Model save
Browse files- README.md +62 -0
- adapter_model.safetensors +1 -1
- all_results.json +8 -0
- runs/May19_15-00-24_gpu4-119-5/events.out.tfevents.1716095069.gpu4-119-5.3541575.0 +2 -2
- train_results.json +8 -0
- trainer_state.json +2718 -0
README.md
ADDED
|
@@ -0,0 +1,62 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
library_name: peft
|
| 4 |
+
tags:
|
| 5 |
+
- trl
|
| 6 |
+
- dpo
|
| 7 |
+
- generated_from_trainer
|
| 8 |
+
base_model: DUAL-GPO/phi-2-sft-lora-ultrachat-merged
|
| 9 |
+
model-index:
|
| 10 |
+
- name: phi-2-ipo-chatml
|
| 11 |
+
results: []
|
| 12 |
+
---
|
| 13 |
+
|
| 14 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 15 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 16 |
+
|
| 17 |
+
# phi-2-ipo-chatml
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [DUAL-GPO/phi-2-sft-lora-ultrachat-merged](https://huggingface.co/DUAL-GPO/phi-2-sft-lora-ultrachat-merged) on the None dataset.
|
| 20 |
+
|
| 21 |
+
## Model description
|
| 22 |
+
|
| 23 |
+
More information needed
|
| 24 |
+
|
| 25 |
+
## Intended uses & limitations
|
| 26 |
+
|
| 27 |
+
More information needed
|
| 28 |
+
|
| 29 |
+
## Training and evaluation data
|
| 30 |
+
|
| 31 |
+
More information needed
|
| 32 |
+
|
| 33 |
+
## Training procedure
|
| 34 |
+
|
| 35 |
+
### Training hyperparameters
|
| 36 |
+
|
| 37 |
+
The following hyperparameters were used during training:
|
| 38 |
+
- learning_rate: 5e-06
|
| 39 |
+
- train_batch_size: 4
|
| 40 |
+
- eval_batch_size: 4
|
| 41 |
+
- seed: 42
|
| 42 |
+
- distributed_type: multi-GPU
|
| 43 |
+
- num_devices: 2
|
| 44 |
+
- gradient_accumulation_steps: 4
|
| 45 |
+
- total_train_batch_size: 32
|
| 46 |
+
- total_eval_batch_size: 8
|
| 47 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 48 |
+
- lr_scheduler_type: cosine
|
| 49 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 50 |
+
- num_epochs: 1
|
| 51 |
+
|
| 52 |
+
### Training results
|
| 53 |
+
|
| 54 |
+
|
| 55 |
+
|
| 56 |
+
### Framework versions
|
| 57 |
+
|
| 58 |
+
- PEFT 0.7.1
|
| 59 |
+
- Transformers 4.36.2
|
| 60 |
+
- Pytorch 2.1.2+cu121
|
| 61 |
+
- Datasets 2.14.6
|
| 62 |
+
- 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:e10f8da8c9ca888d0e2c2f7ebf035cdaf9ab2942f4ffbdef8926e08d3225748d
|
| 3 |
size 335579632
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 2071.8518089414265,
|
| 4 |
+
"train_runtime": 14310.5789,
|
| 5 |
+
"train_samples": 61135,
|
| 6 |
+
"train_samples_per_second": 4.272,
|
| 7 |
+
"train_steps_per_second": 0.133
|
| 8 |
+
}
|
runs/May19_15-00-24_gpu4-119-5/events.out.tfevents.1716095069.gpu4-119-5.3541575.0
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:caaa13a0e8513dfec362daca2ff17f22179ff4dd275de55216c3053b33319d02
|
| 3 |
+
size 126810
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 2071.8518089414265,
|
| 4 |
+
"train_runtime": 14310.5789,
|
| 5 |
+
"train_samples": 61135,
|
| 6 |
+
"train_samples_per_second": 4.272,
|
| 7 |
+
"train_steps_per_second": 0.133
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2718 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9997382884061764,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 1910,
|
| 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": 2.617801047120419e-08,
|
| 14 |
+
"logits/chosen": 0.5248222947120667,
|
| 15 |
+
"logits/rejected": 0.7921571731567383,
|
| 16 |
+
"logps/chosen": -341.40020751953125,
|
| 17 |
+
"logps/rejected": -250.28689575195312,
|
| 18 |
+
"loss": 2500.0,
|
| 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": 2.617801047120419e-07,
|
| 28 |
+
"logits/chosen": 0.6701858043670654,
|
| 29 |
+
"logits/rejected": 0.7214743494987488,
|
| 30 |
+
"logps/chosen": -282.2621765136719,
|
| 31 |
+
"logps/rejected": -253.0035858154297,
|
| 32 |
+
"loss": 2503.3785,
|
| 33 |
+
"rewards/accuracies": 0.3472222089767456,
|
| 34 |
+
"rewards/chosen": -0.0003065296623390168,
|
| 35 |
+
"rewards/margins": -0.0004795099375769496,
|
| 36 |
+
"rewards/rejected": 0.00017298036254942417,
|
| 37 |
+
"step": 10
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.01,
|
| 41 |
+
"learning_rate": 5.235602094240838e-07,
|
| 42 |
+
"logits/chosen": 0.6246625781059265,
|
| 43 |
+
"logits/rejected": 0.6360182166099548,
|
| 44 |
+
"logps/chosen": -314.952392578125,
|
| 45 |
+
"logps/rejected": -267.18927001953125,
|
| 46 |
+
"loss": 2498.9023,
|
| 47 |
+
"rewards/accuracies": 0.4375,
|
| 48 |
+
"rewards/chosen": 0.0006705065025016665,
|
| 49 |
+
"rewards/margins": -0.00035223033046349883,
|
| 50 |
+
"rewards/rejected": 0.0010227367747575045,
|
| 51 |
+
"step": 20
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.02,
|
| 55 |
+
"learning_rate": 7.853403141361258e-07,
|
| 56 |
+
"logits/chosen": 0.621583104133606,
|
| 57 |
+
"logits/rejected": 0.6869794726371765,
|
| 58 |
+
"logps/chosen": -260.4806213378906,
|
| 59 |
+
"logps/rejected": -254.1117401123047,
|
| 60 |
+
"loss": 2501.4164,
|
| 61 |
+
"rewards/accuracies": 0.4375,
|
| 62 |
+
"rewards/chosen": -0.0003680586232803762,
|
| 63 |
+
"rewards/margins": 8.779224299360067e-05,
|
| 64 |
+
"rewards/rejected": -0.00045585090992972255,
|
| 65 |
+
"step": 30
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.02,
|
| 69 |
+
"learning_rate": 1.0471204188481676e-06,
|
| 70 |
+
"logits/chosen": 0.6395634412765503,
|
| 71 |
+
"logits/rejected": 0.7919565439224243,
|
| 72 |
+
"logps/chosen": -240.0687255859375,
|
| 73 |
+
"logps/rejected": -229.44149780273438,
|
| 74 |
+
"loss": 2499.6219,
|
| 75 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 76 |
+
"rewards/chosen": 0.001351092243567109,
|
| 77 |
+
"rewards/margins": 0.0006501252064481378,
|
| 78 |
+
"rewards/rejected": 0.0007009669207036495,
|
| 79 |
+
"step": 40
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.03,
|
| 83 |
+
"learning_rate": 1.3089005235602096e-06,
|
| 84 |
+
"logits/chosen": 0.5841951370239258,
|
| 85 |
+
"logits/rejected": 0.6311219334602356,
|
| 86 |
+
"logps/chosen": -292.2647705078125,
|
| 87 |
+
"logps/rejected": -244.11605834960938,
|
| 88 |
+
"loss": 2499.3211,
|
| 89 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 90 |
+
"rewards/chosen": -0.0004217842943035066,
|
| 91 |
+
"rewards/margins": -0.0002647504734341055,
|
| 92 |
+
"rewards/rejected": -0.00015703373355790973,
|
| 93 |
+
"step": 50
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.03,
|
| 97 |
+
"learning_rate": 1.5706806282722515e-06,
|
| 98 |
+
"logits/chosen": 0.5548856854438782,
|
| 99 |
+
"logits/rejected": 0.6343160271644592,
|
| 100 |
+
"logps/chosen": -250.15615844726562,
|
| 101 |
+
"logps/rejected": -239.6744842529297,
|
| 102 |
+
"loss": 2497.8898,
|
| 103 |
+
"rewards/accuracies": 0.35624998807907104,
|
| 104 |
+
"rewards/chosen": -0.00014420936349779367,
|
| 105 |
+
"rewards/margins": -0.00012744043488055468,
|
| 106 |
+
"rewards/rejected": -1.6768904970376752e-05,
|
| 107 |
+
"step": 60
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.04,
|
| 111 |
+
"learning_rate": 1.8324607329842933e-06,
|
| 112 |
+
"logits/chosen": 0.5806199312210083,
|
| 113 |
+
"logits/rejected": 0.6849480867385864,
|
| 114 |
+
"logps/chosen": -263.5242004394531,
|
| 115 |
+
"logps/rejected": -242.7255096435547,
|
| 116 |
+
"loss": 2495.5641,
|
| 117 |
+
"rewards/accuracies": 0.5,
|
| 118 |
+
"rewards/chosen": -0.0010010639671236277,
|
| 119 |
+
"rewards/margins": 0.0013473776634782553,
|
| 120 |
+
"rewards/rejected": -0.0023484418634325266,
|
| 121 |
+
"step": 70
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.04,
|
| 125 |
+
"learning_rate": 2.094240837696335e-06,
|
| 126 |
+
"logits/chosen": 0.6474049687385559,
|
| 127 |
+
"logits/rejected": 0.6796275973320007,
|
| 128 |
+
"logps/chosen": -265.99652099609375,
|
| 129 |
+
"logps/rejected": -255.4057159423828,
|
| 130 |
+
"loss": 2483.6129,
|
| 131 |
+
"rewards/accuracies": 0.5062500238418579,
|
| 132 |
+
"rewards/chosen": 0.00029454095056280494,
|
| 133 |
+
"rewards/margins": 0.0018218889599666,
|
| 134 |
+
"rewards/rejected": -0.0015273483004420996,
|
| 135 |
+
"step": 80
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.05,
|
| 139 |
+
"learning_rate": 2.356020942408377e-06,
|
| 140 |
+
"logits/chosen": 0.6297029256820679,
|
| 141 |
+
"logits/rejected": 0.6693249344825745,
|
| 142 |
+
"logps/chosen": -259.20013427734375,
|
| 143 |
+
"logps/rejected": -260.8564453125,
|
| 144 |
+
"loss": 2468.257,
|
| 145 |
+
"rewards/accuracies": 0.59375,
|
| 146 |
+
"rewards/chosen": 0.0004229500482324511,
|
| 147 |
+
"rewards/margins": 0.0030678685288876295,
|
| 148 |
+
"rewards/rejected": -0.002644918393343687,
|
| 149 |
+
"step": 90
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.05,
|
| 153 |
+
"learning_rate": 2.617801047120419e-06,
|
| 154 |
+
"logits/chosen": 0.6495063304901123,
|
| 155 |
+
"logits/rejected": 0.6480900645256042,
|
| 156 |
+
"logps/chosen": -265.02508544921875,
|
| 157 |
+
"logps/rejected": -234.58029174804688,
|
| 158 |
+
"loss": 2468.3684,
|
| 159 |
+
"rewards/accuracies": 0.512499988079071,
|
| 160 |
+
"rewards/chosen": -0.000877298996783793,
|
| 161 |
+
"rewards/margins": 0.0035102677065879107,
|
| 162 |
+
"rewards/rejected": -0.0043875668197870255,
|
| 163 |
+
"step": 100
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.06,
|
| 167 |
+
"learning_rate": 2.8795811518324613e-06,
|
| 168 |
+
"logits/chosen": 0.6897019147872925,
|
| 169 |
+
"logits/rejected": 0.7566229701042175,
|
| 170 |
+
"logps/chosen": -304.11102294921875,
|
| 171 |
+
"logps/rejected": -264.30621337890625,
|
| 172 |
+
"loss": 2465.5504,
|
| 173 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 174 |
+
"rewards/chosen": -0.004195456858724356,
|
| 175 |
+
"rewards/margins": 0.0027198302559554577,
|
| 176 |
+
"rewards/rejected": -0.006915287580341101,
|
| 177 |
+
"step": 110
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.06,
|
| 181 |
+
"learning_rate": 3.141361256544503e-06,
|
| 182 |
+
"logits/chosen": 0.6134337186813354,
|
| 183 |
+
"logits/rejected": 0.7231487035751343,
|
| 184 |
+
"logps/chosen": -311.39984130859375,
|
| 185 |
+
"logps/rejected": -257.10650634765625,
|
| 186 |
+
"loss": 2428.1793,
|
| 187 |
+
"rewards/accuracies": 0.643750011920929,
|
| 188 |
+
"rewards/chosen": -0.0012777966912835836,
|
| 189 |
+
"rewards/margins": 0.009219733066856861,
|
| 190 |
+
"rewards/rejected": -0.010497529059648514,
|
| 191 |
+
"step": 120
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.07,
|
| 195 |
+
"learning_rate": 3.403141361256545e-06,
|
| 196 |
+
"logits/chosen": 0.7281027436256409,
|
| 197 |
+
"logits/rejected": 0.7294681668281555,
|
| 198 |
+
"logps/chosen": -287.9706115722656,
|
| 199 |
+
"logps/rejected": -253.9458770751953,
|
| 200 |
+
"loss": 2393.4254,
|
| 201 |
+
"rewards/accuracies": 0.637499988079071,
|
| 202 |
+
"rewards/chosen": 0.0035368993412703276,
|
| 203 |
+
"rewards/margins": 0.01360202394425869,
|
| 204 |
+
"rewards/rejected": -0.010065125301480293,
|
| 205 |
+
"step": 130
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.07,
|
| 209 |
+
"learning_rate": 3.6649214659685865e-06,
|
| 210 |
+
"logits/chosen": 0.6975444555282593,
|
| 211 |
+
"logits/rejected": 0.7464872598648071,
|
| 212 |
+
"logps/chosen": -283.344970703125,
|
| 213 |
+
"logps/rejected": -269.69134521484375,
|
| 214 |
+
"loss": 2391.2451,
|
| 215 |
+
"rewards/accuracies": 0.574999988079071,
|
| 216 |
+
"rewards/chosen": -0.0033713714219629765,
|
| 217 |
+
"rewards/margins": 0.01122850738465786,
|
| 218 |
+
"rewards/rejected": -0.0145998764783144,
|
| 219 |
+
"step": 140
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.08,
|
| 223 |
+
"learning_rate": 3.926701570680629e-06,
|
| 224 |
+
"logits/chosen": 0.6442640423774719,
|
| 225 |
+
"logits/rejected": 0.7055094838142395,
|
| 226 |
+
"logps/chosen": -297.76904296875,
|
| 227 |
+
"logps/rejected": -262.79010009765625,
|
| 228 |
+
"loss": 2339.1273,
|
| 229 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 230 |
+
"rewards/chosen": -0.006125102750957012,
|
| 231 |
+
"rewards/margins": 0.016310054808855057,
|
| 232 |
+
"rewards/rejected": -0.022435154765844345,
|
| 233 |
+
"step": 150
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.08,
|
| 237 |
+
"learning_rate": 4.18848167539267e-06,
|
| 238 |
+
"logits/chosen": 0.6234780550003052,
|
| 239 |
+
"logits/rejected": 0.5896192789077759,
|
| 240 |
+
"logps/chosen": -279.10498046875,
|
| 241 |
+
"logps/rejected": -247.49490356445312,
|
| 242 |
+
"loss": 2336.1186,
|
| 243 |
+
"rewards/accuracies": 0.581250011920929,
|
| 244 |
+
"rewards/chosen": -0.01037096418440342,
|
| 245 |
+
"rewards/margins": 0.019876617938280106,
|
| 246 |
+
"rewards/rejected": -0.030247583985328674,
|
| 247 |
+
"step": 160
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.09,
|
| 251 |
+
"learning_rate": 4.450261780104713e-06,
|
| 252 |
+
"logits/chosen": 0.6194897890090942,
|
| 253 |
+
"logits/rejected": 0.6363841891288757,
|
| 254 |
+
"logps/chosen": -293.1652526855469,
|
| 255 |
+
"logps/rejected": -242.565185546875,
|
| 256 |
+
"loss": 2349.7029,
|
| 257 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 258 |
+
"rewards/chosen": -0.013055374845862389,
|
| 259 |
+
"rewards/margins": 0.02146710641682148,
|
| 260 |
+
"rewards/rejected": -0.03452248126268387,
|
| 261 |
+
"step": 170
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.09,
|
| 265 |
+
"learning_rate": 4.712041884816754e-06,
|
| 266 |
+
"logits/chosen": 0.6403040885925293,
|
| 267 |
+
"logits/rejected": 0.7126356363296509,
|
| 268 |
+
"logps/chosen": -291.5748596191406,
|
| 269 |
+
"logps/rejected": -251.09121704101562,
|
| 270 |
+
"loss": 2273.3473,
|
| 271 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 272 |
+
"rewards/chosen": -0.030034661293029785,
|
| 273 |
+
"rewards/margins": 0.02929743006825447,
|
| 274 |
+
"rewards/rejected": -0.05933208391070366,
|
| 275 |
+
"step": 180
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.1,
|
| 279 |
+
"learning_rate": 4.9738219895287965e-06,
|
| 280 |
+
"logits/chosen": 0.7060214281082153,
|
| 281 |
+
"logits/rejected": 0.7062759399414062,
|
| 282 |
+
"logps/chosen": -251.7637939453125,
|
| 283 |
+
"logps/rejected": -224.1781768798828,
|
| 284 |
+
"loss": 2288.9881,
|
| 285 |
+
"rewards/accuracies": 0.53125,
|
| 286 |
+
"rewards/chosen": -0.039670929312705994,
|
| 287 |
+
"rewards/margins": 0.025101035833358765,
|
| 288 |
+
"rewards/rejected": -0.06477196514606476,
|
| 289 |
+
"step": 190
