Model save
Browse files- README.md +62 -0
- adapter_model.safetensors +1 -1
- all_results.json +8 -0
- runs/May12_13-35-06_gpu4-119-5/events.out.tfevents.1715485023.gpu4-119-5.2489421.0 +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-v4-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-v4-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 sha256:507e7dbf961b259a27f2901b41403ad22916f07bd4f457a26502de4654162163
|
| 3 |
size 167807296
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.22703545004492853,
|
| 4 |
+
"train_runtime": 7888.8763,
|
| 5 |
+
"train_samples": 21000,
|
| 6 |
+
"train_samples_per_second": 2.662,
|
| 7 |
+
"train_steps_per_second": 0.083
|
| 8 |
+
}
|
runs/May12_13-35-06_gpu4-119-5/events.out.tfevents.1715485023.gpu4-119-5.2489421.0
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 sha256:15ddfc3e960972af7db3f580cf5cb8bd5a7b7ae961ae1de08af1a442ffe7eb26
|
| 3 |
+
size 46877
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.22703545004492853,
|
| 4 |
+
"train_runtime": 7888.8763,
|
| 5 |
+
"train_samples": 21000,
|
| 6 |
+
"train_samples_per_second": 2.662,
|
| 7 |
+
"train_steps_per_second": 0.083
|
| 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": 0.040165986865758896,
|
| 15 |
+
"logits/rejected": 0.1715753823518753,
|
| 16 |
+
"logps/chosen": -294.844482421875,
|
| 17 |
+
"logps/rejected": -361.2099914550781,
|
| 18 |
+
"loss": 0.3581,
|
| 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": 0.08047256618738174,
|
| 29 |
+
"logits/rejected": 0.3207971751689911,
|
| 30 |
+
"logps/chosen": -393.2153625488281,
|
| 31 |
+
"logps/rejected": -318.74615478515625,
|
| 32 |
+
"loss": 0.3404,
|
| 33 |
+
"rewards/accuracies": 0.4652777910232544,
|
| 34 |
+
"rewards/chosen": -0.00022995664039626718,
|
| 35 |
+
"rewards/margins": 2.2277235984802246e-05,
|
| 36 |
+
"rewards/rejected": -0.000252234167419374,
|
| 37 |
+
"step": 10
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.03,
|
| 41 |
+
"learning_rate": 1.5151515151515152e-06,
|
| 42 |
+
"logits/chosen": 0.07466734945774078,
|
| 43 |
+
"logits/rejected": 0.23236870765686035,
|
| 44 |
+
"logps/chosen": -332.4886169433594,
|
| 45 |
+
"logps/rejected": -281.853515625,
|
| 46 |
+
"loss": 0.3468,
|
| 47 |
+
"rewards/accuracies": 0.41874998807907104,
|
| 48 |
+
"rewards/chosen": 0.00011478399392217398,
|
| 49 |
+
"rewards/margins": -0.0012641319772228599,
|
| 50 |
+
"rewards/rejected": 0.0013789159711450338,
|
| 51 |
+
"step": 20
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.05,
|
| 55 |
+
"learning_rate": 2.2727272727272728e-06,
|
| 56 |
+
"logits/chosen": 0.07185273617506027,
|
| 57 |
+
"logits/rejected": 0.2604687213897705,
|
| 58 |
+
"logps/chosen": -358.33782958984375,
|
| 59 |
+
"logps/rejected": -292.1524963378906,
|
| 60 |
+
"loss": 0.3502,
|
| 61 |
+
"rewards/accuracies": 0.543749988079071,
|
| 62 |
+
"rewards/chosen": 0.0017305829096585512,
|
| 63 |
+
"rewards/margins": 0.001660021604038775,
|
| 64 |
+
"rewards/rejected": 7.056114554870874e-05,
|
| 65 |
+
"step": 30
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.06,
|
| 69 |
+
"learning_rate": 3.0303030303030305e-06,
|
| 70 |
+
"logits/chosen": 0.1199118122458458,
|
| 71 |
+
"logits/rejected": 0.2392597496509552,
|
| 72 |
+
"logps/chosen": -347.05926513671875,
|
| 73 |
+
"logps/rejected": -286.26080322265625,
|
| 74 |
+
"loss": 0.3353,
|
| 75 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 76 |
+
"rewards/chosen": 0.005740107037127018,
|
| 77 |
+
"rewards/margins": 0.006530737970024347,
|
| 78 |
+
"rewards/rejected": -0.0007906301179900765,
|
| 79 |
+
"step": 40
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.08,
|
| 83 |
+
"learning_rate": 3.7878787878787882e-06,
|
| 84 |
+
"logits/chosen": 0.0741933211684227,
|
| 85 |
+
"logits/rejected": 0.3117237091064453,
|
| 86 |
+
"logps/chosen": -343.31170654296875,
|
| 87 |
+
"logps/rejected": -287.1304931640625,
|
| 88 |
+
"loss": 0.3286,
|
| 89 |
+
"rewards/accuracies": 0.6875,
|
| 90 |
+
"rewards/chosen": 0.016505183652043343,
|
| 91 |
+
"rewards/margins": 0.016020886600017548,
|
| 92 |
+
"rewards/rejected": 0.0004842969647143036,
|
| 93 |
+
"step": 50
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.09,
|
| 97 |
+
"learning_rate": 4.5454545454545455e-06,
|
| 98 |
+
"logits/chosen": 0.0762203261256218,
|
| 99 |
+
"logits/rejected": 0.2750852704048157,
|
| 100 |
+
"logps/chosen": -370.8611755371094,
|
| 101 |
+
"logps/rejected": -302.7222595214844,
|
| 102 |
+
"loss": 0.2924,
|
| 103 |
+
"rewards/accuracies": 0.768750011920929,
|
| 104 |
+
"rewards/chosen": 0.032092947512865067,
|
| 105 |
+
"rewards/margins": 0.04518315941095352,
|
| 106 |
+
"rewards/rejected": -0.013090210035443306,
|
| 107 |
+
"step": 60
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.11,
|
| 111 |
+
"learning_rate": 4.999432965739786e-06,
|
| 112 |
+
