Model save
Browse files- README.md +61 -0
- adapter_model.safetensors +1 -1
- all_results.json +8 -0
- runs/May08_13-21-12_gpu4-119-4/events.out.tfevents.1715138581.gpu4-119-4.214995.0 +2 -2
- train_results.json +8 -0
- trainer_state.json +2172 -0
README.md
ADDED
|
@@ -0,0 +1,61 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: peft
|
| 3 |
+
tags:
|
| 4 |
+
- trl
|
| 5 |
+
- dpo
|
| 6 |
+
- generated_from_trainer
|
| 7 |
+
base_model: DUAL-GPO/zephyr-7b-gpo-final-i0
|
| 8 |
+
model-index:
|
| 9 |
+
- name: zephyr-7b-gpo-v2-i0
|
| 10 |
+
results: []
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 14 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 15 |
+
|
| 16 |
+
# zephyr-7b-gpo-v2-i0
|
| 17 |
+
|
| 18 |
+
This model was trained from scratch on the None dataset.
|
| 19 |
+
|
| 20 |
+
## Model description
|
| 21 |
+
|
| 22 |
+
More information needed
|
| 23 |
+
|
| 24 |
+
## Intended uses & limitations
|
| 25 |
+
|
| 26 |
+
More information needed
|
| 27 |
+
|
| 28 |
+
## Training and evaluation data
|
| 29 |
+
|
| 30 |
+
More information needed
|
| 31 |
+
|
| 32 |
+
## Training procedure
|
| 33 |
+
|
| 34 |
+
### Training hyperparameters
|
| 35 |
+
|
| 36 |
+
The following hyperparameters were used during training:
|
| 37 |
+
- learning_rate: 5e-06
|
| 38 |
+
- train_batch_size: 2
|
| 39 |
+
- eval_batch_size: 2
|
| 40 |
+
- seed: 42
|
| 41 |
+
- distributed_type: multi-GPU
|
| 42 |
+
- num_devices: 3
|
| 43 |
+
- gradient_accumulation_steps: 2
|
| 44 |
+
- total_train_batch_size: 12
|
| 45 |
+
- total_eval_batch_size: 6
|
| 46 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 47 |
+
- lr_scheduler_type: cosine
|
| 48 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 49 |
+
- num_epochs: 1
|
| 50 |
+
|
| 51 |
+
### Training results
|
| 52 |
+
|
| 53 |
+
|
| 54 |
+
|
| 55 |
+
### Framework versions
|
| 56 |
+
|
| 57 |
+
- PEFT 0.7.1
|
| 58 |
+
- Transformers 4.36.2
|
| 59 |
+
- Pytorch 2.1.2+cu118
|
| 60 |
+
- Datasets 2.14.6
|
| 61 |
+
- 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 671150064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3ca78d86d4afd1fabb938f72a4b0cdf799be850b81b12ec3a00e7417fbb089cc
|
| 3 |
size 671150064
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.021083176277421845,
|
| 4 |
+
"train_runtime": 50303.8354,
|
| 5 |
+
"train_samples": 18340,
|
| 6 |
+
"train_samples_per_second": 0.365,
|
| 7 |
+
"train_steps_per_second": 0.03
|
| 8 |
+
}
|
runs/May08_13-21-12_gpu4-119-4/events.out.tfevents.1715138581.gpu4-119-4.214995.0
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 sha256:2413995be0a47dd8733abedda48b1b3f8d2e9be1b5a75c443781edb013ef4753
|
| 3 |
+
size 101780
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.021083176277421845,
|
| 4 |
+
"train_runtime": 50303.8354,
|
| 5 |
+
"train_samples": 18340,
|
| 6 |
+
"train_samples_per_second": 0.365,
|
| 7 |
+
"train_steps_per_second": 0.03
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2172 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9996728819103696,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 1528,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0,
|
| 13 |
+
"learning_rate": 3.267973856209151e-08,
|
| 14 |
+
"logits/chosen": -1.591799020767212,
|
| 15 |
+
"logits/rejected": -1.609328031539917,
|
| 16 |
+
"logps/chosen": -327.9853210449219,
|
| 17 |
+
"logps/rejected": -375.1213684082031,
|
| 18 |
+
"loss": 0.0486,
|
| 19 |
+
"rewards/accuracies": 0.0,
|
| 20 |
+
"rewards/chosen": 0.0,
|
| 21 |
+
"rewards/margins": 0.0,
|
| 22 |
+
"rewards/rejected": 0.0,
|
| 23 |
+
"step": 1
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.01,
|
| 27 |
+
"learning_rate": 3.267973856209151e-07,
|
| 28 |
+
"logits/chosen": -1.8375271558761597,
|
| 29 |
+
"logits/rejected": -1.9081380367279053,
|
| 30 |
+
"logps/chosen": -217.70407104492188,
|
| 31 |
+
"logps/rejected": -291.03619384765625,
|
| 32 |
+
"loss": 0.0481,
|
| 33 |
+
"rewards/accuracies": 0.3888888955116272,
|
| 34 |
+
"rewards/chosen": -0.00010180399840464815,
|
| 35 |
+
"rewards/margins": -1.7401700461050496e-05,
|
| 36 |
+
"rewards/rejected": -8.440230158157647e-05,
|
| 37 |
+
"step": 10
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.01,
|
| 41 |
+
"learning_rate": 6.535947712418302e-07,
|
| 42 |
+
"logits/chosen": -1.8446567058563232,
|
| 43 |
+
"logits/rejected": -1.96743643283844,
|
| 44 |
+
"logps/chosen": -188.84414672851562,
|
| 45 |
+
"logps/rejected": -223.9464569091797,
|
| 46 |
+
"loss": 0.0476,
|
| 47 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 48 |
+
"rewards/chosen": -0.0001967716234503314,
|
| 49 |
+
"rewards/margins": 0.0005254523130133748,
|
| 50 |
+
"rewards/rejected": -0.0007222238346002996,
|
| 51 |
+
"step": 20
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.02,
|
| 55 |
+
"learning_rate": 9.80392156862745e-07,
|
| 56 |
+
"logits/chosen": -2.0779905319213867,
|
| 57 |
+
"logits/rejected": -1.7060468196868896,
|
| 58 |
+
"logps/chosen": -239.03250122070312,
|
| 59 |
+
"logps/rejected": -208.9864501953125,
|
| 60 |
+
"loss": 0.0504,
|
| 61 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 62 |
+
"rewards/chosen": -0.0017802307847887278,
|
| 63 |
+
"rewards/margins": 0.001976901665329933,
|
| 64 |
+
"rewards/rejected": -0.00375713175162673,
|
| 65 |
+
"step": 30
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.03,
|
| 69 |
+
"learning_rate": 1.3071895424836604e-06,
|
| 70 |
+
"logits/chosen": -1.7227962017059326,
|
| 71 |
+
"logits/rejected": -1.7102130651474,
|
| 72 |
+
"logps/chosen": -361.55224609375,
|
| 73 |
+
"logps/rejected": -365.0895690917969,
|
| 74 |
+
"loss": 0.0376,
|
| 75 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 76 |
+
"rewards/chosen": -0.019573379307985306,
|
| 77 |
+
"rewards/margins": 0.011974343098700047,
|
| 78 |
+
"rewards/rejected": -0.031547725200653076,
|
| 79 |
+
"step": 40
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.03,
|
| 83 |
+
"learning_rate": 1.6339869281045753e-06,
|
| 84 |
+
"logits/chosen": -1.8763091564178467,
|
| 85 |
+
"logits/rejected": -1.6960475444793701,
|
| 86 |
+
"logps/chosen": -392.33709716796875,
|
| 87 |
+
"logps/rejected": -353.79827880859375,
|
| 88 |
+
"loss": 0.0283,
|
| 89 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 90 |
+
"rewards/chosen": -0.06343773007392883,
|
| 91 |
+
"rewards/margins": 0.023594681173563004,
|
| 92 |
+
"rewards/rejected": -0.08703242242336273,
|
| 93 |
+
"step": 50
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.04,
|
| 97 |
+
"learning_rate": 1.96078431372549e-06,
|
| 98 |
+
"logits/chosen": -1.6577602624893188,
|
| 99 |
+
"logits/rejected": -1.5122312307357788,
|
| 100 |
+
"logps/chosen": -384.6730651855469,
|
| 101 |
+
"logps/rejected": -499.12677001953125,
|
| 102 |
+
"loss": 0.0236,
|
| 103 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 104 |
+
"rewards/chosen": -0.1024140864610672,
|
| 105 |
+
"rewards/margins": 0.13252298533916473,
|
| 106 |
+
"rewards/rejected": -0.23493707180023193,
|
| 107 |
+
"step": 60
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.05,
|
| 111 |
+
"learning_rate": 2.2875816993464053e-06,
|
| 112 |
+
"logits/chosen": -1.7868068218231201,
|
| 113 |
+
"logits/rejected": -1.6683218479156494,
|
| 114 |
+
"logps/chosen": -391.08642578125,
|
| 115 |
+
"logps/rejected": -484.86737060546875,
|
| 116 |
+
"loss": 0.0204,
|
| 117 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 118 |
+
"rewards/chosen": -0.11379416286945343,
|
| 119 |
+
"rewards/margins": 0.09717617928981781,
|
| 120 |
+
"rewards/rejected": -0.21097035706043243,
|
| 121 |
+
"step": 70
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.05,
|
| 125 |
+
"learning_rate": 2.6143790849673208e-06,
|
| 126 |
+
"logits/chosen": -1.7627947330474854,
|
| 127 |
+
"logits/rejected": -1.526816725730896,
|
| 128 |
+
"logps/chosen": -385.7118835449219,
|
| 129 |
+
"logps/rejected": -395.92108154296875,
|
| 130 |
+
"loss": 0.0225,
|
| 131 |
+
"rewards/accuracies": 0.625,
|
| 132 |
+
"rewards/chosen": -0.10957636684179306,
|
| 133 |
+
"rewards/margins": 0.08020367473363876,
|
| 134 |
+
"rewards/rejected": -0.18978002667427063,
|
| 135 |
+
"step": 80
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.06,
|
| 139 |
+
"learning_rate": 2.9411764705882355e-06,
|
| 140 |
+
"logits/chosen": -1.761884331703186,
|
| 141 |
+
"logits/rejected": -1.7560245990753174,
|
| 142 |
+
"logps/chosen": -301.0171813964844,
|
| 143 |
+
"logps/rejected": -439.34906005859375,
|
| 144 |
+
"loss": 0.0188,
|
| 145 |
+
"rewards/accuracies": 0.675000011920929,
|
| 146 |
+
"rewards/chosen": -0.062449000775814056,
|
| 147 |
+
"rewards/margins": 0.10943855345249176,
|
| 148 |
+
"rewards/rejected": -0.17188753187656403,
|
| 149 |
+
"step": 90
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.07,
|
| 153 |
+
"learning_rate": 3.2679738562091506e-06,
|
| 154 |
+
"logits/chosen": -1.6214962005615234,
|
| 155 |
+
"logits/rejected": -1.6456191539764404,
|
| 156 |
+
"logps/chosen": -258.74237060546875,
|
| 157 |
+
"logps/rejected": -351.51641845703125,
|
| 158 |
+
"loss": 0.0173,
|
| 159 |
+
"rewards/accuracies": 0.675000011920929,
|
| 160 |
+
"rewards/chosen": -0.07586181163787842,
|
| 161 |
+
"rewards/margins": 0.08772234618663788,
|
| 162 |
+
"rewards/rejected": -0.1635841578245163,
|
| 163 |
+
"step": 100
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.07,
|
| 167 |
+
"learning_rate": 3.5947712418300657e-06,
|
| 168 |
+
"logits/chosen": -1.6627695560455322,
|
| 169 |
+
"logits/rejected": -1.5852539539337158,
|
| 170 |
+
"logps/chosen": -331.30145263671875,
|
| 171 |
+
"logps/rejected": -448.34124755859375,
|
| 172 |
+
"loss": 0.0156,
|
| 173 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 174 |
+
"rewards/chosen": -0.12424789369106293,
|
| 175 |
+
"rewards/margins": 0.0813526064157486,
|
| 176 |
+
"rewards/rejected": -0.20560050010681152,
|
| 177 |
+
"step": 110
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.08,
|
| 181 |
+
"learning_rate": 3.92156862745098e-06,
|
| 182 |
+
"logits/chosen": -1.5744519233703613,
|
| 183 |
+
"logits/rejected": -1.479768991470337,
|
| 184 |
+
"logps/chosen": -462.4168395996094,
|
| 185 |
+
"logps/rejected": -554.2327880859375,
|
| 186 |
+
"loss": 0.0216,
|
| 187 |
+
"rewards/accuracies": 0.75,
|
| 188 |
+
"rewards/chosen": -0.2043522298336029,
|
| 189 |
+
"rewards/margins": 0.08921514451503754,
|
| 190 |
+
"rewards/rejected": -0.29356735944747925,
|
| 191 |
+
"step": 120
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.09,
|
| 195 |
+
"learning_rate": 4.2483660130718954e-06,
|
| 196 |
+
"logits/chosen": -1.918543815612793,
|
| 197 |
+
"logits/rejected": -1.7244701385498047,
|
| 198 |
+
"logps/chosen": -366.99761962890625,
|
| 199 |
+
"logps/rejected": -438.06964111328125,
|
| 200 |
+
"loss": 0.0303,
|
| 201 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 202 |
+
"rewards/chosen": -0.11779004335403442,
|
| 203 |
+
"rewards/margins": 0.08299387991428375,
|
| 204 |
+
"rewards/rejected": -0.20078392326831818,
|
| 205 |
+
"step": 130
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.09,
|
| 209 |
+
"learning_rate": 4.5751633986928105e-06,
|
| 210 |
+
"logits/chosen": -1.8474334478378296,
|
| 211 |
+
"logits/rejected": -1.51116144657135,
|
| 212 |
+
"logps/chosen": -443.89129638671875,
|
| 213 |
+
"logps/rejected": -516.485107421875,
|
| 214 |
+
"loss": 0.0285,
|
| 215 |
+
"rewards/accuracies": 0.675000011920929,
|
| 216 |
+
"rewards/chosen": -0.1710434854030609,
|
| 217 |
+
"rewards/margins": 0.08919176459312439,
|
| 218 |
+
