Model save
Browse files- README.md +61 -0
- adapter_model.safetensors +1 -1
- all_results.json +8 -0
- runs/May21_09-26-03_gpu4-119-5/events.out.tfevents.1716247699.gpu4-119-5.1546019.0 +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/phi-2-irepo-chatml-merged-i0
|
| 8 |
+
model-index:
|
| 9 |
+
- name: phi-2-irepo-chatml-v15-i1
|
| 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-irepo-chatml-v15-i1
|
| 17 |
+
|
| 18 |
+
This model is a fine-tuned version of [DUAL-GPO/phi-2-irepo-chatml-merged-i0](https://huggingface.co/DUAL-GPO/phi-2-irepo-chatml-merged-i0) 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: 2
|
| 43 |
+
- gradient_accumulation_steps: 4
|
| 44 |
+
- total_train_batch_size: 32
|
| 45 |
+
- total_eval_batch_size: 8
|
| 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 335579632
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:48ff999f26ed0d5380ee57de2cb3840feb00a7767fb17e7d1eff6b8d56ca19bd
|
| 3 |
size 335579632
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.1911162176847458,
|
| 4 |
+
"train_runtime": 4612.7069,
|
| 5 |
+
"train_samples": 20000,
|
| 6 |
+
"train_samples_per_second": 4.336,
|
| 7 |
+
"train_steps_per_second": 0.135
|
| 8 |
+
}
|
runs/May21_09-26-03_gpu4-119-5/events.out.tfevents.1716247699.gpu4-119-5.1546019.0
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:eb66dafb5beca7efbc2a30288bf25c2472c100b01b33de2328963c8e290d7963
|
| 3 |
+
size 45066
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.1911162176847458,
|
| 4 |
+
"train_runtime": 4612.7069,
|
| 5 |
+
"train_samples": 20000,
|
| 6 |
+
"train_samples_per_second": 4.336,
|
| 7 |
+
"train_steps_per_second": 0.135
|
| 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": 0.0711287260055542,
|
| 15 |
+
"logits/rejected": 0.20400863885879517,
|
| 16 |
+
"logps/chosen": -313.75396728515625,
|
| 17 |
+
"logps/rejected": -420.73980712890625,
|
| 18 |
+
"loss": 0.2285,
|
| 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": 0.2657856047153473,
|
| 29 |
+
"logits/rejected": 0.23175562918186188,
|
| 30 |
+
"logps/chosen": -354.52484130859375,
|
| 31 |
+
"logps/rejected": -365.6746520996094,
|
| 32 |
+
"loss": 0.203,
|
| 33 |
+
"rewards/accuracies": 0.3125,
|
| 34 |
+
"rewards/chosen": -0.0005699184257537127,
|
| 35 |
+
"rewards/margins": 2.8342270525172353e-05,
|
| 36 |
+
"rewards/rejected": -0.0005982607253827155,
|
| 37 |
+
"step": 10
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.03,
|
| 41 |
+
"learning_rate": 1.5873015873015873e-06,
|
| 42 |
+
"logits/chosen": 0.16931554675102234,
|
| 43 |
+
"logits/rejected": 0.265936940908432,
|
| 44 |
+
"logps/chosen": -376.02166748046875,
|
| 45 |
+
"logps/rejected": -396.56524658203125,
|
| 46 |
+
"loss": 0.2032,
|
| 47 |
+
"rewards/accuracies": 0.3062500059604645,
|
| 48 |
+
"rewards/chosen": -0.0009135094587691128,
|
| 49 |
+
"rewards/margins": -1.457883081457112e-05,
|
| 50 |
+
"rewards/rejected": -0.000898930593393743,
|
| 51 |
+
"step": 20
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.05,
|
| 55 |
+
"learning_rate": 2.380952380952381e-06,
|
| 56 |
+
"logits/chosen": 0.1378197968006134,
|
| 57 |
+
"logits/rejected": 0.2801808714866638,
|
| 58 |
+
"logps/chosen": -394.52301025390625,
|
| 59 |
+
"logps/rejected": -406.01971435546875,
|
| 60 |
+
"loss": 0.2045,
|
| 61 |
+
"rewards/accuracies": 0.3499999940395355,
|
| 62 |
+
"rewards/chosen": -0.0023785685189068317,
|
| 63 |
+
"rewards/margins": -4.536235792329535e-05,
|
| 64 |
+
"rewards/rejected": -0.0023332058917731047,
|
| 65 |
+
"step": 30
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.06,
|
| 69 |
+
"learning_rate": 3.1746031746031746e-06,
|
| 70 |
+
"logits/chosen": 0.22031190991401672,
|
| 71 |
+
"logits/rejected": 0.2555253505706787,
|
| 72 |
+
"logps/chosen": -413.9378356933594,
|
| 73 |
+
"logps/rejected": -384.2980041503906,
|
| 74 |
+
"loss": 0.2078,
|
| 75 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 76 |
+
"rewards/chosen": -0.0029828939586877823,
