Instructions to use usr256864/ee_gol with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use usr256864/ee_gol with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("HiTZ/GoLLIE-7B") model = PeftModel.from_pretrained(base_model, "usr256864/ee_gol") - Transformers
How to use usr256864/ee_gol with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="usr256864/ee_gol")# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("usr256864/ee_gol", device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use usr256864/ee_gol with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "usr256864/ee_gol" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "usr256864/ee_gol", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/usr256864/ee_gol
- SGLang
How to use usr256864/ee_gol with SGLang:
Install from pip and serve model
# Install SGLang from pip: pip install sglang # Start the SGLang server: python3 -m sglang.launch_server \ --model-path "usr256864/ee_gol" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "usr256864/ee_gol", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker images
docker run --gpus all \ --shm-size 32g \ -p 30000:30000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ --env "HF_TOKEN=<secret>" \ --ipc=host \ lmsysorg/sglang:latest \ python3 -m sglang.launch_server \ --model-path "usr256864/ee_gol" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "usr256864/ee_gol", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use usr256864/ee_gol with Docker Model Runner:
docker model run hf.co/usr256864/ee_gol
Upload folder using huggingface_hub
Browse files- adapter_model.safetensors +1 -1
- optimizer.pt +1 -1
- rng_state_0.pth +1 -1
- rng_state_2.pth +1 -1
- rng_state_3.pth +1 -1
- rng_state_4.pth +1 -1
- rng_state_5.pth +1 -1
- rng_state_6.pth +1 -1
- rng_state_7.pth +1 -1
- scheduler.pt +1 -1
- trainer_state.json +933 -3
adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 16794200
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:17a1eaa0c2de60b9844d1cc03aff521bbfb4581f7ab3366f9a89e9fae89535ce
|
| 3 |
size 16794200
|
optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 33664331
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6320d8db318aaf99c6b8796d5ef702c67d497a785359999a3379e6800c2346f1
|
| 3 |
size 33664331
|
rng_state_0.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 16325
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a71878d32742b6bb1d7e6f320668e9057eb306b367d448e89261acf1e34b5969
|
| 3 |
size 16325
|
rng_state_2.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 16389
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5451043be89709e3b8e5cde8a1bcc9ff167dd30ef2ebab491d7f2ee7bc224b36
|
| 3 |
size 16389
|
rng_state_3.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 16389
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b32717ec55a8a59baafd0f43933b1acddf22eef248d2231fd3347b42465fa607
|
| 3 |
size 16389
|
rng_state_4.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 16389
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0b613070a232b30a4f8e0d72fb403479556a5ed6861d080a034f21d7ee69c003
|
| 3 |
size 16389
|
rng_state_5.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 16389
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1035ef0576c9b6d7d98a15586cc2c3549a5dfa7fa16332c1ee1fb668547c0fee
|
| 3 |
size 16389
|
rng_state_6.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 16389
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:683799cf324fe7fb2fc880b968b1ca888155d8bb2c08cf2a0760ed3802a8e5fe
|
| 3 |
size 16389
|
rng_state_7.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 16325
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d212ffe5a372e10e4ef7f41fb63a3acf5e7faa0d500fe86883a93de2e93e8f25
|
| 3 |
size 16325
|
scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1465
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f275d9fe3bf7acf1cd62850a289b5eeef9d0373b5a2b0c248e6f9ebb8c2512e6
|
| 3 |
size 1465
|
trainer_state.json
CHANGED
|
@@ -2,9 +2,9 @@
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
-
"epoch":
|
| 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,
|
|
@@ -3758,11 +3758,941 @@
|
|
| 3758 |
"rewards/format_reward_fn/mean": 0.8654715418815613,
|
| 3759 |
"rewards/format_reward_fn/std": 0.3263047467917204,
|
| 3760 |
"step": 2000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3761 |
}
|
| 3762 |
],
|
| 3763 |
"logging_steps": 16,
|
| 3764 |
"max_steps": 10290,
|
| 3765 |
-
"num_input_tokens_seen":
|
| 3766 |
"num_train_epochs": 10,
|
| 3767 |
"save_steps": 500,
|
| 3768 |
"stateful_callbacks": {
|
|
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 2.4295432458697763,
|
| 6 |
"eval_steps": 500,
|
| 7 |
+
"global_step": 2500,
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
|
|
| 3758 |
"rewards/format_reward_fn/mean": 0.8654715418815613,
|
| 3759 |
"rewards/format_reward_fn/std": 0.3263047467917204,
|
| 3760 |
"step": 2000
|
| 3761 |
+
},
|
| 3762 |
+
{
|
| 3763 |
+
"clip_ratio/high_max": 0.0,
|
| 3764 |
+
"clip_ratio/high_mean": 0.0,
|
| 3765 |
+
"clip_ratio/low_mean": 0.0,
|
| 3766 |
+
"clip_ratio/low_min": 0.0,
|
| 3767 |
+
"clip_ratio/region_mean": 0.0,
|
| 3768 |
+
"completions/clipped_ratio": 0.0732421875,
|
| 3769 |
+
"completions/max_length": 255.6875,
|
| 3770 |
+
"completions/max_terminated_length": 246.0625,
|
| 3771 |
+
"completions/mean_length": 203.2724609375,
|
| 3772 |
+
"completions/mean_terminated_length": 199.26644325256348,
|
| 3773 |
+
"completions/min_length": 149.0625,
|
| 3774 |
+
"completions/min_terminated_length": 149.0625,
|
| 3775 |
+
"entropy": 0.08016827004030347,
|
| 3776 |
+
"epoch": 1.9591836734693877,
|
| 3777 |
+
"frac_reward_zero_std": 0.36328125,
|
| 3778 |
+
"grad_norm": 0.28214797377586365,
|
| 3779 |
+
"learning_rate": 5e-05,
|
| 3780 |
+
"loss": 0.0028,
|
| 3781 |
+
"num_tokens": 163027312.0,
|
| 3782 |
+
"reward": 10.820812225341797,
|
| 3783 |
+
"reward_std": 0.9312072917819023,
|
| 3784 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8814437240362167,
|
| 3785 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.2926495522260666,
|
| 3786 |
+
"rewards/event_reward_fn/mean": 9.048828125,
|
| 3787 |
+
"rewards/event_reward_fn/std": 5.542583703994751,
|
| 3788 |
+