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.1,
|
| 293 |
+
"learning_rate": 4.999661831436499e-06,
|
| 294 |
+
"logits/chosen": 0.6065430045127869,
|
| 295 |
+
"logits/rejected": 0.5599089860916138,
|
| 296 |
+
"logps/chosen": -303.3594970703125,
|
| 297 |
+
"logps/rejected": -281.65301513671875,
|
| 298 |
+
"loss": 2323.0852,
|
| 299 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 300 |
+
"rewards/chosen": -0.051525335758924484,
|
| 301 |
+
"rewards/margins": 0.03378116711974144,
|
| 302 |
+
"rewards/rejected": -0.08530650287866592,
|
| 303 |
+
"step": 200
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.11,
|
| 307 |
+
"learning_rate": 4.9984929711403395e-06,
|
| 308 |
+
"logits/chosen": 0.6741048097610474,
|
| 309 |
+
"logits/rejected": 0.7067887187004089,
|
| 310 |
+
"logps/chosen": -256.2054138183594,
|
| 311 |
+
"logps/rejected": -225.0240478515625,
|
| 312 |
+
"loss": 2257.4254,
|
| 313 |
+
"rewards/accuracies": 0.5625,
|
| 314 |
+
"rewards/chosen": -0.05382019281387329,
|
| 315 |
+
"rewards/margins": 0.03237619996070862,
|
| 316 |
+
"rewards/rejected": -0.08619637787342072,
|
| 317 |
+
"step": 210
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.12,
|
| 321 |
+
"learning_rate": 4.996489634487865e-06,
|
| 322 |
+
"logits/chosen": 0.6408742070198059,
|
| 323 |
+
"logits/rejected": 0.7459608316421509,
|
| 324 |
+
"logps/chosen": -280.1336975097656,
|
| 325 |
+
"logps/rejected": -261.680908203125,
|
| 326 |
+
"loss": 2225.691,
|
| 327 |
+
"rewards/accuracies": 0.643750011920929,
|
| 328 |
+
"rewards/chosen": -0.06642502546310425,
|
| 329 |
+
"rewards/margins": 0.03849685937166214,
|
| 330 |
+
"rewards/rejected": -0.10492189228534698,
|
| 331 |
+
"step": 220
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.12,
|
| 335 |
+
"learning_rate": 4.9936524905772466e-06,
|
| 336 |
+
"logits/chosen": 0.511016845703125,
|
| 337 |
+
"logits/rejected": 0.6876882910728455,
|
| 338 |
+
"logps/chosen": -295.7177429199219,
|
| 339 |
+
"logps/rejected": -278.0430908203125,
|
| 340 |
+
"loss": 2274.0725,
|
| 341 |
+
"rewards/accuracies": 0.46875,
|
| 342 |
+
"rewards/chosen": -0.0953628420829773,
|
| 343 |
+
"rewards/margins": 0.025789355859160423,
|
| 344 |
+
"rewards/rejected": -0.12115219980478287,
|
| 345 |
+
"step": 230
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.13,
|
| 349 |
+
"learning_rate": 4.9899824869915e-06,
|
| 350 |
+
"logits/chosen": 0.6104953289031982,
|
| 351 |
+
"logits/rejected": 0.6206714510917664,
|
| 352 |
+
"logps/chosen": -269.36212158203125,
|
| 353 |
+
"logps/rejected": -230.29348754882812,
|
| 354 |
+
"loss": 2182.6342,
|
| 355 |
+
"rewards/accuracies": 0.625,
|
| 356 |
+
"rewards/chosen": -0.0956740602850914,
|
| 357 |
+
"rewards/margins": 0.05557785555720329,
|
| 358 |
+
"rewards/rejected": -0.1512519270181656,
|
| 359 |
+
"step": 240
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.13,
|
| 363 |
+
"learning_rate": 4.985480849482012e-06,
|
| 364 |
+
"logits/chosen": 0.5841513872146606,
|
| 365 |
+
"logits/rejected": 0.7289382815361023,
|
| 366 |
+
"logps/chosen": -297.09185791015625,
|
| 367 |
+
"logps/rejected": -283.0509338378906,
|
| 368 |
+
"loss": 2300.1652,
|
| 369 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 370 |
+
"rewards/chosen": -0.12446895986795425,
|
| 371 |
+
"rewards/margins": 0.020662058144807816,
|
| 372 |
+
"rewards/rejected": -0.14513102173805237,
|
| 373 |
+
"step": 250
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.14,
|
| 377 |
+
"learning_rate": 4.980149081559142e-06,
|
| 378 |
+
"logits/chosen": 0.594504177570343,
|
| 379 |
+
"logits/rejected": 0.659439742565155,
|
| 380 |
+
"logps/chosen": -318.9085693359375,
|
| 381 |
+
"logps/rejected": -282.30548095703125,
|
| 382 |
+
"loss": 2157.8906,
|
| 383 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 384 |
+
"rewards/chosen": -0.0681561678647995,
|
| 385 |
+
"rewards/margins": 0.05495098978281021,
|
| 386 |
+
"rewards/rejected": -0.12310715764760971,
|
| 387 |
+
"step": 260
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.14,
|
| 391 |
+
"learning_rate": 4.9739889639900655e-06,
|
| 392 |
+
"logits/chosen": 0.660437822341919,
|
| 393 |
+
"logits/rejected": 0.6371047496795654,
|
| 394 |
+
"logps/chosen": -278.40045166015625,
|
| 395 |
+
"logps/rejected": -276.32794189453125,
|
| 396 |
+
"loss": 2100.2004,
|
| 397 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 398 |
+
"rewards/chosen": -0.0825929194688797,
|
| 399 |
+
"rewards/margins": 0.06175379827618599,
|
| 400 |
+
"rewards/rejected": -0.1443466991186142,
|
| 401 |
+
"step": 270
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.15,
|
| 405 |
+
"learning_rate": 4.967002554204009e-06,
|
| 406 |
+
"logits/chosen": 0.5346761345863342,
|
| 407 |
+
"logits/rejected": 0.653354287147522,
|
| 408 |
+
"logps/chosen": -274.6492919921875,
|
| 409 |
+
"logps/rejected": -256.0135192871094,
|
| 410 |
+
"loss": 2231.843,
|
| 411 |
+
"rewards/accuracies": 0.550000011920929,
|
| 412 |
+
"rewards/chosen": -0.10598695278167725,
|
| 413 |
+
"rewards/margins": 0.04181862622499466,
|
| 414 |
+
"rewards/rejected": -0.1478056013584137,
|
| 415 |
+
"step": 280
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.15,
|
| 419 |
+
"learning_rate": 4.959192185605089e-06,
|
| 420 |
+
"logits/chosen": 0.6121161580085754,
|
| 421 |
+
"logits/rejected": 0.6786028742790222,
|
| 422 |
+
"logps/chosen": -291.69268798828125,
|
| 423 |
+
"logps/rejected": -270.1987609863281,
|
| 424 |
+
"loss": 2302.0896,
|
| 425 |
+
"rewards/accuracies": 0.5687500238418579,
|
| 426 |
+
"rewards/chosen": -0.11571818590164185,
|
| 427 |
+
"rewards/margins": 0.04277648404240608,
|
| 428 |
+
"rewards/rejected": -0.15849466621875763,
|
| 429 |
+
"step": 290
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.16,
|
| 433 |
+
"learning_rate": 4.950560466792969e-06,
|
| 434 |
+
"logits/chosen": 0.6503465175628662,
|
| 435 |
+
"logits/rejected": 0.6750337481498718,
|
| 436 |
+
"logps/chosen": -297.166259765625,
|
| 437 |
+
"logps/rejected": -264.58905029296875,
|
| 438 |
+
"loss": 2290.2439,
|
| 439 |
+
"rewards/accuracies": 0.59375,
|
| 440 |
+
"rewards/chosen": -0.1382652223110199,
|
| 441 |
+
"rewards/margins": 0.050003498792648315,
|
| 442 |
+
"rewards/rejected": -0.1882687211036682,
|
| 443 |
+
"step": 300
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.16,
|
| 447 |
+
"learning_rate": 4.9411102806916185e-06,
|
| 448 |
+
"logits/chosen": 0.6111949682235718,
|
| 449 |
+
"logits/rejected": 0.5789340734481812,
|
| 450 |
+
"logps/chosen": -335.71209716796875,
|
| 451 |
+
"logps/rejected": -271.94281005859375,
|
| 452 |
+
"loss": 2030.3432,
|
| 453 |
+
"rewards/accuracies": 0.71875,
|
| 454 |
+
"rewards/chosen": -0.1214783638715744,
|
| 455 |
+
"rewards/margins": 0.07279713451862335,
|
| 456 |
+
"rewards/rejected": -0.19427552819252014,
|
| 457 |
+
"step": 310
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.17,
|
| 461 |
+
"learning_rate": 4.930844783586424e-06,
|
| 462 |
+
"logits/chosen": 0.6168066263198853,
|
| 463 |
+
"logits/rejected": 0.6404728293418884,
|
| 464 |
+
"logps/chosen": -267.7217102050781,
|
| 465 |
+
"logps/rejected": -254.76156616210938,
|
| 466 |
+
"loss": 2180.659,
|
| 467 |
+
"rewards/accuracies": 0.606249988079071,
|
| 468 |
+
"rewards/chosen": -0.1399281919002533,
|
| 469 |
+
"rewards/margins": 0.05871574953198433,
|
| 470 |
+
"rewards/rejected": -0.19864396750926971,
|
| 471 |
+
"step": 320
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.17,
|
| 475 |
+
"learning_rate": 4.919767404070033e-06,
|
| 476 |
+
"logits/chosen": 0.6463326215744019,
|
| 477 |
+
"logits/rejected": 0.6055666208267212,
|
| 478 |
+
"logps/chosen": -288.9108581542969,
|
| 479 |
+
"logps/rejected": -271.7781066894531,
|
| 480 |
+
"loss": 2136.0844,
|
| 481 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 482 |
+
"rewards/chosen": -0.18461689352989197,
|
| 483 |
+
"rewards/margins": 0.058290112763643265,
|
| 484 |
+
"rewards/rejected": -0.24290700256824493,
|
| 485 |
+
"step": 330
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.18,
|
| 489 |
+
"learning_rate": 4.907881841897216e-06,
|
| 490 |
+
"logits/chosen": 0.5913775563240051,
|
| 491 |
+
"logits/rejected": 0.6087537407875061,
|
| 492 |
+
"logps/chosen": -348.69317626953125,
|
| 493 |
+
"logps/rejected": -276.8719787597656,
|
| 494 |
+
"loss": 2095.2922,
|
| 495 |
+
"rewards/accuracies": 0.65625,
|
| 496 |
+
"rewards/chosen": -0.21791012585163116,
|
| 497 |
+
"rewards/margins": 0.08299825340509415,
|
| 498 |
+
"rewards/rejected": -0.3009083867073059,
|
| 499 |
+
"step": 340
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.18,
|
| 503 |
+
"learning_rate": 4.89519206674919e-06,
|
| 504 |
+
"logits/chosen": 0.5306932926177979,
|
| 505 |
+
"logits/rejected": 0.5599890947341919,
|
| 506 |
+
"logps/chosen": -277.52801513671875,
|
| 507 |
+
"logps/rejected": -282.8912048339844,
|
| 508 |
+
"loss": 2027.6721,
|
| 509 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 510 |
+
"rewards/chosen": -0.2483474463224411,
|
| 511 |
+
"rewards/margins": 0.08130116015672684,
|
| 512 |
+
"rewards/rejected": -0.32964861392974854,
|
| 513 |
+
"step": 350
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.19,
|
| 517 |
+
"learning_rate": 4.881702316907769e-06,
|
| 518 |
+
"logits/chosen": 0.5017831921577454,
|
| 519 |
+
"logits/rejected": 0.6145971417427063,
|
| 520 |
+
"logps/chosen": -247.94332885742188,
|
| 521 |
+
"logps/rejected": -276.70562744140625,
|
| 522 |
+
"loss": 2096.8201,
|
| 523 |
+
"rewards/accuracies": 0.59375,
|
| 524 |
+
"rewards/chosen": -0.2660738229751587,
|
| 525 |
+
"rewards/margins": 0.07246068120002747,
|
| 526 |
+
"rewards/rejected": -0.33853450417518616,
|
| 527 |
+
"step": 360
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.19,
|
| 531 |
+
"learning_rate": 4.86741709783982e-06,
|
| 532 |
+
"logits/chosen": 0.5044198632240295,
|
| 533 |
+
"logits/rejected": 0.610668957233429,
|
| 534 |
+
"logps/chosen": -370.6680908203125,
|
| 535 |
+
"logps/rejected": -315.87445068359375,
|
| 536 |
+
"loss": 2240.1234,
|
| 537 |
+
"rewards/accuracies": 0.625,
|
| 538 |
+
"rewards/chosen": -0.2816161513328552,
|
| 539 |
+
"rewards/margins": 0.0706692636013031,
|
| 540 |
+
"rewards/rejected": -0.3522854149341583,
|
| 541 |
+
"step": 370
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.2,
|
| 545 |
+
"learning_rate": 4.852341180692471e-06,
|
| 546 |
+
"logits/chosen": 0.5787937045097351,
|
| 547 |
+
"logits/rejected": 0.6844087839126587,
|
| 548 |
+
"logps/chosen": -314.65386962890625,
|
| 549 |
+
"logps/rejected": -279.57464599609375,
|
| 550 |
+
"loss": 2000.0223,
|
| 551 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 552 |
+
"rewards/chosen": -0.28157955408096313,
|
| 553 |
+
"rewards/margins": 0.09683749079704285,
|
| 554 |
+
"rewards/rejected": -0.3784170150756836,
|
| 555 |
+
"step": 380
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.2,
|
| 559 |
+
"learning_rate": 4.836479600699579e-06,
|
| 560 |
+
"logits/chosen": 0.5943303108215332,
|
| 561 |
+
"logits/rejected": 0.5367528200149536,
|
| 562 |
+
"logps/chosen": -326.36322021484375,
|
| 563 |
+
"logps/rejected": -322.28851318359375,
|
| 564 |
+
"loss": 2038.491,
|
| 565 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 566 |
+
"rewards/chosen": -0.31988343596458435,
|
| 567 |
+
"rewards/margins": 0.08327943086624146,
|
| 568 |
+
"rewards/rejected": -0.4031628668308258,
|
| 569 |
+
"step": 390
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.21,
|
| 573 |
+
"learning_rate": 4.819837655500014e-06,
|
| 574 |
+
"logits/chosen": 0.4382709562778473,
|
| 575 |
+
"logits/rejected": 0.5377568006515503,
|
| 576 |
+
"logps/chosen": -268.5506286621094,
|
| 577 |
+
"logps/rejected": -256.25457763671875,
|
| 578 |
+
"loss": 2150.4953,
|
| 579 |
+
"rewards/accuracies": 0.606249988079071,
|
| 580 |
+
"rewards/chosen": -0.3451124131679535,
|
| 581 |
+
"rewards/margins": 0.06852660328149796,
|
| 582 |
+
"rewards/rejected": -0.41363900899887085,
|
| 583 |
+
"step": 400
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.21,
|
| 587 |
+
"learning_rate": 4.802420903368286e-06,
|
| 588 |
+
"logits/chosen": 0.5234228372573853,
|
| 589 |
+
"logits/rejected": 0.5555657744407654,
|
| 590 |
+
"logps/chosen": -306.6415100097656,
|
| 591 |
+
"logps/rejected": -284.44061279296875,
|
| 592 |
+
"loss": 2237.9328,
|
| 593 |
+
"rewards/accuracies": 0.5562499761581421,
|
| 594 |
+
"rewards/chosen": -0.3614524304866791,
|
| 595 |
+
"rewards/margins": 0.05758042261004448,
|
| 596 |
+
"rewards/rejected": -0.41903290152549744,
|
| 597 |
+
"step": 410
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.22,
|
| 601 |
+
"learning_rate": 4.784235161358124e-06,
|
| 602 |
+
"logits/chosen": 0.477716863155365,
|
| 603 |
+
"logits/rejected": 0.5133184194564819,
|
| 604 |
+
"logps/chosen": -334.1138916015625,
|
| 605 |
+
"logps/rejected": -301.9080505371094,
|
| 606 |
+
"loss": 2092.0188,
|
| 607 |
+
"rewards/accuracies": 0.643750011920929,
|
| 608 |
+
"rewards/chosen": -0.3627592921257019,
|
| 609 |
+
"rewards/margins": 0.10072964429855347,
|
| 610 |
+
"rewards/rejected": -0.463488906621933,
|
| 611 |
+
"step": 420
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.23,
|
| 615 |
+
"learning_rate": 4.765286503359632e-06,
|
| 616 |
+
"logits/chosen": 0.5422715544700623,
|
| 617 |
+
"logits/rejected": 0.6406744122505188,
|
| 618 |
+
"logps/chosen": -305.5355529785156,
|
| 619 |
+
"logps/rejected": -282.97210693359375,
|
| 620 |
+
"loss": 2014.3465,
|
| 621 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 622 |
+
"rewards/chosen": -0.29455187916755676,
|
| 623 |
+
"rewards/margins": 0.07915346324443817,
|
| 624 |
+
"rewards/rejected": -0.3737053871154785,
|
| 625 |
+
"step": 430
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.23,
|
| 629 |
+
"learning_rate": 4.745581258070654e-06,
|
| 630 |
+
"logits/chosen": 0.43309369683265686,
|
| 631 |
+
"logits/rejected": 0.5271438360214233,
|
| 632 |
+
"logps/chosen": -287.7671813964844,
|
| 633 |
+
"logps/rejected": -281.0647888183594,
|
| 634 |
+
"loss": 2185.2459,
|
| 635 |
+
"rewards/accuracies": 0.550000011920929,
|
| 636 |
+
"rewards/chosen": -0.34788942337036133,
|
| 637 |
+
"rewards/margins": 0.05820406228303909,
|
| 638 |
+
"rewards/rejected": -0.4060935080051422,
|
| 639 |
+
"step": 440
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.24,
|
| 643 |
+
"learning_rate": 4.725126006883047e-06,
|
| 644 |