|
| 77 |
+
"rewards/margins": 0.0004020760825369507,
|
| 78 |
+
"rewards/rejected": -0.0033849701285362244,
|
| 79 |
+
"step": 40
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.08,
|
| 83 |
+
"learning_rate": 3.968253968253968e-06,
|
| 84 |
+
"logits/chosen": 0.17930440604686737,
|
| 85 |
+
"logits/rejected": 0.27743226289749146,
|
| 86 |
+
"logps/chosen": -437.57220458984375,
|
| 87 |
+
"logps/rejected": -428.25726318359375,
|
| 88 |
+
"loss": 0.2109,
|
| 89 |
+
"rewards/accuracies": 0.4937500059604645,
|
| 90 |
+
"rewards/chosen": -0.0068945749662816525,
|
| 91 |
+
"rewards/margins": 0.0003428836935199797,
|
| 92 |
+
"rewards/rejected": -0.0072374590672552586,
|
| 93 |
+
"step": 50
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.1,
|
| 97 |
+
"learning_rate": 4.761904761904762e-06,
|
| 98 |
+
"logits/chosen": 0.1960189789533615,
|
| 99 |
+
"logits/rejected": 0.20567241311073303,
|
| 100 |
+
"logps/chosen": -412.0870666503906,
|
| 101 |
+
"logps/rejected": -429.5210876464844,
|
| 102 |
+
"loss": 0.2106,
|
| 103 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 104 |
+
"rewards/chosen": -0.01030620839446783,
|
| 105 |
+
"rewards/margins": 0.0008126860484480858,
|
| 106 |
+
"rewards/rejected": -0.011118894442915916,
|
| 107 |
+
"step": 60
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.11,
|
| 111 |
+
"learning_rate": 4.998086282661188e-06,
|
| 112 |
+
"logits/chosen": 0.18384099006652832,
|
| 113 |
+
"logits/rejected": 0.16322512924671173,
|
| 114 |
+
"logps/chosen": -480.0482482910156,
|
| 115 |
+
"logps/rejected": -449.12872314453125,
|
| 116 |
+
"loss": 0.2125,
|
| 117 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 118 |
+
"rewards/chosen": -0.012364747002720833,
|
| 119 |
+
"rewards/margins": 0.0012391259660944343,
|
| 120 |
+
"rewards/rejected": -0.013603871688246727,
|
| 121 |
+
"step": 70
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.13,
|
| 125 |
+
"learning_rate": 4.988720025682995e-06,
|
| 126 |
+
"logits/chosen": 0.17162439227104187,
|
| 127 |
+
"logits/rejected": 0.3024311661720276,
|
| 128 |
+
"logps/chosen": -425.10821533203125,
|
| 129 |
+
"logps/rejected": -405.69903564453125,
|
| 130 |
+
"loss": 0.203,
|
| 131 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 132 |
+
"rewards/chosen": -0.002422564197331667,
|
| 133 |
+
"rewards/margins": 0.0019961665384471416,
|
| 134 |
+
"rewards/rejected": -0.004418730735778809,
|
| 135 |
+
"step": 80
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.14,
|
| 139 |
+
"learning_rate": 4.9715789537359126e-06,
|
| 140 |
+
"logits/chosen": 0.14976395666599274,
|
| 141 |
+
"logits/rejected": 0.293326199054718,
|
| 142 |
+
"logps/chosen": -450.09283447265625,
|
| 143 |
+
"logps/rejected": -463.3335876464844,
|
| 144 |
+
"loss": 0.1928,
|
| 145 |
+
"rewards/accuracies": 0.550000011920929,
|
| 146 |
+
"rewards/chosen": 0.004968429449945688,
|
| 147 |
+
"rewards/margins": 0.004259931854903698,
|
| 148 |
+
"rewards/rejected": 0.0007084974786266685,
|
| 149 |
+
"step": 90
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.16,
|
| 153 |
+
"learning_rate": 4.946716615897932e-06,
|
| 154 |
+
"logits/chosen": 0.14262747764587402,
|
| 155 |
+
"logits/rejected": 0.14231689274311066,
|
| 156 |
+
"logps/chosen": -409.68719482421875,
|
| 157 |
+
"logps/rejected": -385.9997253417969,
|
| 158 |
+
"loss": 0.2044,
|
| 159 |
+
"rewards/accuracies": 0.42500001192092896,
|
| 160 |
+
"rewards/chosen": 0.0011129004415124655,
|
| 161 |
+
"rewards/margins": 0.0027904310263693333,
|
| 162 |
+
"rewards/rejected": -0.001677530468441546,
|
| 163 |
+
"step": 100
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.18,
|
| 167 |
+
"learning_rate": 4.9142106826480114e-06,
|
| 168 |
+
"logits/chosen": 0.09172149002552032,
|
| 169 |
+
"logits/rejected": 0.18723782896995544,
|
| 170 |
+
"logps/chosen": -376.23382568359375,
|
| 171 |
+
"logps/rejected": -382.14312744140625,
|
| 172 |
+
"loss": 0.2117,
|
| 173 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 174 |