"rewards/rejected": -0.2602352499961853,
|
| 219 |
+
"step": 140
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.1,
|
| 223 |
+
"learning_rate": 4.901960784313726e-06,
|
| 224 |
+
"logits/chosen": -1.946266770362854,
|
| 225 |
+
"logits/rejected": -1.6877708435058594,
|
| 226 |
+
"logps/chosen": -403.88897705078125,
|
| 227 |
+
"logps/rejected": -492.328125,
|
| 228 |
+
"loss": 0.0169,
|
| 229 |
+
"rewards/accuracies": 0.699999988079071,
|
| 230 |
+
"rewards/chosen": -0.14448824524879456,
|
| 231 |
+
"rewards/margins": 0.11163143813610077,
|
| 232 |
+
"rewards/rejected": -0.25611966848373413,
|
| 233 |
+
"step": 150
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.1,
|
| 237 |
+
"learning_rate": 4.999680264259825e-06,
|
| 238 |
+
"logits/chosen": -1.8900115489959717,
|
| 239 |
+
"logits/rejected": -1.774910569190979,
|
| 240 |
+
"logps/chosen": -317.9911193847656,
|
| 241 |
+
"logps/rejected": -480.47930908203125,
|
| 242 |
+
"loss": 0.0222,
|
| 243 |
+
"rewards/accuracies": 0.675000011920929,
|
| 244 |
+
"rewards/chosen": -0.12115655094385147,
|
| 245 |
+
"rewards/margins": 0.13366180658340454,
|
| 246 |
+
"rewards/rejected": -0.2548183798789978,
|
| 247 |
+
"step": 160
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.11,
|
| 251 |
+
"learning_rate": 4.998114408534616e-06,
|
| 252 |
+
"logits/chosen": -1.8668384552001953,
|
| 253 |
+
"logits/rejected": -1.594963788986206,
|
| 254 |
+
"logps/chosen": -461.4814453125,
|
| 255 |
+
"logps/rejected": -463.83062744140625,
|
| 256 |
+
"loss": 0.022,
|
| 257 |
+
"rewards/accuracies": 0.675000011920929,
|
| 258 |
+
"rewards/chosen": -0.13391847908496857,
|
| 259 |
+
"rewards/margins": 0.09448060393333435,
|
| 260 |
+
"rewards/rejected": -0.22839908301830292,
|
| 261 |
+
"step": 170
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.12,
|
| 265 |
+
"learning_rate": 4.995244522215781e-06,
|
| 266 |
+
"logits/chosen": -1.4480278491973877,
|
| 267 |
+
"logits/rejected": -1.2824418544769287,
|
| 268 |
+
"logps/chosen": -398.3277587890625,
|
| 269 |
+
"logps/rejected": -441.39776611328125,
|
| 270 |
+
"loss": 0.0323,
|
| 271 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 272 |
+
"rewards/chosen": -0.13857640326023102,
|
| 273 |
+
"rewards/margins": 0.07922663539648056,
|
| 274 |
+
"rewards/rejected": -0.21780303120613098,
|
| 275 |
+
"step": 180
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.12,
|
| 279 |
+
"learning_rate": 4.9910721034010655e-06,
|
| 280 |
+
"logits/chosen": -1.8611648082733154,
|
| 281 |
+
"logits/rejected": -1.5049927234649658,
|
| 282 |
+
"logps/chosen": -442.27532958984375,
|
| 283 |
+
"logps/rejected": -513.6070556640625,
|
| 284 |
+
"loss": 0.0206,
|
| 285 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 286 |
+
"rewards/chosen": -0.10470763593912125,
|
| 287 |
+
"rewards/margins": 0.1342332810163498,
|
| 288 |
+
"rewards/rejected": -0.23894090950489044,
|
| 289 |
+
"step": 190
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.13,
|
| 293 |
+
"learning_rate": 4.985599330117931e-06,
|
| 294 |
+
"logits/chosen": -1.7740875482559204,
|
| 295 |
+
"logits/rejected": -1.6385295391082764,
|
| 296 |
+
"logps/chosen": -340.8525390625,
|
| 297 |
+
"logps/rejected": -373.794677734375,
|
| 298 |
+
"loss": 0.021,
|
| 299 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 300 |
+
"rewards/chosen": -0.11725501716136932,
|
| 301 |
+
"rewards/margins": 0.05946159362792969,
|
| 302 |
+
"rewards/rejected": -0.176716610789299,
|
| 303 |
+
"step": 200
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.14,
|
| 307 |
+
"learning_rate": 4.978829059186611e-06,
|
| 308 |
+
"logits/chosen": -1.6705108880996704,
|
| 309 |
+
"logits/rejected": -1.5942919254302979,
|
| 310 |
+
"logps/chosen": -390.35748291015625,
|
| 311 |
+
"logps/rejected": -478.62713623046875,
|
| 312 |
+
"loss": 0.0167,
|
| 313 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 314 |
+
"rewards/chosen": -0.1433432400226593,
|
| 315 |
+
"rewards/margins": 0.08838480710983276,
|
| 316 |
+
"rewards/rejected": -0.23172803223133087,
|
| 317 |
+
"step": 210
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.14,
|
| 321 |
+
"learning_rate": 4.97076482472884e-06,
|
| 322 |
+
"logits/chosen": -1.8241045475006104,
|
| 323 |
+
"logits/rejected": -1.6329164505004883,
|
| 324 |
+
"logps/chosen": -475.8793029785156,
|
| 325 |
+
"logps/rejected": -482.19854736328125,
|
| 326 |
+
"loss": 0.0205,
|
| 327 |
+
"rewards/accuracies": 0.675000011920929,
|
| 328 |
+
"rewards/chosen": -0.14704802632331848,
|
| 329 |
+
"rewards/margins": 0.05757030099630356,
|
| 330 |
+
"rewards/rejected": -0.20461833477020264,
|
| 331 |
+
"step": 220
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.15,
|
| 335 |
+
"learning_rate": 4.961410836323014e-06,
|
| 336 |
+
"logits/chosen": -1.5026575326919556,
|
| 337 |
+
"logits/rejected": -1.4128906726837158,
|
| 338 |
+
"logps/chosen": -374.00555419921875,
|
| 339 |
+
"logps/rejected": -541.5623168945312,
|
| 340 |
+
"loss": 0.023,
|
| 341 |
+
"rewards/accuracies": 0.574999988079071,
|
| 342 |
+
"rewards/chosen": -0.15363118052482605,
|
| 343 |
+
"rewards/margins": 0.13868795335292816,
|
| 344 |
+
"rewards/rejected": -0.292319118976593,
|
| 345 |
+
"step": 230
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.16,
|
| 349 |
+
"learning_rate": 4.950771976806769e-06,
|
| 350 |
+
"logits/chosen": -1.7081775665283203,
|
| 351 |
+
"logits/rejected": -1.6031742095947266,
|
| 352 |
+
"logps/chosen": -385.6658935546875,
|
| 353 |
+
"logps/rejected": -442.20697021484375,
|
| 354 |
+
"loss": 0.0199,
|
| 355 |
+
"rewards/accuracies": 0.550000011920929,
|
| 356 |
+
"rewards/chosen": -0.145090714097023,
|
| 357 |
+
"rewards/margins": 0.05894991010427475,
|
| 358 |
+
"rewards/rejected": -0.20404061675071716,
|
| 359 |
+
"step": 240
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.16,
|
| 363 |
+
"learning_rate": 4.938853799728112e-06,
|
| 364 |
+
"logits/chosen": -1.7450878620147705,
|
| 365 |
+
"logits/rejected": -1.6923738718032837,
|
| 366 |
+
"logps/chosen": -437.3587341308594,
|
| 367 |
+
"logps/rejected": -505.5862731933594,
|
| 368 |
+
"loss": 0.0166,
|
| 369 |
+
"rewards/accuracies": 0.675000011920929,
|
| 370 |
+
"rewards/chosen": -0.16524454951286316,
|
| 371 |
+
"rewards/margins": 0.10024659335613251,
|
| 372 |
+
"rewards/rejected": -0.2654911279678345,
|
| 373 |
+
"step": 250
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.17,
|
| 377 |
+
"learning_rate": 4.925662526446431e-06,
|
| 378 |
+
"logits/chosen": -1.5717720985412598,
|
| 379 |
+
"logits/rejected": -1.779036521911621,
|
| 380 |
+
"logps/chosen": -432.2626037597656,
|
| 381 |
+
"logps/rejected": -544.822509765625,
|
| 382 |
+
"loss": 0.0236,
|
| 383 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 384 |
+
"rewards/chosen": -0.1972414255142212,
|
| 385 |
+
"rewards/margins": 0.06645989418029785,
|
| 386 |
+
"rewards/rejected": -0.26370128989219666,
|
| 387 |
+
"step": 260
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.18,
|
| 391 |
+
"learning_rate": 4.911205042884912e-06,
|
| 392 |
+
"logits/chosen": -1.7134885787963867,
|
| 393 |
+
"logits/rejected": -1.5530506372451782,
|
| 394 |
+
"logps/chosen": -467.70001220703125,
|
| 395 |
+
"logps/rejected": -521.0452880859375,
|
| 396 |
+
"loss": 0.0153,
|
| 397 |
+
"rewards/accuracies": 0.824999988079071,
|
| 398 |
+
"rewards/chosen": -0.16030625998973846,
|
| 399 |
+
"rewards/margins": 0.11603019386529922,
|
| 400 |
+
"rewards/rejected": -0.2763364017009735,
|
| 401 |
+
"step": 270
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.18,
|
| 405 |
+
"learning_rate": 4.895488895936047e-06,
|
| 406 |
+
"logits/chosen": -1.7987966537475586,
|
| 407 |
+
"logits/rejected": -1.7680168151855469,
|
| 408 |
+
"logps/chosen": -449.7742614746094,
|
| 409 |
+
"logps/rejected": -502.37359619140625,
|
| 410 |
+
"loss": 0.0212,
|
| 411 |
+
"rewards/accuracies": 0.699999988079071,
|
| 412 |
+
"rewards/chosen": -0.11571071296930313,
|
| 413 |
+
"rewards/margins": 0.0691571980714798,
|
| 414 |
+
"rewards/rejected": -0.18486790359020233,
|
| 415 |
+
"step": 280
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.19,
|
| 419 |
+
"learning_rate": 4.8785222895221075e-06,
|
| 420 |
+
"logits/chosen": -1.8355566263198853,
|
| 421 |
+
"logits/rejected": -1.8976913690567017,
|
| 422 |
+
"logps/chosen": -426.9527282714844,
|
| 423 |
+
"logps/rejected": -486.2103576660156,
|
| 424 |
+
"loss": 0.0153,
|
| 425 |
+
"rewards/accuracies": 0.5,
|
| 426 |
+
"rewards/chosen": -0.11018826067447662,
|
| 427 |
+
"rewards/margins": 0.06115338206291199,
|
| 428 |
+
"rewards/rejected": -0.17134162783622742,
|
| 429 |
+
"step": 290
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.2,
|
| 433 |
+
"learning_rate": 4.860314080312651e-06,
|
| 434 |
+
"logits/chosen": -1.716607689857483,
|
| 435 |
+
"logits/rejected": -1.781498908996582,
|
| 436 |
+
"logps/chosen": -279.833740234375,
|
| 437 |
+
"logps/rejected": -413.525146484375,
|
| 438 |
+
"loss": 0.0228,
|
| 439 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 440 |
+
"rewards/chosen": -0.09433625638484955,
|
| 441 |
+
"rewards/margins": 0.10549966245889664,
|
| 442 |
+
"rewards/rejected": -0.19983592629432678,
|
| 443 |
+
"step": 300
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.2,
|
| 447 |
+
"learning_rate": 4.840873773101287e-06,
|
| 448 |
+
"logits/chosen": -1.9624496698379517,
|
| 449 |
+
"logits/rejected": -1.6330894231796265,
|
| 450 |
+
"logps/chosen": -339.2301940917969,
|
| 451 |
+
"logps/rejected": -369.59686279296875,
|
| 452 |
+
"loss": 0.0292,
|
| 453 |
+
"rewards/accuracies": 0.625,
|
| 454 |
+
"rewards/chosen": -0.06697295606136322,
|
| 455 |
+
"rewards/margins": 0.07780499756336212,
|
| 456 |
+
"rewards/rejected": -0.14477796852588654,
|
| 457 |
+
"step": 310
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.21,
|
| 461 |
+
"learning_rate": 4.820211515844116e-06,
|
| 462 |
+
"logits/chosen": -1.9243313074111938,
|
| 463 |
+
"logits/rejected": -1.6555219888687134,
|
| 464 |
+
"logps/chosen": -368.7879943847656,
|
| 465 |
+
"logps/rejected": -404.6373596191406,
|
| 466 |
+
"loss": 0.0257,
|
| 467 |
+
"rewards/accuracies": 0.625,
|
| 468 |
+
"rewards/chosen": -0.10985688120126724,
|
| 469 |
+
"rewards/margins": 0.07864276319742203,
|
| 470 |
+
"rewards/rejected": -0.18849964439868927,
|
| 471 |
+
"step": 320
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.22,
|
| 475 |
+
"learning_rate": 4.798338094362439e-06,
|
| 476 |
+
"logits/chosen": -1.7449579238891602,
|
| 477 |
+
"logits/rejected": -1.4674043655395508,
|
| 478 |
+
"logps/chosen": -423.55255126953125,
|
| 479 |
+
"logps/rejected": -456.8785095214844,
|
| 480 |
+
"loss": 0.0183,
|
| 481 |
+
"rewards/accuracies": 0.75,
|
| 482 |
+
"rewards/chosen": -0.12963487207889557,
|
| 483 |
+
"rewards/margins": 0.09587951004505157,
|
| 484 |
+
"rewards/rejected": -0.22551438212394714,
|
| 485 |
+
"step": 330
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.22,
|
| 489 |
+
"learning_rate": 4.775264926712489e-06,
|
| 490 |
+
"logits/chosen": -1.6274173259735107,
|
| 491 |
+
"logits/rejected": -1.6101815700531006,
|
| 492 |
+
"logps/chosen": -323.04388427734375,
|
| 493 |
+
"logps/rejected": -481.54949951171875,
|
| 494 |
+
"loss": 0.0308,
|
| 495 |
+
"rewards/accuracies": 0.550000011920929,
|
| 496 |
+
"rewards/chosen": -0.1510605365037918,
|
| 497 |
+
"rewards/margins": 0.11264171451330185,
|
| 498 |
+
"rewards/rejected": -0.26370224356651306,
|
| 499 |
+
"step": 340
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.23,
|
| 503 |
