": 0.010600890032947063,
|
| 2817 |
+
"kl": 0.0003261566162109375,
|
| 2818 |
+
"learning_rate": 2.791832395815782e-07,
|
| 2819 |
+
"loss": 0.018,
|
| 2820 |
+
"reward": 0.5254655107855797,
|
| 2821 |
+
"reward_std": 0.9357841089367867,
|
| 2822 |
+
"step": 148
|
| 2823 |
+
},
|
| 2824 |
+
{
|
| 2825 |
+
"clip_fraction": 0.0,
|
| 2826 |
+
"completion_length": 2956.184051513672,
|
| 2827 |
+
"dapo/avg_reward_std": 0.3112214480837186,
|
| 2828 |
+
"dapo/filter_reward_index": 0.0,
|
| 2829 |
+
"dapo/kept_prompts_ratio": 0.44444445334374905,
|
| 2830 |
+
"dapo/num_sampling_attempts": 3.0,
|
| 2831 |
+
"dapo/sampling_efficiency": 42.08333333333333,
|
| 2832 |
+
"dapo/total_prompts_processed": 18.0,
|
| 2833 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2834 |
+
"epoch": 0.1702857142857143,
|
| 2835 |
+
"grad_norm": 0.010259653441607952,
|
| 2836 |
+
"kl": 0.00048613548278808594,
|
| 2837 |
+
"learning_rate": 2.729523361034538e-07,
|
| 2838 |
+
"loss": 0.0339,
|
| 2839 |
+
"reward": 0.6315554305911064,
|
| 2840 |
+
"reward_std": 0.9876029044389725,
|
| 2841 |
+
"step": 149
|
| 2842 |
+
},
|
| 2843 |
+
{
|
| 2844 |
+
"clip_fraction": 0.0,
|
| 2845 |
+
"completion_length": 2855.0520629882812,
|
| 2846 |
+
"dapo/avg_reward_std": 0.32461989257070756,
|
| 2847 |
+
"dapo/filter_reward_index": 0.0,
|
| 2848 |
+
"dapo/kept_prompts_ratio": 0.5370370447635651,
|
| 2849 |
+
"dapo/num_sampling_attempts": 2.25,
|
| 2850 |
+
"dapo/sampling_efficiency": 55.20833333333333,
|
| 2851 |
+
"dapo/total_prompts_processed": 13.5,
|
| 2852 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2853 |
+
"epoch": 0.17142857142857143,
|
| 2854 |
+
"grad_norm": 0.011087276972830296,
|
| 2855 |
+
"kl": 0.0005285739898681641,
|
| 2856 |
+
"learning_rate": 2.6680582402757324e-07,
|
| 2857 |
+
"loss": 0.054,
|
| 2858 |
+
"reward": 0.80087810754776,
|
| 2859 |
+
"reward_std": 1.0038108006119728,
|
| 2860 |
+
"step": 150
|
| 2861 |
+
},
|
| 2862 |
+
{
|
| 2863 |
+
"clip_fraction": 0.0,
|
| 2864 |
+
"completion_length": 2653.5834197998047,
|
| 2865 |
+
"dapo/avg_reward_std": 0.24287073779851198,
|
| 2866 |
+
"dapo/filter_reward_index": 0.0,
|
| 2867 |
+
"dapo/kept_prompts_ratio": 0.36979167396202683,
|
| 2868 |
+
"dapo/num_sampling_attempts": 4.0,
|
| 2869 |
+
"dapo/sampling_efficiency": 30.327380952380953,
|
| 2870 |
+
"dapo/total_prompts_processed": 24.0,
|
| 2871 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2872 |
+
"epoch": 0.17257142857142857,
|
| 2873 |
+
"grad_norm": 0.011102661490440369,
|
| 2874 |
+
"kl": 0.0005296468734741211,
|
| 2875 |
+
"learning_rate": 2.6074557564105724e-07,
|
| 2876 |
+
"loss": 0.0527,
|
| 2877 |
+
"reward": 0.7124785147607327,
|
| 2878 |
+
"reward_std": 0.9657682925462723,
|
| 2879 |
+
"step": 151
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"clip_fraction": 0.0,
|
| 2883 |
+
"completion_length": 2141.173614501953,
|
| 2884 |
+
"dapo/avg_reward_std": 0.25965497308763963,
|
| 2885 |
+
"dapo/filter_reward_index": 0.0,
|
| 2886 |
+
"dapo/kept_prompts_ratio": 0.3333333396706088,
|
| 2887 |
+
"dapo/num_sampling_attempts": 3.625,
|
| 2888 |
+
"dapo/sampling_efficiency": 36.875,
|
| 2889 |
+
"dapo/total_prompts_processed": 21.75,
|
| 2890 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2891 |
+
"epoch": 0.1737142857142857,
|
| 2892 |
+
"grad_norm": 0.01081050094217062,
|
| 2893 |
+
"kl": 0.00039577484130859375,
|
| 2894 |
+
"learning_rate": 2.547734369542718e-07,
|
| 2895 |
+
"loss": 0.0232,
|
| 2896 |
+
"reward": 0.5607589241117239,
|
| 2897 |
+
"reward_std": 0.9106607139110565,
|
| 2898 |
+
"step": 152
|
| 2899 |
+
},
|
| 2900 |
+
{
|
| 2901 |
+
"clip_fraction": 0.0,
|
| 2902 |
+
"completion_length": 2218.2882347106934,
|
| 2903 |
+
"dapo/avg_reward_std": 0.24554675072431564,
|
| 2904 |
+
"dapo/filter_reward_index": 0.0,
|
| 2905 |
+
"dapo/kept_prompts_ratio": 0.39583333767950535,
|
| 2906 |
+
"dapo/num_sampling_attempts": 3.0,
|
| 2907 |
+
"dapo/sampling_efficiency": 48.75,
|
| 2908 |
+
"dapo/total_prompts_processed": 18.0,
|
| 2909 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2910 |
+
"epoch": 0.17485714285714285,
|
| 2911 |
+
"grad_norm": 0.01474699191749096,
|
| 2912 |
+
"kl": 0.000436246395111084,
|
| 2913 |
+
"learning_rate": 2.488912271385139e-07,
|
| 2914 |
+
"loss": 0.0585,
|
| 2915 |
+
"reward": 0.4214355852454901,
|
| 2916 |
+
"reward_std": 0.9400415197014809,
|
| 2917 |
+
"step": 153
|
| 2918 |
+
},
|
| 2919 |
+
{
|
| 2920 |
+
"clip_fraction": 0.0,
|
| 2921 |
+
"completion_length": 2466.3368377685547,
|
| 2922 |
+
"dapo/avg_reward_std": 0.3308070342649113,
|
| 2923 |
+
"dapo/filter_reward_index": 0.0,
|
| 2924 |
+
"dapo/kept_prompts_ratio": 0.46212121776559134,
|
| 2925 |
+
"dapo/num_sampling_attempts": 2.75,
|
| 2926 |
+
"dapo/sampling_efficiency": 38.541666666666664,
|
| 2927 |
+
"dapo/total_prompts_processed": 16.5,
|
| 2928 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2929 |
+
"epoch": 0.176,
|
| 2930 |
+
"grad_norm": 0.011210402473807335,
|
| 2931 |
+
"kl": 0.0004417896270751953,
|
| 2932 |
+
"learning_rate": 2.4310073797187573e-07,
|
| 2933 |
+
"loss": -0.0244,
|
| 2934 |
+
"reward": 0.7323229797184467,
|
| 2935 |
+
"reward_std": 0.9493635967373848,
|
| 2936 |
+
"step": 154
|
| 2937 |
+
},
|
| 2938 |
+
{
|
| 2939 |
+
"clip_fraction": 0.0,
|
| 2940 |
+
"completion_length": 2012.8715438842773,
|
| 2941 |
+
"dapo/avg_reward_std": 0.2809670078754425,
|
| 2942 |
+
"dapo/filter_reward_index": 0.0,
|
| 2943 |
+
