Model save
Browse files- README.md +61 -0
- adapter_model.safetensors +1 -1
- all_results.json +8 -0
- runs/May18_08-24-43_gpu4-119-5/events.out.tfevents.1715984828.gpu4-119-5.2087616.0 +2 -2
- train_results.json +8 -0
- trainer_state.json +912 -0
README.md
ADDED
|
@@ -0,0 +1,61 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: peft
|
| 3 |
+
tags:
|
| 4 |
+
- trl
|
| 5 |
+
- dpo
|
| 6 |
+
- generated_from_trainer
|
| 7 |
+
base_model: DUAL-GPO-2/phi-2-gpo-v34-merged-i1
|
| 8 |
+
model-index:
|
| 9 |
+
- name: phi-2-gpo-v6-i2
|
| 10 |
+
results: []
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 14 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 15 |
+
|
| 16 |
+
# phi-2-gpo-v6-i2
|
| 17 |
+
|
| 18 |
+
This model is a fine-tuned version of [DUAL-GPO-2/phi-2-gpo-v34-merged-i1](https://huggingface.co/DUAL-GPO-2/phi-2-gpo-v34-merged-i1) on the None dataset.
|
| 19 |
+
|
| 20 |
+
## Model description
|
| 21 |
+
|
| 22 |
+
More information needed
|
| 23 |
+
|
| 24 |
+
## Intended uses & limitations
|
| 25 |
+
|
| 26 |
+
More information needed
|
| 27 |
+
|
| 28 |
+
## Training and evaluation data
|
| 29 |
+
|
| 30 |
+
More information needed
|
| 31 |
+
|
| 32 |
+
## Training procedure
|
| 33 |
+
|
| 34 |
+
### Training hyperparameters
|
| 35 |
+
|
| 36 |
+
The following hyperparameters were used during training:
|
| 37 |
+
- learning_rate: 5e-06
|
| 38 |
+
- train_batch_size: 4
|
| 39 |
+
- eval_batch_size: 4
|
| 40 |
+
- seed: 42
|
| 41 |
+
- distributed_type: multi-GPU
|
| 42 |
+
- num_devices: 3
|
| 43 |
+
- gradient_accumulation_steps: 4
|
| 44 |
+
- total_train_batch_size: 48
|
| 45 |
+
- total_eval_batch_size: 12
|
| 46 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 47 |
+
- lr_scheduler_type: cosine
|
| 48 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 49 |
+
- num_epochs: 1
|
| 50 |
+
|
| 51 |
+
### Training results
|
| 52 |
+
|
| 53 |
+
|
| 54 |
+
|
| 55 |
+
### Framework versions
|
| 56 |
+
|
| 57 |
+
- PEFT 0.7.1
|
| 58 |
+
- Transformers 4.36.2
|
| 59 |
+
- Pytorch 2.1.2+cu121
|
| 60 |
+
- Datasets 2.14.6
|
| 61 |
+
- Tokenizers 0.15.2
|
adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 167807296
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:07ecc5e82696068b1cb437de6b495e0215e018f0aedf3b996f068c11a24d1572
|
| 3 |
size 167807296
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.13047290132045747,
|
| 4 |
+
"train_runtime": 8392.5195,
|
| 5 |
+
"train_samples": 30000,
|
| 6 |
+
"train_samples_per_second": 3.575,
|
| 7 |
+
"train_steps_per_second": 0.074
|
| 8 |
+
}
|
runs/May18_08-24-43_gpu4-119-5/events.out.tfevents.1715984828.gpu4-119-5.2087616.0
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 sha256:8261b018d11b68ab0002d125c6922ceb23fb29715c9de215a07aa296baeac165
|
| 3 |
+
size 44993
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.13047290132045747,
|
| 4 |
+
"train_runtime": 8392.5195,
|
| 5 |
+
"train_samples": 30000,
|
| 6 |
+
"train_samples_per_second": 3.575,
|
| 7 |
+
"train_steps_per_second": 0.074
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,912 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 625,
|
| 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": 7.936507936507937e-08,
|
| 14 |
+
"logits/chosen": -0.2010916769504547,
|
| 15 |
+
"logits/rejected": 0.09005054831504822,
|
| 16 |
+
"logps/chosen": -540.942626953125,
|
| 17 |
+
"logps/rejected": -796.8775634765625,
|
| 18 |
+
"loss": 0.2182,
|
| 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": 7.936507936507937e-07,
|
| 28 |
+
"logits/chosen": -0.15247675776481628,
|
| 29 |
+
"logits/rejected": -0.14707261323928833,
|
| 30 |
+
"logps/chosen": -501.8849792480469,
|
| 31 |
+
"logps/rejected": -774.5216064453125,
|
| 32 |
+
"loss": 0.2086,
|
| 33 |
+
"rewards/accuracies": 0.375,
|
| 34 |
+
"rewards/chosen": 3.564198050298728e-05,
|
| 35 |
+
"rewards/margins": -2.1308711438905448e-05,
|
| 36 |
+
"rewards/rejected": 5.695069921785034e-05,
|
| 37 |
+
"step": 10
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.03,
|
| 41 |
+
"learning_rate": 1.5873015873015873e-06,
|
| 42 |
+
"logits/chosen": -0.1420287936925888,
|
| 43 |
+
"logits/rejected": -0.10379602760076523,
|
| 44 |
+
"logps/chosen": -489.9579162597656,
|
| 45 |
+
"logps/rejected": -771.2081298828125,
|
| 46 |
+
"loss": 0.2095,
|
| 47 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 48 |
+
"rewards/chosen": -0.0012544477358460426,
|
| 49 |
+
"rewards/margins": 0.0017321283230558038,
|
| 50 |
+
"rewards/rejected": -0.0029865759424865246,
|
| 51 |
+
"step": 20
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.05,
|
| 55 |
+
"learning_rate": 2.380952380952381e-06,
|
| 56 |
+
"logits/chosen": -0.16283434629440308,
|
| 57 |
+
"logits/rejected": -0.14502206444740295,
|
| 58 |
+
"logps/chosen": -532.1380615234375,
|
| 59 |
+
"logps/rejected": -800.226806640625,
|
| 60 |
+
"loss": 0.1996,
|
| 61 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 62 |