+
"rewards/chosen": 0.0011799463536590338,
|
| 175 |
+
"rewards/margins": 0.006289638578891754,
|
| 176 |
+
"rewards/rejected": -0.005109691992402077,
|
| 177 |
+
"step": 110
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.19,
|
| 181 |
+
"learning_rate": 4.874162703221823e-06,
|
| 182 |
+
"logits/chosen": 0.21056421101093292,
|
| 183 |
+
"logits/rejected": 0.10638086497783661,
|
| 184 |
+
"logps/chosen": -455.8184509277344,
|
| 185 |
+
"logps/rejected": -461.58575439453125,
|
| 186 |
+
"loss": 0.2024,
|
| 187 |
+
"rewards/accuracies": 0.518750011920929,
|
| 188 |
+
"rewards/chosen": 0.0018554453272372484,
|
| 189 |
+
"rewards/margins": 0.007339824922382832,
|
| 190 |
+
"rewards/rejected": -0.005484379827976227,
|
| 191 |
+
"step": 120
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.21,
|
| 195 |
+
"learning_rate": 4.826697788369752e-06,
|
| 196 |
+
"logits/chosen": 0.0843835398554802,
|
| 197 |
+
"logits/rejected": 0.10346312820911407,
|
| 198 |
+
"logps/chosen": -420.7416076660156,
|
| 199 |
+
"logps/rejected": -426.8565979003906,
|
| 200 |
+
"loss": 0.1938,
|
| 201 |
+
"rewards/accuracies": 0.518750011920929,
|
| 202 |
+
"rewards/chosen": 0.0005354422028176486,
|
| 203 |
+
"rewards/margins": 0.012502538040280342,
|
| 204 |
+
"rewards/rejected": -0.011967095546424389,
|
| 205 |
+
"step": 130
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.22,
|
| 209 |
+
"learning_rate": 4.7719642195082224e-06,
|
| 210 |
+
"logits/chosen": 0.03641371801495552,
|
| 211 |
+
"logits/rejected": 0.10479255020618439,
|
| 212 |
+
"logps/chosen": -401.6028137207031,
|
| 213 |
+
"logps/rejected": -464.2025451660156,
|
| 214 |
+
"loss": 0.1958,
|
| 215 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 216 |
+
"rewards/chosen": -0.024544300511479378,
|
| 217 |
+
"rewards/margins": 0.018662814050912857,
|
| 218 |
+
"rewards/rejected": -0.043207116425037384,
|
| 219 |
+
"step": 140
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.24,
|
| 223 |
+
"learning_rate": 4.710132985485355e-06,
|
| 224 |
+
"logits/chosen": 0.11580105125904083,
|
| 225 |
+
"logits/rejected": 0.021813513711094856,
|
| 226 |
+
"logps/chosen": -430.5687561035156,
|
| 227 |
+
"logps/rejected": -515.4051513671875,
|
| 228 |
+
"loss": 0.1976,
|
| 229 |
+
"rewards/accuracies": 0.5687500238418579,
|
| 230 |
+
"rewards/chosen": -0.04684633016586304,
|
| 231 |
+
"rewards/margins": 0.043844569474458694,
|
| 232 |
+
"rewards/rejected": -0.09069089591503143,
|
| 233 |
+
"step": 150
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.26,
|
| 237 |
+
"learning_rate": 4.641397248408122e-06,
|
| 238 |
+
"logits/chosen": -0.0415755994617939,
|
| 239 |
+
"logits/rejected": 0.07411660254001617,
|
| 240 |
+
"logps/chosen": -467.6756286621094,
|
| 241 |
+
"logps/rejected": -472.20001220703125,
|
| 242 |
+
"loss": 0.1972,
|
| 243 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 244 |
+
"rewards/chosen": -0.0470997616648674,
|
| 245 |
+
"rewards/margins": 0.02540501020848751,
|
| 246 |
+
"rewards/rejected": -0.07250477373600006,
|
| 247 |
+
"step": 160
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.27,
|
| 251 |
+
"learning_rate": 4.5659717401997655e-06,
|
| 252 |
+
"logits/chosen": -0.03600798547267914,
|
| 253 |
+
"logits/rejected": 0.027341466397047043,
|
| 254 |
+
"logps/chosen": -504.47894287109375,
|
| 255 |
+
"logps/rejected": -516.9270629882812,
|
| 256 |
+
"loss": 0.1948,
|
| 257 |
+
"rewards/accuracies": 0.5062500238418579,
|
| 258 |
+
"rewards/chosen": -0.05515031889081001,
|
| 259 |
+
"rewards/margins": 0.03458093851804733,
|
| 260 |
+
"rewards/rejected": -0.08973126113414764,
|
| 261 |
+
"step": 170
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.29,
|
| 265 |
+
"learning_rate": 4.4840920917726425e-06,
|
| 266 |
+
"logits/chosen": -0.039528343826532364,
|
| 267 |
+
"logits/rejected": -0.019377198070287704,
|
| 268 |
+
"logps/chosen": -484.7760314941406,
|
| 269 |
+
"logps/rejected": -515.7113037109375,