+
"logits/chosen": 0.45388108491897583,
|
| 645 |
+
"logits/rejected": 0.5083224177360535,
|
| 646 |
+
"logps/chosen": -268.25958251953125,
|
| 647 |
+
"logps/rejected": -266.2535400390625,
|
| 648 |
+
"loss": 2137.6896,
|
| 649 |
+
"rewards/accuracies": 0.550000011920929,
|
| 650 |
+
"rewards/chosen": -0.3282170295715332,
|
| 651 |
+
"rewards/margins": 0.05696592479944229,
|
| 652 |
+
"rewards/rejected": -0.3851829469203949,
|
| 653 |
+
"step": 450
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.24,
|
| 657 |
+
"learning_rate": 4.70392758168454e-06,
|
| 658 |
+
"logits/chosen": 0.5513511896133423,
|
| 659 |
+
"logits/rejected": 0.510746955871582,
|
| 660 |
+
"logps/chosen": -370.2562255859375,
|
| 661 |
+
"logps/rejected": -305.41265869140625,
|
| 662 |
+
"loss": 2088.2188,
|
| 663 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 664 |
+
"rewards/chosen": -0.31895995140075684,
|
| 665 |
+
"rewards/margins": 0.08234255015850067,
|
| 666 |
+
"rewards/rejected": -0.4013025164604187,
|
| 667 |
+
"step": 460
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.25,
|
| 671 |
+
"learning_rate": 4.68199306257695e-06,
|
| 672 |
+
"logits/chosen": 0.5287891626358032,
|
| 673 |
+
"logits/rejected": 0.5082263350486755,
|
| 674 |
+
"logps/chosen": -355.8836364746094,
|
| 675 |
+
"logps/rejected": -318.6692810058594,
|
| 676 |
+
"loss": 2073.2719,
|
| 677 |
+
"rewards/accuracies": 0.643750011920929,
|
| 678 |
+
"rewards/chosen": -0.327432781457901,
|
| 679 |
+
"rewards/margins": 0.10078072547912598,
|
| 680 |
+
"rewards/rejected": -0.42821353673934937,
|
| 681 |
+
"step": 470
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.25,
|
| 685 |
+
"learning_rate": 4.659329775511478e-06,
|
| 686 |
+
"logits/chosen": 0.553450345993042,
|
| 687 |
+
"logits/rejected": 0.5118339657783508,
|
| 688 |
+
"logps/chosen": -307.2048645019531,
|
| 689 |
+
"logps/rejected": -274.0884094238281,
|
| 690 |
+
"loss": 2053.993,
|
| 691 |
+
"rewards/accuracies": 0.612500011920929,
|
| 692 |
+
"rewards/chosen": -0.3153889775276184,
|
| 693 |
+
"rewards/margins": 0.08515409380197525,
|
| 694 |
+
"rewards/rejected": -0.40054306387901306,
|
| 695 |
+
"step": 480
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.26,
|
| 699 |
+
"learning_rate": 4.635945289841902e-06,
|
| 700 |
+
"logits/chosen": 0.49377956986427307,
|
| 701 |
+
"logits/rejected": 0.4585692286491394,
|
| 702 |
+
"logps/chosen": -312.8728942871094,
|
| 703 |
+
"logps/rejected": -320.9603271484375,
|
| 704 |
+
"loss": 2090.7937,
|
| 705 |
+
"rewards/accuracies": 0.606249988079071,
|
| 706 |
+
"rewards/chosen": -0.3384076952934265,
|
| 707 |
+
"rewards/margins": 0.10788736492395401,
|
| 708 |
+
"rewards/rejected": -0.4462950825691223,
|
| 709 |
+
"step": 490
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.26,
|
| 713 |
+
"learning_rate": 4.611847415796476e-06,
|
| 714 |
+
"logits/chosen": 0.4848670959472656,
|
| 715 |
+
"logits/rejected": 0.42902207374572754,
|
| 716 |
+
"logps/chosen": -343.72515869140625,
|
| 717 |
+
"logps/rejected": -293.16815185546875,
|
| 718 |
+
"loss": 2090.8078,
|
| 719 |
+
"rewards/accuracies": 0.574999988079071,
|
| 720 |
+
"rewards/chosen": -0.37585896253585815,
|
| 721 |
+
"rewards/margins": 0.0655459612607956,
|
| 722 |
+
"rewards/rejected": -0.44140490889549255,
|
| 723 |
+
"step": 500
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.27,
|
| 727 |
+
"learning_rate": 4.587044201869378e-06,
|
| 728 |
+
"logits/chosen": 0.4169999957084656,
|
| 729 |
+
"logits/rejected": 0.4936888813972473,
|
| 730 |
+
"logps/chosen": -311.93963623046875,
|
| 731 |
+
"logps/rejected": -320.00408935546875,
|
| 732 |
+
"loss": 2092.8064,
|
| 733 |
+
"rewards/accuracies": 0.5687500238418579,
|
| 734 |
+
"rewards/chosen": -0.42008256912231445,
|
| 735 |
+
"rewards/margins": 0.07488216459751129,
|
| 736 |
+
"rewards/rejected": -0.49496474862098694,
|
| 737 |
+
"step": 510
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.27,
|
| 741 |
+
"learning_rate": 4.561543932132574e-06,
|
| 742 |
+
"logits/chosen": 0.46682390570640564,
|
| 743 |
+
"logits/rejected": 0.4480462074279785,
|
| 744 |
+
"logps/chosen": -331.2510070800781,
|
| 745 |
+
"logps/rejected": -289.58355712890625,
|
| 746 |
+
"loss": 2058.4975,
|
| 747 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 748 |
+
"rewards/chosen": -0.4290853440761566,
|
| 749 |
+
"rewards/margins": 0.08971880376338959,
|
| 750 |
+
"rewards/rejected": -0.5188041925430298,
|
| 751 |
+
"step": 520
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.28,
|
| 755 |
+
"learning_rate": 4.535355123469009e-06,
|
| 756 |
+
"logits/chosen": 0.4296157956123352,
|
| 757 |
+
"logits/rejected": 0.4146192967891693,
|
| 758 |
+
"logps/chosen": -331.16595458984375,
|
| 759 |
+
"logps/rejected": -309.8785705566406,
|
| 760 |
+
"loss": 2106.7164,
|
| 761 |
+
"rewards/accuracies": 0.637499988079071,
|
| 762 |
+
"rewards/chosen": -0.5114426612854004,
|
| 763 |
+
"rewards/margins": 0.09367333352565765,
|
| 764 |
+
"rewards/rejected": -0.6051160097122192,
|
| 765 |
+
"step": 530
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.28,
|
| 769 |
+
"learning_rate": 4.508486522728037e-06,
|
| 770 |
+
"logits/chosen": 0.5072197318077087,
|
| 771 |
+
"logits/rejected": 0.5970960855484009,
|
| 772 |
+
"logps/chosen": -322.4172668457031,
|
| 773 |
+
"logps/rejected": -315.7408752441406,
|
| 774 |
+
"loss": 2065.5746,
|
| 775 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 776 |
+
"rewards/chosen": -0.49308618903160095,
|
| 777 |
+
"rewards/margins": 0.08288516104221344,
|
| 778 |
+
"rewards/rejected": -0.5759714245796204,
|
| 779 |
+
"step": 540
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.29,
|
| 783 |
+
"learning_rate": 4.480947103804044e-06,
|
| 784 |
+
"logits/chosen": 0.5679532289505005,
|
| 785 |
+
"logits/rejected": 0.6028575897216797,
|
| 786 |
+
"logps/chosen": -318.32794189453125,
|
| 787 |
+
"logps/rejected": -275.52716064453125,
|
| 788 |
+
"loss": 2036.5631,
|
| 789 |
+
"rewards/accuracies": 0.699999988079071,
|
| 790 |
+
"rewards/chosen": -0.5034439563751221,
|
| 791 |
+
"rewards/margins": 0.1076350212097168,
|
| 792 |
+
"rewards/rejected": -0.6110790371894836,
|
| 793 |
+
"step": 550
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.29,
|
| 797 |
+
"learning_rate": 4.452746064639239e-06,
|
| 798 |
+
"logits/chosen": 0.3957621455192566,
|
| 799 |
+
"logits/rejected": 0.5071176290512085,
|
| 800 |
+
"logps/chosen": -377.244873046875,
|
| 801 |
+
"logps/rejected": -318.47821044921875,
|
| 802 |
+
"loss": 2035.9066,
|
| 803 |
+
"rewards/accuracies": 0.65625,
|
| 804 |
+
"rewards/chosen": -0.4797874093055725,
|
| 805 |
+
"rewards/margins": 0.0891679972410202,
|
| 806 |
+
"rewards/rejected": -0.5689553022384644,
|
| 807 |
+
"step": 560
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.3,
|
| 811 |
+
"learning_rate": 4.423892824151617e-06,
|
| 812 |
+
"logits/chosen": 0.5428576469421387,
|
| 813 |
+
"logits/rejected": 0.5582025051116943,
|
| 814 |
+
"logps/chosen": -325.265869140625,
|
| 815 |
+
"logps/rejected": -279.44622802734375,
|
| 816 |
+
"loss": 2099.3277,
|
| 817 |
+
"rewards/accuracies": 0.625,
|
| 818 |
+
"rewards/chosen": -0.4889600872993469,
|
| 819 |
+
"rewards/margins": 0.08092018216848373,
|
| 820 |
+
"rewards/rejected": -0.5698802471160889,
|
| 821 |
+
"step": 570
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.3,
|
| 825 |
+
"learning_rate": 4.3943970190891164e-06,
|
| 826 |
+
"logits/chosen": 0.47033435106277466,
|
| 827 |
+
"logits/rejected": 0.4538179039955139,
|
| 828 |
+
"logps/chosen": -338.0190124511719,
|
| 829 |
+
"logps/rejected": -284.2043151855469,
|
| 830 |
+
"loss": 1990.1314,
|
| 831 |
+
"rewards/accuracies": 0.612500011920929,
|
| 832 |
+
"rewards/chosen": -0.5050845742225647,
|
| 833 |
+
"rewards/margins": 0.07665327191352844,
|
| 834 |
+
"rewards/rejected": -0.5817378163337708,
|
| 835 |
+
"step": 580
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.31,
|
| 839 |
+
"learning_rate": 4.364268500811025e-06,
|
| 840 |
+
"logits/chosen": 0.43608421087265015,
|
| 841 |
+
"logits/rejected": 0.588505208492279,
|
| 842 |
+
"logps/chosen": -355.6614685058594,
|
| 843 |
+
"logps/rejected": -311.62603759765625,
|
| 844 |
+
"loss": 1957.0082,
|
| 845 |
+
"rewards/accuracies": 0.668749988079071,
|
| 846 |
+
"rewards/chosen": -0.5208545327186584,
|
| 847 |
+
"rewards/margins": 0.10252735763788223,
|
| 848 |
+
"rewards/rejected": -0.6233818531036377,
|
| 849 |
+
"step": 590
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.31,
|
| 853 |
+
"learning_rate": 4.333517331997704e-06,
|
| 854 |
+
"logits/chosen": 0.5493937730789185,
|
| 855 |
+
"logits/rejected": 0.5916265249252319,
|
| 856 |
+
"logps/chosen": -346.2372131347656,
|
| 857 |
+
"logps/rejected": -297.2783203125,
|
| 858 |
+
"loss": 1937.0328,
|
| 859 |
+
"rewards/accuracies": 0.65625,
|
| 860 |
+
"rewards/chosen": -0.4746910631656647,
|
| 861 |
+
"rewards/margins": 0.10906670242547989,
|
| 862 |
+
"rewards/rejected": -0.583757758140564,
|
| 863 |
+
"step": 600
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.32,
|
| 867 |
+
"learning_rate": 4.302153783289737e-06,
|
| 868 |
+
"logits/chosen": 0.5612093210220337,
|
| 869 |
+
"logits/rejected": 0.5712814331054688,
|
| 870 |
+
"logps/chosen": -299.03192138671875,
|
| 871 |
+
"logps/rejected": -295.932373046875,
|
| 872 |
+
"loss": 1899.6357,
|
| 873 |
+
"rewards/accuracies": 0.643750011920929,
|
| 874 |
+
"rewards/chosen": -0.45807942748069763,
|
| 875 |
+
"rewards/margins": 0.11209867149591446,
|
| 876 |
+
"rewards/rejected": -0.5701780915260315,
|
| 877 |
+
"step": 610
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.32,
|
| 881 |
+
"learning_rate": 4.270188329857613e-06,
|
| 882 |
+
"logits/chosen": 0.4900333285331726,
|
| 883 |
+
"logits/rejected": 0.6181058287620544,
|
| 884 |
+
"logps/chosen": -345.901123046875,
|
| 885 |
+
"logps/rejected": -315.23150634765625,
|
| 886 |
+
"loss": 1829.7244,
|
| 887 |
+
"rewards/accuracies": 0.637499988079071,
|
| 888 |
+
"rewards/chosen": -0.44543027877807617,
|
| 889 |
+
"rewards/margins": 0.11854176223278046,
|
| 890 |
+
"rewards/rejected": -0.563971996307373,
|
| 891 |
+
"step": 620
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.33,
|
| 895 |
+
"learning_rate": 4.237631647903115e-06,
|
| 896 |
+
"logits/chosen": 0.5171164274215698,
|
| 897 |
+
"logits/rejected": 0.6000061631202698,
|
| 898 |
+
"logps/chosen": -305.6175842285156,
|
| 899 |
+
"logps/rejected": -288.61529541015625,
|
| 900 |
+
"loss": 1928.8063,
|
| 901 |
+
"rewards/accuracies": 0.643750011920929,
|
| 902 |
+
"rewards/chosen": -0.41954368352890015,
|
| 903 |
+
"rewards/margins": 0.10498888790607452,
|
| 904 |
+
"rewards/rejected": -0.5245326161384583,
|
| 905 |
+
"step": 630
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.33,
|
| 909 |
+
"learning_rate": 4.204494611093548e-06,
|
| 910 |
+
"logits/chosen": 0.5002217292785645,
|
| 911 |
+
"logits/rejected": 0.4931250512599945,
|
| 912 |
+
"logps/chosen": -374.89447021484375,
|
| 913 |
+
"logps/rejected": -323.4224853515625,
|
| 914 |
+
"loss": 2052.2016,
|
| 915 |
+
"rewards/accuracies": 0.643750011920929,
|
| 916 |
+
"rewards/chosen": -0.4047914445400238,
|
| 917 |
+
"rewards/margins": 0.10931004583835602,
|
| 918 |
+
"rewards/rejected": -0.5141014456748962,
|
| 919 |
+
"step": 640
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.34,
|
| 923 |
+
"learning_rate": 4.170788286930024e-06,
|
| 924 |
+
"logits/chosen": 0.5889243483543396,
|
| 925 |
+
"logits/rejected": 0.49473047256469727,
|
| 926 |
+
"logps/chosen": -311.41204833984375,
|
| 927 |
+
"logps/rejected": -297.2741394042969,
|
| 928 |
+
"loss": 2161.5602,
|
| 929 |
+
"rewards/accuracies": 0.606249988079071,
|
| 930 |
+
"rewards/chosen": -0.4144001007080078,
|
| 931 |
+
"rewards/margins": 0.09458984434604645,
|
| 932 |
+
"rewards/rejected": -0.5089899301528931,
|
| 933 |
+
"step": 650
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.35,
|
| 937 |
+
"learning_rate": 4.136523933051005e-06,
|
| 938 |
+
"logits/chosen": 0.46171918511390686,
|
| 939 |
+
"logits/rejected": 0.5521343350410461,
|
| 940 |
+
"logps/chosen": -260.4045715332031,
|
| 941 |
+
"logps/rejected": -253.0673370361328,
|
| 942 |
+
"loss": 2058.6592,
|
| 943 |
+
"rewards/accuracies": 0.5562499761581421,
|
| 944 |
+
"rewards/chosen": -0.35439711809158325,
|
| 945 |
+
"rewards/margins": 0.0739353746175766,
|
| 946 |
+
"rewards/rejected": -0.42833250761032104,
|
| 947 |
+
"step": 660
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.35,
|
| 951 |
+
"learning_rate": 4.101712993472348e-06,
|
| 952 |
+
"logits/chosen": 0.5095491409301758,
|
| 953 |
+
"logits/rejected": 0.5851413607597351,
|
| 954 |
+
"logps/chosen": -333.0138244628906,
|
| 955 |
+
"logps/rejected": -299.2201232910156,
|
| 956 |
+
"loss": 2114.274,
|
| 957 |
+
"rewards/accuracies": 0.5687500238418579,
|
| 958 |
+
"rewards/chosen": -0.3759954273700714,
|
| 959 |
+
"rewards/margins": 0.07854921370744705,
|
| 960 |
+
"rewards/rejected": -0.45454463362693787,
|
| 961 |
+
"step": 670
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.36,
|
| 965 |
+
"learning_rate": 4.066367094765091e-06,
|
| 966 |
+
"logits/chosen": 0.5400400161743164,
|
| 967 |
+
"logits/rejected": 0.6057177782058716,
|
| 968 |
+
"logps/chosen": -314.9399108886719,
|
| 969 |
+
"logps/rejected": -299.5291442871094,
|
| 970 |
+
"loss": 2165.3699,
|
| 971 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 972 |
+
"rewards/chosen": -0.3974139094352722,
|
| 973 |
+
"rewards/margins": 0.0817367285490036,
|
| 974 |
+
"rewards/rejected": -0.4791506826877594,
|
| 975 |
+
"step": 680
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.36,
|
| 979 |
+
"learning_rate": 4.030498042172277e-06,
|
| 980 |
+
"logits/chosen": 0.46717318892478943,
|
| 981 |
+
"logits/rejected": 0.5683552622795105,
|
| 982 |
+
"logps/chosen": -311.09478759765625,
|
| 983 |
+
"logps/rejected": -286.71856689453125,
|
| 984 |
+
"loss": 1970.309,
|
| 985 |
+
"rewards/accuracies": 0.637499988079071,
|
| 986 |
+
"rewards/chosen": -0.36804935336112976,
|
| 987 |
+
"rewards/margins": 0.09274602681398392,
|
| 988 |
+
"rewards/rejected": -0.46079540252685547,
|
| 989 |
+
"step": 690
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.37,
|
| 993 |
+
"learning_rate": 3.994117815666095e-06,
|
| 994 |
+
"logits/chosen": 0.5168116688728333,
|
| 995 |
+
"logits/rejected": 0.5277969241142273,