"dapo/kept_prompts_ratio": 0.4133333426713943,
|
| 2944 |
+
"dapo/num_sampling_attempts": 3.125,
|
| 2945 |
+
"dapo/sampling_efficiency": 35.83333333333333,
|
| 2946 |
+
"dapo/total_prompts_processed": 18.75,
|
| 2947 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2948 |
+
"epoch": 0.17714285714285713,
|
| 2949 |
+
"grad_norm": 0.01654733158648014,
|
| 2950 |
+
"kl": 0.00036275386810302734,
|
| 2951 |
+
"learning_rate": 2.374037332934512e-07,
|
| 2952 |
+
"loss": 0.0589,
|
| 2953 |
+
"reward": 0.6634213328361511,
|
| 2954 |
+
"reward_std": 0.8785304054617882,
|
| 2955 |
+
"step": 155
|
| 2956 |
+
},
|
| 2957 |
+
{
|
| 2958 |
+
"clip_fraction": 0.0,
|
| 2959 |
+
"completion_length": 2291.3021240234375,
|
| 2960 |
+
"dapo/avg_reward_std": 0.3599580733672432,
|
| 2961 |
+
"dapo/filter_reward_index": 0.0,
|
| 2962 |
+
"dapo/kept_prompts_ratio": 0.44927537182103033,
|
| 2963 |
+
"dapo/num_sampling_attempts": 2.875,
|
| 2964 |
+
"dapo/sampling_efficiency": 45.53571428571428,
|
| 2965 |
+
"dapo/total_prompts_processed": 17.25,
|
| 2966 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2967 |
+
"epoch": 0.1782857142857143,
|
| 2968 |
+
"grad_norm": 0.011936171911656857,
|
| 2969 |
+
"kl": 0.00043827295303344727,
|
| 2970 |
+
"learning_rate": 2.3180194846605364e-07,
|
| 2971 |
+
"loss": 0.0699,
|
| 2972 |
+
"reward": 0.8599490560591221,
|
| 2973 |
+
"reward_std": 0.9719394743442535,
|
| 2974 |
+
"step": 156
|
| 2975 |
+
},
|
| 2976 |
+
{
|
| 2977 |
+
"clip_fraction": 0.0,
|
| 2978 |
+
"completion_length": 2499.791702270508,
|
| 2979 |
+
"dapo/avg_reward_std": 0.3457585884766145,
|
| 2980 |
+
"dapo/filter_reward_index": 0.0,
|
| 2981 |
+
"dapo/kept_prompts_ratio": 0.5378787998448719,
|
| 2982 |
+
"dapo/num_sampling_attempts": 2.75,
|
| 2983 |
+
"dapo/sampling_efficiency": 48.95833333333333,
|
| 2984 |
+
"dapo/total_prompts_processed": 16.5,
|
| 2985 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 2986 |
+
"epoch": 0.17942857142857144,
|
| 2987 |
+
"grad_norm": 0.01289551891386509,
|
| 2988 |
+
"kl": 0.00048601627349853516,
|
| 2989 |
+
"learning_rate": 2.2629708984760706e-07,
|
| 2990 |
+
"loss": 0.0584,
|
| 2991 |
+
"reward": 0.6511420179158449,
|
| 2992 |
+
"reward_std": 0.9461185112595558,
|
| 2993 |
+
"step": 157
|
| 2994 |
+
},
|
| 2995 |
+
{
|
| 2996 |
+
"clip_fraction": 0.0,
|
| 2997 |
+
"completion_length": 2437.9375228881836,
|
| 2998 |
+
"dapo/avg_reward_std": 0.23957703853475637,
|
| 2999 |
+
"dapo/filter_reward_index": 0.0,
|
| 3000 |
+
"dapo/kept_prompts_ratio": 0.36781609829129847,
|
| 3001 |
+
"dapo/num_sampling_attempts": 3.625,
|
| 3002 |
+
"dapo/sampling_efficiency": 37.291666666666664,
|
| 3003 |
+
"dapo/total_prompts_processed": 21.75,
|
| 3004 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3005 |
+
"epoch": 0.18057142857142858,
|
| 3006 |
+
"grad_norm": 0.012769551016390324,
|
| 3007 |
+
"kl": 0.0004298686981201172,
|
| 3008 |
+
"learning_rate": 2.2089083427137329e-07,
|
| 3009 |
+
"loss": 0.0258,
|
| 3010 |
+
"reward": 0.6606059782207012,
|
| 3011 |
+
"reward_std": 0.9088018089532852,
|
| 3012 |
+
"step": 158
|
| 3013 |
+
},
|
| 3014 |
+
{
|
| 3015 |
+
"clip_fraction": 0.0,
|
| 3016 |
+
"completion_length": 1726.5868225097656,
|
| 3017 |
+
"dapo/avg_reward_std": 0.3139249332573103,
|
| 3018 |
+
"dapo/filter_reward_index": 0.0,
|
| 3019 |
+
"dapo/kept_prompts_ratio": 0.3913043543048527,
|
| 3020 |
+
"dapo/num_sampling_attempts": 2.875,
|
| 3021 |
+
"dapo/sampling_efficiency": 55.416666666666664,
|
| 3022 |
+
"dapo/total_prompts_processed": 17.25,
|
| 3023 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3024 |
+
"epoch": 0.18171428571428572,
|
| 3025 |
+
"grad_norm": 0.013688490726053715,
|
| 3026 |
+
"kl": 0.00027683377265930176,
|
| 3027 |
+
"learning_rate": 2.1558482853517253e-07,
|
| 3028 |
+
"loss": 0.0506,
|
| 3029 |
+
"reward": 0.7147123599424958,
|
| 3030 |
+
"reward_std": 0.9531080722808838,
|
| 3031 |
+
"step": 159
|
| 3032 |
+
},
|
| 3033 |
+
{
|
| 3034 |
+
"clip_fraction": 0.0,
|
| 3035 |
+
"completion_length": 1593.003475189209,
|
| 3036 |
+
"dapo/avg_reward_std": 0.2799004193010001,
|
| 3037 |
+
"dapo/filter_reward_index": 0.0,
|
| 3038 |
+
"dapo/kept_prompts_ratio": 0.33908046782016754,
|
| 3039 |
+
"dapo/num_sampling_attempts": 3.625,
|
| 3040 |
+
"dapo/sampling_efficiency": 45.3125,
|
| 3041 |
+
"dapo/total_prompts_processed": 21.75,
|
| 3042 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3043 |
+
"epoch": 0.18285714285714286,
|
| 3044 |
+
"grad_norm": 0.020229365676641464,
|
| 3045 |
+
"kl": 0.00033217668533325195,
|
| 3046 |
+
"learning_rate": 2.1038068889975259e-07,
|
| 3047 |
+
"loss": 0.0296,
|
| 3048 |
+
"reward": 0.7677402682602406,
|
| 3049 |
+
"reward_std": 0.9385703578591347,
|
| 3050 |
+
"step": 160
|
| 3051 |
+
},
|
| 3052 |
+
{
|
| 3053 |
+
"clip_fraction": 0.0,
|
| 3054 |
+
"completion_length": 1877.9444274902344,
|
| 3055 |
+
"dapo/avg_reward_std": 0.36716995636622113,
|
| 3056 |
+
"dapo/filter_reward_index": 0.0,
|
| 3057 |
+
"dapo/kept_prompts_ratio": 0.4930555621782939,
|
| 3058 |
+
"dapo/num_sampling_attempts": 3.0,
|
| 3059 |
+
"dapo/sampling_efficiency": 44.49404761904761,
|
| 3060 |
+
"dapo/total_prompts_processed": 18.0,
|
| 3061 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3062 |
+
"epoch": 0.184,
|
| 3063 |
+
"grad_norm": 0.012556586414575577,
|
| 3064 |
+
"kl": 0.00037413835525512695,
|
| 3065 |
+
"learning_rate": 2.0528000059645995e-07,
|
| 3066 |
+
"loss": 0.0401,
|
| 3067 |