"rewards/format_reward_fn/mean": 0.890540361404419,
|
| 3789 |
+
"rewards/format_reward_fn/std": 0.2912682769820094,
|
| 3790 |
+
"step": 2016
|
| 3791 |
+
},
|
| 3792 |
+
{
|
| 3793 |
+
"clip_ratio/high_max": 0.0,
|
| 3794 |
+
"clip_ratio/high_mean": 0.0,
|
| 3795 |
+
"clip_ratio/low_mean": 0.0,
|
| 3796 |
+
"clip_ratio/low_min": 0.0,
|
| 3797 |
+
"clip_ratio/region_mean": 0.0,
|
| 3798 |
+
"completions/clipped_ratio": 0.080078125,
|
| 3799 |
+
"completions/max_length": 255.8125,
|
| 3800 |
+
"completions/max_terminated_length": 251.4375,
|
| 3801 |
+
"completions/mean_length": 210.033203125,
|
| 3802 |
+
"completions/mean_terminated_length": 206.1150426864624,
|
| 3803 |
+
"completions/min_length": 157.75,
|
| 3804 |
+
"completions/min_terminated_length": 157.75,
|
| 3805 |
+
"entropy": 0.07839876413345337,
|
| 3806 |
+
"epoch": 1.9747327502429544,
|
| 3807 |
+
"frac_reward_zero_std": 0.34765625,
|
| 3808 |
+
"grad_norm": 0.1044137179851532,
|
| 3809 |
+
"learning_rate": 5e-05,
|
| 3810 |
+
"loss": -0.0002,
|
| 3811 |
+
"num_tokens": 164352274.0,
|
| 3812 |
+
"reward": 11.44101220369339,
|
| 3813 |
+
"reward_std": 0.9166238941252232,
|
| 3814 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8769194670021534,
|
| 3815 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.2891712624114007,
|
| 3816 |
+
"rewards/event_reward_fn/mean": 9.6767578125,
|
| 3817 |
+
"rewards/event_reward_fn/std": 5.10790191590786,
|
| 3818 |
+
"rewards/format_reward_fn/mean": 0.8873349130153656,
|
| 3819 |
+
"rewards/format_reward_fn/std": 0.2897007022984326,
|
| 3820 |
+
"step": 2032
|
| 3821 |
+
},
|
| 3822 |
+
{
|
| 3823 |
+
"clip_ratio/high_max": 0.0,
|
| 3824 |
+
"clip_ratio/high_mean": 0.0,
|
| 3825 |
+
"clip_ratio/low_mean": 0.0,
|
| 3826 |
+
"clip_ratio/low_min": 0.0,
|
| 3827 |
+
"clip_ratio/region_mean": 0.0,
|
| 3828 |
+
"completions/clipped_ratio": 0.0849609375,
|
| 3829 |
+
"completions/max_length": 255.125,
|
| 3830 |
+
"completions/max_terminated_length": 249.125,
|
| 3831 |
+
"completions/mean_length": 207.380859375,
|
| 3832 |
+
"completions/mean_terminated_length": 202.79765510559082,
|
| 3833 |
+
"completions/min_length": 162.625,
|
| 3834 |
+
"completions/min_terminated_length": 162.625,
|
| 3835 |
+
"entropy": 0.07377080479636788,
|
| 3836 |
+
"epoch": 1.990281827016521,
|
| 3837 |
+
"frac_reward_zero_std": 0.37890625,
|
| 3838 |
+
"grad_norm": 0.16017131507396698,
|
| 3839 |
+
"learning_rate": 5e-05,
|
| 3840 |
+
"loss": 0.0026,
|
| 3841 |
+
"num_tokens": 165649512.0,
|
| 3842 |
+
"reward": 11.277841091156006,
|
| 3843 |
+
"reward_std": 0.9176982510834932,
|
| 3844 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8867846131324768,
|
| 3845 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.275547455297783,
|
| 3846 |
+
"rewards/event_reward_fn/mean": 9.4931640625,
|
| 3847 |
+
"rewards/event_reward_fn/std": 5.252700716257095,
|
| 3848 |
+
"rewards/format_reward_fn/mean": 0.8978923298418522,
|
| 3849 |
+
"rewards/format_reward_fn/std": 0.27646369859576225,
|
| 3850 |
+
"step": 2048
|
| 3851 |
+
},
|
| 3852 |
+
{
|
| 3853 |
+
"clip_ratio/high_max": 0.0,
|
| 3854 |
+
"clip_ratio/high_mean": 0.0,
|
| 3855 |
+
"clip_ratio/low_mean": 0.0,
|
| 3856 |
+
"clip_ratio/low_min": 0.0,
|
| 3857 |
+
"clip_ratio/region_mean": 0.0,
|
| 3858 |
+
"completions/clipped_ratio": 0.0537109375,
|
| 3859 |
+
"completions/max_length": 255.8125,
|
| 3860 |
+
"completions/max_terminated_length": 249.6875,
|
| 3861 |
+
"completions/mean_length": 204.119140625,
|
| 3862 |
+
"completions/mean_terminated_length": 201.15938472747803,
|
| 3863 |
+
"completions/min_length": 153.9375,
|
| 3864 |
+
"completions/min_terminated_length": 153.9375,
|
| 3865 |
+
"entropy": 0.08090341417118907,
|
| 3866 |
+
"epoch": 2.0058309037900877,
|
| 3867 |
+
"frac_reward_zero_std": 0.3203125,
|
| 3868 |
+
"grad_norm": 0.14778926968574524,
|
| 3869 |
+
"learning_rate": 5e-05,
|
| 3870 |
+
"loss": -0.0041,
|
| 3871 |
+
"num_tokens": 166928214.0,
|
| 3872 |
+
"reward": 10.784139513969421,
|
| 3873 |
+
"reward_std": 0.845683254301548,
|
| 3874 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.9085673242807388,
|
| 3875 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.22585340135265142,
|
| 3876 |
+
"rewards/event_reward_fn/mean": 8.958984375,
|
| 3877 |
+
"rewards/event_reward_fn/std": 5.853749170899391,
|
| 3878 |
+
"rewards/format_reward_fn/mean": 0.9165879562497139,
|
| 3879 |
+
"rewards/format_reward_fn/std": 0.2265966208651662,
|
| 3880 |
+
"step": 2064
|
| 3881 |
+
},
|
| 3882 |
+
{
|
| 3883 |
+
"clip_ratio/high_max": 0.0,
|
| 3884 |
+
"clip_ratio/high_mean": 0.0,
|
| 3885 |
+
"clip_ratio/low_mean": 0.0,
|
| 3886 |
+
"clip_ratio/low_min": 0.0,
|
| 3887 |
+
"clip_ratio/region_mean": 0.0,
|
| 3888 |
+
"completions/clipped_ratio": 0.107421875,
|
| 3889 |
+
"completions/max_length": 255.9375,
|
| 3890 |
+
"completions/max_terminated_length": 252.5625,
|
| 3891 |
+
"completions/mean_length": 217.998046875,
|
| 3892 |
+
"completions/mean_terminated_length": 213.3584222793579,
|
| 3893 |
+
"completions/min_length": 168.625,
|
| 3894 |
+
"completions/min_terminated_length": 168.625,
|
| 3895 |
+
"entropy": 0.07874821173027158,
|
| 3896 |
+
"epoch": 2.021379980563654,
|
| 3897 |
+
"frac_reward_zero_std": 0.3359375,
|
| 3898 |
+
"grad_norm": 0.2152583748102188,
|
| 3899 |
+
"learning_rate": 5e-05,
|
| 3900 |
+
"loss": 0.004,
|
| 3901 |
+
"num_tokens": 168321184.0,
|
| 3902 |
+
"reward": 11.825278103351593,
|
| 3903 |
+
"reward_std": 0.8600351363420486,
|
| 3904 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8512045294046402,
|
| 3905 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.3247975427657366,
|
| 3906 |
+
"rewards/event_reward_fn/mean": 10.119140625,
|
| 3907 |
+
"rewards/event_reward_fn/std": 5.912083759903908,
|
| 3908 |
+
"rewards/format_reward_fn/mean": 0.8549330346286297,
|
| 3909 |
+
"rewards/format_reward_fn/std": 0.32650260720402,
|
| 3910 |
+
"step": 2080
|
| 3911 |
+
},
|
| 3912 |
+
{
|
| 3913 |
+
"clip_ratio/high_max": 0.0,
|
| 3914 |
+
"clip_ratio/high_mean": 0.0,
|
| 3915 |
+
"clip_ratio/low_mean": 0.0,
|
| 3916 |
+
"clip_ratio/low_min": 0.0,
|
| 3917 |
+
"clip_ratio/region_mean": 0.0,
|
| 3918 |
+
"completions/clipped_ratio": 0.103515625,
|
| 3919 |
+
"completions/max_length": 255.6875,
|
