Model save
Browse files- README.md +60 -0
- adapter_model.safetensors +1 -1
- all_results.json +8 -0
- runs/May12_19-12-14_gpu4-119-5/events.out.tfevents.1715505249.gpu4-119-5.2851589.0 +2 -2
- train_results.json +8 -0
- trainer_state.json +1878 -0
README.md
ADDED
|
@@ -0,0 +1,60 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
library_name: peft
|
| 4 |
+
tags:
|
| 5 |
+
- trl
|
| 6 |
+
- dpo
|
| 7 |
+
- generated_from_trainer
|
| 8 |
+
base_model: DUAL-GPO/phi-2-gpo-new-i0
|
| 9 |
+
model-index:
|
| 10 |
+
- name: phi-2-gpo-v6-i1
|
| 11 |
+
results: []
|
| 12 |
+
---
|
| 13 |
+
|
| 14 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 15 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 16 |
+
|
| 17 |
+
# phi-2-gpo-v6-i1
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [DUAL-GPO/phi-2-gpo-new-i0](https://huggingface.co/DUAL-GPO/phi-2-gpo-new-i0) on the None dataset.
|
| 20 |
+
|
| 21 |
+
## Model description
|
| 22 |
+
|
| 23 |
+
More information needed
|
| 24 |
+
|
| 25 |
+
## Intended uses & limitations
|
| 26 |
+
|
| 27 |
+
More information needed
|
| 28 |
+
|
| 29 |
+
## Training and evaluation data
|
| 30 |
+
|
| 31 |
+
More information needed
|
| 32 |
+
|
| 33 |
+
## Training procedure
|
| 34 |
+
|
| 35 |
+
### Training hyperparameters
|
| 36 |
+
|
| 37 |
+
The following hyperparameters were used during training:
|
| 38 |
+
- learning_rate: 5e-06
|
| 39 |
+
- train_batch_size: 4
|
| 40 |
+
- eval_batch_size: 4
|
| 41 |
+
- seed: 42
|
| 42 |
+
- distributed_type: multi-GPU
|
| 43 |
+
- gradient_accumulation_steps: 4
|
| 44 |
+
- total_train_batch_size: 16
|
| 45 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 46 |
+
- lr_scheduler_type: cosine
|
| 47 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 48 |
+
- num_epochs: 1
|
| 49 |
+
|
| 50 |
+
### Training results
|
| 51 |
+
|
| 52 |
+
|
| 53 |
+
|
| 54 |
+
### Framework versions
|
| 55 |
+
|
| 56 |
+
- PEFT 0.7.1
|
| 57 |
+
- Transformers 4.36.2
|
| 58 |
+
- Pytorch 2.1.2+cu121
|
| 59 |
+
- Datasets 2.14.6
|
| 60 |
+
- Tokenizers 0.15.2
|
adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 167807296
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b92a7dfab3d4fa8bf346643c54c27da10de685f346ed4844a446a62f5fba462e
|
| 3 |
size 167807296
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.24134081795175627,
|
| 4 |
+
"train_runtime": 12108.13,
|
| 5 |
+
"train_samples": 21000,
|
| 6 |
+
"train_samples_per_second": 1.734,
|
| 7 |
+
"train_steps_per_second": 0.108
|
| 8 |
+
}
|
runs/May12_19-12-14_gpu4-119-5/events.out.tfevents.1715505249.gpu4-119-5.2851589.0
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a71c5b016668a941218fe8e9d4dbcf18418bbdbfb34932afa63d0eedb9f5320d
|
| 3 |
+
size 88722
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.24134081795175627,
|
| 4 |
+
"train_runtime": 12108.13,
|
| 5 |
+
"train_samples": 21000,
|
| 6 |
+
"train_samples_per_second": 1.734,
|
| 7 |
+
"train_steps_per_second": 0.108
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1878 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9996190476190476,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 1312,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0,
|
| 13 |
+
"learning_rate": 3.787878787878788e-08,
|
| 14 |
+
"logits/chosen": 0.17224127054214478,
|
| 15 |
+
"logits/rejected": 0.18124699592590332,
|
| 16 |
+
"logps/chosen": -379.32623291015625,
|
| 17 |
+
"logps/rejected": -349.5926208496094,
|
| 18 |
+
"loss": 0.2902,
|
| 19 |
+
"rewards/accuracies": 0.0,
|
| 20 |
+
"rewards/chosen": 0.0,
|
| 21 |
+
"rewards/margins": 0.0,
|
| 22 |
+
"rewards/rejected": 0.0,
|
| 23 |
+
"step": 1
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.01,
|
| 27 |
+
"learning_rate": 3.787878787878788e-07,
|
| 28 |
+
"logits/chosen": 0.07856506109237671,
|
| 29 |
+
"logits/rejected": 0.2510358691215515,
|
| 30 |
+
"logps/chosen": -334.9958190917969,
|
| 31 |
+
"logps/rejected": -283.54034423828125,
|
| 32 |
+
"loss": 0.3745,
|
| 33 |
+
"rewards/accuracies": 0.4375,
|
| 34 |
+
"rewards/chosen": 1.0113296411873307e-05,
|
| 35 |
+
"rewards/margins": 5.3244151786202565e-05,
|
| 36 |
+
"rewards/rejected": -4.313084718887694e-05,
|
| 37 |
+
"step": 10
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.02,
|
| 41 |
+
"learning_rate": 7.575757575757576e-07,
|
| 42 |
+
"logits/chosen": 0.07835443317890167,
|
| 43 |
+
"logits/rejected": 0.23690445721149445,
|
| 44 |
+
"logps/chosen": -343.5411682128906,
|
| 45 |
+
"logps/rejected": -300.49774169921875,
|
| 46 |
+
"loss": 0.345,
|
| 47 |
+
"rewards/accuracies": 0.518750011920929,
|
| 48 |
+
"rewards/chosen": 9.991687193178223e-07,
|
| 49 |
+
"rewards/margins": 5.578011041507125e-05,
|
| 50 |
+
"rewards/rejected": -5.478093953570351e-05,
|
| 51 |
+
"step": 20
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.02,
|
| 55 |
+
"learning_rate": 1.1363636363636364e-06,
|
| 56 |
+
"logits/chosen": 0.09362699836492538,
|
| 57 |
+
"logits/rejected": 0.24969105422496796,
|
| 58 |
+
"logps/chosen": -384.11199951171875,
|
| 59 |
+
"logps/rejected": -290.873779296875,
|
| 60 |
+
"loss": 0.3667,
|
| 61 |
+
"rewards/accuracies": 0.53125,
|
| 62 |
+
"rewards/chosen": 0.00015662855003029108,
|
| 63 |
+
"rewards/margins": 9.987165867642034e-06,
|
| 64 |
+
"rewards/rejected": 0.0001466413668822497,
|
| 65 |
+
"step": 30
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.03,
|
| 69 |
+
"learning_rate": 1.5151515151515152e-06,
|
| 70 |
+
"logits/chosen": 0.08275317400693893,
|
| 71 |
+
"logits/rejected": 0.20892572402954102,
|
| 72 |
+
"logps/chosen": -361.24462890625,
|
| 73 |
+
"logps/rejected": -296.53094482421875,
|
| 74 |
+
"loss": 0.3012,
|
| 75 |
+
"rewards/accuracies": 0.512499988079071,
|
| 76 |
+
"rewards/chosen": 0.00019404105842113495,
|
| 77 |
+
"rewards/margins": 0.000132537359604612,
|
| 78 |
+
"rewards/rejected": 6.150371336843818e-05,
|
| 79 |
+
"step": 40
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.04,
|
| 83 |
+
"learning_rate": 1.8939393939393941e-06,
|
| 84 |
+
"logits/chosen": 0.14996236562728882,
|
| 85 |
+
"logits/rejected": 0.2148081511259079,
|
| 86 |
+
"logps/chosen": -339.32916259765625,
|
| 87 |
+
"logps/rejected": -292.900146484375,
|
| 88 |
+
"loss": 0.3391,
|
| 89 |
+
"rewards/accuracies": 0.518750011920929,
|
| 90 |
+
"rewards/chosen": 0.0002820778754539788,
|
| 91 |
+
"rewards/margins": 0.00035016084439121187,
|
| 92 |
+
"rewards/rejected": -6.80829762131907e-05,
|
| 93 |
+
"step": 50
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.05,
|
| 97 |
+
"learning_rate": 2.2727272727272728e-06,
|
| 98 |
+
"logits/chosen": 0.09292325377464294,
|
| 99 |
+
"logits/rejected": 0.25342074036598206,
|
| 100 |
+
"logps/chosen": -357.79498291015625,
|
| 101 |
+
"logps/rejected": -279.56475830078125,
|
| 102 |
+
"loss": 0.3523,
|
| 103 |
+
"rewards/accuracies": 0.643750011920929,
|
| 104 |
+
"rewards/chosen": 0.0005591081571765244,
|
| 105 |
+
"rewards/margins": 0.0007476316532120109,
|
| 106 |
+
"rewards/rejected": -0.00018852358334697783,
|
| 107 |
+
"step": 60
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.05,
|
| 111 |
+
"learning_rate": 2.6515151515151514e-06,
|
| 112 |
+
"logits/chosen": 0.1520170122385025,
|
| 113 |
+
"logits/rejected": 0.23838527500629425,
|
| 114 |
+
"logps/chosen": -344.97381591796875,
|
| 115 |
+
"logps/rejected": -272.6285400390625,
|
| 116 |
+
"loss": 0.333,
|
| 117 |
+
"rewards/accuracies": 0.637499988079071,
|
| 118 |
+
"rewards/chosen": 0.000808493874501437,
|
| 119 |
+
"rewards/margins": 0.0009522804175503552,
|
| 120 |
+
"rewards/rejected": -0.0001437865139450878,
|
| 121 |
+
"step": 70
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.06,
|
| 125 |
+
"learning_rate": 3.0303030303030305e-06,
|
| 126 |
+
"logits/chosen": 0.10541319847106934,
|
| 127 |
+
"logits/rejected": 0.2519112229347229,
|
| 128 |
+
"logps/chosen": -337.99932861328125,
|
| 129 |
+
"logps/rejected": -285.3942565917969,
|
| 130 |
+
"loss": 0.3276,
|
| 131 |
+
"rewards/accuracies": 0.706250011920929,
|
| 132 |
+
"rewards/chosen": 0.0017709163948893547,
|
| 133 |
+
"rewards/margins": 0.0019623935222625732,
|
| 134 |
+
"rewards/rejected": -0.00019147712737321854,
|
| 135 |
+
"step": 80
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.07,
|
| 139 |
+
"learning_rate": 3.409090909090909e-06,
|
| 140 |
+
"logits/chosen": 0.08055099099874496,
|
| 141 |
+
"logits/rejected": 0.23055055737495422,
|
| 142 |
+
"logps/chosen": -358.501220703125,
|
| 143 |
+
"logps/rejected": -301.07373046875,
|
| 144 |
+
"loss": 0.325,
|
| 145 |
+
"rewards/accuracies": 0.71875,
|
| 146 |
+
"rewards/chosen": 0.0031805038452148438,
|
| 147 |
+
"rewards/margins": 0.0035764030180871487,
|
| 148 |
+
"rewards/rejected": -0.0003958995803259313,
|
| 149 |
+
"step": 90
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.08,
|
| 153 |
+
"learning_rate": 3.7878787878787882e-06,
|
| 154 |
+
"logits/chosen": 0.1381727159023285,
|
| 155 |
+
"logits/rejected": 0.36128586530685425,
|
| 156 |
+
"logps/chosen": -408.2177734375,
|
| 157 |
+
"logps/rejected": -295.70391845703125,
|
| 158 |
+
"loss": 0.3551,
|
| 159 |
+
"rewards/accuracies": 0.6875,
|
| 160 |
+
"rewards/chosen": 0.005324442870914936,
|
| 161 |
+
"rewards/margins": 0.006567128002643585,
|
| 162 |
+
"rewards/rejected": -0.00124268583022058,
|
| 163 |
+
"step": 100
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.08,
|
| 167 |
+
"learning_rate": 4.166666666666667e-06,
|
| 168 |
+
"logits/chosen": 0.11063258349895477,
|
| 169 |
+
"logits/rejected": 0.3397473692893982,
|
| 170 |
+
"logps/chosen": -365.5294494628906,
|
| 171 |
+
"logps/rejected": -296.68212890625,
|
| 172 |
+
"loss": 0.3176,
|
| 173 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 174 |
+
"rewards/chosen": 0.0065561323426663876,
|
| 175 |
+
"rewards/margins": 0.009804257191717625,
|
| 176 |
+
"rewards/rejected": -0.0032481239177286625,
|
| 177 |
+
"step": 110
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.09,
|
| 181 |
+