+
"rewards/chosen": -0.007352855056524277,
|
| 63 |
+
"rewards/margins": 0.009955727495253086,
|
| 64 |
+
"rewards/rejected": -0.017308581620454788,
|
| 65 |
+
"step": 30
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.06,
|
| 69 |
+
"learning_rate": 3.1746031746031746e-06,
|
| 70 |
+
"logits/chosen": -0.19371934235095978,
|
| 71 |
+
"logits/rejected": -0.15917012095451355,
|
| 72 |
+
"logps/chosen": -529.4437255859375,
|
| 73 |
+
"logps/rejected": -794.9968872070312,
|
| 74 |
+
"loss": 0.1897,
|
| 75 |
+
"rewards/accuracies": 0.731249988079071,
|
| 76 |
+
"rewards/chosen": -0.028293650597333908,
|
| 77 |
+
"rewards/margins": 0.02773173525929451,
|
| 78 |
+
"rewards/rejected": -0.05602538585662842,
|
| 79 |
+
"step": 40
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.08,
|
| 83 |
+
"learning_rate": 3.968253968253968e-06,
|
| 84 |
+
"logits/chosen": -0.16935278475284576,
|
| 85 |
+
"logits/rejected": -0.12196620553731918,
|
| 86 |
+
"logps/chosen": -556.66064453125,
|
| 87 |
+
"logps/rejected": -913.4127197265625,
|
| 88 |
+
"loss": 0.1623,
|
| 89 |
+
"rewards/accuracies": 0.731249988079071,
|
| 90 |
+
"rewards/chosen": -0.06831763684749603,
|
| 91 |
+
"rewards/margins": 0.09133367240428925,
|
| 92 |
+
"rewards/rejected": -0.15965132415294647,
|
| 93 |
+
"step": 50
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.1,
|
| 97 |
+
"learning_rate": 4.761904761904762e-06,
|
| 98 |
+
"logits/chosen": -0.18819646537303925,
|
| 99 |
+
"logits/rejected": -0.21151557564735413,
|
| 100 |
+
"logps/chosen": -641.8372802734375,
|
| 101 |
+
"logps/rejected": -1064.3094482421875,
|
| 102 |
+
"loss": 0.1532,
|
| 103 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 104 |
+
"rewards/chosen": -0.14167475700378418,
|
| 105 |
+
"rewards/margins": 0.16748657822608948,
|
| 106 |
+
"rewards/rejected": -0.30916133522987366,
|
| 107 |
+
"step": 60
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.11,
|
| 111 |
+
"learning_rate": 4.998086282661188e-06,
|
| 112 |
+
"logits/chosen": -0.19726331532001495,
|
| 113 |
+
"logits/rejected": -0.2387177050113678,
|
| 114 |
+
"logps/chosen": -630.7346801757812,
|
| 115 |
+
"logps/rejected": -1035.373046875,
|
| 116 |
+
"loss": 0.1511,
|
| 117 |
+
"rewards/accuracies": 0.731249988079071,
|
| 118 |
+
"rewards/chosen": -0.13452503085136414,
|
| 119 |
+
"rewards/margins": 0.14926694333553314,
|
| 120 |
+
"rewards/rejected": -0.2837919592857361,
|
| 121 |
+
"step": 70
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.13,
|
| 125 |
+
"learning_rate": 4.988720025682995e-06,
|
| 126 |
+
"logits/chosen": -0.23227183520793915,
|
| 127 |
+
"logits/rejected": -0.15486109256744385,
|
| 128 |
+
"logps/chosen": -612.7208862304688,
|
| 129 |
+
"logps/rejected": -1018.8267822265625,
|
| 130 |
+
"loss": 0.142,
|
| 131 |
+
"rewards/accuracies": 0.737500011920929,
|
| 132 |
+
"rewards/chosen": -0.11648458242416382,
|
| 133 |
+
"rewards/margins": 0.137832373380661,
|
| 134 |
+
"rewards/rejected": -0.25431695580482483,
|
| 135 |
+
"step": 80
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.14,
|
| 139 |
+
"learning_rate": 4.9715789537359126e-06,
|
| 140 |
+
"logits/chosen": -0.269733726978302,
|
| 141 |
+
"logits/rejected": -0.2005140334367752,
|
| 142 |
+
"logps/chosen": -666.8603515625,
|
| 143 |
+
"logps/rejected": -959.771484375,
|
| 144 |
+
"loss": 0.1449,
|
| 145 |
+
"rewards/accuracies": 0.706250011920929,
|
| 146 |
+
"rewards/chosen": -0.1299312561750412,
|
| 147 |
+
"rewards/margins": 0.12538622319698334,
|
| 148 |
+
"rewards/rejected": -0.25531744956970215,
|
| 149 |
+
"step": 90
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.16,
|
| 153 |
+
"learning_rate": 4.946716615897932e-06,
|
| 154 |
+
"logits/chosen": -0.21513569355010986,
|
| 155 |
+
"logits/rejected": -0.22341570258140564,
|
| 156 |
+
"logps/chosen": -636.3090209960938,
|
| 157 |
+
"logps/rejected": -982.4715576171875,
|
| 158 |
+
"loss": 0.1376,
|
| 159 |
+
"rewards/accuracies": 0.768750011920929,
|
| 160 |
+
"rewards/chosen": -0.1491270661354065,
|
| 161 |
+
"rewards/margins": 0.12507781386375427,
|
| 162 |
+
"rewards/rejected": -0.27420490980148315,
|
| 163 |
+
"step": 100
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.18,
|
| 167 |
+
"learning_rate": 4.9142106826480114e-06,
|
| 168 |
+
"logits/chosen": -0.26347213983535767,
|
| 169 |
+
"logits/rejected": -0.1899401843547821,
|
| 170 |
+
"logps/chosen": -682.718017578125,
|
| 171 |
+
"logps/rejected": -1051.3671875,
|
| 172 |
+
"loss": 0.1456,
|
| 173 |
+
"rewards/accuracies": 0.737500011920929,
|
| 174 |
+
"rewards/chosen": -0.1585748940706253,
|
| 175 |
+
"rewards/margins": 0.13942097127437592,
|
| 176 |
+
"rewards/rejected": -0.2979958653450012,
|
| 177 |
+
"step": 110
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.19,
|
| 181 |
+
"learning_rate": 4.874162703221823e-06,
|
| 182 |
+
"logits/chosen": -0.28930234909057617,
|
| 183 |
+
"logits/rejected": -0.21862976253032684,
