Model save
Browse files- README.md +59 -0
- adapter_model.safetensors +1 -1
- all_results.json +8 -0
- runs/May20_19-09-07_gpu4-119-5/events.out.tfevents.1716196294.gpu4-119-5.941181.0 +2 -2
- train_results.json +8 -0
- trainer_state.json +1878 -0
README.md
ADDED
|
@@ -0,0 +1,59 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: peft
|
| 3 |
+
tags:
|
| 4 |
+
- trl
|
| 5 |
+
- dpo
|
| 6 |
+
- generated_from_trainer
|
| 7 |
+
base_model: DUAL-GPO/phi-2-irepo-chatml-merged-i0
|
| 8 |
+
model-index:
|
| 9 |
+
- name: phi-2-irepo-chatml-v6-i1
|
| 10 |
+
results: []
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 14 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 15 |
+
|
| 16 |
+
# phi-2-irepo-chatml-v6-i1
|
| 17 |
+
|
| 18 |
+
This model is a fine-tuned version of [DUAL-GPO/phi-2-irepo-chatml-merged-i0](https://huggingface.co/DUAL-GPO/phi-2-irepo-chatml-merged-i0) on the None dataset.
|
| 19 |
+
|
| 20 |
+
## Model description
|
| 21 |
+
|
| 22 |
+
More information needed
|
| 23 |
+
|
| 24 |
+
## Intended uses & limitations
|
| 25 |
+
|
| 26 |
+
More information needed
|
| 27 |
+
|
| 28 |
+
## Training and evaluation data
|
| 29 |
+
|
| 30 |
+
More information needed
|
| 31 |
+
|
| 32 |
+
## Training procedure
|
| 33 |
+
|
| 34 |
+
### Training hyperparameters
|
| 35 |
+
|
| 36 |
+
The following hyperparameters were used during training:
|
| 37 |
+
- learning_rate: 5e-06
|
| 38 |
+
- train_batch_size: 4
|
| 39 |
+
- eval_batch_size: 4
|
| 40 |
+
- seed: 42
|
| 41 |
+
- distributed_type: multi-GPU
|
| 42 |
+
- gradient_accumulation_steps: 4
|
| 43 |
+
- total_train_batch_size: 16
|
| 44 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 45 |
+
- lr_scheduler_type: cosine
|
| 46 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 47 |
+
- num_epochs: 1
|
| 48 |
+
|
| 49 |
+
### Training results
|
| 50 |
+
|
| 51 |
+
|
| 52 |
+
|
| 53 |
+
### Framework versions
|
| 54 |
+
|
| 55 |
+
- PEFT 0.7.1
|
| 56 |
+
- Transformers 4.36.2
|
| 57 |
+
- Pytorch 2.1.2+cu121
|
| 58 |
+
- Datasets 2.14.6
|
| 59 |
+
- 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 335579632
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e23a501f5b5dd40d12fb0b8eb4b1dbfa05f23050b8cbc8a1b071ec239df0ac9f
|
| 3 |
size 335579632
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.20755663267677513,
|
| 4 |
+
"train_runtime": 11765.5627,
|
| 5 |
+
"train_samples": 21000,
|
| 6 |
+
"train_samples_per_second": 1.785,
|
| 7 |
+
"train_steps_per_second": 0.112
|
| 8 |
+
}
|
runs/May20_19-09-07_gpu4-119-5/events.out.tfevents.1716196294.gpu4-119-5.941181.0
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 sha256:2a7e49606ed870668ddc7cbdf3b237d85ce507cd0898d829c0d9f3b81143dd37
|
| 3 |
+
size 88808
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.20755663267677513,
|
| 4 |
+
"train_runtime": 11765.5627,
|
| 5 |
+
"train_samples": 21000,
|
| 6 |
+
"train_samples_per_second": 1.785,
|
| 7 |
+
"train_steps_per_second": 0.112
|
| 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": 0.3414086103439331,
|
| 15 |
+
"logits/rejected": 0.3618736267089844,
|
| 16 |
+
"logps/chosen": -457.8209228515625,
|
| 17 |
+
"logps/rejected": -477.39923095703125,
|
| 18 |
+
"loss": 0.2223,
|
| 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": 0.3330618441104889,
|
| 29 |
+
"logits/rejected": 0.3205517530441284,
|
| 30 |
+
"logps/chosen": -306.0744323730469,
|
| 31 |
+
"logps/rejected": -326.5745849609375,
|
| 32 |
+
"loss": 0.2099,
|
| 33 |
+
"rewards/accuracies": 0.2777777910232544,
|
| 34 |
+
"rewards/chosen": -0.00023350439732894301,
|
| 35 |
+
"rewards/margins": 6.652800948359072e-05,
|
| 36 |
+
"rewards/rejected": -0.0003000324359163642,
|
| 37 |
+
"step": 10
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.02,
|
| 41 |
+
"learning_rate": 7.575757575757576e-07,
|
| 42 |
+
"logits/chosen": 0.3300573229789734,
|
| 43 |
+
"logits/rejected": 0.3766365945339203,
|
| 44 |
+
"logps/chosen": -294.01226806640625,
|
| 45 |
+
"logps/rejected": -328.8618469238281,
|
| 46 |
+
"loss": 0.2087,
|
| 47 |
+
"rewards/accuracies": 0.3062500059604645,
|
| 48 |
+
"rewards/chosen": 9.059391595656052e-05,
|
| 49 |
+
"rewards/margins": 1.657838947721757e-05,
|
| 50 |
+
"rewards/rejected": 7.401553739327937e-05,
|
| 51 |
+
"step": 20
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.02,
|
| 55 |
+
"learning_rate": 1.1363636363636364e-06,
|
| 56 |
+
"logits/chosen": 0.33157598972320557,
|
| 57 |
+
"logits/rejected": 0.3647070527076721,
|
| 58 |
+
"logps/chosen": -308.50250244140625,
|
| 59 |
+
"logps/rejected": -316.5055236816406,
|
| 60 |
+
"loss": 0.21,
|
| 61 |
+
"rewards/accuracies": 0.26249998807907104,
|
| 62 |
+
"rewards/chosen": -9.306786523666233e-05,
|
| 63 |
+
"rewards/margins": -5.975523890811019e-06,
|
| 64 |
+
"rewards/rejected": -8.709232497494668e-05,
|
| 65 |
+
"step": 30
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.03,
|
| 69 |
+
"learning_rate": 1.5151515151515152e-06,
|
| 70 |
+
"logits/chosen": 0.30125534534454346,
|
| 71 |
+
"logits/rejected": 0.3123665452003479,
|
| 72 |
+
"logps/chosen": -278.3026123046875,
|
| 73 |
+
"logps/rejected": -290.43365478515625,
|
| 74 |
+
"loss": 0.221,
|
| 75 |
+
"rewards/accuracies": 0.20624999701976776,
|
| 76 |
+
"rewards/chosen": -1.4783197912038304e-05,
|
| 77 |
+
"rewards/margins": 5.556107498705387e-05,
|
| 78 |
+
"rewards/rejected": -7.034426380414516e-05,
|
| 79 |
+
"step": 40
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.04,
|
| 83 |
+
"learning_rate": 1.8939393939393941e-06,
|
| 84 |
+
"logits/chosen": 0.33402004837989807,
|
| 85 |
+
"logits/rejected": 0.3797036111354828,
|
| 86 |
+
"logps/chosen": -250.8700408935547,
|
| 87 |
+
"logps/rejected": -274.0110778808594,
|
| 88 |
+
"loss": 0.2,
|
| 89 |
+
"rewards/accuracies": 0.26249998807907104,
|
| 90 |
+
"rewards/chosen": 0.0001303141616517678,
|
| 91 |
+
"rewards/margins": 2.952014074253384e-05,
|
| 92 |
+
"rewards/rejected": 0.00010079403000418097,
|
| 93 |
+
"step": 50
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.05,
|
| 97 |
+
"learning_rate": 2.2727272727272728e-06,
|
| 98 |
+
"logits/chosen": 0.3142412304878235,
|
| 99 |
+
"logits/rejected": 0.3297632336616516,
|
| 100 |
+
"logps/chosen": -261.3359069824219,
|
| 101 |
+
"logps/rejected": -272.4247131347656,
|
| 102 |
+
"loss": 0.216,
|
| 103 |
+
"rewards/accuracies": 0.3125,
|
| 104 |
+
"rewards/chosen": 7.56668159738183e-05,
|
| 105 |
+
"rewards/margins": 0.0001469437702326104,
|
| 106 |
+
"rewards/rejected": -7.12769542587921e-05,
|
| 107 |
+
"step": 60
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.05,
|
| 111 |
+
"learning_rate": 2.6515151515151514e-06,
|
| 112 |
+
"logits/chosen": 0.31074008345603943,
|
| 113 |
+
"logits/rejected": 0.3008107542991638,
|
| 114 |
+
"logps/chosen": -305.2731628417969,
|
| 115 |
+
"logps/rejected": -301.9687805175781,
|
| 116 |
+
"loss": 0.21,
|
| 117 |
+
"rewards/accuracies": 0.25,
|
| 118 |
+
"rewards/chosen": -0.0002583070017863065,
|
| 119 |
+
"rewards/margins": -3.105907671852037e-05,
|
| 120 |
+
"rewards/rejected": -0.0002272479177918285,
|
| 121 |
+
"step": 70
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.06,
|
| 125 |
+
"learning_rate": 3.0303030303030305e-06,
|
| 126 |
+
"logits/chosen": 0.310350626707077,
|
| 127 |
+
"logits/rejected": 0.3315640389919281,
|
| 128 |
+
"logps/chosen": -232.4719696044922,
|
| 129 |
+
"logps/rejected": -277.22833251953125,
|
| 130 |
+
"loss": 0.1962,
|
| 131 |
+
"rewards/accuracies": 0.26875001192092896,
|
| 132 |
+
"rewards/chosen": 9.660933574195951e-05,
|
| 133 |
+
"rewards/margins": 0.00016913158469833434,
|
| 134 |
+
"rewards/rejected": -7.252227078424767e-05,
|
| 135 |
+
"step": 80
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.07,
|
| 139 |
+
"learning_rate": 3.409090909090909e-06,
|
| 140 |
+
"logits/chosen": 0.28212469816207886,
|
| 141 |
+
"logits/rejected": 0.35147595405578613,
|
| 142 |
+
"logps/chosen": -291.93121337890625,
|
| 143 |
+
"logps/rejected": -309.50341796875,
|
| 144 |
+
"loss": 0.2222,
|
| 145 |
+
"rewards/accuracies": 0.3062500059604645,
|
| 146 |
+
"rewards/chosen": -0.0006029252544976771,
|
| 147 |
+
"rewards/margins": 0.00015305924171116203,
|
| 148 |
+
"rewards/rejected": -0.0007559844525530934,
|
| 149 |
+
"step": 90
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.08,
|
| 153 |
+
"learning_rate": 3.7878787878787882e-06,
|
| 154 |
+
"logits/chosen": 0.2799969017505646,
|
| 155 |
+
"logits/rejected": 0.33390945196151733,
|
| 156 |
+
"logps/chosen": -313.59515380859375,
|
| 157 |
+
"logps/rejected": -333.5437316894531,
|
| 158 |
+
"loss": 0.2158,
|
| 159 |
+
"rewards/accuracies": 0.3187499940395355,
|
| 160 |
+
"rewards/chosen": -0.0013320285361260176,
|
| 161 |
+
"rewards/margins": 0.00041345407953485847,
|
| 162 |
+
"rewards/rejected": -0.001745482673868537,
|
| 163 |
+
"step": 100
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.08,
|
| 167 |
+
"learning_rate": 4.166666666666667e-06,
|
| 168 |
+
"logits/chosen": 0.3202076852321625,
|
| 169 |
+
"logits/rejected": 0.31327325105667114,
|
| 170 |
+
"logps/chosen": -339.7900085449219,
|
| 171 |
+
"logps/rejected": -391.5009765625,
|
| 172 |
+
"loss": 0.2034,
|
| 173 |
+
"rewards/accuracies": 0.39375001192092896,
|
| 174 |
+
"rewards/chosen": -0.0016250547487288713,
|
| 175 |
+
"rewards/margins": 0.000587599235586822,
|
| 176 |
+
"rewards/rejected": -0.002212654100731015,
|
| 177 |
+
"step": 110
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.09,
|
| 181 |
+
"learning_rate": 4.5454545454545455e-06,
|
| 182 |
+
"logits/chosen": 0.2636483311653137,
|
| 183 |
+
"logits/rejected": 0.3315524756908417,
|
| 184 |
+
"logps/chosen": -257.42376708984375,
|
| 185 |
+
"logps/rejected": -279.6871643066406,
|
| 186 |
+
"loss": 0.2195,
|
| 187 |
+
"rewards/accuracies": 0.3187499940395355,
|
| 188 |
+
"rewards/chosen": -0.0009949387749657035,
|
| 189 |
+
"rewards/margins": 0.0003876467817462981,
|
| 190 |
+
"rewards/rejected": -0.0013825856149196625,