|
| 270 |
+
"loss": 0.1955,
|
| 271 |
+
"rewards/accuracies": 0.4937500059604645,
|
| 272 |
+
"rewards/chosen": -0.05891401320695877,
|
| 273 |
+
"rewards/margins": 0.03348463773727417,
|
| 274 |
+
"rewards/rejected": -0.09239865839481354,
|
| 275 |
+
"step": 180
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.3,
|
| 279 |
+
"learning_rate": 4.396014096912182e-06,
|
| 280 |
+
"logits/chosen": 0.041537586599588394,
|
| 281 |
+
"logits/rejected": -0.029888898134231567,
|
| 282 |
+
"logps/chosen": -474.3414001464844,
|
| 283 |
+
"logps/rejected": -515.9990234375,
|
| 284 |
+
"loss": 0.1919,
|
| 285 |
+
"rewards/accuracies": 0.512499988079071,
|
| 286 |
+
"rewards/chosen": -0.07046758383512497,
|
| 287 |
+
"rewards/margins": 0.03831402584910393,
|
| 288 |
+
"rewards/rejected": -0.1087816134095192,
|
| 289 |
+
"step": 190
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.32,
|
| 293 |
+
"learning_rate": 4.302012913171584e-06,
|
| 294 |
+
"logits/chosen": -0.017491130158305168,
|
| 295 |
+
"logits/rejected": -0.04960538074374199,
|
| 296 |
+
"logps/chosen": -484.7154846191406,
|
| 297 |
+
"logps/rejected": -524.5064086914062,
|
| 298 |
+
"loss": 0.1935,
|
| 299 |
+
"rewards/accuracies": 0.4937500059604645,
|
| 300 |
+
"rewards/chosen": -0.07989688217639923,
|
| 301 |
+
"rewards/margins": 0.04493844509124756,
|
| 302 |
+
"rewards/rejected": -0.12483533471822739,
|
| 303 |
+
"step": 200
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.34,
|
| 307 |
+
"learning_rate": 4.202382202273702e-06,
|
| 308 |
+
"logits/chosen": -0.13748802244663239,
|
| 309 |
+
"logits/rejected": -0.001641835318878293,
|
| 310 |
+
"logps/chosen": -507.37506103515625,
|
| 311 |
+
"logps/rejected": -535.3721923828125,
|
| 312 |
+
"loss": 0.1804,
|
| 313 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 314 |
+
"rewards/chosen": -0.07145627588033676,
|
| 315 |
+
"rewards/margins": 0.051401056349277496,
|
| 316 |
+
"rewards/rejected": -0.12285731732845306,
|
| 317 |
+
"step": 210
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.35,
|
| 321 |
+
"learning_rate": 4.097433212705492e-06,
|
| 322 |
+
"logits/chosen": -0.021799543872475624,
|
| 323 |
+
"logits/rejected": -0.074518583714962,
|
| 324 |
+
"logps/chosen": -477.99981689453125,
|
| 325 |
+
"logps/rejected": -518.2962646484375,
|
| 326 |
+
"loss": 0.2019,
|
| 327 |
+
"rewards/accuracies": 0.4437499940395355,
|
| 328 |
+
"rewards/chosen": -0.09213604032993317,
|
| 329 |
+
"rewards/margins": 0.04488346725702286,
|
| 330 |
+
"rewards/rejected": -0.13701951503753662,
|
| 331 |
+
"step": 220
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.37,
|
| 335 |
+
"learning_rate": 3.987493807371033e-06,
|
| 336 |
+
"logits/chosen": -0.16067767143249512,
|
| 337 |
+
"logits/rejected": 0.01047535240650177,
|
| 338 |
+
"logps/chosen": -489.7818298339844,
|
| 339 |
+
"logps/rejected": -525.8330688476562,
|
| 340 |
+
"loss": 0.1852,
|
| 341 |
+
"rewards/accuracies": 0.5562499761581421,
|
| 342 |
+
"rewards/chosen": -0.0610770583152771,
|
| 343 |
+
"rewards/margins": 0.04805826395750046,
|
| 344 |
+
"rewards/rejected": -0.10913531482219696,
|
| 345 |
+
"step": 230
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.38,
|
| 349 |
+
"learning_rate": 3.872907439340758e-06,
|
| 350 |
+
"logits/chosen": -0.035437412559986115,
|
| 351 |
+
"logits/rejected": -0.07582642138004303,
|
| 352 |
+
"logps/chosen": -440.7459411621094,
|
| 353 |
+
"logps/rejected": -492.434326171875,
|
| 354 |
+
"loss": 0.177,
|
| 355 |
+
"rewards/accuracies": 0.46875,
|
| 356 |
+
"rewards/chosen": -0.04167747497558594,
|
| 357 |
+
"rewards/margins": 0.04895929619669914,
|
| 358 |
+
"rewards/rejected": -0.09063677489757538,
|
| 359 |
+
"step": 240
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.4,
|
| 363 |
+
"learning_rate": 3.75403207889666e-06,
|
| 364 |
+
"logits/chosen": -0.09768973290920258,
|
| 365 |
+
"logits/rejected": -0.0913015678524971,
|
| 366 |
+
"logps/chosen": -445.75750732421875,
|