|
| 184 |
+
"logps/chosen": -624.7803344726562,
|
| 185 |
+
"logps/rejected": -1044.067626953125,
|
| 186 |
+
"loss": 0.1322,
|
| 187 |
+
"rewards/accuracies": 0.768750011920929,
|
| 188 |
+
"rewards/chosen": -0.1492077112197876,
|
| 189 |
+
"rewards/margins": 0.15546968579292297,
|
| 190 |
+
"rewards/rejected": -0.3046773374080658,
|
| 191 |
+
"step": 120
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.21,
|
| 195 |
+
"learning_rate": 4.826697788369752e-06,
|
| 196 |
+
"logits/chosen": -0.2196696251630783,
|
| 197 |
+
"logits/rejected": -0.22875969111919403,
|
| 198 |
+
"logps/chosen": -621.0065307617188,
|
| 199 |
+
"logps/rejected": -944.654296875,
|
| 200 |
+
"loss": 0.1432,
|
| 201 |
+
"rewards/accuracies": 0.71875,
|
| 202 |
+
"rewards/chosen": -0.15334677696228027,
|
| 203 |
+
"rewards/margins": 0.11417926847934723,
|
| 204 |
+
"rewards/rejected": -0.2675260007381439,
|
| 205 |
+
"step": 130
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.22,
|
| 209 |
+
"learning_rate": 4.7719642195082224e-06,
|
| 210 |
+
"logits/chosen": -0.23348715901374817,
|
| 211 |
+
"logits/rejected": -0.21704678237438202,
|
| 212 |
+
"logps/chosen": -599.64453125,
|
| 213 |
+
"logps/rejected": -991.4964599609375,
|
| 214 |
+
"loss": 0.1471,
|
| 215 |
+
"rewards/accuracies": 0.75,
|
| 216 |
+
"rewards/chosen": -0.13741596043109894,
|
| 217 |
+
"rewards/margins": 0.13189604878425598,
|
| 218 |
+
"rewards/rejected": -0.2693119943141937,
|
| 219 |
+
"step": 140
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.24,
|
| 223 |
+
"learning_rate": 4.710132985485355e-06,
|
| 224 |
+
"logits/chosen": -0.2200555056333542,
|
| 225 |
+
"logits/rejected": -0.20334219932556152,
|
| 226 |
+
"logps/chosen": -661.56884765625,
|
| 227 |
+
"logps/rejected": -1054.980712890625,
|
| 228 |
+
"loss": 0.1376,
|
| 229 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 230 |
+
"rewards/chosen": -0.16635623574256897,
|
| 231 |
+
"rewards/margins": 0.14565840363502502,
|
| 232 |
+
"rewards/rejected": -0.312014639377594,
|
| 233 |
+
"step": 150
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.26,
|
| 237 |
+
"learning_rate": 4.641397248408122e-06,
|
| 238 |
+
"logits/chosen": -0.20752374827861786,
|
| 239 |
+
"logits/rejected": -0.24948814511299133,
|
| 240 |
+
"logps/chosen": -719.3741455078125,
|
| 241 |
+
"logps/rejected": -1020.0384521484375,
|
| 242 |
+
"loss": 0.1338,
|
| 243 |
+
"rewards/accuracies": 0.78125,
|
| 244 |
+
"rewards/chosen": -0.1749049723148346,
|
| 245 |
+
"rewards/margins": 0.13863129913806915,
|
| 246 |
+
"rewards/rejected": -0.31353622674942017,
|
| 247 |
+
"step": 160
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.27,
|
| 251 |
+
"learning_rate": 4.5659717401997655e-06,
|
| 252 |
+
"logits/chosen": -0.2512062191963196,
|
| 253 |
+
"logits/rejected": -0.25224044919013977,
|
| 254 |
+
"logps/chosen": -652.2489624023438,
|
| 255 |
+
"logps/rejected": -1064.6927490234375,
|
| 256 |
+
"loss": 0.1404,
|
| 257 |
+
"rewards/accuracies": 0.78125,
|
| 258 |
+
"rewards/chosen": -0.16943642497062683,
|
| 259 |
+
"rewards/margins": 0.14737632870674133,
|
| 260 |
+
"rewards/rejected": -0.3168127238750458,
|
| 261 |
+
"step": 170
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.29,
|
| 265 |
+
"learning_rate": 4.4840920917726425e-06,
|
| 266 |
+
"logits/chosen": -0.26564353704452515,
|
| 267 |
+
"logits/rejected": -0.31829017400741577,
|
| 268 |
+
"logps/chosen": -643.8062744140625,
|
| 269 |
+
"logps/rejected": -1146.453857421875,
|
| 270 |
+
"loss": 0.1238,
|
| 271 |
+
"rewards/accuracies": 0.78125,
|
| 272 |
+
"rewards/chosen": -0.1721421331167221,
|
| 273 |
+
"rewards/margins": 0.17220233380794525,
|
| 274 |
+
"rewards/rejected": -0.34434446692466736,
|
| 275 |
+
"step": 180
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.3,
|
| 279 |
+
"learning_rate": 4.396014096912182e-06,
|
| 280 |
+
"logits/chosen": -0.30384334921836853,
|
| 281 |
+
"logits/rejected": -0.3049886226654053,
|
| 282 |
+
"logps/chosen": -700.3992919921875,
|
| 283 |
+
"logps/rejected": -1094.860107421875,
|
| 284 |
+
"loss": 0.1293,
|
| 285 |
+
"rewards/accuracies": 0.800000011920929,
|
| 286 |
+
"rewards/chosen": -0.190101757645607,
|
| 287 |
+
"rewards/margins": 0.16493478417396545,
|
| 288 |
+
"rewards/rejected": -0.35503652691841125,
|
| 289 |
+
"step": 190
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.32,
|
| 293 |
+
"learning_rate": 4.302012913171584e-06,
|
| 294 |
+
"logits/chosen": -0.22822122275829315,
|
| 295 |
+
"logits/rejected": -0.21796974539756775,
|
| 296 |
+
"logps/chosen": -703.4880981445312,
|
| 297 |
+
"logps/rejected": -1098.6363525390625,
|
| 298 |
+
"loss": 0.1291,
|
| 299 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 300 |
+
"rewards/chosen": -0.18932084739208221,
|
| 301 |
+
"rewards/margins": 0.18700966238975525,
|
| 302 |
+
"rewards/rejected": -0.37633052468299866,
|
| 303 |
+
"step": 200
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.34,