| 3920 |
+
"completions/max_terminated_length": 251.375,
|
| 3921 |
+
"completions/mean_length": 215.453125,
|
| 3922 |
+
"completions/mean_terminated_length": 210.9893503189087,
|
| 3923 |
+
"completions/min_length": 169.125,
|
| 3924 |
+
"completions/min_terminated_length": 169.125,
|
| 3925 |
+
"entropy": 0.07949487678706646,
|
| 3926 |
+
"epoch": 2.0369290573372205,
|
| 3927 |
+
"frac_reward_zero_std": 0.26953125,
|
| 3928 |
+
"grad_norm": 0.17531728744506836,
|
| 3929 |
+
"learning_rate": 5e-05,
|
| 3930 |
+
"loss": 0.0025,
|
| 3931 |
+
"num_tokens": 169684588.0,
|
| 3932 |
+
"reward": 11.59043002128601,
|
| 3933 |
+
"reward_std": 0.8503808788955212,
|
| 3934 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8546094782650471,
|
| 3935 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.29849119763821363,
|
| 3936 |
+
"rewards/event_reward_fn/mean": 9.8740234375,
|
| 3937 |
+
"rewards/event_reward_fn/std": 5.960100635886192,
|
| 3938 |
+
"rewards/format_reward_fn/mean": 0.8617968708276749,
|
| 3939 |
+
"rewards/format_reward_fn/std": 0.30179503839462996,
|
| 3940 |
+
"step": 2096
|
| 3941 |
+
},
|
| 3942 |
+
{
|
| 3943 |
+
"clip_ratio/high_max": 0.0,
|
| 3944 |
+
"clip_ratio/high_mean": 0.0,
|
| 3945 |
+
"clip_ratio/low_mean": 0.0,
|
| 3946 |
+
"clip_ratio/low_min": 0.0,
|
| 3947 |
+
"clip_ratio/region_mean": 0.0,
|
| 3948 |
+
"completions/clipped_ratio": 0.048828125,
|
| 3949 |
+
"completions/max_length": 254.9375,
|
| 3950 |
+
"completions/max_terminated_length": 249.9375,
|
| 3951 |
+
"completions/mean_length": 208.2587890625,
|
| 3952 |
+
"completions/mean_terminated_length": 205.81217098236084,
|
| 3953 |
+
"completions/min_length": 162.8125,
|
| 3954 |
+
"completions/min_terminated_length": 162.8125,
|
| 3955 |
+
"entropy": 0.08167764730751514,
|
| 3956 |
+
"epoch": 2.052478134110787,
|
| 3957 |
+
"frac_reward_zero_std": 0.30078125,
|
| 3958 |
+
"grad_norm": 0.17342492938041687,
|
| 3959 |
+
"learning_rate": 5e-05,
|
| 3960 |
+
"loss": -0.0046,
|
| 3961 |
+
"num_tokens": 171007325.0,
|
| 3962 |
+
"reward": 11.406825065612793,
|
| 3963 |
+
"reward_std": 0.9372135195881128,
|
| 3964 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8985451720654964,
|
| 3965 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.2477037919452414,
|
| 3966 |
+
"rewards/event_reward_fn/mean": 9.6015625,
|
| 3967 |
+
"rewards/event_reward_fn/std": 5.504318922758102,
|
| 3968 |
+
"rewards/format_reward_fn/mean": 0.9067176692187786,
|
| 3969 |
+
"rewards/format_reward_fn/std": 0.2522282497957349,
|
| 3970 |
+
"step": 2112
|
| 3971 |
+
},
|
| 3972 |
+
{
|
| 3973 |
+
"clip_ratio/high_max": 0.0,
|
| 3974 |
+
"clip_ratio/high_mean": 0.0,
|
| 3975 |
+
"clip_ratio/low_mean": 0.0,
|
| 3976 |
+
"clip_ratio/low_min": 0.0,
|
| 3977 |
+
"clip_ratio/region_mean": 0.0,
|
| 3978 |
+
"completions/clipped_ratio": 0.0595703125,
|
| 3979 |
+
"completions/max_length": 255.5625,
|
| 3980 |
+
"completions/max_terminated_length": 251.4375,
|
| 3981 |
+
"completions/mean_length": 209.2939453125,
|
| 3982 |
+
"completions/mean_terminated_length": 206.3635711669922,
|
| 3983 |
+
"completions/min_length": 164.3125,
|
| 3984 |
+
"completions/min_terminated_length": 164.3125,
|
| 3985 |
+
"entropy": 0.08719462575390935,
|
| 3986 |
+
"epoch": 2.068027210884354,
|
| 3987 |
+
"frac_reward_zero_std": 0.26171875,
|
| 3988 |
+
"grad_norm": 0.21987557411193848,
|
| 3989 |
+
"learning_rate": 5e-05,
|
| 3990 |
+
"loss": -0.0007,
|
| 3991 |
+
"num_tokens": 172342766.0,
|
| 3992 |
+
"reward": 11.441301941871643,
|
| 3993 |
+
"reward_std": 1.013813877478242,
|
| 3994 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8949129357933998,
|
| 3995 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.23817135416902602,
|
| 3996 |
+
"rewards/event_reward_fn/mean": 9.6416015625,
|
| 3997 |
+
"rewards/event_reward_fn/std": 5.478666722774506,
|
| 3998 |
+
"rewards/format_reward_fn/mean": 0.904787328094244,
|
| 3999 |
+
"rewards/format_reward_fn/std": 0.24113686219789088,
|
| 4000 |
+
"step": 2128
|
| 4001 |
+
},
|
| 4002 |
+
{
|
| 4003 |
+
"clip_ratio/high_max": 0.0,
|
| 4004 |
+
"clip_ratio/high_mean": 0.0,
|
| 4005 |
+
"clip_ratio/low_mean": 0.0,
|
| 4006 |
+
"clip_ratio/low_min": 0.0,
|
| 4007 |
+
"clip_ratio/region_mean": 0.0,
|
| 4008 |
+
"completions/clipped_ratio": 0.0751953125,
|
| 4009 |
+
"completions/max_length": 254.6875,
|
| 4010 |
+
"completions/max_terminated_length": 251.3125,
|
| 4011 |
+
"completions/mean_length": 214.8193359375,
|
| 4012 |
+
"completions/mean_terminated_length": 211.62517070770264,
|
| 4013 |
+
"completions/min_length": 171.75,
|
| 4014 |
+
"completions/min_terminated_length": 171.75,
|
| 4015 |
+
"entropy": 0.08591812197118998,
|
| 4016 |
+
"epoch": 2.0835762876579205,
|
| 4017 |
+
"frac_reward_zero_std": 0.25,
|
| 4018 |
+
"grad_norm": 0.12275879085063934,
|
| 4019 |
+
"learning_rate": 5e-05,
|
| 4020 |
+
"loss": 0.0019,
|
| 4021 |
+
"num_tokens": 173671761.0,
|
| 4022 |
+
"reward": 11.561739206314087,
|
| 4023 |
+
"reward_std": 0.9223730489611626,
|
| 4024 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8886811174452305,
|
| 4025 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.24229650711640716,
|
| 4026 |
+
"rewards/event_reward_fn/mean": 9.76953125,
|
| 4027 |
+
"rewards/event_reward_fn/std": 5.791984856128693,
|
| 4028 |
+
"rewards/format_reward_fn/mean": 0.9035267867147923,
|
| 4029 |
+
"rewards/format_reward_fn/std": 0.23866780381649733,
|
| 4030 |
+
"step": 2144
|
| 4031 |
+
},
|
| 4032 |
+
{
|
| 4033 |
+
"clip_ratio/high_max": 0.0,
|
| 4034 |
+
"clip_ratio/high_mean": 0.0,
|
| 4035 |
+
"clip_ratio/low_mean": 0.0,
|
| 4036 |
+
"clip_ratio/low_min": 0.0,
|
| 4037 |
+
"clip_ratio/region_mean": 0.0,
|
| 4038 |
+
"completions/clipped_ratio": 0.0439453125,
|
| 4039 |
+
"completions/max_length": 253.125,
|
| 4040 |
+
"completions/max_terminated_length": 248.3125,
|
| 4041 |
+
"completions/mean_length": 208.9638671875,
|
| 4042 |
+
"completions/mean_terminated_length": 206.84466552734375,
|
| 4043 |
+
"completions/min_length": 168.875,
|
| 4044 |
+
"completions/min_terminated_length": 168.875,
|
| 4045 |
+
"entropy": 0.08218491962179542,
|
| 4046 |
+
"epoch": 2.0991253644314867,
|
| 4047 |
+
"frac_reward_zero_std": 0.3203125,
|
| 4048 |
+
"grad_norm": 0.2145640254020691,