"learning_rate": 4.5454545454545455e-06,
|
| 182 |
+
"logits/chosen": 0.10760724544525146,
|
| 183 |
+
"logits/rejected": 0.3315739035606384,
|
| 184 |
+
"logps/chosen": -369.8674011230469,
|
| 185 |
+
"logps/rejected": -290.553955078125,
|
| 186 |
+
"loss": 0.2862,
|
| 187 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 188 |
+
"rewards/chosen": 0.009247648529708385,
|
| 189 |
+
"rewards/margins": 0.017721759155392647,
|
| 190 |
+
"rewards/rejected": -0.008474110625684261,
|
| 191 |
+
"step": 120
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.1,
|
| 195 |
+
"learning_rate": 4.924242424242425e-06,
|
| 196 |
+
"logits/chosen": 0.19570864737033844,
|
| 197 |
+
"logits/rejected": 0.29995661973953247,
|
| 198 |
+
"logps/chosen": -357.7752685546875,
|
| 199 |
+
"logps/rejected": -298.80865478515625,
|
| 200 |
+
"loss": 0.3091,
|
| 201 |
+
"rewards/accuracies": 0.706250011920929,
|
| 202 |
+
"rewards/chosen": 0.0016540288925170898,
|
| 203 |
+
"rewards/margins": 0.030770784243941307,
|
| 204 |
+
"rewards/rejected": -0.029116755351424217,
|
| 205 |
+
"step": 130
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.11,
|
| 209 |
+
"learning_rate": 4.999432965739786e-06,
|
| 210 |
+
"logits/chosen": 0.18279646337032318,
|
| 211 |
+
"logits/rejected": 0.24193449318408966,
|
| 212 |
+
"logps/chosen": -326.20623779296875,
|
| 213 |
+
"logps/rejected": -340.66204833984375,
|
| 214 |
+
"loss": 0.3146,
|
| 215 |
+
"rewards/accuracies": 0.675000011920929,
|
| 216 |
+
"rewards/chosen": -0.02705274522304535,
|
| 217 |
+
"rewards/margins": 0.038378529250621796,
|
| 218 |
+
"rewards/rejected": -0.06543128192424774,
|
| 219 |
+
"step": 140
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.11,
|
| 223 |
+
"learning_rate": 4.997129829895409e-06,
|
| 224 |
+
"logits/chosen": 0.14213022589683533,
|
| 225 |
+
"logits/rejected": 0.2512076199054718,
|
| 226 |
+
"logps/chosen": -408.73541259765625,
|
| 227 |
+
"logps/rejected": -432.90771484375,
|
| 228 |
+
"loss": 0.2979,
|
| 229 |
+
"rewards/accuracies": 0.71875,
|
| 230 |
+
"rewards/chosen": -0.0731552243232727,
|
| 231 |
+
"rewards/margins": 0.08651129901409149,
|
| 232 |
+
"rewards/rejected": -0.159666508436203,
|
| 233 |
+
"step": 150
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.12,
|
| 237 |
+
"learning_rate": 4.9930567839810125e-06,
|
| 238 |
+
"logits/chosen": 0.17261534929275513,
|
| 239 |
+
"logits/rejected": 0.27892106771469116,
|
| 240 |
+
"logps/chosen": -483.6229553222656,
|
| 241 |
+
"logps/rejected": -527.606201171875,
|
| 242 |
+
"loss": 0.2727,
|
| 243 |
+
"rewards/accuracies": 0.731249988079071,
|
| 244 |
+
"rewards/chosen": -0.121568962931633,
|
| 245 |
+
"rewards/margins": 0.10641299188137054,
|
| 246 |
+
"rewards/rejected": -0.22798196971416473,
|
| 247 |
+
"step": 160
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.13,
|
| 251 |
+
"learning_rate": 4.987216714880929e-06,
|
| 252 |
+
"logits/chosen": 0.19240622222423553,
|
| 253 |
+
"logits/rejected": 0.22908082604408264,
|
| 254 |
+
"logps/chosen": -516.4906005859375,
|
| 255 |
+
"logps/rejected": -531.3421020507812,
|
| 256 |
+
"loss": 0.278,
|
| 257 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 258 |
+
"rewards/chosen": -0.17584487795829773,
|
| 259 |
+
"rewards/margins": 0.10210248082876205,
|
| 260 |
+
"rewards/rejected": -0.27794739603996277,
|
| 261 |
+
"step": 170
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.14,
|
| 265 |
+
"learning_rate": 4.979613761906212e-06,
|
| 266 |
+
"logits/chosen": 0.12470928579568863,
|
| 267 |
+
"logits/rejected": 0.2596343755722046,
|
| 268 |
+
"logps/chosen": -551.5357666015625,
|
| 269 |
+
"logps/rejected": -654.6400756835938,
|
| 270 |
+
"loss": 0.2664,
|
| 271 |
+
"rewards/accuracies": 0.731249988079071,
|
| 272 |
+
"rewards/chosen": -0.21089038252830505,
|
| 273 |
+
"rewards/margins": 0.15304332971572876,
|
| 274 |
+
"rewards/rejected": -0.3639337420463562,
|
| 275 |
+
"step": 180
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.14,
|
| 279 |
+
"learning_rate": 4.970253313860788e-06,
|
| 280 |
+
"logits/chosen": 0.1833106130361557,
|
| 281 |
+
"logits/rejected": 0.2698153257369995,
|
| 282 |
+
"logps/chosen": -562.19677734375,
|
| 283 |
+
"logps/rejected": -634.8347778320312,
|
| 284 |
+
"loss": 0.2832,
|
| 285 |
+
"rewards/accuracies": 0.65625,
|
| 286 |
+
"rewards/chosen": -0.2445872575044632,
|
| 287 |
+
"rewards/margins": 0.11998845636844635,
|
| 288 |
+
"rewards/rejected": -0.36457571387290955,
|
| 289 |
+
"step": 190
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.15,
|
| 293 |
+
"learning_rate": 4.959142005221991e-06,
|
| 294 |
+
"logits/chosen": 0.13080090284347534,
|
| 295 |
+
"logits/rejected": 0.21319513022899628,
|
| 296 |
+
"logps/chosen": -603.390380859375,
|
| 297 |
+
"logps/rejected": -718.3425903320312,
|
| 298 |
+
"loss": 0.2595,
|
| 299 |
+
"rewards/accuracies": 0.768750011920929,
|
| 300 |
+
"rewards/chosen": -0.25654521584510803,
|
| 301 |
+
"rewards/margins": 0.16353540122509003,
|
| 302 |
+
"rewards/rejected": -0.42008060216903687,
|
| 303 |
+
"step": 200
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.16,
|
| 307 |
+
"learning_rate": 4.94628771143819e-06,
|
| 308 |
+
"logits/chosen": 0.17947080731391907,
|
| 309 |
+
"logits/rejected": 0.2654314935207367,
|
| 310 |
+
"logps/chosen": -653.5524291992188,
|
| 311 |
+
"logps/rejected": -726.80126953125,
|
| 312 |
+
"loss": 0.3026,
|
| 313 |
+
"rewards/accuracies": 0.6875,
|
| 314 |
+
"rewards/chosen": -0.2890639305114746,
|
| 315 |
+
"rewards/margins": 0.12797169387340546,
|
| 316 |
+
"rewards/rejected": -0.41703566908836365,
|
| 317 |
+
"step": 210
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.17,
|
| 321 |
+
"learning_rate": 4.931699543346854e-06,
|
| 322 |
+
"logits/chosen": 0.1152615174651146,
|
| 323 |
+
"logits/rejected": 0.25124651193618774,
|
| 324 |
+
"logps/chosen": -588.23046875,
|
| 325 |
+
"logps/rejected": -704.0632934570312,
|
| 326 |
+
"loss": 0.2673,
|
| 327 |
+
"rewards/accuracies": 0.737500011920929,
|
| 328 |
+
"rewards/chosen": -0.2535329759120941,
|
| 329 |
+
"rewards/margins": 0.17576026916503906,
|
| 330 |
+
"rewards/rejected": -0.4292932152748108,
|
| 331 |
+
"step": 220
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.18,
|
| 335 |
+
"learning_rate": 4.9153878407169815e-06,
|
| 336 |
+
"logits/chosen": 0.13138779997825623,
|
| 337 |
+
"logits/rejected": 0.1749398410320282,
|
| 338 |
+
"logps/chosen": -536.669921875,
|
| 339 |
+
"logps/rejected": -619.0616455078125,
|
| 340 |
+
"loss": 0.2609,
|
| 341 |
+
"rewards/accuracies": 0.668749988079071,
|
| 342 |
+
"rewards/chosen": -0.2183808982372284,
|
| 343 |
+
"rewards/margins": 0.12902414798736572,
|
| 344 |
+
"rewards/rejected": -0.3474050462245941,
|
| 345 |
+
"step": 230
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.18,
|
| 349 |
+
"learning_rate": 4.897364164920515e-06,
|
| 350 |
+
"logits/chosen": 0.14722837507724762,
|
| 351 |
+
"logits/rejected": 0.2979207932949066,
|
| 352 |
+
"logps/chosen": -644.4813232421875,
|
| 353 |
+
"logps/rejected": -710.7215576171875,
|
| 354 |
+
"loss": 0.2364,
|
| 355 |
+
"rewards/accuracies": 0.706250011920929,
|
| 356 |
+
"rewards/chosen": -0.2495686262845993,
|
| 357 |
+
"rewards/margins": 0.14611390233039856,
|
| 358 |
+
"rewards/rejected": -0.39568251371383667,
|
| 359 |
+
"step": 240
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.19,
|
| 363 |
+
"learning_rate": 4.8776412907378845e-06,
|
| 364 |
+
"logits/chosen": 0.14406076073646545,
|
| 365 |
+
"logits/rejected": 0.24922068417072296,
|
| 366 |
+
"logps/chosen": -602.2944946289062,
|
| 367 |
+
"logps/rejected": -689.9060668945312,
|
| 368 |
+
"loss": 0.2803,
|
| 369 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 370 |
+
"rewards/chosen": -0.25432199239730835,
|
| 371 |
+
"rewards/margins": 0.14366620779037476,
|
| 372 |
+
"rewards/rejected": -0.3979881703853607,
|
| 373 |
+
"step": 250
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.2,
|
| 377 |
+
"learning_rate": 4.8562331973035396e-06,
|
| 378 |
+
"logits/chosen": 0.14575393497943878,
|
| 379 |
+
"logits/rejected": 0.28581660985946655,
|
| 380 |
+
"logps/chosen": -567.1799926757812,
|
| 381 |
+
"logps/rejected": -649.4906005859375,
|
| 382 |
+
"loss": 0.2541,
|
| 383 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 384 |
+
"rewards/chosen": -0.21804265677928925,
|
| 385 |
+
"rewards/margins": 0.13621987402439117,
|
| 386 |
+
"rewards/rejected": -0.3542625308036804,
|
| 387 |
+
"step": 260
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.21,
|
| 391 |
+
"learning_rate": 4.833155058197842e-06,
|
| 392 |
+
"logits/chosen": 0.20403075218200684,
|
| 393 |
+
"logits/rejected": 0.32390326261520386,
|
| 394 |
+
"logps/chosen": -611.0584106445312,
|
| 395 |
+
"logps/rejected": -643.7525634765625,
|
| 396 |
+
"loss": 0.2827,
|
| 397 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 398 |
+
"rewards/chosen": -0.23078171908855438,
|
| 399 |
+
"rewards/margins": 0.12669074535369873,
|
| 400 |
+
"rewards/rejected": -0.3574724495410919,
|
| 401 |
+
"step": 270
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.21,
|
| 405 |
+
"learning_rate": 4.808423230692374e-06,
|
| 406 |
+
"logits/chosen": 0.16620513796806335,
|
| 407 |
+
"logits/rejected": 0.3157016634941101,
|
| 408 |
+
"logps/chosen": -557.9627075195312,
|
| 409 |
+
"logps/rejected": -638.0758666992188,
|
| 410 |
+
"loss": 0.2304,
|
| 411 |
+
"rewards/accuracies": 0.768750011920929,
|
| 412 |
+
"rewards/chosen": -0.1864185631275177,
|
| 413 |
+
"rewards/margins": 0.15307244658470154,
|
| 414 |
+
"rewards/rejected": -0.33949097990989685,
|
| 415 |
+
"step": 280
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.22,
|
| 419 |
+
"learning_rate": 4.7820552441562625e-06,
|
| 420 |
+
"logits/chosen": 0.18919572234153748,
|
| 421 |
+
"logits/rejected": 0.24673417210578918,
|
| 422 |
+
"logps/chosen": -532.9969482421875,
|
| 423 |
+
"logps/rejected": -591.529296875,
|
| 424 |
+
"loss": 0.2764,
|