+
"learning_rate": 4.751004057225147e-06,
|
| 504 |
+
"logits/chosen": -1.719301462173462,
|
| 505 |
+
"logits/rejected": -1.6450971364974976,
|
| 506 |
+
"logps/chosen": -572.3885498046875,
|
| 507 |
+
"logps/rejected": -655.76953125,
|
| 508 |
+
"loss": 0.0197,
|
| 509 |
+
"rewards/accuracies": 0.625,
|
| 510 |
+
"rewards/chosen": -0.21081407368183136,
|
| 511 |
+
"rewards/margins": 0.08522544801235199,
|
| 512 |
+
"rewards/rejected": -0.29603949189186096,
|
| 513 |
+
"step": 350
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.24,
|
| 517 |
+
"learning_rate": 4.725568150218719e-06,
|
| 518 |
+
"logits/chosen": -1.7547239065170288,
|
| 519 |
+
"logits/rejected": -1.3802999258041382,
|
| 520 |
+
"logps/chosen": -414.50445556640625,
|
| 521 |
+
"logps/rejected": -463.7474670410156,
|
| 522 |
+
"loss": 0.0225,
|
| 523 |
+
"rewards/accuracies": 0.75,
|
| 524 |
+
"rewards/chosen": -0.11397633701562881,
|
| 525 |
+
"rewards/margins": 0.10422484576702118,
|
| 526 |
+
"rewards/rejected": -0.21820120513439178,
|
| 527 |
+
"step": 360
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.24,
|
| 531 |
+
"learning_rate": 4.6989704833880936e-06,
|
| 532 |
+
"logits/chosen": -1.653165578842163,
|
| 533 |
+
"logits/rejected": -1.7715513706207275,
|
| 534 |
+
"logps/chosen": -407.423583984375,
|
| 535 |
+
"logps/rejected": -510.4872131347656,
|
| 536 |
+
"loss": 0.0211,
|
| 537 |
+
"rewards/accuracies": 0.699999988079071,
|
| 538 |
+
"rewards/chosen": -0.1670456975698471,
|
| 539 |
+
"rewards/margins": 0.0627368614077568,
|
| 540 |
+
"rewards/rejected": -0.22978255152702332,
|
| 541 |
+
"step": 370
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.25,
|
| 545 |
+
"learning_rate": 4.671224940873704e-06,
|
| 546 |
+
"logits/chosen": -1.8084532022476196,
|
| 547 |
+
"logits/rejected": -1.6008380651474,
|
| 548 |
+
"logps/chosen": -411.17547607421875,
|
| 549 |
+
"logps/rejected": -478.0519104003906,
|
| 550 |
+
"loss": 0.0215,
|
| 551 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 552 |
+
"rewards/chosen": -0.1790023148059845,
|
| 553 |
+
"rewards/margins": 0.0856521874666214,
|
| 554 |
+
"rewards/rejected": -0.2646545171737671,
|
| 555 |
+
"step": 380
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.26,
|
| 559 |
+
"learning_rate": 4.642346006013925e-06,
|
| 560 |
+
"logits/chosen": -1.7398563623428345,
|
| 561 |
+
"logits/rejected": -1.7137126922607422,
|
| 562 |
+
"logps/chosen": -489.26483154296875,
|
| 563 |
+
"logps/rejected": -501.8857421875,
|
| 564 |
+
"loss": 0.0207,
|
| 565 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 566 |
+
"rewards/chosen": -0.15890449285507202,
|
| 567 |
+
"rewards/margins": 0.04155914857983589,
|
| 568 |
+
"rewards/rejected": -0.20046362280845642,
|
| 569 |
+
"step": 390
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.26,
|
| 573 |
+
"learning_rate": 4.612348753784682e-06,
|
| 574 |
+
"logits/chosen": -1.628105878829956,
|
| 575 |
+
"logits/rejected": -1.6889688968658447,
|
| 576 |
+
"logps/chosen": -348.6719055175781,
|
| 577 |
+
"logps/rejected": -442.29046630859375,
|
| 578 |
+
"loss": 0.0213,
|
| 579 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 580 |
+
"rewards/chosen": -0.1342150866985321,
|
| 581 |
+
"rewards/margins": 0.07590731233358383,
|
| 582 |
+
"rewards/rejected": -0.21012239158153534,
|
| 583 |
+
"step": 400
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.27,
|
| 587 |
+
"learning_rate": 4.5812488429302245e-06,
|
| 588 |
+
"logits/chosen": -1.8471730947494507,
|
| 589 |
+
"logits/rejected": -1.5871822834014893,
|
| 590 |
+
"logps/chosen": -460.2723693847656,
|
| 591 |
+
"logps/rejected": -453.3777770996094,
|
| 592 |
+
"loss": 0.0122,
|
| 593 |
+
"rewards/accuracies": 0.75,
|
| 594 |
+
"rewards/chosen": -0.08317364007234573,
|
| 595 |
+
"rewards/margins": 0.09449958801269531,
|
| 596 |
+
"rewards/rejected": -0.17767322063446045,
|
| 597 |
+
"step": 410
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.27,
|
| 601 |
+
"learning_rate": 4.54906250778917e-06,
|
| 602 |
+
"logits/chosen": -2.0099849700927734,
|
| 603 |
+
"logits/rejected": -1.8552219867706299,
|
| 604 |
+
"logps/chosen": -278.762451171875,
|
| 605 |
+
"logps/rejected": -330.7545166015625,
|
| 606 |
+
"loss": 0.0158,
|
| 607 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 608 |
+
"rewards/chosen": -0.054166488349437714,
|
| 609 |
+
"rewards/margins": 0.0891014039516449,
|
| 610 |
+
"rewards/rejected": -0.1432678997516632,
|
| 611 |
+
"step": 420
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.28,
|
| 615 |
+
"learning_rate": 4.515806549820084e-06,
|
| 616 |
+
"logits/chosen": -1.8863036632537842,
|
| 617 |
+
"logits/rejected": -1.7360244989395142,
|
| 618 |
+
"logps/chosen": -385.07623291015625,
|
| 619 |
+
"logps/rejected": -412.04766845703125,
|
| 620 |
+
"loss": 0.025,
|
| 621 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 622 |
+
"rewards/chosen": -0.10509233176708221,
|
| 623 |
+
"rewards/margins": 0.06545521318912506,
|
| 624 |
+
"rewards/rejected": -0.17054755985736847,
|
| 625 |
+
"step": 430
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.29,
|
| 629 |
+
"learning_rate": 4.48149832883101e-06,
|
| 630 |
+
"logits/chosen": -1.9307434558868408,
|
| 631 |
+
"logits/rejected": -1.6779987812042236,
|
| 632 |
+
"logps/chosen": -407.74383544921875,
|
| 633 |
+
"logps/rejected": -515.69873046875,
|
| 634 |
+
"loss": 0.0205,
|
| 635 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 636 |
+
"rewards/chosen": -0.10622302442789078,
|
| 637 |
+
"rewards/margins": 0.12537193298339844,
|
| 638 |
+
"rewards/rejected": -0.23159494996070862,
|
| 639 |
+
"step": 440
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.29,
|
| 643 |
+
"learning_rate": 4.446155753917559e-06,
|
| 644 |
+
"logits/chosen": -1.9925124645233154,
|
| 645 |
+
"logits/rejected": -1.9259634017944336,
|
| 646 |
+
"logps/chosen": -374.60223388671875,
|
| 647 |
+
"logps/rejected": -428.49969482421875,
|
| 648 |
+
"loss": 0.0217,
|
| 649 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 650 |
+
"rewards/chosen": -0.12692932784557343,
|
| 651 |
+
"rewards/margins": 0.06185678765177727,
|
| 652 |
+
"rewards/rejected": -0.18878613412380219,
|
| 653 |
+
"step": 450
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.3,
|
| 657 |
+
"learning_rate": 4.409797274114245e-06,
|
| 658 |
+
"logits/chosen": -2.01132869720459,
|
| 659 |
+
"logits/rejected": -1.7584896087646484,
|
| 660 |
+
"logps/chosen": -459.43756103515625,
|
| 661 |
+
"logps/rejected": -480.87799072265625,
|
| 662 |
+
"loss": 0.0143,
|
| 663 |
+
"rewards/accuracies": 0.675000011920929,
|
| 664 |
+
"rewards/chosen": -0.12262854725122452,
|
| 665 |
+
"rewards/margins": 0.06466259062290192,
|
| 666 |
+
"rewards/rejected": -0.18729116022586823,
|
| 667 |
+
"step": 460
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.31,
|
| 671 |
+
"learning_rate": 4.372441868763981e-06,
|
| 672 |
+
"logits/chosen": -1.6305385828018188,
|
| 673 |
+
"logits/rejected": -1.368896245956421,
|
| 674 |
+
"logps/chosen": -360.996826171875,
|
| 675 |
+
"logps/rejected": -434.44097900390625,
|
| 676 |
+
"loss": 0.0273,
|
| 677 |
+
"rewards/accuracies": 0.675000011920929,
|
| 678 |
+
"rewards/chosen": -0.1023339256644249,
|
| 679 |
+
"rewards/margins": 0.1018548384308815,
|
| 680 |
+
"rewards/rejected": -0.2041887789964676,
|
| 681 |
+
"step": 470
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.31,
|
| 685 |
+
"learning_rate": 4.334109037610757e-06,
|
| 686 |
+
"logits/chosen": -1.9129701852798462,
|
| 687 |
+
"logits/rejected": -1.6556217670440674,
|
| 688 |
+
"logps/chosen": -380.2204284667969,
|
| 689 |
+
"logps/rejected": -451.8375549316406,
|
| 690 |
+
"loss": 0.0136,
|
| 691 |
+
"rewards/accuracies": 0.75,
|
| 692 |
+
"rewards/chosen": -0.09523477405309677,
|
| 693 |
+
"rewards/margins": 0.0824359580874443,
|
| 694 |
+
"rewards/rejected": -0.17767074704170227,
|
| 695 |
+
"step": 480
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.32,
|
| 699 |
+
"learning_rate": 4.294818790620644e-06,
|
| 700 |
+
"logits/chosen": -1.6334559917449951,
|
| 701 |
+
"logits/rejected": -1.5937422513961792,
|
| 702 |
+
"logps/chosen": -355.0863342285156,
|
| 703 |
+
"logps/rejected": -435.58770751953125,
|
| 704 |
+
"loss": 0.0269,
|
| 705 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 706 |
+
"rewards/chosen": -0.13321994245052338,
|
| 707 |
+
"rewards/margins": 0.054788071662187576,
|
| 708 |
+
"rewards/rejected": -0.18800799548625946,
|
| 709 |
+
"step": 490
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.33,
|
| 713 |
+
"learning_rate": 4.2545916375364835e-06,
|
| 714 |
+
"logits/chosen": -1.7967026233673096,
|
| 715 |
+
"logits/rejected": -1.6936085224151611,
|
| 716 |
+
"logps/chosen": -297.0005798339844,
|
| 717 |
+
"logps/rejected": -450.44097900390625,
|
| 718 |
+
"loss": 0.0191,
|
| 719 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 720 |
+
"rewards/chosen": -0.10439346730709076,
|
| 721 |
+
"rewards/margins": 0.13930422067642212,
|
| 722 |
+
"rewards/rejected": -0.24369768798351288,
|
| 723 |
+
"step": 500
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.33,
|
| 727 |
+
"learning_rate": 4.213448577171676e-06,
|
| 728 |
+
"logits/chosen": -1.8098952770233154,
|
| 729 |
+
"logits/rejected": -1.6606693267822266,
|
| 730 |
+
"logps/chosen": -403.1839294433594,
|
| 731 |
+
"logps/rejected": -525.0617065429688,
|
| 732 |
+
"loss": 0.0235,
|
| 733 |
+
"rewards/accuracies": 0.824999988079071,
|
| 734 |
+
"rewards/chosen": -0.1185646504163742,
|
| 735 |
+
"rewards/margins": 0.13588686287403107,
|
| 736 |
+
"rewards/rejected": -0.25445154309272766,
|
| 737 |
+
"step": 510
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.34,
|
| 741 |
+
"learning_rate": 4.171411086448674e-06,
|
| 742 |
+
"logits/chosen": -1.695320725440979,
|
| 743 |
+
"logits/rejected": -1.7887874841690063,
|
| 744 |
+
"logps/chosen": -334.51434326171875,
|
| 745 |
+
"logps/rejected": -463.96844482421875,
|
| 746 |
+
"loss": 0.0278,
|
| 747 |
+
"rewards/accuracies": 0.699999988079071,
|
| 748 |
+
"rewards/chosen": -0.10001678764820099,
|
| 749 |
+
"rewards/margins": 0.10297022759914398,
|
| 750 |
+
"rewards/rejected": -0.20298703014850616,
|
| 751 |
+
"step": 520
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.35,
|
| 755 |
+
"learning_rate": 4.128501109187903e-06,
|
| 756 |
+
"logits/chosen": -1.88802170753479,
|
| 757 |
+
"logits/rejected": -1.5964090824127197,
|
| 758 |
+
"logps/chosen": -340.66851806640625,
|
| 759 |
+
"logps/rejected": -431.9608459472656,
|
| 760 |
+
"loss": 0.0225,
|
| 761 |
+
"rewards/accuracies": 0.675000011920929,
|
| 762 |
+
"rewards/chosen": -0.09433753788471222,
|
| 763 |
+
"rewards/margins": 0.09536565840244293,
|
| 764 |
+
"rewards/rejected": -0.18970318138599396,
|
| 765 |
+
"step": 530
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.35,
|
| 769 |
+
"learning_rate": 4.084741044652956e-06,
|
| 770 |
+
"logits/chosen": -1.8472003936767578,
|
| 771 |
+
"logits/rejected": -1.7370185852050781,
|
| 772 |
+
"logps/chosen": -400.97662353515625,
|
| 773 |
+
"logps/rejected": -482.4466247558594,
|
| 774 |
+
"loss": 0.0173,
|
| 775 |
+
"rewards/accuracies": 0.800000011920929,
|
| 776 |
+
"rewards/chosen": -0.13318313658237457,
|
| 777 |
+
"rewards/margins": 0.09391706436872482,
|
| 778 |
+
"rewards/rejected": -0.2271002233028412,
|
| 779 |
+
"step": 540
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.36,
|
| 783 |
+
"learning_rate": 4.040153735858041e-06,
|
| 784 |
+
"logits/chosen": -1.7910776138305664,
|
| 785 |