|
| 307 |
+
"learning_rate": 4.202382202273702e-06,
|
| 308 |
+
"logits/chosen": -0.2829930782318115,
|
| 309 |
+
"logits/rejected": -0.2941269278526306,
|
| 310 |
+
"logps/chosen": -712.0001220703125,
|
| 311 |
+
"logps/rejected": -1121.9896240234375,
|
| 312 |
+
"loss": 0.1212,
|
| 313 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 314 |
+
"rewards/chosen": -0.19748489558696747,
|
| 315 |
+
"rewards/margins": 0.1557844579219818,
|
| 316 |
+
"rewards/rejected": -0.3532693684101105,
|
| 317 |
+
"step": 210
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.35,
|
| 321 |
+
"learning_rate": 4.097433212705492e-06,
|
| 322 |
+
"logits/chosen": -0.2563570737838745,
|
| 323 |
+
"logits/rejected": -0.20130082964897156,
|
| 324 |
+
"logps/chosen": -719.8380126953125,
|
| 325 |
+
"logps/rejected": -1181.190185546875,
|
| 326 |
+
"loss": 0.1349,
|
| 327 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 328 |
+
"rewards/chosen": -0.2003975212574005,
|
| 329 |
+
"rewards/margins": 0.1832597851753235,
|
| 330 |
+
"rewards/rejected": -0.3836573362350464,
|
| 331 |
+
"step": 220
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.37,
|
| 335 |
+
"learning_rate": 3.987493807371033e-06,
|
| 336 |
+
"logits/chosen": -0.19761434197425842,
|
| 337 |
+
"logits/rejected": -0.24285447597503662,
|
| 338 |
+
"logps/chosen": -713.5037841796875,
|
| 339 |
+
"logps/rejected": -1127.9371337890625,
|
| 340 |
+
"loss": 0.1377,
|
| 341 |
+
"rewards/accuracies": 0.793749988079071,
|
| 342 |
+
"rewards/chosen": -0.20124192535877228,
|
| 343 |
+
"rewards/margins": 0.16942095756530762,
|
| 344 |
+
"rewards/rejected": -0.3706628680229187,
|
| 345 |
+
"step": 230
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.38,
|
| 349 |
+
"learning_rate": 3.872907439340758e-06,
|
| 350 |
+
"logits/chosen": -0.22070392966270447,
|
| 351 |
+
"logits/rejected": -0.2395116537809372,
|
| 352 |
+
"logps/chosen": -689.1747436523438,
|
| 353 |
+
"logps/rejected": -1246.12646484375,
|
| 354 |
+
"loss": 0.1219,
|
| 355 |
+
"rewards/accuracies": 0.862500011920929,
|
| 356 |
+
"rewards/chosen": -0.18724720180034637,
|
| 357 |
+
"rewards/margins": 0.221848726272583,
|
| 358 |
+
"rewards/rejected": -0.4090959429740906,
|
| 359 |
+
"step": 240
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.4,
|
| 363 |
+
"learning_rate": 3.75403207889666e-06,
|
| 364 |
+
"logits/chosen": -0.2952207028865814,
|
| 365 |
+
"logits/rejected": -0.286382257938385,
|
| 366 |
+
"logps/chosen": -680.5881958007812,
|
| 367 |
+
"logps/rejected": -1075.620849609375,
|
| 368 |
+
"loss": 0.1351,
|
| 369 |
+
"rewards/accuracies": 0.78125,
|
| 370 |
+
"rewards/chosen": -0.20417407155036926,
|
| 371 |
+
"rewards/margins": 0.16537046432495117,
|
| 372 |
+
"rewards/rejected": -0.36954453587532043,
|
| 373 |
+
"step": 250
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.42,
|
| 377 |
+
"learning_rate": 3.631239095225417e-06,
|
| 378 |
+
"logits/chosen": -0.28091198205947876,
|
| 379 |
+
"logits/rejected": -0.25016146898269653,
|
| 380 |
+
"logps/chosen": -717.04150390625,
|
| 381 |
+
"logps/rejected": -1180.3306884765625,
|
| 382 |
+
"loss": 0.1185,
|
| 383 |
+
"rewards/accuracies": 0.78125,
|
| 384 |
+
"rewards/chosen": -0.2006029188632965,
|
| 385 |
+
"rewards/margins": 0.1787930279970169,
|
| 386 |
+
"rewards/rejected": -0.3793959617614746,
|
| 387 |
+
"step": 260
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.43,
|
| 391 |
+
"learning_rate": 3.5049120962530608e-06,
|
| 392 |
+
"logits/chosen": -0.23318138718605042,
|
| 393 |
+
"logits/rejected": -0.1964143067598343,
|
| 394 |
+
"logps/chosen": -722.5912475585938,
|
| 395 |
+
"logps/rejected": -1105.978759765625,
|
| 396 |
+
"loss": 0.1251,
|
| 397 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 398 |
+
"rewards/chosen": -0.20104451477527618,
|
| 399 |
+
"rewards/margins": 0.16880543529987335,
|
| 400 |
+
"rewards/rejected": -0.36984992027282715,
|
| 401 |
+
"step": 270
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.45,
|
| 405 |
+
"learning_rate": 3.3754457302455464e-06,
|
| 406 |
+
"logits/chosen": -0.21553221344947815,
|
| 407 |
+
"logits/rejected": -0.2545991837978363,
|
| 408 |
+
"logps/chosen": -708.6617431640625,
|
| 409 |
+
"logps/rejected": -1200.54248046875,
|
| 410 |
+
"loss": 0.1241,
|
| 411 |
+
"rewards/accuracies": 0.824999988079071,
|
| 412 |
+
"rewards/chosen": -0.1849481165409088,
|
| 413 |
+
"rewards/margins": 0.22220449149608612,
|
| 414 |
+
"rewards/rejected": -0.40715259313583374,
|
| 415 |
+
"step": 280
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.46,
|
| 419 |
+
"learning_rate": 3.2432444529190714e-06,
|
| 420 |
+
"logits/chosen": -0.2878536283969879,
|
| 421 |
+
"logits/rejected": -0.240807443857193,
|
| 422 |
+
"logps/chosen": -685.4102783203125,
|
| 423 |
+
"logps/rejected": -1073.6112060546875,
|
| 424 |
+
"loss": 0.1278,
|
| 425 |
+
"rewards/accuracies": 0.8187500238418579,
|
| 426 |
+