"logits/chosen": 0.07090188562870026,
|
| 113 |
+
"logits/rejected": 0.25236833095550537,
|
| 114 |
+
"logps/chosen": -321.5715026855469,
|
| 115 |
+
"logps/rejected": -301.0986022949219,
|
| 116 |
+
"loss": 0.2803,
|
| 117 |
+
"rewards/accuracies": 0.731249988079071,
|
| 118 |
+
"rewards/chosen": 0.05498770996928215,
|
| 119 |
+
"rewards/margins": 0.0691528171300888,
|
| 120 |
+
"rewards/rejected": -0.014165110886096954,
|
| 121 |
+
"step": 70
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.12,
|
| 125 |
+
"learning_rate": 4.9930567839810125e-06,
|
| 126 |
+
"logits/chosen": 0.10318852961063385,
|
| 127 |
+
"logits/rejected": 0.2712559401988983,
|
| 128 |
+
"logps/chosen": -358.31500244140625,
|
| 129 |
+
"logps/rejected": -302.6133117675781,
|
| 130 |
+
"loss": 0.2346,
|
| 131 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 132 |
+
"rewards/chosen": 0.08596866577863693,
|
| 133 |
+
"rewards/margins": 0.1306326687335968,
|
| 134 |
+
"rewards/rejected": -0.044663988053798676,
|
| 135 |
+
"step": 80
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.14,
|
| 139 |
+
"learning_rate": 4.979613761906212e-06,
|
| 140 |
+
"logits/chosen": 0.1258704960346222,
|
| 141 |
+
"logits/rejected": 0.2653108239173889,
|
| 142 |
+
"logps/chosen": -316.4566345214844,
|
| 143 |
+
"logps/rejected": -285.59442138671875,
|
| 144 |
+
"loss": 0.2181,
|
| 145 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 146 |
+
"rewards/chosen": 0.10472643375396729,
|
| 147 |
+
"rewards/margins": 0.16386187076568604,
|
| 148 |
+
"rewards/rejected": -0.05913544446229935,
|
| 149 |
+
"step": 90
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.15,
|
| 153 |
+
"learning_rate": 4.959142005221991e-06,
|
| 154 |
+
"logits/chosen": 0.14865969121456146,
|
| 155 |
+
"logits/rejected": 0.2514886260032654,
|
| 156 |
+
"logps/chosen": -316.22650146484375,
|
| 157 |
+
"logps/rejected": -298.5392150878906,
|
| 158 |
+
"loss": 0.2379,
|
| 159 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 160 |
+
"rewards/chosen": 0.09984966367483139,
|
| 161 |
+
"rewards/margins": 0.20773954689502716,
|
| 162 |
+
"rewards/rejected": -0.10788986831903458,
|
| 163 |
+
"step": 100
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.17,
|
| 167 |
+
"learning_rate": 4.931699543346854e-06,
|
| 168 |
+
"logits/chosen": 0.1114228144288063,
|
| 169 |
+
"logits/rejected": 0.2948494553565979,
|
| 170 |
+
"logps/chosen": -338.43450927734375,
|
| 171 |
+
"logps/rejected": -285.25494384765625,
|
| 172 |
+
"loss": 0.2098,
|
| 173 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 174 |
+
"rewards/chosen": 0.1445818692445755,
|
| 175 |
+
"rewards/margins": 0.19616642594337463,
|
| 176 |
+
"rewards/rejected": -0.05158457159996033,
|
| 177 |
+
"step": 110
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.18,
|
| 181 |
+
"learning_rate": 4.897364164920515e-06,
|
| 182 |
+
"logits/chosen": 0.10053505003452301,
|
| 183 |
+
"logits/rejected": 0.27514562010765076,
|
| 184 |
+
"logps/chosen": -338.99151611328125,
|
| 185 |
+
"logps/rejected": -297.2850646972656,
|
| 186 |
+
"loss": 0.2161,
|
| 187 |
+
"rewards/accuracies": 0.699999988079071,
|
| 188 |
+
"rewards/chosen": 0.14088504016399384,
|
| 189 |
+
"rewards/margins": 0.18905261158943176,
|
| 190 |
+
"rewards/rejected": -0.048167549073696136,
|
| 191 |
+
"step": 120
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.2,
|
| 195 |
+
"learning_rate": 4.8562331973035396e-06,
|
| 196 |
+
"logits/chosen": 0.08438269048929214,
|
| 197 |
+
"logits/rejected": 0.23364977538585663,
|
| 198 |
+
"logps/chosen": -322.02117919921875,
|
| 199 |
+
"logps/rejected": -303.7391357421875,
|
| 200 |
+
"loss": 0.2438,
|
| 201 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 202 |
+
"rewards/chosen": 0.146646648645401,
|
| 203 |
+
"rewards/margins": 0.1478952169418335,
|
| 204 |
+
"rewards/rejected": -0.0012485686456784606,
|
| 205 |
+
"step": 130
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.21,
|
| 209 |
+
"learning_rate": 4.808423230692374e-06,
|
| 210 |
+
"logits/chosen": 0.09936638176441193,
|
| 211 |
+
"logits/rejected": 0.24736297130584717,
|
| 212 |
+
"logps/chosen": -310.3603515625,
|
| 213 |
+
"logps/rejected": -278.47320556640625,
|
| 214 |
+
"loss": 0.222,
|
| 215 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 216 |
+
"rewards/chosen": 0.18732748925685883,
|
| 217 |
+
"rewards/margins": 0.1890837401151657,
|
| 218 |
+
"rewards/rejected": -0.0017562557477504015,
|
| 219 |
+
"step": 140
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.23,
|
| 223 |
+
"learning_rate": 4.754069787631761e-06,
|
| 224 |
+
"logits/chosen": 0.13391128182411194,
|
| 225 |
+
"logits/rejected": 0.2741778492927551,
|
| 226 |
+
"logps/chosen": -366.7373046875,
|
| 227 |
+
"logps/rejected": -295.2054138183594,
|
| 228 |
+
"loss": 0.2071,
|
| 229 |
+
"rewards/accuracies": 0.768750011920929,
|
| 230 |