|
| 191 |
+
"step": 120
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.1,
|
| 195 |
+
"learning_rate": 4.924242424242425e-06,
|
| 196 |
+
"logits/chosen": 0.25745201110839844,
|
| 197 |
+
"logits/rejected": 0.3214574456214905,
|
| 198 |
+
"logps/chosen": -302.13787841796875,
|
| 199 |
+
"logps/rejected": -334.5384826660156,
|
| 200 |
+
"loss": 0.1975,
|
| 201 |
+
"rewards/accuracies": 0.36250001192092896,
|
| 202 |
+
"rewards/chosen": -0.0025944842491298914,
|
| 203 |
+
"rewards/margins": 0.0014145069289952517,
|
| 204 |
+
"rewards/rejected": -0.004008991178125143,
|
| 205 |
+
"step": 130
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.11,
|
| 209 |
+
"learning_rate": 4.999432965739786e-06,
|
| 210 |
+
"logits/chosen": 0.26108819246292114,
|
| 211 |
+
"logits/rejected": 0.21493813395500183,
|
| 212 |
+
"logps/chosen": -321.00067138671875,
|
| 213 |
+
"logps/rejected": -349.4842834472656,
|
| 214 |
+
"loss": 0.2053,
|
| 215 |
+
"rewards/accuracies": 0.3687500059604645,
|
| 216 |
+
"rewards/chosen": -0.003485953900963068,
|
| 217 |
+
"rewards/margins": 0.0010465523228049278,
|
| 218 |
+
"rewards/rejected": -0.0045325057581067085,
|
| 219 |
+
"step": 140
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.11,
|
| 223 |
+
"learning_rate": 4.997129829895409e-06,
|
| 224 |
+
"logits/chosen": 0.16764743626117706,
|
| 225 |
+
"logits/rejected": 0.20418021082878113,
|
| 226 |
+
"logps/chosen": -216.89199829101562,
|
| 227 |
+
"logps/rejected": -234.42935180664062,
|
| 228 |
+
"loss": 0.208,
|
| 229 |
+
"rewards/accuracies": 0.2562499940395355,
|
| 230 |
+
"rewards/chosen": -0.0010050686541944742,
|
| 231 |
+
"rewards/margins": 0.0011736680753529072,
|
| 232 |
+
"rewards/rejected": -0.0021787367295473814,
|
| 233 |
+
"step": 150
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.12,
|
| 237 |
+
"learning_rate": 4.9930567839810125e-06,
|
| 238 |
+
"logits/chosen": 0.13134139776229858,
|
| 239 |
+
"logits/rejected": 0.14908884465694427,
|
| 240 |
+
"logps/chosen": -268.61419677734375,
|
| 241 |
+
"logps/rejected": -301.322265625,
|
| 242 |
+
"loss": 0.2203,
|
| 243 |
+
"rewards/accuracies": 0.3375000059604645,
|
| 244 |
+
"rewards/chosen": -0.0038432697765529156,
|
| 245 |
+
"rewards/margins": 0.005182007793337107,
|
| 246 |
+
"rewards/rejected": -0.009025278501212597,
|
| 247 |
+
"step": 160
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.13,
|
| 251 |
+
"learning_rate": 4.987216714880929e-06,
|
| 252 |
+
"logits/chosen": 0.1437113881111145,
|
| 253 |
+
"logits/rejected": 0.13215014338493347,
|
| 254 |
+
"logps/chosen": -259.3898010253906,
|
| 255 |
+
"logps/rejected": -267.3370056152344,
|
| 256 |
+
"loss": 0.2012,
|
| 257 |
+
"rewards/accuracies": 0.2750000059604645,
|
| 258 |
+
"rewards/chosen": -0.01510892529040575,
|
| 259 |
+
"rewards/margins": 0.0029346179217100143,
|
| 260 |
+
"rewards/rejected": -0.01804354228079319,
|
| 261 |
+
"step": 170
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.14,
|
| 265 |
+
"learning_rate": 4.979613761906212e-06,
|
| 266 |
+
"logits/chosen": 0.07960179448127747,
|
| 267 |
+
"logits/rejected": 0.08153261244297028,
|
| 268 |
+
"logps/chosen": -357.0888366699219,
|
| 269 |
+
"logps/rejected": -368.5904846191406,
|
| 270 |
+
"loss": 0.2207,
|
| 271 |
+
"rewards/accuracies": 0.29374998807907104,
|
| 272 |
+
"rewards/chosen": -0.021911371499300003,
|
| 273 |
+
"rewards/margins": 0.0033388130832463503,
|
| 274 |
+
"rewards/rejected": -0.025250181555747986,
|
| 275 |
+
"step": 180
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.14,
|
| 279 |
+
"learning_rate": 4.970253313860788e-06,
|
| 280 |
+
"logits/chosen": 0.12686966359615326,
|
| 281 |
+
"logits/rejected": 0.092551589012146,
|
| 282 |
+
"logps/chosen": -323.23162841796875,
|
| 283 |
+
"logps/rejected": -354.8710632324219,
|
| 284 |
+
"loss": 0.2065,
|
| 285 |
+
"rewards/accuracies": 0.32499998807907104,
|
| 286 |
+
"rewards/chosen": -0.027633970603346825,
|
| 287 |
+
"rewards/margins": 0.007219144143164158,
|
| 288 |
+
"rewards/rejected": -0.03485311195254326,
|
| 289 |
+
"step": 190
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.15,
|
| 293 |
+
"learning_rate": 4.959142005221991e-06,
|
| 294 |
+
"logits/chosen": 0.030479246750473976,
|
| 295 |
+
"logits/rejected": 0.03255733102560043,
|
| 296 |
+
"logps/chosen": -324.7137145996094,
|
| 297 |
+
"logps/rejected": -360.15802001953125,
|
| 298 |
+
"loss": 0.2118,
|
| 299 |
+
"rewards/accuracies": 0.29374998807907104,
|
| 300 |
+
"rewards/chosen": -0.036979734897613525,
|
| 301 |
+
"rewards/margins": 0.0006652610609307885,
|
| 302 |
+
"rewards/rejected": -0.037644993513822556,
|
| 303 |
+
"step": 200
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.16,
|
| 307 |
+
"learning_rate": 4.94628771143819e-06,
|
| 308 |
+
"logits/chosen": 0.011406493373215199,
|
| 309 |
+
"logits/rejected": 0.0717320591211319,
|
| 310 |
+
"logps/chosen": -335.43402099609375,
|
| 311 |
+
"logps/rejected": -334.8843078613281,
|
| 312 |
+
"loss": 0.2127,
|
| 313 |
+
"rewards/accuracies": 0.32499998807907104,
|
| 314 |
+
"rewards/chosen": -0.02695571817457676,
|
| 315 |
+
"rewards/margins": 0.0005523961153812706,
|
| 316 |
+
"rewards/rejected": -0.027508113533258438,
|
| 317 |
+
"step": 210
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.17,
|
| 321 |
+
"learning_rate": 4.931699543346854e-06,
|
| 322 |
+
"logits/chosen": 0.11805645376443863,
|
| 323 |
+
"logits/rejected": 0.13258619606494904,
|
| 324 |
+
"logps/chosen": -321.3063049316406,
|
| 325 |
+
"logps/rejected": -337.80084228515625,
|
| 326 |
+
"loss": 0.221,
|
| 327 |
+
"rewards/accuracies": 0.2874999940395355,
|
| 328 |
+
"rewards/chosen": -0.009172406047582626,
|
| 329 |
+
"rewards/margins": 0.0013926494866609573,
|
| 330 |
+
"rewards/rejected": -0.010565054602921009,
|
| 331 |
+
"step": 220
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.18,
|
| 335 |
+
"learning_rate": 4.9153878407169815e-06,
|
| 336 |
+
"logits/chosen": 0.04168776422739029,
|
| 337 |
+
"logits/rejected": 0.052837539464235306,
|
| 338 |
+
"logps/chosen": -321.91546630859375,
|
| 339 |
+
"logps/rejected": -337.6319274902344,
|
| 340 |
+
"loss": 0.2142,
|
| 341 |
+
"rewards/accuracies": 0.36250001192092896,
|
| 342 |
+
"rewards/chosen": 0.00017253603436984122,
|
| 343 |
+
"rewards/margins": 0.0029792343266308308,
|
| 344 |
+
"rewards/rejected": -0.002806698437780142,
|
| 345 |
+
"step": 230
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.18,
|
| 349 |
+
"learning_rate": 4.897364164920515e-06,
|
| 350 |
+
"logits/chosen": 0.0712689757347107,
|
| 351 |
+
"logits/rejected": 0.07041509449481964,
|
| 352 |
+
"logps/chosen": -275.91632080078125,
|
| 353 |
+
"logps/rejected": -272.06927490234375,
|
| 354 |
+
"loss": 0.2087,
|
| 355 |
+
"rewards/accuracies": 0.2562499940395355,
|
| 356 |
+
"rewards/chosen": 0.0034352331422269344,
|
| 357 |
+
"rewards/margins": 0.0012143031926825643,
|
| 358 |
+
"rewards/rejected": 0.002220930065959692,
|
| 359 |
+
"step": 240
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.19,
|
| 363 |
+
"learning_rate": 4.8776412907378845e-06,
|
| 364 |
+
"logits/chosen": 0.0784018337726593,
|
| 365 |
+
"logits/rejected": 0.07138744741678238,
|
| 366 |
+
"logps/chosen": -270.2529296875,
|
| 367 |
+
"logps/rejected": -276.2494812011719,
|
| 368 |
+
"loss": 0.2158,
|
| 369 |
+
"rewards/accuracies": 0.29374998807907104,
|
| 370 |
+
"rewards/chosen": 0.004147468134760857,
|
| 371 |
+
"rewards/margins": 0.0022368025965988636,
|
| 372 |
+
"rewards/rejected": 0.0019106656545773149,
|
| 373 |
+
"step": 250
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.2,
|
| 377 |
+
"learning_rate": 4.8562331973035396e-06,
|
| 378 |
+
"logits/chosen": 0.04922521486878395,
|
| 379 |
+
"logits/rejected": 0.03360098600387573,
|
| 380 |
+
"logps/chosen": -256.4879455566406,
|
| 381 |
+
"logps/rejected": -295.20220947265625,
|
| 382 |
+
"loss": 0.2052,
|
| 383 |
+
"rewards/accuracies": 0.3062500059604645,
|
| 384 |
+
"rewards/chosen": 0.0018207021057605743,
|
| 385 |
+
"rewards/margins": -4.667430039262399e-05,
|
| 386 |
+
"rewards/rejected": 0.0018673762679100037,
|
| 387 |
+
"step": 260
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.21,
|
| 391 |
+
"learning_rate": 4.833155058197842e-06,
|
| 392 |
+
"logits/chosen": 0.003297035349532962,
|
| 393 |
+
"logits/rejected": 0.05275546759366989,
|
| 394 |
+
"logps/chosen": -241.99462890625,
|
| 395 |
+
"logps/rejected": -263.92120361328125,
|
| 396 |
+
"loss": 0.2039,
|
| 397 |
+
"rewards/accuracies": 0.29374998807907104,
|
| 398 |
+
"rewards/chosen": 0.005943115334957838,
|
| 399 |
+
"rewards/margins": 0.004178833216428757,
|
| 400 |
+
"rewards/rejected": 0.0017642822349444032,
|
| 401 |
+
"step": 270
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.21,
|
| 405 |
+
"learning_rate": 4.808423230692374e-06,
|
| 406 |
+
"logits/chosen": 0.05449346452951431,
|
| 407 |
+
"logits/rejected": 0.06165250390768051,
|
| 408 |
+
"logps/chosen": -253.69369506835938,
|
| 409 |
+
"logps/rejected": -293.31585693359375,
|
| 410 |
+
"loss": 0.2029,
|
| 411 |
+
"rewards/accuracies": 0.3499999940395355,
|
| 412 |
+
"rewards/chosen": 0.012908290140330791,
|
| 413 |
+
"rewards/margins": 0.007234405726194382,
|
| 414 |
+
"rewards/rejected": 0.005673886742442846,
|
| 415 |
+
"step": 280
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.22,
|
| 419 |
+
"learning_rate": 4.7820552441562625e-06,
|
| 420 |
+
"logits/chosen": -0.005410021636635065,
|
| 421 |
+
"logits/rejected": 0.010305285453796387,
|
| 422 |
+
"logps/chosen": -270.04083251953125,
|
| 423 |
+
"logps/rejected": -316.2156066894531,
|
| 424 |
+
"loss": 0.1928,
|
| 425 |
+
"rewards/accuracies": 0.35624998807907104,
|
| 426 |
+
"rewards/chosen": 0.010301874950528145,
|
| 427 |
+
"rewards/margins": 0.007835518568754196,
|
| 428 |
+
"rewards/rejected": 0.002466357545927167,
|
| 429 |
+
"step": 290
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.23,
|
| 433 |
+
"learning_rate": 4.754069787631761e-06,
|
