Model save
Browse files- README.md +62 -0
- adapter_model.safetensors +1 -1
- all_results.json +8 -0
- runs/May12_23-12-56_gpu4-119-5/events.out.tfevents.1715519696.gpu4-119-5.3177525.0 +2 -2
- train_results.json +8 -0
- trainer_state.json +954 -0
README.md
ADDED
|
@@ -0,0 +1,62 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
library_name: peft
|
| 4 |
+
tags:
|
| 5 |
+
- trl
|
| 6 |
+
- dpo
|
| 7 |
+
- generated_from_trainer
|
| 8 |
+
base_model: DUAL-GPO/phi-2-gpo-new-i0
|
| 9 |
+
model-index:
|
| 10 |
+
- name: phi-2-gpo-v7-i1
|
| 11 |
+
results: []
|
| 12 |
+
---
|
| 13 |
+
|
| 14 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 15 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 16 |
+
|
| 17 |
+
# phi-2-gpo-v7-i1
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [DUAL-GPO/phi-2-gpo-new-i0](https://huggingface.co/DUAL-GPO/phi-2-gpo-new-i0) on the None dataset.
|
| 20 |
+
|
| 21 |
+
## Model description
|
| 22 |
+
|
| 23 |
+
More information needed
|
| 24 |
+
|
| 25 |
+
## Intended uses & limitations
|
| 26 |
+
|
| 27 |
+
More information needed
|
| 28 |
+
|
| 29 |
+
## Training and evaluation data
|
| 30 |
+
|
| 31 |
+
More information needed
|
| 32 |
+
|
| 33 |
+
## Training procedure
|
| 34 |
+
|
| 35 |
+
### Training hyperparameters
|
| 36 |
+
|
| 37 |
+
The following hyperparameters were used during training:
|
| 38 |
+
- learning_rate: 5e-06
|
| 39 |
+
- train_batch_size: 4
|
| 40 |
+
- eval_batch_size: 4
|
| 41 |
+
- seed: 42
|
| 42 |
+
- distributed_type: multi-GPU
|
| 43 |
+
- num_devices: 2
|
| 44 |
+
- gradient_accumulation_steps: 4
|
| 45 |
+
- total_train_batch_size: 32
|
| 46 |
+
- total_eval_batch_size: 8
|
| 47 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 48 |
+
- lr_scheduler_type: cosine
|
| 49 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 50 |
+
- num_epochs: 1
|
| 51 |
+
|
| 52 |
+
### Training results
|
| 53 |
+
|
| 54 |
+
|
| 55 |
+
|
| 56 |
+
### Framework versions
|
| 57 |
+
|
| 58 |
+
- PEFT 0.7.1
|
| 59 |
+
- Transformers 4.36.2
|
| 60 |
+
- Pytorch 2.1.2+cu121
|
| 61 |
+
- Datasets 2.14.6
|
| 62 |
+
- Tokenizers 0.15.2
|
adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 167807296
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b7d58d2df658fc0b321035fb16102cc10abbcc0052211a0a9b1b0596d89689fc
|
| 3 |
size 167807296
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.19985946376876132,
|
| 4 |
+
"train_runtime": 6634.0696,
|
| 5 |
+
"train_samples": 21000,
|
| 6 |
+
"train_samples_per_second": 3.165,
|
| 7 |
+
"train_steps_per_second": 0.099
|
| 8 |
+
}
|
runs/May12_23-12-56_gpu4-119-5/events.out.tfevents.1715519696.gpu4-119-5.3177525.0
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 sha256:5c7241998259e7e97f1885c1d346624b894f451d24cce8498a902db961dddaa4
|
| 3 |
+
size 40025
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.19985946376876132,
|
| 4 |
+
"train_runtime": 6634.0696,
|
| 5 |
+
"train_samples": 21000,
|
| 6 |
+
"train_samples_per_second": 3.165,
|
| 7 |
+
"train_steps_per_second": 0.099
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,954 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9996190476190476,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 656,
|
| 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": 7.575757575757576e-08,
|
| 14 |
+
"logits/chosen": 0.04974596947431564,
|
| 15 |
+
"logits/rejected": 0.2963363230228424,
|
| 16 |
+
"logps/chosen": -446.36370849609375,
|
| 17 |
+
"logps/rejected": -275.23162841796875,
|
| 18 |
+
"loss": 0.3488,
|
| 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": 0.02,
|
| 27 |
+
"learning_rate": 7.575757575757576e-07,
|
| 28 |
+
"logits/chosen": 0.07917162775993347,
|
| 29 |
+
"logits/rejected": 0.2545989155769348,
|
| 30 |
+
"logps/chosen": -351.4217529296875,
|
| 31 |
+
"logps/rejected": -305.8712463378906,
|
| 32 |
+
"loss": 0.3403,
|
| 33 |
+
"rewards/accuracies": 0.3472222089767456,
|
| 34 |
+
"rewards/chosen": -1.3770493296760833e-06,
|
| 35 |
+
"rewards/margins": -5.5950724345166236e-05,
|
| 36 |
+
"rewards/rejected": 5.4573672969127074e-05,
|
| 37 |
+
"step": 10
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.03,
|
| 41 |
+
"learning_rate": 1.5151515151515152e-06,
|
| 42 |
+
"logits/chosen": 0.08743862062692642,
|
| 43 |
+
"logits/rejected": 0.2663661539554596,
|
| 44 |
+
"logps/chosen": -356.5379333496094,
|
| 45 |
+
"logps/rejected": -272.34771728515625,
|
| 46 |
+
"loss": 0.3613,
|
| 47 |
+
"rewards/accuracies": 0.574999988079071,
|
| 48 |
+
"rewards/chosen": 8.02798240329139e-05,
|
| 49 |
+
"rewards/margins": 0.0002525355666875839,
|
| 50 |
+
"rewards/rejected": -0.00017225573537871242,
|
| 51 |
+
"step": 20
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.05,
|
| 55 |
+
"learning_rate": 2.2727272727272728e-06,
|
| 56 |
+
"logits/chosen": 0.10032109916210175,
|
| 57 |
+
"logits/rejected": 0.2204272300004959,
|
| 58 |
+
"logps/chosen": -320.36749267578125,
|
| 59 |
+
"logps/rejected": -265.65338134765625,
|
| 60 |
+
"loss": 0.3542,
|
| 61 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 62 |
+
"rewards/chosen": 0.00015481823356822133,
|
| 63 |
+
"rewards/margins": 7.609631575178355e-05,