"rewards/chosen": -0.19928036630153656,
|
| 427 |
+
"rewards/margins": 0.167768195271492,
|
| 428 |
+
"rewards/rejected": -0.36704859137535095,
|
| 429 |
+
"step": 290
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.48,
|
| 433 |
+
"learning_rate": 3.1087212639117057e-06,
|
| 434 |
+
"logits/chosen": -0.22543080151081085,
|
| 435 |
+
"logits/rejected": -0.25053372979164124,
|
| 436 |
+
"logps/chosen": -713.0574340820312,
|
| 437 |
+
"logps/rejected": -1165.50244140625,
|
| 438 |
+
"loss": 0.1199,
|
| 439 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 440 |
+
"rewards/chosen": -0.20086824893951416,
|
| 441 |
+
"rewards/margins": 0.19784381985664368,
|
| 442 |
+
"rewards/rejected": -0.39871203899383545,
|
| 443 |
+
"step": 300
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.5,
|
| 447 |
+
"learning_rate": 2.9722964165636263e-06,
|
| 448 |
+
"logits/chosen": -0.2477823793888092,
|
| 449 |
+
"logits/rejected": -0.23944005370140076,
|
| 450 |
+
"logps/chosen": -741.0659790039062,
|
| 451 |
+
"logps/rejected": -1150.4737548828125,
|
| 452 |
+
"loss": 0.1334,
|
| 453 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 454 |
+
"rewards/chosen": -0.21726730465888977,
|
| 455 |
+
"rewards/margins": 0.18077899515628815,
|
| 456 |
+
"rewards/rejected": -0.3980463147163391,
|
| 457 |
+
"step": 310
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.51,
|
| 461 |
+
"learning_rate": 2.8343961050366275e-06,
|
| 462 |
+
"logits/chosen": -0.206426739692688,
|
| 463 |
+
"logits/rejected": -0.294622004032135,
|
| 464 |
+
"logps/chosen": -698.1879272460938,
|
| 465 |
+
"logps/rejected": -1074.371337890625,
|
| 466 |
+
"loss": 0.1279,
|
| 467 |
+
"rewards/accuracies": 0.762499988079071,
|
| 468 |
+
"rewards/chosen": -0.1868043690919876,
|
| 469 |
+
"rewards/margins": 0.1699267327785492,
|
| 470 |
+
"rewards/rejected": -0.3567310869693756,
|
| 471 |
+
"step": 320
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.53,
|
| 475 |
+
"learning_rate": 2.695451132874385e-06,
|
| 476 |
+
"logits/chosen": -0.2724359631538391,
|
| 477 |
+
"logits/rejected": -0.21938621997833252,
|
| 478 |
+
"logps/chosen": -708.3114013671875,
|
| 479 |
+
"logps/rejected": -1099.7318115234375,
|
| 480 |
+
"loss": 0.1196,
|
| 481 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 482 |
+
"rewards/chosen": -0.18547296524047852,
|
| 483 |
+
"rewards/margins": 0.1704384684562683,
|
| 484 |
+
"rewards/rejected": -0.3559114336967468,
|
| 485 |
+
"step": 330
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.54,
|
| 489 |
+
"learning_rate": 2.5558955671628964e-06,
|
| 490 |
+
"logits/chosen": -0.25434714555740356,
|
| 491 |
+
"logits/rejected": -0.25090768933296204,
|
| 492 |
+
"logps/chosen": -695.5904541015625,
|
| 493 |
+
"logps/rejected": -1176.5301513671875,
|
| 494 |
+
"loss": 0.131,
|
| 495 |
+
"rewards/accuracies": 0.768750011920929,
|
| 496 |
+
"rewards/chosen": -0.21099016070365906,
|
| 497 |
+
"rewards/margins": 0.20053663849830627,
|
| 498 |
+
"rewards/rejected": -0.4115268290042877,
|
| 499 |
+
"step": 340
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.56,
|
| 503 |
+
"learning_rate": 2.4161653824955654e-06,
|
| 504 |
+
"logits/chosen": -0.28552037477493286,
|
| 505 |
+
"logits/rejected": -0.24306419491767883,
|
| 506 |
+
"logps/chosen": -764.7316284179688,
|
| 507 |
+
"logps/rejected": -1231.4193115234375,
|
| 508 |
+
"loss": 0.1166,
|
| 509 |
+
"rewards/accuracies": 0.8187500238418579,
|
| 510 |
+
"rewards/chosen": -0.24171674251556396,
|
| 511 |
+
"rewards/margins": 0.21239931881427765,
|
| 512 |
+
"rewards/rejected": -0.4541160464286804,
|
| 513 |
+
"step": 350
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.58,
|
| 517 |
+
"learning_rate": 2.2766970989791697e-06,
|
| 518 |
+
"logits/chosen": -0.25456491112709045,
|
| 519 |
+
"logits/rejected": -0.21178500354290009,
|
| 520 |
+
"logps/chosen": -765.3328247070312,
|
| 521 |
+
"logps/rejected": -1261.208984375,
|
| 522 |
+
"loss": 0.1103,
|
| 523 |
+
"rewards/accuracies": 0.824999988079071,
|
| 524 |
+
"rewards/chosen": -0.22941815853118896,
|
| 525 |
+
"rewards/margins": 0.21526849269866943,
|
| 526 |
+
"rewards/rejected": -0.4446867108345032,
|
| 527 |
+
"step": 360
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.59,
|
| 531 |
+
"learning_rate": 2.1379264185356545e-06,
|
| 532 |
+
"logits/chosen": -0.2724260687828064,
|
| 533 |
+
"logits/rejected": -0.18538828194141388,
|
| 534 |
+
"logps/chosen": -741.7236328125,
|
| 535 |
+
"logps/rejected": -1078.375732421875,
|
| 536 |
+
"loss": 0.1161,
|
| 537 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 538 |
+
"rewards/chosen": -0.21514591574668884,
|
| 539 |
+
"rewards/margins": 0.16624660789966583,
|
| 540 |
+
"rewards/rejected": -0.38139253854751587,
|
| 541 |
+
"step": 370
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.61,
|
| 545 |
+
"learning_rate": 2.000286863759934e-06,
|
| 546 |
+
"logits/chosen": -0.27519670128822327,
|
| 547 |
+