+
"rewards/chosen": 0.23995013535022736,
|
| 231 |
+
"rewards/margins": 0.24667362868785858,
|
| 232 |
+
"rewards/rejected": -0.006723466329276562,
|
| 233 |
+
"step": 150
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.24,
|
| 237 |
+
"learning_rate": 4.693326938861367e-06,
|
| 238 |
+
"logits/chosen": 0.08330532908439636,
|
| 239 |
+
"logits/rejected": 0.23301962018013,
|
| 240 |
+
"logps/chosen": -295.6157531738281,
|
| 241 |
+
"logps/rejected": -277.6016540527344,
|
| 242 |
+
"loss": 0.2514,
|
| 243 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 244 |
+
"rewards/chosen": 0.19748535752296448,
|
| 245 |
+
"rewards/margins": 0.15948796272277832,
|
| 246 |
+
"rewards/rejected": 0.03799740970134735,
|
| 247 |
+
"step": 160
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.26,
|
| 251 |
+
"learning_rate": 4.626366866585528e-06,
|
| 252 |
+
"logits/chosen": 0.17037127912044525,
|
| 253 |
+
"logits/rejected": 0.3279545307159424,
|
| 254 |
+
"logps/chosen": -371.6747131347656,
|
| 255 |
+
"logps/rejected": -300.25250244140625,
|
| 256 |
+
"loss": 0.226,
|
| 257 |
+
"rewards/accuracies": 0.737500011920929,
|
| 258 |
+
"rewards/chosen": 0.2174214869737625,
|
| 259 |
+
"rewards/margins": 0.1945430338382721,
|
| 260 |
+
"rewards/rejected": 0.022878441959619522,
|
| 261 |
+
"step": 170
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.27,
|
| 265 |
+
"learning_rate": 4.553379376404085e-06,
|
| 266 |
+
"logits/chosen": 0.12852030992507935,
|
| 267 |
+
"logits/rejected": 0.23063895106315613,
|
| 268 |
+
"logps/chosen": -339.1563720703125,
|
| 269 |
+
"logps/rejected": -277.93841552734375,
|
| 270 |
+
"loss": 0.1992,
|
| 271 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 272 |
+
"rewards/chosen": 0.20981720089912415,
|
| 273 |
+
"rewards/margins": 0.22972619533538818,
|
| 274 |
+
"rewards/rejected": -0.019909001886844635,
|
| 275 |
+
"step": 180
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.29,
|
| 279 |
+
"learning_rate": 4.474571359287791e-06,
|
| 280 |
+
"logits/chosen": 0.08759725093841553,
|
| 281 |
+
"logits/rejected": 0.26223450899124146,
|
| 282 |
+
"logps/chosen": -315.0005187988281,
|
| 283 |
+
"logps/rejected": -266.51715087890625,
|
| 284 |
+
"loss": 0.2166,
|
| 285 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 286 |
+
"rewards/chosen": 0.19832256436347961,
|
| 287 |
+
"rewards/margins": 0.2093769758939743,
|
| 288 |
+
"rewards/rejected": -0.011054400354623795,
|
| 289 |
+
"step": 190
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.3,
|
| 293 |
+
"learning_rate": 4.3901662051233755e-06,
|
| 294 |
+
"logits/chosen": 0.1616448611021042,
|
| 295 |
+
"logits/rejected": 0.30854135751724243,
|
| 296 |
+
"logps/chosen": -382.2667236328125,
|
| 297 |
+
"logps/rejected": -286.1771240234375,
|
| 298 |
+
"loss": 0.2198,
|
| 299 |
+
"rewards/accuracies": 0.731249988079071,
|
| 300 |
+
"rewards/chosen": 0.19347664713859558,
|
| 301 |
+
"rewards/margins": 0.20408186316490173,
|
| 302 |
+
"rewards/rejected": -0.010605214163661003,
|
| 303 |
+
"step": 200
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.32,
|
| 307 |
+
"learning_rate": 4.30040316949064e-06,
|
| 308 |
+
"logits/chosen": 0.14471155405044556,
|
| 309 |
+
"logits/rejected": 0.26751500368118286,
|
| 310 |
+
"logps/chosen": -300.742919921875,
|
| 311 |
+
"logps/rejected": -251.2882080078125,
|
| 312 |
+
"loss": 0.23,
|
| 313 |
+
"rewards/accuracies": 0.675000011920929,
|
| 314 |
+
"rewards/chosen": 0.1704670488834381,
|
| 315 |
+
"rewards/margins": 0.18817836046218872,
|
| 316 |
+
"rewards/rejected": -0.0177113339304924,
|
| 317 |
+
"step": 210
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.34,
|
| 321 |
+
"learning_rate": 4.205536695466524e-06,
|
| 322 |
+
"logits/chosen": 0.16468700766563416,
|
| 323 |
+
"logits/rejected": 0.24206213653087616,
|
| 324 |
+
"logps/chosen": -294.42767333984375,
|
| 325 |
+
"logps/rejected": -272.92010498046875,
|
| 326 |
+
"loss": 0.2529,
|
| 327 |
+
"rewards/accuracies": 0.643750011920929,
|
| 328 |
+
"rewards/chosen": 0.17376992106437683,
|
| 329 |
+
"rewards/margins": 0.16116927564144135,
|
| 330 |
+
"rewards/rejected": 0.012600669637322426,
|
| 331 |
+
"step": 220
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.35,
|
| 335 |
+
"learning_rate": 4.105835692378557e-06,
|
| 336 |
+
"logits/chosen": 0.10714814811944962,
|
| 337 |
+
"logits/rejected": 0.2699413597583771,
|
| 338 |
+
"logps/chosen": -304.6961364746094,
|
| 339 |
+
"logps/rejected": -285.6314392089844,
|
| 340 |
+
"loss": 0.1973,
|
| 341 |
+
"rewards/accuracies": 0.731249988079071,
|
| 342 |
+
"rewards/chosen": 0.1876644492149353,
|
| 343 |
+
"rewards/margins": 0.20867836475372314,
|
| 344 |
+
"rewards/rejected": -0.02101389318704605,
|
| 345 |
+
"step": 230
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.37,
|
| 349 |
+
"learning_rate": 4.001582773552153e-06,