|
| 64 |
+
"rewards/rejected": 7.872191054048017e-05,
|
| 65 |
+
"step": 30
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.06,
|
| 69 |
+
"learning_rate": 3.0303030303030305e-06,
|
| 70 |
+
"logits/chosen": 0.1070769876241684,
|
| 71 |
+
"logits/rejected": 0.2916509211063385,
|
| 72 |
+
"logps/chosen": -365.18646240234375,
|
| 73 |
+
"logps/rejected": -253.8933868408203,
|
| 74 |
+
"loss": 0.3552,
|
| 75 |
+
"rewards/accuracies": 0.612500011920929,
|
| 76 |
+
"rewards/chosen": 0.0006457852432504296,
|
| 77 |
+
"rewards/margins": 0.0006201790529303253,
|
| 78 |
+
"rewards/rejected": 2.560619941505138e-05,
|
| 79 |
+
"step": 40
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.08,
|
| 83 |
+
"learning_rate": 3.7878787878787882e-06,
|
| 84 |
+
"logits/chosen": 0.08771614730358124,
|
| 85 |
+
"logits/rejected": 0.28780585527420044,
|
| 86 |
+
"logps/chosen": -370.36212158203125,
|
| 87 |
+
"logps/rejected": -294.9586181640625,
|
| 88 |
+
"loss": 0.3338,
|
| 89 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 90 |
+
"rewards/chosen": 0.0016288382466882467,
|
| 91 |
+
"rewards/margins": 0.0014908717712387443,
|
| 92 |
+
"rewards/rejected": 0.000137966446345672,
|
| 93 |
+
"step": 50
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.09,
|
| 97 |
+
"learning_rate": 4.5454545454545455e-06,
|
| 98 |
+
"logits/chosen": 0.06850675493478775,
|
| 99 |
+
"logits/rejected": 0.276347815990448,
|
| 100 |
+
"logps/chosen": -375.48883056640625,
|
| 101 |
+
"logps/rejected": -298.6650695800781,
|
| 102 |
+
"loss": 0.3422,
|
| 103 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 104 |
+
"rewards/chosen": 0.004066063556820154,
|
| 105 |
+
"rewards/margins": 0.004035579971969128,
|
| 106 |
+
"rewards/rejected": 3.0483581213047728e-05,
|
| 107 |
+
"step": 60
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.11,
|
| 111 |
+
"learning_rate": 4.999432965739786e-06,
|
| 112 |
+
"logits/chosen": 0.1320158988237381,
|
| 113 |
+
"logits/rejected": 0.3291288912296295,
|
| 114 |
+
"logps/chosen": -330.6249084472656,
|
| 115 |
+
"logps/rejected": -303.25128173828125,
|
| 116 |
+
"loss": 0.3197,
|
| 117 |
+
"rewards/accuracies": 0.731249988079071,
|
| 118 |
+
"rewards/chosen": 0.007968532852828503,
|
| 119 |
+
"rewards/margins": 0.007708103861659765,
|
| 120 |
+
"rewards/rejected": 0.0002604298642836511,
|
| 121 |
+
"step": 70
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.12,
|
| 125 |
+
"learning_rate": 4.9930567839810125e-06,
|
| 126 |
+
"logits/chosen": 0.09887860715389252,
|
| 127 |
+
"logits/rejected": 0.31491416692733765,
|
| 128 |
+
"logps/chosen": -353.16217041015625,
|
| 129 |
+
"logps/rejected": -297.11651611328125,
|
| 130 |
+
"loss": 0.3248,
|
| 131 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 132 |
+
"rewards/chosen": 0.01291077770292759,
|
| 133 |
+
"rewards/margins": 0.018067900091409683,
|
| 134 |
+
"rewards/rejected": -0.005157124251127243,
|
| 135 |
+
"step": 80
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.14,
|
| 139 |
+
"learning_rate": 4.979613761906212e-06,
|
| 140 |
+
"logits/chosen": 0.17029765248298645,
|
| 141 |
+
"logits/rejected": 0.260185182094574,
|
| 142 |
+
"logps/chosen": -352.32379150390625,
|
| 143 |
+
"logps/rejected": -288.6719665527344,
|
| 144 |
+
"loss": 0.2939,
|
| 145 |
+
"rewards/accuracies": 0.706250011920929,
|
| 146 |
+
"rewards/chosen": 0.007807808928191662,
|
| 147 |
+
"rewards/margins": 0.023743372410535812,
|
| 148 |
+
"rewards/rejected": -0.015935566276311874,
|
| 149 |
+
"step": 90
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.15,
|
| 153 |
+
"learning_rate": 4.959142005221991e-06,
|
| 154 |
+
"logits/chosen": 0.1353963315486908,
|
| 155 |
+
"logits/rejected": 0.2721942365169525,
|
| 156 |
+
"logps/chosen": -363.1653137207031,
|
| 157 |
+
"logps/rejected": -354.89312744140625,
|
| 158 |
+
"loss": 0.2981,
|
| 159 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 160 |
+
"rewards/chosen": -0.01541908085346222,
|
| 161 |
+
"rewards/margins": 0.04708936810493469,
|
| 162 |
+
"rewards/rejected": -0.06250844895839691,
|
| 163 |
+
"step": 100
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.17,
|
| 167 |
+
"learning_rate": 4.931699543346854e-06,
|
| 168 |
+
"logits/chosen": 0.20096346735954285,
|
| 169 |
+
"logits/rejected": 0.3208036422729492,
|
| 170 |
+
"logps/chosen": -412.08306884765625,
|
| 171 |
+
"logps/rejected": -410.16162109375,
|
| 172 |
+
"loss": 0.2759,
|
| 173 |
+
"rewards/accuracies": 0.706250011920929,
|
| 174 |
+
"rewards/chosen": -0.06705033034086227,
|
| 175 |
+
"rewards/margins": 0.07482309639453888,
|
| 176 |
+
"rewards/rejected": -0.14187341928482056,
|
| 177 |
+
"step": 110
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.18,
|
| 181 |
+
"learning_rate": 4.897364164920515e-06,
|
| 182 |
+
"logits/chosen": 0.15477022528648376,
|
| 183 |
+
"logits/rejected": 0.24951288104057312,
|
| 184 |
+
"logps/chosen": -500.6244201660156,
|
| 185 |
+
"logps/rejected": -552.2005004882812,
|
| 186 |
+
"loss": 0.284,
|
| 187 |
+
"rewards/accuracies": 0.643750011920929,
|
| 188 |
+
"rewards/chosen": -0.15765079855918884,
|
| 189 |
+
"rewards/margins": 0.10079488903284073,