+
"reward": 0.6385876163840294,
|
| 3068 |
+
"reward_std": 0.9741755276918411,
|
| 3069 |
+
"step": 161
|
| 3070 |
+
},
|
| 3071 |
+
{
|
| 3072 |
+
"clip_fraction": 0.0,
|
| 3073 |
+
"completion_length": 2543.1145782470703,
|
| 3074 |
+
"dapo/avg_reward_std": 0.20304026060244618,
|
| 3075 |
+
"dapo/filter_reward_index": 0.0,
|
| 3076 |
+
"dapo/kept_prompts_ratio": 0.28921569226419225,
|
| 3077 |
+
"dapo/num_sampling_attempts": 4.25,
|
| 3078 |
+
"dapo/sampling_efficiency": 26.249999999999996,
|
| 3079 |
+
"dapo/total_prompts_processed": 25.5,
|
| 3080 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3081 |
+
"epoch": 0.18514285714285714,
|
| 3082 |
+
"grad_norm": 0.010984732769429684,
|
| 3083 |
+
"kl": 0.0005058050155639648,
|
| 3084 |
+
"learning_rate": 2.0028431734436308e-07,
|
| 3085 |
+
"loss": 0.0214,
|
| 3086 |
+
"reward": 0.8138710260391235,
|
| 3087 |
+
"reward_std": 0.937220610678196,
|
| 3088 |
+
"step": 162
|
| 3089 |
+
},
|
| 3090 |
+
{
|
| 3091 |
+
"clip_fraction": 0.0,
|
| 3092 |
+
"completion_length": 2579.7916946411133,
|
| 3093 |
+
"dapo/avg_reward_std": 0.2669851701049244,
|
| 3094 |
+
"dapo/filter_reward_index": 0.0,
|
| 3095 |
+
"dapo/kept_prompts_ratio": 0.3333333397612852,
|
| 3096 |
+
"dapo/num_sampling_attempts": 4.25,
|
| 3097 |
+
"dapo/sampling_efficiency": 31.38888888888889,
|
| 3098 |
+
"dapo/total_prompts_processed": 25.5,
|
| 3099 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3100 |
+
"epoch": 0.18628571428571428,
|
| 3101 |
+
"grad_norm": 0.01393849402666092,
|
| 3102 |
+
"kl": 0.0005407929420471191,
|
| 3103 |
+
"learning_rate": 1.9539516087697517e-07,
|
| 3104 |
+
"loss": 0.0557,
|
| 3105 |
+
"reward": 0.6086596520617604,
|
| 3106 |
+
"reward_std": 0.9360831007361412,
|
| 3107 |
+
"step": 163
|
| 3108 |
+
},
|
| 3109 |
+
{
|
| 3110 |
+
"clip_fraction": 0.0,
|
| 3111 |
+
"completion_length": 2303.781295776367,
|
| 3112 |
+
"dapo/avg_reward_std": 0.2889538109302521,
|
| 3113 |
+
"dapo/filter_reward_index": 0.0,
|
| 3114 |
+
"dapo/kept_prompts_ratio": 0.40384616129673445,
|
| 3115 |
+
"dapo/num_sampling_attempts": 3.25,
|
| 3116 |
+
"dapo/sampling_efficiency": 41.041666666666664,
|
| 3117 |
+
"dapo/total_prompts_processed": 19.5,
|
| 3118 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3119 |
+
"epoch": 0.18742857142857142,
|
| 3120 |
+
"grad_norm": 0.012467012740671635,
|
| 3121 |
+
"kl": 0.0005753040313720703,
|
| 3122 |
+
"learning_rate": 1.9061402047871833e-07,
|
| 3123 |
+
"loss": 0.0286,
|
| 3124 |
+
"reward": 0.7579413987696171,
|
| 3125 |
+
"reward_std": 0.966604009270668,
|
| 3126 |
+
"step": 164
|
| 3127 |
+
},
|
| 3128 |
+
{
|
| 3129 |
+
"clip_fraction": 0.0,
|
| 3130 |
+
"completion_length": 2215.8715744018555,
|
| 3131 |
+
"dapo/avg_reward_std": 0.2284111071910177,
|
| 3132 |
+
"dapo/filter_reward_index": 0.0,
|
| 3133 |
+
"dapo/kept_prompts_ratio": 0.3630952446588448,
|
| 3134 |
+
"dapo/num_sampling_attempts": 3.5,
|
| 3135 |
+
"dapo/sampling_efficiency": 49.72222222222222,
|
| 3136 |
+
"dapo/total_prompts_processed": 21.0,
|
| 3137 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3138 |
+
"epoch": 0.18857142857142858,
|
| 3139 |
+
"grad_norm": 0.013376005925238132,
|
| 3140 |
+
"kl": 0.00038570165634155273,
|
| 3141 |
+
"learning_rate": 1.8594235253127372e-07,
|
| 3142 |
+
"loss": 0.0737,
|
| 3143 |
+
"reward": 0.6369971446692944,
|
| 3144 |
+
"reward_std": 0.944696456193924,
|
| 3145 |
+
"step": 165
|
| 3146 |
+
},
|
| 3147 |
+
{
|
| 3148 |
+
"clip_fraction": 0.0,
|
| 3149 |
+
"completion_length": 2194.999984741211,
|
| 3150 |
+
"dapo/avg_reward_std": 0.35230770577555115,
|
| 3151 |
+
"dapo/filter_reward_index": 0.0,
|
| 3152 |
+
"dapo/kept_prompts_ratio": 0.5289855158847311,
|
| 3153 |
+
"dapo/num_sampling_attempts": 2.875,
|
| 3154 |
+
"dapo/sampling_efficiency": 40.416666666666664,
|
| 3155 |
+
"dapo/total_prompts_processed": 17.25,
|
| 3156 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3157 |
+
"epoch": 0.18971428571428572,
|
| 3158 |
+
"grad_norm": 0.00896221399307251,
|
| 3159 |
+
"kl": 0.0004324018955230713,
|
| 3160 |
+
"learning_rate": 1.8138158006995363e-07,
|
| 3161 |
+
"loss": -0.0087,
|
| 3162 |
+
"reward": 0.770520705729723,
|
| 3163 |
+
"reward_std": 0.9258415997028351,
|
| 3164 |
+
"step": 166
|
| 3165 |
+
},
|
| 3166 |
+
{
|
| 3167 |
+
"clip_fraction": 0.0,
|
| 3168 |
+
"completion_length": 2363.9861373901367,
|
| 3169 |
+
"dapo/avg_reward_std": 0.23058613193662544,
|
| 3170 |
+
"dapo/filter_reward_index": 0.0,
|
| 3171 |
+
"dapo/kept_prompts_ratio": 0.2763157930029066,
|
| 3172 |
+
"dapo/num_sampling_attempts": 4.75,
|
| 3173 |
+
"dapo/sampling_efficiency": 30.44642857142857,
|
| 3174 |
+
"dapo/total_prompts_processed": 28.5,
|
| 3175 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3176 |
+
"epoch": 0.19085714285714286,
|
| 3177 |
+
"grad_norm": 0.011913989670574665,
|
| 3178 |
+
"kl": 0.0005799531936645508,
|
| 3179 |
+
"learning_rate": 1.7693309235023127e-07,
|
| 3180 |
+
"loss": 0.0282,
|
| 3181 |
+
"reward": 0.8937316909432411,
|
| 3182 |
+
"reward_std": 0.9134809225797653,
|
| 3183 |
+
"step": 167
|
| 3184 |
+
},
|
| 3185 |
+
{
|
| 3186 |
+
"clip_fraction": 0.0,
|
| 3187 |
+
"completion_length": 1846.3229217529297,
|
| 3188 |
+
"dapo/avg_reward_std": 0.2788652099412063,
|
| 3189 |
+
"dapo/filter_reward_index": 0.0,
|
| 3190 |
+
"dapo/kept_prompts_ratio": 0.37931035459041595,
|
| 3191 |
+
"dapo/num_sampling_attempts": 3.625,
|
| 3192 |
+