| 434 |
+
"logits/chosen": -0.05866356939077377,
|
| 435 |
+
"logits/rejected": -0.09499181061983109,
|
| 436 |
+
"logps/chosen": -276.0648193359375,
|
| 437 |
+
"logps/rejected": -317.32037353515625,
|
| 438 |
+
"loss": 0.203,
|
| 439 |
+
"rewards/accuracies": 0.32499998807907104,
|
| 440 |
+
"rewards/chosen": 0.008139651268720627,
|
| 441 |
+
"rewards/margins": 0.008907916024327278,
|
| 442 |
+
"rewards/rejected": -0.0007682637078687549,
|
| 443 |
+
"step": 300
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.24,
|
| 447 |
+
"learning_rate": 4.724486696587862e-06,
|
| 448 |
+
"logits/chosen": -0.08903555572032928,
|
| 449 |
+
"logits/rejected": -0.07178256660699844,
|
| 450 |
+
"logps/chosen": -291.4513244628906,
|
| 451 |
+
"logps/rejected": -304.63726806640625,
|
| 452 |
+
"loss": 0.2106,
|
| 453 |
+
"rewards/accuracies": 0.26875001192092896,
|
| 454 |
+
"rewards/chosen": 0.001623433781787753,
|
| 455 |
+
"rewards/margins": 0.0021732584573328495,
|
| 456 |
+
"rewards/rejected": -0.0005498243262991309,
|
| 457 |
+
"step": 310
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.24,
|
| 461 |
+
"learning_rate": 4.693326938861367e-06,
|
| 462 |
+
"logits/chosen": -0.12407398223876953,
|
| 463 |
+
"logits/rejected": -0.06067138910293579,
|
| 464 |
+
"logps/chosen": -249.7843475341797,
|
| 465 |
+
"logps/rejected": -290.00408935546875,
|
| 466 |
+
"loss": 0.2082,
|
| 467 |
+
"rewards/accuracies": 0.34375,
|
| 468 |
+
"rewards/chosen": 0.010884495452046394,
|
| 469 |
+
"rewards/margins": 0.008355050347745419,
|
| 470 |
+
"rewards/rejected": 0.002529443707317114,
|
| 471 |
+
"step": 320
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.25,
|
| 475 |
+
"learning_rate": 4.660612599795343e-06,
|
| 476 |
+
"logits/chosen": -0.11239596456289291,
|
| 477 |
+
"logits/rejected": -0.12089481204748154,
|
| 478 |
+
"logps/chosen": -290.61737060546875,
|
| 479 |
+
"logps/rejected": -312.4447937011719,
|
| 480 |
+
"loss": 0.2046,
|
| 481 |
+
"rewards/accuracies": 0.30000001192092896,
|
| 482 |
+
"rewards/chosen": 0.007386790122836828,
|
| 483 |
+
"rewards/margins": 0.009295441210269928,
|
| 484 |
+
"rewards/rejected": -0.0019086506217718124,
|
| 485 |
+
"step": 330
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.26,
|
| 489 |
+
"learning_rate": 4.626366866585528e-06,
|
| 490 |
+
"logits/chosen": -0.12021325528621674,
|
| 491 |
+
"logits/rejected": -0.11920841783285141,
|
| 492 |
+
"logps/chosen": -334.55303955078125,
|
| 493 |
+
"logps/rejected": -341.6528625488281,
|
| 494 |
+
"loss": 0.2095,
|
| 495 |
+
"rewards/accuracies": 0.32499998807907104,
|
| 496 |
+
"rewards/chosen": 0.005449886433780193,
|
| 497 |
+
"rewards/margins": -0.0001580852986080572,
|
| 498 |
+
"rewards/rejected": 0.0056079719215631485,
|
| 499 |
+
"step": 340
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.27,
|
| 503 |
+
"learning_rate": 4.590614011845758e-06,
|
| 504 |
+
"logits/chosen": -0.0888114720582962,
|
| 505 |
+
"logits/rejected": -0.11601515114307404,
|
| 506 |
+
"logps/chosen": -275.80316162109375,
|
| 507 |
+
"logps/rejected": -300.4934997558594,
|
| 508 |
+
"loss": 0.1974,
|
| 509 |
+
"rewards/accuracies": 0.3687500059604645,
|
| 510 |
+
"rewards/chosen": 0.01670825481414795,
|
| 511 |
+
"rewards/margins": 0.00685430783778429,
|
| 512 |
+
"rewards/rejected": 0.009853946976363659,
|
| 513 |
+
"step": 350
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.27,
|
| 517 |
+
"learning_rate": 4.553379376404085e-06,
|
| 518 |
+
"logits/chosen": -0.1058160662651062,
|
| 519 |
+
"logits/rejected": -0.11791692674160004,
|
| 520 |
+
"logps/chosen": -266.61322021484375,
|
| 521 |
+
"logps/rejected": -279.0782470703125,
|
| 522 |
+
"loss": 0.1983,
|
| 523 |
+
"rewards/accuracies": 0.3125,
|
| 524 |
+
"rewards/chosen": 0.021091196686029434,
|
| 525 |
+
"rewards/margins": 0.004850545898079872,
|
| 526 |
+
"rewards/rejected": 0.01624065265059471,
|
| 527 |
+
"step": 360
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.28,
|
| 531 |
+
"learning_rate": 4.514689351341751e-06,
|
| 532 |
+
"logits/chosen": -0.16150256991386414,
|
| 533 |
+
"logits/rejected": -0.1245572566986084,
|
| 534 |
+
"logps/chosen": -277.8587341308594,
|
| 535 |
+
"logps/rejected": -272.55792236328125,
|
| 536 |
+
"loss": 0.208,
|
| 537 |
+
"rewards/accuracies": 0.3062500059604645,
|
| 538 |
+
"rewards/chosen": 0.017171379178762436,
|
| 539 |
+
"rewards/margins": 8.804257959127426e-05,
|
| 540 |
+
"rewards/rejected": 0.017083335667848587,
|
| 541 |
+
"step": 370
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.29,
|
| 545 |
+
"learning_rate": 4.474571359287791e-06,
|
| 546 |
+
"logits/chosen": -0.1678142547607422,
|
| 547 |
+
"logits/rejected": -0.1356990933418274,
|
| 548 |
+
"logps/chosen": -280.3147277832031,
|
| 549 |
+
"logps/rejected": -318.61224365234375,
|
| 550 |
+
"loss": 0.2096,
|
| 551 |
+
"rewards/accuracies": 0.375,
|
| 552 |
+
"rewards/chosen": 0.024694612249732018,
|
| 553 |
+
"rewards/margins": 0.010597179643809795,
|
| 554 |
+
"rewards/rejected": 0.014097435399889946,
|
| 555 |
+
"step": 380
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.3,
|
| 559 |
+
"learning_rate": 4.4330538349824684e-06,
|
| 560 |
+
"logits/chosen": -0.12732991576194763,
|
| 561 |
+
"logits/rejected": -0.12050570547580719,
|
| 562 |
+
"logps/chosen": -287.99615478515625,
|
| 563 |
+
"logps/rejected": -286.3104553222656,
|
| 564 |
+
"loss": 0.2079,
|
| 565 |
+
"rewards/accuracies": 0.29374998807907104,
|
| 566 |
+
"rewards/chosen": 0.017507802695035934,
|
| 567 |
+
"rewards/margins": 1.922808587551117e-05,
|
| 568 |
+
"rewards/rejected": 0.017488572746515274,
|
| 569 |
+
"step": 390
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.3,
|
| 573 |
+
"learning_rate": 4.3901662051233755e-06,
|
| 574 |
+
"logits/chosen": -0.18693214654922485,
|
| 575 |
+
"logits/rejected": -0.11668721586465836,
|
| 576 |
+
"logps/chosen": -293.45611572265625,
|
| 577 |
+
"logps/rejected": -291.56011962890625,
|
| 578 |
+
"loss": 0.2156,
|
| 579 |
+
"rewards/accuracies": 0.2562499940395355,
|
| 580 |
+
"rewards/chosen": 0.011068233288824558,
|
| 581 |
+
"rewards/margins": -0.005082954186946154,
|
| 582 |
+
"rewards/rejected": 0.01615118607878685,
|
| 583 |
+
"step": 400
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.31,
|
| 587 |
+
"learning_rate": 4.345938867508439e-06,
|
| 588 |
+
"logits/chosen": -0.06379786878824234,
|
| 589 |
+
"logits/rejected": -0.07422082126140594,
|
| 590 |
+
"logps/chosen": -299.4942626953125,
|
| 591 |
+
"logps/rejected": -322.50201416015625,
|
| 592 |
+
"loss": 0.209,
|
| 593 |
+
"rewards/accuracies": 0.2874999940395355,
|
| 594 |
+
"rewards/chosen": 0.015014107339084148,
|
| 595 |
+
"rewards/margins": 0.003927754703909159,
|
| 596 |
+
"rewards/rejected": 0.011086350306868553,
|
| 597 |
+
"step": 410
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.32,
|
| 601 |
+
"learning_rate": 4.30040316949064e-06,
|
| 602 |
+
"logits/chosen": -0.0699755996465683,
|
| 603 |
+
"logits/rejected": -0.08232170343399048,
|
| 604 |
+
"logps/chosen": -288.906494140625,
|
| 605 |
+
"logps/rejected": -310.1832580566406,
|
| 606 |
+
"loss": 0.2186,
|
| 607 |
+
"rewards/accuracies": 0.3375000059604645,
|
| 608 |
+
"rewards/chosen": 0.015213333070278168,
|
| 609 |
+
"rewards/margins": 0.0060978480614721775,
|
| 610 |
+
"rewards/rejected": 0.009115484543144703,
|
| 611 |
+
"step": 420
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.33,
|
| 615 |
+
"learning_rate": 4.253591385759705e-06,
|
| 616 |
+
"logits/chosen": -0.10615179687738419,
|
| 617 |
+
"logits/rejected": -0.06481219828128815,
|
| 618 |
+
"logps/chosen": -278.0174255371094,
|
| 619 |
+
"logps/rejected": -334.2236328125,
|
| 620 |
+
"loss": 0.2108,
|
| 621 |
+
"rewards/accuracies": 0.3499999940395355,
|
| 622 |
+
"rewards/chosen": 0.012311488389968872,
|
| 623 |
+
"rewards/margins": 0.00595985259860754,
|
| 624 |
+
"rewards/rejected": 0.006351633928716183,
|
| 625 |
+
"step": 430
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.34,
|
| 629 |
+
"learning_rate": 4.205536695466524e-06,
|
| 630 |
+
"logits/chosen": -0.13743802905082703,
|
| 631 |
+
"logits/rejected": -0.133576899766922,
|
| 632 |
+
"logps/chosen": -274.42333984375,
|
| 633 |
+
"logps/rejected": -305.0398864746094,
|
| 634 |
+
"loss": 0.1894,
|
| 635 |
+
"rewards/accuracies": 0.28125,
|
| 636 |
+
"rewards/chosen": 0.00840366818010807,
|
| 637 |
+
"rewards/margins": 0.004010985605418682,
|
| 638 |
+
"rewards/rejected": 0.004392682109028101,
|
| 639 |
+
"step": 440
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.34,
|
| 643 |
+
"learning_rate": 4.15627315870651e-06,
|
| 644 |
+
"logits/chosen": -0.2103540152311325,
|
| 645 |
+
"logits/rejected": -0.19342456758022308,
|
| 646 |
+
"logps/chosen": -282.6641540527344,
|
| 647 |
+
"logps/rejected": -310.85113525390625,
|
| 648 |
+
"loss": 0.1974,
|
| 649 |
+
"rewards/accuracies": 0.3187499940395355,
|
| 650 |
+
"rewards/chosen": 0.00448573287576437,
|
| 651 |
+
"rewards/margins": 0.0029663294553756714,
|
| 652 |
+
"rewards/rejected": 0.0015194038860499859,
|
| 653 |
+
"step": 450
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.35,
|
| 657 |
+
"learning_rate": 4.105835692378557e-06,
|
| 658 |
+
"logits/chosen": -0.1941739320755005,
|
| 659 |
+
"logits/rejected": -0.22001402080059052,
|
| 660 |
+
"logps/chosen": -267.8673095703125,
|
| 661 |
+
"logps/rejected": -265.15618896484375,
|
| 662 |
+
"loss": 0.2101,
|
| 663 |
+
"rewards/accuracies": 0.2562499940395355,
|
| 664 |
+
"rewards/chosen": 0.007565724663436413,
|
| 665 |
+
"rewards/margins": 0.0029779626056551933,
|
| 666 |
+
"rewards/rejected": 0.004587762989103794,
|
| 667 |
+
"step": 460
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.36,
|
| 671 |
+
"learning_rate": 4.05426004543672e-06,
|
| 672 |
+
"logits/chosen": -0.2290777713060379,
|
| 673 |
+
"logits/rejected": -0.22092656791210175,
|
| 674 |
+
"logps/chosen": -267.7095947265625,
|
| 675 |
+
"logps/rejected": -307.23468017578125,
|