|
| 190 |
+
"rewards/rejected": -0.258445680141449,
|
| 191 |
+
"step": 120
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.2,
|
| 195 |
+
"learning_rate": 4.8562331973035396e-06,
|
| 196 |
+
"logits/chosen": 0.15251179039478302,
|
| 197 |
+
"logits/rejected": 0.2451750487089157,
|
| 198 |
+
"logps/chosen": -500.58416748046875,
|
| 199 |
+
"logps/rejected": -580.9321899414062,
|
| 200 |
+
"loss": 0.2784,
|
| 201 |
+
"rewards/accuracies": 0.737500011920929,
|
| 202 |
+
"rewards/chosen": -0.15634217858314514,
|
| 203 |
+
"rewards/margins": 0.12097392231225967,
|
| 204 |
+
"rewards/rejected": -0.27731606364250183,
|
| 205 |
+
"step": 130
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.21,
|
| 209 |
+
"learning_rate": 4.808423230692374e-06,
|
| 210 |
+
"logits/chosen": 0.14251364767551422,
|
| 211 |
+
"logits/rejected": 0.3138013184070587,
|
| 212 |
+
"logps/chosen": -508.4844665527344,
|
| 213 |
+
"logps/rejected": -568.8987426757812,
|
| 214 |
+
"loss": 0.2638,
|
| 215 |
+
"rewards/accuracies": 0.737500011920929,
|
| 216 |
+
"rewards/chosen": -0.15639568865299225,
|
| 217 |
+
"rewards/margins": 0.1208159551024437,
|
| 218 |
+
"rewards/rejected": -0.27721160650253296,
|
| 219 |
+
"step": 140
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.23,
|
| 223 |
+
"learning_rate": 4.754069787631761e-06,
|
| 224 |
+
"logits/chosen": 0.2015986144542694,
|
| 225 |
+
"logits/rejected": 0.28302785754203796,
|
| 226 |
+
"logps/chosen": -529.3695068359375,
|
| 227 |
+
"logps/rejected": -604.30859375,
|
| 228 |
+
"loss": 0.251,
|
| 229 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 230 |
+
"rewards/chosen": -0.18331554532051086,
|
| 231 |
+
"rewards/margins": 0.1465958058834076,
|
| 232 |
+
"rewards/rejected": -0.32991132140159607,
|
| 233 |
+
"step": 150
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.24,
|
| 237 |
+
"learning_rate": 4.693326938861367e-06,
|
| 238 |
+
"logits/chosen": 0.19963108003139496,
|
| 239 |
+
"logits/rejected": 0.2948302626609802,
|
| 240 |
+
"logps/chosen": -566.9232788085938,
|
| 241 |
+
"logps/rejected": -621.8565673828125,
|
| 242 |
+
"loss": 0.2514,
|
| 243 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 244 |
+
"rewards/chosen": -0.2004656344652176,
|
| 245 |
+
"rewards/margins": 0.12107620388269424,
|
| 246 |
+
"rewards/rejected": -0.32154184579849243,
|
| 247 |
+
"step": 160
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.26,
|
| 251 |
+
"learning_rate": 4.626366866585528e-06,
|
| 252 |
+
"logits/chosen": 0.20774176716804504,
|
| 253 |
+
"logits/rejected": 0.31210097670555115,
|
| 254 |
+
"logps/chosen": -525.9526977539062,
|
| 255 |
+
"logps/rejected": -598.7525634765625,
|
| 256 |
+
"loss": 0.2523,
|
| 257 |
+
"rewards/accuracies": 0.699999988079071,
|
| 258 |
+
"rewards/chosen": -0.2060777246952057,
|
| 259 |
+
"rewards/margins": 0.1383783519268036,
|
| 260 |
+
"rewards/rejected": -0.3444560766220093,
|
| 261 |
+
"step": 170
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.27,
|
| 265 |
+
"learning_rate": 4.553379376404085e-06,
|
| 266 |
+
"logits/chosen": 0.1882271021604538,
|
| 267 |
+
"logits/rejected": 0.2884698808193207,
|
| 268 |
+
"logps/chosen": -622.5642700195312,
|
| 269 |
+
"logps/rejected": -703.97998046875,
|
| 270 |
+
"loss": 0.2686,
|
| 271 |
+
"rewards/accuracies": 0.75,
|
| 272 |
+
"rewards/chosen": -0.2503646910190582,
|
| 273 |
+
"rewards/margins": 0.14028559625148773,
|
| 274 |
+
"rewards/rejected": -0.39065021276474,
|
| 275 |
+
"step": 180
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.29,
|
| 279 |
+
"learning_rate": 4.474571359287791e-06,
|
| 280 |
+
"logits/chosen": 0.1948501169681549,
|
| 281 |
+
"logits/rejected": 0.3517269492149353,
|
| 282 |
+
"logps/chosen": -605.8263549804688,
|
| 283 |
+
"logps/rejected": -685.9619750976562,
|
| 284 |
+
"loss": 0.2324,
|
| 285 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 286 |
+
"rewards/chosen": -0.21189594268798828,
|
| 287 |
+
"rewards/margins": 0.19469039142131805,
|
| 288 |
+
"rewards/rejected": -0.40658634901046753,
|
| 289 |
+
"step": 190
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.3,
|
| 293 |
+
"learning_rate": 4.3901662051233755e-06,
|
| 294 |
+
"logits/chosen": 0.1501813530921936,
|
| 295 |
+
"logits/rejected": 0.29967957735061646,
|
| 296 |
+
"logps/chosen": -588.287841796875,
|
| 297 |
+
"logps/rejected": -692.4317626953125,
|
| 298 |
+
"loss": 0.232,
|
| 299 |
+
"rewards/accuracies": 0.762499988079071,
|
| 300 |
+
"rewards/chosen": -0.21489660441875458,
|
| 301 |
+
"rewards/margins": 0.20667514204978943,
|
| 302 |
+
"rewards/rejected": -0.4215717315673828,
|
| 303 |
+
"step": 200
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.32,
|
| 307 |
+
"learning_rate": 4.30040316949064e-06,
|
| 308 |
+
"logits/chosen": 0.1845458298921585,
|
| 309 |
+
"logits/rejected": 0.25728127360343933,
|
| 310 |
+
"logps/chosen": -593.4507446289062,
|
| 311 |
+
"logps/rejected": -647.9249267578125,
|
| 312 |
+
"loss": 0.268,
|
| 313 |
+
"rewards/accuracies": 0.65625,
|
| 314 |
+
"rewards/chosen": -0.2259581983089447,
|
| 315 |
+
"rewards/margins": 0.1573750227689743,
|
| 316 |
+