"dapo/sampling_efficiency": 33.03571428571428,
|
| 3193 |
+
"dapo/total_prompts_processed": 21.75,
|
| 3194 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3195 |
+
"epoch": 0.192,
|
| 3196 |
+
"grad_norm": 0.013345438055694103,
|
| 3197 |
+
"kl": 0.00038933753967285156,
|
| 3198 |
+
"learning_rate": 1.7259824442455923e-07,
|
| 3199 |
+
"loss": 0.0657,
|
| 3200 |
+
"reward": 0.5173812105786055,
|
| 3201 |
+
"reward_std": 0.9046202600002289,
|
| 3202 |
+
"step": 168
|
| 3203 |
+
},
|
| 3204 |
+
{
|
| 3205 |
+
"clip_fraction": 0.0,
|
| 3206 |
+
"completion_length": 1632.9965515136719,
|
| 3207 |
+
"dapo/avg_reward_std": 0.33004767837978544,
|
| 3208 |
+
"dapo/filter_reward_index": 0.0,
|
| 3209 |
+
"dapo/kept_prompts_ratio": 0.49206350318023134,
|
| 3210 |
+
"dapo/num_sampling_attempts": 2.625,
|
| 3211 |
+
"dapo/sampling_efficiency": 51.45833333333333,
|
| 3212 |
+
"dapo/total_prompts_processed": 15.75,
|
| 3213 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3214 |
+
"epoch": 0.19314285714285714,
|
| 3215 |
+
"grad_norm": 0.016018711030483246,
|
| 3216 |
+
"kl": 0.0004235506057739258,
|
| 3217 |
+
"learning_rate": 1.6837835672960831e-07,
|
| 3218 |
+
"loss": -0.0266,
|
| 3219 |
+
"reward": 0.7293304707854986,
|
| 3220 |
+
"reward_std": 0.9580913484096527,
|
| 3221 |
+
"step": 169
|
| 3222 |
+
},
|
| 3223 |
+
{
|
| 3224 |
+
"clip_fraction": 0.0,
|
| 3225 |
+
"completion_length": 2218.357666015625,
|
| 3226 |
+
"dapo/avg_reward_std": 0.30882045084779913,
|
| 3227 |
+
"dapo/filter_reward_index": 0.0,
|
| 3228 |
+
"dapo/kept_prompts_ratio": 0.4242424287579276,
|
| 3229 |
+
"dapo/num_sampling_attempts": 2.75,
|
| 3230 |
+
"dapo/sampling_efficiency": 49.37499999999999,
|
| 3231 |
+
"dapo/total_prompts_processed": 16.5,
|
| 3232 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3233 |
+
"epoch": 0.19428571428571428,
|
| 3234 |
+
"grad_norm": 0.012691031210124493,
|
| 3235 |
+
"kl": 0.0005915164947509766,
|
| 3236 |
+
"learning_rate": 1.6427471468404952e-07,
|
| 3237 |
+
"loss": 0.0375,
|
| 3238 |
+
"reward": 0.731636168435216,
|
| 3239 |
+
"reward_std": 0.9506037011742592,
|
| 3240 |
+
"step": 170
|
| 3241 |
+
},
|
| 3242 |
+
{
|
| 3243 |
+
"clip_fraction": 0.0,
|
| 3244 |
+
"completion_length": 2086.989585876465,
|
| 3245 |
+
"dapo/avg_reward_std": 0.26685478786627453,
|
| 3246 |
+
"dapo/filter_reward_index": 0.0,
|
| 3247 |
+
"dapo/kept_prompts_ratio": 0.372222230831782,
|
| 3248 |
+
"dapo/num_sampling_attempts": 3.75,
|
| 3249 |
+
"dapo/sampling_efficiency": 36.45833333333333,
|
| 3250 |
+
"dapo/total_prompts_processed": 22.5,
|
| 3251 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3252 |
+
"epoch": 0.19542857142857142,
|
| 3253 |
+
"grad_norm": 0.0107533298432827,
|
| 3254 |
+
"kl": 0.00045359134674072266,
|
| 3255 |
+
"learning_rate": 1.6028856829700258e-07,
|
| 3256 |
+
"loss": 0.0268,
|
| 3257 |
+
"reward": 0.6401270348578691,
|
| 3258 |
+
"reward_std": 0.9421326443552971,
|
| 3259 |
+
"step": 171
|
| 3260 |
+
},
|
| 3261 |
+
{
|
| 3262 |
+
"clip_fraction": 0.0,
|
| 3263 |
+
"completion_length": 1523.298625946045,
|
| 3264 |
+
"dapo/avg_reward_std": 0.2958875367274651,
|
| 3265 |
+
"dapo/filter_reward_index": 0.0,
|
| 3266 |
+
"dapo/kept_prompts_ratio": 0.4294871888481654,
|
| 3267 |
+
"dapo/num_sampling_attempts": 3.25,
|
| 3268 |
+
"dapo/sampling_efficiency": 41.041666666666664,
|
| 3269 |
+
"dapo/total_prompts_processed": 19.5,
|
| 3270 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3271 |
+
"epoch": 0.19657142857142856,
|
| 3272 |
+
"grad_norm": 0.02487981878221035,
|
| 3273 |
+
"kl": 0.00044208765029907227,
|
| 3274 |
+
"learning_rate": 1.5642113178727193e-07,
|
| 3275 |
+
"loss": 0.0215,
|
| 3276 |
+
"reward": 0.5742892920970917,
|
| 3277 |
+
"reward_std": 0.9192508533596992,
|
| 3278 |
+
"step": 172
|
| 3279 |
+
},
|
| 3280 |
+
{
|
| 3281 |
+
"clip_fraction": 0.0,
|
| 3282 |
+
"completion_length": 2197.4722290039062,
|
| 3283 |
+
"dapo/avg_reward_std": 0.33716599914160644,
|
| 3284 |
+
"dapo/filter_reward_index": 0.0,
|
| 3285 |
+
"dapo/kept_prompts_ratio": 0.4545454619960351,
|
| 3286 |
+
"dapo/num_sampling_attempts": 2.75,
|
| 3287 |
+
"dapo/sampling_efficiency": 45.20833333333333,
|
| 3288 |
+
"dapo/total_prompts_processed": 16.5,
|
| 3289 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3290 |
+
"epoch": 0.1977142857142857,
|
| 3291 |
+
"grad_norm": 0.00999497715383768,
|
| 3292 |
+
"kl": 0.0006158351898193359,
|
| 3293 |
+
"learning_rate": 1.5267358321348285e-07,
|
| 3294 |
+
"loss": -0.0198,
|
| 3295 |
+
"reward": 0.6909432113170624,
|
| 3296 |
+
"reward_std": 0.9331774786114693,
|
| 3297 |
+
"step": 173
|
| 3298 |
+
},
|
| 3299 |
+
{
|
| 3300 |
+
"clip_fraction": 0.0,
|
| 3301 |
+
"completion_length": 2469.1909942626953,
|
| 3302 |
+
"dapo/avg_reward_std": 0.31674497947096825,
|
| 3303 |
+
"dapo/filter_reward_index": 0.0,
|
| 3304 |
+
"dapo/kept_prompts_ratio": 0.4722222325702508,
|
| 3305 |
+
"dapo/num_sampling_attempts": 3.0,
|
| 3306 |
+
"dapo/sampling_efficiency": 47.61904761904762,
|
| 3307 |
+
"dapo/total_prompts_processed": 18.0,
|
| 3308 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3309 |
+
"epoch": 0.19885714285714284,
|
| 3310 |
+
"grad_norm": 0.027324816212058067,
|
| 3311 |
+
"kl": 0.0005202293395996094,
|
| 3312 |
+
"learning_rate": 1.4904706411523448e-07,
|
| 3313 |
+
"loss": 0.1381,
|
| 3314 |
+
"reward": 0.7919853329658508,
|
| 3315 |
+
"reward_std": 0.9734821692109108,
|
| 3316 |
+
"step": 174
|
| 3317 |
+
},
|