| 425 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 426 |
+
"rewards/chosen": -0.19002854824066162,
|
| 427 |
+
"rewards/margins": 0.12101318687200546,
|
| 428 |
+
"rewards/rejected": -0.3110417425632477,
|
| 429 |
+
"step": 290
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.23,
|
| 433 |
+
"learning_rate": 4.754069787631761e-06,
|
| 434 |
+
"logits/chosen": 0.15965518355369568,
|
| 435 |
+
"logits/rejected": 0.3048693537712097,
|
| 436 |
+
"logps/chosen": -550.864990234375,
|
| 437 |
+
"logps/rejected": -679.41357421875,
|
| 438 |
+
"loss": 0.25,
|
| 439 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 440 |
+
"rewards/chosen": -0.21647436916828156,
|
| 441 |
+
"rewards/margins": 0.16393651068210602,
|
| 442 |
+
"rewards/rejected": -0.3804108500480652,
|
| 443 |
+
"step": 300
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.24,
|
| 447 |
+
"learning_rate": 4.724486696587862e-06,
|
| 448 |
+
"logits/chosen": 0.17140202224254608,
|
| 449 |
+
"logits/rejected": 0.21409063041210175,
|
| 450 |
+
"logps/chosen": -635.85791015625,
|
| 451 |
+
"logps/rejected": -747.188232421875,
|
| 452 |
+
"loss": 0.258,
|
| 453 |
+
"rewards/accuracies": 0.706250011920929,
|
| 454 |
+
"rewards/chosen": -0.29053109884262085,
|
| 455 |
+
"rewards/margins": 0.15401865541934967,
|
| 456 |
+
"rewards/rejected": -0.4445497393608093,
|
| 457 |
+
"step": 310
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.24,
|
| 461 |
+
"learning_rate": 4.693326938861367e-06,
|
| 462 |
+
"logits/chosen": 0.19796046614646912,
|
| 463 |
+
"logits/rejected": 0.25672799348831177,
|
| 464 |
+
"logps/chosen": -643.25537109375,
|
| 465 |
+
"logps/rejected": -801.0643310546875,
|
| 466 |
+
"loss": 0.2427,
|
| 467 |
+
"rewards/accuracies": 0.75,
|
| 468 |
+
"rewards/chosen": -0.31196296215057373,
|
| 469 |
+
"rewards/margins": 0.19956240057945251,
|
| 470 |
+
"rewards/rejected": -0.5115253925323486,
|
| 471 |
+
"step": 320
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.25,
|
| 475 |
+
"learning_rate": 4.660612599795343e-06,
|
| 476 |
+
"logits/chosen": 0.11408114433288574,
|
| 477 |
+
"logits/rejected": 0.22174029052257538,
|
| 478 |
+
"logps/chosen": -647.49658203125,
|
| 479 |
+
"logps/rejected": -756.5814208984375,
|
| 480 |
+
"loss": 0.2472,
|
| 481 |
+
"rewards/accuracies": 0.699999988079071,
|
| 482 |
+
"rewards/chosen": -0.29143720865249634,
|
| 483 |
+
"rewards/margins": 0.1790235936641693,
|
| 484 |
+
"rewards/rejected": -0.47046083211898804,
|
| 485 |
+
"step": 330
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.26,
|
| 489 |
+
"learning_rate": 4.626366866585528e-06,
|
| 490 |
+
"logits/chosen": 0.08651003241539001,
|
| 491 |
+
"logits/rejected": 0.29140934348106384,
|
| 492 |
+
"logps/chosen": -647.8692626953125,
|
| 493 |
+
"logps/rejected": -736.3836669921875,
|
| 494 |
+
"loss": 0.2449,
|
| 495 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 496 |
+
"rewards/chosen": -0.28671354055404663,
|
| 497 |
+
"rewards/margins": 0.16735221445560455,
|
| 498 |
+
"rewards/rejected": -0.45406574010849,
|
| 499 |
+
"step": 340
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.27,
|
| 503 |
+
"learning_rate": 4.590614011845758e-06,
|
| 504 |
+
"logits/chosen": 0.14404548704624176,
|
| 505 |
+
"logits/rejected": 0.26181453466415405,
|
| 506 |
+
"logps/chosen": -638.6263427734375,
|
| 507 |
+
"logps/rejected": -730.7149658203125,
|
| 508 |
+
"loss": 0.2148,
|
| 509 |
+
"rewards/accuracies": 0.71875,
|
| 510 |
+
"rewards/chosen": -0.2571524977684021,
|
| 511 |
+
"rewards/margins": 0.1780269742012024,
|
| 512 |
+
"rewards/rejected": -0.4351794719696045,
|
| 513 |
+
"step": 350
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.27,
|
| 517 |
+
"learning_rate": 4.553379376404085e-06,
|
| 518 |
+
"logits/chosen": 0.17719906568527222,
|
| 519 |
+
"logits/rejected": 0.17130860686302185,
|
| 520 |
+
"logps/chosen": -565.2915649414062,
|
| 521 |
+
"logps/rejected": -659.3453369140625,
|
| 522 |
+
"loss": 0.252,
|
| 523 |
+
"rewards/accuracies": 0.71875,
|
| 524 |
+
"rewards/chosen": -0.23078814148902893,
|
| 525 |
+
"rewards/margins": 0.16517826914787292,
|
| 526 |
+
"rewards/rejected": -0.39596638083457947,
|
| 527 |
+
"step": 360
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.28,
|
| 531 |
+
"learning_rate": 4.514689351341751e-06,
|
| 532 |
+
"logits/chosen": 0.18946215510368347,
|
| 533 |
+
"logits/rejected": 0.2520686089992523,
|
| 534 |
+
"logps/chosen": -707.220458984375,
|
| 535 |
+
"logps/rejected": -799.5474853515625,
|
| 536 |
+
"loss": 0.2296,
|
| 537 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 538 |
+
"rewards/chosen": -0.29390206933021545,
|
| 539 |
+
"rewards/margins": 0.1939956247806549,
|
| 540 |
+
"rewards/rejected": -0.48789769411087036,
|
| 541 |
+
"step": 370
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.29,
|
| 545 |
+
"learning_rate": 4.474571359287791e-06,
|
| 546 |
+
"logits/chosen": 0.1794353723526001,
|
| 547 |
+
"logits/rejected": 0.20889122784137726,
|
| 548 |
+
"logps/chosen": -659.88720703125,
|
| 549 |
+
"logps/rejected": -792.5078125,
|
| 550 |
+
"loss": 0.2767,
|
| 551 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 552 |
+
"rewards/chosen": -0.32972732186317444,
|
| 553 |
+
"rewards/margins": 0.16797736287117004,
|
| 554 |
+
"rewards/rejected": -0.4977046847343445,
|
| 555 |
+
"step": 380
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.3,
|
| 559 |
+
"learning_rate": 4.4330538349824684e-06,
|
| 560 |
+
"logits/chosen": 0.15636876225471497,
|
| 561 |
+
"logits/rejected": 0.2390608787536621,
|
| 562 |
+
"logps/chosen": -675.450927734375,
|
| 563 |
+
"logps/rejected": -797.5559692382812,
|
| 564 |
+
"loss": 0.2478,
|
| 565 |
+
"rewards/accuracies": 0.731249988079071,
|
| 566 |
+
"rewards/chosen": -0.3252645432949066,
|
| 567 |
+
"rewards/margins": 0.17893439531326294,
|
| 568 |
+
"rewards/rejected": -0.5041989684104919,
|
| 569 |
+
"step": 390
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.3,
|
| 573 |
+
"learning_rate": 4.3901662051233755e-06,
|
| 574 |
+
"logits/chosen": 0.1320812702178955,
|
| 575 |
+
"logits/rejected": 0.25596413016319275,
|
| 576 |
+
"logps/chosen": -722.6912841796875,
|
| 577 |
+
"logps/rejected": -806.4666748046875,
|
| 578 |
+
"loss": 0.2426,
|
| 579 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 580 |
+
"rewards/chosen": -0.33014723658561707,
|
| 581 |
+
"rewards/margins": 0.17146170139312744,
|
| 582 |
+
"rewards/rejected": -0.5016089677810669,
|
| 583 |
+
"step": 400
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.31,
|
| 587 |
+
"learning_rate": 4.345938867508439e-06,
|
| 588 |
+
"logits/chosen": 0.15700757503509521,
|
| 589 |
+
"logits/rejected": 0.2542612552642822,
|
| 590 |
+
"logps/chosen": -715.8280029296875,
|
| 591 |
+
"logps/rejected": -812.6696166992188,
|
| 592 |
+
"loss": 0.2483,
|
| 593 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 594 |
+
"rewards/chosen": -0.34785327315330505,
|
| 595 |
+
"rewards/margins": 0.17259207367897034,
|
| 596 |
+
"rewards/rejected": -0.5204453468322754,
|
| 597 |
+
"step": 410
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.32,
|
| 601 |
+
"learning_rate": 4.30040316949064e-06,
|
| 602 |
+
"logits/chosen": 0.12117477506399155,
|
| 603 |
+
"logits/rejected": 0.21865728497505188,
|
| 604 |
+
"logps/chosen": -640.8743896484375,
|
| 605 |
+
"logps/rejected": -720.25439453125,
|
| 606 |
+
"loss": 0.2724,
|
| 607 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 608 |
+
"rewards/chosen": -0.30353325605392456,
|
| 609 |
+
"rewards/margins": 0.136855810880661,
|
| 610 |
+
"rewards/rejected": -0.44038906693458557,
|
| 611 |
+
"step": 420
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.33,
|
| 615 |
+
"learning_rate": 4.253591385759705e-06,
|
| 616 |
+
"logits/chosen": 0.1048569455742836,
|
| 617 |
+
"logits/rejected": 0.2262450009584427,
|
| 618 |
+
"logps/chosen": -654.50634765625,
|
| 619 |
+
"logps/rejected": -762.971435546875,
|
| 620 |
+
"loss": 0.2437,
|
| 621 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 622 |
+
"rewards/chosen": -0.30736270546913147,
|
| 623 |
+
"rewards/margins": 0.18053661286830902,
|
| 624 |
+
"rewards/rejected": -0.48789939284324646,
|
| 625 |
+
"step": 430
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.34,
|
| 629 |
+
"learning_rate": 4.205536695466524e-06,
|
| 630 |
+
"logits/chosen": 0.1525915265083313,
|
| 631 |
+
"logits/rejected": 0.27372902631759644,
|
| 632 |
+
"logps/chosen": -654.9791259765625,
|
| 633 |
+
"logps/rejected": -811.6696166992188,
|
| 634 |
+
"loss": 0.2447,
|
| 635 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 636 |
+
"rewards/chosen": -0.319768488407135,
|
| 637 |
+
"rewards/margins": 0.19004443287849426,
|
| 638 |
+
"rewards/rejected": -0.5098129510879517,
|
| 639 |
+
"step": 440
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.34,
|
| 643 |
+
"learning_rate": 4.15627315870651e-06,
|
| 644 |
+
"logits/chosen": 0.136866956949234,
|
| 645 |
+
"logits/rejected": 0.21996262669563293,
|
| 646 |
+
"logps/chosen": -676.5054931640625,
|
| 647 |
+
"logps/rejected": -751.3809204101562,
|
| 648 |
+
"loss": 0.2517,
|
| 649 |
+
"rewards/accuracies": 0.699999988079071,
|
| 650 |
+
"rewards/chosen": -0.3240048289299011,
|
| 651 |
+
"rewards/margins": 0.1569635421037674,
|
| 652 |
+
"rewards/rejected": -0.48096832633018494,
|
| 653 |
+
"step": 450
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.35,
|
| 657 |
+
"learning_rate": 4.105835692378557e-06,
|
| 658 |
+
"logits/chosen": 0.17371432483196259,
|
| 659 |
+
"logits/rejected": 0.2416466474533081,
|
| 660 |
+
"logps/chosen": -675.8497314453125,
|
| 661 |
+
"logps/rejected": -758.9337158203125,
|
| 662 |
+
"loss": 0.2008,
|
| 663 |
+
"rewards/accuracies": 0.731249988079071,
|
| 664 |
+
"rewards/chosen": -0.28624922037124634,
|
| 665 |
+
"rewards/margins": 0.17619088292121887,
|
| 666 |
+
"rewards/rejected": -0.4624401032924652,
|
| 667 |
+
"step": 460
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.36,