|
| 4049 |
+
"learning_rate": 5e-05,
|
| 4050 |
+
"loss": 0.0021,
|
| 4051 |
+
"num_tokens": 175000300.0,
|
| 4052 |
+
"reward": 11.832590639591217,
|
| 4053 |
+
"reward_std": 0.8098492994904518,
|
| 4054 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.9014225117862225,
|
| 4055 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.24191926792263985,
|
| 4056 |
+
"rewards/event_reward_fn/mean": 10.02734375,
|
| 4057 |
+
"rewards/event_reward_fn/std": 5.564895883202553,
|
| 4058 |
+
"rewards/format_reward_fn/mean": 0.9038244113326073,
|
| 4059 |
+
"rewards/format_reward_fn/std": 0.2553019989281893,
|
| 4060 |
+
"step": 2160
|
| 4061 |
+
},
|
| 4062 |
+
{
|
| 4063 |
+
"clip_ratio/high_max": 0.0,
|
| 4064 |
+
"clip_ratio/high_mean": 0.0,
|
| 4065 |
+
"clip_ratio/low_mean": 0.0,
|
| 4066 |
+
"clip_ratio/low_min": 0.0,
|
| 4067 |
+
"clip_ratio/region_mean": 0.0,
|
| 4068 |
+
"completions/clipped_ratio": 0.060546875,
|
| 4069 |
+
"completions/max_length": 255.0625,
|
| 4070 |
+
"completions/max_terminated_length": 251.375,
|
| 4071 |
+
"completions/mean_length": 212.275390625,
|
| 4072 |
+
"completions/mean_terminated_length": 209.4294490814209,
|
| 4073 |
+
"completions/min_length": 172.875,
|
| 4074 |
+
"completions/min_terminated_length": 172.875,
|
| 4075 |
+
"entropy": 0.08632302051410079,
|
| 4076 |
+
"epoch": 2.1146744412050533,
|
| 4077 |
+
"frac_reward_zero_std": 0.27734375,
|
| 4078 |
+
"grad_norm": 0.12819725275039673,
|
| 4079 |
+
"learning_rate": 5e-05,
|
| 4080 |
+
"loss": -0.0027,
|
| 4081 |
+
"num_tokens": 176397358.0,
|
| 4082 |
+
"reward": 11.750000655651093,
|
| 4083 |
+
"reward_std": 1.0350622907280922,
|
| 4084 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8829724602401257,
|
| 4085 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.26738651166670024,
|
| 4086 |
+
"rewards/event_reward_fn/mean": 9.9794921875,
|
| 4087 |
+
"rewards/event_reward_fn/std": 5.499541476368904,
|
| 4088 |
+
"rewards/format_reward_fn/mean": 0.8875359706580639,
|
| 4089 |
+
"rewards/format_reward_fn/std": 0.27499296236783266,
|
| 4090 |
+
"step": 2176
|
| 4091 |
+
},
|
| 4092 |
+
{
|
| 4093 |
+
"clip_ratio/high_max": 0.0,
|
| 4094 |
+
"clip_ratio/high_mean": 0.0,
|
| 4095 |
+
"clip_ratio/low_mean": 0.0,
|
| 4096 |
+
"clip_ratio/low_min": 0.0,
|
| 4097 |
+
"clip_ratio/region_mean": 0.0,
|
| 4098 |
+
"completions/clipped_ratio": 0.0771484375,
|
| 4099 |
+
"completions/max_length": 253.9375,
|
| 4100 |
+
"completions/max_terminated_length": 248.5,
|
| 4101 |
+
"completions/mean_length": 211.8388671875,
|
| 4102 |
+
"completions/mean_terminated_length": 208.1127519607544,
|
| 4103 |
+
"completions/min_length": 169.125,
|
| 4104 |
+
"completions/min_terminated_length": 169.125,
|
| 4105 |
+
"entropy": 0.09303951309993863,
|
| 4106 |
+
"epoch": 2.13022351797862,
|
| 4107 |
+
"frac_reward_zero_std": 0.28515625,
|
| 4108 |
+
"grad_norm": 0.22441552579402924,
|
| 4109 |
+
"learning_rate": 5e-05,
|
| 4110 |
+
"loss": -0.0018,
|
| 4111 |
+
"num_tokens": 177793437.0,
|
| 4112 |
+
"reward": 11.069317996501923,
|
| 4113 |
+
"reward_std": 1.0597262904047966,
|
| 4114 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8652172312140465,
|
| 4115 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.28404899896122515,
|
| 4116 |
+
"rewards/event_reward_fn/mean": 9.337890625,
|
| 4117 |
+
"rewards/event_reward_fn/std": 5.563613697886467,
|
| 4118 |
+
"rewards/format_reward_fn/mean": 0.8662101663649082,
|
| 4119 |
+
"rewards/format_reward_fn/std": 0.2897696476429701,
|
| 4120 |
+
"step": 2192
|
| 4121 |
+
},
|
| 4122 |
+
{
|
| 4123 |
+
"clip_ratio/high_max": 0.0,
|
| 4124 |
+
"clip_ratio/high_mean": 0.0,
|
| 4125 |
+
"clip_ratio/low_mean": 0.0,
|
| 4126 |
+
"clip_ratio/low_min": 0.0,
|
| 4127 |
+
"clip_ratio/region_mean": 0.0,
|
| 4128 |
+
"completions/clipped_ratio": 0.1318359375,
|
| 4129 |
+
"completions/max_length": 256.0,
|
| 4130 |
+
"completions/max_terminated_length": 253.8125,
|
| 4131 |
+
"completions/mean_length": 223.1318359375,
|
| 4132 |
+
"completions/mean_terminated_length": 218.348069190979,
|
| 4133 |
+
"completions/min_length": 178.5625,
|
| 4134 |
+
"completions/min_terminated_length": 178.5625,
|
| 4135 |
+
"entropy": 0.08407490560784936,
|
| 4136 |
+
"epoch": 2.1457725947521866,
|
| 4137 |
+
"frac_reward_zero_std": 0.2734375,
|
| 4138 |
+
"grad_norm": 0.20042170584201813,
|
| 4139 |
+
"learning_rate": 5e-05,
|
| 4140 |
+
"loss": 0.0042,
|
| 4141 |
+
"num_tokens": 179158840.0,
|
| 4142 |
+
"reward": 11.508206486701965,
|
| 4143 |
+
"reward_std": 0.8616610933095217,
|
| 4144 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8278331160545349,
|
| 4145 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.32595117576420307,
|
| 4146 |
+
"rewards/event_reward_fn/mean": 9.8291015625,
|
| 4147 |
+
"rewards/event_reward_fn/std": 5.2151205241680145,
|
| 4148 |
+
"rewards/format_reward_fn/mean": 0.8512718565762043,
|
| 4149 |
+
"rewards/format_reward_fn/std": 0.33205954916775227,
|
| 4150 |
+
"step": 2208
|
| 4151 |
+
},
|
| 4152 |
+
{
|
| 4153 |
+
"clip_ratio/high_max": 0.0,
|
| 4154 |
+
"clip_ratio/high_mean": 0.0,
|
| 4155 |
+
"clip_ratio/low_mean": 0.0,
|
| 4156 |
+
"clip_ratio/low_min": 0.0,
|
| 4157 |
+
"clip_ratio/region_mean": 0.0,
|
| 4158 |
+
"completions/clipped_ratio": 0.05078125,
|
| 4159 |
+
"completions/max_length": 254.6875,
|
| 4160 |
+
"completions/max_terminated_length": 247.75,
|
| 4161 |
+
"completions/mean_length": 205.880859375,
|
| 4162 |
+
"completions/mean_terminated_length": 203.19688606262207,
|
| 4163 |
+
"completions/min_length": 166.625,
|
| 4164 |
+
"completions/min_terminated_length": 166.625,
|
| 4165 |
+
"entropy": 0.07565503777004778,
|
| 4166 |
+
"epoch": 2.1613216715257533,
|
| 4167 |
+
"frac_reward_zero_std": 0.31640625,
|
| 4168 |
+
"grad_norm": 0.15642929077148438,
|
| 4169 |
+
"learning_rate": 5e-05,
|
| 4170 |
+
"loss": -0.0029,
|
| 4171 |
+
"num_tokens": 180532374.0,
|
| 4172 |
+
"reward": 11.076000154018402,
|
| 4173 |
+
"reward_std": 0.8950711917132139,
|
| 4174 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.9075712524354458,
|
| 4175 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.24491250491701066,
|
| 4176 |
+
"rewards/event_reward_fn/mean": 9.248046875,
|
| 4177 |
+