"logits/rejected": -0.2832408845424652,
|
| 548 |
+
"logps/chosen": -682.8367309570312,
|
| 549 |
+
"logps/rejected": -1131.5003662109375,
|
| 550 |
+
"loss": 0.1198,
|
| 551 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 552 |
+
"rewards/chosen": -0.20102819800376892,
|
| 553 |
+
"rewards/margins": 0.18376831710338593,
|
| 554 |
+
"rewards/rejected": -0.38479650020599365,
|
| 555 |
+
"step": 380
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.62,
|
| 559 |
+
"learning_rate": 1.8642084235859764e-06,
|
| 560 |
+
"logits/chosen": -0.2652140259742737,
|
| 561 |
+
"logits/rejected": -0.2889960706233978,
|
| 562 |
+
"logps/chosen": -731.1627807617188,
|
| 563 |
+
"logps/rejected": -1137.467041015625,
|
| 564 |
+
"loss": 0.1208,
|
| 565 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 566 |
+
"rewards/chosen": -0.19975660741329193,
|
| 567 |
+
"rewards/margins": 0.21678335964679718,
|
| 568 |
+
"rewards/rejected": -0.4165399968624115,
|
| 569 |
+
"step": 390
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.64,
|
| 573 |
+
"learning_rate": 1.7301162099921013e-06,
|
| 574 |
+
"logits/chosen": -0.26221686601638794,
|
| 575 |
+
"logits/rejected": -0.2641231119632721,
|
| 576 |
+
"logps/chosen": -755.5460205078125,
|
| 577 |
+
"logps/rejected": -1093.4853515625,
|
| 578 |
+
"loss": 0.1228,
|
| 579 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 580 |
+
"rewards/chosen": -0.21798157691955566,
|
| 581 |
+
"rewards/margins": 0.17120864987373352,
|
| 582 |
+
"rewards/rejected": -0.3891902267932892,
|
| 583 |
+
"step": 400
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.66,
|
| 587 |
+
"learning_rate": 1.5984291299420117e-06,
|
| 588 |
+
"logits/chosen": -0.2737795114517212,
|
| 589 |
+
"logits/rejected": -0.26781201362609863,
|
| 590 |
+
"logps/chosen": -703.7815551757812,
|
| 591 |
+
"logps/rejected": -1142.2947998046875,
|
| 592 |
+
"loss": 0.1225,
|
| 593 |
+
"rewards/accuracies": 0.78125,
|
| 594 |
+
"rewards/chosen": -0.2116294652223587,
|
| 595 |
+
"rewards/margins": 0.18610379099845886,
|
| 596 |
+
"rewards/rejected": -0.39773327112197876,
|
| 597 |
+
"step": 410
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.67,
|
| 601 |
+
"learning_rate": 1.4695585767104092e-06,
|
| 602 |
+
"logits/chosen": -0.29544955492019653,
|
| 603 |
+
"logits/rejected": -0.2773270308971405,
|
| 604 |
+
"logps/chosen": -686.2037353515625,
|
| 605 |
+
"logps/rejected": -1173.82958984375,
|
| 606 |
+
"loss": 0.1319,
|
| 607 |
+
"rewards/accuracies": 0.762499988079071,
|
| 608 |
+
"rewards/chosen": -0.22286900877952576,
|
| 609 |
+
"rewards/margins": 0.183340385556221,
|
| 610 |
+
"rewards/rejected": -0.40620937943458557,
|
| 611 |
+
"step": 420
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.69,
|
| 615 |
+
"learning_rate": 1.3439071446815452e-06,
|
| 616 |
+
"logits/chosen": -0.2415129840373993,
|
| 617 |
+
"logits/rejected": -0.27339568734169006,
|
| 618 |
+
"logps/chosen": -718.8626708984375,
|
| 619 |
+
"logps/rejected": -1240.2945556640625,
|
| 620 |
+
"loss": 0.1158,
|
| 621 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 622 |
+
"rewards/chosen": -0.21683505177497864,
|
| 623 |
+
"rewards/margins": 0.2326025515794754,
|
| 624 |
+
"rewards/rejected": -0.44943755865097046,
|
| 625 |
+
"step": 430
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.7,
|
| 629 |
+
"learning_rate": 1.2218673716356919e-06,
|
| 630 |
+
"logits/chosen": -0.27473658323287964,
|
| 631 |
+
"logits/rejected": -0.2448565512895584,
|
| 632 |
+
"logps/chosen": -734.1240844726562,
|
| 633 |
+
"logps/rejected": -1157.1583251953125,
|
| 634 |
+
"loss": 0.1151,
|
| 635 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 636 |
+
"rewards/chosen": -0.22244882583618164,
|
| 637 |
+
"rewards/margins": 0.20606684684753418,
|
| 638 |
+
"rewards/rejected": -0.4285156726837158,
|
| 639 |
+
"step": 440
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.72,
|
| 643 |
+
"learning_rate": 1.103820512452661e-06,
|
| 644 |
+
"logits/chosen": -0.24782344698905945,
|
| 645 |
+
"logits/rejected": -0.26655644178390503,
|
| 646 |
+
"logps/chosen": -712.7034912109375,
|
| 647 |
+
"logps/rejected": -1142.8284912109375,
|
| 648 |
+
"loss": 0.1273,
|
| 649 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 650 |
+
"rewards/chosen": -0.21107585728168488,
|
| 651 |
+
"rewards/margins": 0.19358885288238525,
|
| 652 |
+
"rewards/rejected": -0.40466469526290894,
|
| 653 |
+
"step": 450
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.74,
|
| 657 |
+
"learning_rate": 9.901353480633468e-07,
|
| 658 |
+
"logits/chosen": -0.2719441056251526,
|
| 659 |
+
"logits/rejected": -0.26369568705558777,
|
| 660 |
+
"logps/chosen": -686.2820434570312,
|
| 661 |
+
"logps/rejected": -1074.0950927734375,
|
| 662 |
+
"loss": 0.1207,
|
| 663 |
+
"rewards/accuracies": 0.8187500238418579,
|
| 664 |
+
"rewards/chosen": -0.21289560198783875,
|
| 665 |
+
"rewards/margins": 0.17246314883232117,
|
| 666 |
+