+
"logits/rejected": -1.4051631689071655,
|
| 786 |
+
"logps/chosen": -472.873291015625,
|
| 787 |
+
"logps/rejected": -538.3961791992188,
|
| 788 |
+
"loss": 0.0215,
|
| 789 |
+
"rewards/accuracies": 0.75,
|
| 790 |
+
"rewards/chosen": -0.20243683457374573,
|
| 791 |
+
"rewards/margins": 0.11980029195547104,
|
| 792 |
+
"rewards/rejected": -0.322237104177475,
|
| 793 |
+
"step": 550
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.37,
|
| 797 |
+
"learning_rate": 3.9947624576437975e-06,
|
| 798 |
+
"logits/chosen": -1.8611959218978882,
|
| 799 |
+
"logits/rejected": -1.5981895923614502,
|
| 800 |
+
"logps/chosen": -379.29010009765625,
|
| 801 |
+
"logps/rejected": -485.96649169921875,
|
| 802 |
+
"loss": 0.0163,
|
| 803 |
+
"rewards/accuracies": 0.699999988079071,
|
| 804 |
+
"rewards/chosen": -0.125335693359375,
|
| 805 |
+
"rewards/margins": 0.11214844882488251,
|
| 806 |
+
"rewards/rejected": -0.2374841421842575,
|
| 807 |
+
"step": 560
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.37,
|
| 811 |
+
"learning_rate": 3.948590904527689e-06,
|
| 812 |
+
"logits/chosen": -1.879133939743042,
|
| 813 |
+
"logits/rejected": -1.824703574180603,
|
| 814 |
+
"logps/chosen": -372.97430419921875,
|
| 815 |
+
"logps/rejected": -450.3955078125,
|
| 816 |
+
"loss": 0.0187,
|
| 817 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 818 |
+
"rewards/chosen": -0.11131548881530762,
|
| 819 |
+
"rewards/margins": 0.07067877054214478,
|
| 820 |
+
"rewards/rejected": -0.1819942742586136,
|
| 821 |
+
"step": 570
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.38,
|
| 825 |
+
"learning_rate": 3.901663178335318e-06,
|
| 826 |
+
"logits/chosen": -1.7198855876922607,
|
| 827 |
+
"logits/rejected": -1.7563533782958984,
|
| 828 |
+
"logps/chosen": -334.19659423828125,
|
| 829 |
+
"logps/rejected": -416.8275451660156,
|
| 830 |
+
"loss": 0.024,
|
| 831 |
+
"rewards/accuracies": 0.625,
|
| 832 |
+
"rewards/chosen": -0.1272047609090805,
|
| 833 |
+
"rewards/margins": 0.06788735836744308,
|
| 834 |
+
"rewards/rejected": -0.195092111825943,
|
| 835 |
+
"step": 580
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.39,
|
| 839 |
+
"learning_rate": 3.854003775619142e-06,
|
| 840 |
+
"logits/chosen": -1.9564396142959595,
|
| 841 |
+
"logits/rejected": -1.7008472681045532,
|
| 842 |
+
"logps/chosen": -400.6410827636719,
|
| 843 |
+
"logps/rejected": -497.25433349609375,
|
| 844 |
+
"loss": 0.0167,
|
| 845 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 846 |
+
"rewards/chosen": -0.1666775941848755,
|
| 847 |
+
"rewards/margins": 0.10049338638782501,
|
| 848 |
+
"rewards/rejected": -0.2671709656715393,
|
| 849 |
+
"step": 590
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.39,
|
| 853 |
+
"learning_rate": 3.805637574871115e-06,
|
| 854 |
+
"logits/chosen": -1.6202869415283203,
|
| 855 |
+
"logits/rejected": -1.6643507480621338,
|
| 856 |
+
"logps/chosen": -424.1668395996094,
|
| 857 |
+
"logps/rejected": -568.9484252929688,
|
| 858 |
+
"loss": 0.0161,
|
| 859 |
+
"rewards/accuracies": 0.75,
|
| 860 |
+
"rewards/chosen": -0.1753002405166626,
|
| 861 |
+
"rewards/margins": 0.11582837998867035,
|
| 862 |
+
"rewards/rejected": -0.29112863540649414,
|
| 863 |
+
"step": 600
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.4,
|
| 867 |
+
"learning_rate": 3.7565898235359717e-06,
|
| 868 |
+
"logits/chosen": -1.801203966140747,
|
| 869 |
+
"logits/rejected": -1.4760253429412842,
|
| 870 |
+
"logps/chosen": -466.33258056640625,
|
| 871 |
+
"logps/rejected": -551.4314575195312,
|
| 872 |
+
"loss": 0.0156,
|
| 873 |
+
"rewards/accuracies": 0.675000011920929,
|
| 874 |
+
"rewards/chosen": -0.1938503533601761,
|
| 875 |
+
"rewards/margins": 0.11616852134466171,
|
| 876 |
+
"rewards/rejected": -0.3100188672542572,
|
| 877 |
+
"step": 610
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.41,
|
| 881 |
+
"learning_rate": 3.7068861248319127e-06,
|
| 882 |
+
"logits/chosen": -1.7439998388290405,
|
| 883 |
+
"logits/rejected": -1.5612130165100098,
|
| 884 |
+
"logps/chosen": -405.24310302734375,
|
| 885 |
+
"logps/rejected": -510.33258056640625,
|
| 886 |
+
"loss": 0.0156,
|
| 887 |
+
"rewards/accuracies": 0.800000011920929,
|
| 888 |
+
"rewards/chosen": -0.1456567347049713,
|
| 889 |
+
"rewards/margins": 0.11440464109182358,
|
| 890 |
+
"rewards/rejected": -0.2600614130496979,
|
| 891 |
+
"step": 620
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.41,
|
| 895 |
+
"learning_rate": 3.6565524243855695e-06,
|
| 896 |
+
"logits/chosen": -1.8692495822906494,
|
| 897 |
+
"logits/rejected": -1.547742247581482,
|
| 898 |
+
"logps/chosen": -383.50909423828125,
|
| 899 |
+
"logps/rejected": -442.6529846191406,
|
| 900 |
+
"loss": 0.0148,
|
| 901 |
+
"rewards/accuracies": 0.699999988079071,
|
| 902 |
+
"rewards/chosen": -0.11476878821849823,
|
| 903 |
+
"rewards/margins": 0.10948886722326279,
|
| 904 |
+
"rewards/rejected": -0.2242576777935028,
|
| 905 |
+
"step": 630
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.42,
|
| 909 |
+
"learning_rate": 3.6056149966882325e-06,
|
| 910 |
+
"logits/chosen": -1.894252061843872,
|
| 911 |
+
"logits/rejected": -1.7816390991210938,
|
| 912 |
+
"logps/chosen": -362.1920471191406,
|
| 913 |
+
"logps/rejected": -425.6106872558594,
|
| 914 |
+
"loss": 0.0233,
|
| 915 |
+
"rewards/accuracies": 0.699999988079071,
|
| 916 |
+
"rewards/chosen": -0.11883090436458588,
|
| 917 |
+
"rewards/margins": 0.07852541655302048,
|
| 918 |
+
"rewards/rejected": -0.19735632836818695,
|
| 919 |
+
"step": 640
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.43,
|
| 923 |
+
"learning_rate": 3.554100431380414e-06,
|
| 924 |
+
"logits/chosen": -1.7054016590118408,
|
| 925 |
+
"logits/rejected": -1.615083932876587,
|
| 926 |
+
"logps/chosen": -408.5704345703125,
|
| 927 |
+
"logps/rejected": -545.007568359375,
|
| 928 |
+
"loss": 0.0212,
|
| 929 |
+
"rewards/accuracies": 0.800000011920929,
|
| 930 |
+
"rewards/chosen": -0.12973299622535706,
|
| 931 |
+
"rewards/margins": 0.11624966561794281,
|
| 932 |
+
"rewards/rejected": -0.24598267674446106,
|
| 933 |
+
"step": 650
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.43,
|
| 937 |
+
"learning_rate": 3.5020356193718934e-06,
|
| 938 |
+
"logits/chosen": -1.8543107509613037,
|
| 939 |
+
"logits/rejected": -1.5251867771148682,
|
| 940 |
+
"logps/chosen": -361.87066650390625,
|
| 941 |
+
"logps/rejected": -412.02947998046875,
|
| 942 |
+
"loss": 0.0176,
|
| 943 |
+
"rewards/accuracies": 0.675000011920929,
|
| 944 |
+
"rewards/chosen": -0.10308834165334702,
|
| 945 |
+
"rewards/margins": 0.07799768447875977,
|
| 946 |
+
"rewards/rejected": -0.18108603358268738,
|
| 947 |
+
"step": 660
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.44,
|
| 951 |
+
"learning_rate": 3.4494477388045035e-06,
|
| 952 |
+
"logits/chosen": -1.6554620265960693,
|
| 953 |
+
"logits/rejected": -1.5981006622314453,
|
| 954 |
+
"logps/chosen": -466.32623291015625,
|
| 955 |
+
"logps/rejected": -572.7677612304688,
|
| 956 |
+
"loss": 0.0317,
|
| 957 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 958 |
+
"rewards/chosen": -0.16580899059772491,
|
| 959 |
+
"rewards/margins": 0.10259596258401871,
|
| 960 |
+
"rewards/rejected": -0.26840493083000183,
|
| 961 |
+
"step": 670
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.44,
|
| 965 |
+
"learning_rate": 3.3963642408649783e-06,
|
| 966 |
+
"logits/chosen": -1.904425024986267,
|
| 967 |
+
"logits/rejected": -1.8261890411376953,
|
| 968 |
+
"logps/chosen": -458.0430603027344,
|
| 969 |
+
"logps/rejected": -514.0013427734375,
|
| 970 |
+
"loss": 0.0147,
|
| 971 |
+
"rewards/accuracies": 0.675000011920929,
|
| 972 |
+
"rewards/chosen": -0.12817558646202087,
|
| 973 |
+
"rewards/margins": 0.08143512904644012,
|
| 974 |
+
"rewards/rejected": -0.2096107006072998,
|
| 975 |
+
"step": 680
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.45,
|
| 979 |
+
"learning_rate": 3.3428128354552727e-06,
|
| 980 |
+
"logits/chosen": -1.8188130855560303,
|
| 981 |
+
"logits/rejected": -1.619821548461914,
|
| 982 |
+
"logps/chosen": -363.2087097167969,
|
| 983 |
+
"logps/rejected": -493.71844482421875,
|
| 984 |
+
"loss": 0.0282,
|
| 985 |
+
"rewards/accuracies": 0.699999988079071,
|
| 986 |
+
"rewards/chosen": -0.09969594329595566,
|
| 987 |
+
"rewards/margins": 0.11581007391214371,
|
| 988 |
+
"rewards/rejected": -0.21550603210926056,
|
| 989 |
+
"step": 690
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.46,
|
| 993 |
+
"learning_rate": 3.2888214767278246e-06,
|
| 994 |
+
"logits/chosen": -1.7933197021484375,
|
| 995 |
+
"logits/rejected": -1.7083864212036133,
|
| 996 |
+
"logps/chosen": -396.96832275390625,
|
| 997 |
+
"logps/rejected": -464.05462646484375,
|
| 998 |
+
"loss": 0.0156,
|
| 999 |
+
"rewards/accuracies": 0.625,
|
| 1000 |
+
"rewards/chosen": -0.1241300106048584,
|
| 1001 |
+
"rewards/margins": 0.06391225755214691,
|
| 1002 |
+
"rewards/rejected": -0.1880422681570053,
|
| 1003 |
+
"step": 700
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.46,
|
| 1007 |
+
"learning_rate": 3.2344183484933247e-06,
|
| 1008 |
+
"logits/chosen": -1.863983392715454,
|
| 1009 |
+
"logits/rejected": -1.502881646156311,
|
| 1010 |
+
"logps/chosen": -404.3202209472656,
|
| 1011 |
+
"logps/rejected": -506.7728576660156,
|
| 1012 |
+
"loss": 0.0139,
|
| 1013 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1014 |
+
"rewards/chosen": -0.1348416805267334,
|
| 1015 |
+
"rewards/margins": 0.10185597836971283,
|
| 1016 |
+
"rewards/rejected": -0.23669764399528503,
|
| 1017 |
+
"step": 710
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.47,
|
| 1021 |
+
"learning_rate": 3.179631849508597e-06,
|
| 1022 |
+
"logits/chosen": -1.732456922531128,
|
| 1023 |
+
"logits/rejected": -1.5833830833435059,
|
| 1024 |
+
"logps/chosen": -370.2289123535156,
|
| 1025 |
+
"logps/rejected": -396.13299560546875,
|
| 1026 |
+
"loss": 0.0196,
|
| 1027 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 1028 |
+
"rewards/chosen": -0.10813581943511963,
|
| 1029 |
+
"rewards/margins": 0.049150481820106506,
|
| 1030 |
+
"rewards/rejected": -0.15728630125522614,
|
| 1031 |
+
"step": 720
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.48,
|
| 1035 |
+
"learning_rate": 3.1244905786522796e-06,
|
| 1036 |
+
"logits/chosen": -1.7448046207427979,
|
| 1037 |
+
"logits/rejected": -1.7952296733856201,
|
| 1038 |
+
"logps/chosen": -374.96612548828125,
|
| 1039 |
+
"logps/rejected": -485.88427734375,
|
| 1040 |
+
"loss": 0.0221,
|
| 1041 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1042 |
+
"rewards/chosen": -0.12416452169418335,
|
| 1043 |
+
"rewards/margins": 0.07095464318990707,
|
| 1044 |
+
"rewards/rejected": -0.19511918723583221,
|
| 1045 |
+
"step": 730
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.48,
|
| 1049 |
+
"learning_rate": 3.0690233199960393e-06,
|
| 1050 |
+
"logits/chosen": -1.8050892353057861,
|
| 1051 |
+
"logits/rejected": -1.51162588596344,
|
| 1052 |
+
"logps/chosen": -443.24102783203125,
|
| 1053 |
+
"logps/rejected": -499.46417236328125,
|
| 1054 |
+
"loss": 0.0134,
|
| 1055 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1056 |
+
"rewards/chosen": -0.14171595871448517,
|
| 1057 |
+
"rewards/margins": 0.12594462931156158,
|
| 1058 |
+
"rewards/rejected": -0.26766055822372437,
|
| 1059 |
+
"step": 740
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.49,
|
| 1063 |
+
"learning_rate": 3.0132590277791163e-06,
|
| 1064 |
+
"logits/chosen": -2.1219112873077393,
|
| 1065 |
+
"logits/rejected": -1.8001441955566406,
|