"rewards/event_reward_fn/std": 5.206828847527504,
|
| 4178 |
+
"rewards/format_reward_fn/mean": 0.9203820116817951,
|
| 4179 |
+
"rewards/format_reward_fn/std": 0.24691881332546473,
|
| 4180 |
+
"step": 2224
|
| 4181 |
+
},
|
| 4182 |
+
{
|
| 4183 |
+
"clip_ratio/high_max": 0.0,
|
| 4184 |
+
"clip_ratio/high_mean": 0.0,
|
| 4185 |
+
"clip_ratio/low_mean": 0.0,
|
| 4186 |
+
"clip_ratio/low_min": 0.0,
|
| 4187 |
+
"clip_ratio/region_mean": 0.0,
|
| 4188 |
+
"completions/clipped_ratio": 0.0498046875,
|
| 4189 |
+
"completions/max_length": 253.6875,
|
| 4190 |
+
"completions/max_terminated_length": 248.5625,
|
| 4191 |
+
"completions/mean_length": 204.2744140625,
|
| 4192 |
+
"completions/mean_terminated_length": 201.6161012649536,
|
| 4193 |
+
"completions/min_length": 162.0625,
|
| 4194 |
+
"completions/min_terminated_length": 162.0625,
|
| 4195 |
+
"entropy": 0.07292898930609226,
|
| 4196 |
+
"epoch": 2.17687074829932,
|
| 4197 |
+
"frac_reward_zero_std": 0.35546875,
|
| 4198 |
+
"grad_norm": 0.20497596263885498,
|
| 4199 |
+
"learning_rate": 5e-05,
|
| 4200 |
+
"loss": -0.0018,
|
| 4201 |
+
"num_tokens": 181834967.0,
|
| 4202 |
+
"reward": 11.745607078075409,
|
| 4203 |
+
"reward_std": 0.847535029053688,
|
| 4204 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8826499357819557,
|
| 4205 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.2826285846531391,
|
| 4206 |
+
"rewards/event_reward_fn/mean": 9.96875,
|
| 4207 |
+
"rewards/event_reward_fn/std": 5.558514207601547,
|
| 4208 |
+
"rewards/format_reward_fn/mean": 0.8942071311175823,
|
| 4209 |
+
"rewards/format_reward_fn/std": 0.2861539525911212,
|
| 4210 |
+
"step": 2240
|
| 4211 |
+
},
|
| 4212 |
+
{
|
| 4213 |
+
"clip_ratio/high_max": 0.0,
|
| 4214 |
+
"clip_ratio/high_mean": 0.0,
|
| 4215 |
+
"clip_ratio/low_mean": 0.0,
|
| 4216 |
+
"clip_ratio/low_min": 0.0,
|
| 4217 |
+
"clip_ratio/region_mean": 0.0,
|
| 4218 |
+
"completions/clipped_ratio": 0.0615234375,
|
| 4219 |
+
"completions/max_length": 255.1875,
|
| 4220 |
+
"completions/max_terminated_length": 250.8125,
|
| 4221 |
+
"completions/mean_length": 212.1318359375,
|
| 4222 |
+
"completions/mean_terminated_length": 209.29450035095215,
|
| 4223 |
+
"completions/min_length": 174.1875,
|
| 4224 |
+
"completions/min_terminated_length": 174.1875,
|
| 4225 |
+
"entropy": 0.07543019764125347,
|
| 4226 |
+
"epoch": 2.192419825072886,
|
| 4227 |
+
"frac_reward_zero_std": 0.32421875,
|
| 4228 |
+
"grad_norm": 0.10211238265037537,
|
| 4229 |
+
"learning_rate": 5e-05,
|
| 4230 |
+
"loss": -0.0013,
|
| 4231 |
+
"num_tokens": 183156678.0,
|
| 4232 |
+
"reward": 11.885134816169739,
|
| 4233 |
+
"reward_std": 0.7728028316050768,
|
| 4234 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.9035747610032558,
|
| 4235 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.24522930546663702,
|
| 4236 |
+
"rewards/event_reward_fn/mean": 10.0634765625,
|
| 4237 |
+
"rewards/event_reward_fn/std": 5.761056482791901,
|
| 4238 |
+
"rewards/format_reward_fn/mean": 0.9180834665894508,
|
| 4239 |
+
"rewards/format_reward_fn/std": 0.24741819500923157,
|
| 4240 |
+
"step": 2256
|
| 4241 |
+
},
|
| 4242 |
+
{
|
| 4243 |
+
"clip_ratio/high_max": 0.0,
|
| 4244 |
+
"clip_ratio/high_mean": 0.0,
|
| 4245 |
+
"clip_ratio/low_mean": 0.0,
|
| 4246 |
+
"clip_ratio/low_min": 0.0,
|
| 4247 |
+
"clip_ratio/region_mean": 0.0,
|
| 4248 |
+
"completions/clipped_ratio": 0.125,
|
| 4249 |
+
"completions/max_length": 255.5,
|
| 4250 |
+
"completions/max_terminated_length": 252.125,
|
| 4251 |
+
"completions/mean_length": 217.482421875,
|
| 4252 |
+
"completions/mean_terminated_length": 212.08176708221436,
|
| 4253 |
+
"completions/min_length": 176.875,
|
| 4254 |
+
"completions/min_terminated_length": 176.875,
|
| 4255 |
+
"entropy": 0.07963799126446247,
|
| 4256 |
+
"epoch": 2.207968901846453,
|
| 4257 |
+
"frac_reward_zero_std": 0.30078125,
|
| 4258 |
+
"grad_norm": 0.2214186191558838,
|
| 4259 |
+
"learning_rate": 5e-05,
|
| 4260 |
+
"loss": 0.0004,
|
| 4261 |
+
"num_tokens": 184464124.0,
|
| 4262 |
+
"reward": 11.618954241275787,
|
| 4263 |
+
"reward_std": 0.9292504880577326,
|
| 4264 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8386580236256123,
|
| 4265 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.32745907083153725,
|
| 4266 |
+
"rewards/event_reward_fn/mean": 9.9248046875,
|
| 4267 |
+
"rewards/event_reward_fn/std": 6.44027054309845,
|
| 4268 |
+
"rewards/format_reward_fn/mean": 0.8554915376007557,
|
| 4269 |
+
"rewards/format_reward_fn/std": 0.3294975752942264,
|
| 4270 |
+
"step": 2272
|
| 4271 |
+
},
|
| 4272 |
+
{
|
| 4273 |
+
"clip_ratio/high_max": 0.0,
|
| 4274 |
+
"clip_ratio/high_mean": 0.0,
|
| 4275 |
+
"clip_ratio/low_mean": 0.0,
|
| 4276 |
+
"clip_ratio/low_min": 0.0,
|
| 4277 |
+
"clip_ratio/region_mean": 0.0,
|
| 4278 |
+
"completions/clipped_ratio": 0.0498046875,
|
| 4279 |
+
"completions/max_length": 254.4375,
|
| 4280 |
+
"completions/max_terminated_length": 249.375,
|
| 4281 |
+
"completions/mean_length": 206.373046875,
|
| 4282 |
+
"completions/mean_terminated_length": 203.74801063537598,
|
| 4283 |
+
"completions/min_length": 163.625,
|
| 4284 |
+
"completions/min_terminated_length": 163.625,
|
| 4285 |
+
"entropy": 0.0740656116977334,
|
| 4286 |
+
"epoch": 2.2235179786200194,
|
| 4287 |
+
"frac_reward_zero_std": 0.39453125,
|
| 4288 |
+
"grad_norm": 0.15276597440242767,
|
| 4289 |
+
"learning_rate": 5e-05,
|
| 4290 |
+
"loss": -0.0032,
|
| 4291 |
+
"num_tokens": 185812542.0,
|
| 4292 |
+
"reward": 11.359130620956421,
|
| 4293 |
+
"reward_std": 0.6935872584581375,
|
| 4294 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.9103271588683128,
|
| 4295 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.23343394591938704,
|
| 4296 |
+
"rewards/event_reward_fn/mean": 9.5244140625,
|
| 4297 |
+
"rewards/event_reward_fn/std": 5.075364321470261,
|
| 4298 |
+
"rewards/format_reward_fn/mean": 0.9243892580270767,
|
| 4299 |
+
"rewards/format_reward_fn/std": 0.2387167038396001,
|
| 4300 |
+
"step": 2288
|
| 4301 |
+
},
|
| 4302 |
+
{
|
| 4303 |
+
"clip_ratio/high_max": 0.0,
|
| 4304 |
+
"clip_ratio/high_mean": 0.0,
|
| 4305 |
+
"clip_ratio/low_mean": 0.0,
|
| 4306 |
+
"clip_ratio/low_min": 0.0,
|
| 4307 |
+
"clip_ratio/region_mean": 0.0,
|
| 4308 |
+
"completions/clipped_ratio": 0.0625,