|
| 350 |
+
"logits/chosen": 0.06082786247134209,
|
| 351 |
+
"logits/rejected": 0.29382848739624023,
|
| 352 |
+
"logps/chosen": -378.81732177734375,
|
| 353 |
+
"logps/rejected": -309.0159912109375,
|
| 354 |
+
"loss": 0.1942,
|
| 355 |
+
"rewards/accuracies": 0.800000011920929,
|
| 356 |
+
"rewards/chosen": 0.16840167343616486,
|
| 357 |
+
"rewards/margins": 0.24110493063926697,
|
| 358 |
+
"rewards/rejected": -0.0727032721042633,
|
| 359 |
+
"step": 240
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.38,
|
| 363 |
+
"learning_rate": 3.893073455212438e-06,
|
| 364 |
+
"logits/chosen": 0.1116786003112793,
|
| 365 |
+
"logits/rejected": 0.2517862915992737,
|
| 366 |
+
"logps/chosen": -320.56365966796875,
|
| 367 |
+
"logps/rejected": -278.5521545410156,
|
| 368 |
+
"loss": 0.2346,
|
| 369 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 370 |
+
"rewards/chosen": 0.17214366793632507,
|
| 371 |
+
"rewards/margins": 0.2073363959789276,
|
| 372 |
+
"rewards/rejected": -0.03519275039434433,
|
| 373 |
+
"step": 250
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.4,
|
| 377 |
+
"learning_rate": 3.7806153188114027e-06,
|
| 378 |
+
"logits/chosen": 0.11448470503091812,
|
| 379 |
+
"logits/rejected": 0.22813239693641663,
|
| 380 |
+
"logps/chosen": -279.91162109375,
|
| 381 |
+
"logps/rejected": -260.7430114746094,
|
| 382 |
+
"loss": 0.2402,
|
| 383 |
+
"rewards/accuracies": 0.699999988079071,
|
| 384 |
+
"rewards/chosen": 0.1783895492553711,
|
| 385 |
+
"rewards/margins": 0.17652130126953125,
|
| 386 |
+
"rewards/rejected": 0.001868226332589984,
|
| 387 |
+
"step": 260
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.41,
|
| 391 |
+
"learning_rate": 3.6645271391548542e-06,
|
| 392 |
+
"logits/chosen": 0.06586723029613495,
|
| 393 |
+
"logits/rejected": 0.21862812340259552,
|
| 394 |
+
"logps/chosen": -302.5594787597656,
|
| 395 |
+
"logps/rejected": -270.1592712402344,
|
| 396 |
+
"loss": 0.212,
|
| 397 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 398 |
+
"rewards/chosen": 0.1828652024269104,
|
| 399 |
+
"rewards/margins": 0.2052091658115387,
|
| 400 |
+
"rewards/rejected": -0.022343963384628296,
|
| 401 |
+
"step": 270
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.43,
|
| 405 |
+
"learning_rate": 3.5451379808006014e-06,
|
| 406 |
+
"logits/chosen": 0.1570790708065033,
|
| 407 |
+
"logits/rejected": 0.288757860660553,
|
| 408 |
+
"logps/chosen": -347.58013916015625,
|
| 409 |
+
"logps/rejected": -294.15716552734375,
|
| 410 |
+
"loss": 0.1946,
|
| 411 |
+
"rewards/accuracies": 0.762499988079071,
|
| 412 |
+
"rewards/chosen": 0.21899041533470154,
|
| 413 |
+
"rewards/margins": 0.2574598491191864,
|
| 414 |
+
"rewards/rejected": -0.03846944123506546,
|
| 415 |
+
"step": 280
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.44,
|
| 419 |
+
"learning_rate": 3.4227862652892106e-06,
|
| 420 |
+
"logits/chosen": 0.10818709433078766,
|
| 421 |
+
"logits/rejected": 0.23891910910606384,
|
| 422 |
+
"logps/chosen": -348.48370361328125,
|
| 423 |
+
"logps/rejected": -307.1477966308594,
|
| 424 |
+
"loss": 0.201,
|
| 425 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 426 |
+
"rewards/chosen": 0.1834907829761505,
|
| 427 |
+
"rewards/margins": 0.22308149933815002,
|
| 428 |
+
"rewards/rejected": -0.03959069401025772,
|
| 429 |
+
"step": 290
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.46,
|
| 433 |
+
"learning_rate": 3.2978188118513814e-06,
|
| 434 |
+
"logits/chosen": 0.11096982657909393,
|
| 435 |
+
"logits/rejected": 0.24753287434577942,
|
| 436 |
+
"logps/chosen": -296.8543701171875,
|
| 437 |
+
"logps/rejected": -281.43310546875,
|
| 438 |
+
"loss": 0.2306,
|
| 439 |
+
"rewards/accuracies": 0.706250011920929,
|
| 440 |
+
"rewards/chosen": 0.19204005599021912,
|
| 441 |
+
"rewards/margins": 0.18743662536144257,
|
| 442 |
+
"rewards/rejected": 0.0046034445986151695,
|
| 443 |
+
"step": 300
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.47,
|
| 447 |
+
"learning_rate": 3.1705898543111576e-06,
|
| 448 |
+
"logits/chosen": 0.07924026995897293,
|
| 449 |
+
"logits/rejected": 0.2654581665992737,
|
| 450 |
+
"logps/chosen": -327.4444274902344,
|
| 451 |
+
"logps/rejected": -319.9097595214844,
|
| 452 |
+
"loss": 0.2183,
|
| 453 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 454 |
+
"rewards/chosen": 0.1851603388786316,
|
| 455 |
+
"rewards/margins": 0.23894396424293518,
|
| 456 |
+
"rewards/rejected": -0.05378361791372299,
|
| 457 |
+
"step": 310
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.49,
|
| 461 |
+
"learning_rate": 3.041460036971664e-06,
|
| 462 |
+
"logits/chosen": 0.09540507942438126,
|
| 463 |
+
"logits/rejected": 0.22888918220996857,
|
| 464 |
+
"logps/chosen": -334.81689453125,
|
| 465 |
+
"logps/rejected": -270.1705017089844,
|
| 466 |
+
"loss": 0.2204,
|
| 467 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 468 |