| 1066 |
+
"logps/chosen": -450.737060546875,
|
| 1067 |
+
"logps/rejected": -471.3848571777344,
|
| 1068 |
+
"loss": 0.0161,
|
| 1069 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1070 |
+
"rewards/chosen": -0.16006456315517426,
|
| 1071 |
+
"rewards/margins": 0.05473906546831131,
|
| 1072 |
+
"rewards/rejected": -0.21480365097522736,
|
| 1073 |
+
"step": 750
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.5,
|
| 1077 |
+
"learning_rate": 2.9572268112940354e-06,
|
| 1078 |
+
"logits/chosen": -1.9351593255996704,
|
| 1079 |
+
"logits/rejected": -1.6238371133804321,
|
| 1080 |
+
"logps/chosen": -431.12701416015625,
|
| 1081 |
+
"logps/rejected": -458.659423828125,
|
| 1082 |
+
"loss": 0.0207,
|
| 1083 |
+
"rewards/accuracies": 0.675000011920929,
|
| 1084 |
+
"rewards/chosen": -0.1289026290178299,
|
| 1085 |
+
"rewards/margins": 0.08564122021198273,
|
| 1086 |
+
"rewards/rejected": -0.21454384922981262,
|
| 1087 |
+
"step": 760
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.5,
|
| 1091 |
+
"learning_rate": 2.9009559196913882e-06,
|
| 1092 |
+
"logits/chosen": -1.8791773319244385,
|
| 1093 |
+
"logits/rejected": -1.7700687646865845,
|
| 1094 |
+
"logps/chosen": -350.3512268066406,
|
| 1095 |
+
"logps/rejected": -448.70513916015625,
|
| 1096 |
+
"loss": 0.0153,
|
| 1097 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1098 |
+
"rewards/chosen": -0.13599291443824768,
|
| 1099 |
+
"rewards/margins": 0.08505571633577347,
|
| 1100 |
+
"rewards/rejected": -0.22104862332344055,
|
| 1101 |
+
"step": 770
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.51,
|
| 1105 |
+
"learning_rate": 2.844475726711595e-06,
|
| 1106 |
+
"logits/chosen": -1.8688617944717407,
|
| 1107 |
+
"logits/rejected": -1.6239147186279297,
|
| 1108 |
+
"logps/chosen": -379.91534423828125,
|
| 1109 |
+
"logps/rejected": -455.30792236328125,
|
| 1110 |
+
"loss": 0.0186,
|
| 1111 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1112 |
+
"rewards/chosen": -0.1507028043270111,
|
| 1113 |
+
"rewards/margins": 0.09271397441625595,
|
| 1114 |
+
"rewards/rejected": -0.24341675639152527,
|
| 1115 |
+
"step": 780
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.52,
|
| 1119 |
+
"learning_rate": 2.7878157153516446e-06,
|
| 1120 |
+
"logits/chosen": -1.8980302810668945,
|
| 1121 |
+
"logits/rejected": -1.4565205574035645,
|
| 1122 |
+
"logps/chosen": -368.5240478515625,
|
| 1123 |
+
"logps/rejected": -475.89013671875,
|
| 1124 |
+
"loss": 0.0203,
|
| 1125 |
+
"rewards/accuracies": 0.675000011920929,
|
| 1126 |
+
"rewards/chosen": -0.12418484687805176,
|
| 1127 |
+
"rewards/margins": 0.12127965688705444,
|
| 1128 |
+
"rewards/rejected": -0.2454645186662674,
|
| 1129 |
+
"step": 790
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.52,
|
| 1133 |
+
"learning_rate": 2.731005462474787e-06,
|
| 1134 |
+
"logits/chosen": -2.0655064582824707,
|
| 1135 |
+
"logits/rejected": -1.778184175491333,
|
| 1136 |
+
"logps/chosen": -362.5508728027344,
|
| 1137 |
+
"logps/rejected": -415.6459045410156,
|
| 1138 |
+
"loss": 0.0307,
|
| 1139 |
+
"rewards/accuracies": 0.75,
|
| 1140 |
+
"rewards/chosen": -0.11265746504068375,
|
| 1141 |
+
"rewards/margins": 0.09785433113574982,
|
| 1142 |
+
"rewards/rejected": -0.21051180362701416,
|
| 1143 |
+
"step": 800
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.53,
|
| 1147 |
+
"learning_rate": 2.67407462337124e-06,
|
| 1148 |
+
"logits/chosen": -1.669070839881897,
|
| 1149 |
+
"logits/rejected": -1.6044425964355469,
|
| 1150 |
+
"logps/chosen": -367.7320556640625,
|
| 1151 |
+
"logps/rejected": -497.50970458984375,
|
| 1152 |
+
"loss": 0.0199,
|
| 1153 |
+
"rewards/accuracies": 0.625,
|
| 1154 |
+
"rewards/chosen": -0.12311694771051407,
|
| 1155 |
+
"rewards/margins": 0.12384019047021866,
|
| 1156 |
+
"rewards/rejected": -0.24695713818073273,
|
| 1157 |
+
"step": 810
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.54,
|
| 1161 |
+
"learning_rate": 2.617052916277952e-06,
|
| 1162 |
+
"logits/chosen": -2.0958499908447266,
|
| 1163 |
+
"logits/rejected": -1.8438915014266968,
|
| 1164 |
+
"logps/chosen": -428.4202575683594,
|
| 1165 |
+
"logps/rejected": -517.01025390625,
|
| 1166 |
+
"loss": 0.0196,
|
| 1167 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 1168 |
+
"rewards/chosen": -0.11365973949432373,
|
| 1169 |
+
"rewards/margins": 0.0948016420006752,
|
| 1170 |
+
"rewards/rejected": -0.20846137404441833,
|
| 1171 |
+
"step": 820
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.54,
|
| 1175 |
+
"learning_rate": 2.5599701068654985e-06,
|
| 1176 |
+
"logits/chosen": -1.7646852731704712,
|
| 1177 |
+
"logits/rejected": -1.435537338256836,
|
| 1178 |
+
"logps/chosen": -422.580810546875,
|
| 1179 |
+
"logps/rejected": -500.8734436035156,
|
| 1180 |
+
"loss": 0.0201,
|
| 1181 |
+
"rewards/accuracies": 0.625,
|
| 1182 |
+
"rewards/chosen": -0.14953741431236267,
|
| 1183 |
+
"rewards/margins": 0.11128922551870346,
|
| 1184 |
+
"rewards/rejected": -0.2608266770839691,
|
| 1185 |
+
"step": 830
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.55,
|
| 1189 |
+
"learning_rate": 2.5028559927002326e-06,
|
| 1190 |
+
"logits/chosen": -2.0240190029144287,
|
| 1191 |
+
"logits/rejected": -1.542840600013733,
|
| 1192 |
+
"logps/chosen": -451.66876220703125,
|
| 1193 |
+
"logps/rejected": -516.1002197265625,
|
| 1194 |
+
"loss": 0.0174,
|
| 1195 |
+
"rewards/accuracies": 0.699999988079071,
|
| 1196 |
+
"rewards/chosen": -0.1411553919315338,
|
| 1197 |
+
"rewards/margins": 0.13159087300300598,
|
| 1198 |
+
"rewards/rejected": -0.2727462351322174,
|
| 1199 |
+
"step": 840
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.56,
|
| 1203 |
+
"learning_rate": 2.4457403876897756e-06,
|
| 1204 |
+
"logits/chosen": -1.6757514476776123,
|
| 1205 |
+
"logits/rejected": -1.4124213457107544,
|
| 1206 |
+
"logps/chosen": -414.86883544921875,
|
| 1207 |
+
"logps/rejected": -541.7101440429688,
|
| 1208 |
+
"loss": 0.0145,
|
| 1209 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 1210 |
+
"rewards/chosen": -0.15753154456615448,
|
| 1211 |
+
"rewards/margins": 0.1300159990787506,
|
| 1212 |
+
"rewards/rejected": -0.2875475585460663,
|
| 1213 |
+
"step": 850
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.56,
|
| 1217 |
+
"learning_rate": 2.388653106519975e-06,
|
| 1218 |
+
"logits/chosen": -1.888702630996704,
|
| 1219 |
+
"logits/rejected": -1.7808609008789062,
|
| 1220 |
+
"logps/chosen": -413.09295654296875,
|
| 1221 |
+
"logps/rejected": -521.931884765625,
|
| 1222 |
+
"loss": 0.0194,
|
| 1223 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1224 |
+
"rewards/chosen": -0.15543514490127563,
|
| 1225 |
+
"rewards/margins": 0.09803062677383423,
|
| 1226 |
+
"rewards/rejected": -0.2534657418727875,
|
| 1227 |
+
"step": 860
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.57,
|
| 1231 |
+
"learning_rate": 2.331623949091467e-06,
|
| 1232 |
+
"logits/chosen": -1.8230218887329102,
|
| 1233 |
+
"logits/rejected": -1.629063367843628,
|
| 1234 |
+
"logps/chosen": -441.42034912109375,
|
| 1235 |
+
"logps/rejected": -485.918212890625,
|
| 1236 |
+
"loss": 0.0239,
|
| 1237 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1238 |
+
"rewards/chosen": -0.12905916571617126,
|
| 1239 |
+
"rewards/margins": 0.10947714000940323,
|
| 1240 |
+
"rewards/rejected": -0.2385363131761551,
|
| 1241 |
+
"step": 870
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.58,
|
| 1245 |
+
"learning_rate": 2.2746826849639513e-06,
|
| 1246 |
+
"logits/chosen": -2.148059129714966,
|
| 1247 |
+
"logits/rejected": -1.7364435195922852,
|
| 1248 |
+
"logps/chosen": -456.3299255371094,
|
| 1249 |
+
"logps/rejected": -466.22515869140625,
|
| 1250 |
+
"loss": 0.0189,
|
| 1251 |
+
"rewards/accuracies": 0.699999988079071,
|
| 1252 |
+
"rewards/chosen": -0.1590055525302887,
|
| 1253 |
+
"rewards/margins": 0.06117180734872818,
|
| 1254 |
+
"rewards/rejected": -0.22017733752727509,
|
| 1255 |
+
"step": 880
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.58,
|
| 1259 |
+
"learning_rate": 2.2178590378162957e-06,
|
| 1260 |
+
"logits/chosen": -1.765479326248169,
|
| 1261 |
+
"logits/rejected": -1.6112077236175537,
|
| 1262 |
+
"logps/chosen": -437.97381591796875,
|
| 1263 |
+
"logps/rejected": -574.2666015625,
|
| 1264 |
+
"loss": 0.0227,
|
| 1265 |
+
"rewards/accuracies": 0.675000011920929,
|
| 1266 |
+
"rewards/chosen": -0.1820591539144516,
|
| 1267 |
+
"rewards/margins": 0.1335388422012329,
|
| 1268 |
+
"rewards/rejected": -0.3155979812145233,
|
| 1269 |
+
"step": 890
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.59,
|
| 1273 |
+
"learning_rate": 2.1611826699306104e-06,
|
| 1274 |
+
"logits/chosen": -1.8870223760604858,
|
| 1275 |
+
"logits/rejected": -1.520627737045288,
|
| 1276 |
+
"logps/chosen": -420.63323974609375,
|
| 1277 |
+
"logps/rejected": -466.6917419433594,
|
| 1278 |
+
"loss": 0.0155,
|
| 1279 |
+
"rewards/accuracies": 0.625,
|
| 1280 |
+
"rewards/chosen": -0.15777216851711273,
|
| 1281 |
+
"rewards/margins": 0.05299310013651848,
|
| 1282 |
+
"rewards/rejected": -0.2107652723789215,
|
| 1283 |
+
"step": 900
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.6,
|
| 1287 |
+
"learning_rate": 2.1046831667083483e-06,
|
| 1288 |
+
"logits/chosen": -1.8595775365829468,
|
| 1289 |
+
"logits/rejected": -1.6033302545547485,
|
| 1290 |
+
"logps/chosen": -453.5380859375,
|
| 1291 |
+
"logps/rejected": -464.77398681640625,
|
| 1292 |
+
"loss": 0.0291,
|
| 1293 |
+
"rewards/accuracies": 0.699999988079071,
|
| 1294 |
+
"rewards/chosen": -0.16525158286094666,
|
| 1295 |
+
"rewards/margins": 0.07691897451877594,
|
| 1296 |
+
"rewards/rejected": -0.2421705275774002,
|
| 1297 |
+
"step": 910
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.6,
|
| 1301 |
+
"learning_rate": 2.048390021226559e-06,
|
| 1302 |
+
"logits/chosen": -1.9039933681488037,
|
| 1303 |
+
"logits/rejected": -1.672969102859497,
|
| 1304 |
+
"logps/chosen": -395.4986572265625,
|
| 1305 |
+
"logps/rejected": -436.7999572753906,
|
| 1306 |
+
"loss": 0.0178,
|
| 1307 |
+
"rewards/accuracies": 0.675000011920929,
|
| 1308 |
+
"rewards/chosen": -0.1351805031299591,
|
| 1309 |
+
"rewards/margins": 0.08283836394548416,
|
| 1310 |
+
"rewards/rejected": -0.21801885962486267,
|
| 1311 |
+
"step": 920
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.61,
|
| 1315 |
+
"learning_rate": 1.9923326188423212e-06,
|
| 1316 |
+
"logits/chosen": -1.8809058666229248,
|
| 1317 |
+
"logits/rejected": -1.6939058303833008,
|
| 1318 |
+
"logps/chosen": -359.4674987792969,
|
| 1319 |
+
"logps/rejected": -431.02752685546875,
|
| 1320 |
+
"loss": 0.0137,
|
| 1321 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1322 |
+
"rewards/chosen": -0.11502976715564728,
|
| 1323 |
+
"rewards/margins": 0.09757871180772781,
|
| 1324 |
+
"rewards/rejected": -0.2126084864139557,
|
| 1325 |
+
"step": 930
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 0.61,
|
| 1329 |
+
"learning_rate": 1.936540221853415e-06,
|
| 1330 |
+
"logits/chosen": -1.687034249305725,
|
| 1331 |
+
"logits/rejected": -1.7980533838272095,
|
| 1332 |
+
"logps/chosen": -385.8014221191406,
|
| 1333 |
+
"logps/rejected": -513.2298583984375,
|
| 1334 |
+
"loss": 0.0179,
|
| 1335 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 1336 |
+
"rewards/chosen": -0.12754614651203156,
|
| 1337 |
+
"rewards/margins": 0.10141207277774811,
|
| 1338 |
+
"rewards/rejected": -0.22895821928977966,
|
| 1339 |
+
"step": 940
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.62,
|
| 1343 |
+
"learning_rate": 1.8810419542232245e-06,