"rewards/rejected": -0.3833332657814026,
|
| 317 |
+
"step": 210
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.34,
|
| 321 |
+
"learning_rate": 4.205536695466524e-06,
|
| 322 |
+
"logits/chosen": 0.19619445502758026,
|
| 323 |
+
"logits/rejected": 0.289485901594162,
|
| 324 |
+
"logps/chosen": -596.2908325195312,
|
| 325 |
+
"logps/rejected": -673.9340209960938,
|
| 326 |
+
"loss": 0.2279,
|
| 327 |
+
"rewards/accuracies": 0.737500011920929,
|
| 328 |
+
"rewards/chosen": -0.20733025670051575,
|
| 329 |
+
"rewards/margins": 0.1769087016582489,
|
| 330 |
+
"rewards/rejected": -0.38423892855644226,
|
| 331 |
+
"step": 220
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.35,
|
| 335 |
+
"learning_rate": 4.105835692378557e-06,
|
| 336 |
+
"logits/chosen": 0.18007412552833557,
|
| 337 |
+
"logits/rejected": 0.23454514145851135,
|
| 338 |
+
"logps/chosen": -605.221923828125,
|
| 339 |
+
"logps/rejected": -682.0065307617188,
|
| 340 |
+
"loss": 0.2477,
|
| 341 |
+
"rewards/accuracies": 0.71875,
|
| 342 |
+
"rewards/chosen": -0.2464592009782791,
|
| 343 |
+
"rewards/margins": 0.1562257707118988,
|
| 344 |
+
"rewards/rejected": -0.4026849865913391,
|
| 345 |
+
"step": 230
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.37,
|
| 349 |
+
"learning_rate": 4.001582773552153e-06,
|
| 350 |
+
"logits/chosen": 0.17503593862056732,
|
| 351 |
+
"logits/rejected": 0.26662883162498474,
|
| 352 |
+
"logps/chosen": -545.9107666015625,
|
| 353 |
+
"logps/rejected": -640.1486206054688,
|
| 354 |
+
"loss": 0.2225,
|
| 355 |
+
"rewards/accuracies": 0.675000011920929,
|
| 356 |
+
"rewards/chosen": -0.23691585659980774,
|
| 357 |
+
"rewards/margins": 0.16106419265270233,
|
| 358 |
+
"rewards/rejected": -0.39798006415367126,
|
| 359 |
+
"step": 240
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.38,
|
| 363 |
+
"learning_rate": 3.893073455212438e-06,
|
| 364 |
+
"logits/chosen": 0.200395867228508,
|
| 365 |
+
"logits/rejected": 0.3584834933280945,
|
| 366 |
+
"logps/chosen": -608.1827392578125,
|
| 367 |
+
"logps/rejected": -740.0325317382812,
|
| 368 |
+
"loss": 0.2478,
|
| 369 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 370 |
+
"rewards/chosen": -0.26550978422164917,
|
| 371 |
+
"rewards/margins": 0.1945229470729828,
|
| 372 |
+
"rewards/rejected": -0.46003276109695435,
|
| 373 |
+
"step": 250
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.4,
|
| 377 |
+
"learning_rate": 3.7806153188114027e-06,
|
| 378 |
+
"logits/chosen": 0.1786949336528778,
|
| 379 |
+
"logits/rejected": 0.2547721266746521,
|
| 380 |
+
"logps/chosen": -599.2467041015625,
|
| 381 |
+
"logps/rejected": -727.934326171875,
|
| 382 |
+
"loss": 0.2601,
|
| 383 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 384 |
+
"rewards/chosen": -0.22916042804718018,
|
| 385 |
+
"rewards/margins": 0.17522968351840973,
|
| 386 |
+
"rewards/rejected": -0.40439003705978394,
|
| 387 |
+
"step": 260
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.41,
|
| 391 |
+
"learning_rate": 3.6645271391548542e-06,
|
| 392 |
+
"logits/chosen": 0.18216630816459656,
|
| 393 |
+
"logits/rejected": 0.24754111468791962,
|
| 394 |
+
"logps/chosen": -618.3128051757812,
|
| 395 |
+
"logps/rejected": -640.1773681640625,
|
| 396 |
+
"loss": 0.2761,
|
| 397 |
+
"rewards/accuracies": 0.675000011920929,
|
| 398 |
+
"rewards/chosen": -0.22649279236793518,
|
| 399 |
+
"rewards/margins": 0.14542898535728455,
|
| 400 |
+
"rewards/rejected": -0.37192174792289734,
|
| 401 |
+
"step": 270
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.43,
|
| 405 |
+
"learning_rate": 3.5451379808006014e-06,
|
| 406 |
+
"logits/chosen": 0.19754137098789215,
|
| 407 |
+
"logits/rejected": 0.27186593413352966,
|
| 408 |
+
"logps/chosen": -573.2451171875,
|
| 409 |
+
"logps/rejected": -707.2796020507812,
|
| 410 |
+
"loss": 0.2209,
|
| 411 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 412 |
+
"rewards/chosen": -0.2302563190460205,
|
| 413 |
+
"rewards/margins": 0.19838322699069977,
|
| 414 |
+
"rewards/rejected": -0.42863956093788147,
|
| 415 |
+
"step": 280
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.44,
|
| 419 |
+
"learning_rate": 3.4227862652892106e-06,
|
| 420 |
+
"logits/chosen": 0.16008315980434418,
|
| 421 |
+
"logits/rejected": 0.27827057242393494,
|
| 422 |
+
"logps/chosen": -662.4371948242188,
|
| 423 |
+
"logps/rejected": -759.597900390625,
|
| 424 |
+
"loss": 0.2408,
|
| 425 |
+
"rewards/accuracies": 0.75,
|
| 426 |
+
"rewards/chosen": -0.2687516212463379,
|
| 427 |
+
"rewards/margins": 0.17680145800113678,
|
| 428 |
+
"rewards/rejected": -0.4455530643463135,
|
| 429 |
+
"step": 290
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.46,
|
| 433 |
+
"learning_rate": 3.2978188118513814e-06,
|
| 434 |
+
"logits/chosen": 0.17593641579151154,
|
| 435 |
+
"logits/rejected": 0.2789141535758972,
|
| 436 |
+
"logps/chosen": -642.6029052734375,
|
| 437 |
+
"logps/rejected": -753.4479370117188,
|
| 438 |
+
"loss": 0.2304,
|
| 439 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 440 |
+
"rewards/chosen": -0.27796030044555664,
|
| 441 |
+
"rewards/margins": 0.1859740912914276,
|
| 442 |
+
"rewards/rejected": -0.46393436193466187,