|
| 671 |
+
"learning_rate": 4.05426004543672e-06,
|
| 672 |
+
"logits/chosen": 0.12746404111385345,
|
| 673 |
+
"logits/rejected": 0.2692243754863739,
|
| 674 |
+
"logps/chosen": -697.2105712890625,
|
| 675 |
+
"logps/rejected": -788.6505126953125,
|
| 676 |
+
"loss": 0.2271,
|
| 677 |
+
"rewards/accuracies": 0.768750011920929,
|
| 678 |
+
"rewards/chosen": -0.3282613158226013,
|
| 679 |
+
"rewards/margins": 0.18242862820625305,
|
| 680 |
+
"rewards/rejected": -0.510689914226532,
|
| 681 |
+
"step": 470
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.37,
|
| 685 |
+
"learning_rate": 4.001582773552153e-06,
|
| 686 |
+
"logits/chosen": 0.19310589134693146,
|
| 687 |
+
"logits/rejected": 0.24649909138679504,
|
| 688 |
+
"logps/chosen": -676.9974365234375,
|
| 689 |
+
"logps/rejected": -746.8987426757812,
|
| 690 |
+
"loss": 0.2236,
|
| 691 |
+
"rewards/accuracies": 0.731249988079071,
|
| 692 |
+
"rewards/chosen": -0.31384724378585815,
|
| 693 |
+
"rewards/margins": 0.16469937562942505,
|
| 694 |
+
"rewards/rejected": -0.478546679019928,
|
| 695 |
+
"step": 480
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.37,
|
| 699 |
+
"learning_rate": 3.947841213203262e-06,
|
| 700 |
+
"logits/chosen": 0.15690350532531738,
|
| 701 |
+
"logits/rejected": 0.300692081451416,
|
| 702 |
+
"logps/chosen": -735.6814575195312,
|
| 703 |
+
"logps/rejected": -848.1085815429688,
|
| 704 |
+
"loss": 0.2125,
|
| 705 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 706 |
+
"rewards/chosen": -0.34843936562538147,
|
| 707 |
+
"rewards/margins": 0.20286861062049866,
|
| 708 |
+
"rewards/rejected": -0.5513080358505249,
|
| 709 |
+
"step": 490
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.38,
|
| 713 |
+
"learning_rate": 3.893073455212438e-06,
|
| 714 |
+
"logits/chosen": 0.18581806123256683,
|
| 715 |
+
"logits/rejected": 0.28662142157554626,
|
| 716 |
+
"logps/chosen": -699.9567260742188,
|
| 717 |
+
"logps/rejected": -875.0812377929688,
|
| 718 |
+
"loss": 0.2029,
|
| 719 |
+
"rewards/accuracies": 0.706250011920929,
|
| 720 |
+
"rewards/chosen": -0.35302621126174927,
|
| 721 |
+
"rewards/margins": 0.2367512434720993,
|
| 722 |
+
"rewards/rejected": -0.589777410030365,
|
| 723 |
+
"step": 500
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.39,
|
| 727 |
+
"learning_rate": 3.837318317748134e-06,
|
| 728 |
+
"logits/chosen": 0.2250034064054489,
|
| 729 |
+
"logits/rejected": 0.25657814741134644,
|
| 730 |
+
"logps/chosen": -629.2000732421875,
|
| 731 |
+
"logps/rejected": -729.8438720703125,
|
| 732 |
+
"loss": 0.2881,
|
| 733 |
+
"rewards/accuracies": 0.699999988079071,
|
| 734 |
+
"rewards/chosen": -0.2759498357772827,
|
| 735 |
+
"rewards/margins": 0.16293799877166748,
|
| 736 |
+
"rewards/rejected": -0.43888789415359497,
|
| 737 |
+
"step": 510
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.4,
|
| 741 |
+
"learning_rate": 3.7806153188114027e-06,
|
| 742 |
+
"logits/chosen": 0.15463793277740479,
|
| 743 |
+
"logits/rejected": 0.26033270359039307,
|
| 744 |
+
"logps/chosen": -536.32861328125,
|
| 745 |
+
"logps/rejected": -633.5609130859375,
|
| 746 |
+
"loss": 0.2673,
|
| 747 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 748 |
+
"rewards/chosen": -0.219068244099617,
|
| 749 |
+
"rewards/margins": 0.13708043098449707,
|
| 750 |
+
"rewards/rejected": -0.3561486601829529,
|
| 751 |
+
"step": 520
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.4,
|
| 755 |
+
"learning_rate": 3.7230046482264256e-06,
|
| 756 |
+
"logits/chosen": 0.1772613823413849,
|
| 757 |
+
"logits/rejected": 0.3070564270019531,
|
| 758 |
+
"logps/chosen": -638.0252685546875,
|
| 759 |
+
"logps/rejected": -699.7626953125,
|
| 760 |
+
"loss": 0.2509,
|
| 761 |
+
"rewards/accuracies": 0.737500011920929,
|
| 762 |
+
"rewards/chosen": -0.24965958297252655,
|
| 763 |
+
"rewards/margins": 0.15243235230445862,
|
| 764 |
+
"rewards/rejected": -0.402091920375824,
|
| 765 |
+
"step": 530
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.41,
|
| 769 |
+
"learning_rate": 3.6645271391548542e-06,
|
| 770 |
+
"logits/chosen": 0.1249006986618042,
|
| 771 |
+
"logits/rejected": 0.20080497860908508,
|
| 772 |
+
"logps/chosen": -620.0784912109375,
|
| 773 |
+
"logps/rejected": -716.3678588867188,
|
| 774 |
+
"loss": 0.2308,
|
| 775 |
+
"rewards/accuracies": 0.668749988079071,
|
| 776 |
+
"rewards/chosen": -0.2955653667449951,
|
| 777 |
+
"rewards/margins": 0.15287812054157257,
|
| 778 |
+
"rewards/rejected": -0.4484435021877289,
|
| 779 |
+
"step": 540
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.42,
|
| 783 |
+
"learning_rate": 3.6052242391541746e-06,
|
| 784 |
+
"logits/chosen": 0.1353389322757721,
|
| 785 |
+
"logits/rejected": 0.22389094531536102,
|
| 786 |
+
"logps/chosen": -614.9147338867188,
|
| 787 |
+
"logps/rejected": -771.8558349609375,
|
| 788 |
+
"loss": 0.2102,
|
| 789 |
+
"rewards/accuracies": 0.737500011920929,
|
| 790 |
+
"rewards/chosen": -0.2764993906021118,
|
| 791 |
+
"rewards/margins": 0.18298561871051788,
|
| 792 |
+
"rewards/rejected": -0.4594849944114685,
|
| 793 |
+
"step": 550
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.43,
|
| 797 |
+
"learning_rate": 3.5451379808006014e-06,
|
| 798 |
+
"logits/chosen": 0.17011868953704834,
|
| 799 |
+
"logits/rejected": 0.27556750178337097,
|
| 800 |
+
"logps/chosen": -658.0120849609375,
|
| 801 |
+
"logps/rejected": -755.5858764648438,
|
| 802 |
+
"loss": 0.2345,
|
| 803 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 804 |
+
"rewards/chosen": -0.29901498556137085,
|
| 805 |
+
"rewards/margins": 0.18377789855003357,
|
| 806 |
+
"rewards/rejected": -0.48279285430908203,
|
| 807 |
+
"step": 560
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.43,
|
| 811 |
+
"learning_rate": 3.484310951897323e-06,
|
| 812 |
+
"logits/chosen": 0.1761491745710373,
|
| 813 |
+
"logits/rejected": 0.30945947766304016,
|
| 814 |
+
"logps/chosen": -717.3760986328125,
|
| 815 |
+
"logps/rejected": -787.5486450195312,
|
| 816 |
+
"loss": 0.2483,
|
| 817 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 818 |
+
"rewards/chosen": -0.32107868790626526,
|
| 819 |
+
"rewards/margins": 0.1693280041217804,
|
| 820 |
+
"rewards/rejected": -0.49040669202804565,
|
| 821 |
+
"step": 570
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.44,
|
| 825 |
+
"learning_rate": 3.4227862652892106e-06,
|
| 826 |
+
"logits/chosen": 0.20576810836791992,
|
| 827 |
+
"logits/rejected": 0.26615187525749207,
|
| 828 |
+
"logps/chosen": -668.9432983398438,
|
| 829 |
+
"logps/rejected": -773.5794067382812,
|
| 830 |
+
"loss": 0.2502,
|
| 831 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 832 |
+
"rewards/chosen": -0.31081390380859375,
|
| 833 |
+
"rewards/margins": 0.15438678860664368,
|
| 834 |
+
"rewards/rejected": -0.46520066261291504,
|
| 835 |
+
"step": 580
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.45,
|
| 839 |
+
"learning_rate": 3.3606075283054005e-06,
|
| 840 |
+
"logits/chosen": 0.18598072230815887,
|
| 841 |
+
"logits/rejected": 0.2837832570075989,
|
| 842 |
+
"logps/chosen": -603.8995361328125,
|
| 843 |
+
"logps/rejected": -735.7412719726562,
|
| 844 |
+
"loss": 0.208,
|
| 845 |
+
"rewards/accuracies": 0.75,
|
| 846 |
+
"rewards/chosen": -0.2771075367927551,
|
| 847 |
+
"rewards/margins": 0.1824244260787964,
|
| 848 |
+
"rewards/rejected": -0.4595320224761963,
|
| 849 |
+
"step": 590
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.46,
|
| 853 |
+
"learning_rate": 3.2978188118513814e-06,
|
| 854 |
+
"logits/chosen": 0.1718396097421646,
|
| 855 |
+
"logits/rejected": 0.3636724054813385,
|
| 856 |
+
"logps/chosen": -653.9473266601562,
|
| 857 |
+
"logps/rejected": -759.1709594726562,
|
| 858 |
+
"loss": 0.2263,
|
| 859 |
+
"rewards/accuracies": 0.737500011920929,
|
| 860 |
+
"rewards/chosen": -0.28921762108802795,
|
| 861 |
+
"rewards/margins": 0.19078542292118073,
|
| 862 |
+
"rewards/rejected": -0.4800030589103699,
|
| 863 |
+
"step": 600
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.46,
|
| 867 |
+
"learning_rate": 3.234464619172522e-06,
|
| 868 |
+
"logits/chosen": 0.1497826725244522,
|
| 869 |
+
"logits/rejected": 0.2777649164199829,
|
| 870 |
+
"logps/chosen": -638.6043090820312,
|
| 871 |
+
"logps/rejected": -779.666748046875,
|
| 872 |
+
"loss": 0.223,
|
| 873 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 874 |
+
"rewards/chosen": -0.28172770142555237,
|
| 875 |
+
"rewards/margins": 0.22779683768749237,
|
| 876 |
+
"rewards/rejected": -0.5095245242118835,
|
| 877 |
+
"step": 610
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.47,
|
| 881 |
+
"learning_rate": 3.1705898543111576e-06,
|
| 882 |
+
"logits/chosen": 0.13684847950935364,
|
| 883 |
+
"logits/rejected": 0.24817728996276855,
|
| 884 |
+
"logps/chosen": -634.4373779296875,
|
| 885 |
+
"logps/rejected": -796.2052612304688,
|
| 886 |
+
"loss": 0.2257,
|
| 887 |
+
"rewards/accuracies": 0.731249988079071,
|
| 888 |
+
"rewards/chosen": -0.2933855652809143,
|
| 889 |
+
"rewards/margins": 0.20933008193969727,
|
| 890 |
+
"rewards/rejected": -0.5027156472206116,
|
| 891 |
+
"step": 620
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.48,
|
| 895 |
+
"learning_rate": 3.106239790279606e-06,
|
| 896 |
+
"logits/chosen": 0.16658630967140198,
|
| 897 |
+
"logits/rejected": 0.2954631447792053,
|
| 898 |
+
"logps/chosen": -654.8861083984375,
|
| 899 |
+
"logps/rejected": -780.149169921875,
|
| 900 |
+
"loss": 0.2097,
|
| 901 |
+
"rewards/accuracies": 0.737500011920929,
|
| 902 |
+
"rewards/chosen": -0.3087119460105896,
|
| 903 |
+
"rewards/margins": 0.1822930872440338,
|
| 904 |
+
"rewards/rejected": -0.4910050928592682,
|
| 905 |
+
"step": 630
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.49,
|
| 909 |
+
"learning_rate": 3.041460036971664e-06,
|
| 910 |
+
"logits/chosen": 0.13499195873737335,
|
| 911 |
+
"logits/rejected": 0.27665549516677856,
|
| 912 |
+
"logps/chosen": -655.3276977539062,
|
| 913 |
+
"logps/rejected": -844.3450927734375,