| 676 |
+
"loss": 0.2034,
|
| 677 |
+
"rewards/accuracies": 0.3499999940395355,
|
| 678 |
+
"rewards/chosen": 0.013477683067321777,
|
| 679 |
+
"rewards/margins": 0.006231679115444422,
|
| 680 |
+
"rewards/rejected": 0.007246003951877356,
|
| 681 |
+
"step": 470
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.37,
|
| 685 |
+
"learning_rate": 4.001582773552153e-06,
|
| 686 |
+
"logits/chosen": -0.2154918909072876,
|
| 687 |
+
"logits/rejected": -0.2263382226228714,
|
| 688 |
+
"logps/chosen": -236.22079467773438,
|
| 689 |
+
"logps/rejected": -281.47308349609375,
|
| 690 |
+
"loss": 0.208,
|
| 691 |
+
"rewards/accuracies": 0.3062500059604645,
|
| 692 |
+
"rewards/chosen": 0.008721614256501198,
|
| 693 |
+
"rewards/margins": 0.004514150787144899,
|
| 694 |
+
"rewards/rejected": 0.004207463003695011,
|
| 695 |
+
"step": 480
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.37,
|
| 699 |
+
"learning_rate": 3.947841213203262e-06,
|
| 700 |
+
"logits/chosen": -0.25622615218162537,
|
| 701 |
+
"logits/rejected": -0.28361567854881287,
|
| 702 |
+
"logps/chosen": -307.4781494140625,
|
| 703 |
+
"logps/rejected": -311.0342712402344,
|
| 704 |
+
"loss": 0.2118,
|
| 705 |
+
"rewards/accuracies": 0.3125,
|
| 706 |
+
"rewards/chosen": 0.00395470205694437,
|
| 707 |
+
"rewards/margins": 0.005559145472943783,
|
| 708 |
+
"rewards/rejected": -0.001604443765245378,
|
| 709 |
+
"step": 490
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.38,
|
| 713 |
+
"learning_rate": 3.893073455212438e-06,
|
| 714 |
+
"logits/chosen": -0.2454405277967453,
|
| 715 |
+
"logits/rejected": -0.23036327958106995,
|
| 716 |
+
"logps/chosen": -305.7126159667969,
|
| 717 |
+
"logps/rejected": -320.9441833496094,
|
| 718 |
+
"loss": 0.1978,
|
| 719 |
+
"rewards/accuracies": 0.29374998807907104,
|
| 720 |
+
"rewards/chosen": 0.006379422731697559,
|
| 721 |
+
"rewards/margins": 0.0032655359245836735,
|
| 722 |
+
"rewards/rejected": 0.003113886807113886,
|
| 723 |
+
"step": 500
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.39,
|
| 727 |
+
"learning_rate": 3.837318317748134e-06,
|
| 728 |
+
"logits/chosen": -0.23187056183815002,
|
| 729 |
+
"logits/rejected": -0.25806108117103577,
|
| 730 |
+
"logps/chosen": -295.49847412109375,
|
| 731 |
+
"logps/rejected": -312.9036560058594,
|
| 732 |
+
"loss": 0.2095,
|
| 733 |
+
"rewards/accuracies": 0.35624998807907104,
|
| 734 |
+
"rewards/chosen": 0.010056470520794392,
|
| 735 |
+
"rewards/margins": 0.005588938947767019,
|
| 736 |
+
"rewards/rejected": 0.004467530641704798,
|
| 737 |
+
"step": 510
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.4,
|
| 741 |
+
"learning_rate": 3.7806153188114027e-06,
|
| 742 |
+
"logits/chosen": -0.28555235266685486,
|
| 743 |
+
"logits/rejected": -0.27182719111442566,
|
| 744 |
+
"logps/chosen": -278.7008361816406,
|
| 745 |
+
"logps/rejected": -286.72845458984375,
|
| 746 |
+
"loss": 0.2112,
|
| 747 |
+
"rewards/accuracies": 0.3062500059604645,
|
| 748 |
+
"rewards/chosen": 0.0004449138359632343,
|
| 749 |
+
"rewards/margins": 0.0027928180061280727,
|
| 750 |
+
"rewards/rejected": -0.002347904024645686,
|
| 751 |
+
"step": 520
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.4,
|
| 755 |
+
"learning_rate": 3.7230046482264256e-06,
|
| 756 |
+
"logits/chosen": -0.29695600271224976,
|
| 757 |
+
"logits/rejected": -0.30071455240249634,
|
| 758 |
+
"logps/chosen": -303.6511535644531,
|
| 759 |
+
"logps/rejected": -311.7902526855469,
|
| 760 |
+
"loss": 0.2241,
|
| 761 |
+
"rewards/accuracies": 0.3062500059604645,
|
| 762 |
+
"rewards/chosen": 0.0028660064563155174,
|
| 763 |
+
"rewards/margins": -0.001624327152967453,
|
| 764 |
+
"rewards/rejected": 0.00449033360928297,
|
| 765 |
+
"step": 530
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.41,
|
| 769 |
+
"learning_rate": 3.6645271391548542e-06,
|
| 770 |
+
"logits/chosen": -0.2697906196117401,
|
| 771 |
+
"logits/rejected": -0.2420235425233841,
|
| 772 |
+
"logps/chosen": -287.1878662109375,
|
| 773 |
+
"logps/rejected": -289.3710632324219,
|
| 774 |
+
"loss": 0.1866,
|
| 775 |
+
"rewards/accuracies": 0.29374998807907104,
|
| 776 |
+
"rewards/chosen": 0.010242622345685959,
|
| 777 |
+
"rewards/margins": 0.0021990591194480658,
|
| 778 |
+
"rewards/rejected": 0.008043562062084675,
|
| 779 |
+
"step": 540
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.42,
|
| 783 |
+
"learning_rate": 3.6052242391541746e-06,
|
| 784 |
+
"logits/chosen": -0.2508285641670227,
|
| 785 |
+
"logits/rejected": -0.22416014969348907,
|
| 786 |
+
"logps/chosen": -282.50885009765625,
|
| 787 |
+
"logps/rejected": -304.0210876464844,
|
| 788 |
+
"loss": 0.1951,
|
| 789 |
+
"rewards/accuracies": 0.29374998807907104,
|
| 790 |
+
"rewards/chosen": 0.014059506356716156,
|
| 791 |
+
"rewards/margins": 0.006549051962792873,
|
| 792 |
+
"rewards/rejected": 0.007510454393923283,
|
| 793 |
+
"step": 550
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.43,
|
| 797 |
+
"learning_rate": 3.5451379808006014e-06,
|
| 798 |
+
"logits/chosen": -0.2925412654876709,
|
| 799 |
+
"logits/rejected": -0.27184224128723145,
|
| 800 |
+
"logps/chosen": -226.45083618164062,
|
| 801 |
+
"logps/rejected": -252.47207641601562,
|
| 802 |
+
"loss": 0.2026,
|
| 803 |
+
"rewards/accuracies": 0.26249998807907104,
|
| 804 |
+
"rewards/chosen": 0.006832784973084927,
|
| 805 |
+
"rewards/margins": 0.006407728884369135,
|
| 806 |
+
"rewards/rejected": 0.00042505591409280896,
|
| 807 |
+
"step": 560
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.43,
|
| 811 |
+
"learning_rate": 3.484310951897323e-06,
|
| 812 |
+
"logits/chosen": -0.280683308839798,
|
| 813 |
+
"logits/rejected": -0.2360977679491043,
|
| 814 |
+
"logps/chosen": -272.83953857421875,
|
| 815 |
+
"logps/rejected": -299.12847900390625,
|
| 816 |
+
"loss": 0.1933,
|
| 817 |
+
"rewards/accuracies": 0.32499998807907104,
|
| 818 |
+
"rewards/chosen": 0.0044427914544939995,
|
| 819 |
+
"rewards/margins": 0.007019062992185354,
|
| 820 |
+
"rewards/rejected": -0.0025762729346752167,
|
| 821 |
+
"step": 570
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.44,
|
| 825 |
+
"learning_rate": 3.4227862652892106e-06,
|
| 826 |
+
"logits/chosen": -0.23029498755931854,
|
| 827 |
+
"logits/rejected": -0.30575060844421387,
|
| 828 |
+
"logps/chosen": -284.2434997558594,
|
| 829 |
+
"logps/rejected": -325.15606689453125,
|
| 830 |
+
"loss": 0.2113,
|
| 831 |
+
"rewards/accuracies": 0.30000001192092896,
|
| 832 |
+
"rewards/chosen": -0.0008473257767036557,
|
| 833 |
+
"rewards/margins": 0.0063803717494010925,
|
| 834 |
+
"rewards/rejected": -0.007227697875350714,
|
| 835 |
+
"step": 580
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.45,
|
| 839 |
+
"learning_rate": 3.3606075283054005e-06,
|
| 840 |
+
"logits/chosen": -0.21216976642608643,
|
| 841 |
+
"logits/rejected": -0.24321305751800537,
|
| 842 |
+
"logps/chosen": -289.8013610839844,
|
| 843 |
+
"logps/rejected": -292.233642578125,
|
| 844 |
+
"loss": 0.2076,
|
| 845 |
+
"rewards/accuracies": 0.2750000059604645,
|
| 846 |
+
"rewards/chosen": -0.006494040135294199,
|
| 847 |
+
"rewards/margins": 0.0029291484970599413,
|
| 848 |
+
"rewards/rejected": -0.009423188865184784,
|
| 849 |
+
"step": 590
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.46,
|
| 853 |
+
"learning_rate": 3.2978188118513814e-06,
|
| 854 |
+
"logits/chosen": -0.3345397114753723,
|
| 855 |
+
"logits/rejected": -0.2698872983455658,
|
| 856 |
+
"logps/chosen": -288.9801025390625,
|
| 857 |
+
"logps/rejected": -325.5001525878906,
|
| 858 |
+
"loss": 0.2204,
|
| 859 |
+
"rewards/accuracies": 0.3375000059604645,
|
| 860 |
+
"rewards/chosen": -0.005164691712707281,
|
| 861 |
+
"rewards/margins": 0.006596796214580536,
|
| 862 |
+
"rewards/rejected": -0.011761486530303955,
|
| 863 |
+
"step": 600
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.46,
|
| 867 |
+
"learning_rate": 3.234464619172522e-06,
|
| 868 |
+
"logits/chosen": -0.28216275572776794,
|
| 869 |
+
"logits/rejected": -0.36517441272735596,
|
| 870 |
+
"logps/chosen": -275.3839416503906,
|
| 871 |
+
"logps/rejected": -297.3681640625,
|
| 872 |
+
"loss": 0.2018,
|
| 873 |
+
"rewards/accuracies": 0.30000001192092896,
|
| 874 |
+
"rewards/chosen": -0.007699407637119293,
|
| 875 |
+
"rewards/margins": 0.0056534623727202415,
|
| 876 |
+
"rewards/rejected": -0.01335287094116211,
|
| 877 |
+
"step": 610
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.47,
|
| 881 |
+
"learning_rate": 3.1705898543111576e-06,
|
| 882 |
+
"logits/chosen": -0.32491374015808105,
|
| 883 |
+
"logits/rejected": -0.31058216094970703,
|
| 884 |
+
"logps/chosen": -318.451904296875,
|
| 885 |
+
"logps/rejected": -336.5116271972656,
|
| 886 |
+
"loss": 0.2052,
|
| 887 |
+
"rewards/accuracies": 0.32499998807907104,
|
| 888 |
+
"rewards/chosen": -0.007857877761125565,
|
| 889 |
+
"rewards/margins": 0.005408051423728466,
|
| 890 |
+
"rewards/rejected": -0.013265928253531456,
|
| 891 |
+
"step": 620
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.48,
|
| 895 |
+
"learning_rate": 3.106239790279606e-06,
|
| 896 |
+
"logits/chosen": -0.3316212594509125,
|
| 897 |
+
"logits/rejected": -0.27467650175094604,
|
| 898 |
+
"logps/chosen": -306.818359375,
|
| 899 |
+
"logps/rejected": -310.139892578125,
|
| 900 |
+
"loss": 0.2198,
|
| 901 |
+
"rewards/accuracies": 0.30000001192092896,
|
| 902 |
+
"rewards/chosen": -0.008624967187643051,
|
| 903 |
+
"rewards/margins": 0.0013632235350087285,
|
| 904 |
+
"rewards/rejected": -0.009988190606236458,
|
| 905 |
+
"step": 630
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.49,
|
| 909 |
+
"learning_rate": 3.041460036971664e-06,
|
| 910 |
+
"logits/chosen": -0.30510929226875305,
|
| 911 |
+
"logits/rejected": -0.2651751637458801,
|
| 912 |
+
"logps/chosen": -289.76177978515625,
|
| 913 |
+
"logps/rejected": -306.5693664550781,
|
| 914 |
+
"loss": 0.2091,
|
| 915 |
+
"rewards/accuracies": 0.3187499940395355,
|
| 916 |
+
"rewards/chosen": -0.0034012228716164827,
|
| 917 |
+