|
| 443 |
+
"step": 300
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.47,
|
| 447 |
+
"learning_rate": 3.1705898543111576e-06,
|
| 448 |
+
"logits/chosen": 0.13969933986663818,
|
| 449 |
+
"logits/rejected": 0.3079048693180084,
|
| 450 |
+
"logps/chosen": -630.0859375,
|
| 451 |
+
"logps/rejected": -765.4630737304688,
|
| 452 |
+
"loss": 0.2384,
|
| 453 |
+
"rewards/accuracies": 0.762499988079071,
|
| 454 |
+
"rewards/chosen": -0.2862771153450012,
|
| 455 |
+
"rewards/margins": 0.19564750790596008,
|
| 456 |
+
"rewards/rejected": -0.4819245934486389,
|
| 457 |
+
"step": 310
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.49,
|
| 461 |
+
"learning_rate": 3.041460036971664e-06,
|
| 462 |
+
"logits/chosen": 0.14904941618442535,
|
| 463 |
+
"logits/rejected": 0.27918586134910583,
|
| 464 |
+
"logps/chosen": -619.8759765625,
|
| 465 |
+
"logps/rejected": -759.2853393554688,
|
| 466 |
+
"loss": 0.2347,
|
| 467 |
+
"rewards/accuracies": 0.762499988079071,
|
| 468 |
+
"rewards/chosen": -0.27361705899238586,
|
| 469 |
+
"rewards/margins": 0.19538012146949768,
|
| 470 |
+
"rewards/rejected": -0.4689972400665283,
|
| 471 |
+
"step": 320
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.5,
|
| 475 |
+
"learning_rate": 2.910795392329649e-06,
|
| 476 |
+
"logits/chosen": 0.1377524435520172,
|
| 477 |
+
"logits/rejected": 0.2504701614379883,
|
| 478 |
+
"logps/chosen": -647.4981689453125,
|
| 479 |
+
"logps/rejected": -789.5301513671875,
|
| 480 |
+
"loss": 0.2352,
|
| 481 |
+
"rewards/accuracies": 0.768750011920929,
|
| 482 |
+
"rewards/chosen": -0.2706158757209778,
|
| 483 |
+
"rewards/margins": 0.18975581228733063,
|
| 484 |
+
"rewards/rejected": -0.460371732711792,
|
| 485 |
+
"step": 330
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.52,
|
| 489 |
+
"learning_rate": 2.7789663035166035e-06,
|
| 490 |
+
"logits/chosen": 0.18703623116016388,
|
| 491 |
+
"logits/rejected": 0.3294333815574646,
|
| 492 |
+
"logps/chosen": -656.440673828125,
|
| 493 |
+
"logps/rejected": -793.1935424804688,
|
| 494 |
+
"loss": 0.2072,
|
| 495 |
+
"rewards/accuracies": 0.78125,
|
| 496 |
+
"rewards/chosen": -0.2833554744720459,
|
| 497 |
+
"rewards/margins": 0.21432232856750488,
|
| 498 |
+
"rewards/rejected": -0.4976778030395508,
|
| 499 |
+
"step": 340
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.53,
|
| 503 |
+
"learning_rate": 2.6463464544075344e-06,
|
| 504 |
+
"logits/chosen": 0.15668293833732605,
|
| 505 |
+
"logits/rejected": 0.2470695674419403,
|
| 506 |
+
"logps/chosen": -603.0676879882812,
|
| 507 |
+
"logps/rejected": -736.9429931640625,
|
| 508 |
+
"loss": 0.227,
|
| 509 |
+
"rewards/accuracies": 0.699999988079071,
|
| 510 |
+
"rewards/chosen": -0.27748292684555054,
|
| 511 |
+
"rewards/margins": 0.1968688815832138,
|
| 512 |
+
"rewards/rejected": -0.4743518829345703,
|
| 513 |
+
"step": 350
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.55,
|
| 517 |
+
"learning_rate": 2.513311770373421e-06,
|
| 518 |
+
"logits/chosen": 0.09050649404525757,
|
| 519 |
+
"logits/rejected": 0.27548736333847046,
|
| 520 |
+
"logps/chosen": -645.054443359375,
|
| 521 |
+
"logps/rejected": -780.7376708984375,
|
| 522 |
+
"loss": 0.2277,
|
| 523 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 524 |
+
"rewards/chosen": -0.27279263734817505,
|
| 525 |
+
"rewards/margins": 0.2085564136505127,
|
| 526 |
+
"rewards/rejected": -0.48134905099868774,
|
| 527 |
+
"step": 360
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.56,
|
| 531 |
+
"learning_rate": 2.380239352679908e-06,
|
| 532 |
+
"logits/chosen": 0.11263048648834229,
|
| 533 |
+
"logits/rejected": 0.23795035481452942,
|
| 534 |
+
"logps/chosen": -635.0528564453125,
|
| 535 |
+
"logps/rejected": -752.4202270507812,
|
| 536 |
+
"loss": 0.2295,
|
| 537 |
+
"rewards/accuracies": 0.731249988079071,
|
| 538 |
+
"rewards/chosen": -0.28419169783592224,
|
| 539 |
+
"rewards/margins": 0.18686431646347046,
|
| 540 |
+
"rewards/rejected": -0.4710559844970703,
|
| 541 |
+
"step": 370
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.58,
|
| 545 |
+
"learning_rate": 2.247506409552795e-06,
|
| 546 |
+
"logits/chosen": 0.18179914355278015,
|
| 547 |
+
"logits/rejected": 0.2870942950248718,
|
| 548 |
+
"logps/chosen": -632.6707153320312,
|
| 549 |
+
"logps/rejected": -745.6337890625,
|
| 550 |
+
"loss": 0.21,
|
| 551 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 552 |
+
"rewards/chosen": -0.2822064459323883,
|
| 553 |
+
"rewards/margins": 0.18581680953502655,
|
| 554 |
+
"rewards/rejected": -0.4680232107639313,
|
| 555 |
+
"step": 380
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.59,
|
| 559 |
+
"learning_rate": 2.1154891869403436e-06,
|
| 560 |
+
"logits/chosen": 0.1858624517917633,
|
| 561 |
+
"logits/rejected": 0.24371235072612762,
|
| 562 |
+
"logps/chosen": -612.2597045898438,
|
| 563 |
+
"logps/rejected": -773.7772216796875,
|
| 564 |
+
"loss": 0.24,
|
| 565 |
+
"rewards/accuracies": 0.706250011920929,
|
| 566 |
+
"rewards/chosen": -0.2878335416316986,
|
| 567 |
+
"rewards/margins": 0.1961011290550232,
|
| 568 |
+
"rewards/rejected": -0.4839346408843994,
|
| 569 |
+
"step": 390