|
| 1344 |
+
"logits/chosen": -1.832686424255371,
|
| 1345 |
+
"logits/rejected": -1.7866284847259521,
|
| 1346 |
+
"logps/chosen": -375.6191711425781,
|
| 1347 |
+
"logps/rejected": -473.59747314453125,
|
| 1348 |
+
"loss": 0.0143,
|
| 1349 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 1350 |
+
"rewards/chosen": -0.13123920559883118,
|
| 1351 |
+
"rewards/margins": 0.07887652516365051,
|
| 1352 |
+
"rewards/rejected": -0.2101157158613205,
|
| 1353 |
+
"step": 950
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 0.63,
|
| 1357 |
+
"learning_rate": 1.8258667863778573e-06,
|
| 1358 |
+
"logits/chosen": -1.8020689487457275,
|
| 1359 |
+
"logits/rejected": -1.5563127994537354,
|
| 1360 |
+
"logps/chosen": -391.3179931640625,
|
| 1361 |
+
"logps/rejected": -437.33990478515625,
|
| 1362 |
+
"loss": 0.0253,
|
| 1363 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1364 |
+
"rewards/chosen": -0.15354005992412567,
|
| 1365 |
+
"rewards/margins": 0.07208997011184692,
|
| 1366 |
+
"rewards/rejected": -0.2256300449371338,
|
| 1367 |
+
"step": 960
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.63,
|
| 1371 |
+
"learning_rate": 1.7710435200834126e-06,
|
| 1372 |
+
"logits/chosen": -1.8209571838378906,
|
| 1373 |
+
"logits/rejected": -1.6538410186767578,
|
| 1374 |
+
"logps/chosen": -382.71331787109375,
|
| 1375 |
+
"logps/rejected": -439.09765625,
|
| 1376 |
+
"loss": 0.0171,
|
| 1377 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 1378 |
+
"rewards/chosen": -0.13624554872512817,
|
| 1379 |
+
"rewards/margins": 0.08001069724559784,
|
| 1380 |
+
"rewards/rejected": -0.2162562608718872,
|
| 1381 |
+
"step": 970
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 0.64,
|
| 1385 |
+
"learning_rate": 1.7166007734112808e-06,
|
| 1386 |
+
"logits/chosen": -1.6404680013656616,
|
| 1387 |
+
"logits/rejected": -1.4822646379470825,
|
| 1388 |
+
"logps/chosen": -394.8628845214844,
|
| 1389 |
+
"logps/rejected": -502.18524169921875,
|
| 1390 |
+
"loss": 0.0203,
|
| 1391 |
+
"rewards/accuracies": 0.699999988079071,
|
| 1392 |
+
"rewards/chosen": -0.15191176533699036,
|
| 1393 |
+
"rewards/margins": 0.10607035458087921,
|
| 1394 |
+
"rewards/rejected": -0.2579821050167084,
|
| 1395 |
+
"step": 980
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 0.65,
|
| 1399 |
+
"learning_rate": 1.6625669657993483e-06,
|
| 1400 |
+
"logits/chosen": -1.8440077304840088,
|
| 1401 |
+
"logits/rejected": -1.6627609729766846,
|
| 1402 |
+
"logps/chosen": -379.8057556152344,
|
| 1403 |
+
"logps/rejected": -463.258544921875,
|
| 1404 |
+
"loss": 0.0367,
|
| 1405 |
+
"rewards/accuracies": 0.699999988079071,
|
| 1406 |
+
"rewards/chosen": -0.11155873537063599,
|
| 1407 |
+
"rewards/margins": 0.09738429635763168,
|
| 1408 |
+
"rewards/rejected": -0.20894300937652588,
|
| 1409 |
+
"step": 990
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 0.65,
|
| 1413 |
+
"learning_rate": 1.6089703032168736e-06,
|
| 1414 |
+
"logits/chosen": -1.7317651510238647,
|
| 1415 |
+
"logits/rejected": -1.6905380487442017,
|
| 1416 |
+
"logps/chosen": -416.31622314453125,
|
| 1417 |
+
"logps/rejected": -504.07415771484375,
|
| 1418 |
+
"loss": 0.0191,
|
| 1419 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 1420 |
+
"rewards/chosen": -0.13056668639183044,
|
| 1421 |
+
"rewards/margins": 0.1028464064002037,
|
| 1422 |
+
"rewards/rejected": -0.23341310024261475,
|
| 1423 |
+
"step": 1000
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 0.66,
|
| 1427 |
+
"learning_rate": 1.55583876344081e-06,
|
| 1428 |
+
"logits/chosen": -1.8492292165756226,
|
| 1429 |
+
"logits/rejected": -1.7530572414398193,
|
| 1430 |
+
"logps/chosen": -425.2975158691406,
|
| 1431 |
+
"logps/rejected": -527.4212646484375,
|
| 1432 |
+
"loss": 0.0189,
|
| 1433 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 1434 |
+
"rewards/chosen": -0.14273019134998322,
|
| 1435 |
+
"rewards/margins": 0.0867033302783966,
|
| 1436 |
+
"rewards/rejected": -0.22943353652954102,
|
| 1437 |
+
"step": 1010
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 0.67,
|
| 1441 |
+
"learning_rate": 1.5032000814512372e-06,
|
| 1442 |
+
"logits/chosen": -1.8472225666046143,
|
| 1443 |
+
"logits/rejected": -1.8299192190170288,
|
| 1444 |
+
"logps/chosen": -367.6811828613281,
|
| 1445 |
+
"logps/rejected": -439.68975830078125,
|
| 1446 |
+
"loss": 0.0232,
|
| 1447 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 1448 |
+
"rewards/chosen": -0.14622075855731964,
|
| 1449 |
+
"rewards/margins": 0.07023419439792633,
|
| 1450 |
+
"rewards/rejected": -0.21645493805408478,
|
| 1451 |
+
"step": 1020
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 0.67,
|
| 1455 |
+
"learning_rate": 1.4510817349535323e-06,
|
| 1456 |
+
"logits/chosen": -1.8069846630096436,
|
| 1457 |
+
"logits/rejected": -1.733162522315979,
|
| 1458 |
+
"logps/chosen": -352.75054931640625,
|
| 1459 |
+
"logps/rejected": -439.2781677246094,
|
| 1460 |
+
"loss": 0.023,
|
| 1461 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1462 |
+
"rewards/chosen": -0.12202004343271255,
|
| 1463 |
+
"rewards/margins": 0.07520598918199539,
|
| 1464 |
+
"rewards/rejected": -0.19722604751586914,
|
| 1465 |
+
"step": 1030
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 0.68,
|
| 1469 |
+
"learning_rate": 1.3995109300348537e-06,
|
| 1470 |
+
"logits/chosen": -1.7219600677490234,
|
| 1471 |
+
"logits/rejected": -1.5683410167694092,
|
| 1472 |
+
"logps/chosen": -397.7275085449219,
|
| 1473 |
+
"logps/rejected": -526.4014282226562,
|
| 1474 |
+
"loss": 0.0233,
|
| 1475 |
+
"rewards/accuracies": 0.824999988079071,
|
| 1476 |
+
"rewards/chosen": -0.15109822154045105,
|
| 1477 |
+
"rewards/margins": 0.12963151931762695,
|
| 1478 |
+
"rewards/rejected": -0.280729740858078,
|
| 1479 |
+
"step": 1040
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 0.69,
|
| 1483 |
+
"learning_rate": 1.348514586962389e-06,
|
| 1484 |
+
"logits/chosen": -1.9068981409072876,
|
| 1485 |
+
"logits/rejected": -1.6539924144744873,
|
| 1486 |
+
"logps/chosen": -451.5171813964844,
|
| 1487 |
+
"logps/rejected": -513.7973022460938,
|
| 1488 |
+
"loss": 0.0224,
|
| 1489 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 1490 |
+
"rewards/chosen": -0.18701951205730438,
|
| 1491 |
+
"rewards/margins": 0.063180111348629,
|
| 1492 |
+
"rewards/rejected": -0.25019964575767517,
|
| 1493 |
+
"step": 1050
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 0.69,
|
| 1497 |
+
"learning_rate": 1.2981193261308284e-06,
|
| 1498 |
+
"logits/chosen": -1.601199746131897,
|
| 1499 |
+
"logits/rejected": -1.4806240797042847,
|
| 1500 |
+
"logps/chosen": -401.56243896484375,
|
| 1501 |
+
"logps/rejected": -576.2053833007812,
|
| 1502 |
+
"loss": 0.0208,
|
| 1503 |
+
"rewards/accuracies": 0.75,
|
| 1504 |
+
"rewards/chosen": -0.1684935986995697,
|
| 1505 |
+
"rewards/margins": 0.1360427886247635,
|
| 1506 |
+
"rewards/rejected": -0.304536372423172,
|
| 1507 |
+
"step": 1060
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 0.7,
|
| 1511 |
+
"learning_rate": 1.2483514541663501e-06,
|
| 1512 |
+
"logits/chosen": -1.7805836200714111,
|
| 1513 |
+
"logits/rejected": -1.636907935142517,
|
| 1514 |
+
"logps/chosen": -396.3351135253906,
|
| 1515 |
+
"logps/rejected": -458.74725341796875,
|
| 1516 |
+
"loss": 0.0119,
|
| 1517 |
+
"rewards/accuracies": 0.675000011920929,
|
| 1518 |
+
"rewards/chosen": -0.13784000277519226,
|
| 1519 |
+
"rewards/margins": 0.08524910360574722,
|
| 1520 |
+
"rewards/rejected": -0.22308912873268127,
|
| 1521 |
+
"step": 1070
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 0.71,
|
| 1525 |
+
"learning_rate": 1.1992369501944096e-06,
|
| 1526 |
+
"logits/chosen": -1.8869998455047607,
|
| 1527 |
+
"logits/rejected": -1.8042247295379639,
|
| 1528 |
+
"logps/chosen": -363.39788818359375,
|
| 1529 |
+
"logps/rejected": -539.3111572265625,
|
| 1530 |
+
"loss": 0.0213,
|
| 1531 |
+
"rewards/accuracies": 0.699999988079071,
|
| 1532 |
+
"rewards/chosen": -0.14907002449035645,
|
| 1533 |
+
"rewards/margins": 0.14250054955482483,
|
| 1534 |
+
"rewards/rejected": -0.29157060384750366,
|
| 1535 |
+
"step": 1080
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 0.71,
|
| 1539 |
+
"learning_rate": 1.1508014522784803e-06,
|
| 1540 |
+
"logits/chosen": -1.8102096319198608,
|
| 1541 |
+
"logits/rejected": -1.695726752281189,
|
| 1542 |
+
"logps/chosen": -443.92327880859375,
|
| 1543 |
+
"logps/rejected": -519.6880493164062,
|
| 1544 |
+
"loss": 0.0199,
|
| 1545 |
+
"rewards/accuracies": 0.800000011920929,
|
| 1546 |
+
"rewards/chosen": -0.1804676353931427,
|
| 1547 |
+
"rewards/margins": 0.09004662185907364,
|
| 1548 |
+
"rewards/rejected": -0.27051427960395813,
|
| 1549 |
+
"step": 1090
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 0.72,
|
| 1553 |
+
"learning_rate": 1.1030702440368319e-06,
|
| 1554 |
+
"logits/chosen": -1.5621861219406128,
|
| 1555 |
+
"logits/rejected": -1.5408356189727783,
|
| 1556 |
+
"logps/chosen": -452.7931213378906,
|
| 1557 |
+
"logps/rejected": -742.4008178710938,
|
| 1558 |
+
"loss": 0.0272,
|
| 1559 |
+
"rewards/accuracies": 0.75,
|
| 1560 |
+
"rewards/chosen": -0.18929113447666168,
|
| 1561 |
+
"rewards/margins": 0.12937882542610168,
|
| 1562 |
+
"rewards/rejected": -0.31866997480392456,
|
| 1563 |
+
"step": 1100
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 0.73,
|
| 1567 |
+
"learning_rate": 1.0560682414443315e-06,
|
| 1568 |
+
"logits/chosen": -1.8824927806854248,
|
| 1569 |
+
"logits/rejected": -1.7471214532852173,
|
| 1570 |
+
"logps/chosen": -460.31640625,
|
| 1571 |
+
"logps/rejected": -539.2716064453125,
|
| 1572 |
+
"loss": 0.0114,
|
| 1573 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 1574 |
+
"rewards/chosen": -0.1245940774679184,
|
| 1575 |
+
"rewards/margins": 0.12403732538223267,
|
| 1576 |
+
"rewards/rejected": -0.24863143265247345,
|
| 1577 |
+
"step": 1110
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 0.73,
|
| 1581 |
+
"learning_rate": 1.009819979826156e-06,
|
| 1582 |
+
"logits/chosen": -1.9311294555664062,
|
| 1583 |
+
"logits/rejected": -1.5975712537765503,
|
| 1584 |
+
"logps/chosen": -442.8894958496094,
|
| 1585 |
+
"logps/rejected": -503.81048583984375,
|
| 1586 |
+
"loss": 0.0304,
|
| 1587 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 1588 |
+
"rewards/chosen": -0.1375322788953781,
|
| 1589 |
+
"rewards/margins": 0.08977995812892914,
|
| 1590 |
+
"rewards/rejected": -0.22731228172779083,
|
| 1591 |
+
"step": 1120
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 0.74,
|
| 1595 |
+
"learning_rate": 9.643496010502054e-07,
|
| 1596 |
+
"logits/chosen": -1.8107373714447021,
|
| 1597 |
+
"logits/rejected": -1.6052519083023071,
|
| 1598 |
+
"logps/chosen": -426.5626525878906,
|
| 1599 |
+
"logps/rejected": -484.64971923828125,
|
| 1600 |
+
"loss": 0.0149,
|
| 1601 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1602 |
+
"rewards/chosen": -0.13215045630931854,
|
| 1603 |
+
"rewards/margins": 0.09739480167627335,
|
| 1604 |
+
"rewards/rejected": -0.2295452356338501,
|
| 1605 |
+
"step": 1130
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 0.75,
|
| 1609 |
+
"learning_rate": 9.196808409249086e-07,
|
| 1610 |
+
"logits/chosen": -1.901611566543579,
|
| 1611 |
+
"logits/rejected": -1.9235025644302368,
|
| 1612 |
+
"logps/chosen": -372.97808837890625,
|
| 1613 |
+
"logps/rejected": -540.9254150390625,
|
| 1614 |
+
"loss": 0.0176,
|
| 1615 |
+
"rewards/accuracies": 0.699999988079071,
|
| 1616 |
+
"rewards/chosen": -0.11672798544168472,
|
| 1617 |
+