+
"rewards/chosen": 0.15298177301883698,
|
| 469 |
+
"rewards/margins": 0.2034742832183838,
|
| 470 |
+
"rewards/rejected": -0.05049251392483711,
|
| 471 |
+
"step": 320
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.5,
|
| 475 |
+
"learning_rate": 2.910795392329649e-06,
|
| 476 |
+
"logits/chosen": 0.11527317762374878,
|
| 477 |
+
"logits/rejected": 0.21196472644805908,
|
| 478 |
+
"logps/chosen": -315.2773132324219,
|
| 479 |
+
"logps/rejected": -283.607421875,
|
| 480 |
+
"loss": 0.1907,
|
| 481 |
+
"rewards/accuracies": 0.706250011920929,
|
| 482 |
+
"rewards/chosen": 0.19066214561462402,
|
| 483 |
+
"rewards/margins": 0.22805961966514587,
|
| 484 |
+
"rewards/rejected": -0.037397462874650955,
|
| 485 |
+
"step": 330
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.52,
|
| 489 |
+
"learning_rate": 2.7789663035166035e-06,
|
| 490 |
+
"logits/chosen": 0.054509587585926056,
|
| 491 |
+
"logits/rejected": 0.21440072357654572,
|
| 492 |
+
"logps/chosen": -327.4574279785156,
|
| 493 |
+
"logps/rejected": -277.25958251953125,
|
| 494 |
+
"loss": 0.201,
|
| 495 |
+
"rewards/accuracies": 0.737500011920929,
|
| 496 |
+
"rewards/chosen": 0.16702668368816376,
|
| 497 |
+
"rewards/margins": 0.22345618903636932,
|
| 498 |
+
"rewards/rejected": -0.05642951279878616,
|
| 499 |
+
"step": 340
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.53,
|
| 503 |
+
"learning_rate": 2.6463464544075344e-06,
|
| 504 |
+
"logits/chosen": 0.0780135840177536,
|
| 505 |
+
"logits/rejected": 0.2536531686782837,
|
| 506 |
+
"logps/chosen": -305.2668151855469,
|
| 507 |
+
"logps/rejected": -291.6106872558594,
|
| 508 |
+
"loss": 0.1948,
|
| 509 |
+
"rewards/accuracies": 0.793749988079071,
|
| 510 |
+
"rewards/chosen": 0.21040305495262146,
|
| 511 |
+
"rewards/margins": 0.27535635232925415,
|
| 512 |
+
"rewards/rejected": -0.06495330482721329,
|
| 513 |
+
"step": 350
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.55,
|
| 517 |
+
"learning_rate": 2.513311770373421e-06,
|
| 518 |
+
"logits/chosen": 0.12684503197669983,
|
| 519 |
+
"logits/rejected": 0.23648087680339813,
|
| 520 |
+
"logps/chosen": -299.1307067871094,
|
| 521 |
+
"logps/rejected": -288.98614501953125,
|
| 522 |
+
"loss": 0.2112,
|
| 523 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 524 |
+
"rewards/chosen": 0.18101270496845245,
|
| 525 |
+
"rewards/margins": 0.24970810115337372,
|
| 526 |
+
"rewards/rejected": -0.06869538873434067,
|
| 527 |
+
"step": 360
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.56,
|
| 531 |
+
"learning_rate": 2.380239352679908e-06,
|
| 532 |
+
"logits/chosen": 0.050291478633880615,
|
| 533 |
+
"logits/rejected": 0.2234155833721161,
|
| 534 |
+
"logps/chosen": -296.54595947265625,
|
| 535 |
+
"logps/rejected": -269.51702880859375,
|
| 536 |
+
"loss": 0.2095,
|
| 537 |
+
"rewards/accuracies": 0.6875,
|
| 538 |
+
"rewards/chosen": 0.17090751230716705,
|
| 539 |
+
"rewards/margins": 0.19086746871471405,
|
| 540 |
+
"rewards/rejected": -0.019959963858127594,
|
| 541 |
+
"step": 370
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.58,
|
| 545 |
+
"learning_rate": 2.247506409552795e-06,
|
| 546 |
+
"logits/chosen": 0.10957841575145721,
|
| 547 |
+
"logits/rejected": 0.21787157654762268,
|
| 548 |
+
"logps/chosen": -305.3380432128906,
|
| 549 |
+
"logps/rejected": -283.4737854003906,
|
| 550 |
+
"loss": 0.2027,
|
| 551 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 552 |
+
"rewards/chosen": 0.17482581734657288,
|
| 553 |
+
"rewards/margins": 0.22671441733837128,
|
| 554 |
+
"rewards/rejected": -0.0518886037170887,
|
| 555 |
+
"step": 380
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.59,
|
| 559 |
+
"learning_rate": 2.1154891869403436e-06,
|
| 560 |
+
"logits/chosen": 0.1384754180908203,
|
| 561 |
+
"logits/rejected": 0.25176170468330383,
|
| 562 |
+
"logps/chosen": -354.0063171386719,
|
| 563 |
+
"logps/rejected": -307.35455322265625,
|
| 564 |
+
"loss": 0.2073,
|
| 565 |
+
"rewards/accuracies": 0.793749988079071,
|
| 566 |
+
"rewards/chosen": 0.2005975991487503,
|
| 567 |
+
"rewards/margins": 0.27450570464134216,
|
| 568 |
+
"rewards/rejected": -0.07390810549259186,
|
| 569 |
+
"step": 390
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.61,
|
| 573 |
+
"learning_rate": 1.9845619020032552e-06,
|
| 574 |
+
"logits/chosen": 0.09188776463270187,
|
| 575 |
+
"logits/rejected": 0.2664358913898468,
|
| 576 |
+
"logps/chosen": -328.11083984375,
|
| 577 |
+
"logps/rejected": -286.30419921875,
|
| 578 |
+
"loss": 0.1895,
|
| 579 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 580 |
+
"rewards/chosen": 0.1963554322719574,
|
| 581 |
+
"rewards/margins": 0.23746831715106964,
|
| 582 |
+
"rewards/rejected": -0.04111289232969284,
|
| 583 |
+
"step": 400
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.62,
|
| 587 |
+