| 3318 |
+
{
|
| 3319 |
+
"clip_fraction": 0.0,
|
| 3320 |
+
"completion_length": 2290.7292098999023,
|
| 3321 |
+
"dapo/avg_reward_std": 0.2796748812709536,
|
| 3322 |
+
"dapo/filter_reward_index": 0.0,
|
| 3323 |
+
"dapo/kept_prompts_ratio": 0.41071429369705065,
|
| 3324 |
+
"dapo/num_sampling_attempts": 3.5,
|
| 3325 |
+
"dapo/sampling_efficiency": 31.041666666666664,
|
| 3326 |
+
"dapo/total_prompts_processed": 21.0,
|
| 3327 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3328 |
+
"epoch": 0.2,
|
| 3329 |
+
"grad_norm": 0.011332061141729355,
|
| 3330 |
+
"kl": 0.000499039888381958,
|
| 3331 |
+
"learning_rate": 1.4554267916537495e-07,
|
| 3332 |
+
"loss": 0.0026,
|
| 3333 |
+
"reward": 0.5971913021057844,
|
| 3334 |
+
"reward_std": 0.9767839089035988,
|
| 3335 |
+
"step": 175
|
| 3336 |
+
},
|
| 3337 |
+
{
|
| 3338 |
+
"clip_fraction": 0.0,
|
| 3339 |
+
"completion_length": 2643.475685119629,
|
| 3340 |
+
"dapo/avg_reward_std": 0.30459834399976227,
|
| 3341 |
+
"dapo/filter_reward_index": 0.0,
|
| 3342 |
+
"dapo/kept_prompts_ratio": 0.4649122922044051,
|
| 3343 |
+
"dapo/num_sampling_attempts": 2.375,
|
| 3344 |
+
"dapo/sampling_efficiency": 54.58333333333333,
|
| 3345 |
+
"dapo/total_prompts_processed": 14.25,
|
| 3346 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3347 |
+
"epoch": 0.20114285714285715,
|
| 3348 |
+
"grad_norm": 0.011058920994400978,
|
| 3349 |
+
"kl": 0.0006421804428100586,
|
| 3350 |
+
"learning_rate": 1.4216149583350755e-07,
|
| 3351 |
+
"loss": 0.0243,
|
| 3352 |
+
"reward": 0.801079198718071,
|
| 3353 |
+
"reward_std": 1.0328236892819405,
|
| 3354 |
+
"step": 176
|
| 3355 |
+
},
|
| 3356 |
+
{
|
| 3357 |
+
"clip_fraction": 0.0,
|
| 3358 |
+
"completion_length": 2657.517364501953,
|
| 3359 |
+
"dapo/avg_reward_std": 0.268055671826005,
|
| 3360 |
+
"dapo/filter_reward_index": 0.0,
|
| 3361 |
+
"dapo/kept_prompts_ratio": 0.3072916711680591,
|
| 3362 |
+
"dapo/num_sampling_attempts": 4.0,
|
| 3363 |
+
"dapo/sampling_efficiency": 30.32738095238095,
|
| 3364 |
+
"dapo/total_prompts_processed": 24.0,
|
| 3365 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3366 |
+
"epoch": 0.2022857142857143,
|
| 3367 |
+
"grad_norm": 0.012514113448560238,
|
| 3368 |
+
"kl": 0.0006227493286132812,
|
| 3369 |
+
"learning_rate": 1.3890454406082956e-07,
|
| 3370 |
+
"loss": 0.066,
|
| 3371 |
+
"reward": 0.5342087037861347,
|
| 3372 |
+
"reward_std": 0.9403787776827812,
|
| 3373 |
+
"step": 177
|
| 3374 |
+
},
|
| 3375 |
+
{
|
| 3376 |
+
"clip_fraction": 0.0,
|
| 3377 |
+
"completion_length": 1730.2395935058594,
|
| 3378 |
+
"dapo/avg_reward_std": 0.22906314557598484,
|
| 3379 |
+
"dapo/filter_reward_index": 0.0,
|
| 3380 |
+
"dapo/kept_prompts_ratio": 0.354838716406976,
|
| 3381 |
+
"dapo/num_sampling_attempts": 2.625,
|
| 3382 |
+
"dapo/sampling_efficiency": 49.99999999999999,
|
| 3383 |
+
"dapo/total_prompts_processed": 15.75,
|
| 3384 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3385 |
+
"epoch": 0.20342857142857143,
|
| 3386 |
+
"grad_norm": 0.013909725472331047,
|
| 3387 |
+
"kl": 0.0004641413688659668,
|
| 3388 |
+
"learning_rate": 1.3577281594640182e-07,
|
| 3389 |
+
"loss": -0.0032,
|
| 3390 |
+
"reward": 0.817855941131711,
|
| 3391 |
+
"reward_std": 0.9715805351734161,
|
| 3392 |
+
"step": 178
|
| 3393 |
+
},
|
| 3394 |
+
{
|
| 3395 |
+
"clip_fraction": 0.0,
|
| 3396 |
+
"completion_length": 1916.9652633666992,
|
| 3397 |
+
"dapo/avg_reward_std": 0.33905652307328726,
|
| 3398 |
+
"dapo/filter_reward_index": 0.0,
|
| 3399 |
+
"dapo/kept_prompts_ratio": 0.5000000085149493,
|
| 3400 |
+
"dapo/num_sampling_attempts": 2.625,
|
| 3401 |
+
"dapo/sampling_efficiency": 49.99999999999999,
|
| 3402 |
+
"dapo/total_prompts_processed": 15.75,
|
| 3403 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3404 |
+
"epoch": 0.20457142857142857,
|
| 3405 |
+
"grad_norm": 0.010170280002057552,
|
| 3406 |
+
"kl": 0.00033092498779296875,
|
| 3407 |
+
"learning_rate": 1.3276726544494571e-07,
|
| 3408 |
+
"loss": 0.0153,
|
| 3409 |
+
"reward": 0.6332587338984013,
|
| 3410 |
+
"reward_std": 0.9844094663858414,
|
| 3411 |
+
"step": 179
|
| 3412 |
+
},
|
| 3413 |
+
{
|
| 3414 |
+
"clip_fraction": 0.0,
|
| 3415 |
+
"completion_length": 2013.7534942626953,
|
| 3416 |
+
"dapo/avg_reward_std": 0.4115603660282336,
|
| 3417 |
+
"dapo/filter_reward_index": 0.0,
|
| 3418 |
+
"dapo/kept_prompts_ratio": 0.5175438719360452,
|
| 3419 |
+
"dapo/num_sampling_attempts": 2.375,
|
| 3420 |
+
"dapo/sampling_efficiency": 48.95833333333333,
|
| 3421 |
+
"dapo/total_prompts_processed": 14.25,
|
| 3422 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3423 |
+
"epoch": 0.2057142857142857,
|
| 3424 |
+
"grad_norm": 0.010059732012450695,
|
| 3425 |
+
"kl": 0.0004872828722000122,
|
| 3426 |
+
"learning_rate": 1.2988880807625927e-07,
|
| 3427 |
+
"loss": 0.012,
|
| 3428 |
+
"reward": 0.7964395936578512,
|
| 3429 |
+
"reward_std": 0.9064052030444145,
|
| 3430 |
+
"step": 180
|
| 3431 |
+
},
|
| 3432 |
+
{
|
| 3433 |
+
"clip_fraction": 0.0,
|
| 3434 |
+
"completion_length": 2538.3159713745117,
|
| 3435 |
+
"dapo/avg_reward_std": 0.3185795678032769,
|
| 3436 |
+
"dapo/filter_reward_index": 0.0,
|
| 3437 |
+
"dapo/kept_prompts_ratio": 0.3703703775450035,
|
| 3438 |
+
"dapo/num_sampling_attempts": 3.375,
|
| 3439 |
+
"dapo/sampling_efficiency": 40.0,
|
| 3440 |
+
"dapo/total_prompts_processed": 20.25,
|
| 3441 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3442 |
+