"rewards/rejected": -0.3853587806224823,
|
| 667 |
+
"step": 460
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.75,
|
| 671 |
+
"learning_rate": 8.811670333701544e-07,
|
| 672 |
+
"logits/chosen": -0.2672122120857239,
|
| 673 |
+
"logits/rejected": -0.2347395420074463,
|
| 674 |
+
"logps/chosen": -752.274169921875,
|
| 675 |
+
"logps/rejected": -1232.9998779296875,
|
| 676 |
+
"loss": 0.1098,
|
| 677 |
+
"rewards/accuracies": 0.831250011920929,
|
| 678 |
+
"rewards/chosen": -0.2272864133119583,
|
| 679 |
+
"rewards/margins": 0.21149499714374542,
|
| 680 |
+
"rewards/rejected": -0.4387814402580261,
|
| 681 |
+
"step": 470
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.77,
|
| 685 |
+
"learning_rate": 7.772559877354341e-07,
|
| 686 |
+
"logits/chosen": -0.24777153134346008,
|
| 687 |
+
"logits/rejected": -0.2399456799030304,
|
| 688 |
+
"logps/chosen": -736.2609252929688,
|
| 689 |
+
"logps/rejected": -1130.094970703125,
|
| 690 |
+
"loss": 0.1168,
|
| 691 |
+
"rewards/accuracies": 0.793749988079071,
|
| 692 |
+
"rewards/chosen": -0.22519740462303162,
|
| 693 |
+
"rewards/margins": 0.17224054038524628,
|
| 694 |
+
"rewards/rejected": -0.3974379599094391,
|
| 695 |
+
"step": 480
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.78,
|
| 699 |
+
"learning_rate": 6.787268315040604e-07,
|
| 700 |
+
"logits/chosen": -0.2684328854084015,
|
| 701 |
+
"logits/rejected": -0.24628722667694092,
|
| 702 |
+
"logps/chosen": -689.9044799804688,
|
| 703 |
+
"logps/rejected": -1155.0946044921875,
|
| 704 |
+
"loss": 0.1174,
|
| 705 |
+
"rewards/accuracies": 0.800000011920929,
|
| 706 |
+
"rewards/chosen": -0.21575360000133514,
|
| 707 |
+
"rewards/margins": 0.1959463357925415,
|
| 708 |
+
"rewards/rejected": -0.41169995069503784,
|
| 709 |
+
"step": 490
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.8,
|
| 713 |
+
"learning_rate": 5.858873718824829e-07,
|
| 714 |
+
"logits/chosen": -0.26447051763534546,
|
| 715 |
+
"logits/rejected": -0.26331770420074463,
|
| 716 |
+
"logps/chosen": -707.0438842773438,
|
| 717 |
+
"logps/rejected": -1152.1182861328125,
|
| 718 |
+
"loss": 0.1172,
|
| 719 |
+
"rewards/accuracies": 0.8187500238418579,
|
| 720 |
+
"rewards/chosen": -0.20341674983501434,
|
| 721 |
+
"rewards/margins": 0.19374600052833557,
|
| 722 |
+
"rewards/rejected": -0.3971627652645111,
|
| 723 |
+
"step": 500
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.82,
|
| 727 |
+
"learning_rate": 4.990276413423817e-07,
|
| 728 |
+
"logits/chosen": -0.23977680504322052,
|
| 729 |
+
"logits/rejected": -0.22561779618263245,
|
| 730 |
+
"logps/chosen": -729.1976318359375,
|
| 731 |
+
"logps/rejected": -1142.5386962890625,
|
| 732 |
+
"loss": 0.1182,
|
| 733 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 734 |
+
"rewards/chosen": -0.21028919517993927,
|
| 735 |
+
"rewards/margins": 0.1970864236354828,
|
| 736 |
+
"rewards/rejected": -0.40737563371658325,
|
| 737 |
+
"step": 510
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.83,
|
| 741 |
+
"learning_rate": 4.184189915529796e-07,
|
| 742 |
+
"logits/chosen": -0.26442182064056396,
|
| 743 |
+
"logits/rejected": -0.25705209374427795,
|
| 744 |
+
"logps/chosen": -737.6463623046875,
|
| 745 |
+
"logps/rejected": -1247.0609130859375,
|
| 746 |
+
"loss": 0.1102,
|
| 747 |
+
"rewards/accuracies": 0.78125,
|
| 748 |
+
"rewards/chosen": -0.2177223414182663,
|
| 749 |
+
"rewards/margins": 0.20199593901634216,
|
| 750 |
+
"rewards/rejected": -0.41971832513809204,
|
| 751 |
+
"step": 520
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.85,
|
| 755 |
+
"learning_rate": 3.4431324567258176e-07,
|
| 756 |
+
"logits/chosen": -0.26106202602386475,
|
| 757 |
+
"logits/rejected": -0.2545424997806549,
|
| 758 |
+
"logps/chosen": -700.1468505859375,
|
| 759 |
+
"logps/rejected": -1189.3382568359375,
|
| 760 |
+
"loss": 0.1211,
|
| 761 |
+
"rewards/accuracies": 0.800000011920929,
|
| 762 |
+
"rewards/chosen": -0.20997758209705353,
|
| 763 |
+
"rewards/margins": 0.2070033997297287,
|
| 764 |
+
"rewards/rejected": -0.4169809818267822,
|
| 765 |
+
"step": 530
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.86,
|
| 769 |
+
"learning_rate": 2.769419116476052e-07,
|
| 770 |
+
"logits/chosen": -0.2785791754722595,
|
| 771 |
+
"logits/rejected": -0.26757797598838806,
|
| 772 |
+
"logps/chosen": -753.193115234375,
|
| 773 |
+
"logps/rejected": -1155.57763671875,
|
| 774 |
+
"loss": 0.1094,
|
| 775 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 776 |
+
"rewards/chosen": -0.21952767670154572,
|
| 777 |
+
"rewards/margins": 0.1966555118560791,
|
| 778 |
+
"rewards/rejected": -0.4161831736564636,
|
| 779 |
+
"step": 540
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.88,
|
| 783 |
+
"learning_rate": 2.1651545897676512e-07,
|
| 784 |
+
"logits/chosen": -0.21319365501403809,
|
| 785 |
+
"logits/rejected": -0.20390382409095764,
|
| 786 |
+
"logps/chosen": -736.0490112304688,
|
| 787 |
+