"rewards/margins": 0.11135835945606232,
|
| 1618 |
+
"rewards/rejected": -0.22808635234832764,
|
| 1619 |
+
"step": 1140
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 0.75,
|
| 1623 |
+
"learning_rate": 8.758370168089797e-07,
|
| 1624 |
+
"logits/chosen": -1.8253875970840454,
|
| 1625 |
+
"logits/rejected": -1.6909106969833374,
|
| 1626 |
+
"logps/chosen": -321.0955505371094,
|
| 1627 |
+
"logps/rejected": -385.1728515625,
|
| 1628 |
+
"loss": 0.0159,
|
| 1629 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1630 |
+
"rewards/chosen": -0.12681838870048523,
|
| 1631 |
+
"rewards/margins": 0.06274305284023285,
|
| 1632 |
+
"rewards/rejected": -0.18956145644187927,
|
| 1633 |
+
"step": 1150
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 0.76,
|
| 1637 |
+
"learning_rate": 8.328410154396318e-07,
|
| 1638 |
+
"logits/chosen": -1.8118044137954712,
|
| 1639 |
+
"logits/rejected": -1.8118581771850586,
|
| 1640 |
+
"logps/chosen": -379.05615234375,
|
| 1641 |
+
"logps/rejected": -470.08184814453125,
|
| 1642 |
+
"loss": 0.0176,
|
| 1643 |
+
"rewards/accuracies": 0.675000011920929,
|
| 1644 |
+
"rewards/chosen": -0.12152254581451416,
|
| 1645 |
+
"rewards/margins": 0.06952939927577972,
|
| 1646 |
+
"rewards/rejected": -0.19105194509029388,
|
| 1647 |
+
"step": 1160
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.77,
|
| 1651 |
+
"learning_rate": 7.907152809855529e-07,
|
| 1652 |
+
"logits/chosen": -1.877508521080017,
|
| 1653 |
+
"logits/rejected": -1.6081901788711548,
|
| 1654 |
+
"logps/chosen": -366.0731201171875,
|
| 1655 |
+
"logps/rejected": -417.31683349609375,
|
| 1656 |
+
"loss": 0.0202,
|
| 1657 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1658 |
+
"rewards/chosen": -0.12633651494979858,
|
| 1659 |
+
"rewards/margins": 0.06398525089025497,
|
| 1660 |
+
"rewards/rejected": -0.19032175838947296,
|
| 1661 |
+
"step": 1170
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 0.77,
|
| 1665 |
+
"learning_rate": 7.494818033309207e-07,
|
| 1666 |
+
"logits/chosen": -1.7752797603607178,
|
| 1667 |
+
"logits/rejected": -1.6302067041397095,
|
| 1668 |
+
"logps/chosen": -360.30023193359375,
|
| 1669 |
+
"logps/rejected": -443.12908935546875,
|
| 1670 |
+
"loss": 0.0261,
|
| 1671 |
+
"rewards/accuracies": 0.574999988079071,
|
| 1672 |
+
"rewards/chosen": -0.12191496789455414,
|
| 1673 |
+
"rewards/margins": 0.07757655531167984,
|
| 1674 |
+
"rewards/rejected": -0.19949153065681458,
|
| 1675 |
+
"step": 1180
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 0.78,
|
| 1679 |
+
"learning_rate": 7.091621065965521e-07,
|
| 1680 |
+
"logits/chosen": -1.7687801122665405,
|
| 1681 |
+
"logits/rejected": -1.4826685190200806,
|
| 1682 |
+
"logps/chosen": -366.3409118652344,
|
| 1683 |
+
"logps/rejected": -412.891357421875,
|
| 1684 |
+
"loss": 0.0155,
|
| 1685 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 1686 |
+
"rewards/chosen": -0.1527003049850464,
|
| 1687 |
+
"rewards/margins": 0.08991221338510513,
|
| 1688 |
+
"rewards/rejected": -0.24261251091957092,
|
| 1689 |
+
"step": 1190
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 0.79,
|
| 1693 |
+
"learning_rate": 6.697772379041823e-07,
|
| 1694 |
+
"logits/chosen": -2.0650315284729004,
|
| 1695 |
+
"logits/rejected": -1.6713231801986694,
|
| 1696 |
+
"logps/chosen": -448.5660095214844,
|
| 1697 |
+
"logps/rejected": -507.5440368652344,
|
| 1698 |
+
"loss": 0.0299,
|
| 1699 |
+
"rewards/accuracies": 0.800000011920929,
|
| 1700 |
+
"rewards/chosen": -0.14657150208950043,
|
| 1701 |
+
"rewards/margins": 0.1087670549750328,
|
| 1702 |
+
"rewards/rejected": -0.255338579416275,
|
| 1703 |
+
"step": 1200
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 0.79,
|
| 1707 |
+
"learning_rate": 6.313477563897466e-07,
|
| 1708 |
+
"logits/chosen": -1.8000714778900146,
|
| 1709 |
+
"logits/rejected": -1.6599605083465576,
|
| 1710 |
+
"logps/chosen": -451.50140380859375,
|
| 1711 |
+
"logps/rejected": -568.575927734375,
|
| 1712 |
+
"loss": 0.0186,
|
| 1713 |
+
"rewards/accuracies": 0.699999988079071,
|
| 1714 |
+
"rewards/chosen": -0.18340735137462616,
|
| 1715 |
+
"rewards/margins": 0.10143464803695679,
|
| 1716 |
+
"rewards/rejected": -0.28484198451042175,
|
| 1717 |
+
"step": 1210
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 0.8,
|
| 1721 |
+
"learning_rate": 5.9389372247138e-07,
|
| 1722 |
+
"logits/chosen": -1.7278544902801514,
|
| 1723 |
+
"logits/rejected": -1.328420639038086,
|
| 1724 |
+
"logps/chosen": -498.06280517578125,
|
| 1725 |
+
"logps/rejected": -581.17041015625,
|
| 1726 |
+
"loss": 0.0193,
|
| 1727 |
+
"rewards/accuracies": 0.675000011920929,
|
| 1728 |
+
"rewards/chosen": -0.1877037137746811,
|
| 1729 |
+
"rewards/margins": 0.12751582264900208,
|
| 1730 |
+
"rewards/rejected": -0.315219521522522,
|
| 1731 |
+
"step": 1220
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 0.8,
|
| 1735 |
+
"learning_rate": 5.574346873777714e-07,
|
| 1736 |
+
"logits/chosen": -1.5999044179916382,
|
| 1737 |
+
"logits/rejected": -1.5874285697937012,
|
| 1738 |
+
"logps/chosen": -431.46185302734375,
|
| 1739 |
+
"logps/rejected": -516.9653930664062,
|
| 1740 |
+
"loss": 0.0167,
|
| 1741 |
+
"rewards/accuracies": 0.699999988079071,
|
| 1742 |
+
"rewards/chosen": -0.16077804565429688,
|
| 1743 |
+
"rewards/margins": 0.07463011145591736,
|
| 1744 |
+
"rewards/rejected": -0.23540815711021423,
|
| 1745 |
+
"step": 1230
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 0.81,
|
| 1749 |
+
"learning_rate": 5.219896829422927e-07,
|
| 1750 |
+
"logits/chosen": -1.6012115478515625,
|
| 1751 |
+
"logits/rejected": -1.5557236671447754,
|
| 1752 |
+
"logps/chosen": -308.88348388671875,
|
| 1753 |
+
"logps/rejected": -417.7029724121094,
|
| 1754 |
+
"loss": 0.0222,
|
| 1755 |
+
"rewards/accuracies": 0.699999988079071,
|
| 1756 |
+
"rewards/chosen": -0.14151635766029358,
|
| 1757 |
+
"rewards/margins": 0.10681549459695816,
|
| 1758 |
+
"rewards/rejected": -0.24833182990550995,
|
| 1759 |
+
"step": 1240
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 0.82,
|
| 1763 |
+
"learning_rate": 4.875772116682817e-07,
|
| 1764 |
+
"logits/chosen": -1.9544742107391357,
|
| 1765 |
+
"logits/rejected": -1.5280301570892334,
|
| 1766 |
+
"logps/chosen": -457.99078369140625,
|
| 1767 |
+
"logps/rejected": -516.620849609375,
|
| 1768 |
+
"loss": 0.0222,
|
| 1769 |
+
"rewards/accuracies": 0.625,
|
| 1770 |
+
"rewards/chosen": -0.13625690340995789,
|
| 1771 |
+
"rewards/margins": 0.0871158018708229,
|
| 1772 |
+
"rewards/rejected": -0.22337272763252258,
|
| 1773 |
+
"step": 1250
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 0.82,
|
| 1777 |
+
"learning_rate": 4.542152370706149e-07,
|
| 1778 |
+
"logits/chosen": -1.8728841543197632,
|
| 1779 |
+
"logits/rejected": -1.6816173791885376,
|
| 1780 |
+
"logps/chosen": -441.79693603515625,
|
| 1781 |
+
"logps/rejected": -553.9568481445312,
|
| 1782 |
+
"loss": 0.0127,
|
| 1783 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1784 |
+
"rewards/chosen": -0.16504457592964172,
|
| 1785 |
+
"rewards/margins": 0.11639133840799332,
|
| 1786 |
+
"rewards/rejected": -0.28143590688705444,
|
| 1787 |
+
"step": 1260
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 0.83,
|
| 1791 |
+
"learning_rate": 4.2192117429865067e-07,
|
| 1792 |
+
"logits/chosen": -2.061262845993042,
|
| 1793 |
+
"logits/rejected": -1.6205766201019287,
|
| 1794 |
+
"logps/chosen": -463.18408203125,
|
| 1795 |
+
"logps/rejected": -434.10662841796875,
|
| 1796 |
+
"loss": 0.0173,
|
| 1797 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1798 |
+
"rewards/chosen": -0.15750975906848907,
|
| 1799 |
+
"rewards/margins": 0.07638247311115265,
|
| 1800 |
+
"rewards/rejected": -0.23389224708080292,
|
| 1801 |
+
"step": 1270
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 0.84,
|
| 1805 |
+
"learning_rate": 3.907118810454172e-07,
|
| 1806 |
+
"logits/chosen": -1.746106505393982,
|
| 1807 |
+
"logits/rejected": -1.5782015323638916,
|
| 1808 |
+
"logps/chosen": -492.3348083496094,
|
| 1809 |
+
"logps/rejected": -519.789306640625,
|
| 1810 |
+
"loss": 0.0184,
|
| 1811 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1812 |
+
"rewards/chosen": -0.17052312195301056,
|
| 1813 |
+
"rewards/margins": 0.0672903060913086,
|
| 1814 |
+
"rewards/rejected": -0.23781342804431915,
|
| 1815 |
+
"step": 1280
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 0.84,
|
| 1819 |
+
"learning_rate": 3.6060364874779455e-07,
|
| 1820 |
+
"logits/chosen": -1.9814525842666626,
|
| 1821 |
+
"logits/rejected": -1.5197412967681885,
|
| 1822 |
+
"logps/chosen": -486.7601623535156,
|
| 1823 |
+
"logps/rejected": -499.59320068359375,
|
| 1824 |
+
"loss": 0.0193,
|
| 1825 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 1826 |
+
"rewards/chosen": -0.15768514573574066,
|
| 1827 |
+
"rewards/margins": 0.09786301851272583,
|
| 1828 |
+
"rewards/rejected": -0.2555481493473053,
|
| 1829 |
+
"step": 1290
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 0.85,
|
| 1833 |
+
"learning_rate": 3.3161219408229026e-07,
|
| 1834 |
+
"logits/chosen": -1.8380473852157593,
|
| 1835 |
+
"logits/rejected": -1.5035631656646729,
|
| 1836 |
+
"logps/chosen": -428.0066833496094,
|
| 1837 |
+
"logps/rejected": -499.0686950683594,
|
| 1838 |
+
"loss": 0.0252,
|
| 1839 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1840 |
+
"rewards/chosen": -0.15843157470226288,
|
| 1841 |
+
"rewards/margins": 0.11070273071527481,
|
| 1842 |
+
"rewards/rejected": -0.2691343128681183,
|
| 1843 |
+
"step": 1300
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 0.86,
|
| 1847 |
+
"learning_rate": 3.0375265076083796e-07,
|
| 1848 |
+
"logits/chosen": -1.7933199405670166,
|
| 1849 |
+
"logits/rejected": -1.6641219854354858,
|
| 1850 |
+
"logps/chosen": -398.84716796875,
|
| 1851 |
+
"logps/rejected": -450.40704345703125,
|
| 1852 |
+
"loss": 0.0219,
|
| 1853 |
+
"rewards/accuracies": 0.625,
|
| 1854 |
+
"rewards/chosen": -0.1423635184764862,
|
| 1855 |
+
"rewards/margins": 0.06088986247777939,
|
| 1856 |
+
"rewards/rejected": -0.2032533884048462,
|
| 1857 |
+
"step": 1310
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 0.86,
|
| 1861 |
+
"learning_rate": 2.7703956163091153e-07,
|
| 1862 |
+
"logits/chosen": -1.6546907424926758,
|
| 1863 |
+
"logits/rejected": -1.491092562675476,
|
| 1864 |
+
"logps/chosen": -442.4610290527344,
|
| 1865 |
+
"logps/rejected": -543.3446044921875,
|
| 1866 |
+
"loss": 0.0214,
|
| 1867 |
+
"rewards/accuracies": 0.675000011920929,
|
| 1868 |
+
"rewards/chosen": -0.1697242110967636,
|
| 1869 |
+
"rewards/margins": 0.11211826652288437,
|
| 1870 |
+
"rewards/rejected": -0.28184249997138977,
|
| 1871 |
+
"step": 1320
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 0.87,
|
| 1875 |
+
"learning_rate": 2.514868710840723e-07,
|
| 1876 |
+
"logits/chosen": -1.7892532348632812,
|
| 1877 |
+
"logits/rejected": -1.673905611038208,
|
| 1878 |
+
"logps/chosen": -385.91119384765625,
|
| 1879 |
+
"logps/rejected": -513.5677490234375,
|
| 1880 |
+
"loss": 0.0256,
|
| 1881 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 1882 |
+
"rewards/chosen": -0.13928866386413574,
|
| 1883 |
+
"rewards/margins": 0.10927131026983261,
|
| 1884 |
+
"rewards/rejected": -0.24855998158454895,
|
| 1885 |
+
"step": 1330
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 0.88,
|
| 1889 |
+
"learning_rate": 2.271079177769117e-07,
|
| 1890 |
+
"logits/chosen": -1.801029920578003,
|
| 1891 |
+
"logits/rejected": -1.7066704034805298,
|
| 1892 |
+
"logps/chosen": -447.7923889160156,
|
| 1893 |