| 367 |
+
"logps/rejected": -473.06671142578125,
|
| 368 |
+
"loss": 0.1765,
|
| 369 |
+
"rewards/accuracies": 0.45625001192092896,
|
| 370 |
+
"rewards/chosen": -0.04659276083111763,
|
| 371 |
+
"rewards/margins": 0.03554708510637283,
|
| 372 |
+
"rewards/rejected": -0.08213984221220016,
|
| 373 |
+
"step": 250
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.42,
|
| 377 |
+
"learning_rate": 3.631239095225417e-06,
|
| 378 |
+
"logits/chosen": -0.002584155648946762,
|
| 379 |
+
"logits/rejected": -0.0503731295466423,
|
| 380 |
+
"logps/chosen": -443.8072204589844,
|
| 381 |
+
"logps/rejected": -522.3037109375,
|
| 382 |
+
"loss": 0.1903,
|
| 383 |
+
"rewards/accuracies": 0.5,
|
| 384 |
+
"rewards/chosen": -0.057579755783081055,
|
| 385 |
+
"rewards/margins": 0.042577676475048065,
|
| 386 |
+
"rewards/rejected": -0.10015741735696793,
|
| 387 |
+
"step": 260
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.43,
|
| 391 |
+
"learning_rate": 3.5049120962530608e-06,
|
| 392 |
+
"logits/chosen": -0.07919616252183914,
|
| 393 |
+
"logits/rejected": -0.051559675484895706,
|
| 394 |
+
"logps/chosen": -442.90020751953125,
|
| 395 |
+
"logps/rejected": -506.71026611328125,
|
| 396 |
+
"loss": 0.1801,
|
| 397 |
+
"rewards/accuracies": 0.48124998807907104,
|
| 398 |
+
"rewards/chosen": -0.04894575849175453,
|
| 399 |
+
"rewards/margins": 0.060835689306259155,
|
| 400 |
+
"rewards/rejected": -0.1097814291715622,
|
| 401 |
+
"step": 270
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.45,
|
| 405 |
+
"learning_rate": 3.3754457302455464e-06,
|
| 406 |
+
"logits/chosen": -0.05536097288131714,
|
| 407 |
+
"logits/rejected": -0.05594850331544876,
|
| 408 |
+
"logps/chosen": -491.963623046875,
|
| 409 |
+
"logps/rejected": -476.1817932128906,
|
| 410 |
+
"loss": 0.1987,
|
| 411 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 412 |
+
"rewards/chosen": -0.05905281379818916,
|
| 413 |
+
"rewards/margins": 0.03382394462823868,
|
| 414 |
+
"rewards/rejected": -0.09287675470113754,
|
| 415 |
+
"step": 280
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.46,
|
| 419 |
+
"learning_rate": 3.2432444529190714e-06,
|
| 420 |
+
"logits/chosen": -0.06521332263946533,
|
| 421 |
+
"logits/rejected": 0.04030764847993851,
|
| 422 |
+
"logps/chosen": -503.37115478515625,
|
| 423 |
+
"logps/rejected": -539.033935546875,
|
| 424 |
+
"loss": 0.1758,
|
| 425 |
+
"rewards/accuracies": 0.48124998807907104,
|
| 426 |
+
"rewards/chosen": -0.06568136066198349,
|
| 427 |
+
"rewards/margins": 0.04754118248820305,
|
| 428 |
+
"rewards/rejected": -0.11322255432605743,
|
| 429 |
+
"step": 290
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.48,
|
| 433 |
+
"learning_rate": 3.1087212639117057e-06,
|
| 434 |
+
"logits/chosen": -0.007789143826812506,
|
| 435 |
+
"logits/rejected": -0.08723556995391846,
|
| 436 |
+
"logps/chosen": -448.15185546875,
|
| 437 |
+
"logps/rejected": -472.66046142578125,
|
| 438 |
+
"loss": 0.195,
|
| 439 |
+
"rewards/accuracies": 0.4937500059604645,
|
| 440 |
+
"rewards/chosen": -0.04991251975297928,
|
| 441 |
+
"rewards/margins": 0.03361000493168831,
|
| 442 |
+
"rewards/rejected": -0.08352252840995789,
|
| 443 |
+
"step": 300
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.5,
|
| 447 |
+
"learning_rate": 2.9722964165636263e-06,
|
| 448 |
+
"logits/chosen": -0.10513603687286377,
|
| 449 |
+
"logits/rejected": -0.02427390217781067,
|
| 450 |
+
"logps/chosen": -456.4302673339844,
|
| 451 |
+
"logps/rejected": -480.01904296875,
|
| 452 |
+
"loss": 0.1777,
|
| 453 |
+
"rewards/accuracies": 0.46875,
|
| 454 |
+
"rewards/chosen": -0.04909784346818924,
|
| 455 |
+
"rewards/margins": 0.032266296446323395,
|
| 456 |
+
"rewards/rejected": -0.08136413991451263,
|
| 457 |
+
"step": 310
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.51,
|
| 461 |
+
"learning_rate": 2.8343961050366275e-06,
|
| 462 |
+
"logits/chosen": -0.06911730021238327,
|
| 463 |
+
"logits/rejected": -0.02331097424030304,
|
| 464 |
+