|
| 4309 |
+
"completions/max_length": 254.5,
|
| 4310 |
+
"completions/max_terminated_length": 250.9375,
|
| 4311 |
+
"completions/mean_length": 210.4521484375,
|
| 4312 |
+
"completions/mean_terminated_length": 207.25661849975586,
|
| 4313 |
+
"completions/min_length": 169.8125,
|
| 4314 |
+
"completions/min_terminated_length": 169.8125,
|
| 4315 |
+
"entropy": 0.07526633841916919,
|
| 4316 |
+
"epoch": 2.239067055393586,
|
| 4317 |
+
"frac_reward_zero_std": 0.390625,
|
| 4318 |
+
"grad_norm": 0.17438335716724396,
|
| 4319 |
+
"learning_rate": 5e-05,
|
| 4320 |
+
"loss": 0.0007,
|
| 4321 |
+
"num_tokens": 187174565.0,
|
| 4322 |
+
"reward": 11.606376469135284,
|
| 4323 |
+
"reward_std": 0.8281035982072353,
|
| 4324 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8901838399469852,
|
| 4325 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.2384987068362534,
|
| 4326 |
+
"rewards/event_reward_fn/mean": 9.806640625,
|
| 4327 |
+
"rewards/event_reward_fn/std": 5.699323073029518,
|
| 4328 |
+
"rewards/format_reward_fn/mean": 0.9095519706606865,
|
| 4329 |
+
"rewards/format_reward_fn/std": 0.2393078247550875,
|
| 4330 |
+
"step": 2304
|
| 4331 |
+
},
|
| 4332 |
+
{
|
| 4333 |
+
"clip_ratio/high_max": 0.0,
|
| 4334 |
+
"clip_ratio/high_mean": 0.0,
|
| 4335 |
+
"clip_ratio/low_mean": 0.0,
|
| 4336 |
+
"clip_ratio/low_min": 0.0,
|
| 4337 |
+
"clip_ratio/region_mean": 0.0,
|
| 4338 |
+
"completions/clipped_ratio": 0.0322265625,
|
| 4339 |
+
"completions/max_length": 249.0,
|
| 4340 |
+
"completions/max_terminated_length": 243.6875,
|
| 4341 |
+
"completions/mean_length": 203.9052734375,
|
| 4342 |
+
"completions/mean_terminated_length": 202.31891632080078,
|
| 4343 |
+
"completions/min_length": 159.4375,
|
| 4344 |
+
"completions/min_terminated_length": 159.4375,
|
| 4345 |
+
"entropy": 0.07370157795958221,
|
| 4346 |
+
"epoch": 2.2546161321671527,
|
| 4347 |
+
"frac_reward_zero_std": 0.390625,
|
| 4348 |
+
"grad_norm": 0.12816278636455536,
|
| 4349 |
+
"learning_rate": 5e-05,
|
| 4350 |
+
"loss": -0.0022,
|
| 4351 |
+
"num_tokens": 188422072.0,
|
| 4352 |
+
"reward": 11.180573999881744,
|
| 4353 |
+
"reward_std": 0.7691474985331297,
|
| 4354 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.9313375540077686,
|
| 4355 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.16660339455120265,
|
| 4356 |
+
"rewards/event_reward_fn/mean": 9.302734375,
|
| 4357 |
+
"rewards/event_reward_fn/std": 5.455415144562721,
|
| 4358 |
+
"rewards/format_reward_fn/mean": 0.946502048522234,
|
| 4359 |
+
"rewards/format_reward_fn/std": 0.16397117311134934,
|
| 4360 |
+
"step": 2320
|
| 4361 |
+
},
|
| 4362 |
+
{
|
| 4363 |
+
"clip_ratio/high_max": 0.0,
|
| 4364 |
+
"clip_ratio/high_mean": 0.0,
|
| 4365 |
+
"clip_ratio/low_mean": 0.0,
|
| 4366 |
+
"clip_ratio/low_min": 0.0,
|
| 4367 |
+
"clip_ratio/region_mean": 0.0,
|
| 4368 |
+
"completions/clipped_ratio": 0.0419921875,
|
| 4369 |
+
"completions/max_length": 250.875,
|
| 4370 |
+
"completions/max_terminated_length": 246.125,
|
| 4371 |
+
"completions/mean_length": 201.8203125,
|
| 4372 |
+
"completions/mean_terminated_length": 199.46971607208252,
|
| 4373 |
+
"completions/min_length": 160.6875,
|
| 4374 |
+
"completions/min_terminated_length": 160.6875,
|
| 4375 |
+
"entropy": 0.0713472084607929,
|
| 4376 |
+
"epoch": 2.2701652089407194,
|
| 4377 |
+
"frac_reward_zero_std": 0.45703125,
|
| 4378 |
+
"grad_norm": 0.15345498919487,
|
| 4379 |
+
"learning_rate": 5e-05,
|
| 4380 |
+
"loss": 0.0001,
|
| 4381 |
+
"num_tokens": 189749628.0,
|
| 4382 |
+
"reward": 11.676891207695007,
|
| 4383 |
+
"reward_std": 0.7641248423606157,
|
| 4384 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.9076020866632462,
|
| 4385 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.25307453935965896,
|
| 4386 |
+
"rewards/event_reward_fn/mean": 9.8544921875,
|
| 4387 |
+
"rewards/event_reward_fn/std": 5.772381603717804,
|
| 4388 |
+
"rewards/format_reward_fn/mean": 0.9147970490157604,
|
| 4389 |
+
"rewards/format_reward_fn/std": 0.25177112873643637,
|
| 4390 |
+
"step": 2336
|
| 4391 |
+
},
|
| 4392 |
+
{
|
| 4393 |
+
"clip_ratio/high_max": 0.0,
|
| 4394 |
+
"clip_ratio/high_mean": 0.0,
|
| 4395 |
+
"clip_ratio/low_mean": 0.0,
|
| 4396 |
+
"clip_ratio/low_min": 0.0,
|
| 4397 |
+
"clip_ratio/region_mean": 0.0,
|
| 4398 |
+
"completions/clipped_ratio": 0.052734375,
|
| 4399 |
+
"completions/max_length": 253.0,
|
| 4400 |
+
"completions/max_terminated_length": 245.9375,
|
| 4401 |
+
"completions/mean_length": 206.1015625,
|
| 4402 |
+
"completions/mean_terminated_length": 203.27950382232666,
|
| 4403 |
+
"completions/min_length": 162.625,
|
| 4404 |
+
"completions/min_terminated_length": 162.625,
|
| 4405 |
+
"entropy": 0.06717114523053169,
|
| 4406 |
+
"epoch": 2.2857142857142856,
|
| 4407 |
+
"frac_reward_zero_std": 0.3984375,
|
| 4408 |
+
"grad_norm": 0.14800839126110077,
|
| 4409 |
+
"learning_rate": 5e-05,
|
| 4410 |
+
"loss": -0.0018,
|
| 4411 |
+
"num_tokens": 191059400.0,
|
| 4412 |
+
"reward": 11.288125574588776,
|
| 4413 |
+
"reward_std": 0.8216591961681843,
|
| 4414 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.9098481498658657,
|
| 4415 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.2527642482891679,
|
| 4416 |
+
"rewards/event_reward_fn/mean": 9.45703125,
|
| 4417 |
+
"rewards/event_reward_fn/std": 6.013585805892944,
|
| 4418 |
+
"rewards/format_reward_fn/mean": 0.9212462790310383,
|
| 4419 |
+
"rewards/format_reward_fn/std": 0.255017863586545,
|
| 4420 |
+
"step": 2352
|
| 4421 |
+
},
|
| 4422 |
+
{
|
| 4423 |
+
"clip_ratio/high_max": 0.0,
|
| 4424 |
+
"clip_ratio/high_mean": 0.0,
|
| 4425 |
+
"clip_ratio/low_mean": 0.0,
|
| 4426 |
+
"clip_ratio/low_min": 0.0,
|
| 4427 |
+
"clip_ratio/region_mean": 0.0,
|
| 4428 |
+
"completions/clipped_ratio": 0.119140625,
|
| 4429 |
+
"completions/max_length": 256.0,
|
| 4430 |
+
"completions/max_terminated_length": 252.3125,
|
| 4431 |
+
"completions/mean_length": 222.3125,
|
| 4432 |
+
"completions/mean_terminated_length": 217.90027904510498,
|
| 4433 |
+
"completions/min_length": 175.625,
|
| 4434 |
+
"completions/min_terminated_length": 175.625,
|
| 4435 |
+
"entropy": 0.0761020069476217,
|
| 4436 |
+
"epoch": 2.3012633624878522,
|
| 4437 |
+
"frac_reward_zero_std": 0.39453125,