"logps/rejected": -1217.077880859375,
|
| 788 |
+
"loss": 0.1175,
|
| 789 |
+
"rewards/accuracies": 0.800000011920929,
|
| 790 |
+
"rewards/chosen": -0.21741120517253876,
|
| 791 |
+
"rewards/margins": 0.21254892647266388,
|
| 792 |
+
"rewards/rejected": -0.42996007204055786,
|
| 793 |
+
"step": 550
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.9,
|
| 797 |
+
"learning_rate": 1.6322266119983222e-07,
|
| 798 |
+
"logits/chosen": -0.24104240536689758,
|
| 799 |
+
"logits/rejected": -0.2172488421201706,
|
| 800 |
+
"logps/chosen": -765.8303833007812,
|
| 801 |
+
"logps/rejected": -1205.9599609375,
|
| 802 |
+
"loss": 0.11,
|
| 803 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 804 |
+
"rewards/chosen": -0.21993395686149597,
|
| 805 |
+
"rewards/margins": 0.19853533804416656,
|
| 806 |
+
"rewards/rejected": -0.41846928000450134,
|
| 807 |
+
"step": 560
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.91,
|
| 811 |
+
"learning_rate": 1.1723000616502167e-07,
|
| 812 |
+
"logits/chosen": -0.26349014043807983,
|
| 813 |
+
"logits/rejected": -0.3146611452102661,
|
| 814 |
+
"logps/chosen": -755.0218505859375,
|
| 815 |
+
"logps/rejected": -1200.8770751953125,
|
| 816 |
+
"loss": 0.1129,
|
| 817 |
+
"rewards/accuracies": 0.8125,
|
| 818 |
+
"rewards/chosen": -0.21991780400276184,
|
| 819 |
+
"rewards/margins": 0.22101625800132751,
|
| 820 |
+
"rewards/rejected": -0.44093409180641174,
|
| 821 |
+
"step": 570
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.93,
|
| 825 |
+
"learning_rate": 7.868117591737585e-08,
|
| 826 |
+
"logits/chosen": -0.20703573524951935,
|
| 827 |
+
"logits/rejected": -0.21116182208061218,
|
| 828 |
+
"logps/chosen": -741.433837890625,
|
| 829 |
+
"logps/rejected": -1168.3992919921875,
|
| 830 |
+
"loss": 0.1149,
|
| 831 |
+
"rewards/accuracies": 0.8187500238418579,
|
| 832 |
+
"rewards/chosen": -0.2117561548948288,
|
| 833 |
+
"rewards/margins": 0.19863612949848175,
|
| 834 |
+
"rewards/rejected": -0.41039222478866577,
|
| 835 |
+
"step": 580
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.94,
|
| 839 |
+
"learning_rate": 4.769659783295383e-08,
|
| 840 |
+
"logits/chosen": -0.23616275191307068,
|
| 841 |
+
"logits/rejected": -0.2213321179151535,
|
| 842 |
+
"logps/chosen": -802.316650390625,
|
| 843 |
+
"logps/rejected": -1224.220458984375,
|
| 844 |
+
"loss": 0.1123,
|
| 845 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 846 |
+
"rewards/chosen": -0.2234838753938675,
|
| 847 |
+
"rewards/margins": 0.21354734897613525,
|
| 848 |
+
"rewards/rejected": -0.43703120946884155,
|
| 849 |
+
"step": 590
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.96,
|
| 853 |
+
"learning_rate": 2.4373068401120358e-08,
|
| 854 |
+
"logits/chosen": -0.25830012559890747,
|
| 855 |
+
"logits/rejected": -0.24621865153312683,
|
| 856 |
+
"logps/chosen": -802.4405517578125,
|
| 857 |
+
"logps/rejected": -1169.389892578125,
|
| 858 |
+
"loss": 0.1136,
|
| 859 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 860 |
+
"rewards/chosen": -0.23679308593273163,
|
| 861 |
+
"rewards/margins": 0.1784859597682953,
|
| 862 |
+
"rewards/rejected": -0.4152790606021881,
|
| 863 |
+
"step": 600
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.98,
|
| 867 |
+
"learning_rate": 8.78345083022425e-09,
|
| 868 |
+
"logits/chosen": -0.24250511825084686,
|
| 869 |
+
"logits/rejected": -0.24350687861442566,
|
| 870 |
+
"logps/chosen": -742.060546875,
|
| 871 |
+
"logps/rejected": -1132.1630859375,
|
| 872 |
+
"loss": 0.1141,
|
| 873 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 874 |
+
"rewards/chosen": -0.2040189504623413,
|
| 875 |
+
"rewards/margins": 0.19322746992111206,
|
| 876 |
+
"rewards/rejected": -0.39724642038345337,
|
| 877 |
+
"step": 610
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.99,
|
| 881 |
+
"learning_rate": 9.764474213677654e-10,
|
| 882 |
+
"logits/chosen": -0.2109394073486328,
|
| 883 |
+
"logits/rejected": -0.22563035786151886,
|
| 884 |
+
"logps/chosen": -744.6256103515625,
|
| 885 |
+
"logps/rejected": -1185.656982421875,
|
| 886 |
+
"loss": 0.1153,
|
| 887 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 888 |
+
"rewards/chosen": -0.23164710402488708,
|
| 889 |
+
"rewards/margins": 0.20466098189353943,
|
| 890 |
+
"rewards/rejected": -0.4363080859184265,
|
| 891 |
+
"step": 620
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 1.0,
|
| 895 |
+
"step": 625,
|
| 896 |
+
"total_flos": 0.0,
|
| 897 |
+
"train_loss": 0.13047290132045747,
|
| 898 |
+
"train_runtime": 8392.5195,
|
| 899 |
+
"train_samples_per_second": 3.575,
|
| 900 |
+
"train_steps_per_second": 0.074
|
| 901 |
+
}
|
| 902 |
+
],
|
| 903 |
+
"logging_steps": 10,
|
| 904 |
+
"max_steps": 625,
|
| 905 |
+
"num_input_tokens_seen": 0,
|
| 906 |
+
"num_train_epochs": 1,
|
| 907 |
+
"save_steps": 100,
|
| 908 |
+
"total_flos": 0.0,
|
| 909 |
+
"train_batch_size": 4,
|
| 910 |
+
"trial_name": null,
|
| 911 |
+
"trial_params": null
|
| 912 |
+
}
|