"rewards/margins": 0.005689640529453754,
|
| 918 |
+
"rewards/rejected": -0.009090864099562168,
|
| 919 |
+
"step": 640
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.5,
|
| 923 |
+
"learning_rate": 2.976296508835326e-06,
|
| 924 |
+
"logits/chosen": -0.27622857689857483,
|
| 925 |
+
"logits/rejected": -0.3029255270957947,
|
| 926 |
+
"logps/chosen": -254.8801727294922,
|
| 927 |
+
"logps/rejected": -280.85552978515625,
|
| 928 |
+
"loss": 0.2118,
|
| 929 |
+
"rewards/accuracies": 0.26875001192092896,
|
| 930 |
+
"rewards/chosen": 0.003135355655103922,
|
| 931 |
+
"rewards/margins": 0.00885056797415018,
|
| 932 |
+
"rewards/rejected": -0.005715211853384972,
|
| 933 |
+
"step": 650
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.5,
|
| 937 |
+
"learning_rate": 2.910795392329649e-06,
|
| 938 |
+
"logits/chosen": -0.29605865478515625,
|
| 939 |
+
"logits/rejected": -0.27770885825157166,
|
| 940 |
+
"logps/chosen": -277.80804443359375,
|
| 941 |
+
"logps/rejected": -309.5155029296875,
|
| 942 |
+
"loss": 0.2133,
|
| 943 |
+
"rewards/accuracies": 0.3687500059604645,
|
| 944 |
+
"rewards/chosen": 0.0010512445587664843,
|
| 945 |
+
"rewards/margins": 0.008048586547374725,
|
| 946 |
+
"rewards/rejected": -0.006997342221438885,
|
| 947 |
+
"step": 660
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.51,
|
| 951 |
+
"learning_rate": 2.8450031131888147e-06,
|
| 952 |
+
"logits/chosen": -0.35795870423316956,
|
| 953 |
+
"logits/rejected": -0.37449702620506287,
|
| 954 |
+
"logps/chosen": -289.4441833496094,
|
| 955 |
+
"logps/rejected": -332.01055908203125,
|
| 956 |
+
"loss": 0.1977,
|
| 957 |
+
"rewards/accuracies": 0.33125001192092896,
|
| 958 |
+
"rewards/chosen": 0.004084504209458828,
|
| 959 |
+
"rewards/margins": 0.011728955432772636,
|
| 960 |
+
"rewards/rejected": -0.0076444498263299465,
|
| 961 |
+
"step": 670
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.52,
|
| 965 |
+
"learning_rate": 2.7789663035166035e-06,
|
| 966 |
+
"logits/chosen": -0.35911503434181213,
|
| 967 |
+
"logits/rejected": -0.3697941303253174,
|
| 968 |
+
"logps/chosen": -251.0248565673828,
|
| 969 |
+
"logps/rejected": -283.61279296875,
|
| 970 |
+
"loss": 0.2036,
|
| 971 |
+
"rewards/accuracies": 0.32499998807907104,
|
| 972 |
+
"rewards/chosen": 0.0005092100473120809,
|
| 973 |
+
"rewards/margins": 0.011741384863853455,
|
| 974 |
+
"rewards/rejected": -0.011232174932956696,
|
| 975 |
+
"step": 680
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.53,
|
| 979 |
+
"learning_rate": 2.7127317687345973e-06,
|
| 980 |
+
"logits/chosen": -0.4329107403755188,
|
| 981 |
+
"logits/rejected": -0.38218554854393005,
|
| 982 |
+
"logps/chosen": -275.99176025390625,
|
| 983 |
+
"logps/rejected": -282.9251708984375,
|
| 984 |
+
"loss": 0.2255,
|
| 985 |
+
"rewards/accuracies": 0.29374998807907104,
|
| 986 |
+
"rewards/chosen": -0.0015629607951268554,
|
| 987 |
+
"rewards/margins": 0.005656304769217968,
|
| 988 |
+
"rewards/rejected": -0.007219265215098858,
|
| 989 |
+
"step": 690
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.53,
|
| 993 |
+
"learning_rate": 2.6463464544075344e-06,
|
| 994 |
+
"logits/chosen": -0.38624441623687744,
|
| 995 |
+
"logits/rejected": -0.4270317554473877,
|
| 996 |
+
"logps/chosen": -313.55047607421875,
|
| 997 |
+
"logps/rejected": -360.2120361328125,
|
| 998 |
+
"loss": 0.1998,
|
| 999 |
+
"rewards/accuracies": 0.33125001192092896,
|
| 1000 |
+
"rewards/chosen": -0.0040723588317632675,
|
| 1001 |
+
"rewards/margins": 0.013303135521709919,
|
| 1002 |
+
"rewards/rejected": -0.01737549528479576,
|
| 1003 |
+
"step": 700
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.54,
|
| 1007 |
+
"learning_rate": 2.579857412969345e-06,
|
| 1008 |
+
"logits/chosen": -0.4271603226661682,
|
| 1009 |
+
"logits/rejected": -0.4234057366847992,
|
| 1010 |
+
"logps/chosen": -304.10211181640625,
|
| 1011 |
+
"logps/rejected": -354.1361389160156,
|
| 1012 |
+
"loss": 0.1979,
|
| 1013 |
+
"rewards/accuracies": 0.36250001192092896,
|
| 1014 |
+
"rewards/chosen": -0.007728138472884893,
|
| 1015 |
+
"rewards/margins": 0.016483409330248833,
|
| 1016 |
+
"rewards/rejected": -0.024211544543504715,
|
| 1017 |
+
"step": 710
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.55,
|
| 1021 |
+
"learning_rate": 2.513311770373421e-06,
|
| 1022 |
+
"logits/chosen": -0.5024863481521606,
|
| 1023 |
+
"logits/rejected": -0.47420111298561096,
|
| 1024 |
+
"logps/chosen": -317.60174560546875,
|
| 1025 |
+
"logps/rejected": -362.4761657714844,
|
| 1026 |
+
"loss": 0.192,
|
| 1027 |
+
"rewards/accuracies": 0.36250001192092896,
|
| 1028 |
+
"rewards/chosen": -0.011080259457230568,
|
| 1029 |
+
"rewards/margins": 0.010900650173425674,
|
| 1030 |
+
"rewards/rejected": -0.021980909630656242,
|
| 1031 |
+
"step": 720
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.56,
|
| 1035 |
+
"learning_rate": 2.446756692690804e-06,
|
| 1036 |
+
"logits/chosen": -0.4976615905761719,
|
| 1037 |
+
"logits/rejected": -0.4921053349971771,
|
| 1038 |
+
"logps/chosen": -306.67547607421875,
|
| 1039 |
+
"logps/rejected": -361.6275939941406,
|
| 1040 |
+
"loss": 0.2069,
|
| 1041 |
+
"rewards/accuracies": 0.36250001192092896,
|
| 1042 |
+
"rewards/chosen": -0.006612093187868595,
|
| 1043 |
+
"rewards/margins": 0.014301796443760395,
|
| 1044 |
+
"rewards/rejected": -0.02091388963162899,
|
| 1045 |
+
"step": 730
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.56,
|
| 1049 |
+
"learning_rate": 2.380239352679908e-06,
|
| 1050 |
+
"logits/chosen": -0.45196834206581116,
|
| 1051 |
+
"logits/rejected": -0.48185840249061584,
|
| 1052 |
+
"logps/chosen": -313.2336730957031,
|
| 1053 |
+
"logps/rejected": -313.48260498046875,
|
| 1054 |
+
"loss": 0.2149,
|
| 1055 |
+
"rewards/accuracies": 0.2874999940395355,
|
| 1056 |
+
"rewards/chosen": -0.010375128127634525,
|
| 1057 |
+
"rewards/margins": 0.004926495254039764,
|
| 1058 |
+
"rewards/rejected": -0.015301624312996864,
|
| 1059 |
+
"step": 740
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.57,
|
| 1063 |
+
"learning_rate": 2.313806896351529e-06,
|
| 1064 |
+
"logits/chosen": -0.5004686117172241,
|
| 1065 |
+
"logits/rejected": -0.47765883803367615,
|
| 1066 |
+
"logps/chosen": -274.7811279296875,
|
| 1067 |
+
"logps/rejected": -311.44207763671875,
|
| 1068 |
+
"loss": 0.1904,
|
| 1069 |
+
"rewards/accuracies": 0.33125001192092896,
|
| 1070 |
+
"rewards/chosen": 0.002466167788952589,
|
| 1071 |
+
"rewards/margins": 0.012883700430393219,
|
| 1072 |
+
"rewards/rejected": -0.010417533107101917,
|
| 1073 |
+
"step": 750
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.58,
|
| 1077 |
+
"learning_rate": 2.247506409552795e-06,
|
| 1078 |
+
"logits/chosen": -0.563784122467041,
|
| 1079 |
+
"logits/rejected": -0.5083848237991333,
|
| 1080 |
+
"logps/chosen": -286.28277587890625,
|
| 1081 |
+
"logps/rejected": -303.66925048828125,
|
| 1082 |
+
"loss": 0.2037,
|
| 1083 |
+
"rewards/accuracies": 0.30000001192092896,
|
| 1084 |
+
"rewards/chosen": -0.0052384668961167336,
|
| 1085 |
+
"rewards/margins": 0.004287950228899717,
|
| 1086 |
+
"rewards/rejected": -0.009526416659355164,
|
| 1087 |
+
"step": 760
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.59,
|
| 1091 |
+
"learning_rate": 2.1813848845937695e-06,
|
| 1092 |
+
"logits/chosen": -0.5057908296585083,
|
| 1093 |
+
"logits/rejected": -0.48972076177597046,
|
| 1094 |
+
"logps/chosen": -297.547607421875,
|
| 1095 |
+
"logps/rejected": -294.28863525390625,
|
| 1096 |
+
"loss": 0.2101,
|
| 1097 |
+
"rewards/accuracies": 0.28125,
|
| 1098 |
+
"rewards/chosen": -0.0030983686447143555,
|
| 1099 |
+
"rewards/margins": 0.0027166479267179966,
|
| 1100 |
+
"rewards/rejected": -0.00581501517444849,
|
| 1101 |
+
"step": 770
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.59,
|
| 1105 |
+
"learning_rate": 2.1154891869403436e-06,
|
| 1106 |
+
"logits/chosen": -0.472622811794281,
|
| 1107 |
+
"logits/rejected": -0.4686339497566223,
|
| 1108 |
+
"logps/chosen": -241.92630004882812,
|
| 1109 |
+
"logps/rejected": -279.33062744140625,
|
| 1110 |
+
"loss": 0.2091,
|
| 1111 |
+
"rewards/accuracies": 0.3125,
|
| 1112 |
+
"rewards/chosen": 0.009422892704606056,
|
| 1113 |
+
"rewards/margins": 0.009348099119961262,
|
| 1114 |
+
"rewards/rejected": 7.479395571863279e-05,
|
| 1115 |
+
"step": 780
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.6,
|
| 1119 |
+
"learning_rate": 2.0498660219970395e-06,
|
| 1120 |
+
"logits/chosen": -0.44067639112472534,
|
| 1121 |
+
"logits/rejected": -0.46692126989364624,
|
| 1122 |
+
"logps/chosen": -285.769287109375,
|
| 1123 |
+
"logps/rejected": -305.04296875,
|
| 1124 |
+
"loss": 0.2271,
|
| 1125 |
+
"rewards/accuracies": 0.3187499940395355,
|
| 1126 |
+
"rewards/chosen": 0.008178983815014362,
|
| 1127 |
+
"rewards/margins": 0.007535718381404877,
|
| 1128 |
+
"rewards/rejected": 0.0006432650843635201,
|
| 1129 |
+
"step": 790
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.61,
|
| 1133 |
+
"learning_rate": 1.9845619020032552e-06,
|
| 1134 |
+
"logits/chosen": -0.4349413514137268,
|
| 1135 |
+
"logits/rejected": -0.4484075605869293,
|
| 1136 |
+
"logps/chosen": -305.90972900390625,
|
| 1137 |
+
"logps/rejected": -287.2285461425781,
|
| 1138 |
+
"loss": 0.2105,
|
| 1139 |
+
"rewards/accuracies": 0.3062500059604645,
|
| 1140 |
+
"rewards/chosen": 0.006356480531394482,
|
| 1141 |
+
"rewards/margins": 0.0014576372923329473,
|
| 1142 |
+
"rewards/rejected": 0.004898843355476856,
|
| 1143 |
+
"step": 800
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.62,
|
| 1147 |
+
"learning_rate": 1.9196231130664282e-06,
|
| 1148 |
+
"logits/chosen": -0.45880216360092163,
|
| 1149 |
+
"logits/rejected": -0.44850587844848633,
|
| 1150 |
+
"logps/chosen": -263.9127502441406,
|
| 1151 |
+
"logps/rejected": -292.25177001953125,
|
| 1152 |
+
"loss": 0.2033,
|
| 1153 |
+
"rewards/accuracies": 0.3125,
|
| 1154 |
+
"rewards/chosen": 0.0159307774156332,
|
| 1155 |
+
"rewards/margins": 0.009099806658923626,
|