"logps/chosen": -412.3163146972656,
|
| 465 |
+
"logps/rejected": -419.2935485839844,
|
| 466 |
+
"loss": 0.1807,
|
| 467 |
+
"rewards/accuracies": 0.45625001192092896,
|
| 468 |
+
"rewards/chosen": -0.031493835151195526,
|
| 469 |
+
"rewards/margins": 0.03578699007630348,
|
| 470 |
+
"rewards/rejected": -0.06728082150220871,
|
| 471 |
+
"step": 320
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.53,
|
| 475 |
+
"learning_rate": 2.695451132874385e-06,
|
| 476 |
+
"logits/chosen": 0.030731942504644394,
|
| 477 |
+
"logits/rejected": -0.028069671243429184,
|
| 478 |
+
"logps/chosen": -477.44512939453125,
|
| 479 |
+
"logps/rejected": -501.154541015625,
|
| 480 |
+
"loss": 0.1925,
|
| 481 |
+
"rewards/accuracies": 0.518750011920929,
|
| 482 |
+
"rewards/chosen": -0.045379411429166794,
|
| 483 |
+
"rewards/margins": 0.04541900008916855,
|
| 484 |
+
"rewards/rejected": -0.09079841524362564,
|
| 485 |
+
"step": 330
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.54,
|
| 489 |
+
"learning_rate": 2.5558955671628964e-06,
|
| 490 |
+
"logits/chosen": -0.0826629176735878,
|
| 491 |
+
"logits/rejected": -0.04008691757917404,
|
| 492 |
+
"logps/chosen": -429.1664123535156,
|
| 493 |
+
"logps/rejected": -467.46173095703125,
|
| 494 |
+
"loss": 0.1838,
|
| 495 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 496 |
+
"rewards/chosen": -0.05404701828956604,
|
| 497 |
+
"rewards/margins": 0.04224228113889694,
|
| 498 |
+
"rewards/rejected": -0.09628931432962418,
|
| 499 |
+
"step": 340
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.56,
|
| 503 |
+
"learning_rate": 2.4161653824955654e-06,
|
| 504 |
+
"logits/chosen": -0.056598931550979614,
|
| 505 |
+
"logits/rejected": -0.10974061489105225,
|
| 506 |
+
"logps/chosen": -513.0817260742188,
|
| 507 |
+
"logps/rejected": -546.1875,
|
| 508 |
+
"loss": 0.1864,
|
| 509 |
+
"rewards/accuracies": 0.5,
|
| 510 |
+
"rewards/chosen": -0.06718280166387558,
|
| 511 |
+
"rewards/margins": 0.05270420387387276,
|
| 512 |
+
"rewards/rejected": -0.11988700926303864,
|
| 513 |
+
"step": 350
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.58,
|
| 517 |
+
"learning_rate": 2.2766970989791697e-06,
|
| 518 |
+
"logits/chosen": -0.04662217199802399,
|
| 519 |
+
"logits/rejected": -0.05574822425842285,
|
| 520 |
+
"logps/chosen": -480.88848876953125,
|
| 521 |
+
"logps/rejected": -530.1246337890625,
|
| 522 |
+
"loss": 0.1915,
|
| 523 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 524 |
+
"rewards/chosen": -0.06772033870220184,
|
| 525 |
+
"rewards/margins": 0.03862085938453674,
|
| 526 |
+
"rewards/rejected": -0.10634119808673859,
|
| 527 |
+
"step": 360
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.59,
|
| 531 |
+
"learning_rate": 2.1379264185356545e-06,
|
| 532 |
+
"logits/chosen": -0.020046677440404892,
|
| 533 |
+
"logits/rejected": 0.09056108444929123,
|
| 534 |
+
"logps/chosen": -486.93115234375,
|
| 535 |
+
"logps/rejected": -528.6192626953125,
|
| 536 |
+
"loss": 0.1814,
|
| 537 |
+
"rewards/accuracies": 0.4937500059604645,
|
| 538 |
+
"rewards/chosen": -0.054186802357435226,
|
| 539 |
+
"rewards/margins": 0.05737306550145149,
|
| 540 |
+
"rewards/rejected": -0.11155986785888672,
|
| 541 |
+
"step": 370
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.61,
|
| 545 |
+
"learning_rate": 2.000286863759934e-06,
|
| 546 |
+
"logits/chosen": -0.09826477617025375,
|
| 547 |
+
"logits/rejected": -0.07946722954511642,
|
| 548 |
+
"logps/chosen": -470.00811767578125,
|
| 549 |
+
"logps/rejected": -511.309814453125,
|
| 550 |
+
"loss": 0.1828,
|
| 551 |
+
"rewards/accuracies": 0.512499988079071,
|
| 552 |
+
"rewards/chosen": -0.06328146904706955,
|
| 553 |
+
"rewards/margins": 0.05404907464981079,
|
| 554 |
+
"rewards/rejected": -0.11733055114746094,
|
| 555 |
+
"step": 380
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.62,
|
| 559 |
+
"learning_rate": 1.8642084235859764e-06,
|
| 560 |
+
"logits/chosen": -0.15358105301856995,
|
| 561 |
+