|
| 996 |
+
"logps/chosen": -331.3283996582031,
|
| 997 |
+
"logps/rejected": -316.9101257324219,
|
| 998 |
+
"loss": 2094.1586,
|
| 999 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 1000 |
+
"rewards/chosen": -0.3705582618713379,
|
| 1001 |
+
"rewards/margins": 0.07725582271814346,
|
| 1002 |
+
"rewards/rejected": -0.44781407713890076,
|
| 1003 |
+
"step": 700
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.37,
|
| 1007 |
+
"learning_rate": 3.957238565946672e-06,
|
| 1008 |
+
"logits/chosen": 0.4891355633735657,
|
| 1009 |
+
"logits/rejected": 0.5829272270202637,
|
| 1010 |
+
"logps/chosen": -341.33270263671875,
|
| 1011 |
+
"logps/rejected": -326.12237548828125,
|
| 1012 |
+
"loss": 2015.8363,
|
| 1013 |
+
"rewards/accuracies": 0.668749988079071,
|
| 1014 |
+
"rewards/chosen": -0.3461948037147522,
|
| 1015 |
+
"rewards/margins": 0.1047135442495346,
|
| 1016 |
+
"rewards/rejected": -0.450908362865448,
|
| 1017 |
+
"step": 710
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.38,
|
| 1021 |
+
"learning_rate": 3.919872610383831e-06,
|
| 1022 |
+
"logits/chosen": 0.5361688733100891,
|
| 1023 |
+
"logits/rejected": 0.4950820505619049,
|
| 1024 |
+
"logps/chosen": -301.51470947265625,
|
| 1025 |
+
"logps/rejected": -264.515625,
|
| 1026 |
+
"loss": 2067.884,
|
| 1027 |
+
"rewards/accuracies": 0.637499988079071,
|
| 1028 |
+
"rewards/chosen": -0.31028127670288086,
|
| 1029 |
+
"rewards/margins": 0.07867135107517242,
|
| 1030 |
+
"rewards/rejected": -0.38895267248153687,
|
| 1031 |
+
"step": 720
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.38,
|
| 1035 |
+
"learning_rate": 3.882032428903195e-06,
|
| 1036 |
+
"logits/chosen": 0.5467667579650879,
|
| 1037 |
+
"logits/rejected": 0.6214498281478882,
|
| 1038 |
+
"logps/chosen": -310.25006103515625,
|
| 1039 |
+
"logps/rejected": -304.9784240722656,
|
| 1040 |
+
"loss": 1854.108,
|
| 1041 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 1042 |
+
"rewards/chosen": -0.31836968660354614,
|
| 1043 |
+
"rewards/margins": 0.10814164578914642,
|
| 1044 |
+
"rewards/rejected": -0.42651137709617615,
|
| 1045 |
+
"step": 730
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.39,
|
| 1049 |
+
"learning_rate": 3.84373065981799e-06,
|
| 1050 |
+
"logits/chosen": 0.5906549692153931,
|
| 1051 |
+
"logits/rejected": 0.5477146506309509,
|
| 1052 |
+
"logps/chosen": -336.2082214355469,
|
| 1053 |
+
"logps/rejected": -320.98419189453125,
|
| 1054 |
+
"loss": 1870.9252,
|
| 1055 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 1056 |
+
"rewards/chosen": -0.4078896641731262,
|
| 1057 |
+
"rewards/margins": 0.12890958786010742,
|
| 1058 |
+
"rewards/rejected": -0.5367991924285889,
|
| 1059 |
+
"step": 740
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.39,
|
| 1063 |
+
"learning_rate": 3.8049800956079552e-06,
|
| 1064 |
+
"logits/chosen": 0.5637394189834595,
|
| 1065 |
+
"logits/rejected": 0.5951513648033142,
|
| 1066 |
+
"logps/chosen": -312.58380126953125,
|
| 1067 |
+
"logps/rejected": -294.388671875,
|
| 1068 |
+
"loss": 2077.7766,
|
| 1069 |
+
"rewards/accuracies": 0.581250011920929,
|
| 1070 |
+
"rewards/chosen": -0.4279128909111023,
|
| 1071 |
+
"rewards/margins": 0.0825895294547081,
|
| 1072 |
+
"rewards/rejected": -0.5105024576187134,
|
| 1073 |
+
"step": 750
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.4,
|
| 1077 |
+
"learning_rate": 3.765793678646753e-06,
|
| 1078 |
+
"logits/chosen": 0.49519261717796326,
|
| 1079 |
+
"logits/rejected": 0.6189637184143066,
|
| 1080 |
+
"logps/chosen": -306.1346435546875,
|
| 1081 |
+
"logps/rejected": -292.5857238769531,
|
| 1082 |
+
"loss": 1969.0184,
|
| 1083 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 1084 |
+
"rewards/chosen": -0.36586111783981323,
|
| 1085 |
+
"rewards/margins": 0.12026441097259521,
|
| 1086 |
+
"rewards/rejected": -0.48612555861473083,
|
| 1087 |
+
"step": 760
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.4,
|
| 1091 |
+
"learning_rate": 3.726184496879323e-06,
|
| 1092 |
+
"logits/chosen": 0.4921692907810211,
|
| 1093 |
+
"logits/rejected": 0.501875102519989,
|
| 1094 |
+
"logps/chosen": -299.2643127441406,
|
| 1095 |
+
"logps/rejected": -286.10308837890625,
|
| 1096 |
+
"loss": 2031.8379,
|
| 1097 |
+
"rewards/accuracies": 0.612500011920929,
|
| 1098 |
+
"rewards/chosen": -0.4021153450012207,
|
| 1099 |
+
"rewards/margins": 0.09204810857772827,
|
| 1100 |
+
"rewards/rejected": -0.494163453578949,
|
| 1101 |
+
"step": 770
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.41,
|
| 1105 |
+
"learning_rate": 3.686165779450619e-06,
|
| 1106 |
+
"logits/chosen": 0.4793570935726166,
|
| 1107 |
+
"logits/rejected": 0.529563307762146,
|
| 1108 |
+
"logps/chosen": -292.5688171386719,
|
| 1109 |
+
"logps/rejected": -271.9722595214844,
|
| 1110 |
+
"loss": 2042.7758,
|
| 1111 |
+
"rewards/accuracies": 0.5625,
|
| 1112 |
+
"rewards/chosen": -0.35684093832969666,
|
| 1113 |
+
"rewards/margins": 0.09431228041648865,
|
| 1114 |
+
"rewards/rejected": -0.4511532187461853,
|
| 1115 |
+
"step": 780
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.41,
|
| 1119 |
+
"learning_rate": 3.645750892287178e-06,
|
| 1120 |
+
"logits/chosen": 0.5109771490097046,
|
| 1121 |
+
"logits/rejected": 0.5633312463760376,
|
| 1122 |
+
"logps/chosen": -340.49176025390625,
|
| 1123 |
+
"logps/rejected": -291.06756591796875,
|
| 1124 |
+
"loss": 1950.8371,
|
| 1125 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 1126 |
+
"rewards/chosen": -0.33940690755844116,
|
| 1127 |
+
"rewards/margins": 0.10259196907281876,
|
| 1128 |
+
"rewards/rejected": -0.4419988691806793,
|
| 1129 |
+
"step": 790
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.42,
|
| 1133 |
+
"learning_rate": 3.604953333633009e-06,
|
| 1134 |
+
"logits/chosen": 0.5391252636909485,
|
| 1135 |
+
"logits/rejected": 0.5337890982627869,
|
| 1136 |
+
"logps/chosen": -289.8169860839844,
|
| 1137 |
+
"logps/rejected": -260.87677001953125,
|
| 1138 |
+
"loss": 1936.325,
|
| 1139 |
+
"rewards/accuracies": 0.612500011920929,
|
| 1140 |
+
"rewards/chosen": -0.32662123441696167,
|
| 1141 |
+
"rewards/margins": 0.08759995549917221,
|
| 1142 |
+
"rewards/rejected": -0.4142211973667145,
|
| 1143 |
+
"step": 800
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.42,
|
| 1147 |
+
"learning_rate": 3.56378672954129e-06,
|
| 1148 |
+
"logits/chosen": 0.5750656127929688,
|
| 1149 |
+
"logits/rejected": 0.6396702527999878,
|
| 1150 |
+
"logps/chosen": -315.6180114746094,
|
| 1151 |
+
"logps/rejected": -275.87847900390625,
|
| 1152 |
+
"loss": 1988.6426,
|
| 1153 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1154 |
+
"rewards/chosen": -0.3871499300003052,
|
| 1155 |
+
"rewards/margins": 0.11068111658096313,
|
| 1156 |
+
"rewards/rejected": -0.4978310167789459,
|
| 1157 |
+
"step": 810
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.43,
|
| 1161 |
+
"learning_rate": 3.5222648293233806e-06,
|
| 1162 |
+
"logits/chosen": 0.5217655897140503,
|
| 1163 |
+
"logits/rejected": 0.5655398368835449,
|
| 1164 |
+
"logps/chosen": -312.867919921875,
|
| 1165 |
+
"logps/rejected": -288.189697265625,
|
| 1166 |
+
"loss": 2051.176,
|
| 1167 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1168 |
+
"rewards/chosen": -0.36912721395492554,
|
| 1169 |
+
"rewards/margins": 0.11765513569116592,
|
| 1170 |
+
"rewards/rejected": -0.48678237199783325,
|
| 1171 |
+
"step": 820
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.43,
|
| 1175 |
+
"learning_rate": 3.4804015009566573e-06,
|
| 1176 |
+
"logits/chosen": 0.5041700005531311,
|
| 1177 |
+
"logits/rejected": 0.5614610314369202,
|
| 1178 |
+
"logps/chosen": -299.4610595703125,
|
| 1179 |
+
"logps/rejected": -294.58990478515625,
|
| 1180 |
+
"loss": 1867.1797,
|
| 1181 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 1182 |
+
"rewards/chosen": -0.3925167918205261,
|
| 1183 |
+
"rewards/margins": 0.11273722350597382,
|
| 1184 |
+
"rewards/rejected": -0.5052539706230164,
|
| 1185 |
+
"step": 830
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.44,
|
| 1189 |
+
"learning_rate": 3.4382107264527244e-06,
|
| 1190 |
+
"logits/chosen": 0.5158332586288452,
|
| 1191 |
+
"logits/rejected": 0.5967845916748047,
|
| 1192 |
+
"logps/chosen": -292.78387451171875,
|
| 1193 |
+
"logps/rejected": -286.22259521484375,
|
| 1194 |
+
"loss": 2125.177,
|
| 1195 |
+
"rewards/accuracies": 0.581250011920929,
|
| 1196 |
+
"rewards/chosen": -0.38172584772109985,
|
| 1197 |
+
"rewards/margins": 0.08153903484344482,
|
| 1198 |
+
"rewards/rejected": -0.4632648825645447,
|
| 1199 |
+
"step": 840
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.44,
|
| 1203 |
+
"learning_rate": 3.3957065971875387e-06,
|
| 1204 |
+
"logits/chosen": 0.5155011415481567,
|
| 1205 |
+
"logits/rejected": 0.6147416234016418,
|
| 1206 |
+
"logps/chosen": -309.46826171875,
|
| 1207 |
+
"logps/rejected": -292.97857666015625,
|
| 1208 |
+
"loss": 2040.8613,
|
| 1209 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1210 |
+
"rewards/chosen": -0.3627835512161255,
|
| 1211 |
+
"rewards/margins": 0.08301069587469101,
|
| 1212 |
+
"rewards/rejected": -0.4457942843437195,
|
| 1213 |
+
"step": 850
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.45,
|
| 1217 |
+
"learning_rate": 3.352903309194999e-06,
|
| 1218 |
+
"logits/chosen": 0.49975937604904175,
|
| 1219 |
+
"logits/rejected": 0.5919264554977417,
|
| 1220 |
+
"logps/chosen": -314.7627868652344,
|
| 1221 |
+
"logps/rejected": -304.3371276855469,
|
| 1222 |
+
"loss": 2078.6633,
|
| 1223 |
+
"rewards/accuracies": 0.643750011920929,
|
| 1224 |
+
"rewards/chosen": -0.3559049963951111,
|
| 1225 |
+
"rewards/margins": 0.08720938861370087,
|
| 1226 |
+
"rewards/rejected": -0.44311434030532837,
|
| 1227 |
+
"step": 860
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.46,
|
| 1231 |
+
"learning_rate": 3.309815158425591e-06,
|
| 1232 |
+
"logits/chosen": 0.5549314618110657,
|
| 1233 |
+
"logits/rejected": 0.45847368240356445,
|
| 1234 |
+
"logps/chosen": -302.06573486328125,
|
| 1235 |
+
"logps/rejected": -272.84136962890625,
|
| 1236 |
+
"loss": 2036.8318,
|
| 1237 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1238 |
+
"rewards/chosen": -0.35362687706947327,
|
| 1239 |
+
"rewards/margins": 0.08619710803031921,
|
| 1240 |
+
"rewards/rejected": -0.4398239552974701,
|
| 1241 |
+
"step": 870
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.46,
|
| 1245 |
+
"learning_rate": 3.266456535971654e-06,
|
| 1246 |
+
"logits/chosen": 0.5452480912208557,
|
| 1247 |
+
"logits/rejected": 0.6346302628517151,
|
| 1248 |
+
"logps/chosen": -334.15423583984375,
|
| 1249 |
+
"logps/rejected": -286.4046630859375,
|
| 1250 |
+
"loss": 2093.8855,
|
| 1251 |
+
"rewards/accuracies": 0.637499988079071,
|
| 1252 |
+
"rewards/chosen": -0.37248533964157104,
|
| 1253 |
+
"rewards/margins": 0.08576939254999161,
|
| 1254 |
+
"rewards/rejected": -0.45825472474098206,
|
| 1255 |
+
"step": 880
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.47,
|
| 1259 |
+
"learning_rate": 3.2228419232608692e-06,
|
| 1260 |
+
"logits/chosen": 0.535306990146637,
|
| 1261 |
+
"logits/rejected": 0.4712342321872711,
|
| 1262 |
+
"logps/chosen": -282.85968017578125,
|
| 1263 |
+
"logps/rejected": -274.203125,
|
| 1264 |
+
"loss": 2098.0096,
|
| 1265 |
+
"rewards/accuracies": 0.612500011920929,
|
| 1266 |
+
"rewards/chosen": -0.34499186277389526,
|
| 1267 |
+
"rewards/margins": 0.08891385793685913,
|
| 1268 |
+
"rewards/rejected": -0.433905690908432,
|
| 1269 |
+
"step": 890
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.47,
|
| 1273 |
+
"learning_rate": 3.1789858872195888e-06,
|
| 1274 |
+
"logits/chosen": 0.6342380046844482,
|
| 1275 |
+
"logits/rejected": 0.629281759262085,
|
| 1276 |
+
"logps/chosen": -286.6862487792969,
|
| 1277 |
+
"logps/rejected": -271.05792236328125,
|
| 1278 |
+
"loss": 1953.5867,
|
| 1279 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 1280 |
+
"rewards/chosen": -0.3626910150051117,
|
| 1281 |
+
"rewards/margins": 0.0837903618812561,
|
| 1282 |
+
"rewards/rejected": -0.4464813768863678,
|
| 1283 |
+
"step": 900
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.48,
|
| 1287 |
+
"learning_rate": 3.1349030754075945e-06,
|
| 1288 |
+
"logits/chosen": 0.5246872901916504,
|
| 1289 |
+
"logits/rejected": 0.5165098905563354,
|
| 1290 |
+
"logps/chosen": -342.14508056640625,
|
| 1291 |
+
"logps/rejected": -293.86541748046875,
|
| 1292 |
+
"loss": 1883.041,
|
| 1293 |
+
"rewards/accuracies": 0.5687500238418579,
|
| 1294 |
+
"rewards/chosen": -0.3446907103061676,
|
| 1295 |
+
"rewards/margins": 0.09030432999134064,
|
| 1296 |
+
"rewards/rejected": -0.43499502539634705,
|
| 1297 |
+
"step": 910
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.48,
|
| 1301 |
+
"learning_rate": 3.0906082111259313e-06,
|
| 1302 |
+
"logits/chosen": 0.5085369944572449,
|
| 1303 |
+
"logits/rejected": 0.6210469603538513,
|
| 1304 |
+
"logps/chosen": -324.135498046875,
|
| 1305 |
+
"logps/rejected": -288.7994079589844,
|
| 1306 |
+
"loss": 2054.7223,
|
| 1307 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1308 |
+
"rewards/chosen": -0.4166959822177887,
|
| 1309 |
+
"rewards/margins": 0.06678648293018341,
|
| 1310 |
+
"rewards/rejected": -0.4834825098514557,
|
| 1311 |
+
"step": 920
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.49,
|
| 1315 |
+
"learning_rate": 3.046116088499449e-06,
|
| 1316 |
+
"logits/chosen": 0.4991762638092041,
|
| 1317 |
+
"logits/rejected": 0.5716468691825867,
|
| 1318 |
+
"logps/chosen": -322.47100830078125,
|
| 1319 |
+
"logps/rejected": -305.85247802734375,
|
| 1320 |
+
"loss": 1875.9535,
|
| 1321 |
+
"rewards/accuracies": 0.643750011920929,
|
| 1322 |
+
"rewards/chosen": -0.4109939634799957,
|
| 1323 |
+
"rewards/margins": 0.1062377318739891,
|
| 1324 |
+
"rewards/rejected": -0.5172317028045654,
|
| 1325 |
+
"step": 930
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 0.49,
|
| 1329 |
+
"learning_rate": 3.0014415675356813e-06,
|
| 1330 |
+
"logits/chosen": 0.5478680729866028,
|
| 1331 |
+
"logits/rejected": 0.5703476667404175,
|
| 1332 |
+
"logps/chosen": -349.9249572753906,
|
| 1333 |
+
"logps/rejected": -293.72625732421875,
|
| 1334 |
+
"loss": 2017.5168,
|
| 1335 |
+
"rewards/accuracies": 0.637499988079071,
|
| 1336 |
+
"rewards/chosen": -0.42143529653549194,
|
| 1337 |
+
"rewards/margins": 0.10333251953125,
|
| 1338 |
+
"rewards/rejected": -0.5247678160667419,