+
"logps/rejected": -510.22088623046875,
|
| 1894 |
+
"loss": 0.016,
|
| 1895 |
+
"rewards/accuracies": 0.699999988079071,
|
| 1896 |
+
"rewards/chosen": -0.14237356185913086,
|
| 1897 |
+
"rewards/margins": 0.09682431071996689,
|
| 1898 |
+
"rewards/rejected": -0.23919788002967834,
|
| 1899 |
+
"step": 1340
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 0.88,
|
| 1903 |
+
"learning_rate": 2.0391542766819456e-07,
|
| 1904 |
+
"logits/chosen": -1.6222654581069946,
|
| 1905 |
+
"logits/rejected": -1.5031228065490723,
|
| 1906 |
+
"logps/chosen": -444.59112548828125,
|
| 1907 |
+
"logps/rejected": -575.5186157226562,
|
| 1908 |
+
"loss": 0.0206,
|
| 1909 |
+
"rewards/accuracies": 0.824999988079071,
|
| 1910 |
+
"rewards/chosen": -0.17282214760780334,
|
| 1911 |
+
"rewards/margins": 0.13414433598518372,
|
| 1912 |
+
"rewards/rejected": -0.30696648359298706,
|
| 1913 |
+
"step": 1350
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 0.89,
|
| 1917 |
+
"learning_rate": 1.8192150737583264e-07,
|
| 1918 |
+
"logits/chosen": -1.8754476308822632,
|
| 1919 |
+
"logits/rejected": -1.4290292263031006,
|
| 1920 |
+
"logps/chosen": -465.01922607421875,
|
| 1921 |
+
"logps/rejected": -613.7750244140625,
|
| 1922 |
+
"loss": 0.0151,
|
| 1923 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1924 |
+
"rewards/chosen": -0.17955392599105835,
|
| 1925 |
+
"rewards/margins": 0.1857127696275711,
|
| 1926 |
+
"rewards/rejected": -0.3652666509151459,
|
| 1927 |
+
"step": 1360
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"epoch": 0.9,
|
| 1931 |
+
"learning_rate": 1.61137637857158e-07,
|
| 1932 |
+
"logits/chosen": -1.8329002857208252,
|
| 1933 |
+
"logits/rejected": -1.6551681756973267,
|
| 1934 |
+
"logps/chosen": -404.65765380859375,
|
| 1935 |
+
"logps/rejected": -481.35931396484375,
|
| 1936 |
+
"loss": 0.0176,
|
| 1937 |
+
"rewards/accuracies": 0.675000011920929,
|
| 1938 |
+
"rewards/chosen": -0.15663662552833557,
|
| 1939 |
+
"rewards/margins": 0.08110956102609634,
|
| 1940 |
+
"rewards/rejected": -0.2377461940050125,
|
| 1941 |
+
"step": 1370
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 0.9,
|
| 1945 |
+
"learning_rate": 1.415746684157951e-07,
|
| 1946 |
+
"logits/chosen": -1.815999984741211,
|
| 1947 |
+
"logits/rejected": -1.5772883892059326,
|
| 1948 |
+
"logps/chosen": -508.98248291015625,
|
| 1949 |
+
"logps/rejected": -616.5827026367188,
|
| 1950 |
+
"loss": 0.0129,
|
| 1951 |
+
"rewards/accuracies": 0.824999988079071,
|
| 1952 |
+
"rewards/chosen": -0.16146263480186462,
|
| 1953 |
+
"rewards/margins": 0.15336495637893677,
|
| 1954 |
+
"rewards/rejected": -0.3148276209831238,
|
| 1955 |
+
"step": 1380
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 0.91,
|
| 1959 |
+
"learning_rate": 1.232428110382586e-07,
|
| 1960 |
+
"logits/chosen": -1.705453634262085,
|
| 1961 |
+
"logits/rejected": -1.7126257419586182,
|
| 1962 |
+
"logps/chosen": -443.9436950683594,
|
| 1963 |
+
"logps/rejected": -507.04681396484375,
|
| 1964 |
+
"loss": 0.0185,
|
| 1965 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 1966 |
+
"rewards/chosen": -0.14823515713214874,
|
| 1967 |
+
"rewards/margins": 0.06383778899908066,
|
| 1968 |
+
"rewards/rejected": -0.2120729386806488,
|
| 1969 |
+
"step": 1390
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 0.92,
|
| 1973 |
+
"learning_rate": 1.0615163506323856e-07,
|
| 1974 |
+
"logits/chosen": -1.734545111656189,
|
| 1975 |
+
"logits/rejected": -1.7701743841171265,
|
| 1976 |
+
"logps/chosen": -467.75225830078125,
|
| 1977 |
+
"logps/rejected": -597.7060546875,
|
| 1978 |
+
"loss": 0.0213,
|
| 1979 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 1980 |
+
"rewards/chosen": -0.17752481997013092,
|
| 1981 |
+
"rewards/margins": 0.08124741166830063,
|
| 1982 |
+
"rewards/rejected": -0.25877219438552856,
|
| 1983 |
+
"step": 1400
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 0.92,
|
| 1987 |
+
"learning_rate": 9.031006218634342e-08,
|
| 1988 |
+
"logits/chosen": -1.8693500757217407,
|
| 1989 |
+
"logits/rejected": -1.7246434688568115,
|
| 1990 |
+
"logps/chosen": -405.7983703613281,
|
| 1991 |
+
"logps/rejected": -492.0767517089844,
|
| 1992 |
+
"loss": 0.0167,
|
| 1993 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 1994 |
+
"rewards/chosen": -0.14290982484817505,
|
| 1995 |
+
"rewards/margins": 0.08994929492473602,
|
| 1996 |
+
"rewards/rejected": -0.23285913467407227,
|
| 1997 |
+
"step": 1410
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 0.93,
|
| 2001 |
+
"learning_rate": 7.572636180292831e-08,
|
| 2002 |
+
"logits/chosen": -1.9143638610839844,
|
| 2003 |
+
"logits/rejected": -1.8357505798339844,
|
| 2004 |
+
"logps/chosen": -359.0616149902344,
|
| 2005 |
+
"logps/rejected": -476.2394104003906,
|
| 2006 |
+
"loss": 0.0251,
|
| 2007 |
+
"rewards/accuracies": 0.574999988079071,
|
| 2008 |
+
"rewards/chosen": -0.14140944182872772,
|
| 2009 |
+
"rewards/margins": 0.09983612596988678,
|
| 2010 |
+
"rewards/rejected": -0.2412455528974533,
|
| 2011 |
+
"step": 1420
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 0.94,
|
| 2015 |
+
"learning_rate": 6.240814669141559e-08,
|
| 2016 |
+
"logits/chosen": -1.6147396564483643,
|
| 2017 |
+
"logits/rejected": -1.5888813734054565,
|
| 2018 |
+
"logps/chosen": -426.96929931640625,
|
| 2019 |
+
"logps/rejected": -460.4154357910156,
|
| 2020 |
+
"loss": 0.023,
|
| 2021 |
+
"rewards/accuracies": 0.625,
|
| 2022 |
+
"rewards/chosen": -0.1609814316034317,
|
| 2023 |
+
"rewards/margins": 0.04549731686711311,
|
| 2024 |
+
"rewards/rejected": -0.20647874474525452,
|
| 2025 |
+
"step": 1430
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 0.94,
|
| 2029 |
+
"learning_rate": 5.036236903938285e-08,
|
| 2030 |
+
"logits/chosen": -1.7797266244888306,
|
| 2031 |
+
"logits/rejected": -1.4338808059692383,
|
| 2032 |
+
"logps/chosen": -383.97021484375,
|
| 2033 |
+
"logps/rejected": -458.4593811035156,
|
| 2034 |
+
"loss": 0.0131,
|
| 2035 |
+
"rewards/accuracies": 0.75,
|
| 2036 |
+
"rewards/chosen": -0.14667093753814697,
|
| 2037 |
+
"rewards/margins": 0.11462962627410889,
|
| 2038 |
+
"rewards/rejected": -0.26130059361457825,
|
| 2039 |
+
"step": 1440
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 0.95,
|
| 2043 |
+
"learning_rate": 3.959531681447859e-08,
|
| 2044 |
+
"logits/chosen": -1.8102188110351562,
|
| 2045 |
+
"logits/rejected": -1.7282259464263916,
|
| 2046 |
+
"logps/chosen": -432.2884826660156,
|
| 2047 |
+
"logps/rejected": -555.3939208984375,
|
| 2048 |
+
"loss": 0.0139,
|
| 2049 |
+
"rewards/accuracies": 0.800000011920929,
|
| 2050 |
+
"rewards/chosen": -0.18346509337425232,
|
| 2051 |
+
"rewards/margins": 0.11375024169683456,
|
| 2052 |
+
"rewards/rejected": -0.2972153425216675,
|
| 2053 |
+
"step": 1450
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 0.96,
|
| 2057 |
+
"learning_rate": 3.0112610482064544e-08,
|
| 2058 |
+
"logits/chosen": -1.774643898010254,
|
| 2059 |
+
"logits/rejected": -1.499586582183838,
|
| 2060 |
+
"logps/chosen": -427.5147399902344,
|
| 2061 |
+
"logps/rejected": -525.1546630859375,
|
| 2062 |
+
"loss": 0.0193,
|
| 2063 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 2064 |
+
"rewards/chosen": -0.1594984531402588,
|
| 2065 |
+
"rewards/margins": 0.09911755472421646,
|
| 2066 |
+
"rewards/rejected": -0.25861600041389465,
|
| 2067 |
+
"step": 1460
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 0.96,
|
| 2071 |
+
"learning_rate": 2.1919200071301715e-08,
|
| 2072 |
+
"logits/chosen": -1.7066888809204102,
|
| 2073 |
+
"logits/rejected": -1.6820920705795288,
|
| 2074 |
+
"logps/chosen": -494.503662109375,
|
| 2075 |
+
"logps/rejected": -650.3525390625,
|
| 2076 |
+
"loss": 0.0154,
|
| 2077 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 2078 |
+
"rewards/chosen": -0.19812996685504913,
|
| 2079 |
+
"rewards/margins": 0.12928268313407898,
|
| 2080 |
+
"rewards/rejected": -0.3274126648902893,
|
| 2081 |
+
"step": 1470
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 0.97,
|
| 2085 |
+
"learning_rate": 1.50193625912029e-08,
|
| 2086 |
+
"logits/chosen": -1.6772607564926147,
|
| 2087 |
+
"logits/rejected": -1.3972876071929932,
|
| 2088 |
+
"logps/chosen": -391.4365539550781,
|
| 2089 |
+
"logps/rejected": -551.2503662109375,
|
| 2090 |
+
"loss": 0.0233,
|
| 2091 |
+
"rewards/accuracies": 0.800000011920929,
|
| 2092 |
+
"rewards/chosen": -0.17442724108695984,
|
| 2093 |
+
"rewards/margins": 0.14638181030750275,
|
| 2094 |
+
"rewards/rejected": -0.3208090364933014,
|
| 2095 |
+
"step": 1480
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 0.97,
|
| 2099 |
+
"learning_rate": 9.416699798010521e-09,
|
| 2100 |
+
"logits/chosen": -1.8336042165756226,
|
| 2101 |
+
"logits/rejected": -1.3778246641159058,
|
| 2102 |
+
"logps/chosen": -412.4085388183594,
|
| 2103 |
+
"logps/rejected": -431.26690673828125,
|
| 2104 |
+
"loss": 0.0284,
|
| 2105 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 2106 |
+
"rewards/chosen": -0.14829722046852112,
|
| 2107 |
+
"rewards/margins": 0.08893366903066635,
|
| 2108 |
+
"rewards/rejected": -0.23723092675209045,
|
| 2109 |
+
"step": 1490
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 0.98,
|
| 2113 |
+
"learning_rate": 5.114136315058083e-09,
|
| 2114 |
+
"logits/chosen": -1.5414822101593018,
|
| 2115 |
+
"logits/rejected": -1.5957967042922974,
|
| 2116 |
+
"logps/chosen": -371.25933837890625,
|
| 2117 |
+
"logps/rejected": -498.68695068359375,
|
| 2118 |
+
"loss": 0.0276,
|
| 2119 |
+
"rewards/accuracies": 0.625,
|
| 2120 |
+
"rewards/chosen": -0.15921153128147125,
|
| 2121 |
+
"rewards/margins": 0.08158004283905029,
|
| 2122 |
+
"rewards/rejected": -0.24079158902168274,
|
| 2123 |
+
"step": 1500
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 0.99,
|
| 2127 |
+
"learning_rate": 2.113918106098345e-09,
|
| 2128 |
+
"logits/chosen": -1.774679183959961,
|
| 2129 |
+
"logits/rejected": -1.5650732517242432,
|
| 2130 |
+
"logps/chosen": -484.24627685546875,
|
| 2131 |
+
"logps/rejected": -497.2215270996094,
|
| 2132 |
+
"loss": 0.0197,
|
| 2133 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 2134 |
+
"rewards/chosen": -0.16433146595954895,
|
| 2135 |
+
"rewards/margins": 0.07303442806005478,
|
| 2136 |
+
"rewards/rejected": -0.23736588656902313,
|
| 2137 |
+
"step": 1510
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 0.99,
|
| 2141 |
+
"learning_rate": 4.176113028983575e-10,
|
| 2142 |
+
"logits/chosen": -2.0026144981384277,
|
| 2143 |
+
"logits/rejected": -1.8274720907211304,
|
| 2144 |
+
"logps/chosen": -379.2737731933594,
|
| 2145 |
+
"logps/rejected": -428.299072265625,
|
| 2146 |
+
"loss": 0.0196,
|
| 2147 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 2148 |
+
"rewards/chosen": -0.13032202422618866,
|
| 2149 |
+
"rewards/margins": 0.08789823949337006,
|
| 2150 |
+
"rewards/rejected": -0.2182202786207199,
|
| 2151 |
+
"step": 1520
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 1.0,
|
| 2155 |
+
"step": 1528,
|
| 2156 |
+
"total_flos": 0.0,
|
| 2157 |
+
"train_loss": 0.021083176277421845,
|
| 2158 |
+
"train_runtime": 50303.8354,
|
| 2159 |
+
"train_samples_per_second": 0.365,
|
| 2160 |
+
"train_steps_per_second": 0.03
|
| 2161 |
+
}
|
| 2162 |
+
],
|
| 2163 |
+
"logging_steps": 10,
|
| 2164 |
+
"max_steps": 1528,
|
| 2165 |
+
"num_input_tokens_seen": 0,
|
| 2166 |
+
"num_train_epochs": 1,
|
| 2167 |
+
"save_steps": 100,
|
| 2168 |
+
"total_flos": 0.0,
|
| 2169 |
+
"train_batch_size": 2,
|
| 2170 |
+
"trial_name": null,
|
| 2171 |
+
"trial_params": null
|
| 2172 |
+
}
|