"learning_rate": 1.8550956823554708e-06,
|
| 588 |
+
"logits/chosen": 0.055462319403886795,
|
| 589 |
+
"logits/rejected": 0.18411260843276978,
|
| 590 |
+
"logps/chosen": -326.14703369140625,
|
| 591 |
+
"logps/rejected": -275.77069091796875,
|
| 592 |
+
"loss": 0.235,
|
| 593 |
+
"rewards/accuracies": 0.768750011920929,
|
| 594 |
+
"rewards/chosen": 0.1863834410905838,
|
| 595 |
+
"rewards/margins": 0.22951778769493103,
|
| 596 |
+
"rewards/rejected": -0.043134383857250214,
|
| 597 |
+
"step": 410
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.64,
|
| 601 |
+
"learning_rate": 1.7274575140626318e-06,
|
| 602 |
+
"logits/chosen": 0.09742958843708038,
|
| 603 |
+
"logits/rejected": 0.23133966326713562,
|
| 604 |
+
"logps/chosen": -334.5967712402344,
|
| 605 |
+
"logps/rejected": -302.8537902832031,
|
| 606 |
+
"loss": 0.2048,
|
| 607 |
+
"rewards/accuracies": 0.800000011920929,
|
| 608 |
+
"rewards/chosen": 0.21493402123451233,
|
| 609 |
+
"rewards/margins": 0.26205217838287354,
|
| 610 |
+
"rewards/rejected": -0.047118157148361206,
|
| 611 |
+
"step": 420
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.66,
|
| 615 |
+
"learning_rate": 1.6020092013802002e-06,
|
| 616 |
+
"logits/chosen": 0.061180900782346725,
|
| 617 |
+
"logits/rejected": 0.22411946952342987,
|
| 618 |
+
"logps/chosen": -326.1264953613281,
|
| 619 |
+
"logps/rejected": -270.6374206542969,
|
| 620 |
+
"loss": 0.2211,
|
| 621 |
+
"rewards/accuracies": 0.768750011920929,
|
| 622 |
+
"rewards/chosen": 0.19443106651306152,
|
| 623 |
+
"rewards/margins": 0.2181394100189209,
|
| 624 |
+
"rewards/rejected": -0.023708324879407883,
|
| 625 |
+
"step": 430
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.67,
|
| 629 |
+
"learning_rate": 1.4791063411799938e-06,
|
| 630 |
+
"logits/chosen": 0.11679482460021973,
|
| 631 |
+
"logits/rejected": 0.20075193047523499,
|
| 632 |
+
"logps/chosen": -313.3033752441406,
|
| 633 |
+
"logps/rejected": -301.5423278808594,
|
| 634 |
+
"loss": 0.1941,
|
| 635 |
+
"rewards/accuracies": 0.737500011920929,
|
| 636 |
+
"rewards/chosen": 0.18806949257850647,
|
| 637 |
+
"rewards/margins": 0.22217002511024475,
|
| 638 |
+
"rewards/rejected": -0.03410057723522186,
|
| 639 |
+
"step": 440
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.69,
|
| 643 |
+
"learning_rate": 1.3590973149722103e-06,
|
| 644 |
+
"logits/chosen": 0.11450199782848358,
|
| 645 |
+
"logits/rejected": 0.2375846803188324,
|
| 646 |
+
"logps/chosen": -324.9442443847656,
|
| 647 |
+
"logps/rejected": -282.5760498046875,
|
| 648 |
+
"loss": 0.2164,
|
| 649 |
+
"rewards/accuracies": 0.737500011920929,
|
| 650 |
+
"rewards/chosen": 0.1744813621044159,
|
| 651 |
+
"rewards/margins": 0.21334946155548096,
|
| 652 |
+
"rewards/rejected": -0.03886810690164566,
|
| 653 |
+
"step": 450
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.7,
|
| 657 |
+
"learning_rate": 1.2423223013801946e-06,
|
| 658 |
+
"logits/chosen": 0.10633231699466705,
|
| 659 |
+
"logits/rejected": 0.2275623381137848,
|
| 660 |
+
"logps/chosen": -359.95428466796875,
|
| 661 |
+
"logps/rejected": -296.4113464355469,
|
| 662 |
+
"loss": 0.1969,
|
| 663 |
+
"rewards/accuracies": 0.762499988079071,
|
| 664 |
+
"rewards/chosen": 0.21253260970115662,
|
| 665 |
+
"rewards/margins": 0.24374982714653015,
|
| 666 |
+
"rewards/rejected": -0.03121720813214779,
|
| 667 |
+
"step": 460
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.72,
|
| 671 |
+
"learning_rate": 1.1291123118671665e-06,
|
| 672 |
+
"logits/chosen": 0.03388429060578346,
|
| 673 |
+
"logits/rejected": 0.17446021735668182,
|
| 674 |
+
"logps/chosen": -308.4808349609375,
|
| 675 |
+
"logps/rejected": -263.65283203125,
|
| 676 |
+
"loss": 0.2128,
|
| 677 |
+
"rewards/accuracies": 0.800000011920929,
|
| 678 |
+
"rewards/chosen": 0.21648752689361572,
|
| 679 |
+
"rewards/margins": 0.22755059599876404,
|
| 680 |
+
"rewards/rejected": -0.011063081212341785,
|
| 681 |
+
"step": 470
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.73,
|
| 685 |
+
"learning_rate": 1.019788252448267e-06,
|
| 686 |
+
"logits/chosen": 0.12791678309440613,
|
| 687 |
+
"logits/rejected": 0.28256458044052124,
|
| 688 |
+
"logps/chosen": -379.8475341796875,
|
| 689 |
+
"logps/rejected": -303.40545654296875,
|
| 690 |
+
"loss": 0.182,
|
| 691 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 692 |
+
"rewards/chosen": 0.22206160426139832,
|
| 693 |
+
"rewards/margins": 0.2580808699131012,
|
| 694 |
+
"rewards/rejected": -0.036019258201122284,
|
| 695 |
+
"step": 480
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.75,
|
| 699 |
+
"learning_rate": 9.146600140475945e-07,
|
| 700 |
+
"logits/chosen": 0.11331765353679657,
|
| 701 |
+
"logits/rejected": 0.2163075953722,
|
| 702 |
+
"logps/chosen": -366.92926025390625,
|
| 703 |
+
"logps/rejected": -308.08148193359375,
|
| 704 |
+
"loss": 0.2333,
|
| 705 |
+