|
| 4438 |
+
"grad_norm": 0.24630595743656158,
|
| 4439 |
+
"learning_rate": 5e-05,
|
| 4440 |
+
"loss": 0.0025,
|
| 4441 |
+
"num_tokens": 192350156.0,
|
| 4442 |
+
"reward": 11.402837812900543,
|
| 4443 |
+
"reward_std": 0.8105970397591591,
|
| 4444 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8388476483523846,
|
| 4445 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.32804084848612547,
|
| 4446 |
+
"rewards/event_reward_fn/mean": 9.7060546875,
|
| 4447 |
+
"rewards/event_reward_fn/std": 5.547398820519447,
|
| 4448 |
+
"rewards/format_reward_fn/mean": 0.8579354099929333,
|
| 4449 |
+
"rewards/format_reward_fn/std": 0.33059023320674896,
|
| 4450 |
+
"step": 2368
|
| 4451 |
+
},
|
| 4452 |
+
{
|
| 4453 |
+
"clip_ratio/high_max": 0.0,
|
| 4454 |
+
"clip_ratio/high_mean": 0.0,
|
| 4455 |
+
"clip_ratio/low_mean": 0.0,
|
| 4456 |
+
"clip_ratio/low_min": 0.0,
|
| 4457 |
+
"clip_ratio/region_mean": 0.0,
|
| 4458 |
+
"completions/clipped_ratio": 0.0537109375,
|
| 4459 |
+
"completions/max_length": 253.0625,
|
| 4460 |
+
"completions/max_terminated_length": 248.125,
|
| 4461 |
+
"completions/mean_length": 211.1552734375,
|
| 4462 |
+
"completions/mean_terminated_length": 208.6152687072754,
|
| 4463 |
+
"completions/min_length": 169.0,
|
| 4464 |
+
"completions/min_terminated_length": 169.0,
|
| 4465 |
+
"entropy": 0.07612166181206703,
|
| 4466 |
+
"epoch": 2.316812439261419,
|
| 4467 |
+
"frac_reward_zero_std": 0.390625,
|
| 4468 |
+
"grad_norm": 0.08691530674695969,
|
| 4469 |
+
"learning_rate": 5e-05,
|
| 4470 |
+
"loss": 0.0024,
|
| 4471 |
+
"num_tokens": 193665307.0,
|
| 4472 |
+
"reward": 11.495616167783737,
|
| 4473 |
+
"reward_std": 0.7880423050373793,
|
| 4474 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.902044016867876,
|
| 4475 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.23614092892967165,
|
| 4476 |
+
"rewards/event_reward_fn/mean": 9.6865234375,
|
| 4477 |
+
"rewards/event_reward_fn/std": 6.028887152671814,
|
| 4478 |
+
"rewards/format_reward_fn/mean": 0.9070489220321178,
|
| 4479 |
+
"rewards/format_reward_fn/std": 0.24180734669789672,
|
| 4480 |
+
"step": 2384
|
| 4481 |
+
},
|
| 4482 |
+
{
|
| 4483 |
+
"clip_ratio/high_max": 0.0,
|
| 4484 |
+
"clip_ratio/high_mean": 0.0,
|
| 4485 |
+
"clip_ratio/low_mean": 0.0,
|
| 4486 |
+
"clip_ratio/low_min": 0.0,
|
| 4487 |
+
"clip_ratio/region_mean": 0.0,
|
| 4488 |
+
"completions/clipped_ratio": 0.0185546875,
|
| 4489 |
+
"completions/max_length": 246.4375,
|
| 4490 |
+
"completions/max_terminated_length": 242.0625,
|
| 4491 |
+
"completions/mean_length": 193.6845703125,
|
| 4492 |
+
"completions/mean_terminated_length": 192.51368141174316,
|
| 4493 |
+
"completions/min_length": 148.6875,
|
| 4494 |
+
"completions/min_terminated_length": 148.6875,
|
| 4495 |
+
"entropy": 0.07832195260562003,
|
| 4496 |
+
"epoch": 2.3323615160349855,
|
| 4497 |
+
"frac_reward_zero_std": 0.40234375,
|
| 4498 |
+
"grad_norm": 0.1274639219045639,
|
| 4499 |
+
"learning_rate": 5e-05,
|
| 4500 |
+
"loss": -0.0048,
|
| 4501 |
+
"num_tokens": 194920676.0,
|
| 4502 |
+
"reward": 11.40235447883606,
|
| 4503 |
+
"reward_std": 0.7331925742328167,
|
| 4504 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.9401694796979427,
|
| 4505 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.17933110590092838,
|
| 4506 |
+
"rewards/event_reward_fn/mean": 9.5068359375,
|
| 4507 |
+
"rewards/event_reward_fn/std": 5.196439817547798,
|
| 4508 |
+
"rewards/format_reward_fn/mean": 0.9553493969142437,
|
| 4509 |
+
"rewards/format_reward_fn/std": 0.172615127870813,
|
| 4510 |
+
"step": 2400
|
| 4511 |
+
},
|
| 4512 |
+
{
|
| 4513 |
+
"clip_ratio/high_max": 0.0,
|
| 4514 |
+
"clip_ratio/high_mean": 0.0,
|
| 4515 |
+
"clip_ratio/low_mean": 0.0,
|
| 4516 |
+
"clip_ratio/low_min": 0.0,
|
| 4517 |
+
"clip_ratio/region_mean": 0.0,
|
| 4518 |
+
"completions/clipped_ratio": 0.0234375,
|
| 4519 |
+
"completions/max_length": 246.9375,
|
| 4520 |
+
"completions/max_terminated_length": 243.6875,
|
| 4521 |
+
"completions/mean_length": 197.6689453125,
|
| 4522 |
+
"completions/mean_terminated_length": 196.28760814666748,
|
| 4523 |
+
"completions/min_length": 158.0,
|
| 4524 |
+
"completions/min_terminated_length": 158.0,
|
| 4525 |
+
"entropy": 0.07638159766793251,
|
| 4526 |
+
"epoch": 2.347910592808552,
|
| 4527 |
+
"frac_reward_zero_std": 0.3671875,
|
| 4528 |
+
"grad_norm": 0.18019680678844452,
|
| 4529 |
+
"learning_rate": 5e-05,
|
| 4530 |
+
"loss": -0.002,
|
| 4531 |
+
"num_tokens": 196239697.0,
|
| 4532 |
+
"reward": 11.403923392295837,
|
| 4533 |
+
"reward_std": 0.7415321134030819,
|
| 4534 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.9454397931694984,
|
| 4535 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.15600819129031152,
|
| 4536 |
+
"rewards/event_reward_fn/mean": 9.4951171875,
|
| 4537 |
+
"rewards/event_reward_fn/std": 5.713589310646057,
|
| 4538 |
+
"rewards/format_reward_fn/mean": 0.9633664786815643,
|
| 4539 |
+
"rewards/format_reward_fn/std": 0.14195893332362175,
|
| 4540 |
+
"step": 2416
|
| 4541 |
+
},
|
| 4542 |
+
{
|
| 4543 |
+
"clip_ratio/high_max": 0.0,
|
| 4544 |
+
"clip_ratio/high_mean": 0.0,
|
| 4545 |
+
"clip_ratio/low_mean": 0.0,
|
| 4546 |
+
"clip_ratio/low_min": 0.0,
|
| 4547 |
+
"clip_ratio/region_mean": 0.0,
|
| 4548 |
+
"completions/clipped_ratio": 0.0927734375,
|
| 4549 |
+
"completions/max_length": 255.75,
|
| 4550 |
+
"completions/max_terminated_length": 252.25,
|
| 4551 |
+
"completions/mean_length": 216.240234375,
|
| 4552 |
+
"completions/mean_terminated_length": 212.49735260009766,
|
| 4553 |
+
"completions/min_length": 171.4375,
|
| 4554 |
+
"completions/min_terminated_length": 171.4375,
|
| 4555 |
+
"entropy": 0.07729306910187006,
|
| 4556 |
+
"epoch": 2.3634596695821184,
|
| 4557 |
+
"frac_reward_zero_std": 0.35546875,
|
| 4558 |
+
"grad_norm": 0.1351216435432434,
|
| 4559 |
+
"learning_rate": 5e-05,
|
| 4560 |
+
"loss": 0.0052,
|
| 4561 |
+
"num_tokens": 197585395.0,
|
| 4562 |
+
"reward": 11.078547358512878,
|
| 4563 |
+
"reward_std": 0.8885079212486744,
|
| 4564 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8728507719933987,
|
| 4565 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.2855207370594144,
|
| 4566 |
+
"rewards/event_reward_fn/mean": 9.3125,