|
| 1339 |
+
"step": 940
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.5,
|
| 1343 |
+
"learning_rate": 2.9565995691617242e-06,
|
| 1344 |
+
"logits/chosen": 0.5334219932556152,
|
| 1345 |
+
"logits/rejected": 0.5596605539321899,
|
| 1346 |
+
"logps/chosen": -266.18426513671875,
|
| 1347 |
+
"logps/rejected": -274.7750244140625,
|
| 1348 |
+
"loss": 2175.4988,
|
| 1349 |
+
"rewards/accuracies": 0.550000011920929,
|
| 1350 |
+
"rewards/chosen": -0.4017234444618225,
|
| 1351 |
+
"rewards/margins": 0.06810633838176727,
|
| 1352 |
+
"rewards/rejected": -0.469829797744751,
|
| 1353 |
+
"step": 950
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 0.5,
|
| 1357 |
+
"learning_rate": 2.9116050702407706e-06,
|
| 1358 |
+
"logits/chosen": 0.602304995059967,
|
| 1359 |
+
"logits/rejected": 0.5311247110366821,
|
| 1360 |
+
"logps/chosen": -277.86669921875,
|
| 1361 |
+
"logps/rejected": -269.7265319824219,
|
| 1362 |
+
"loss": 2177.3363,
|
| 1363 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 1364 |
+
"rewards/chosen": -0.4107086658477783,
|
| 1365 |
+
"rewards/margins": 0.04552067071199417,
|
| 1366 |
+
"rewards/rejected": -0.4562292993068695,
|
| 1367 |
+
"step": 960
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.51,
|
| 1371 |
+
"learning_rate": 2.8664730985699537e-06,
|
| 1372 |
+
"logits/chosen": 0.503399670124054,
|
| 1373 |
+
"logits/rejected": 0.639790415763855,
|
| 1374 |
+
"logps/chosen": -273.44622802734375,
|
| 1375 |
+
"logps/rejected": -260.8460693359375,
|
| 1376 |
+
"loss": 2045.4781,
|
| 1377 |
+
"rewards/accuracies": 0.550000011920929,
|
| 1378 |
+
"rewards/chosen": -0.35943537950515747,
|
| 1379 |
+
"rewards/margins": 0.06276627629995346,
|
| 1380 |
+
"rewards/rejected": -0.42220163345336914,
|
| 1381 |
+
"step": 970
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 0.51,
|
| 1385 |
+
"learning_rate": 2.8212187278611907e-06,
|
| 1386 |
+
"logits/chosen": 0.5885658860206604,
|
| 1387 |
+
"logits/rejected": 0.5756844282150269,
|
| 1388 |
+
"logps/chosen": -325.32952880859375,
|
| 1389 |
+
"logps/rejected": -282.21453857421875,
|
| 1390 |
+
"loss": 1891.5383,
|
| 1391 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 1392 |
+
"rewards/chosen": -0.33539459109306335,
|
| 1393 |
+
"rewards/margins": 0.12442357838153839,
|
| 1394 |
+
"rewards/rejected": -0.45981818437576294,
|
| 1395 |
+
"step": 980
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 0.52,
|
| 1399 |
+
"learning_rate": 2.7758570727066843e-06,
|
| 1400 |
+
"logits/chosen": 0.6374679803848267,
|
| 1401 |
+
"logits/rejected": 0.6624254584312439,
|
| 1402 |
+
"logps/chosen": -295.37738037109375,
|
| 1403 |
+
"logps/rejected": -270.3449401855469,
|
| 1404 |
+
"loss": 1947.8027,
|
| 1405 |
+
"rewards/accuracies": 0.606249988079071,
|
| 1406 |
+
"rewards/chosen": -0.31836217641830444,
|
| 1407 |
+
"rewards/margins": 0.11628556251525879,
|
| 1408 |
+
"rewards/rejected": -0.43464773893356323,
|
| 1409 |
+
"step": 990
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 0.52,
|
| 1413 |
+
"learning_rate": 2.730403283530767e-06,
|
| 1414 |
+
"logits/chosen": 0.5361552238464355,
|
| 1415 |
+
"logits/rejected": 0.5510913133621216,
|
| 1416 |
+
"logps/chosen": -301.1198425292969,
|
| 1417 |
+
"logps/rejected": -294.21234130859375,
|
| 1418 |
+
"loss": 2041.7438,
|
| 1419 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 1420 |
+
"rewards/chosen": -0.34723925590515137,
|
| 1421 |
+
"rewards/margins": 0.09882111847400665,
|
| 1422 |
+
"rewards/rejected": -0.44606032967567444,
|
| 1423 |
+
"step": 1000
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 0.53,
|
| 1427 |
+
"learning_rate": 2.6848725415297888e-06,
|
| 1428 |
+
"logits/chosen": 0.4825092852115631,
|
| 1429 |
+
"logits/rejected": 0.5265048742294312,
|
| 1430 |
+
"logps/chosen": -304.63140869140625,
|
| 1431 |
+
"logps/rejected": -276.94464111328125,
|
| 1432 |
+
"loss": 1806.9965,
|
| 1433 |
+
"rewards/accuracies": 0.675000011920929,
|
| 1434 |
+
"rewards/chosen": -0.360734760761261,
|
| 1435 |
+
"rewards/margins": 0.12996909022331238,
|
| 1436 |
+
"rewards/rejected": -0.49070388078689575,
|
| 1437 |
+
"step": 1010
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 0.53,
|
| 1441 |
+
"learning_rate": 2.639280053601719e-06,
|
| 1442 |
+
"logits/chosen": 0.5663384199142456,
|
| 1443 |
+
"logits/rejected": 0.6073741912841797,
|
| 1444 |
+
"logps/chosen": -316.7065124511719,
|
| 1445 |
+
"logps/rejected": -322.63494873046875,
|
| 1446 |
+
"loss": 2131.3072,
|
| 1447 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1448 |
+
"rewards/chosen": -0.40306559205055237,
|
| 1449 |
+
"rewards/margins": 0.07923634350299835,
|
| 1450 |
+
"rewards/rejected": -0.48230189085006714,
|
| 1451 |
+
"step": 1020
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 0.54,
|
| 1455 |
+
"learning_rate": 2.59364104726716e-06,
|
| 1456 |
+
"logits/chosen": 0.5961092710494995,
|
| 1457 |
+
"logits/rejected": 0.6114310622215271,
|
| 1458 |
+
"logps/chosen": -321.1643981933594,
|
| 1459 |
+
"logps/rejected": -285.0345153808594,
|
| 1460 |
+
"loss": 1963.3941,
|
| 1461 |
+
"rewards/accuracies": 0.625,
|
| 1462 |
+
"rewards/chosen": -0.3763844072818756,
|
| 1463 |
+
"rewards/margins": 0.11694605648517609,
|
| 1464 |
+
"rewards/rejected": -0.4933304190635681,
|
| 1465 |
+
"step": 1030
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 0.54,
|
| 1469 |
+
"learning_rate": 2.547970765583491e-06,
|
| 1470 |
+
"logits/chosen": 0.5694259405136108,
|
| 1471 |
+
"logits/rejected": 0.5002479553222656,
|
| 1472 |
+
"logps/chosen": -326.5146484375,
|
| 1473 |
+
"logps/rejected": -302.0265808105469,
|
| 1474 |
+
"loss": 2105.7691,
|
| 1475 |
+
"rewards/accuracies": 0.643750011920929,
|
| 1476 |
+
"rewards/chosen": -0.40365272760391235,
|
| 1477 |
+
"rewards/margins": 0.09866581857204437,
|
| 1478 |
+
"rewards/rejected": -0.5023185014724731,
|
| 1479 |
+
"step": 1040
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 0.55,
|
| 1483 |
+
"learning_rate": 2.502284462053799e-06,
|
| 1484 |
+
"logits/chosen": 0.48704004287719727,
|
| 1485 |
+
"logits/rejected": 0.5804616808891296,
|
| 1486 |
+
"logps/chosen": -332.91162109375,
|
| 1487 |
+
"logps/rejected": -303.14617919921875,
|
| 1488 |
+
"loss": 1834.8215,
|
| 1489 |
+
"rewards/accuracies": 0.675000011920929,
|
| 1490 |
+
"rewards/chosen": -0.3842596411705017,
|
| 1491 |
+
"rewards/margins": 0.12224143743515015,
|
| 1492 |
+
"rewards/rejected": -0.5065010786056519,
|
| 1493 |
+
"step": 1050
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 0.55,
|
| 1497 |
+
"learning_rate": 2.456597395532338e-06,
|
| 1498 |
+
"logits/chosen": 0.4348418116569519,
|
| 1499 |
+
"logits/rejected": 0.5593420267105103,
|
| 1500 |
+
"logps/chosen": -335.1719055175781,
|
| 1501 |
+
"logps/rejected": -308.69830322265625,
|
| 1502 |
+
"loss": 1984.7055,
|
| 1503 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 1504 |
+
"rewards/chosen": -0.3781774044036865,
|
| 1505 |
+
"rewards/margins": 0.09801622480154037,
|
| 1506 |
+
"rewards/rejected": -0.4761936068534851,
|
| 1507 |
+
"step": 1060
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 0.56,
|
| 1511 |
+
"learning_rate": 2.4109248251281953e-06,
|
| 1512 |
+
"logits/chosen": 0.5774113535881042,
|
| 1513 |
+
"logits/rejected": 0.6447092294692993,
|
| 1514 |
+
"logps/chosen": -314.6346740722656,
|
| 1515 |
+
"logps/rejected": -309.253173828125,
|
| 1516 |
+
"loss": 2091.5516,
|
| 1517 |
+
"rewards/accuracies": 0.625,
|
| 1518 |
+
"rewards/chosen": -0.3814038932323456,
|
| 1519 |
+
"rewards/margins": 0.10535021126270294,
|
| 1520 |
+
"rewards/rejected": -0.48675408959388733,
|
| 1521 |
+
"step": 1070
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 0.57,
|
| 1525 |
+
"learning_rate": 2.365282005108875e-06,
|
| 1526 |
+
"logits/chosen": 0.5727014541625977,
|
| 1527 |
+
"logits/rejected": 0.5744349360466003,
|
| 1528 |
+
"logps/chosen": -324.27154541015625,
|
| 1529 |
+
"logps/rejected": -283.59686279296875,
|
| 1530 |
+
"loss": 1978.2016,
|
| 1531 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 1532 |
+
"rewards/chosen": -0.3670370876789093,
|
| 1533 |
+
"rewards/margins": 0.101667121052742,
|
| 1534 |
+
"rewards/rejected": -0.46870413422584534,
|
| 1535 |
+
"step": 1080
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 0.57,
|
| 1539 |
+
"learning_rate": 2.319684179805491e-06,
|
| 1540 |
+
"logits/chosen": 0.51801598072052,
|
| 1541 |
+
"logits/rejected": 0.5624555349349976,
|
| 1542 |
+
"logps/chosen": -300.46734619140625,
|
| 1543 |
+
"logps/rejected": -287.84307861328125,
|
| 1544 |
+
"loss": 1968.3572,
|
| 1545 |
+
"rewards/accuracies": 0.625,
|
| 1546 |
+
"rewards/chosen": -0.34267181158065796,
|
| 1547 |
+
"rewards/margins": 0.11022396385669708,
|
| 1548 |
+
"rewards/rejected": -0.45289579033851624,
|
| 1549 |
+
"step": 1090
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 0.58,
|
| 1553 |
+
"learning_rate": 2.2741465785212905e-06,
|
| 1554 |
+
"logits/chosen": 0.5614932775497437,
|
| 1555 |
+
"logits/rejected": 0.5673717856407166,
|
| 1556 |
+
"logps/chosen": -306.4109191894531,
|
| 1557 |
+
"logps/rejected": -280.79205322265625,
|
| 1558 |
+
"loss": 2043.6598,
|
| 1559 |
+
"rewards/accuracies": 0.574999988079071,
|
| 1560 |
+
"rewards/chosen": -0.3364596962928772,
|
| 1561 |
+
"rewards/margins": 0.0902407318353653,
|
| 1562 |
+
"rewards/rejected": -0.42670050263404846,
|
| 1563 |
+
"step": 1100
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 0.58,
|
| 1567 |
+
"learning_rate": 2.2286844104451848e-06,
|
| 1568 |
+
"logits/chosen": 0.5359445810317993,
|
| 1569 |
+
"logits/rejected": 0.6252576112747192,
|
| 1570 |
+
"logps/chosen": -322.9229736328125,
|
| 1571 |
+
"logps/rejected": -268.3149108886719,
|
| 1572 |
+
"loss": 2021.5035,
|
| 1573 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1574 |
+
"rewards/chosen": -0.3492088317871094,
|
| 1575 |
+
"rewards/margins": 0.09384563565254211,
|
| 1576 |
+
"rewards/rejected": -0.4430544972419739,
|
| 1577 |
+
"step": 1110
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 0.59,
|
| 1581 |
+
"learning_rate": 2.183312859572008e-06,
|
| 1582 |
+
"logits/chosen": 0.4629904627799988,
|
| 1583 |
+
"logits/rejected": 0.6062875390052795,
|
| 1584 |
+
"logps/chosen": -305.40814208984375,
|
| 1585 |
+
"logps/rejected": -251.60107421875,
|
| 1586 |
+
"loss": 2015.4504,
|
| 1587 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1588 |
+
"rewards/chosen": -0.33251887559890747,
|
| 1589 |
+
"rewards/margins": 0.09136182069778442,
|
| 1590 |
+
"rewards/rejected": -0.4238806664943695,
|
| 1591 |
+
"step": 1120
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 0.59,
|
| 1595 |
+
"learning_rate": 2.1380470796311843e-06,
|
| 1596 |
+
"logits/chosen": 0.5688912868499756,
|
| 1597 |
+
"logits/rejected": 0.5831522941589355,
|
| 1598 |
+
"logps/chosen": -300.88079833984375,
|
| 1599 |
+
"logps/rejected": -278.1544494628906,
|
| 1600 |
+
"loss": 2065.1121,
|
| 1601 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 1602 |
+
"rewards/chosen": -0.35972604155540466,
|
| 1603 |
+
"rewards/margins": 0.08174722641706467,
|
| 1604 |
+
"rewards/rejected": -0.4414733052253723,
|
| 1605 |
+
"step": 1130
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 0.6,
|
| 1609 |
+
"learning_rate": 2.092902189025507e-06,
|
| 1610 |
+
"logits/chosen": 0.5642494559288025,
|
| 1611 |
+
"logits/rejected": 0.5858667492866516,
|
| 1612 |
+
"logps/chosen": -290.29217529296875,
|
| 1613 |
+
"logps/rejected": -271.4120788574219,
|
| 1614 |
+
"loss": 1997.6766,
|
| 1615 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 1616 |
+
"rewards/chosen": -0.33833402395248413,
|
| 1617 |
+
"rewards/margins": 0.11596833169460297,
|
| 1618 |
+
"rewards/rejected": -0.4543024003505707,
|
| 1619 |
+
"step": 1140
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 0.6,
|
| 1623 |
+
"learning_rate": 2.0478932657817105e-06,
|
| 1624 |
+
"logits/chosen": 0.6026065349578857,
|
| 1625 |
+
"logits/rejected": 0.637354850769043,
|
| 1626 |
+
"logps/chosen": -318.32611083984375,
|
| 1627 |
+
"logps/rejected": -278.863037109375,
|
| 1628 |
+
"loss": 2020.9238,
|
| 1629 |
+
"rewards/accuracies": 0.625,
|
| 1630 |
+
"rewards/chosen": -0.3429223895072937,
|
| 1631 |
+
"rewards/margins": 0.10237185657024384,
|
| 1632 |
+
"rewards/rejected": -0.4452942907810211,
|
| 1633 |
+
"step": 1150
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 0.61,
|
| 1637 |
+
"learning_rate": 2.0030353425145376e-06,
|
| 1638 |
+
"logits/chosen": 0.5387909412384033,
|
| 1639 |
+
"logits/rejected": 0.5283801555633545,
|
| 1640 |
+
"logps/chosen": -281.98944091796875,
|
| 1641 |
+
"logps/rejected": -282.2474060058594,
|
| 1642 |
+
"loss": 2121.0545,
|
| 1643 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1644 |
+
"rewards/chosen": -0.40112051367759705,
|
| 1645 |
+
"rewards/margins": 0.060013122856616974,
|
| 1646 |
+
"rewards/rejected": -0.4611336290836334,
|
| 1647 |
+
"step": 1160
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.61,
|
| 1651 |
+
"learning_rate": 1.958343401405964e-06,
|
| 1652 |
+
"logits/chosen": 0.5407668352127075,
|
| 1653 |
+
"logits/rejected": 0.5168458223342896,
|
| 1654 |
+
"logps/chosen": -271.4942321777344,
|
| 1655 |
+
"logps/rejected": -262.18310546875,
|
| 1656 |
+
"loss": 1999.9113,
|
| 1657 |
+
"rewards/accuracies": 0.581250011920929,
|
| 1658 |
+
"rewards/chosen": -0.3525828719139099,
|
| 1659 |
+
"rewards/margins": 0.07946565002202988,
|
| 1660 |
+
"rewards/rejected": -0.4320485591888428,
|
| 1661 |
+
"step": 1170
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 0.62,
|
| 1665 |
+
"learning_rate": 1.9138323692012734e-06,
|
| 1666 |
+
"logits/chosen": 0.5828490257263184,
|
| 1667 |
+
"logits/rejected": 0.6618238091468811,
|
| 1668 |
+
"logps/chosen": -355.33355712890625,
|
| 1669 |
+
"logps/rejected": -293.50323486328125,
|
| 1670 |
+
"loss": 2052.1723,
|
| 1671 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 1672 |
+
"rewards/chosen": -0.3654106557369232,
|
| 1673 |
+
"rewards/margins": 0.09124873578548431,
|
| 1674 |
+
"rewards/rejected": -0.45665937662124634,
|
| 1675 |
+
"step": 1180
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 0.62,
|
| 1679 |
+
"learning_rate": 1.8695171122236443e-06,
|
| 1680 |
+
"logits/chosen": 0.5557527542114258,
|
| 1681 |
+
"logits/rejected": 0.6298555135726929,
|
| 1682 |
+
"logps/chosen": -312.16094970703125,
|