|
| 914 |
+
"loss": 0.2251,
|
| 915 |
+
"rewards/accuracies": 0.78125,
|
| 916 |
+
"rewards/chosen": -0.3263702094554901,
|
| 917 |
+
"rewards/margins": 0.22413411736488342,
|
| 918 |
+
"rewards/rejected": -0.5505043268203735,
|
| 919 |
+
"step": 640
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.5,
|
| 923 |
+
"learning_rate": 2.976296508835326e-06,
|
| 924 |
+
"logits/chosen": 0.16133855283260345,
|
| 925 |
+
"logits/rejected": 0.256599485874176,
|
| 926 |
+
"logps/chosen": -660.0172119140625,
|
| 927 |
+
"logps/rejected": -769.0188598632812,
|
| 928 |
+
"loss": 0.2228,
|
| 929 |
+
"rewards/accuracies": 0.71875,
|
| 930 |
+
"rewards/chosen": -0.3122148811817169,
|
| 931 |
+
"rewards/margins": 0.18627096712589264,
|
| 932 |
+
"rewards/rejected": -0.49848586320877075,
|
| 933 |
+
"step": 650
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.5,
|
| 937 |
+
"learning_rate": 2.910795392329649e-06,
|
| 938 |
+
"logits/chosen": 0.1639477163553238,
|
| 939 |
+
"logits/rejected": 0.24110262095928192,
|
| 940 |
+
"logps/chosen": -739.8809204101562,
|
| 941 |
+
"logps/rejected": -895.7755126953125,
|
| 942 |
+
"loss": 0.2189,
|
| 943 |
+
"rewards/accuracies": 0.78125,
|
| 944 |
+
"rewards/chosen": -0.33884376287460327,
|
| 945 |
+
"rewards/margins": 0.23572292923927307,
|
| 946 |
+
"rewards/rejected": -0.5745667219161987,
|
| 947 |
+
"step": 660
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.51,
|
| 951 |
+
"learning_rate": 2.8450031131888147e-06,
|
| 952 |
+
"logits/chosen": 0.13471753895282745,
|
| 953 |
+
"logits/rejected": 0.27453264594078064,
|
| 954 |
+
"logps/chosen": -660.6944580078125,
|
| 955 |
+
"logps/rejected": -760.9178466796875,
|
| 956 |
+
"loss": 0.2316,
|
| 957 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 958 |
+
"rewards/chosen": -0.32461482286453247,
|
| 959 |
+
"rewards/margins": 0.18133333325386047,
|
| 960 |
+
"rewards/rejected": -0.5059481859207153,
|
| 961 |
+
"step": 670
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.52,
|
| 965 |
+
"learning_rate": 2.7789663035166035e-06,
|
| 966 |
+
"logits/chosen": 0.07190994918346405,
|
| 967 |
+
"logits/rejected": 0.22703304886817932,
|
| 968 |
+
"logps/chosen": -670.5394287109375,
|
| 969 |
+
"logps/rejected": -823.9978637695312,
|
| 970 |
+
"loss": 0.2036,
|
| 971 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 972 |
+
"rewards/chosen": -0.3125961720943451,
|
| 973 |
+
"rewards/margins": 0.2316323220729828,
|
| 974 |
+
"rewards/rejected": -0.5442285537719727,
|
| 975 |
+
"step": 680
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.53,
|
| 979 |
+
"learning_rate": 2.7127317687345973e-06,
|
| 980 |
+
"logits/chosen": 0.07092462480068207,
|
| 981 |
+
"logits/rejected": 0.26120471954345703,
|
| 982 |
+
"logps/chosen": -671.69189453125,
|
| 983 |
+
"logps/rejected": -811.6243286132812,
|
| 984 |
+
"loss": 0.2018,
|
| 985 |
+
"rewards/accuracies": 0.800000011920929,
|
| 986 |
+
"rewards/chosen": -0.30282872915267944,
|
| 987 |
+
"rewards/margins": 0.21179743111133575,
|
| 988 |
+
"rewards/rejected": -0.514626145362854,
|
| 989 |
+
"step": 690
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.53,
|
| 993 |
+
"learning_rate": 2.6463464544075344e-06,
|
| 994 |
+
"logits/chosen": 0.14573340117931366,
|
| 995 |
+
"logits/rejected": 0.23916473984718323,
|
| 996 |
+
"logps/chosen": -684.8336181640625,
|
| 997 |
+
"logps/rejected": -777.0270385742188,
|
| 998 |
+
"loss": 0.1987,
|
| 999 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 1000 |
+
"rewards/chosen": -0.3061198592185974,
|
| 1001 |
+
"rewards/margins": 0.19952335953712463,
|
| 1002 |
+
"rewards/rejected": -0.5056431889533997,
|
| 1003 |
+
"step": 700
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.54,
|
| 1007 |
+
"learning_rate": 2.579857412969345e-06,
|
| 1008 |
+
"logits/chosen": 0.11962984502315521,
|
| 1009 |
+
"logits/rejected": 0.23462708294391632,
|
| 1010 |
+
"logps/chosen": -730.8359375,
|
| 1011 |
+
"logps/rejected": -856.9847412109375,
|
| 1012 |
+
"loss": 0.237,
|
| 1013 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1014 |
+
"rewards/chosen": -0.37815654277801514,
|
| 1015 |
+
"rewards/margins": 0.18362250924110413,
|
| 1016 |
+
"rewards/rejected": -0.5617790818214417,
|
| 1017 |
+
"step": 710
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.55,
|
| 1021 |
+
"learning_rate": 2.513311770373421e-06,
|
| 1022 |
+
"logits/chosen": 0.11839403957128525,
|
| 1023 |
+
"logits/rejected": 0.16228660941123962,
|
| 1024 |
+
"logps/chosen": -644.6649780273438,
|
| 1025 |
+
"logps/rejected": -824.513671875,
|
| 1026 |
+
"loss": 0.2276,
|
| 1027 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1028 |
+
"rewards/chosen": -0.32936322689056396,
|
| 1029 |
+
"rewards/margins": 0.2137618511915207,
|
| 1030 |
+
"rewards/rejected": -0.5431250929832458,
|
| 1031 |
+
"step": 720
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.56,
|
| 1035 |
+
"learning_rate": 2.446756692690804e-06,
|
| 1036 |
+
"logits/chosen": 0.13511911034584045,
|
| 1037 |
+
"logits/rejected": 0.2399536371231079,
|
| 1038 |
+
"logps/chosen": -717.3846435546875,
|
| 1039 |
+
"logps/rejected": -858.6282958984375,
|
| 1040 |
+
"loss": 0.2365,
|
| 1041 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1042 |
+
"rewards/chosen": -0.3688200116157532,
|
| 1043 |
+
"rewards/margins": 0.2047378122806549,
|
| 1044 |
+
"rewards/rejected": -0.5735577940940857,
|
| 1045 |
+
"step": 730
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.56,
|
| 1049 |
+
"learning_rate": 2.380239352679908e-06,
|
| 1050 |
+
"logits/chosen": 0.14800789952278137,
|
| 1051 |
+
"logits/rejected": 0.2702687382698059,
|
| 1052 |
+
"logps/chosen": -674.1021728515625,
|
| 1053 |
+
"logps/rejected": -830.0198974609375,
|
| 1054 |
+
"loss": 0.2284,
|
| 1055 |
+
"rewards/accuracies": 0.71875,
|
| 1056 |
+
"rewards/chosen": -0.31319013237953186,
|
| 1057 |
+
"rewards/margins": 0.2306814193725586,
|
| 1058 |
+
"rewards/rejected": -0.5438715219497681,
|
| 1059 |
+
"step": 740
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.57,
|
| 1063 |
+
"learning_rate": 2.313806896351529e-06,
|
| 1064 |
+
"logits/chosen": 0.07657043635845184,
|
| 1065 |
+
"logits/rejected": 0.2402833253145218,
|
| 1066 |
+
"logps/chosen": -611.2683715820312,
|
| 1067 |
+
"logps/rejected": -762.6300048828125,
|
| 1068 |
+
"loss": 0.2543,
|
| 1069 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 1070 |
+
"rewards/chosen": -0.3043960928916931,
|
| 1071 |
+
"rewards/margins": 0.188668355345726,
|
| 1072 |
+
"rewards/rejected": -0.49306440353393555,
|
| 1073 |
+
"step": 750
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.58,
|
| 1077 |
+
"learning_rate": 2.247506409552795e-06,
|
| 1078 |
+
"logits/chosen": 0.11987291276454926,
|
| 1079 |
+
"logits/rejected": 0.1521257907152176,
|
| 1080 |
+
"logps/chosen": -654.2996826171875,
|
| 1081 |
+
"logps/rejected": -776.6150512695312,
|
| 1082 |
+
"loss": 0.248,
|
| 1083 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 1084 |
+
"rewards/chosen": -0.2802131772041321,
|
| 1085 |
+
"rewards/margins": 0.1828833967447281,
|
| 1086 |
+
"rewards/rejected": -0.463096559047699,
|
| 1087 |
+
"step": 760
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.59,
|
| 1091 |
+
"learning_rate": 2.1813848845937695e-06,
|
| 1092 |
+
"logits/chosen": 0.07216247916221619,
|
| 1093 |
+
"logits/rejected": 0.16098852455615997,
|
| 1094 |
+
"logps/chosen": -612.78759765625,
|
| 1095 |
+
"logps/rejected": -761.1871948242188,
|
| 1096 |
+
"loss": 0.2486,
|
| 1097 |
+
"rewards/accuracies": 0.675000011920929,
|
| 1098 |
+
"rewards/chosen": -0.31463655829429626,
|
| 1099 |
+
"rewards/margins": 0.16312028467655182,
|
| 1100 |
+
"rewards/rejected": -0.4777568280696869,
|
| 1101 |
+
"step": 770
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.59,
|
| 1105 |
+
"learning_rate": 2.1154891869403436e-06,
|
| 1106 |
+
"logits/chosen": 0.056468479335308075,
|
| 1107 |
+
"logits/rejected": 0.21281781792640686,
|
| 1108 |
+
"logps/chosen": -676.2210693359375,
|
| 1109 |
+
"logps/rejected": -826.8810424804688,
|
| 1110 |
+
"loss": 0.202,
|
| 1111 |
+
"rewards/accuracies": 0.793749988079071,
|
| 1112 |
+
"rewards/chosen": -0.31608521938323975,
|
| 1113 |
+
"rewards/margins": 0.22052684426307678,
|
| 1114 |
+
"rewards/rejected": -0.5366120934486389,
|
| 1115 |
+
"step": 780
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.6,
|
| 1119 |
+
"learning_rate": 2.0498660219970395e-06,
|
| 1120 |
+
"logits/chosen": 0.14304831624031067,
|
| 1121 |
+
"logits/rejected": 0.2865908741950989,
|
| 1122 |
+
"logps/chosen": -720.0322875976562,
|
| 1123 |
+
"logps/rejected": -834.4865112304688,
|
| 1124 |
+
"loss": 0.2143,
|
| 1125 |
+
"rewards/accuracies": 0.706250011920929,
|
| 1126 |
+
"rewards/chosen": -0.3310641646385193,
|
| 1127 |
+
"rewards/margins": 0.21627302467823029,
|
| 1128 |
+
"rewards/rejected": -0.5473372340202332,
|
| 1129 |
+
"step": 790
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.61,
|
| 1133 |
+
"learning_rate": 1.9845619020032552e-06,
|
| 1134 |
+
"logits/chosen": 0.0913710817694664,
|
| 1135 |
+
"logits/rejected": 0.2621229588985443,
|
| 1136 |
+
"logps/chosen": -681.4385986328125,
|
| 1137 |
+
"logps/rejected": -793.8607788085938,
|
| 1138 |
+
"loss": 0.2232,
|
| 1139 |
+
"rewards/accuracies": 0.706250011920929,
|
| 1140 |
+
"rewards/chosen": -0.33381250500679016,
|
| 1141 |
+
"rewards/margins": 0.18326430022716522,
|
| 1142 |
+
"rewards/rejected": -0.5170767903327942,
|
| 1143 |
+
"step": 800
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.62,
|
| 1147 |
+
"learning_rate": 1.9196231130664282e-06,
|
| 1148 |
+
"logits/chosen": 0.12347328662872314,
|
| 1149 |
+
"logits/rejected": 0.22142863273620605,
|
| 1150 |
+
"logps/chosen": -656.6223754882812,
|
| 1151 |
+
"logps/rejected": -849.30908203125,
|
| 1152 |
+
"loss": 0.2207,
|
| 1153 |
+
"rewards/accuracies": 0.737500011920929,
|
| 1154 |
+
"rewards/chosen": -0.32625263929367065,
|
| 1155 |
+