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.61,
|
| 573 |
+
"learning_rate": 1.9845619020032552e-06,
|
| 574 |
+
"logits/chosen": 0.11868009716272354,
|
| 575 |
+
"logits/rejected": 0.2442179173231125,
|
| 576 |
+
"logps/chosen": -628.7079467773438,
|
| 577 |
+
"logps/rejected": -782.9298095703125,
|
| 578 |
+
"loss": 0.2128,
|
| 579 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 580 |
+
"rewards/chosen": -0.2861320376396179,
|
| 581 |
+
"rewards/margins": 0.2013685703277588,
|
| 582 |
+
"rewards/rejected": -0.4875006079673767,
|
| 583 |
+
"step": 400
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.62,
|
| 587 |
+
"learning_rate": 1.8550956823554708e-06,
|
| 588 |
+
"logits/chosen": 0.15941022336483002,
|
| 589 |
+
"logits/rejected": 0.26665717363357544,
|
| 590 |
+
"logps/chosen": -611.2957763671875,
|
| 591 |
+
"logps/rejected": -745.6875610351562,
|
| 592 |
+
"loss": 0.2419,
|
| 593 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 594 |
+
"rewards/chosen": -0.26627546548843384,
|
| 595 |
+
"rewards/margins": 0.20409245789051056,
|
| 596 |
+
"rewards/rejected": -0.4703678488731384,
|
| 597 |
+
"step": 410
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.64,
|
| 601 |
+
"learning_rate": 1.7274575140626318e-06,
|
| 602 |
+
"logits/chosen": 0.12057618796825409,
|
| 603 |
+
"logits/rejected": 0.2439500391483307,
|
| 604 |
+
"logps/chosen": -563.8010864257812,
|
| 605 |
+
"logps/rejected": -753.3555908203125,
|
| 606 |
+
"loss": 0.2241,
|
| 607 |
+
"rewards/accuracies": 0.75,
|
| 608 |
+
"rewards/chosen": -0.22670452296733856,
|
| 609 |
+
"rewards/margins": 0.22902043163776398,
|
| 610 |
+
"rewards/rejected": -0.45572495460510254,
|
| 611 |
+
"step": 420
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.66,
|
| 615 |
+
"learning_rate": 1.6020092013802002e-06,
|
| 616 |
+
"logits/chosen": 0.1292915642261505,
|
| 617 |
+
"logits/rejected": 0.27418118715286255,
|
| 618 |
+
"logps/chosen": -623.4757080078125,
|
| 619 |
+
"logps/rejected": -712.2957153320312,
|
| 620 |
+
"loss": 0.2196,
|
| 621 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 622 |
+
"rewards/chosen": -0.24320180714130402,
|
| 623 |
+
"rewards/margins": 0.19605949521064758,
|
| 624 |
+
"rewards/rejected": -0.4392613470554352,
|
| 625 |
+
"step": 430
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.67,
|
| 629 |
+
"learning_rate": 1.4791063411799938e-06,
|
| 630 |
+
"logits/chosen": 0.10819476842880249,
|
| 631 |
+
"logits/rejected": 0.2511535882949829,
|
| 632 |
+
"logps/chosen": -570.8446044921875,
|
| 633 |
+
"logps/rejected": -685.5501708984375,
|
| 634 |
+
"loss": 0.2124,
|
| 635 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 636 |
+
"rewards/chosen": -0.22663471102714539,
|
| 637 |
+
"rewards/margins": 0.20201578736305237,
|
| 638 |
+
"rewards/rejected": -0.42865046858787537,
|
| 639 |
+
"step": 440
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.69,
|
| 643 |
+
"learning_rate": 1.3590973149722103e-06,
|
| 644 |
+
"logits/chosen": 0.15888772904872894,
|
| 645 |
+
"logits/rejected": 0.2368732988834381,
|
| 646 |
+
"logps/chosen": -589.4254150390625,
|
| 647 |
+
"logps/rejected": -735.7716674804688,
|
| 648 |
+
"loss": 0.2038,
|
| 649 |
+
"rewards/accuracies": 0.731249988079071,
|
| 650 |
+
"rewards/chosen": -0.26341313123703003,
|
| 651 |
+
"rewards/margins": 0.1875002086162567,
|
| 652 |
+
"rewards/rejected": -0.45091336965560913,
|
| 653 |
+
"step": 450
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.7,
|
| 657 |
+
"learning_rate": 1.2423223013801946e-06,
|
| 658 |
+
"logits/chosen": 0.14813843369483948,
|
| 659 |
+
"logits/rejected": 0.21993982791900635,
|
| 660 |
+
"logps/chosen": -589.8951416015625,
|
| 661 |
+
"logps/rejected": -729.8585205078125,
|
| 662 |
+
"loss": 0.2504,
|
| 663 |
+
"rewards/accuracies": 0.675000011920929,
|
| 664 |
+
"rewards/chosen": -0.2774823307991028,
|
| 665 |
+
"rewards/margins": 0.1887308657169342,
|
| 666 |
+
"rewards/rejected": -0.4662131667137146,
|
| 667 |
+
"step": 460
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.72,
|
| 671 |
+
"learning_rate": 1.1291123118671665e-06,
|
| 672 |
+
"logits/chosen": 0.09890522062778473,
|
| 673 |
+
"logits/rejected": 0.2421140968799591,
|
| 674 |
+
"logps/chosen": -639.2291870117188,
|
| 675 |
+
"logps/rejected": -729.2130126953125,
|
| 676 |
+
"loss": 0.1871,
|
| 677 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 678 |
+
"rewards/chosen": -0.2574080526828766,
|
| 679 |
+
"rewards/margins": 0.20436222851276398,
|
| 680 |
+
"rewards/rejected": -0.46177029609680176,
|
| 681 |
+
"step": 470
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.73,
|
| 685 |
+
"learning_rate": 1.019788252448267e-06,
|
| 686 |
+
"logits/chosen": 0.1626007854938507,
|
| 687 |
+
"logits/rejected": 0.2534940242767334,
|
| 688 |
+
"logps/chosen": -614.7600708007812,
|
| 689 |
+
"logps/rejected": -789.6011352539062,
|
| 690 |
+
"loss": 0.202,
|
| 691 |
+
"rewards/accuracies": 0.737500011920929,
|
| 692 |
+
"rewards/chosen": -0.27494579553604126,
|
| 693 |
+
"rewards/margins": 0.21227261424064636,
|
| 694 |
+
"rewards/rejected": -0.48721843957901,
|
| 695 |
+
"step": 480