| 1683 |
+
"logps/rejected": -288.3011779785156,
|
| 1684 |
+
"loss": 1976.1777,
|
| 1685 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 1686 |
+
"rewards/chosen": -0.3318810760974884,
|
| 1687 |
+
"rewards/margins": 0.12243027985095978,
|
| 1688 |
+
"rewards/rejected": -0.4543113708496094,
|
| 1689 |
+
"step": 1190
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 0.63,
|
| 1693 |
+
"learning_rate": 1.8254124314089225e-06,
|
| 1694 |
+
"logits/chosen": 0.49655452370643616,
|
| 1695 |
+
"logits/rejected": 0.6068973541259766,
|
| 1696 |
+
"logps/chosen": -294.2212829589844,
|
| 1697 |
+
"logps/rejected": -318.2776794433594,
|
| 1698 |
+
"loss": 2004.0912,
|
| 1699 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 1700 |
+
"rewards/chosen": -0.33194708824157715,
|
| 1701 |
+
"rewards/margins": 0.09898443520069122,
|
| 1702 |
+
"rewards/rejected": -0.43093156814575195,
|
| 1703 |
+
"step": 1200
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 0.63,
|
| 1707 |
+
"learning_rate": 1.781533057362221e-06,
|
| 1708 |
+
"logits/chosen": 0.5596984028816223,
|
| 1709 |
+
"logits/rejected": 0.5853079557418823,
|
| 1710 |
+
"logps/chosen": -322.6697082519531,
|
| 1711 |
+
"logps/rejected": -272.7930603027344,
|
| 1712 |
+
"loss": 1955.9035,
|
| 1713 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 1714 |
+
"rewards/chosen": -0.369145929813385,
|
| 1715 |
+
"rewards/margins": 0.12208880484104156,
|
| 1716 |
+
"rewards/rejected": -0.491234689950943,
|
| 1717 |
+
"step": 1210
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 0.64,
|
| 1721 |
+
"learning_rate": 1.7378936454380277e-06,
|
| 1722 |
+
"logits/chosen": 0.5217611193656921,
|
| 1723 |
+
"logits/rejected": 0.5789054036140442,
|
| 1724 |
+
"logps/chosen": -336.248779296875,
|
| 1725 |
+
"logps/rejected": -306.63067626953125,
|
| 1726 |
+
"loss": 1938.6953,
|
| 1727 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 1728 |
+
"rewards/chosen": -0.3923150599002838,
|
| 1729 |
+
"rewards/margins": 0.09614621102809906,
|
| 1730 |
+
"rewards/rejected": -0.48846131563186646,
|
| 1731 |
+
"step": 1220
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 0.64,
|
| 1735 |
+
"learning_rate": 1.6945087708454273e-06,
|
| 1736 |
+
"logits/chosen": 0.5122044682502747,
|
| 1737 |
+
"logits/rejected": 0.54069584608078,
|
| 1738 |
+
"logps/chosen": -292.97711181640625,
|
| 1739 |
+
"logps/rejected": -275.40069580078125,
|
| 1740 |
+
"loss": 2029.4551,
|
| 1741 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 1742 |
+
"rewards/chosen": -0.4053107798099518,
|
| 1743 |
+
"rewards/margins": 0.08846473693847656,
|
| 1744 |
+
"rewards/rejected": -0.49377545714378357,
|
| 1745 |
+
"step": 1230
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 0.65,
|
| 1749 |
+
"learning_rate": 1.651392923780105e-06,
|
| 1750 |
+
"logits/chosen": 0.5897082090377808,
|
| 1751 |
+
"logits/rejected": 0.5889968276023865,
|
| 1752 |
+
"logps/chosen": -332.0108947753906,
|
| 1753 |
+
"logps/rejected": -292.29119873046875,
|
| 1754 |
+
"loss": 1970.8734,
|
| 1755 |
+
"rewards/accuracies": 0.606249988079071,
|
| 1756 |
+
"rewards/chosen": -0.38728633522987366,
|
| 1757 |
+
"rewards/margins": 0.11461566388607025,
|
| 1758 |
+
"rewards/rejected": -0.5019019842147827,
|
| 1759 |
+
"step": 1240
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 0.65,
|
| 1763 |
+
"learning_rate": 1.608560504584737e-06,
|
| 1764 |
+
"logits/chosen": 0.5301613807678223,
|
| 1765 |
+
"logits/rejected": 0.6047118902206421,
|
| 1766 |
+
"logps/chosen": -335.94598388671875,
|
| 1767 |
+
"logps/rejected": -308.5770263671875,
|
| 1768 |
+
"loss": 1999.6207,
|
| 1769 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 1770 |
+
"rewards/chosen": -0.41463202238082886,
|
| 1771 |
+
"rewards/margins": 0.1080860048532486,
|
| 1772 |
+
"rewards/rejected": -0.522718071937561,
|
| 1773 |
+
"step": 1250
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 0.66,
|
| 1777 |
+
"learning_rate": 1.5660258189393945e-06,
|
| 1778 |
+
"logits/chosen": 0.46664172410964966,
|
| 1779 |
+
"logits/rejected": 0.6063970923423767,
|
| 1780 |
+
"logps/chosen": -292.8914489746094,
|
| 1781 |
+
"logps/rejected": -287.7743835449219,
|
| 1782 |
+
"loss": 1998.873,
|
| 1783 |
+
"rewards/accuracies": 0.59375,
|
| 1784 |
+
"rewards/chosen": -0.41233953833580017,
|
| 1785 |
+
"rewards/margins": 0.0787600427865982,
|
| 1786 |
+
"rewards/rejected": -0.49109959602355957,
|
| 1787 |
+
"step": 1260
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 0.66,
|
| 1791 |
+
"learning_rate": 1.5238030730835578e-06,
|
| 1792 |
+
"logits/chosen": 0.5474873185157776,
|
| 1793 |
+
"logits/rejected": 0.6035341024398804,
|
| 1794 |
+
"logps/chosen": -316.2177734375,
|
| 1795 |
+
"logps/rejected": -293.90936279296875,
|
| 1796 |
+
"loss": 2059.1195,
|
| 1797 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 1798 |
+
"rewards/chosen": -0.4276704788208008,
|
| 1799 |
+
"rewards/margins": 0.09831003844738007,
|
| 1800 |
+
"rewards/rejected": -0.525980532169342,
|
| 1801 |
+
"step": 1270
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 0.67,
|
| 1805 |
+
"learning_rate": 1.4819063690713565e-06,
|
| 1806 |
+
"logits/chosen": 0.631860077381134,
|
| 1807 |
+
"logits/rejected": 0.604947566986084,
|
| 1808 |
+
"logps/chosen": -280.0582275390625,
|
| 1809 |
+
"logps/rejected": -259.1899108886719,
|
| 1810 |
+
"loss": 1954.3059,
|
| 1811 |
+
"rewards/accuracies": 0.5625,
|
| 1812 |
+
"rewards/chosen": -0.42631417512893677,
|
| 1813 |
+
"rewards/margins": 0.09757888317108154,
|
| 1814 |
+
"rewards/rejected": -0.5238930583000183,
|
| 1815 |
+
"step": 1280
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 0.68,
|
| 1819 |
+
"learning_rate": 1.4403497000615885e-06,
|
| 1820 |
+
"logits/chosen": 0.5674291253089905,
|
| 1821 |
+
"logits/rejected": 0.5523722767829895,
|
| 1822 |
+
"logps/chosen": -313.6990661621094,
|
| 1823 |
+
"logps/rejected": -303.3889465332031,
|
| 1824 |
+
"loss": 1995.9789,
|
| 1825 |
+
"rewards/accuracies": 0.612500011920929,
|
| 1826 |
+
"rewards/chosen": -0.4305340647697449,
|
| 1827 |
+
"rewards/margins": 0.10536874830722809,
|
| 1828 |
+
"rewards/rejected": -0.5359027981758118,
|
| 1829 |
+
"step": 1290
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 0.68,
|
| 1833 |
+
"learning_rate": 1.3991469456441273e-06,
|
| 1834 |
+
"logits/chosen": 0.5443872809410095,
|
| 1835 |
+
"logits/rejected": 0.629615843296051,
|
| 1836 |
+
"logps/chosen": -343.49072265625,
|
| 1837 |
+
"logps/rejected": -309.73480224609375,
|
| 1838 |
+
"loss": 1837.8883,
|
| 1839 |
+
"rewards/accuracies": 0.699999988079071,
|
| 1840 |
+
"rewards/chosen": -0.43309682607650757,
|
| 1841 |
+
"rewards/margins": 0.1318664252758026,
|
| 1842 |
+
"rewards/rejected": -0.5649632811546326,
|
| 1843 |
+
"step": 1300
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 0.69,
|
| 1847 |
+
"learning_rate": 1.3583118672042441e-06,
|
| 1848 |
+
"logits/chosen": 0.43076056241989136,
|
| 1849 |
+
"logits/rejected": 0.4598851799964905,
|
| 1850 |
+
"logps/chosen": -333.4667053222656,
|
| 1851 |
+
"logps/rejected": -277.95086669921875,
|
| 1852 |
+
"loss": 1996.2734,
|
| 1853 |
+
"rewards/accuracies": 0.637499988079071,
|
| 1854 |
+
"rewards/chosen": -0.44571250677108765,
|
| 1855 |
+
"rewards/margins": 0.10792438685894012,
|
| 1856 |
+
"rewards/rejected": -0.553636908531189,
|
| 1857 |
+
"step": 1310
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 0.69,
|
| 1861 |
+
"learning_rate": 1.3178581033264218e-06,
|
| 1862 |
+
"logits/chosen": 0.5987478494644165,
|
| 1863 |
+
"logits/rejected": 0.571110725402832,
|
| 1864 |
+
"logps/chosen": -277.62542724609375,
|
| 1865 |
+
"logps/rejected": -271.3750305175781,
|
| 1866 |
+
"loss": 2000.3379,
|
| 1867 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1868 |
+
"rewards/chosen": -0.4611617624759674,
|
| 1869 |
+
"rewards/margins": 0.09958922863006592,
|
| 1870 |
+
"rewards/rejected": -0.5607509613037109,
|
| 1871 |
+
"step": 1320
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 0.7,
|
| 1875 |
+
"learning_rate": 1.2777991652391757e-06,
|
| 1876 |
+
"logits/chosen": 0.5397824048995972,
|
| 1877 |
+
"logits/rejected": 0.5152544379234314,
|
| 1878 |
+
"logps/chosen": -285.2364501953125,
|
| 1879 |
+
"logps/rejected": -273.76214599609375,
|
| 1880 |
+
"loss": 1866.8488,
|
| 1881 |
+
"rewards/accuracies": 0.668749988079071,
|
| 1882 |
+
"rewards/chosen": -0.44938772916793823,
|
| 1883 |
+
"rewards/margins": 0.10914802551269531,
|
| 1884 |
+
"rewards/rejected": -0.5585357546806335,
|
| 1885 |
+
"step": 1330
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 0.7,
|
| 1889 |
+
"learning_rate": 1.2381484323024178e-06,
|
| 1890 |
+
"logits/chosen": 0.4960288107395172,
|
| 1891 |
+
"logits/rejected": 0.5434283018112183,
|
| 1892 |
+
"logps/chosen": -351.0420837402344,
|
| 1893 |
+
"logps/rejected": -310.51483154296875,
|
| 1894 |
+
"loss": 2042.0777,
|
| 1895 |
+
"rewards/accuracies": 0.675000011920929,
|
| 1896 |
+
"rewards/chosen": -0.45654526352882385,
|
| 1897 |
+
"rewards/margins": 0.11480891704559326,
|
| 1898 |
+
"rewards/rejected": -0.5713542103767395,
|
| 1899 |
+
"step": 1340
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 0.71,
|
| 1903 |
+
"learning_rate": 1.1989191475388518e-06,
|
| 1904 |
+
"logits/chosen": 0.5680415034294128,
|
| 1905 |
+
"logits/rejected": 0.5574635863304138,
|
| 1906 |
+
"logps/chosen": -313.9222106933594,
|
| 1907 |
+
"logps/rejected": -292.32843017578125,
|
| 1908 |
+
"loss": 2134.3227,
|
| 1909 |
+
"rewards/accuracies": 0.550000011920929,
|
| 1910 |
+
"rewards/chosen": -0.43487685918807983,
|
| 1911 |
+
"rewards/margins": 0.0749758630990982,
|
| 1912 |
+
"rewards/rejected": -0.5098527669906616,
|
| 1913 |
+
"step": 1350
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 0.71,
|
| 1917 |
+
"learning_rate": 1.160124413210918e-06,
|
| 1918 |
+
"logits/chosen": 0.49776411056518555,
|
| 1919 |
+
"logits/rejected": 0.5919879078865051,
|
| 1920 |
+
"logps/chosen": -329.8394775390625,
|
| 1921 |
+
"logps/rejected": -288.6294860839844,
|
| 1922 |
+
"loss": 1978.9926,
|
| 1923 |
+
"rewards/accuracies": 0.606249988079071,
|
| 1924 |
+
"rewards/chosen": -0.38458770513534546,
|
| 1925 |
+
"rewards/margins": 0.08599748462438583,
|
| 1926 |
+
"rewards/rejected": -0.4705851972103119,
|
| 1927 |
+
"step": 1360
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"epoch": 0.72,
|
| 1931 |
+
"learning_rate": 1.1217771864447396e-06,
|
| 1932 |
+
"logits/chosen": 0.4789052903652191,
|
| 1933 |
+
"logits/rejected": 0.5946076512336731,
|
| 1934 |
+
"logps/chosen": -343.82440185546875,
|
| 1935 |
+
"logps/rejected": -312.8480224609375,
|
| 1936 |
+
"loss": 1838.2496,
|
| 1937 |
+
"rewards/accuracies": 0.675000011920929,
|
| 1938 |
+
"rewards/chosen": -0.41200247406959534,
|
| 1939 |
+
"rewards/margins": 0.13190023601055145,
|
| 1940 |
+
"rewards/rejected": -0.5439027547836304,
|
| 1941 |
+
"step": 1370
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 0.72,
|
| 1945 |
+
"learning_rate": 1.08389027490255e-06,
|
| 1946 |
+
"logits/chosen": 0.518727719783783,
|
| 1947 |
+
"logits/rejected": 0.5448898673057556,
|
| 1948 |
+
"logps/chosen": -308.6346130371094,
|
| 1949 |
+
"logps/rejected": -262.6923828125,
|
| 1950 |
+
"loss": 2192.4686,
|
| 1951 |
+
"rewards/accuracies": 0.637499988079071,
|
| 1952 |
+
"rewards/chosen": -0.45256251096725464,
|
| 1953 |
+
"rewards/margins": 0.08636941015720367,
|
| 1954 |
+
"rewards/rejected": -0.5389319062232971,
|
| 1955 |
+
"step": 1380
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 0.73,
|
| 1959 |
+
"learning_rate": 1.046476332505036e-06,
|
| 1960 |
+
"logits/chosen": 0.5441064834594727,
|
| 1961 |
+
"logits/rejected": 0.6444369554519653,
|
| 1962 |
+
"logps/chosen": -301.1471862792969,
|
| 1963 |
+
"logps/rejected": -272.04296875,
|
| 1964 |
+
"loss": 2016.5557,
|
| 1965 |
+
"rewards/accuracies": 0.637499988079071,
|
| 1966 |
+
"rewards/chosen": -0.3905426263809204,
|
| 1967 |
+
"rewards/margins": 0.11363613605499268,
|
| 1968 |
+
"rewards/rejected": -0.5041787028312683,
|
| 1969 |
+
"step": 1390
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 0.73,
|
| 1973 |
+
"learning_rate": 1.0095478552050348e-06,
|
| 1974 |
+
"logits/chosen": 0.5234050750732422,
|
| 1975 |
+
"logits/rejected": 0.6232806444168091,
|
| 1976 |
+
"logps/chosen": -278.05462646484375,
|
| 1977 |
+
"logps/rejected": -237.6455078125,
|
| 1978 |
+
"loss": 1943.2496,
|
| 1979 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1980 |
+
"rewards/chosen": -0.4108065664768219,
|
| 1981 |
+
"rewards/margins": 0.09128499031066895,
|
| 1982 |
+
"rewards/rejected": -0.5020915269851685,
|
| 1983 |
+
"step": 1400
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 0.74,
|
| 1987 |
+
"learning_rate": 9.731171768139808e-07,
|
| 1988 |
+
"logits/chosen": 0.5350190997123718,
|
| 1989 |
+
"logits/rejected": 0.5485085248947144,
|
| 1990 |
+
"logps/chosen": -287.7547607421875,
|
| 1991 |
+
"logps/rejected": -263.4801025390625,
|
| 1992 |
+
"loss": 2019.2094,
|
| 1993 |
+
"rewards/accuracies": 0.625,
|
| 1994 |
+
"rewards/chosen": -0.3987637162208557,
|
| 1995 |
+
"rewards/margins": 0.07709892094135284,
|
| 1996 |
+
"rewards/rejected": -0.47586268186569214,
|
| 1997 |
+
"step": 1410
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 0.74,
|
| 2001 |
+
"learning_rate": 9.371964648825221e-07,
|
| 2002 |
+
"logits/chosen": 0.5190201997756958,
|
| 2003 |
+
"logits/rejected": 0.5698527693748474,
|
| 2004 |
+
"logps/chosen": -345.8193664550781,
|
| 2005 |
+
"logps/rejected": -291.22821044921875,
|
| 2006 |
+
"loss": 1915.1037,
|
| 2007 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 2008 |
+
"rewards/chosen": -0.39381805062294006,
|
| 2009 |
+
"rewards/margins": 0.11871011555194855,
|
| 2010 |
+
"rewards/rejected": -0.512528121471405,
|
| 2011 |
+
"step": 1420
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 0.75,
|
| 2015 |
+
"learning_rate": 9.017977166366445e-07,
|
| 2016 |
+
"logits/chosen": 0.5621356964111328,
|
| 2017 |
+
"logits/rejected": 0.574070155620575,
|
| 2018 |
+
"logps/chosen": -301.27276611328125,
|
| 2019 |
+
"logps/rejected": -268.485595703125,
|
| 2020 |
+
"loss": 2027.6629,
|
| 2021 |
+
"rewards/accuracies": 0.606249988079071,
|
| 2022 |
+
"rewards/chosen": -0.4134562611579895,
|
| 2023 |
+
"rewards/margins": 0.08553650230169296,
|
| 2024 |
+
"rewards/rejected": -0.49899277091026306,
|