"rewards/accuracies": 0.71875,
|
| 706 |
+
"rewards/chosen": 0.1822303682565689,
|
| 707 |
+
"rewards/margins": 0.232683464884758,
|
| 708 |
+
"rewards/rejected": -0.050453104078769684,
|
| 709 |
+
"step": 490
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.76,
|
| 713 |
+
"learning_rate": 8.140255940787059e-07,
|
| 714 |
+
"logits/chosen": 0.1049843281507492,
|
| 715 |
+
"logits/rejected": 0.20622961223125458,
|
| 716 |
+
"logps/chosen": -283.49822998046875,
|
| 717 |
+
"logps/rejected": -262.69561767578125,
|
| 718 |
+
"loss": 0.2244,
|
| 719 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 720 |
+
"rewards/chosen": 0.19038459658622742,
|
| 721 |
+
"rewards/margins": 0.19227764010429382,
|
| 722 |
+
"rewards/rejected": -0.001893045729957521,
|
| 723 |
+
"step": 500
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.78,
|
| 727 |
+
"learning_rate": 7.181702517385789e-07,
|
| 728 |
+
"logits/chosen": 0.06920811533927917,
|
| 729 |
+
"logits/rejected": 0.22420334815979004,
|
| 730 |
+
"logps/chosen": -313.43670654296875,
|
| 731 |
+
"logps/rejected": -280.72442626953125,
|
| 732 |
+
"loss": 0.2521,
|
| 733 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 734 |
+
"rewards/chosen": 0.1772725135087967,
|
| 735 |
+
"rewards/margins": 0.1995859146118164,
|
| 736 |
+
"rewards/rejected": -0.022313417866826057,
|
| 737 |
+
"step": 510
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.79,
|
| 741 |
+
"learning_rate": 6.273656994094232e-07,
|
| 742 |
+
"logits/chosen": 0.0857834741473198,
|
| 743 |
+
"logits/rejected": 0.1759118139743805,
|
| 744 |
+
"logps/chosen": -308.9149475097656,
|
| 745 |
+
"logps/rejected": -283.34747314453125,
|
| 746 |
+
"loss": 0.1955,
|
| 747 |
+
"rewards/accuracies": 0.762499988079071,
|
| 748 |
+
"rewards/chosen": 0.2155129462480545,
|
| 749 |
+
"rewards/margins": 0.22527408599853516,
|
| 750 |
+
"rewards/rejected": -0.009761162102222443,
|
| 751 |
+
"step": 520
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.81,
|
| 755 |
+
"learning_rate": 5.418693324604082e-07,
|
| 756 |
+
"logits/chosen": 0.057937733829021454,
|
| 757 |
+
"logits/rejected": 0.24161191284656525,
|
| 758 |
+
"logps/chosen": -333.49603271484375,
|
| 759 |
+
"logps/rejected": -280.8927917480469,
|
| 760 |
+
"loss": 0.2266,
|
| 761 |
+
"rewards/accuracies": 0.793749988079071,
|
| 762 |
+
"rewards/chosen": 0.21429701149463654,
|
| 763 |
+
"rewards/margins": 0.22269944846630096,
|
| 764 |
+
"rewards/rejected": -0.008402440696954727,
|
| 765 |
+
"step": 530
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.82,
|
| 769 |
+
"learning_rate": 4.619234996325314e-07,
|
| 770 |
+
"logits/chosen": 0.07015545666217804,
|
| 771 |
+
"logits/rejected": 0.2190779447555542,
|
| 772 |
+
"logps/chosen": -344.3935852050781,
|
| 773 |
+
"logps/rejected": -303.2052917480469,
|
| 774 |
+
"loss": 0.2045,
|
| 775 |
+
"rewards/accuracies": 0.824999988079071,
|
| 776 |
+
"rewards/chosen": 0.21076282858848572,
|
| 777 |
+
"rewards/margins": 0.25355157256126404,
|
| 778 |
+
"rewards/rejected": -0.04278876259922981,
|
| 779 |
+
"step": 540
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.84,
|
| 783 |
+
"learning_rate": 3.877548160747768e-07,
|
| 784 |
+
"logits/chosen": 0.08561773598194122,
|
| 785 |
+
"logits/rejected": 0.2999951243400574,
|
| 786 |
+
"logps/chosen": -330.68475341796875,
|
| 787 |
+
"logps/rejected": -281.82275390625,
|
| 788 |
+
"loss": 0.2013,
|
| 789 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 790 |
+
"rewards/chosen": 0.20333750545978546,
|
| 791 |
+
"rewards/margins": 0.23657293617725372,
|
| 792 |
+
"rewards/rejected": -0.033235400915145874,
|
| 793 |
+
"step": 550
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.85,
|
| 797 |
+
"learning_rate": 3.195735209788528e-07,
|
| 798 |
+
"logits/chosen": 0.13458076119422913,
|
| 799 |
+
"logits/rejected": 0.27361050248146057,
|
| 800 |
+
"logps/chosen": -318.90203857421875,
|
| 801 |
+
"logps/rejected": -284.5428466796875,
|
| 802 |
+
"loss": 0.2178,
|
| 803 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 804 |
+
"rewards/chosen": 0.19831877946853638,
|
| 805 |
+
"rewards/margins": 0.23214980959892273,
|
| 806 |
+
"rewards/rejected": -0.03383101895451546,
|
| 807 |
+
"step": 560
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.87,
|
| 811 |
+
"learning_rate": 2.5757288163336806e-07,
|
| 812 |
+
"logits/chosen": 0.11606297641992569,
|
| 813 |
+
"logits/rejected": 0.22765600681304932,
|
| 814 |
+
"logps/chosen": -313.04278564453125,
|
| 815 |
+
"logps/rejected": -281.8597717285156,
|
| 816 |
+
"loss": 0.1959,
|
| 817 |
+
"rewards/accuracies": 0.78125,
|
| 818 |
+
"rewards/chosen": 0.21330790221691132,
|
| 819 |
+
"rewards/margins": 0.25364676117897034,
|
| 820 |
+
"rewards/rejected": -0.04033887758851051,
|
| 821 |
+
"step": 570
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.88,
|
| 825 |
+