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.75,
|
| 699 |
+
"learning_rate": 9.146600140475945e-07,
|
| 700 |
+
"logits/chosen": 0.18194663524627686,
|
| 701 |
+
"logits/rejected": 0.2571627199649811,
|
| 702 |
+
"logps/chosen": -695.4674682617188,
|
| 703 |
+
"logps/rejected": -824.3109130859375,
|
| 704 |
+
"loss": 0.2242,
|
| 705 |
+
"rewards/accuracies": 0.768750011920929,
|
| 706 |
+
"rewards/chosen": -0.31008249521255493,
|
| 707 |
+
"rewards/margins": 0.20844268798828125,
|
| 708 |
+
"rewards/rejected": -0.518525242805481,
|
| 709 |
+
"step": 490
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.76,
|
| 713 |
+
"learning_rate": 8.140255940787059e-07,
|
| 714 |
+
"logits/chosen": 0.1600276529788971,
|
| 715 |
+
"logits/rejected": 0.24592892825603485,
|
| 716 |
+
"logps/chosen": -621.3486328125,
|
| 717 |
+
"logps/rejected": -787.8779296875,
|
| 718 |
+
"loss": 0.2379,
|
| 719 |
+
"rewards/accuracies": 0.737500011920929,
|
| 720 |
+
"rewards/chosen": -0.2887078523635864,
|
| 721 |
+
"rewards/margins": 0.21262690424919128,
|
| 722 |
+
"rewards/rejected": -0.5013347864151001,
|
| 723 |
+
"step": 500
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.78,
|
| 727 |
+
"learning_rate": 7.181702517385789e-07,
|
| 728 |
+
"logits/chosen": 0.11802725493907928,
|
| 729 |
+
"logits/rejected": 0.24163658916950226,
|
| 730 |
+
"logps/chosen": -626.5403442382812,
|
| 731 |
+
"logps/rejected": -774.00830078125,
|
| 732 |
+
"loss": 0.219,
|
| 733 |
+
"rewards/accuracies": 0.800000011920929,
|
| 734 |
+
"rewards/chosen": -0.27267104387283325,
|
| 735 |
+
"rewards/margins": 0.21761374175548553,
|
| 736 |
+
"rewards/rejected": -0.4902847409248352,
|
| 737 |
+
"step": 510
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.79,
|
| 741 |
+
"learning_rate": 6.273656994094232e-07,
|
| 742 |
+
"logits/chosen": 0.1450013965368271,
|
| 743 |
+
"logits/rejected": 0.23478934168815613,
|
| 744 |
+
"logps/chosen": -608.7362060546875,
|
| 745 |
+
"logps/rejected": -739.3218994140625,
|
| 746 |
+
"loss": 0.2294,
|
| 747 |
+
"rewards/accuracies": 0.762499988079071,
|
| 748 |
+
"rewards/chosen": -0.2396283596754074,
|
| 749 |
+
"rewards/margins": 0.21917715668678284,
|
| 750 |
+
"rewards/rejected": -0.45880550146102905,
|
| 751 |
+
"step": 520
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.81,
|
| 755 |
+
"learning_rate": 5.418693324604082e-07,
|
| 756 |
+
"logits/chosen": 0.16014720499515533,
|
| 757 |
+
"logits/rejected": 0.25679388642311096,
|
| 758 |
+
"logps/chosen": -598.508056640625,
|
| 759 |
+
"logps/rejected": -752.1389770507812,
|
| 760 |
+
"loss": 0.2315,
|
| 761 |
+
"rewards/accuracies": 0.762499988079071,
|
| 762 |
+
"rewards/chosen": -0.26550406217575073,
|
| 763 |
+
"rewards/margins": 0.20078308880329132,
|
| 764 |
+
"rewards/rejected": -0.46628713607788086,
|
| 765 |
+
"step": 530
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.82,
|
| 769 |
+
"learning_rate": 4.619234996325314e-07,
|
| 770 |
+
"logits/chosen": 0.10893194377422333,
|
| 771 |
+
"logits/rejected": 0.18960845470428467,
|
| 772 |
+
"logps/chosen": -589.2452392578125,
|
| 773 |
+
"logps/rejected": -712.2008666992188,
|
| 774 |
+
"loss": 0.2139,
|
| 775 |
+
"rewards/accuracies": 0.768750011920929,
|
| 776 |
+
"rewards/chosen": -0.2257637232542038,
|
| 777 |
+
"rewards/margins": 0.20952418446540833,
|
| 778 |
+
"rewards/rejected": -0.4352878928184509,
|
| 779 |
+
"step": 540
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.84,
|
| 783 |
+
"learning_rate": 3.877548160747768e-07,
|
| 784 |
+
"logits/chosen": 0.1536407321691513,
|
| 785 |
+
"logits/rejected": 0.25428491830825806,
|
| 786 |
+
"logps/chosen": -667.2605590820312,
|
| 787 |
+
"logps/rejected": -712.8155517578125,
|
| 788 |
+
"loss": 0.2543,
|
| 789 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 790 |
+
"rewards/chosen": -0.26264914870262146,
|
| 791 |
+
"rewards/margins": 0.1710439920425415,
|
| 792 |
+
"rewards/rejected": -0.4336931109428406,
|
| 793 |
+
"step": 550
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.85,
|
| 797 |
+
"learning_rate": 3.195735209788528e-07,
|
| 798 |
+
"logits/chosen": 0.1331530511379242,
|
| 799 |
+
"logits/rejected": 0.25019171833992004,
|
| 800 |
+
"logps/chosen": -614.9050903320312,
|
| 801 |
+
"logps/rejected": -703.8809814453125,
|
| 802 |
+
"loss": 0.2492,
|
| 803 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 804 |
+
"rewards/chosen": -0.27054914832115173,
|
| 805 |
+
"rewards/margins": 0.14634455740451813,
|
| 806 |
+
"rewards/rejected": -0.4168936610221863,
|
| 807 |
+
"step": 560
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.87,
|
| 811 |
+
"learning_rate": 2.5757288163336806e-07,
|
| 812 |
+
"logits/chosen": 0.15042927861213684,
|
| 813 |
+
"logits/rejected": 0.2752520442008972,
|
| 814 |
+
"logps/chosen": -605.6216430664062,
|
| 815 |
+
"logps/rejected": -749.6766357421875,
|
| 816 |
+
"loss": 0.2349,
|
| 817 |
+
"rewards/accuracies": 0.731249988079071,
|
| 818 |
+
"rewards/chosen": -0.2648725211620331,
|
| 819 |
+
"rewards/margins": 0.18051207065582275,
|
| 820 |
+
"rewards/rejected": -0.44538459181785583,
|
| 821 |
+
"step": 570
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.88,