"epoch": 0.20685714285714285,
|
| 3443 |
+
"grad_norm": 0.009190794080495834,
|
| 3444 |
+
"kl": 0.0005941390991210938,
|
| 3445 |
+
"learning_rate": 1.2713832064634125e-07,
|
| 3446 |
+
"loss": -0.0091,
|
| 3447 |
+
"reward": 0.6052752519026399,
|
| 3448 |
+
"reward_std": 0.9398948326706886,
|
| 3449 |
+
"step": 181
|
| 3450 |
+
},
|
| 3451 |
+
{
|
| 3452 |
+
"clip_fraction": 0.0,
|
| 3453 |
+
"completion_length": 1992.0277557373047,
|
| 3454 |
+
"dapo/avg_reward_std": 0.30058977752923965,
|
| 3455 |
+
"dapo/filter_reward_index": 0.0,
|
| 3456 |
+
"dapo/kept_prompts_ratio": 0.3392857238650322,
|
| 3457 |
+
"dapo/num_sampling_attempts": 3.5,
|
| 3458 |
+
"dapo/sampling_efficiency": 45.32738095238095,
|
| 3459 |
+
"dapo/total_prompts_processed": 21.0,
|
| 3460 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3461 |
+
"epoch": 0.208,
|
| 3462 |
+
"grad_norm": 0.017918387427926064,
|
| 3463 |
+
"kl": 0.00043332576751708984,
|
| 3464 |
+
"learning_rate": 1.2451664098030743e-07,
|
| 3465 |
+
"loss": 0.0782,
|
| 3466 |
+
"reward": 0.7308525424450636,
|
| 3467 |
+
"reward_std": 0.8988610878586769,
|
| 3468 |
+
"step": 182
|
| 3469 |
+
},
|
| 3470 |
+
{
|
| 3471 |
+
"clip_fraction": 0.0,
|
| 3472 |
+
"completion_length": 2368.312515258789,
|
| 3473 |
+
"dapo/avg_reward_std": 0.2227620858213176,
|
| 3474 |
+
"dapo/filter_reward_index": 0.0,
|
| 3475 |
+
"dapo/kept_prompts_ratio": 0.40579710317694623,
|
| 3476 |
+
"dapo/num_sampling_attempts": 2.875,
|
| 3477 |
+
"dapo/sampling_efficiency": 48.33333333333333,
|
| 3478 |
+
"dapo/total_prompts_processed": 17.25,
|
| 3479 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3480 |
+
"epoch": 0.20914285714285713,
|
| 3481 |
+
"grad_norm": 0.01093615498393774,
|
| 3482 |
+
"kl": 0.0005226731300354004,
|
| 3483 |
+
"learning_rate": 1.220245676671809e-07,
|
| 3484 |
+
"loss": -0.0097,
|
| 3485 |
+
"reward": 0.6296821031719446,
|
| 3486 |
+
"reward_std": 0.9496165588498116,
|
| 3487 |
+
"step": 183
|
| 3488 |
+
},
|
| 3489 |
+
{
|
| 3490 |
+
"clip_fraction": 0.0,
|
| 3491 |
+
"completion_length": 1855.0486297607422,
|
| 3492 |
+
"dapo/avg_reward_std": 0.3308859848976135,
|
| 3493 |
+
"dapo/filter_reward_index": 0.0,
|
| 3494 |
+
"dapo/kept_prompts_ratio": 0.4133333384990692,
|
| 3495 |
+
"dapo/num_sampling_attempts": 3.125,
|
| 3496 |
+
"dapo/sampling_efficiency": 41.666666666666664,
|
| 3497 |
+
"dapo/total_prompts_processed": 18.75,
|
| 3498 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3499 |
+
"epoch": 0.2102857142857143,
|
| 3500 |
+
"grad_norm": 0.013805963099002838,
|
| 3501 |
+
"kl": 0.0004195570945739746,
|
| 3502 |
+
"learning_rate": 1.1966285981663407e-07,
|
| 3503 |
+
"loss": 0.0542,
|
| 3504 |
+
"reward": 0.8230033777654171,
|
| 3505 |
+
"reward_std": 0.9269852489233017,
|
| 3506 |
+
"step": 184
|
| 3507 |
+
},
|
| 3508 |
+
{
|
| 3509 |
+
"clip_fraction": 0.0,
|
| 3510 |
+
"completion_length": 2737.260452270508,
|
| 3511 |
+
"dapo/avg_reward_std": 0.3074522775908311,
|
| 3512 |
+
"dapo/filter_reward_index": 0.0,
|
| 3513 |
+
"dapo/kept_prompts_ratio": 0.45138889489074546,
|
| 3514 |
+
"dapo/num_sampling_attempts": 3.0,
|
| 3515 |
+
"dapo/sampling_efficiency": 49.37499999999999,
|
| 3516 |
+
"dapo/total_prompts_processed": 18.0,
|
| 3517 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3518 |
+
"epoch": 0.21142857142857144,
|
| 3519 |
+
"grad_norm": 0.01179632730782032,
|
| 3520 |
+
"kl": 0.0006718635559082031,
|
| 3521 |
+
"learning_rate": 1.1743223682775649e-07,
|
| 3522 |
+
"loss": 0.0529,
|
| 3523 |
+
"reward": 0.6228375509381294,
|
| 3524 |
+
"reward_std": 0.9775977432727814,
|
| 3525 |
+
"step": 185
|
| 3526 |
+
},
|
| 3527 |
+
{
|
| 3528 |
+
"clip_fraction": 0.0,
|
| 3529 |
+
"completion_length": 2526.899368286133,
|
| 3530 |
+
"dapo/avg_reward_std": 0.2964219942688942,
|
| 3531 |
+
"dapo/filter_reward_index": 0.0,
|
| 3532 |
+
"dapo/kept_prompts_ratio": 0.48333334401249883,
|
| 3533 |
+
"dapo/num_sampling_attempts": 2.5,
|
| 3534 |
+
"dapo/sampling_efficiency": 58.33333333333333,
|
| 3535 |
+
"dapo/total_prompts_processed": 15.0,
|
| 3536 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3537 |
+
"epoch": 0.21257142857142858,
|
| 3538 |
+
"grad_norm": 0.014796112664043903,
|
| 3539 |
+
"kl": 0.0005816221237182617,
|
| 3540 |
+
"learning_rate": 1.1533337816991931e-07,
|
| 3541 |
+
"loss": 0.088,
|
| 3542 |
+
"reward": 0.8448536917567253,
|
| 3543 |
+
"reward_std": 0.9608767181634903,
|
| 3544 |
+
"step": 186
|
| 3545 |
+
},
|
| 3546 |
+
{
|
| 3547 |
+
"clip_fraction": 0.0,
|
| 3548 |
+
"completion_length": 2288.274345397949,
|
| 3549 |
+
"dapo/avg_reward_std": 0.3166468055159957,
|
| 3550 |
+
"dapo/filter_reward_index": 0.0,
|
| 3551 |
+
"dapo/kept_prompts_ratio": 0.34567901823255753,
|
| 3552 |
+
"dapo/num_sampling_attempts": 3.375,
|
| 3553 |
+
"dapo/sampling_efficiency": 40.972222222222214,
|
| 3554 |
+
"dapo/total_prompts_processed": 20.25,
|
| 3555 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3556 |
+
"epoch": 0.21371428571428572,
|
| 3557 |
+
"grad_norm": 0.011898735538125038,
|
| 3558 |
+
"kl": 0.000521540641784668,
|
| 3559 |
+
"learning_rate": 1.1336692317580158e-07,
|
| 3560 |
+
"loss": 0.0415,
|
| 3561 |
+
"reward": 0.7687236070632935,
|
| 3562 |
+
"reward_std": 0.9334599822759628,
|
| 3563 |
+
"step": 187
|
| 3564 |
+
},
|
| 3565 |
+
{
|
| 3566 |
+
"clip_fraction": 0.0,
|
| 3567 |
+
"completion_length": 2432.531265258789,
|
| 3568 |
+
"dapo/avg_reward_std": 0.28751447051763535,
|