| 2025 |
+
"step": 1430
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 0.75,
|
| 2029 |
+
"learning_rate": 8.669327549707096e-07,
|
| 2030 |
+
"logits/chosen": 0.5577922463417053,
|
| 2031 |
+
"logits/rejected": 0.6254245042800903,
|
| 2032 |
+
"logps/chosen": -307.9657897949219,
|
| 2033 |
+
"logps/rejected": -278.01068115234375,
|
| 2034 |
+
"loss": 1957.0965,
|
| 2035 |
+
"rewards/accuracies": 0.65625,
|
| 2036 |
+
"rewards/chosen": -0.38403937220573425,
|
| 2037 |
+
"rewards/margins": 0.10890159755945206,
|
| 2038 |
+
"rewards/rejected": -0.4929409921169281,
|
| 2039 |
+
"step": 1440
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 0.76,
|
| 2043 |
+
"learning_rate": 8.326132244986932e-07,
|
| 2044 |
+
"logits/chosen": 0.5165051817893982,
|
| 2045 |
+
"logits/rejected": 0.463472455739975,
|
| 2046 |
+
"logps/chosen": -325.6259460449219,
|
| 2047 |
+
"logps/rejected": -282.59246826171875,
|
| 2048 |
+
"loss": 1931.7,
|
| 2049 |
+
"rewards/accuracies": 0.668749988079071,
|
| 2050 |
+
"rewards/chosen": -0.3894692063331604,
|
| 2051 |
+
"rewards/margins": 0.11373750865459442,
|
| 2052 |
+
"rewards/rejected": -0.5032067894935608,
|
| 2053 |
+
"step": 1450
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 0.76,
|
| 2057 |
+
"learning_rate": 7.988505876649863e-07,
|
| 2058 |
+
"logits/chosen": 0.5601966977119446,
|
| 2059 |
+
"logits/rejected": 0.5657812356948853,
|
| 2060 |
+
"logps/chosen": -329.68304443359375,
|
| 2061 |
+
"logps/rejected": -279.58612060546875,
|
| 2062 |
+
"loss": 2044.8211,
|
| 2063 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 2064 |
+
"rewards/chosen": -0.4129094183444977,
|
| 2065 |
+
"rewards/margins": 0.08905430138111115,
|
| 2066 |
+
"rewards/rejected": -0.5019636750221252,
|
| 2067 |
+
"step": 1460
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 0.77,
|
| 2071 |
+
"learning_rate": 7.656561209160248e-07,
|
| 2072 |
+
"logits/chosen": 0.5333635210990906,
|
| 2073 |
+
"logits/rejected": 0.6559829711914062,
|
| 2074 |
+
"logps/chosen": -334.6247863769531,
|
| 2075 |
+
"logps/rejected": -293.94403076171875,
|
| 2076 |
+
"loss": 1924.4891,
|
| 2077 |
+
"rewards/accuracies": 0.668749988079071,
|
| 2078 |
+
"rewards/chosen": -0.3982022702693939,
|
| 2079 |
+
"rewards/margins": 0.12732462584972382,
|
| 2080 |
+
"rewards/rejected": -0.5255268812179565,
|
| 2081 |
+
"step": 1470
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 0.77,
|
| 2085 |
+
"learning_rate": 7.330409109340563e-07,
|
| 2086 |
+
"logits/chosen": 0.44074922800064087,
|
| 2087 |
+
"logits/rejected": 0.5341771245002747,
|
| 2088 |
+
"logps/chosen": -325.88763427734375,
|
| 2089 |
+
"logps/rejected": -301.38427734375,
|
| 2090 |
+
"loss": 2064.7137,
|
| 2091 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 2092 |
+
"rewards/chosen": -0.39723503589630127,
|
| 2093 |
+
"rewards/margins": 0.09469970315694809,
|
| 2094 |
+
"rewards/rejected": -0.49193471670150757,
|
| 2095 |
+
"step": 1480
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 0.78,
|
| 2099 |
+
"learning_rate": 7.010158509342682e-07,
|
| 2100 |
+
"logits/chosen": 0.6108728647232056,
|
| 2101 |
+
"logits/rejected": 0.5816964507102966,
|
| 2102 |
+
"logps/chosen": -280.2751159667969,
|
| 2103 |
+
"logps/rejected": -272.6864929199219,
|
| 2104 |
+
"loss": 2012.7195,
|
| 2105 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 2106 |
+
"rewards/chosen": -0.39345765113830566,
|
| 2107 |
+
"rewards/margins": 0.10830279439687729,
|
| 2108 |
+
"rewards/rejected": -0.5017604827880859,
|
| 2109 |
+
"step": 1490
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 0.79,
|
| 2113 |
+
"learning_rate": 6.695916370265529e-07,
|
| 2114 |
+
"logits/chosen": 0.5402048826217651,
|
| 2115 |
+
"logits/rejected": 0.5511065721511841,
|
| 2116 |
+
"logps/chosen": -293.0211486816406,
|
| 2117 |
+
"logps/rejected": -299.3897399902344,
|
| 2118 |
+
"loss": 2051.2686,
|
| 2119 |
+
"rewards/accuracies": 0.637499988079071,
|
| 2120 |
+
"rewards/chosen": -0.38481405377388,
|
| 2121 |
+
"rewards/margins": 0.09063725918531418,
|
| 2122 |
+
"rewards/rejected": -0.4754512906074524,
|
| 2123 |
+
"step": 1500
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 0.79,
|
| 2127 |
+
"learning_rate": 6.387787646430854e-07,
|
| 2128 |
+
"logits/chosen": 0.4866926670074463,
|
| 2129 |
+
"logits/rejected": 0.553752064704895,
|
| 2130 |
+
"logps/chosen": -323.8518371582031,
|
| 2131 |
+
"logps/rejected": -315.7661437988281,
|
| 2132 |
+
"loss": 1892.2789,
|
| 2133 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 2134 |
+
"rewards/chosen": -0.40690216422080994,
|
| 2135 |
+
"rewards/margins": 0.1288827359676361,
|
| 2136 |
+
"rewards/rejected": -0.535784900188446,
|
| 2137 |
+
"step": 1510
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 0.8,
|
| 2141 |
+
"learning_rate": 6.085875250329401e-07,
|
| 2142 |
+
"logits/chosen": 0.5334519743919373,
|
| 2143 |
+
"logits/rejected": 0.629740834236145,
|
| 2144 |
+
"logps/chosen": -342.43341064453125,
|
| 2145 |
+
"logps/rejected": -304.27178955078125,
|
| 2146 |
+
"loss": 1996.1777,
|
| 2147 |
+
"rewards/accuracies": 0.59375,
|
| 2148 |
+
"rewards/chosen": -0.4247370660305023,
|
| 2149 |
+
"rewards/margins": 0.07812504470348358,
|
| 2150 |
+
"rewards/rejected": -0.5028620958328247,
|
| 2151 |
+
"step": 1520
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 0.8,
|
| 2155 |
+
"learning_rate": 5.79028001824894e-07,
|
| 2156 |
+
"logits/chosen": 0.536971926689148,
|
| 2157 |
+
"logits/rejected": 0.6389668583869934,
|
| 2158 |
+
"logps/chosen": -344.0328063964844,
|
| 2159 |
+
"logps/rejected": -307.81195068359375,
|
| 2160 |
+
"loss": 2099.0133,
|
| 2161 |
+
"rewards/accuracies": 0.65625,
|
| 2162 |
+
"rewards/chosen": -0.4052516520023346,
|
| 2163 |
+
"rewards/margins": 0.1204022616147995,
|
| 2164 |
+
"rewards/rejected": -0.5256539583206177,
|
| 2165 |
+
"step": 1530
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 0.81,
|
| 2169 |
+
"learning_rate": 5.501100676595761e-07,
|
| 2170 |
+
"logits/chosen": 0.501252293586731,
|
| 2171 |
+
"logits/rejected": 0.582123875617981,
|
| 2172 |
+
"logps/chosen": -350.5921936035156,
|
| 2173 |
+
"logps/rejected": -305.71551513671875,
|
| 2174 |
+
"loss": 2080.3439,
|
| 2175 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 2176 |
+
"rewards/chosen": -0.386319637298584,
|
| 2177 |
+
"rewards/margins": 0.1132698804140091,
|
| 2178 |
+
"rewards/rejected": -0.4995895326137543,
|
| 2179 |
+
"step": 1540
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 0.81,
|
| 2183 |
+
"learning_rate": 5.218433808920884e-07,
|
| 2184 |
+
"logits/chosen": 0.5797770619392395,
|
| 2185 |
+
"logits/rejected": 0.5843828916549683,
|
| 2186 |
+
"logps/chosen": -311.27313232421875,
|
| 2187 |
+
"logps/rejected": -272.329833984375,
|
| 2188 |
+
"loss": 1840.9139,
|
| 2189 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 2190 |
+
"rewards/chosen": -0.4069671630859375,
|
| 2191 |
+
"rewards/margins": 0.10061927884817123,
|
| 2192 |
+
"rewards/rejected": -0.5075864195823669,
|
| 2193 |
+
"step": 1550
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 0.82,
|
| 2197 |
+
"learning_rate": 4.942373823661928e-07,
|
| 2198 |
+
"logits/chosen": 0.4943598806858063,
|
| 2199 |
+
"logits/rejected": 0.5981740355491638,
|
| 2200 |
+
"logps/chosen": -339.5876770019531,
|
| 2201 |
+
"logps/rejected": -323.6890563964844,
|
| 2202 |
+
"loss": 2091.2941,
|
| 2203 |
+
"rewards/accuracies": 0.606249988079071,
|
| 2204 |
+
"rewards/chosen": -0.42342454195022583,
|
| 2205 |
+
"rewards/margins": 0.09093138575553894,
|
| 2206 |
+
"rewards/rejected": -0.5143559575080872,
|
| 2207 |
+
"step": 1560
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 0.82,
|
| 2211 |
+
"learning_rate": 4.6730129226114363e-07,
|
| 2212 |
+
"logits/chosen": 0.5075241327285767,
|
| 2213 |
+
"logits/rejected": 0.5773884654045105,
|
| 2214 |
+
"logps/chosen": -301.22906494140625,
|
| 2215 |
+
"logps/rejected": -278.94293212890625,
|
| 2216 |
+
"loss": 2206.433,
|
| 2217 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 2218 |
+
"rewards/chosen": -0.4186634421348572,
|
| 2219 |
+
"rewards/margins": 0.08012911677360535,
|
| 2220 |
+
"rewards/rejected": -0.4987925887107849,
|
| 2221 |
+
"step": 1570
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 0.83,
|
| 2225 |
+
"learning_rate": 4.4104410701222703e-07,
|
| 2226 |
+
"logits/chosen": 0.47776398062705994,
|
| 2227 |
+
"logits/rejected": 0.6287192106246948,
|
| 2228 |
+
"logps/chosen": -321.4864196777344,
|
| 2229 |
+
"logps/rejected": -324.2401428222656,
|
| 2230 |
+
"loss": 2054.8516,
|
| 2231 |
+
"rewards/accuracies": 0.643750011920929,
|
| 2232 |
+
"rewards/chosen": -0.4205222725868225,
|
| 2233 |
+
"rewards/margins": 0.08562308549880981,
|
| 2234 |
+
"rewards/rejected": -0.5061453580856323,
|
| 2235 |
+
"step": 1580
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 0.83,
|
| 2239 |
+
"learning_rate": 4.154745963060197e-07,
|
| 2240 |
+
"logits/chosen": 0.5759707093238831,
|
| 2241 |
+
"logits/rejected": 0.6366346478462219,
|
| 2242 |
+
"logps/chosen": -310.05487060546875,
|
| 2243 |
+
"logps/rejected": -309.42303466796875,
|
| 2244 |
+
"loss": 1960.8133,
|
| 2245 |
+
"rewards/accuracies": 0.668749988079071,
|
| 2246 |
+
"rewards/chosen": -0.40062469244003296,
|
| 2247 |
+
"rewards/margins": 0.12129988521337509,
|
| 2248 |
+
"rewards/rejected": -0.521924614906311,
|
| 2249 |
+
"step": 1590
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 0.84,
|
| 2253 |
+
"learning_rate": 3.9060130015138863e-07,
|
| 2254 |
+
"logits/chosen": 0.5741788148880005,
|
| 2255 |
+
"logits/rejected": 0.6168212294578552,
|
| 2256 |
+
"logps/chosen": -313.01531982421875,
|
| 2257 |
+
"logps/rejected": -290.6816101074219,
|
| 2258 |
+
"loss": 1803.5051,
|
| 2259 |
+
"rewards/accuracies": 0.668749988079071,
|
| 2260 |
+
"rewards/chosen": -0.39632636308670044,
|
| 2261 |
+
"rewards/margins": 0.0985620766878128,
|
| 2262 |
+
"rewards/rejected": -0.49488845467567444,
|
| 2263 |
+
"step": 1600
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 0.84,
|
| 2267 |
+
"learning_rate": 3.664325260271953e-07,
|
| 2268 |
+
"logits/chosen": 0.5977301001548767,
|
| 2269 |
+
"logits/rejected": 0.5980079770088196,
|
| 2270 |
+
"logps/chosen": -305.8802185058594,
|
| 2271 |
+
"logps/rejected": -245.7880859375,
|
| 2272 |
+
"loss": 1887.7617,
|
| 2273 |
+
"rewards/accuracies": 0.59375,
|
| 2274 |
+
"rewards/chosen": -0.38649195432662964,
|
| 2275 |
+
"rewards/margins": 0.1288624256849289,
|
| 2276 |
+
"rewards/rejected": -0.5153544545173645,
|
| 2277 |
+
"step": 1610
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 0.85,
|
| 2281 |
+
"learning_rate": 3.429763461076677e-07,
|
| 2282 |
+
"logits/chosen": 0.4545535147190094,
|
| 2283 |
+
"logits/rejected": 0.656732439994812,
|
| 2284 |
+
"logps/chosen": -348.40594482421875,
|
| 2285 |
+
"logps/rejected": -304.6084899902344,
|
| 2286 |
+
"loss": 2087.083,
|
| 2287 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 2288 |
+
"rewards/chosen": -0.39987581968307495,
|
| 2289 |
+
"rewards/margins": 0.09592723846435547,
|
| 2290 |
+
"rewards/rejected": -0.4958030581474304,
|
| 2291 |
+
"step": 1620
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 0.85,
|
| 2295 |
+
"learning_rate": 3.202405945663556e-07,
|
| 2296 |
+
"logits/chosen": 0.5257354974746704,
|
| 2297 |
+
"logits/rejected": 0.5875495076179504,
|
| 2298 |
+
"logps/chosen": -321.76397705078125,
|
| 2299 |
+
"logps/rejected": -286.9283142089844,
|
| 2300 |
+
"loss": 2075.5064,
|
| 2301 |
+
"rewards/accuracies": 0.5625,
|
| 2302 |
+
"rewards/chosen": -0.39552539587020874,
|
| 2303 |
+
"rewards/margins": 0.10027774423360825,
|
| 2304 |
+
"rewards/rejected": -0.4958031177520752,
|
| 2305 |
+
"step": 1630
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 0.86,
|
| 2309 |
+
"learning_rate": 2.982328649595856e-07,
|
| 2310 |
+
"logits/chosen": 0.4653220772743225,
|
| 2311 |
+
"logits/rejected": 0.5480079650878906,
|
| 2312 |
+
"logps/chosen": -309.8276672363281,
|
| 2313 |
+
"logps/rejected": -274.49029541015625,
|
| 2314 |
+
"loss": 1948.8629,
|
| 2315 |
+
"rewards/accuracies": 0.65625,
|
| 2316 |
+
"rewards/chosen": -0.3729521930217743,
|
| 2317 |
+
"rewards/margins": 0.1282852441072464,
|
| 2318 |
+
"rewards/rejected": -0.5012374520301819,
|
| 2319 |
+
"step": 1640
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 0.86,
|
| 2323 |
+
"learning_rate": 2.7696050769026954e-07,
|
| 2324 |
+
"logits/chosen": 0.5885381698608398,
|
| 2325 |
+
"logits/rejected": 0.5182594060897827,
|
| 2326 |
+
"logps/chosen": -282.14544677734375,
|
| 2327 |
+
"logps/rejected": -316.86639404296875,
|
| 2328 |
+
"loss": 1897.5818,
|
| 2329 |
+
"rewards/accuracies": 0.668749988079071,
|
| 2330 |
+
"rewards/chosen": -0.4102315902709961,
|
| 2331 |
+
"rewards/margins": 0.10423590242862701,
|
| 2332 |
+
"rewards/rejected": -0.5144674777984619,
|
| 2333 |
+
"step": 1650
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 0.87,
|
| 2337 |
+
"learning_rate": 2.564306275529341e-07,
|
| 2338 |
+
"logits/chosen": 0.5538032650947571,
|
| 2339 |
+
"logits/rejected": 0.6108819246292114,
|
| 2340 |
+
"logps/chosen": -317.8446960449219,
|
| 2341 |
+
"logps/rejected": -322.9876403808594,
|
| 2342 |
+
"loss": 1870.0818,
|
| 2343 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 2344 |
+
"rewards/chosen": -0.40971916913986206,
|
| 2345 |
+
"rewards/margins": 0.10465312004089355,
|
| 2346 |
+
"rewards/rejected": -0.5143723487854004,
|
| 2347 |
+
"step": 1660
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 0.87,
|
| 2351 |
+
"learning_rate": 2.3665008136077332e-07,
|
| 2352 |
+
"logits/chosen": 0.5508009791374207,
|
| 2353 |
+
"logits/rejected": 0.6037713885307312,
|
| 2354 |
+
"logps/chosen": -357.86907958984375,
|
| 2355 |
+
"logps/rejected": -307.35723876953125,
|
| 2356 |
+
"loss": 2033.6316,
|
| 2357 |
+
"rewards/accuracies": 0.625,
|
| 2358 |
+
"rewards/chosen": -0.4212369918823242,
|
| 2359 |
+
"rewards/margins": 0.09935277700424194,
|
| 2360 |
+
"rewards/rejected": -0.5205897688865662,
|
| 2361 |
+
"step": 1670
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 0.88,
|
| 2365 |
+
"learning_rate": 2.1762547565553293e-07,
|
| 2366 |
+
"logits/chosen": 0.5105335712432861,
|
| 2367 |
+
"logits/rejected": 0.5201815366744995,
|
| 2368 |
+
"logps/chosen": -263.2312316894531,
|
| 2369 |
+
"logps/rejected": -266.52801513671875,
|