"rewards/margins": 0.2373245507478714,
|
| 1156 |
+
"rewards/rejected": -0.5635771751403809,
|
| 1157 |
+
"step": 810
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.62,
|
| 1161 |
+
"learning_rate": 1.8550956823554708e-06,
|
| 1162 |
+
"logits/chosen": 0.12958547472953796,
|
| 1163 |
+
"logits/rejected": 0.20247094333171844,
|
| 1164 |
+
"logps/chosen": -676.3146362304688,
|
| 1165 |
+
"logps/rejected": -852.1516723632812,
|
| 1166 |
+
"loss": 0.2231,
|
| 1167 |
+
"rewards/accuracies": 0.793749988079071,
|
| 1168 |
+
"rewards/chosen": -0.3258567452430725,
|
| 1169 |
+
"rewards/margins": 0.22926822304725647,
|
| 1170 |
+
"rewards/rejected": -0.5551249384880066,
|
| 1171 |
+
"step": 820
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.63,
|
| 1175 |
+
"learning_rate": 1.7910253454777346e-06,
|
| 1176 |
+
"logits/chosen": 0.1100487932562828,
|
| 1177 |
+
"logits/rejected": 0.178089901804924,
|
| 1178 |
+
"logps/chosen": -628.2322387695312,
|
| 1179 |
+
"logps/rejected": -760.2506103515625,
|
| 1180 |
+
"loss": 0.2364,
|
| 1181 |
+
"rewards/accuracies": 0.75,
|
| 1182 |
+
"rewards/chosen": -0.3000151813030243,
|
| 1183 |
+
"rewards/margins": 0.20657257735729218,
|
| 1184 |
+
"rewards/rejected": -0.5065878033638,
|
| 1185 |
+
"step": 830
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.64,
|
| 1189 |
+
"learning_rate": 1.7274575140626318e-06,
|
| 1190 |
+
"logits/chosen": 0.06159307807683945,
|
| 1191 |
+
"logits/rejected": 0.21589604020118713,
|
| 1192 |
+
"logps/chosen": -658.613525390625,
|
| 1193 |
+
"logps/rejected": -839.2703247070312,
|
| 1194 |
+
"loss": 0.2066,
|
| 1195 |
+
"rewards/accuracies": 0.762499988079071,
|
| 1196 |
+
"rewards/chosen": -0.3004804253578186,
|
| 1197 |
+
"rewards/margins": 0.23831859230995178,
|
| 1198 |
+
"rewards/rejected": -0.5387989282608032,
|
| 1199 |
+
"step": 840
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.65,
|
| 1203 |
+
"learning_rate": 1.6644372435748823e-06,
|
| 1204 |
+
"logits/chosen": 0.13025906682014465,
|
| 1205 |
+
"logits/rejected": 0.25792089104652405,
|
| 1206 |
+
"logps/chosen": -657.7048950195312,
|
| 1207 |
+
"logps/rejected": -741.805908203125,
|
| 1208 |
+
"loss": 0.2056,
|
| 1209 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 1210 |
+
"rewards/chosen": -0.3026004731655121,
|
| 1211 |
+
"rewards/margins": 0.18632087111473083,
|
| 1212 |
+
"rewards/rejected": -0.4889214038848877,
|
| 1213 |
+
"step": 850
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.66,
|
| 1217 |
+
"learning_rate": 1.6020092013802002e-06,
|
| 1218 |
+
"logits/chosen": 0.13131192326545715,
|
| 1219 |
+
"logits/rejected": 0.2267313450574875,
|
| 1220 |
+
"logps/chosen": -609.1980590820312,
|
| 1221 |
+
"logps/rejected": -762.2578735351562,
|
| 1222 |
+
"loss": 0.2173,
|
| 1223 |
+
"rewards/accuracies": 0.71875,
|
| 1224 |
+
"rewards/chosen": -0.2911318838596344,
|
| 1225 |
+
"rewards/margins": 0.19050107896327972,
|
| 1226 |
+
"rewards/rejected": -0.48163294792175293,
|
| 1227 |
+
"step": 860
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.66,
|
| 1231 |
+
"learning_rate": 1.5402176350860653e-06,
|
| 1232 |
+
"logits/chosen": 0.1238790899515152,
|
| 1233 |
+
"logits/rejected": 0.215298131108284,
|
| 1234 |
+
"logps/chosen": -654.3424072265625,
|
| 1235 |
+
"logps/rejected": -766.5152587890625,
|
| 1236 |
+
"loss": 0.2276,
|
| 1237 |
+
"rewards/accuracies": 0.6875,
|
| 1238 |
+
"rewards/chosen": -0.311137855052948,
|
| 1239 |
+
"rewards/margins": 0.18526899814605713,
|
| 1240 |
+
"rewards/rejected": -0.4964068531990051,
|
| 1241 |
+
"step": 870
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.67,
|
| 1245 |
+
"learning_rate": 1.4791063411799938e-06,
|
| 1246 |
+
"logits/chosen": 0.07573570311069489,
|
| 1247 |
+
"logits/rejected": 0.16694195568561554,
|
| 1248 |
+
"logps/chosen": -611.1826782226562,
|
| 1249 |
+
"logps/rejected": -747.8428955078125,
|
| 1250 |
+
"loss": 0.225,
|
| 1251 |
+
"rewards/accuracies": 0.737500011920929,
|
| 1252 |
+
"rewards/chosen": -0.28111016750335693,
|
| 1253 |
+
"rewards/margins": 0.19318345189094543,
|
| 1254 |
+
"rewards/rejected": -0.47429361939430237,
|
| 1255 |
+
"step": 880
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.68,
|
| 1259 |
+
"learning_rate": 1.4187186339875697e-06,
|
| 1260 |
+
"logits/chosen": 0.14150698482990265,
|
| 1261 |
+
"logits/rejected": 0.19218984246253967,
|
| 1262 |
+
"logps/chosen": -653.740478515625,
|
| 1263 |
+
"logps/rejected": -790.3258666992188,
|
| 1264 |
+
"loss": 0.1959,
|
| 1265 |
+
"rewards/accuracies": 0.768750011920929,
|
| 1266 |
+
"rewards/chosen": -0.2978840172290802,
|
| 1267 |
+
"rewards/margins": 0.19249173998832703,
|
| 1268 |
+
"rewards/rejected": -0.4903757572174072,
|
| 1269 |
+
"step": 890
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.69,
|
| 1273 |
+
"learning_rate": 1.3590973149722103e-06,
|
| 1274 |
+
"logits/chosen": 0.07968850433826447,
|
| 1275 |
+
"logits/rejected": 0.24394066631793976,
|
| 1276 |
+
"logps/chosen": -660.4170532226562,
|
| 1277 |
+
"logps/rejected": -787.5786743164062,
|
| 1278 |
+
"loss": 0.2189,
|
| 1279 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 1280 |
+
"rewards/chosen": -0.3150777518749237,
|
| 1281 |
+
"rewards/margins": 0.19490757584571838,
|
| 1282 |
+
"rewards/rejected": -0.5099853277206421,
|
| 1283 |
+
"step": 900
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.69,
|
| 1287 |
+
"learning_rate": 1.300284642398445e-06,
|
| 1288 |
+
"logits/chosen": 0.1339327096939087,
|
| 1289 |
+
"logits/rejected": 0.19529958069324493,
|
| 1290 |
+
"logps/chosen": -637.8862915039062,
|
| 1291 |
+
"logps/rejected": -804.20263671875,
|
| 1292 |
+
"loss": 0.2179,
|
| 1293 |
+
"rewards/accuracies": 0.75,
|
| 1294 |
+
"rewards/chosen": -0.30193617939949036,
|
| 1295 |
+
"rewards/margins": 0.2071986198425293,
|
| 1296 |
+
"rewards/rejected": -0.5091347694396973,
|
| 1297 |
+
"step": 910
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.7,
|
| 1301 |
+
"learning_rate": 1.2423223013801946e-06,
|
| 1302 |
+
"logits/chosen": 0.09140697866678238,
|
| 1303 |
+
"logits/rejected": 0.23906917870044708,
|
| 1304 |
+
"logps/chosen": -615.2650146484375,
|
| 1305 |
+
"logps/rejected": -794.8267211914062,
|
| 1306 |
+
"loss": 0.2345,
|
| 1307 |
+
"rewards/accuracies": 0.737500011920929,
|
| 1308 |
+
"rewards/chosen": -0.290291965007782,
|
| 1309 |
+
"rewards/margins": 0.21678297221660614,
|
| 1310 |
+
"rewards/rejected": -0.5070749521255493,
|
| 1311 |
+
"step": 920
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.71,
|
| 1315 |
+
"learning_rate": 1.1852513743352886e-06,
|
| 1316 |
+
"logits/chosen": 0.09063401818275452,
|
| 1317 |
+
"logits/rejected": 0.18860659003257751,
|
| 1318 |
+
"logps/chosen": -603.1500854492188,
|
| 1319 |
+
"logps/rejected": -759.2166748046875,
|
| 1320 |
+
"loss": 0.1973,
|
| 1321 |
+
"rewards/accuracies": 0.762499988079071,
|
| 1322 |
+
"rewards/chosen": -0.26861757040023804,
|
| 1323 |
+
"rewards/margins": 0.2037418633699417,
|
| 1324 |
+
"rewards/rejected": -0.47235947847366333,
|
| 1325 |
+
"step": 930
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 0.72,
|
| 1329 |
+
"learning_rate": 1.1291123118671665e-06,
|
| 1330 |
+
"logits/chosen": 0.082930788397789,
|
| 1331 |
+
"logits/rejected": 0.177327498793602,
|
| 1332 |
+
"logps/chosen": -685.0097045898438,
|
| 1333 |
+
"logps/rejected": -753.899658203125,
|
| 1334 |
+
"loss": 0.2698,
|
| 1335 |
+
"rewards/accuracies": 0.668749988079071,
|
| 1336 |
+
"rewards/chosen": -0.31101205945014954,
|
| 1337 |
+
"rewards/margins": 0.16297343373298645,
|
| 1338 |
+
"rewards/rejected": -0.473985493183136,
|
| 1339 |
+
"step": 940
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.72,
|
| 1343 |
+
"learning_rate": 1.073944904094385e-06,
|
| 1344 |
+
"logits/chosen": 0.12176795303821564,
|
| 1345 |
+
"logits/rejected": 0.24055452644824982,
|
| 1346 |
+
"logps/chosen": -727.9547729492188,
|
| 1347 |
+
"logps/rejected": -807.3270874023438,
|
| 1348 |
+
"loss": 0.2284,
|
| 1349 |
+
"rewards/accuracies": 0.731249988079071,
|
| 1350 |
+
"rewards/chosen": -0.3418019115924835,
|
| 1351 |
+
"rewards/margins": 0.1853446215391159,
|
| 1352 |
+
"rewards/rejected": -0.5271465182304382,
|
| 1353 |
+
"step": 950
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 0.73,
|
| 1357 |
+
"learning_rate": 1.019788252448267e-06,
|
| 1358 |
+
"logits/chosen": 0.1278046816587448,
|
| 1359 |
+
"logits/rejected": 0.18657834827899933,
|
| 1360 |
+
"logps/chosen": -629.8650512695312,
|
| 1361 |
+
"logps/rejected": -775.0892333984375,
|
| 1362 |
+
"loss": 0.2073,
|
| 1363 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1364 |
+
"rewards/chosen": -0.31090590357780457,
|
| 1365 |
+
"rewards/margins": 0.19522252678871155,
|
| 1366 |
+
"rewards/rejected": -0.5061284899711609,
|
| 1367 |
+
"step": 960
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.74,
|
| 1371 |
+
"learning_rate": 9.66680741958685e-07,
|
| 1372 |
+
"logits/chosen": 0.06174594908952713,
|
| 1373 |
+
"logits/rejected": 0.21714496612548828,
|
| 1374 |
+
"logps/chosen": -706.654052734375,
|
| 1375 |
+
"logps/rejected": -845.1712646484375,
|
| 1376 |
+
"loss": 0.2177,
|
| 1377 |
+
"rewards/accuracies": 0.737500011920929,
|
| 1378 |
+
"rewards/chosen": -0.33396342396736145,
|
| 1379 |
+
"rewards/margins": 0.21265073120594025,
|
| 1380 |
+
"rewards/rejected": -0.5466141700744629,
|
| 1381 |
+
"step": 970
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 0.75,
|
| 1385 |
+
"learning_rate": 9.146600140475945e-07,
|
| 1386 |
+
"logits/chosen": 0.09592024236917496,
|
| 1387 |
+
"logits/rejected": 0.1995583027601242,
|
| 1388 |
+
"logps/chosen": -628.862548828125,
|
| 1389 |
+
"logps/rejected": -779.5205688476562,
|
| 1390 |
+
"loss": 0.2233,
|
| 1391 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 1392 |
+
"rewards/chosen": -0.3146194815635681,
|
| 1393 |
+
"rewards/margins": 0.17806106805801392,
|
| 1394 |
+
"rewards/rejected": -0.49268054962158203,
|
| 1395 |
+
"step": 980