| 3569 |
+
"dapo/filter_reward_index": 0.0,
|
| 3570 |
+
"dapo/kept_prompts_ratio": 0.4513888979951541,
|
| 3571 |
+
"dapo/num_sampling_attempts": 3.0,
|
| 3572 |
+
"dapo/sampling_efficiency": 53.33333333333333,
|
| 3573 |
+
"dapo/total_prompts_processed": 18.0,
|
| 3574 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3575 |
+
"epoch": 0.21485714285714286,
|
| 3576 |
+
"grad_norm": 0.010497819632291794,
|
| 3577 |
+
"kl": 0.0007112026214599609,
|
| 3578 |
+
"learning_rate": 1.1153347084664419e-07,
|
| 3579 |
+
"loss": 0.0185,
|
| 3580 |
+
"reward": 0.7899295631796122,
|
| 3581 |
+
"reward_std": 0.9512373134493828,
|
| 3582 |
+
"step": 188
|
| 3583 |
+
},
|
| 3584 |
+
{
|
| 3585 |
+
"clip_fraction": 0.0,
|
| 3586 |
+
"completion_length": 1948.9167022705078,
|
| 3587 |
+
"dapo/avg_reward_std": 0.30568089832862216,
|
| 3588 |
+
"dapo/filter_reward_index": 0.0,
|
| 3589 |
+
"dapo/kept_prompts_ratio": 0.46527779040237266,
|
| 3590 |
+
"dapo/num_sampling_attempts": 3.0,
|
| 3591 |
+
"dapo/sampling_efficiency": 36.87499999999999,
|
| 3592 |
+
"dapo/total_prompts_processed": 18.0,
|
| 3593 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3594 |
+
"epoch": 0.216,
|
| 3595 |
+
"grad_norm": 0.013562222942709923,
|
| 3596 |
+
"kl": 0.0006091594696044922,
|
| 3597 |
+
"learning_rate": 1.0983357966978745e-07,
|
| 3598 |
+
"loss": 0.0388,
|
| 3599 |
+
"reward": 0.6485428418964148,
|
| 3600 |
+
"reward_std": 0.9110815972089767,
|
| 3601 |
+
"step": 189
|
| 3602 |
+
},
|
| 3603 |
+
{
|
| 3604 |
+
"clip_fraction": 0.0,
|
| 3605 |
+
"completion_length": 2494.395866394043,
|
| 3606 |
+
"dapo/avg_reward_std": 0.27111421525478363,
|
| 3607 |
+
"dapo/filter_reward_index": 0.0,
|
| 3608 |
+
"dapo/kept_prompts_ratio": 0.3863636404275894,
|
| 3609 |
+
"dapo/num_sampling_attempts": 2.75,
|
| 3610 |
+
"dapo/sampling_efficiency": 45.20833333333333,
|
| 3611 |
+
"dapo/total_prompts_processed": 16.5,
|
| 3612 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3613 |
+
"epoch": 0.21714285714285714,
|
| 3614 |
+
"grad_norm": 0.00931188277900219,
|
| 3615 |
+
"kl": 0.0006044209003448486,
|
| 3616 |
+
"learning_rate": 1.0826776744855121e-07,
|
| 3617 |
+
"loss": 0.0024,
|
| 3618 |
+
"reward": 0.5944220442324877,
|
| 3619 |
+
"reward_std": 0.9433802142739296,
|
| 3620 |
+
"step": 190
|
| 3621 |
+
},
|
| 3622 |
+
{
|
| 3623 |
+
"clip_fraction": 0.0,
|
| 3624 |
+
"completion_length": 2601.7569427490234,
|
| 3625 |
+
"dapo/avg_reward_std": 0.3233232215046883,
|
| 3626 |
+
"dapo/filter_reward_index": 0.0,
|
| 3627 |
+
"dapo/kept_prompts_ratio": 0.49166667386889457,
|
| 3628 |
+
"dapo/num_sampling_attempts": 2.5,
|
| 3629 |
+
"dapo/sampling_efficiency": 49.375,
|
| 3630 |
+
"dapo/total_prompts_processed": 15.0,
|
| 3631 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3632 |
+
"epoch": 0.21828571428571428,
|
| 3633 |
+
"grad_norm": 0.011869938112795353,
|
| 3634 |
+
"kl": 0.0006383061408996582,
|
| 3635 |
+
"learning_rate": 1.068365111445064e-07,
|
| 3636 |
+
"loss": 0.0221,
|
| 3637 |
+
"reward": 0.5644997656345367,
|
| 3638 |
+
"reward_std": 0.9473884925246239,
|
| 3639 |
+
"step": 191
|
| 3640 |
+
},
|
| 3641 |
+
{
|
| 3642 |
+
"clip_fraction": 0.0,
|
| 3643 |
+
"completion_length": 1624.8541564941406,
|
| 3644 |
+
"dapo/avg_reward_std": 0.33193936944007874,
|
| 3645 |
+
"dapo/filter_reward_index": 0.0,
|
| 3646 |
+
"dapo/kept_prompts_ratio": 0.46969697692177514,
|
| 3647 |
+
"dapo/num_sampling_attempts": 2.75,
|
| 3648 |
+
"dapo/sampling_efficiency": 44.791666666666664,
|
| 3649 |
+
"dapo/total_prompts_processed": 16.5,
|
| 3650 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3651 |
+
"epoch": 0.21942857142857142,
|
| 3652 |
+
"grad_norm": 0.011828861199319363,
|
| 3653 |
+
"kl": 0.0003381967544555664,
|
| 3654 |
+
"learning_rate": 1.0554024673218806e-07,
|
| 3655 |
+
"loss": -0.0125,
|
| 3656 |
+
"reward": 0.7034952798858285,
|
| 3657 |
+
"reward_std": 0.9275326952338219,
|
| 3658 |
+
"step": 192
|
| 3659 |
+
},
|
| 3660 |
+
{
|
| 3661 |
+
"clip_fraction": 0.0,
|
| 3662 |
+
"completion_length": 2333.607650756836,
|
| 3663 |
+
"dapo/avg_reward_std": 0.4260722654206412,
|
| 3664 |
+
"dapo/filter_reward_index": 0.0,
|
| 3665 |
+
"dapo/kept_prompts_ratio": 0.6309523891125407,
|
| 3666 |
+
"dapo/num_sampling_attempts": 1.75,
|
| 3667 |
+
"dapo/sampling_efficiency": 70.83333333333333,
|
| 3668 |
+
"dapo/total_prompts_processed": 10.5,
|
| 3669 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3670 |
+
"epoch": 0.22057142857142858,
|
| 3671 |
+
"grad_norm": 0.010871903039515018,
|
| 3672 |
+
"kl": 0.0005550980567932129,
|
| 3673 |
+
"learning_rate": 1.0437936906629334e-07,
|
| 3674 |
+
"loss": -0.004,
|
| 3675 |
+
"reward": 0.4316184278577566,
|
| 3676 |
+
"reward_std": 0.9555172920227051,
|
| 3677 |
+
"step": 193
|
| 3678 |
+
},
|
| 3679 |
+
{
|
| 3680 |
+
"clip_fraction": 0.0,
|
| 3681 |
+
"completion_length": 2939.9097442626953,
|
| 3682 |
+
"dapo/avg_reward_std": 0.2783619257119986,
|
| 3683 |
+
"dapo/filter_reward_index": 0.0,
|
| 3684 |
+
"dapo/kept_prompts_ratio": 0.3846153932122084,
|
| 3685 |
+
"dapo/num_sampling_attempts": 3.25,
|
| 3686 |
+
"dapo/sampling_efficiency": 47.39583333333333,
|
| 3687 |
+
"dapo/total_prompts_processed": 19.5,
|
| 3688 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3689 |
+
"epoch": 0.22171428571428572,
|
| 3690 |
+
"grad_norm": 0.014206220395863056,
|
| 3691 |
+
"kl": 0.0007078647613525391,
|
| 3692 |
+
"learning_rate": 1.0335423176140511e-07,