"logits/rejected": -0.07723913341760635,
|
| 562 |
+
"logps/chosen": -438.9100646972656,
|
| 563 |
+
"logps/rejected": -479.63092041015625,
|
| 564 |
+
"loss": 0.1972,
|
| 565 |
+
"rewards/accuracies": 0.46875,
|
| 566 |
+
"rewards/chosen": -0.05276118963956833,
|
| 567 |
+
"rewards/margins": 0.04282676801085472,
|
| 568 |
+
"rewards/rejected": -0.09558796137571335,
|
| 569 |
+
"step": 390
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.64,
|
| 573 |
+
"learning_rate": 1.7301162099921013e-06,
|
| 574 |
+
"logits/chosen": -0.06951303780078888,
|
| 575 |
+
"logits/rejected": -0.16843798756599426,
|
| 576 |
+
"logps/chosen": -422.1139221191406,
|
| 577 |
+
"logps/rejected": -459.0763244628906,
|
| 578 |
+
"loss": 0.1995,
|
| 579 |
+
"rewards/accuracies": 0.45625001192092896,
|
| 580 |
+
"rewards/chosen": -0.05717768147587776,
|
| 581 |
+
"rewards/margins": 0.040271807461977005,
|
| 582 |
+
"rewards/rejected": -0.09744948893785477,
|
| 583 |
+
"step": 400
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.66,
|
| 587 |
+
"learning_rate": 1.5984291299420117e-06,
|
| 588 |
+
"logits/chosen": -0.1256350576877594,
|
| 589 |
+
"logits/rejected": -0.08233270049095154,
|
| 590 |
+
"logps/chosen": -404.26263427734375,
|
| 591 |
+
"logps/rejected": -455.4290466308594,
|
| 592 |
+
"loss": 0.1892,
|
| 593 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 594 |
+
"rewards/chosen": -0.04233751818537712,
|
| 595 |
+
"rewards/margins": 0.05102572590112686,
|
| 596 |
+
"rewards/rejected": -0.09336324036121368,
|
| 597 |
+
"step": 410
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.67,
|
| 601 |
+
"learning_rate": 1.4695585767104092e-06,
|
| 602 |
+
"logits/chosen": -0.0601225309073925,
|
| 603 |
+
"logits/rejected": -0.12005972862243652,
|
| 604 |
+
"logps/chosen": -482.4237365722656,
|
| 605 |
+
"logps/rejected": -531.4822998046875,
|
| 606 |
+
"loss": 0.1715,
|
| 607 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 608 |
+
"rewards/chosen": -0.045725829899311066,
|
| 609 |
+
"rewards/margins": 0.04561304301023483,
|
| 610 |
+
"rewards/rejected": -0.0913388729095459,
|
| 611 |
+
"step": 420
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.69,
|
| 615 |
+
"learning_rate": 1.3439071446815452e-06,
|
| 616 |
+
"logits/chosen": -0.15923841297626495,
|
| 617 |
+
"logits/rejected": 0.011970462277531624,
|
| 618 |
+
"logps/chosen": -441.70648193359375,
|
| 619 |
+
"logps/rejected": -458.97052001953125,
|
| 620 |
+
"loss": 0.186,
|
| 621 |
+
"rewards/accuracies": 0.4937500059604645,
|
| 622 |
+
"rewards/chosen": -0.03852088749408722,
|
| 623 |
+
"rewards/margins": 0.038460638374090195,
|
| 624 |
+
"rewards/rejected": -0.07698152959346771,
|
| 625 |
+
"step": 430
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.7,
|
| 629 |
+
"learning_rate": 1.2218673716356919e-06,
|
| 630 |
+
"logits/chosen": -0.05337870121002197,
|
| 631 |
+
"logits/rejected": -0.16203683614730835,
|
| 632 |
+
"logps/chosen": -411.7693786621094,
|
| 633 |
+
"logps/rejected": -472.7538146972656,
|
| 634 |
+
"loss": 0.1868,
|
| 635 |
+
"rewards/accuracies": 0.53125,
|
| 636 |
+
"rewards/chosen": -0.04310641810297966,
|
| 637 |
+
"rewards/margins": 0.053300343453884125,
|
| 638 |
+
"rewards/rejected": -0.09640677273273468,
|
| 639 |
+
"step": 440
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.72,
|
| 643 |
+
"learning_rate": 1.103820512452661e-06,
|
| 644 |
+
"logits/chosen": -0.16054467856884003,
|
| 645 |
+
"logits/rejected": -0.10418369621038437,
|
| 646 |
+
"logps/chosen": -471.15740966796875,
|
| 647 |
+
"logps/rejected": -499.28997802734375,
|
| 648 |
+
"loss": 0.1866,
|
| 649 |
+
"rewards/accuracies": 0.5,
|
| 650 |
+
"rewards/chosen": -0.044102419167757034,
|
| 651 |
+
"rewards/margins": 0.04977239668369293,
|
| 652 |
+
"rewards/rejected": -0.09387481212615967,
|
| 653 |
+
"step": 450
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.74,
|
| 657 |
+
"learning_rate": 9.901353480633468e-07,
|
| 658 |
+
"logits/chosen": -0.01945580169558525,