| 1156 |
+
"rewards/rejected": 0.006830970756709576,
|
| 1157 |
+
"step": 810
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.62,
|
| 1161 |
+
"learning_rate": 1.8550956823554708e-06,
|
| 1162 |
+
"logits/chosen": -0.4190438687801361,
|
| 1163 |
+
"logits/rejected": -0.39324015378952026,
|
| 1164 |
+
"logps/chosen": -297.383056640625,
|
| 1165 |
+
"logps/rejected": -307.9434814453125,
|
| 1166 |
+
"loss": 0.2115,
|
| 1167 |
+
"rewards/accuracies": 0.2750000059604645,
|
| 1168 |
+
"rewards/chosen": 0.008881737478077412,
|
| 1169 |
+
"rewards/margins": 0.0028643650002777576,
|
| 1170 |
+
"rewards/rejected": 0.006017371080815792,
|
| 1171 |
+
"step": 820
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.63,
|
| 1175 |
+
"learning_rate": 1.7910253454777346e-06,
|
| 1176 |
+
"logits/chosen": -0.4339013993740082,
|
| 1177 |
+
"logits/rejected": -0.4780551791191101,
|
| 1178 |
+
"logps/chosen": -264.0601501464844,
|
| 1179 |
+
"logps/rejected": -309.7989807128906,
|
| 1180 |
+
"loss": 0.2071,
|
| 1181 |
+
"rewards/accuracies": 0.3687500059604645,
|
| 1182 |
+
"rewards/chosen": 0.01502480823546648,
|
| 1183 |
+
"rewards/margins": 0.013155427761375904,
|
| 1184 |
+
"rewards/rejected": 0.0018693817546591163,
|
| 1185 |
+
"step": 830
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.64,
|
| 1189 |
+
"learning_rate": 1.7274575140626318e-06,
|
| 1190 |
+
"logits/chosen": -0.413571834564209,
|
| 1191 |
+
"logits/rejected": -0.4285905361175537,
|
| 1192 |
+
"logps/chosen": -263.08453369140625,
|
| 1193 |
+
"logps/rejected": -286.2958068847656,
|
| 1194 |
+
"loss": 0.2184,
|
| 1195 |
+
"rewards/accuracies": 0.25,
|
| 1196 |
+
"rewards/chosen": 0.008341384120285511,
|
| 1197 |
+
"rewards/margins": 0.0012967393267899752,
|
| 1198 |
+
"rewards/rejected": 0.007044644560664892,
|
| 1199 |
+
"step": 840
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.65,
|
| 1203 |
+
"learning_rate": 1.6644372435748823e-06,
|
| 1204 |
+
"logits/chosen": -0.408627986907959,
|
| 1205 |
+
"logits/rejected": -0.4011750817298889,
|
| 1206 |
+
"logps/chosen": -229.0798797607422,
|
| 1207 |
+
"logps/rejected": -262.96075439453125,
|
| 1208 |
+
"loss": 0.2043,
|
| 1209 |
+
"rewards/accuracies": 0.2750000059604645,
|
| 1210 |
+
"rewards/chosen": 0.0140065373852849,
|
| 1211 |
+
"rewards/margins": 0.007881510071456432,
|
| 1212 |
+
"rewards/rejected": 0.006125027313828468,
|
| 1213 |
+
"step": 850
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.66,
|
| 1217 |
+
"learning_rate": 1.6020092013802002e-06,
|
| 1218 |
+
"logits/chosen": -0.43513408303260803,
|
| 1219 |
+
"logits/rejected": -0.4182719588279724,
|
| 1220 |
+
"logps/chosen": -283.6298828125,
|
| 1221 |
+
"logps/rejected": -299.7771301269531,
|
| 1222 |
+
"loss": 0.2094,
|
| 1223 |
+
"rewards/accuracies": 0.3125,
|
| 1224 |
+
"rewards/chosen": 0.015776053071022034,
|
| 1225 |
+
"rewards/margins": 0.005827675573527813,
|
| 1226 |
+
"rewards/rejected": 0.00994837749749422,
|
| 1227 |
+
"step": 860
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.66,
|
| 1231 |
+
"learning_rate": 1.5402176350860653e-06,
|
| 1232 |
+
"logits/chosen": -0.4543350338935852,
|
| 1233 |
+
"logits/rejected": -0.43857789039611816,
|
| 1234 |
+
"logps/chosen": -258.48822021484375,
|
| 1235 |
+
"logps/rejected": -309.0599365234375,
|
| 1236 |
+
"loss": 0.2144,
|
| 1237 |
+
"rewards/accuracies": 0.36250001192092896,
|
| 1238 |
+
"rewards/chosen": 0.01519864983856678,
|
| 1239 |
+
"rewards/margins": 0.0170084610581398,
|
| 1240 |
+
"rewards/rejected": -0.0018098097061738372,
|
| 1241 |
+
"step": 870
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.67,
|
| 1245 |
+
"learning_rate": 1.4791063411799938e-06,
|
| 1246 |
+
"logits/chosen": -0.4367285370826721,
|
| 1247 |
+
"logits/rejected": -0.4124454855918884,
|
| 1248 |
+
"logps/chosen": -258.4836120605469,
|
| 1249 |
+
"logps/rejected": -256.78948974609375,
|
| 1250 |
+
"loss": 0.2125,
|
| 1251 |
+
"rewards/accuracies": 0.26249998807907104,
|
| 1252 |
+
"rewards/chosen": 0.010697437450289726,
|
| 1253 |
+
"rewards/margins": 0.0008281940827146173,
|
| 1254 |
+
"rewards/rejected": 0.009869244880974293,
|
| 1255 |
+
"step": 880
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.68,
|
| 1259 |
+
"learning_rate": 1.4187186339875697e-06,
|
| 1260 |
+
"logits/chosen": -0.3717727065086365,
|
| 1261 |
+
"logits/rejected": -0.3890214264392853,
|
| 1262 |
+
"logps/chosen": -238.35391235351562,
|
| 1263 |
+
"logps/rejected": -270.5140686035156,
|
| 1264 |
+
"loss": 0.202,
|
| 1265 |
+
"rewards/accuracies": 0.26875001192092896,
|
| 1266 |
+
"rewards/chosen": 0.015574706718325615,
|
| 1267 |
+
"rewards/margins": 0.009048102423548698,
|
| 1268 |
+
"rewards/rejected": 0.006526602897793055,
|
| 1269 |
+
"step": 890
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.69,
|
| 1273 |
+
"learning_rate": 1.3590973149722103e-06,
|
| 1274 |
+
"logits/chosen": -0.4208298325538635,
|
| 1275 |
+
"logits/rejected": -0.4319419860839844,
|
| 1276 |
+
"logps/chosen": -272.0270690917969,
|
| 1277 |
+
"logps/rejected": -303.1417541503906,
|
| 1278 |
+
"loss": 0.2083,
|
| 1279 |
+
"rewards/accuracies": 0.32499998807907104,
|
| 1280 |
+
"rewards/chosen": 0.015811121091246605,
|
| 1281 |
+
"rewards/margins": 0.010845163837075233,
|
| 1282 |
+
"rewards/rejected": 0.004965959116816521,
|
| 1283 |
+
"step": 900
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.69,
|
| 1287 |
+
"learning_rate": 1.300284642398445e-06,
|
| 1288 |
+
"logits/chosen": -0.47379741072654724,
|
| 1289 |
+
"logits/rejected": -0.4520508348941803,
|
| 1290 |
+
"logps/chosen": -262.06292724609375,
|
| 1291 |
+
"logps/rejected": -311.0146789550781,
|
| 1292 |
+
"loss": 0.1942,
|
| 1293 |
+
"rewards/accuracies": 0.3499999940395355,
|
| 1294 |
+
"rewards/chosen": 0.014805925078690052,
|
| 1295 |
+
"rewards/margins": 0.01618744060397148,
|
| 1296 |
+
"rewards/rejected": -0.0013815152924507856,
|
| 1297 |
+
"step": 910
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.7,
|
| 1301 |
+
"learning_rate": 1.2423223013801946e-06,
|
| 1302 |
+
"logits/chosen": -0.45992302894592285,
|
| 1303 |
+
"logits/rejected": -0.4493522644042969,
|
| 1304 |
+
"logps/chosen": -277.81231689453125,
|
| 1305 |
+
"logps/rejected": -303.0189514160156,
|
| 1306 |
+
"loss": 0.2059,
|
| 1307 |
+
"rewards/accuracies": 0.30000001192092896,
|
| 1308 |
+
"rewards/chosen": 0.007296368479728699,
|
| 1309 |
+
"rewards/margins": 0.007703060749918222,
|
| 1310 |
+
"rewards/rejected": -0.00040669291047379375,
|
| 1311 |
+
"step": 920
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.71,
|
| 1315 |
+
"learning_rate": 1.1852513743352886e-06,
|
| 1316 |
+
"logits/chosen": -0.4434199929237366,
|
| 1317 |
+
"logits/rejected": -0.4254834055900574,
|
| 1318 |
+
"logps/chosen": -236.1838836669922,
|
| 1319 |
+
"logps/rejected": -280.4830627441406,
|
| 1320 |
+
"loss": 0.2005,
|
| 1321 |
+
"rewards/accuracies": 0.2874999940395355,
|
| 1322 |
+
"rewards/chosen": 0.00997086614370346,
|
| 1323 |
+
"rewards/margins": 0.012648127973079681,
|
| 1324 |
+
"rewards/rejected": -0.0026772604323923588,
|
| 1325 |
+
"step": 930
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 0.72,
|
| 1329 |
+
"learning_rate": 1.1291123118671665e-06,
|
| 1330 |
+
"logits/chosen": -0.4678889811038971,
|
| 1331 |
+
"logits/rejected": -0.4497453570365906,
|
| 1332 |
+
"logps/chosen": -260.39630126953125,
|
| 1333 |
+
"logps/rejected": -310.59783935546875,
|
| 1334 |
+
"loss": 0.2033,
|
| 1335 |
+
"rewards/accuracies": 0.29374998807907104,
|
| 1336 |
+
"rewards/chosen": 0.013568739406764507,
|
| 1337 |
+
"rewards/margins": 0.013490801677107811,
|
| 1338 |
+
"rewards/rejected": 7.793791883159429e-05,
|
| 1339 |
+
"step": 940
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.72,
|
| 1343 |
+
"learning_rate": 1.073944904094385e-06,
|
| 1344 |
+
"logits/chosen": -0.4590323865413666,
|
| 1345 |
+
"logits/rejected": -0.4857376217842102,
|
| 1346 |
+
"logps/chosen": -250.05941772460938,
|
| 1347 |
+
"logps/rejected": -269.81182861328125,
|
| 1348 |
+
"loss": 0.2081,
|
| 1349 |
+
"rewards/accuracies": 0.26875001192092896,
|
| 1350 |
+
"rewards/chosen": 0.009169178083539009,
|
| 1351 |
+
"rewards/margins": 0.0009465640177950263,
|
| 1352 |
+
"rewards/rejected": 0.008222613483667374,
|
| 1353 |
+
"step": 950
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 0.73,
|
| 1357 |
+
"learning_rate": 1.019788252448267e-06,
|
| 1358 |
+
"logits/chosen": -0.42771902680397034,
|
| 1359 |
+
"logits/rejected": -0.4784491956233978,
|
| 1360 |
+
"logps/chosen": -286.91033935546875,
|
| 1361 |
+
"logps/rejected": -356.3207092285156,
|
| 1362 |
+
"loss": 0.2176,
|
| 1363 |
+
"rewards/accuracies": 0.36250001192092896,
|
| 1364 |
+
"rewards/chosen": 0.009989687241613865,
|
| 1365 |
+
"rewards/margins": 0.01874762959778309,
|
| 1366 |
+
"rewards/rejected": -0.008757943287491798,
|
| 1367 |
+
"step": 960
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.74,
|
| 1371 |
+
"learning_rate": 9.66680741958685e-07,
|
| 1372 |
+
"logits/chosen": -0.5021045804023743,
|
| 1373 |
+
"logits/rejected": -0.4944073557853699,
|
| 1374 |
+
"logps/chosen": -256.72845458984375,
|
| 1375 |
+
"logps/rejected": -282.7173767089844,
|
| 1376 |
+
"loss": 0.2023,
|
| 1377 |
+
"rewards/accuracies": 0.33125001192092896,
|
| 1378 |
+
"rewards/chosen": 0.008730728179216385,
|
| 1379 |
+
"rewards/margins": 0.010655781254172325,
|
| 1380 |
+
"rewards/rejected": -0.001925053307786584,
|
| 1381 |
+
"step": 970
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 0.75,
|
| 1385 |
+
"learning_rate": 9.146600140475945e-07,
|
| 1386 |
+
"logits/chosen": -0.49039751291275024,
|
| 1387 |
+
"logits/rejected": -0.4602015018463135,
|
| 1388 |
+
"logps/chosen": -319.5450134277344,
|
| 1389 |
+
"logps/rejected": -356.46478271484375,
|
| 1390 |
+
"loss": 0.2141,
|
| 1391 |
+
"rewards/accuracies": 0.3375000059604645,
|
| 1392 |
+
"rewards/chosen": 0.0025001950562000275,
|
| 1393 |
+
"rewards/margins": 0.0142419608309865,