"learning_rate": 2.019286455866981e-07,
|
| 826 |
+
"logits/chosen": 0.057523488998413086,
|
| 827 |
+
"logits/rejected": 0.19763953983783722,
|
| 828 |
+
"logps/chosen": -299.98626708984375,
|
| 829 |
+
"logps/rejected": -252.2355194091797,
|
| 830 |
+
"loss": 0.2448,
|
| 831 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 832 |
+
"rewards/chosen": 0.2040693461894989,
|
| 833 |
+
"rewards/margins": 0.18340806663036346,
|
| 834 |
+
"rewards/rejected": 0.020661287009716034,
|
| 835 |
+
"step": 580
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.9,
|
| 839 |
+
"learning_rate": 1.5279854247146703e-07,
|
| 840 |
+
"logits/chosen": 0.10874730348587036,
|
| 841 |
+
"logits/rejected": 0.26722806692123413,
|
| 842 |
+
"logps/chosen": -326.7654724121094,
|
| 843 |
+
"logps/rejected": -274.41162109375,
|
| 844 |
+
"loss": 0.2041,
|
| 845 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 846 |
+
"rewards/chosen": 0.2091568410396576,
|
| 847 |
+
"rewards/margins": 0.2396632879972458,
|
| 848 |
+
"rewards/rejected": -0.030506467446684837,
|
| 849 |
+
"step": 590
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.91,
|
| 853 |
+
"learning_rate": 1.1032183690276754e-07,
|
| 854 |
+
"logits/chosen": 0.12037558853626251,
|
| 855 |
+
"logits/rejected": 0.19973725080490112,
|
| 856 |
+
"logps/chosen": -314.14910888671875,
|
| 857 |
+
"logps/rejected": -270.01043701171875,
|
| 858 |
+
"loss": 0.2167,
|
| 859 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 860 |
+
"rewards/chosen": 0.21347804367542267,
|
| 861 |
+
"rewards/margins": 0.22940710186958313,
|
| 862 |
+
"rewards/rejected": -0.015929043292999268,
|
| 863 |
+
"step": 600
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.93,
|
| 867 |
+
"learning_rate": 7.46189337174788e-08,
|
| 868 |
+
"logits/chosen": 0.08730605989694595,
|
| 869 |
+
"logits/rejected": 0.2364271879196167,
|
| 870 |
+
"logps/chosen": -289.814453125,
|
| 871 |
+
"logps/rejected": -251.5435791015625,
|
| 872 |
+
"loss": 0.2221,
|
| 873 |
+
"rewards/accuracies": 0.71875,
|
| 874 |
+
"rewards/chosen": 0.19059757888317108,
|
| 875 |
+
"rewards/margins": 0.19377049803733826,
|
| 876 |
+
"rewards/rejected": -0.0031729289330542088,
|
| 877 |
+
"step": 610
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.94,
|
| 881 |
+
"learning_rate": 4.579103667367385e-08,
|
| 882 |
+
"logits/chosen": 0.13279291987419128,
|
| 883 |
+
"logits/rejected": 0.2206091433763504,
|
| 884 |
+
"logps/chosen": -342.04498291015625,
|
| 885 |
+
"logps/rejected": -276.22613525390625,
|
| 886 |
+
"loss": 0.1984,
|
| 887 |
+
"rewards/accuracies": 0.71875,
|
| 888 |
+
"rewards/chosen": 0.19695451855659485,
|
| 889 |
+
"rewards/margins": 0.2266121655702591,
|
| 890 |
+
"rewards/rejected": -0.029657626524567604,
|
| 891 |
+
"step": 620
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.96,
|
| 895 |
+
"learning_rate": 2.3919861577572924e-08,
|
| 896 |
+
"logits/chosen": 0.0992569848895073,
|
| 897 |
+
"logits/rejected": 0.26812419295310974,
|
| 898 |
+
"logps/chosen": -336.68817138671875,
|
| 899 |
+
"logps/rejected": -261.6818542480469,
|
| 900 |
+
"loss": 0.2111,
|
| 901 |
+
"rewards/accuracies": 0.762499988079071,
|
| 902 |
+
"rewards/chosen": 0.21192236244678497,
|
| 903 |
+
"rewards/margins": 0.2388623207807541,
|
| 904 |
+
"rewards/rejected": -0.026939954608678818,
|
| 905 |
+
"step": 630
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.98,
|
| 909 |
+
"learning_rate": 9.067404651211808e-09,
|
| 910 |
+
"logits/chosen": 0.04493387043476105,
|
| 911 |
+
"logits/rejected": 0.1637609452009201,
|
| 912 |
+
"logps/chosen": -332.1258544921875,
|
| 913 |
+
"logps/rejected": -283.83709716796875,
|
| 914 |
+
"loss": 0.1939,
|
| 915 |
+
"rewards/accuracies": 0.800000011920929,
|
| 916 |
+
"rewards/chosen": 0.20134110748767853,
|
| 917 |
+
"rewards/margins": 0.23969101905822754,
|
| 918 |
+
"rewards/rejected": -0.03834990784525871,
|
| 919 |
+
"step": 640
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.99,
|
| 923 |
+
"learning_rate": 1.2757667974155896e-09,
|
| 924 |
+
"logits/chosen": 0.1145804151892662,
|
| 925 |
+
"logits/rejected": 0.23529252409934998,
|
| 926 |
+
"logps/chosen": -351.08197021484375,
|
| 927 |
+
"logps/rejected": -293.947998046875,
|
| 928 |
+
"loss": 0.2055,
|
| 929 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 930 |
+
"rewards/chosen": 0.2066545933485031,
|
| 931 |
+
"rewards/margins": 0.22757765650749207,
|
| 932 |
+
"rewards/rejected": -0.020923063158988953,
|
| 933 |
+
"step": 650
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 1.0,
|
| 937 |
+
"step": 656,
|
| 938 |
+
"total_flos": 0.0,
|
| 939 |
+
"train_loss": 0.22703545004492853,
|
| 940 |
+
"train_runtime": 7888.8763,
|
| 941 |
+
"train_samples_per_second": 2.662,
|
| 942 |
+
"train_steps_per_second": 0.083
|
| 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 |
+
}
|