| 2370 |
+
"loss": 2212.5818,
|
| 2371 |
+
"rewards/accuracies": 0.5,
|
| 2372 |
+
"rewards/chosen": -0.4164590835571289,
|
| 2373 |
+
"rewards/margins": 0.04953977093100548,
|
| 2374 |
+
"rewards/rejected": -0.4659988284111023,
|
| 2375 |
+
"step": 1680
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 0.88,
|
| 2379 |
+
"learning_rate": 1.993631645009747e-07,
|
| 2380 |
+
"logits/chosen": 0.5417348742485046,
|
| 2381 |
+
"logits/rejected": 0.5702573657035828,
|
| 2382 |
+
"logps/chosen": -342.66839599609375,
|
| 2383 |
+
"logps/rejected": -324.40557861328125,
|
| 2384 |
+
"loss": 2001.3066,
|
| 2385 |
+
"rewards/accuracies": 0.65625,
|
| 2386 |
+
"rewards/chosen": -0.42196089029312134,
|
| 2387 |
+
"rewards/margins": 0.09920088946819305,
|
| 2388 |
+
"rewards/rejected": -0.5211617350578308,
|
| 2389 |
+
"step": 1690
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 0.89,
|
| 2393 |
+
"learning_rate": 1.818692473606748e-07,
|
| 2394 |
+
"logits/chosen": 0.5118446350097656,
|
| 2395 |
+
"logits/rejected": 0.6164907217025757,
|
| 2396 |
+
"logps/chosen": -285.0245056152344,
|
| 2397 |
+
"logps/rejected": -270.7235412597656,
|
| 2398 |
+
"loss": 1936.7527,
|
| 2399 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 2400 |
+
"rewards/chosen": -0.39391833543777466,
|
| 2401 |
+
"rewards/margins": 0.10235867649316788,
|
| 2402 |
+
"rewards/rejected": -0.49627700448036194,
|
| 2403 |
+
"step": 1700
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 0.9,
|
| 2407 |
+
"learning_rate": 1.6514956706084885e-07,
|
| 2408 |
+
"logits/chosen": 0.5716298818588257,
|
| 2409 |
+
"logits/rejected": 0.533431351184845,
|
| 2410 |
+
"logps/chosen": -298.3558654785156,
|
| 2411 |
+
"logps/rejected": -315.1064453125,
|
| 2412 |
+
"loss": 2030.7699,
|
| 2413 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 2414 |
+
"rewards/chosen": -0.444709837436676,
|
| 2415 |
+
"rewards/margins": 0.07402367889881134,
|
| 2416 |
+
"rewards/rejected": -0.5187335014343262,
|
| 2417 |
+
"step": 1710
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 0.9,
|
| 2421 |
+
"learning_rate": 1.4920970783889737e-07,
|
| 2422 |
+
"logits/chosen": 0.41161495447158813,
|
| 2423 |
+
"logits/rejected": 0.5431499481201172,
|
| 2424 |
+
"logps/chosen": -335.1160583496094,
|
| 2425 |
+
"logps/rejected": -285.2781677246094,
|
| 2426 |
+
"loss": 2117.8803,
|
| 2427 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 2428 |
+
"rewards/chosen": -0.382477343082428,
|
| 2429 |
+
"rewards/margins": 0.10167716443538666,
|
| 2430 |
+
"rewards/rejected": -0.48415452241897583,
|
| 2431 |
+
"step": 1720
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 0.91,
|
| 2435 |
+
"learning_rate": 1.340549934783164e-07,
|
| 2436 |
+
"logits/chosen": 0.5718288421630859,
|
| 2437 |
+
"logits/rejected": 0.5717177987098694,
|
| 2438 |
+
"logps/chosen": -286.6661682128906,
|
| 2439 |
+
"logps/rejected": -286.0840759277344,
|
| 2440 |
+
"loss": 2009.7559,
|
| 2441 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 2442 |
+
"rewards/chosen": -0.40986448526382446,
|
| 2443 |
+
"rewards/margins": 0.08955219388008118,
|
| 2444 |
+
"rewards/rejected": -0.49941664934158325,
|
| 2445 |
+
"step": 1730
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 0.91,
|
| 2449 |
+
"learning_rate": 1.196904855305961e-07,
|
| 2450 |
+
"logits/chosen": 0.5395318865776062,
|
| 2451 |
+
"logits/rejected": 0.6408040523529053,
|
| 2452 |
+
"logps/chosen": -318.8950500488281,
|
| 2453 |
+
"logps/rejected": -298.7578125,
|
| 2454 |
+
"loss": 2034.2352,
|
| 2455 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 2456 |
+
"rewards/chosen": -0.4017771780490875,
|
| 2457 |
+
"rewards/margins": 0.11808891594409943,
|
| 2458 |
+
"rewards/rejected": -0.5198661088943481,
|
| 2459 |
+
"step": 1740
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 0.92,
|
| 2463 |
+
"learning_rate": 1.0612098162470302e-07,
|
| 2464 |
+
"logits/chosen": 0.4985600411891937,
|
| 2465 |
+
"logits/rejected": 0.5432295799255371,
|
| 2466 |
+
"logps/chosen": -286.3262634277344,
|
| 2467 |
+
"logps/rejected": -289.0979919433594,
|
| 2468 |
+
"loss": 1954.6828,
|
| 2469 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 2470 |
+
"rewards/chosen": -0.39057815074920654,
|
| 2471 |
+
"rewards/margins": 0.0826260969042778,
|
| 2472 |
+
"rewards/rejected": -0.47320422530174255,
|
| 2473 |
+
"step": 1750
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 0.92,
|
| 2477 |
+
"learning_rate": 9.335101386471285e-08,
|
| 2478 |
+
"logits/chosen": 0.5725753307342529,
|
| 2479 |
+
"logits/rejected": 0.6116470098495483,
|
| 2480 |
+
"logps/chosen": -335.08843994140625,
|
| 2481 |
+
"logps/rejected": -312.4408264160156,
|
| 2482 |
+
"loss": 2050.2887,
|
| 2483 |
+
"rewards/accuracies": 0.668749988079071,
|
| 2484 |
+
"rewards/chosen": -0.40838685631752014,
|
| 2485 |
+
"rewards/margins": 0.0921243280172348,
|
| 2486 |
+
"rewards/rejected": -0.5005111694335938,
|
| 2487 |
+
"step": 1760
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 0.93,
|
| 2491 |
+
"learning_rate": 8.138484731612273e-08,
|
| 2492 |
+
"logits/chosen": 0.6018707156181335,
|
| 2493 |
+
"logits/rejected": 0.5222383141517639,
|
| 2494 |
+
"logps/chosen": -253.42379760742188,
|
| 2495 |
+
"logps/rejected": -274.9010925292969,
|
| 2496 |
+
"loss": 2175.2457,
|
| 2497 |
+
"rewards/accuracies": 0.574999988079071,
|
| 2498 |
+
"rewards/chosen": -0.3636482059955597,
|
| 2499 |
+
"rewards/margins": 0.072720468044281,
|
| 2500 |
+
"rewards/rejected": -0.4363686442375183,
|
| 2501 |
+
"step": 1770
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 0.93,
|
| 2505 |
+
"learning_rate": 7.022647858135501e-08,
|
| 2506 |
+
"logits/chosen": 0.5884579420089722,
|
| 2507 |
+
"logits/rejected": 0.5674210786819458,
|
| 2508 |
+
"logps/chosen": -338.01837158203125,
|
| 2509 |
+
"logps/rejected": -324.35980224609375,
|
| 2510 |
+
"loss": 2028.5854,
|
| 2511 |
+
"rewards/accuracies": 0.625,
|
| 2512 |
+
"rewards/chosen": -0.42924776673316956,
|
| 2513 |
+
"rewards/margins": 0.09629428386688232,
|
| 2514 |
+
"rewards/rejected": -0.5255420804023743,
|
| 2515 |
+
"step": 1780
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 0.94,
|
| 2519 |
+
"learning_rate": 5.987963446492384e-08,
|
| 2520 |
+
"logits/chosen": 0.5171129703521729,
|
| 2521 |
+
"logits/rejected": 0.6137627363204956,
|
| 2522 |
+
"logps/chosen": -315.68609619140625,
|
| 2523 |
+
"logps/rejected": -290.91619873046875,
|
| 2524 |
+
"loss": 2091.5113,
|
| 2525 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 2526 |
+
"rewards/chosen": -0.41310811042785645,
|
| 2527 |
+
"rewards/margins": 0.09795816987752914,
|
| 2528 |
+
"rewards/rejected": -0.5110663175582886,
|
| 2529 |
+
"step": 1790
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 0.94,
|
| 2533 |
+
"learning_rate": 5.034777072871394e-08,
|
| 2534 |
+
"logits/chosen": 0.5506697297096252,
|
| 2535 |
+
"logits/rejected": 0.5700551867485046,
|
| 2536 |
+
"logps/chosen": -306.2592468261719,
|
| 2537 |
+
"logps/rejected": -284.780029296875,
|
| 2538 |
+
"loss": 2043.7551,
|
| 2539 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 2540 |
+
"rewards/chosen": -0.3679637312889099,
|
| 2541 |
+
"rewards/margins": 0.10389117151498795,
|
| 2542 |
+
"rewards/rejected": -0.4718549847602844,
|
| 2543 |
+
"step": 1800
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 0.95,
|
| 2547 |
+
"learning_rate": 4.163407093778243e-08,
|
| 2548 |
+
"logits/chosen": 0.4897570013999939,
|
| 2549 |
+
"logits/rejected": 0.565123438835144,
|
| 2550 |
+
"logps/chosen": -326.5579833984375,
|
| 2551 |
+
"logps/rejected": -310.0860595703125,
|
| 2552 |
+
"loss": 1972.2996,
|
| 2553 |
+
"rewards/accuracies": 0.612500011920929,
|
| 2554 |
+
"rewards/chosen": -0.4172098636627197,
|
| 2555 |
+
"rewards/margins": 0.09725239127874374,
|
| 2556 |
+
"rewards/rejected": -0.5144622921943665,
|
| 2557 |
+
"step": 1810
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 0.95,
|
| 2561 |
+
"learning_rate": 3.37414453970758e-08,
|
| 2562 |
+
"logits/chosen": 0.4947708249092102,
|
| 2563 |
+
"logits/rejected": 0.6116417050361633,
|
| 2564 |
+
"logps/chosen": -369.77789306640625,
|
| 2565 |
+
"logps/rejected": -310.3260192871094,
|
| 2566 |
+
"loss": 1986.4135,
|
| 2567 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 2568 |
+
"rewards/chosen": -0.3952573835849762,
|
| 2569 |
+
"rewards/margins": 0.10770855844020844,
|
| 2570 |
+
"rewards/rejected": -0.5029659271240234,
|
| 2571 |
+
"step": 1820
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 0.96,
|
| 2575 |
+
"learning_rate": 2.6672530179410183e-08,
|
| 2576 |
+
"logits/chosen": 0.49206972122192383,
|
| 2577 |
+
"logits/rejected": 0.6127623915672302,
|
| 2578 |
+
"logps/chosen": -326.89813232421875,
|
| 2579 |
+
"logps/rejected": -278.120849609375,
|
| 2580 |
+
"loss": 1947.9119,
|
| 2581 |
+
"rewards/accuracies": 0.625,
|
| 2582 |
+
"rewards/chosen": -0.4299922585487366,
|
| 2583 |
+
"rewards/margins": 0.10199449956417084,
|
| 2584 |
+
"rewards/rejected": -0.5319867730140686,
|
| 2585 |
+
"step": 1830
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 0.96,
|
| 2589 |
+
"learning_rate": 2.04296862450451e-08,
|
| 2590 |
+
"logits/chosen": 0.5761113166809082,
|
| 2591 |
+
"logits/rejected": 0.5587304830551147,
|
| 2592 |
+
"logps/chosen": -328.021240234375,
|
| 2593 |
+
"logps/rejected": -305.7353210449219,
|
| 2594 |
+
"loss": 1986.5195,
|
| 2595 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 2596 |
+
"rewards/chosen": -0.4139803349971771,
|
| 2597 |
+
"rewards/margins": 0.09623098373413086,
|
| 2598 |
+
"rewards/rejected": -0.5102113485336304,
|
| 2599 |
+
"step": 1840
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 0.97,
|
| 2603 |
+
"learning_rate": 1.501499865314171e-08,
|
| 2604 |
+
"logits/chosen": 0.533464252948761,
|
| 2605 |
+
"logits/rejected": 0.5722233057022095,
|
| 2606 |
+
"logps/chosen": -323.434814453125,
|
| 2607 |
+
"logps/rejected": -295.4156188964844,
|
| 2608 |
+
"loss": 1949.7422,
|
| 2609 |
+
"rewards/accuracies": 0.59375,
|
| 2610 |
+
"rewards/chosen": -0.3916718363761902,
|
| 2611 |
+
"rewards/margins": 0.10514561086893082,
|
| 2612 |
+
"rewards/rejected": -0.4968174397945404,
|
| 2613 |
+
"step": 1850
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 0.97,
|
| 2617 |
+
"learning_rate": 1.0430275865371265e-08,
|
| 2618 |
+
"logits/chosen": 0.5557342171669006,
|
| 2619 |
+
"logits/rejected": 0.5775563716888428,
|
| 2620 |
+
"logps/chosen": -320.87158203125,
|
| 2621 |
+
"logps/rejected": -306.94561767578125,
|
| 2622 |
+
"loss": 1832.5422,
|
| 2623 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 2624 |
+
"rewards/chosen": -0.39497238397598267,
|
| 2625 |
+
"rewards/margins": 0.13699549436569214,
|
| 2626 |
+
"rewards/rejected": -0.53196781873703,
|
| 2627 |
+
"step": 1860
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 0.98,
|
| 2631 |
+
"learning_rate": 6.677049141901315e-09,
|
| 2632 |
+
"logits/chosen": 0.5044248104095459,
|
| 2633 |
+
"logits/rejected": 0.5328904986381531,
|
| 2634 |
+
"logps/chosen": -296.86865234375,
|
| 2635 |
+
"logps/rejected": -287.9560546875,
|
| 2636 |
+
"loss": 1983.6154,
|
| 2637 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 2638 |
+
"rewards/chosen": -0.39695119857788086,
|
| 2639 |
+
"rewards/margins": 0.09827554225921631,
|
| 2640 |
+
"rewards/rejected": -0.49522677063941956,
|
| 2641 |
+
"step": 1870
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 0.98,
|
| 2645 |
+
"learning_rate": 3.756572029968708e-09,
|
| 2646 |
+
"logits/chosen": 0.481318861246109,
|
| 2647 |
+
"logits/rejected": 0.6070166230201721,
|
| 2648 |
+
"logps/chosen": -329.2176818847656,
|
| 2649 |
+
"logps/rejected": -315.81561279296875,
|
| 2650 |
+
"loss": 1846.1697,
|
| 2651 |
+
"rewards/accuracies": 0.668749988079071,
|
| 2652 |
+
"rewards/chosen": -0.404033899307251,
|
| 2653 |
+
"rewards/margins": 0.11680416762828827,
|
| 2654 |
+
"rewards/rejected": -0.5208381414413452,
|
| 2655 |
+
"step": 1880
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 0.99,
|
| 2659 |
+
"learning_rate": 1.6698199452053199e-09,
|
| 2660 |
+
"logits/chosen": 0.5581511855125427,
|
| 2661 |
+
"logits/rejected": 0.5964113473892212,
|
| 2662 |
+
"logps/chosen": -296.1249694824219,
|
| 2663 |
+
"logps/rejected": -285.54205322265625,
|
| 2664 |
+
"loss": 1886.4951,
|
| 2665 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 2666 |
+
"rewards/chosen": -0.41859906911849976,
|
| 2667 |
+
"rewards/margins": 0.10544377565383911,
|
| 2668 |
+
"rewards/rejected": -0.5240427851676941,
|
| 2669 |
+
"step": 1890
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 0.99,
|
| 2673 |
+
"learning_rate": 4.1748984585560094e-10,
|
| 2674 |
+
"logits/chosen": 0.5473756194114685,
|
| 2675 |
+
"logits/rejected": 0.5344858765602112,
|
| 2676 |
+
"logps/chosen": -310.27728271484375,
|
| 2677 |
+
"logps/rejected": -314.9624328613281,
|
| 2678 |
+
"loss": 2101.4789,
|
| 2679 |
+
"rewards/accuracies": 0.574999988079071,
|
| 2680 |
+
"rewards/chosen": -0.41071709990501404,
|
| 2681 |
+
"rewards/margins": 0.09045806527137756,
|
| 2682 |
+
"rewards/rejected": -0.5011752247810364,
|
| 2683 |
+
"step": 1900
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 1.0,
|
| 2687 |
+
"learning_rate": 0.0,
|
| 2688 |
+
"logits/chosen": 0.6098369359970093,
|
| 2689 |
+
"logits/rejected": 0.6253079771995544,
|
| 2690 |
+
"logps/chosen": -299.2195739746094,
|
| 2691 |
+
"logps/rejected": -283.7442932128906,
|
| 2692 |
+
"loss": 2031.7352,
|
| 2693 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 2694 |
+
"rewards/chosen": -0.4044032096862793,
|
| 2695 |
+
"rewards/margins": 0.08278089016675949,
|
| 2696 |
+
"rewards/rejected": -0.4871840476989746,
|
| 2697 |
+
"step": 1910
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 1.0,
|
| 2701 |
+
"step": 1910,
|
| 2702 |
+
"total_flos": 0.0,
|
| 2703 |
+
"train_loss": 2071.8518089414265,
|
| 2704 |
+
"train_runtime": 14310.5789,
|
| 2705 |
+
"train_samples_per_second": 4.272,
|
| 2706 |
+
"train_steps_per_second": 0.133
|
| 2707 |
+
}
|
| 2708 |
+
],
|
| 2709 |
+
"logging_steps": 10,
|
| 2710 |
+
"max_steps": 1910,
|
| 2711 |
+
"num_input_tokens_seen": 0,
|
| 2712 |
+
"num_train_epochs": 1,
|
| 2713 |
+
"save_steps": 100,
|
| 2714 |
+
"total_flos": 0.0,
|
| 2715 |
+
"train_batch_size": 4,
|
| 2716 |
+
"trial_name": null,
|
| 2717 |
+
"trial_params": null
|
| 2718 |
+
}
|