|
| 4567 |
+
"rewards/event_reward_fn/std": 5.477887436747551,
|
| 4568 |
+
"rewards/format_reward_fn/mean": 0.8931966163218021,
|
| 4569 |
+
"rewards/format_reward_fn/std": 0.2850890662521124,
|
| 4570 |
+
"step": 2432
|
| 4571 |
+
},
|
| 4572 |
+
{
|
| 4573 |
+
"clip_ratio/high_max": 0.0,
|
| 4574 |
+
"clip_ratio/high_mean": 0.0,
|
| 4575 |
+
"clip_ratio/low_mean": 0.0,
|
| 4576 |
+
"clip_ratio/low_min": 0.0,
|
| 4577 |
+
"clip_ratio/region_mean": 0.0,
|
| 4578 |
+
"completions/clipped_ratio": 0.1064453125,
|
| 4579 |
+
"completions/max_length": 256.0,
|
| 4580 |
+
"completions/max_terminated_length": 254.25,
|
| 4581 |
+
"completions/mean_length": 222.671875,
|
| 4582 |
+
"completions/mean_terminated_length": 218.8065061569214,
|
| 4583 |
+
"completions/min_length": 170.75,
|
| 4584 |
+
"completions/min_terminated_length": 170.75,
|
| 4585 |
+
"entropy": 0.07714059529826045,
|
| 4586 |
+
"epoch": 2.379008746355685,
|
| 4587 |
+
"frac_reward_zero_std": 0.359375,
|
| 4588 |
+
"grad_norm": 0.14632916450500488,
|
| 4589 |
+
"learning_rate": 5e-05,
|
| 4590 |
+
"loss": 0.0014,
|
| 4591 |
+
"num_tokens": 198898579.0,
|
| 4592 |
+
"reward": 10.960965871810913,
|
| 4593 |
+
"reward_std": 0.8031115736812353,
|
| 4594 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8641464188694954,
|
| 4595 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.3086322648450732,
|
| 4596 |
+
"rewards/event_reward_fn/mean": 9.2119140625,
|
| 4597 |
+
"rewards/event_reward_fn/std": 5.528163373470306,
|
| 4598 |
+
"rewards/format_reward_fn/mean": 0.8849051333963871,
|
| 4599 |
+
"rewards/format_reward_fn/std": 0.31144819781184196,
|
| 4600 |
+
"step": 2448
|
| 4601 |
+
},
|
| 4602 |
+
{
|
| 4603 |
+
"clip_ratio/high_max": 0.0,
|
| 4604 |
+
"clip_ratio/high_mean": 0.0,
|
| 4605 |
+
"clip_ratio/low_mean": 0.0,
|
| 4606 |
+
"clip_ratio/low_min": 0.0,
|
| 4607 |
+
"clip_ratio/region_mean": 0.0,
|
| 4608 |
+
"completions/clipped_ratio": 0.0419921875,
|
| 4609 |
+
"completions/max_length": 251.8125,
|
| 4610 |
+
"completions/max_terminated_length": 247.875,
|
| 4611 |
+
"completions/mean_length": 210.1025390625,
|
| 4612 |
+
"completions/mean_terminated_length": 208.211838722229,
|
| 4613 |
+
"completions/min_length": 161.5625,
|
| 4614 |
+
"completions/min_terminated_length": 161.5625,
|
| 4615 |
+
"entropy": 0.07773100049234927,
|
| 4616 |
+
"epoch": 2.3945578231292517,
|
| 4617 |
+
"frac_reward_zero_std": 0.37890625,
|
| 4618 |
+
"grad_norm": 0.16118626296520233,
|
| 4619 |
+
"learning_rate": 5e-05,
|
| 4620 |
+
"loss": -0.0013,
|
| 4621 |
+
"num_tokens": 200201952.0,
|
| 4622 |
+
"reward": 10.956557631492615,
|
| 4623 |
+
"reward_std": 0.7257527317851782,
|
| 4624 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.9221826978027821,
|
| 4625 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.20850562094710767,
|
| 4626 |
+
"rewards/event_reward_fn/mean": 9.0927734375,
|
| 4627 |
+
"rewards/event_reward_fn/std": 5.140145808458328,
|
| 4628 |
+
"rewards/format_reward_fn/mean": 0.9416015632450581,
|
| 4629 |
+
"rewards/format_reward_fn/std": 0.1953780883923173,
|
| 4630 |
+
"step": 2464
|
| 4631 |
+
},
|
| 4632 |
+
{
|
| 4633 |
+
"clip_ratio/high_max": 0.0,
|
| 4634 |
+
"clip_ratio/high_mean": 0.0,
|
| 4635 |
+
"clip_ratio/low_mean": 0.0,
|
| 4636 |
+
"clip_ratio/low_min": 0.0,
|
| 4637 |
+
"clip_ratio/region_mean": 0.0,
|
| 4638 |
+
"completions/clipped_ratio": 0.0224609375,
|
| 4639 |
+
"completions/max_length": 245.9375,
|
| 4640 |
+
"completions/max_terminated_length": 240.75,
|
| 4641 |
+
"completions/mean_length": 202.029296875,
|
| 4642 |
+
"completions/mean_terminated_length": 200.80841064453125,
|
| 4643 |
+
"completions/min_length": 159.5,
|
| 4644 |
+
"completions/min_terminated_length": 159.5,
|
| 4645 |
+
"entropy": 0.0753114647231996,
|
| 4646 |
+
"epoch": 2.4101068999028183,
|
| 4647 |
+
"frac_reward_zero_std": 0.41015625,
|
| 4648 |
+
"grad_norm": 0.15847159922122955,
|
| 4649 |
+
"learning_rate": 5e-05,
|
| 4650 |
+
"loss": -0.0026,
|
| 4651 |
+
"num_tokens": 201522430.0,
|
| 4652 |
+
"reward": 11.323036313056946,
|
| 4653 |
+
"reward_std": 0.7887390460819006,
|
| 4654 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.9379655607044697,
|
| 4655 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.17426727525889874,
|
| 4656 |
+
"rewards/event_reward_fn/mean": 9.4345703125,
|
| 4657 |
+
"rewards/event_reward_fn/std": 5.223158270120621,
|
| 4658 |
+
"rewards/format_reward_fn/mean": 0.9505006894469261,
|
| 4659 |
+
"rewards/format_reward_fn/std": 0.1740011121146381,
|
| 4660 |
+
"step": 2480
|
| 4661 |
+
},
|
| 4662 |
+
{
|
| 4663 |
+
"clip_ratio/high_max": 0.0,
|
| 4664 |
+
"clip_ratio/high_mean": 0.0,
|
| 4665 |
+
"clip_ratio/low_mean": 0.0,
|
| 4666 |
+
"clip_ratio/low_min": 0.0,
|
| 4667 |
+
"clip_ratio/region_mean": 0.0,
|
| 4668 |
+
"completions/clipped_ratio": 0.025390625,
|
| 4669 |
+
"completions/max_length": 250.125,
|
| 4670 |
+
"completions/max_terminated_length": 246.9375,
|
| 4671 |
+
"completions/mean_length": 205.435546875,
|
| 4672 |
+
"completions/mean_terminated_length": 204.02826595306396,
|
| 4673 |
+
"completions/min_length": 164.875,
|
| 4674 |
+
"completions/min_terminated_length": 164.875,
|
| 4675 |
+
"entropy": 0.07844416983425617,
|
| 4676 |
+
"epoch": 2.425655976676385,
|
| 4677 |
+
"frac_reward_zero_std": 0.34375,
|
| 4678 |
+
"grad_norm": 0.0907289907336235,
|
| 4679 |
+
"learning_rate": 5e-05,
|
| 4680 |
+
"loss": -0.0048,
|
| 4681 |
+
"num_tokens": 202805596.0,
|
| 4682 |
+
"reward": 11.417901694774628,
|
| 4683 |
+
"reward_std": 0.9192124493420124,
|
| 4684 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.9360267631709576,
|
| 4685 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.18138159497175366,
|
| 4686 |
+
"rewards/event_reward_fn/mean": 9.52734375,
|
| 4687 |
+
"rewards/event_reward_fn/std": 5.153972968459129,
|
| 4688 |
+
"rewards/format_reward_fn/mean": 0.9545312523841858,
|
| 4689 |
+
"rewards/format_reward_fn/std": 0.17224382143467665,
|
| 4690 |
+
"step": 2496
|
| 4691 |
}
|
| 4692 |
],
|
| 4693 |
"logging_steps": 16,
|
| 4694 |
"max_steps": 10290,
|
| 4695 |
+
"num_input_tokens_seen": 203156199,
|
| 4696 |
"num_train_epochs": 10,
|
| 4697 |
"save_steps": 500,
|
| 4698 |
"stateful_callbacks": {
|