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 0.75,
|
| 1399 |
+
"learning_rate": 8.637629398496378e-07,
|
| 1400 |
+
"logits/chosen": 0.06787069141864777,
|
| 1401 |
+
"logits/rejected": 0.19926394522190094,
|
| 1402 |
+
"logps/chosen": -675.4822387695312,
|
| 1403 |
+
"logps/rejected": -792.4046630859375,
|
| 1404 |
+
"loss": 0.2406,
|
| 1405 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 1406 |
+
"rewards/chosen": -0.31674814224243164,
|
| 1407 |
+
"rewards/margins": 0.19142326712608337,
|
| 1408 |
+
"rewards/rejected": -0.5081714391708374,
|
| 1409 |
+
"step": 990
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 0.76,
|
| 1413 |
+
"learning_rate": 8.140255940787059e-07,
|
| 1414 |
+
"logits/chosen": 0.14530155062675476,
|
| 1415 |
+
"logits/rejected": 0.2198048084974289,
|
| 1416 |
+
"logps/chosen": -646.3609008789062,
|
| 1417 |
+
"logps/rejected": -774.7366333007812,
|
| 1418 |
+
"loss": 0.2234,
|
| 1419 |
+
"rewards/accuracies": 0.737500011920929,
|
| 1420 |
+
"rewards/chosen": -0.30126315355300903,
|
| 1421 |
+
"rewards/margins": 0.1865108758211136,
|
| 1422 |
+
"rewards/rejected": -0.4877740442752838,
|
| 1423 |
+
"step": 1000
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 0.77,
|
| 1427 |
+
"learning_rate": 7.654832294589776e-07,
|
| 1428 |
+
"logits/chosen": 0.14776812493801117,
|
| 1429 |
+
"logits/rejected": 0.18824756145477295,
|
| 1430 |
+
"logps/chosen": -676.3685302734375,
|
| 1431 |
+
"logps/rejected": -820.2717895507812,
|
| 1432 |
+
"loss": 0.242,
|
| 1433 |
+
"rewards/accuracies": 0.78125,
|
| 1434 |
+
"rewards/chosen": -0.3261755406856537,
|
| 1435 |
+
"rewards/margins": 0.19091561436653137,
|
| 1436 |
+
"rewards/rejected": -0.5170911550521851,
|
| 1437 |
+
"step": 1010
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 0.78,
|
| 1441 |
+
"learning_rate": 7.181702517385789e-07,
|
| 1442 |
+
"logits/chosen": 0.15526942908763885,
|
| 1443 |
+
"logits/rejected": 0.20596864819526672,
|
| 1444 |
+
"logps/chosen": -720.7452392578125,
|
| 1445 |
+
"logps/rejected": -820.1512451171875,
|
| 1446 |
+
"loss": 0.2423,
|
| 1447 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 1448 |
+
"rewards/chosen": -0.33318132162094116,
|
| 1449 |
+
"rewards/margins": 0.173573836684227,
|
| 1450 |
+
"rewards/rejected": -0.5067551732063293,
|
| 1451 |
+
"step": 1020
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 0.78,
|
| 1455 |
+
"learning_rate": 6.721201953035511e-07,
|
| 1456 |
+
"logits/chosen": 0.09918368607759476,
|
| 1457 |
+
"logits/rejected": 0.23762516677379608,
|
| 1458 |
+
"logps/chosen": -690.7906494140625,
|
| 1459 |
+
"logps/rejected": -799.6129150390625,
|
| 1460 |
+
"loss": 0.222,
|
| 1461 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 1462 |
+
"rewards/chosen": -0.31717661023139954,
|
| 1463 |
+
"rewards/margins": 0.19311536848545074,
|
| 1464 |
+
"rewards/rejected": -0.5102919340133667,
|
| 1465 |
+
"step": 1030
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 0.79,
|
| 1469 |
+
"learning_rate": 6.273656994094232e-07,
|
| 1470 |
+
"logits/chosen": 0.08452818542718887,
|
| 1471 |
+
"logits/rejected": 0.15428626537322998,
|
| 1472 |
+
"logps/chosen": -674.3331298828125,
|
| 1473 |
+
"logps/rejected": -852.9098510742188,
|
| 1474 |
+
"loss": 0.212,
|
| 1475 |
+
"rewards/accuracies": 0.793749988079071,
|
| 1476 |
+
"rewards/chosen": -0.3417903482913971,
|
| 1477 |
+
"rewards/margins": 0.20718708634376526,
|
| 1478 |
+
"rewards/rejected": -0.5489774942398071,
|
| 1479 |
+
"step": 1040
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 0.8,
|
| 1483 |
+
"learning_rate": 5.839384850472359e-07,
|
| 1484 |
+
"logits/chosen": 0.11137109994888306,
|
| 1485 |
+
"logits/rejected": 0.23986658453941345,
|
| 1486 |
+
"logps/chosen": -681.7805786132812,
|
| 1487 |
+
"logps/rejected": -840.8401489257812,
|
| 1488 |
+
"loss": 0.2013,
|
| 1489 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 1490 |
+
"rewards/chosen": -0.33553558588027954,
|
| 1491 |
+
"rewards/margins": 0.22832973301410675,
|
| 1492 |
+
"rewards/rejected": -0.5638653039932251,
|
| 1493 |
+
"step": 1050
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 0.81,
|
| 1497 |
+
"learning_rate": 5.418693324604082e-07,
|
| 1498 |
+
"logits/chosen": 0.054320525377988815,
|
| 1499 |
+
"logits/rejected": 0.2136838734149933,
|
| 1500 |
+
"logps/chosen": -713.266357421875,
|
| 1501 |
+
"logps/rejected": -872.5158081054688,
|
| 1502 |
+
"loss": 0.1902,
|
| 1503 |
+
"rewards/accuracies": 0.800000011920929,
|
| 1504 |
+
"rewards/chosen": -0.3621819019317627,
|
| 1505 |
+
"rewards/margins": 0.21344491839408875,
|
| 1506 |
+
"rewards/rejected": -0.5756268501281738,
|
| 1507 |
+
"step": 1060
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 0.82,
|
| 1511 |
+
"learning_rate": 5.01188059328386e-07,
|
| 1512 |
+
"logits/chosen": 0.10655899345874786,
|
| 1513 |
+
"logits/rejected": 0.20958073437213898,
|
| 1514 |
+
"logps/chosen": -671.4863891601562,
|
| 1515 |
+
"logps/rejected": -796.0863647460938,
|
| 1516 |
+
"loss": 0.1935,
|
| 1517 |
+
"rewards/accuracies": 0.706250011920929,
|
| 1518 |
+
"rewards/chosen": -0.31292790174484253,
|
| 1519 |
+
"rewards/margins": 0.21162652969360352,
|
| 1520 |
+
"rewards/rejected": -0.5245543718338013,
|
| 1521 |
+
"step": 1070
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 0.82,
|
| 1525 |
+
"learning_rate": 4.619234996325314e-07,
|
| 1526 |
+
"logits/chosen": 0.1319715678691864,
|
| 1527 |
+
"logits/rejected": 0.21460673213005066,
|
| 1528 |
+
"logps/chosen": -709.2652587890625,
|
| 1529 |
+
"logps/rejected": -860.7151489257812,
|
| 1530 |
+
"loss": 0.2324,
|
| 1531 |
+
"rewards/accuracies": 0.75,
|
| 1532 |
+
"rewards/chosen": -0.3620796799659729,
|
| 1533 |
+
"rewards/margins": 0.1911153942346573,
|
| 1534 |
+
"rewards/rejected": -0.5531951189041138,
|
| 1535 |
+
"step": 1080
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 0.83,
|
| 1539 |
+
"learning_rate": 4.241034832192434e-07,
|
| 1540 |
+
"logits/chosen": 0.10131983458995819,
|
| 1541 |
+
"logits/rejected": 0.1885637789964676,
|
| 1542 |
+
"logps/chosen": -677.9981079101562,
|
| 1543 |
+
"logps/rejected": -864.3063354492188,
|
| 1544 |
+
"loss": 0.2159,
|
| 1545 |
+
"rewards/accuracies": 0.793749988079071,
|
| 1546 |
+
"rewards/chosen": -0.3420425057411194,
|
| 1547 |
+
"rewards/margins": 0.23228974640369415,
|
| 1548 |
+
"rewards/rejected": -0.5743322372436523,
|
| 1549 |
+
"step": 1090
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 0.84,
|
| 1553 |
+
"learning_rate": 3.877548160747768e-07,
|
| 1554 |
+
"logits/chosen": 0.11466535180807114,
|
| 1555 |
+
"logits/rejected": 0.24565303325653076,
|
| 1556 |
+
"logps/chosen": -680.638671875,
|
| 1557 |
+
"logps/rejected": -798.11376953125,
|
| 1558 |
+
"loss": 0.1963,
|
| 1559 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 1560 |
+
"rewards/chosen": -0.30832600593566895,
|
| 1561 |
+
"rewards/margins": 0.23151281476020813,
|
| 1562 |
+
"rewards/rejected": -0.5398387908935547,
|
| 1563 |
+
"step": 1100
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 0.85,
|
| 1567 |
+
"learning_rate": 3.529032613257574e-07,
|
| 1568 |
+
"logits/chosen": 0.10837472975254059,
|
| 1569 |
+
"logits/rejected": 0.17577466368675232,
|
| 1570 |
+
"logps/chosen": -682.72216796875,
|
| 1571 |
+
"logps/rejected": -828.9464111328125,
|
| 1572 |
+
"loss": 0.2376,
|
| 1573 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 1574 |
+
"rewards/chosen": -0.3374750316143036,
|
| 1575 |
+
"rewards/margins": 0.20901791751384735,
|
| 1576 |
+
"rewards/rejected": -0.5464929342269897,
|
| 1577 |
+
"step": 1110
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 0.85,
|
| 1581 |
+
"learning_rate": 3.195735209788528e-07,
|
| 1582 |
+
"logits/chosen": 0.1015692800283432,
|
| 1583 |
+
"logits/rejected": 0.1785646677017212,
|
| 1584 |
+
"logps/chosen": -639.4000244140625,
|
| 1585 |
+
"logps/rejected": -781.9521484375,
|
| 1586 |
+
"loss": 0.2455,
|
| 1587 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 1588 |
+
"rewards/chosen": -0.33476072549819946,
|
| 1589 |
+
"rewards/margins": 0.18203067779541016,
|
| 1590 |
+
"rewards/rejected": -0.5167914032936096,
|
| 1591 |
+
"step": 1120
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 0.86,
|
| 1595 |
+
"learning_rate": 2.8778921841253774e-07,
|
| 1596 |
+
"logits/chosen": 0.06415721774101257,
|
| 1597 |
+
"logits/rejected": 0.2145998477935791,
|
| 1598 |
+
"logps/chosen": -690.4867553710938,
|
| 1599 |
+
"logps/rejected": -865.8243408203125,
|
| 1600 |
+
"loss": 0.1726,
|
| 1601 |
+
"rewards/accuracies": 0.8125,
|
| 1602 |
+
"rewards/chosen": -0.3243727684020996,
|
| 1603 |
+
"rewards/margins": 0.251103937625885,
|
| 1604 |
+
"rewards/rejected": -0.5754767656326294,
|
| 1605 |
+
"step": 1130
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 0.87,
|
| 1609 |
+
"learning_rate": 2.5757288163336806e-07,
|
| 1610 |
+
"logits/chosen": 0.08277393132448196,
|
| 1611 |
+
"logits/rejected": 0.20556513965129852,
|
| 1612 |
+
"logps/chosen": -701.6412353515625,
|
| 1613 |
+
"logps/rejected": -884.2093505859375,
|
| 1614 |
+
"loss": 0.2016,
|
| 1615 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 1616 |
+
"rewards/chosen": -0.34318509697914124,
|
| 1617 |
+
"rewards/margins": 0.23703384399414062,
|
| 1618 |
+
"rewards/rejected": -0.5802189707756042,
|
| 1619 |
+
"step": 1140
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 0.88,
|
| 1623 |
+
"learning_rate": 2.2894592730863336e-07,
|
| 1624 |
+
"logits/chosen": 0.08898656070232391,
|
| 1625 |
+
"logits/rejected": 0.19494621455669403,
|
| 1626 |
+
"logps/chosen": -663.3425903320312,
|
| 1627 |
+
"logps/rejected": -838.2100830078125,
|
| 1628 |
+
"loss": 0.2081,
|
| 1629 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 1630 |
+
"rewards/chosen": -0.3233157992362976,
|
| 1631 |
+
"rewards/margins": 0.22612051665782928,
|
| 1632 |
+
"rewards/rejected": -0.5494363903999329,
|
| 1633 |
+
"step": 1150
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 0.88,