|
| 1394 |
+
"rewards/rejected": -0.011741766706109047,
|
| 1395 |
+
"step": 980
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 0.75,
|
| 1399 |
+
"learning_rate": 8.637629398496378e-07,
|
| 1400 |
+
"logits/chosen": -0.4612942636013031,
|
| 1401 |
+
"logits/rejected": -0.4650956988334656,
|
| 1402 |
+
"logps/chosen": -314.468017578125,
|
| 1403 |
+
"logps/rejected": -336.6165466308594,
|
| 1404 |
+
"loss": 0.2201,
|
| 1405 |
+
"rewards/accuracies": 0.3125,
|
| 1406 |
+
"rewards/chosen": 0.004640791565179825,
|
| 1407 |
+
"rewards/margins": 0.011841602623462677,
|
| 1408 |
+
"rewards/rejected": -0.0072008101269602776,
|
| 1409 |
+
"step": 990
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 0.76,
|
| 1413 |
+
"learning_rate": 8.140255940787059e-07,
|
| 1414 |
+
"logits/chosen": -0.4051927924156189,
|
| 1415 |
+
"logits/rejected": -0.41572269797325134,
|
| 1416 |
+
"logps/chosen": -266.1190185546875,
|
| 1417 |
+
"logps/rejected": -307.9905090332031,
|
| 1418 |
+
"loss": 0.1983,
|
| 1419 |
+
"rewards/accuracies": 0.3062500059604645,
|
| 1420 |
+
"rewards/chosen": 0.005232260562479496,
|
| 1421 |
+
"rewards/margins": 0.011333976872265339,
|
| 1422 |
+
"rewards/rejected": -0.006101716309785843,
|
| 1423 |
+
"step": 1000
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 0.77,
|
| 1427 |
+
"learning_rate": 7.654832294589776e-07,
|
| 1428 |
+
"logits/chosen": -0.433448851108551,
|
| 1429 |
+
"logits/rejected": -0.4530878961086273,
|
| 1430 |
+
"logps/chosen": -264.8626403808594,
|
| 1431 |
+
"logps/rejected": -305.2262878417969,
|
| 1432 |
+
"loss": 0.2114,
|
| 1433 |
+
"rewards/accuracies": 0.3187499940395355,
|
| 1434 |
+
"rewards/chosen": 0.004321249667555094,
|
| 1435 |
+
"rewards/margins": 0.009284446947276592,
|
| 1436 |
+
"rewards/rejected": -0.004963197745382786,
|
| 1437 |
+
"step": 1010
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 0.78,
|
| 1441 |
+
"learning_rate": 7.181702517385789e-07,
|
| 1442 |
+
"logits/chosen": -0.470588743686676,
|
| 1443 |
+
"logits/rejected": -0.4782032370567322,
|
| 1444 |
+
"logps/chosen": -274.71527099609375,
|
| 1445 |
+
"logps/rejected": -307.9129943847656,
|
| 1446 |
+
"loss": 0.1997,
|
| 1447 |
+
"rewards/accuracies": 0.3125,
|
| 1448 |
+
"rewards/chosen": 0.007287648506462574,
|
| 1449 |
+
"rewards/margins": 0.00972357951104641,
|
| 1450 |
+
"rewards/rejected": -0.0024359312374144793,
|
| 1451 |
+
"step": 1020
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 0.78,
|
| 1455 |
+
"learning_rate": 6.721201953035511e-07,
|
| 1456 |
+
"logits/chosen": -0.4322214126586914,
|
| 1457 |
+
"logits/rejected": -0.458743155002594,
|
| 1458 |
+
"logps/chosen": -262.1369323730469,
|
| 1459 |
+
"logps/rejected": -275.8923034667969,
|
| 1460 |
+
"loss": 0.2108,
|
| 1461 |
+
"rewards/accuracies": 0.23749999701976776,
|
| 1462 |
+
"rewards/chosen": 0.0005865416023880243,
|
| 1463 |
+
"rewards/margins": 0.000941948383115232,
|
| 1464 |
+
"rewards/rejected": -0.00035540637327358127,
|
| 1465 |
+
"step": 1030
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 0.79,
|
| 1469 |
+
"learning_rate": 6.273656994094232e-07,
|
| 1470 |
+
"logits/chosen": -0.4811081886291504,
|
| 1471 |
+
"logits/rejected": -0.40350890159606934,
|
| 1472 |
+
"logps/chosen": -272.11419677734375,
|
| 1473 |
+
"logps/rejected": -307.12371826171875,
|
| 1474 |
+
"loss": 0.1991,
|
| 1475 |
+
"rewards/accuracies": 0.32499998807907104,
|
| 1476 |
+
"rewards/chosen": 0.005647097248584032,
|
| 1477 |
+
"rewards/margins": 0.007634321693331003,
|
| 1478 |
+
"rewards/rejected": -0.0019872249104082584,
|
| 1479 |
+
"step": 1040
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 0.8,
|
| 1483 |
+
"learning_rate": 5.839384850472359e-07,
|
| 1484 |
+
"logits/chosen": -0.46519404649734497,
|
| 1485 |
+
"logits/rejected": -0.40799570083618164,
|
| 1486 |
+
"logps/chosen": -289.98883056640625,
|
| 1487 |
+
"logps/rejected": -292.3999938964844,
|
| 1488 |
+
"loss": 0.2163,
|
| 1489 |
+
"rewards/accuracies": 0.29374998807907104,
|
| 1490 |
+
"rewards/chosen": 0.008130472153425217,
|
| 1491 |
+
"rewards/margins": 0.002877952065318823,
|
| 1492 |
+
"rewards/rejected": 0.005252520553767681,
|
| 1493 |
+
"step": 1050
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 0.81,
|
| 1497 |
+
"learning_rate": 5.418693324604082e-07,
|
| 1498 |
+
"logits/chosen": -0.5294805765151978,
|
| 1499 |
+
"logits/rejected": -0.5169610977172852,
|
| 1500 |
+
"logps/chosen": -272.34686279296875,
|
| 1501 |
+
"logps/rejected": -307.16412353515625,
|
| 1502 |
+
"loss": 0.1938,
|
| 1503 |
+
"rewards/accuracies": 0.35624998807907104,
|
| 1504 |
+
"rewards/chosen": 0.012028077617287636,
|
| 1505 |
+
"rewards/margins": 0.013656511902809143,
|
| 1506 |
+
"rewards/rejected": -0.0016284309094771743,
|
| 1507 |
+
"step": 1060
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 0.82,
|
| 1511 |
+
"learning_rate": 5.01188059328386e-07,
|
| 1512 |
+
"logits/chosen": -0.5478223562240601,
|
| 1513 |
+
"logits/rejected": -0.4814823269844055,
|
| 1514 |
+
"logps/chosen": -289.8566589355469,
|
| 1515 |
+
"logps/rejected": -289.57525634765625,
|
| 1516 |
+
"loss": 0.2065,
|
| 1517 |
+
"rewards/accuracies": 0.28125,
|
| 1518 |
+
"rewards/chosen": 0.008872238919138908,
|
| 1519 |
+
"rewards/margins": 0.006515635643154383,
|
| 1520 |
+
"rewards/rejected": 0.0023566028103232384,
|
| 1521 |
+
"step": 1070
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 0.82,
|
| 1525 |
+
"learning_rate": 4.619234996325314e-07,
|
| 1526 |
+
"logits/chosen": -0.4742591977119446,
|
| 1527 |
+
"logits/rejected": -0.45757928490638733,
|
| 1528 |
+
"logps/chosen": -328.6953125,
|
| 1529 |
+
"logps/rejected": -340.1205139160156,
|
| 1530 |
+
"loss": 0.2115,
|
| 1531 |
+
"rewards/accuracies": 0.32499998807907104,
|
| 1532 |
+
"rewards/chosen": 0.006806342396885157,
|
| 1533 |
+
"rewards/margins": 0.0063201361335814,
|
| 1534 |
+
"rewards/rejected": 0.000486206088680774,
|
| 1535 |
+
"step": 1080
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 0.83,
|
| 1539 |
+
"learning_rate": 4.241034832192434e-07,
|
| 1540 |
+
"logits/chosen": -0.5209555625915527,
|
| 1541 |
+
"logits/rejected": -0.5151633024215698,
|
| 1542 |
+
"logps/chosen": -276.89581298828125,
|
| 1543 |
+
"logps/rejected": -273.1256103515625,
|
| 1544 |
+
"loss": 0.2135,
|
| 1545 |
+
"rewards/accuracies": 0.24375000596046448,
|
| 1546 |
+
"rewards/chosen": 0.0027219385374337435,
|
| 1547 |
+
"rewards/margins": 0.0031283677089959383,
|
| 1548 |
+
"rewards/rejected": -0.00040642902604304254,
|
| 1549 |
+
"step": 1090
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 0.84,
|
| 1553 |
+
"learning_rate": 3.877548160747768e-07,
|
| 1554 |
+
"logits/chosen": -0.4666425585746765,
|
| 1555 |
+
"logits/rejected": -0.48001235723495483,
|
| 1556 |
+
"logps/chosen": -280.6470031738281,
|
| 1557 |
+
"logps/rejected": -315.48443603515625,
|
| 1558 |
+
"loss": 0.2048,
|
| 1559 |
+
"rewards/accuracies": 0.3187499940395355,
|
| 1560 |
+
"rewards/chosen": 0.002273663878440857,
|
| 1561 |
+
"rewards/margins": 0.008108812384307384,
|
| 1562 |
+
"rewards/rejected": -0.0058351485058665276,
|
| 1563 |
+
"step": 1100
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 0.85,
|
| 1567 |
+
"learning_rate": 3.529032613257574e-07,
|
| 1568 |
+
"logits/chosen": -0.4815247654914856,
|
| 1569 |
+
"logits/rejected": -0.45564061403274536,
|
| 1570 |
+
"logps/chosen": -300.01959228515625,
|
| 1571 |
+
"logps/rejected": -330.40899658203125,
|
| 1572 |
+
"loss": 0.2076,
|
| 1573 |
+
"rewards/accuracies": 0.32499998807907104,
|
| 1574 |
+
"rewards/chosen": 0.0031171857845038176,
|
| 1575 |
+
"rewards/margins": 0.009788629598915577,
|
| 1576 |
+
"rewards/rejected": -0.006671445909887552,
|
| 1577 |
+
"step": 1110
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 0.85,
|
| 1581 |
+
"learning_rate": 3.195735209788528e-07,
|
| 1582 |
+
"logits/chosen": -0.4933408796787262,
|
| 1583 |
+
"logits/rejected": -0.5491820573806763,
|
| 1584 |
+
"logps/chosen": -258.11822509765625,
|
| 1585 |
+
"logps/rejected": -296.2974853515625,
|
| 1586 |
+
"loss": 0.213,
|
| 1587 |
+
"rewards/accuracies": 0.3062500059604645,
|
| 1588 |
+
"rewards/chosen": 0.008377974852919579,
|
| 1589 |
+
"rewards/margins": 0.012130925431847572,
|
| 1590 |
+
"rewards/rejected": -0.003752949880436063,
|
| 1591 |
+
"step": 1120
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 0.86,
|
| 1595 |
+
"learning_rate": 2.8778921841253774e-07,
|
| 1596 |
+
"logits/chosen": -0.43251729011535645,
|
| 1597 |
+
"logits/rejected": -0.5054982900619507,
|
| 1598 |
+
"logps/chosen": -281.94403076171875,
|
| 1599 |
+
"logps/rejected": -280.0823669433594,
|
| 1600 |
+
"loss": 0.2204,
|
| 1601 |
+
"rewards/accuracies": 0.23125000298023224,
|
| 1602 |
+
"rewards/chosen": 0.0010522855445742607,
|
| 1603 |
+
"rewards/margins": 0.00011768531840061769,
|
| 1604 |
+
"rewards/rejected": 0.0009346003644168377,
|
| 1605 |
+
"step": 1130
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 0.87,
|
| 1609 |
+
"learning_rate": 2.5757288163336806e-07,
|
| 1610 |
+
"logits/chosen": -0.4674537181854248,
|
| 1611 |
+
"logits/rejected": -0.43040981888771057,
|
| 1612 |
+
"logps/chosen": -267.8955078125,
|
| 1613 |
+
"logps/rejected": -268.64312744140625,
|
| 1614 |
+
"loss": 0.2134,
|
| 1615 |
+
"rewards/accuracies": 0.2750000059604645,
|
| 1616 |
+
"rewards/chosen": -0.00031215575290843844,
|
| 1617 |
+
"rewards/margins": -0.00015194570005405694,
|
| 1618 |
+
"rewards/rejected": -0.00016021067858673632,
|
| 1619 |
+
"step": 1140
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 0.88,
|
| 1623 |
+
"learning_rate": 2.2894592730863336e-07,
|
| 1624 |
+
"logits/chosen": -0.4933416247367859,
|
| 1625 |
+
"logits/rejected": -0.4184231758117676,
|
| 1626 |
+
"logps/chosen": -274.43487548828125,
|
| 1627 |
+
"logps/rejected": -287.72259521484375,
|
| 1628 |
+
"loss": 0.2058,
|
| 1629 |
+
"rewards/accuracies": 0.29374998807907104,
|
| 1630 |
+
"rewards/chosen": 0.012380121275782585,