|
| 3693 |
+
"loss": 0.0805,
|
| 3694 |
+
"reward": 0.7283875979483128,
|
| 3695 |
+
"reward_std": 0.9719515442848206,
|
| 3696 |
+
"step": 194
|
| 3697 |
+
},
|
| 3698 |
+
{
|
| 3699 |
+
"clip_fraction": 0.0,
|
| 3700 |
+
"completion_length": 1945.9653244018555,
|
| 3701 |
+
"dapo/avg_reward_std": 0.3208765654187453,
|
| 3702 |
+
"dapo/filter_reward_index": 0.0,
|
| 3703 |
+
"dapo/kept_prompts_ratio": 0.5438596551355562,
|
| 3704 |
+
"dapo/num_sampling_attempts": 2.375,
|
| 3705 |
+
"dapo/sampling_efficiency": 60.416666666666664,
|
| 3706 |
+
"dapo/total_prompts_processed": 14.25,
|
| 3707 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3708 |
+
"epoch": 0.22285714285714286,
|
| 3709 |
+
"grad_norm": 0.015090257860720158,
|
| 3710 |
+
"kl": 0.000569462776184082,
|
| 3711 |
+
"learning_rate": 1.0246514708427701e-07,
|
| 3712 |
+
"loss": -0.021,
|
| 3713 |
+
"reward": 0.5579635920003057,
|
| 3714 |
+
"reward_std": 0.9634370356798172,
|
| 3715 |
+
"step": 195
|
| 3716 |
+
},
|
| 3717 |
+
{
|
| 3718 |
+
"clip_fraction": 0.0,
|
| 3719 |
+
"completion_length": 2212.5902709960938,
|
| 3720 |
+
"dapo/avg_reward_std": 0.23615881362382105,
|
| 3721 |
+
"dapo/filter_reward_index": 0.0,
|
| 3722 |
+
"dapo/kept_prompts_ratio": 0.2696078498573864,
|
| 3723 |
+
"dapo/num_sampling_attempts": 4.25,
|
| 3724 |
+
"dapo/sampling_efficiency": 27.916666666666664,
|
| 3725 |
+
"dapo/total_prompts_processed": 25.5,
|
| 3726 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3727 |
+
"epoch": 0.224,
|
| 3728 |
+
"grad_norm": 0.012650169432163239,
|
| 3729 |
+
"kl": 0.0005346536636352539,
|
| 3730 |
+
"learning_rate": 1.017123858587145e-07,
|
| 3731 |
+
"loss": 0.0756,
|
| 3732 |
+
"reward": 0.6994661018252373,
|
| 3733 |
+
"reward_std": 0.9281085133552551,
|
| 3734 |
+
"step": 196
|
| 3735 |
+
},
|
| 3736 |
+
{
|
| 3737 |
+
"clip_fraction": 0.0,
|
| 3738 |
+
"completion_length": 2392.7742919921875,
|
| 3739 |
+
"dapo/avg_reward_std": 0.3088900530338287,
|
| 3740 |
+
"dapo/filter_reward_index": 0.0,
|
| 3741 |
+
"dapo/kept_prompts_ratio": 0.406666676402092,
|
| 3742 |
+
"dapo/num_sampling_attempts": 3.125,
|
| 3743 |
+
"dapo/sampling_efficiency": 45.3125,
|
| 3744 |
+
"dapo/total_prompts_processed": 18.75,
|
| 3745 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3746 |
+
"epoch": 0.22514285714285714,
|
| 3747 |
+
"grad_norm": 0.01346337329596281,
|
| 3748 |
+
"kl": 0.0006176233291625977,
|
| 3749 |
+
"learning_rate": 1.0109617738307911e-07,
|
| 3750 |
+
"loss": 0.0523,
|
| 3751 |
+
"reward": 0.6644653081893921,
|
| 3752 |
+
"reward_std": 0.9385305866599083,
|
| 3753 |
+
"step": 197
|
| 3754 |
+
},
|
| 3755 |
+
{
|
| 3756 |
+
"clip_fraction": 0.0,
|
| 3757 |
+
"completion_length": 2743.819465637207,
|
| 3758 |
+
"dapo/avg_reward_std": 0.3153854298591614,
|
| 3759 |
+
"dapo/filter_reward_index": 0.0,
|
| 3760 |
+
"dapo/kept_prompts_ratio": 0.4133333432674408,
|
| 3761 |
+
"dapo/num_sampling_attempts": 3.125,
|
| 3762 |
+
"dapo/sampling_efficiency": 43.75,
|
| 3763 |
+
"dapo/total_prompts_processed": 18.75,
|
| 3764 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3765 |
+
"epoch": 0.22628571428571428,
|
| 3766 |
+
"grad_norm": 0.010797293856739998,
|
| 3767 |
+
"kl": 0.000672459602355957,
|
| 3768 |
+
"learning_rate": 1.0061670936044178e-07,
|
| 3769 |
+
"loss": 0.04,
|
| 3770 |
+
"reward": 0.5658168056979775,
|
| 3771 |
+
"reward_std": 0.9682240337133408,
|
| 3772 |
+
"step": 198
|
| 3773 |
+
},
|
| 3774 |
+
{
|
| 3775 |
+
"clip_fraction": 0.0,
|
| 3776 |
+
"completion_length": 2336.80558013916,
|
| 3777 |
+
"dapo/avg_reward_std": 0.3246711401835732,
|
| 3778 |
+
"dapo/filter_reward_index": 0.0,
|
| 3779 |
+
"dapo/kept_prompts_ratio": 0.4855072530715362,
|
| 3780 |
+
"dapo/num_sampling_attempts": 2.875,
|
| 3781 |
+
"dapo/sampling_efficiency": 41.666666666666664,
|
| 3782 |
+
"dapo/total_prompts_processed": 17.25,
|
| 3783 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3784 |
+
"epoch": 0.22742857142857142,
|
| 3785 |
+
"grad_norm": 0.011765834875404835,
|
| 3786 |
+
"kl": 0.00055694580078125,
|
| 3787 |
+
"learning_rate": 1.002741278414069e-07,
|
| 3788 |
+
"loss": 0.0308,
|
| 3789 |
+
"reward": 0.6460054386407137,
|
| 3790 |
+
"reward_std": 0.9711420610547066,
|
| 3791 |
+
"step": 199
|
| 3792 |
+
},
|
| 3793 |
+
{
|
| 3794 |
+
"clip_fraction": 0.0,
|
| 3795 |
+
"completion_length": 2571.1875228881836,
|
| 3796 |
+
"dapo/avg_reward_std": 0.29997331152359646,
|
| 3797 |
+
"dapo/filter_reward_index": 0.0,
|
| 3798 |
+
"dapo/kept_prompts_ratio": 0.486111119389534,
|
| 3799 |
+
"dapo/num_sampling_attempts": 3.0,
|
| 3800 |
+
"dapo/sampling_efficiency": 39.285714285714285,
|
| 3801 |
+
"dapo/total_prompts_processed": 18.0,
|
| 3802 |
+
"dapo/valid_prompts_collected": 6.0,
|
| 3803 |
+
"epoch": 0.22857142857142856,
|
| 3804 |
+
"grad_norm": 0.009876573458313942,
|
| 3805 |
+
"kl": 0.0005443096160888672,
|
| 3806 |
+
"learning_rate": 1.0006853717962393e-07,
|
| 3807 |
+
"loss": 0.0268,
|
| 3808 |
+
"reward": 0.5957941338419914,
|
| 3809 |
+
"reward_std": 0.992652915418148,
|
| 3810 |
+
"step": 200
|
| 3811 |
+
},
|
| 3812 |
+
{
|
| 3813 |
+
"epoch": 0.22857142857142856,
|
| 3814 |
+
"step": 200,
|
| 3815 |
"total_flos": 0.0,
|
| 3816 |
+
"train_loss": 0.01698429927288089,
|
| 3817 |
+
"train_runtime": 137940.7556,
|
| 3818 |
+
"train_samples_per_second": 0.07,
|
| 3819 |
"train_steps_per_second": 0.001
|
| 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,
|