|
| 825 |
+
"learning_rate": 2.019286455866981e-07,
|
| 826 |
+
"logits/chosen": 0.09744317829608917,
|
| 827 |
+
"logits/rejected": 0.30391809344291687,
|
| 828 |
+
"logps/chosen": -656.5165405273438,
|
| 829 |
+
"logps/rejected": -782.3225708007812,
|
| 830 |
+
"loss": 0.2347,
|
| 831 |
+
"rewards/accuracies": 0.762499988079071,
|
| 832 |
+
"rewards/chosen": -0.26627975702285767,
|
| 833 |
+
"rewards/margins": 0.2162231206893921,
|
| 834 |
+
"rewards/rejected": -0.48250293731689453,
|
| 835 |
+
"step": 580
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.9,
|
| 839 |
+
"learning_rate": 1.5279854247146703e-07,
|
| 840 |
+
"logits/chosen": 0.10827809572219849,
|
| 841 |
+
"logits/rejected": 0.23128509521484375,
|
| 842 |
+
"logps/chosen": -597.4005126953125,
|
| 843 |
+
"logps/rejected": -727.5601196289062,
|
| 844 |
+
"loss": 0.2356,
|
| 845 |
+
"rewards/accuracies": 0.75,
|
| 846 |
+
"rewards/chosen": -0.2508590817451477,
|
| 847 |
+
"rewards/margins": 0.19355928897857666,
|
| 848 |
+
"rewards/rejected": -0.44441837072372437,
|
| 849 |
+
"step": 590
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.91,
|
| 853 |
+
"learning_rate": 1.1032183690276754e-07,
|
| 854 |
+
"logits/chosen": 0.1284581571817398,
|
| 855 |
+
"logits/rejected": 0.22490856051445007,
|
| 856 |
+
"logps/chosen": -608.4317016601562,
|
| 857 |
+
"logps/rejected": -756.5374755859375,
|
| 858 |
+
"loss": 0.2282,
|
| 859 |
+
"rewards/accuracies": 0.737500011920929,
|
| 860 |
+
"rewards/chosen": -0.27711886167526245,
|
| 861 |
+
"rewards/margins": 0.19119976460933685,
|
| 862 |
+
"rewards/rejected": -0.4683186411857605,
|
| 863 |
+
"step": 600
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.93,
|
| 867 |
+
"learning_rate": 7.46189337174788e-08,
|
| 868 |
+
"logits/chosen": 0.1757223904132843,
|
| 869 |
+
"logits/rejected": 0.26904186606407166,
|
| 870 |
+
"logps/chosen": -603.6378173828125,
|
| 871 |
+
"logps/rejected": -660.9766235351562,
|
| 872 |
+
"loss": 0.2443,
|
| 873 |
+
"rewards/accuracies": 0.706250011920929,
|
| 874 |
+
"rewards/chosen": -0.24105675518512726,
|
| 875 |
+
"rewards/margins": 0.1722828447818756,
|
| 876 |
+
"rewards/rejected": -0.4133395552635193,
|
| 877 |
+
"step": 610
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.94,
|
| 881 |
+
"learning_rate": 4.579103667367385e-08,
|
| 882 |
+
"logits/chosen": 0.10849970579147339,
|
| 883 |
+
"logits/rejected": 0.19708193838596344,
|
| 884 |
+
"logps/chosen": -565.9713134765625,
|
| 885 |
+
"logps/rejected": -679.5427856445312,
|
| 886 |
+
"loss": 0.2211,
|
| 887 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 888 |
+
"rewards/chosen": -0.2512280344963074,
|
| 889 |
+
"rewards/margins": 0.17560149729251862,
|
| 890 |
+
"rewards/rejected": -0.4268294870853424,
|
| 891 |
+
"step": 620
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.96,
|
| 895 |
+
"learning_rate": 2.3919861577572924e-08,
|
| 896 |
+
"logits/chosen": 0.13952800631523132,
|
| 897 |
+
"logits/rejected": 0.20136961340904236,
|
| 898 |
+
"logps/chosen": -621.4708862304688,
|
| 899 |
+
"logps/rejected": -748.71240234375,
|
| 900 |
+
"loss": 0.2234,
|
| 901 |
+
"rewards/accuracies": 0.768750011920929,
|
| 902 |
+
"rewards/chosen": -0.2641890048980713,
|
| 903 |
+
"rewards/margins": 0.19575798511505127,
|
| 904 |
+
"rewards/rejected": -0.45994701981544495,
|
| 905 |
+
"step": 630
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.98,
|
| 909 |
+
"learning_rate": 9.067404651211808e-09,
|
| 910 |
+
"logits/chosen": 0.18144121766090393,
|
| 911 |
+
"logits/rejected": 0.23935365676879883,
|
| 912 |
+
"logps/chosen": -601.0186767578125,
|
| 913 |
+
"logps/rejected": -719.6766357421875,
|
| 914 |
+
"loss": 0.2458,
|
| 915 |
+
"rewards/accuracies": 0.731249988079071,
|
| 916 |
+
"rewards/chosen": -0.26298612356185913,
|
| 917 |
+
"rewards/margins": 0.16590112447738647,
|
| 918 |
+
"rewards/rejected": -0.428887277841568,
|
| 919 |
+
"step": 640
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.99,
|
| 923 |
+
"learning_rate": 1.2757667974155896e-09,
|
| 924 |
+
"logits/chosen": 0.13537321984767914,
|
| 925 |
+
"logits/rejected": 0.2640915811061859,
|
| 926 |
+
"logps/chosen": -576.1346435546875,
|
| 927 |
+
"logps/rejected": -705.7242431640625,
|
| 928 |
+
"loss": 0.2243,
|
| 929 |
+
"rewards/accuracies": 0.75,
|
| 930 |
+
"rewards/chosen": -0.22819241881370544,
|
| 931 |
+
"rewards/margins": 0.1957322061061859,
|
| 932 |
+
"rewards/rejected": -0.42392459511756897,
|
| 933 |
+
"step": 650
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 1.0,
|
| 937 |
+
"step": 656,
|
| 938 |
+
"total_flos": 0.0,
|
| 939 |
+
"train_loss": 0.19985946376876132,
|
| 940 |
+
"train_runtime": 6634.0696,
|
| 941 |
+
"train_samples_per_second": 3.165,
|
| 942 |
+
"train_steps_per_second": 0.099
|
| 943 |
+
}
|
| 944 |
+
],
|
| 945 |
+
"logging_steps": 10,
|
| 946 |
+
"max_steps": 656,
|
| 947 |
+
"num_input_tokens_seen": 0,
|
| 948 |
+
"num_train_epochs": 1,
|
| 949 |
+
"save_steps": 100,
|
| 950 |
+
"total_flos": 0.0,
|
| 951 |
+
"train_batch_size": 4,
|
| 952 |
+
"trial_name": null,
|
| 953 |
+
"trial_params": null
|
| 954 |
+
}
|