|
| 1631 |
+
"rewards/margins": 0.007005924824625254,
|
| 1632 |
+
"rewards/rejected": 0.005374195985496044,
|
| 1633 |
+
"step": 1150
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 0.88,
|
| 1637 |
+
"learning_rate": 2.019286455866981e-07,
|
| 1638 |
+
"logits/chosen": -0.5218170881271362,
|
| 1639 |
+
"logits/rejected": -0.46864786744117737,
|
| 1640 |
+
"logps/chosen": -298.1045227050781,
|
| 1641 |
+
"logps/rejected": -322.55181884765625,
|
| 1642 |
+
"loss": 0.1925,
|
| 1643 |
+
"rewards/accuracies": 0.32499998807907104,
|
| 1644 |
+
"rewards/chosen": -0.0008321186760440469,
|
| 1645 |
+
"rewards/margins": 0.004535870626568794,
|
| 1646 |
+
"rewards/rejected": -0.005367989186197519,
|
| 1647 |
+
"step": 1160
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.89,
|
| 1651 |
+
"learning_rate": 1.7654018571579557e-07,
|
| 1652 |
+
"logits/chosen": -0.48008307814598083,
|
| 1653 |
+
"logits/rejected": -0.4698426127433777,
|
| 1654 |
+
"logps/chosen": -307.2906494140625,
|
| 1655 |
+
"logps/rejected": -356.3848571777344,
|
| 1656 |
+
"loss": 0.2072,
|
| 1657 |
+
"rewards/accuracies": 0.3812499940395355,
|
| 1658 |
+
"rewards/chosen": 0.013308033347129822,
|
| 1659 |
+
"rewards/margins": 0.019972598180174828,
|
| 1660 |
+
"rewards/rejected": -0.00666456576436758,
|
| 1661 |
+
"step": 1170
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 0.9,
|
| 1665 |
+
"learning_rate": 1.5279854247146703e-07,
|
| 1666 |
+
"logits/chosen": -0.46893006563186646,
|
| 1667 |
+
"logits/rejected": -0.4585728645324707,
|
| 1668 |
+
"logps/chosen": -340.1523132324219,
|
| 1669 |
+
"logps/rejected": -351.4723205566406,
|
| 1670 |
+
"loss": 0.2135,
|
| 1671 |
+
"rewards/accuracies": 0.3187499940395355,
|
| 1672 |
+
"rewards/chosen": 0.0029886148404330015,
|
| 1673 |
+
"rewards/margins": 0.005356112495064735,
|
| 1674 |
+
"rewards/rejected": -0.002367498353123665,
|
| 1675 |
+
"step": 1180
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 0.91,
|
| 1679 |
+
"learning_rate": 1.307205434022671e-07,
|
| 1680 |
+
"logits/chosen": -0.4998309016227722,
|
| 1681 |
+
"logits/rejected": -0.49961596727371216,
|
| 1682 |
+
"logps/chosen": -272.09832763671875,
|
| 1683 |
+
"logps/rejected": -272.3662109375,
|
| 1684 |
+
"loss": 0.1976,
|
| 1685 |
+
"rewards/accuracies": 0.26875001192092896,
|
| 1686 |
+
"rewards/chosen": 0.0032546469010412693,
|
| 1687 |
+
"rewards/margins": 0.003977665212005377,
|
| 1688 |
+
"rewards/rejected": -0.0007230177288874984,
|
| 1689 |
+
"step": 1190
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 0.91,
|
| 1693 |
+
"learning_rate": 1.1032183690276754e-07,
|
| 1694 |
+
"logits/chosen": -0.518333911895752,
|
| 1695 |
+
"logits/rejected": -0.5002647638320923,
|
| 1696 |
+
"logps/chosen": -280.9651794433594,
|
| 1697 |
+
"logps/rejected": -300.6822204589844,
|
| 1698 |
+
"loss": 0.2098,
|
| 1699 |
+
"rewards/accuracies": 0.26875001192092896,
|
| 1700 |
+
"rewards/chosen": 0.002680544275790453,
|
| 1701 |
+
"rewards/margins": 0.003743491368368268,
|
| 1702 |
+
"rewards/rejected": -0.001062947092577815,
|
| 1703 |
+
"step": 1200
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 0.92,
|
| 1707 |
+
"learning_rate": 9.161688112232836e-08,
|
| 1708 |
+
"logits/chosen": -0.46787554025650024,
|
| 1709 |
+
"logits/rejected": -0.45076021552085876,
|
| 1710 |
+
"logps/chosen": -282.44903564453125,
|
| 1711 |
+
"logps/rejected": -328.4521789550781,
|
| 1712 |
+
"loss": 0.2076,
|
| 1713 |
+
"rewards/accuracies": 0.35624998807907104,
|
| 1714 |
+
"rewards/chosen": 0.012471871450543404,
|
| 1715 |
+
"rewards/margins": 0.01295357383787632,
|
| 1716 |
+
"rewards/rejected": -0.00048170099034905434,
|
| 1717 |
+
"step": 1210
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 0.93,
|
| 1721 |
+
"learning_rate": 7.46189337174788e-08,
|
| 1722 |
+
"logits/chosen": -0.4574583172798157,
|
| 1723 |
+
"logits/rejected": -0.48262113332748413,
|
| 1724 |
+
"logps/chosen": -287.9381408691406,
|
| 1725 |
+
"logps/rejected": -307.37017822265625,
|
| 1726 |
+
"loss": 0.2135,
|
| 1727 |
+
"rewards/accuracies": 0.33125001192092896,
|
| 1728 |
+
"rewards/chosen": 0.01407331507652998,
|
| 1729 |
+
"rewards/margins": 0.009314117953181267,
|
| 1730 |
+
"rewards/rejected": 0.004759198985993862,
|
| 1731 |
+
"step": 1220
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 0.94,
|
| 1735 |
+
"learning_rate": 5.934004245518793e-08,
|
| 1736 |
+
"logits/chosen": -0.5343510508537292,
|
| 1737 |
+
"logits/rejected": -0.42381030321121216,
|
| 1738 |
+
"logps/chosen": -301.0657958984375,
|
| 1739 |
+
"logps/rejected": -314.10260009765625,
|
| 1740 |
+
"loss": 0.2144,
|
| 1741 |
+
"rewards/accuracies": 0.33125001192092896,
|
| 1742 |
+
"rewards/chosen": 0.011122044175863266,
|
| 1743 |
+
"rewards/margins": 0.00779320765286684,
|
| 1744 |
+
"rewards/rejected": 0.0033288367558270693,
|
| 1745 |
+
"step": 1230
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 0.94,
|
| 1749 |
+
"learning_rate": 4.579103667367385e-08,
|
| 1750 |
+
"logits/chosen": -0.5401466488838196,
|
| 1751 |
+
"logits/rejected": -0.4905480444431305,
|
| 1752 |
+
"logps/chosen": -302.44818115234375,
|
| 1753 |
+
"logps/rejected": -316.00640869140625,
|
| 1754 |
+
"loss": 0.2009,
|
| 1755 |
+
"rewards/accuracies": 0.3125,
|
| 1756 |
+
"rewards/chosen": 0.009371964260935783,
|
| 1757 |
+
"rewards/margins": 0.005004683043807745,
|
| 1758 |
+
"rewards/rejected": 0.004367280751466751,
|
| 1759 |
+
"step": 1240
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 0.95,
|
| 1763 |
+
"learning_rate": 3.398151960681162e-08,
|
| 1764 |
+
"logits/chosen": -0.4864376485347748,
|
| 1765 |
+
"logits/rejected": -0.474257230758667,
|
| 1766 |
+
"logps/chosen": -272.18450927734375,
|
| 1767 |
+
"logps/rejected": -293.8534851074219,
|
| 1768 |
+
"loss": 0.212,
|
| 1769 |
+
"rewards/accuracies": 0.29374998807907104,
|
| 1770 |
+
"rewards/chosen": 0.010555078275501728,
|
| 1771 |
+
"rewards/margins": 0.007184301503002644,
|
| 1772 |
+
"rewards/rejected": 0.003370775608345866,
|
| 1773 |
+
"step": 1250
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 0.96,
|
| 1777 |
+
"learning_rate": 2.3919861577572924e-08,
|
| 1778 |
+
"logits/chosen": -0.48694509267807007,
|
| 1779 |
+
"logits/rejected": -0.4881502091884613,
|
| 1780 |
+
"logps/chosen": -285.6047668457031,
|
| 1781 |
+
"logps/rejected": -344.7032470703125,
|
| 1782 |
+
"loss": 0.207,
|
| 1783 |
+
"rewards/accuracies": 0.3375000059604645,
|
| 1784 |
+
"rewards/chosen": 0.010286719538271427,
|
| 1785 |
+
"rewards/margins": 0.011380873620510101,
|
| 1786 |
+
"rewards/rejected": -0.0010941538494080305,
|
| 1787 |
+
"step": 1260
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 0.97,
|
| 1791 |
+
"learning_rate": 1.5613194065327854e-08,
|
| 1792 |
+
"logits/chosen": -0.4984816610813141,
|
| 1793 |
+
"logits/rejected": -0.4886694550514221,
|
| 1794 |
+
"logps/chosen": -341.03369140625,
|
| 1795 |
+
"logps/rejected": -376.7138977050781,
|
| 1796 |
+
"loss": 0.1955,
|
| 1797 |
+
"rewards/accuracies": 0.38749998807907104,
|
| 1798 |
+
"rewards/chosen": 0.008053203113377094,
|
| 1799 |
+
"rewards/margins": 0.012355836108326912,
|
| 1800 |
+
"rewards/rejected": -0.004302632994949818,
|
| 1801 |
+
"step": 1270
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 0.98,
|
| 1805 |
+
"learning_rate": 9.067404651211808e-09,
|
| 1806 |
+
"logits/chosen": -0.47365856170654297,
|
| 1807 |
+
"logits/rejected": -0.463469922542572,
|
| 1808 |
+
"logps/chosen": -296.1658630371094,
|
| 1809 |
+
"logps/rejected": -324.7825012207031,
|
| 1810 |
+
"loss": 0.2016,
|
| 1811 |
+
"rewards/accuracies": 0.35624998807907104,
|
| 1812 |
+
"rewards/chosen": 0.00670830812305212,
|
| 1813 |
+
"rewards/margins": 0.009644337929785252,
|
| 1814 |
+
"rewards/rejected": -0.0029360298067331314,
|
| 1815 |
+
"step": 1280
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 0.98,
|
| 1819 |
+
"learning_rate": 4.287132845137709e-09,
|
| 1820 |
+
"logits/chosen": -0.4601454734802246,
|
| 1821 |
+
"logits/rejected": -0.44960451126098633,
|
| 1822 |
+
"logps/chosen": -305.7701110839844,
|
| 1823 |
+
"logps/rejected": -349.0350341796875,
|
| 1824 |
+
"loss": 0.1976,
|
| 1825 |
+
"rewards/accuracies": 0.3499999940395355,
|
| 1826 |
+
"rewards/chosen": 0.007751205004751682,
|
| 1827 |
+
"rewards/margins": 0.011196482926607132,
|
| 1828 |
+
"rewards/rejected": -0.0034452793188393116,
|
| 1829 |
+
"step": 1290
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 0.99,
|
| 1833 |
+
"learning_rate": 1.2757667974155896e-09,
|
| 1834 |
+
"logits/chosen": -0.4950791001319885,
|
| 1835 |
+
"logits/rejected": -0.47172340750694275,
|
| 1836 |
+
"logps/chosen": -312.46453857421875,
|
| 1837 |
+
"logps/rejected": -315.9293212890625,
|
| 1838 |
+
"loss": 0.2221,
|
| 1839 |
+
"rewards/accuracies": 0.2874999940395355,
|
| 1840 |
+
"rewards/chosen": 0.00022761887521483004,
|
| 1841 |
+
"rewards/margins": -0.004159080795943737,
|
| 1842 |
+
"rewards/rejected": 0.0043866997584700584,
|
| 1843 |
+
"step": 1300
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 1.0,
|
| 1847 |
+
"learning_rate": 3.544089730633804e-11,
|
| 1848 |
+
"logits/chosen": -0.48454493284225464,
|
| 1849 |
+
"logits/rejected": -0.46828117966651917,
|
| 1850 |
+
"logps/chosen": -290.0190734863281,
|
| 1851 |
+
"logps/rejected": -295.50665283203125,
|
| 1852 |
+
"loss": 0.1916,
|
| 1853 |
+
"rewards/accuracies": 0.29374998807907104,
|
| 1854 |
+
"rewards/chosen": 0.0072511411271989346,
|
| 1855 |
+
"rewards/margins": 0.00542284082621336,
|
| 1856 |
+
"rewards/rejected": 0.001828300068154931,
|
| 1857 |
+
"step": 1310
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 1.0,
|
| 1861 |
+
"step": 1312,
|
| 1862 |
+
"total_flos": 0.0,
|
| 1863 |
+
"train_loss": 0.20755663267677513,
|
| 1864 |
+
"train_runtime": 11765.5627,
|
| 1865 |
+
"train_samples_per_second": 1.785,
|
| 1866 |
+
"train_steps_per_second": 0.112
|
| 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 |
+
}
|