|
| 1637 |
+
"learning_rate": 2.019286455866981e-07,
|
| 1638 |
+
"logits/chosen": 0.09930244833230972,
|
| 1639 |
+
"logits/rejected": 0.25030988454818726,
|
| 1640 |
+
"logps/chosen": -665.492919921875,
|
| 1641 |
+
"logps/rejected": -791.5203857421875,
|
| 1642 |
+
"loss": 0.2154,
|
| 1643 |
+
"rewards/accuracies": 0.706250011920929,
|
| 1644 |
+
"rewards/chosen": -0.3297829031944275,
|
| 1645 |
+
"rewards/margins": 0.19049373269081116,
|
| 1646 |
+
"rewards/rejected": -0.520276665687561,
|
| 1647 |
+
"step": 1160
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.89,
|
| 1651 |
+
"learning_rate": 1.7654018571579557e-07,
|
| 1652 |
+
"logits/chosen": 0.07714973390102386,
|
| 1653 |
+
"logits/rejected": 0.1515274941921234,
|
| 1654 |
+
"logps/chosen": -710.2567749023438,
|
| 1655 |
+
"logps/rejected": -828.8133544921875,
|
| 1656 |
+
"loss": 0.2217,
|
| 1657 |
+
"rewards/accuracies": 0.71875,
|
| 1658 |
+
"rewards/chosen": -0.3351573944091797,
|
| 1659 |
+
"rewards/margins": 0.20143434405326843,
|
| 1660 |
+
"rewards/rejected": -0.5365917086601257,
|
| 1661 |
+
"step": 1170
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 0.9,
|
| 1665 |
+
"learning_rate": 1.5279854247146703e-07,
|
| 1666 |
+
"logits/chosen": 0.08655952662229538,
|
| 1667 |
+
"logits/rejected": 0.18316319584846497,
|
| 1668 |
+
"logps/chosen": -664.911376953125,
|
| 1669 |
+
"logps/rejected": -804.0281982421875,
|
| 1670 |
+
"loss": 0.2363,
|
| 1671 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 1672 |
+
"rewards/chosen": -0.3494827151298523,
|
| 1673 |
+
"rewards/margins": 0.19858424365520477,
|
| 1674 |
+
"rewards/rejected": -0.5480669140815735,
|
| 1675 |
+
"step": 1180
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 0.91,
|
| 1679 |
+
"learning_rate": 1.307205434022671e-07,
|
| 1680 |
+
"logits/chosen": 0.0854581817984581,
|
| 1681 |
+
"logits/rejected": 0.21030446887016296,
|
| 1682 |
+
"logps/chosen": -680.3010864257812,
|
| 1683 |
+
"logps/rejected": -877.068359375,
|
| 1684 |
+
"loss": 0.1977,
|
| 1685 |
+
"rewards/accuracies": 0.793749988079071,
|
| 1686 |
+
"rewards/chosen": -0.33833056688308716,
|
| 1687 |
+
"rewards/margins": 0.23077313601970673,
|
| 1688 |
+
"rewards/rejected": -0.5691036581993103,
|
| 1689 |
+
"step": 1190
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 0.91,
|
| 1693 |
+
"learning_rate": 1.1032183690276754e-07,
|
| 1694 |
+
"logits/chosen": 0.07660888135433197,
|
| 1695 |
+
"logits/rejected": 0.1921808272600174,
|
| 1696 |
+
"logps/chosen": -673.0133056640625,
|
| 1697 |
+
"logps/rejected": -855.6375732421875,
|
| 1698 |
+
"loss": 0.2076,
|
| 1699 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 1700 |
+
"rewards/chosen": -0.32487696409225464,
|
| 1701 |
+
"rewards/margins": 0.2425091564655304,
|
| 1702 |
+
"rewards/rejected": -0.5673861503601074,
|
| 1703 |
+
"step": 1200
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 0.92,
|
| 1707 |
+
"learning_rate": 9.161688112232836e-08,
|
| 1708 |
+
"logits/chosen": 0.10119873285293579,
|
| 1709 |
+
"logits/rejected": 0.22447247803211212,
|
| 1710 |
+
"logps/chosen": -711.1212158203125,
|
| 1711 |
+
"logps/rejected": -857.6790771484375,
|
| 1712 |
+
"loss": 0.2041,
|
| 1713 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1714 |
+
"rewards/chosen": -0.35159099102020264,
|
| 1715 |
+
"rewards/margins": 0.22090163826942444,
|
| 1716 |
+
"rewards/rejected": -0.5724925994873047,
|
| 1717 |
+
"step": 1210
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 0.93,
|
| 1721 |
+
"learning_rate": 7.46189337174788e-08,
|
| 1722 |
+
"logits/chosen": 0.0487164705991745,
|
| 1723 |
+
"logits/rejected": 0.1770685911178589,
|
| 1724 |
+
"logps/chosen": -655.6456298828125,
|
| 1725 |
+
"logps/rejected": -825.86572265625,
|
| 1726 |
+
"loss": 0.1868,
|
| 1727 |
+
"rewards/accuracies": 0.737500011920929,
|
| 1728 |
+
"rewards/chosen": -0.3402259647846222,
|
| 1729 |
+
"rewards/margins": 0.2170572578907013,
|
| 1730 |
+
"rewards/rejected": -0.5572832822799683,
|
| 1731 |
+
"step": 1220
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 0.94,
|
| 1735 |
+
"learning_rate": 5.934004245518793e-08,
|
| 1736 |
+
"logits/chosen": 0.09085109829902649,
|
| 1737 |
+
"logits/rejected": 0.22920957207679749,
|
| 1738 |
+
"logps/chosen": -656.9745483398438,
|
| 1739 |
+
"logps/rejected": -807.7351684570312,
|
| 1740 |
+
"loss": 0.2086,
|
| 1741 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 1742 |
+
"rewards/chosen": -0.3091570734977722,
|
| 1743 |
+
"rewards/margins": 0.2053556740283966,
|
| 1744 |
+
"rewards/rejected": -0.5145127177238464,
|
| 1745 |
+
"step": 1230
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 0.94,
|
| 1749 |
+
"learning_rate": 4.579103667367385e-08,
|
| 1750 |
+
"logits/chosen": 0.10511846840381622,
|
| 1751 |
+
"logits/rejected": 0.22097325325012207,
|
| 1752 |
+
"logps/chosen": -677.5391235351562,
|
| 1753 |
+
"logps/rejected": -815.5825805664062,
|
| 1754 |
+
"loss": 0.2188,
|
| 1755 |
+
"rewards/accuracies": 0.75,
|
| 1756 |
+
"rewards/chosen": -0.3394906222820282,
|
| 1757 |
+
"rewards/margins": 0.2081030309200287,
|
| 1758 |
+
"rewards/rejected": -0.5475937128067017,
|
| 1759 |
+
"step": 1240
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 0.95,
|
| 1763 |
+
"learning_rate": 3.398151960681162e-08,
|
| 1764 |
+
"logits/chosen": 0.04612383991479874,
|
| 1765 |
+
"logits/rejected": 0.16620075702667236,
|
| 1766 |
+
"logps/chosen": -679.2454833984375,
|
| 1767 |
+
"logps/rejected": -800.4119262695312,
|
| 1768 |
+
"loss": 0.2539,
|
| 1769 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 1770 |
+
"rewards/chosen": -0.35038450360298157,
|
| 1771 |
+
"rewards/margins": 0.1713865101337433,
|
| 1772 |
+
"rewards/rejected": -0.5217710733413696,
|
| 1773 |
+
"step": 1250
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 0.96,
|
| 1777 |
+
"learning_rate": 2.3919861577572924e-08,
|
| 1778 |
+
"logits/chosen": 0.11022261530160904,
|
| 1779 |
+
"logits/rejected": 0.166605606675148,
|
| 1780 |
+
"logps/chosen": -690.1370239257812,
|
| 1781 |
+
"logps/rejected": -797.578369140625,
|
| 1782 |
+
"loss": 0.2181,
|
| 1783 |
+
"rewards/accuracies": 0.768750011920929,
|
| 1784 |
+
"rewards/chosen": -0.32749998569488525,
|
| 1785 |
+
"rewards/margins": 0.20777150988578796,
|
| 1786 |
+
"rewards/rejected": -0.5352715253829956,
|
| 1787 |
+
"step": 1260
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 0.97,
|
| 1791 |
+
"learning_rate": 1.5613194065327854e-08,
|
| 1792 |
+
"logits/chosen": 0.052184127271175385,
|
| 1793 |
+
"logits/rejected": 0.22973528504371643,
|
| 1794 |
+
"logps/chosen": -629.1040649414062,
|
| 1795 |
+
"logps/rejected": -766.4088134765625,
|
| 1796 |
+
"loss": 0.1872,
|
| 1797 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 1798 |
+
"rewards/chosen": -0.2912130355834961,
|
| 1799 |
+
"rewards/margins": 0.22654838860034943,
|
| 1800 |
+
"rewards/rejected": -0.5177614092826843,
|
| 1801 |
+
"step": 1270
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 0.98,
|
| 1805 |
+
"learning_rate": 9.067404651211808e-09,
|
| 1806 |
+
"logits/chosen": 0.11986621469259262,
|
| 1807 |
+
"logits/rejected": 0.21088480949401855,
|
| 1808 |
+
"logps/chosen": -687.1105346679688,
|
| 1809 |
+
"logps/rejected": -877.3173828125,
|
| 1810 |
+
"loss": 0.2028,
|
| 1811 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1812 |
+
"rewards/chosen": -0.3289037048816681,
|
| 1813 |
+
"rewards/margins": 0.23036351799964905,
|
| 1814 |
+
"rewards/rejected": -0.5592672228813171,
|
| 1815 |
+
"step": 1280
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 0.98,
|
| 1819 |
+
"learning_rate": 4.287132845137709e-09,
|
| 1820 |
+
"logits/chosen": 0.08242613077163696,
|
| 1821 |
+
"logits/rejected": 0.20786967873573303,
|
| 1822 |
+
"logps/chosen": -682.5579223632812,
|
| 1823 |
+
"logps/rejected": -803.6607666015625,
|
| 1824 |
+
"loss": 0.2108,
|
| 1825 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 1826 |
+
"rewards/chosen": -0.32676413655281067,
|
| 1827 |
+
"rewards/margins": 0.19439134001731873,
|
| 1828 |
+
"rewards/rejected": -0.5211554765701294,
|
| 1829 |
+
"step": 1290
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 0.99,
|
| 1833 |
+
"learning_rate": 1.2757667974155896e-09,
|
| 1834 |
+
"logits/chosen": 0.03548216074705124,
|
| 1835 |
+
"logits/rejected": 0.12074669450521469,
|
| 1836 |
+
"logps/chosen": -665.2953491210938,
|
| 1837 |
+
"logps/rejected": -846.1324462890625,
|
| 1838 |
+
"loss": 0.2049,
|
| 1839 |
+
"rewards/accuracies": 0.737500011920929,
|
| 1840 |
+
"rewards/chosen": -0.3495555818080902,
|
| 1841 |
+
"rewards/margins": 0.1981724500656128,
|
| 1842 |
+
"rewards/rejected": -0.5477280020713806,
|
| 1843 |
+
"step": 1300
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 1.0,
|
| 1847 |
+
"learning_rate": 3.544089730633804e-11,
|
| 1848 |
+
"logits/chosen": 0.10499806702136993,
|
| 1849 |
+
"logits/rejected": 0.2223556488752365,
|
| 1850 |
+
"logps/chosen": -656.1048583984375,
|
| 1851 |
+
"logps/rejected": -808.8086547851562,
|
| 1852 |
+
"loss": 0.1796,
|
| 1853 |
+
"rewards/accuracies": 0.71875,
|
| 1854 |
+
"rewards/chosen": -0.3139723837375641,
|
| 1855 |
+
"rewards/margins": 0.22497034072875977,
|
| 1856 |
+
"rewards/rejected": -0.5389427542686462,
|
| 1857 |
+
"step": 1310
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 1.0,
|
| 1861 |
+
"step": 1312,
|
| 1862 |
+
"total_flos": 0.0,
|
| 1863 |
+
"train_loss": 0.24134081795175627,
|
| 1864 |
+
"train_runtime": 12108.13,
|
| 1865 |
+
"train_samples_per_second": 1.734,
|
| 1866 |
+
"train_steps_per_second": 0.108
|
| 1867 |
+
}
|
| 1868 |
+
],
|
| 1869 |
+
"logging_steps": 10,
|
| 1870 |
+
"max_steps": 1312,
|
| 1871 |
+
"num_input_tokens_seen": 0,
|
| 1872 |
+
"num_train_epochs": 1,
|
| 1873 |
+
"save_steps": 100,
|
| 1874 |
+
"total_flos": 0.0,
|
| 1875 |
+
"train_batch_size": 4,
|
| 1876 |
+
"trial_name": null,
|
| 1877 |
+
"trial_params": null
|
| 1878 |
+
}
|