|
| 659 |
+
"logits/rejected": -0.08020860701799393,
|
| 660 |
+
"logps/chosen": -478.95135498046875,
|
| 661 |
+
"logps/rejected": -510.05645751953125,
|
| 662 |
+
"loss": 0.1876,
|
| 663 |
+
"rewards/accuracies": 0.4937500059604645,
|
| 664 |
+
"rewards/chosen": -0.05099078267812729,
|
| 665 |
+
"rewards/margins": 0.04500190168619156,
|
| 666 |
+
"rewards/rejected": -0.09599269926548004,
|
| 667 |
+
"step": 460
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.75,
|
| 671 |
+
"learning_rate": 8.811670333701544e-07,
|
| 672 |
+
"logits/chosen": -0.09503830224275589,
|
| 673 |
+
"logits/rejected": -0.12620458006858826,
|
| 674 |
+
"logps/chosen": -442.7535095214844,
|
| 675 |
+
"logps/rejected": -478.792236328125,
|
| 676 |
+
"loss": 0.1797,
|
| 677 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 678 |
+
"rewards/chosen": -0.039722852408885956,
|
| 679 |
+
"rewards/margins": 0.04783398285508156,
|
| 680 |
+
"rewards/rejected": -0.08755683898925781,
|
| 681 |
+
"step": 470
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.77,
|
| 685 |
+
"learning_rate": 7.772559877354341e-07,
|
| 686 |
+
"logits/chosen": -0.1489885151386261,
|
| 687 |
+
"logits/rejected": -0.052212201058864594,
|
| 688 |
+
"logps/chosen": -408.1684265136719,
|
| 689 |
+
"logps/rejected": -442.21466064453125,
|
| 690 |
+
"loss": 0.1972,
|
| 691 |
+
"rewards/accuracies": 0.4312500059604645,
|
| 692 |
+
"rewards/chosen": -0.05573519319295883,
|
| 693 |
+
"rewards/margins": 0.033387087285518646,
|
| 694 |
+
"rewards/rejected": -0.08912228047847748,
|
| 695 |
+
"step": 480
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.78,
|
| 699 |
+
"learning_rate": 6.787268315040604e-07,
|
| 700 |
+
"logits/chosen": -0.05187498405575752,
|
| 701 |
+
"logits/rejected": -0.04689168184995651,
|
| 702 |
+
"logps/chosen": -473.5567932128906,
|
| 703 |
+
"logps/rejected": -553.9025268554688,
|
| 704 |
+
"loss": 0.1722,
|
| 705 |
+
"rewards/accuracies": 0.543749988079071,
|
| 706 |
+
"rewards/chosen": -0.047236062586307526,
|
| 707 |
+
"rewards/margins": 0.06369349360466003,
|
| 708 |
+
"rewards/rejected": -0.11092956364154816,
|
| 709 |
+
"step": 490
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.8,
|
| 713 |
+
"learning_rate": 5.858873718824829e-07,
|
| 714 |
+
"logits/chosen": -0.1269344836473465,
|
| 715 |
+
"logits/rejected": -0.12827740609645844,
|
| 716 |
+
"logps/chosen": -436.95001220703125,
|
| 717 |
+
"logps/rejected": -530.323486328125,
|
| 718 |
+
"loss": 0.1828,
|
| 719 |
+
"rewards/accuracies": 0.581250011920929,
|
| 720 |
+
"rewards/chosen": -0.03250400722026825,
|
| 721 |
+
"rewards/margins": 0.06746237725019455,
|
| 722 |
+
"rewards/rejected": -0.0999663770198822,
|
| 723 |
+
"step": 500
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.82,
|
| 727 |
+
"learning_rate": 4.990276413423817e-07,
|
| 728 |
+
"logits/chosen": -0.05357852578163147,
|
| 729 |
+
"logits/rejected": -0.0927656888961792,
|
| 730 |
+
"logps/chosen": -499.13262939453125,
|
| 731 |
+
"logps/rejected": -548.8838500976562,
|
| 732 |
+
"loss": 0.1697,
|
| 733 |
+
"rewards/accuracies": 0.574999988079071,
|
| 734 |
+
"rewards/chosen": -0.04802858829498291,
|
| 735 |
+
"rewards/margins": 0.05423368886113167,
|
| 736 |
+
"rewards/rejected": -0.10226227343082428,
|
| 737 |
+
"step": 510
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.83,
|
| 741 |
+
"learning_rate": 4.184189915529796e-07,
|
| 742 |
+
"logits/chosen": -0.11798721551895142,
|
| 743 |
+
"logits/rejected": -0.09332814067602158,
|
| 744 |
+
"logps/chosen": -465.55816650390625,
|
| 745 |
+
"logps/rejected": -533.0569458007812,
|
| 746 |
+
"loss": 0.1967,
|
| 747 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 748 |
+
"rewards/chosen": -0.042215488851070404,
|
| 749 |
+
"rewards/margins": 0.05858848616480827,
|
| 750 |
+
"rewards/rejected": -0.10080397129058838,
|
| 751 |
+
"step": 520
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.85,
|
| 755 |
+
"learning_rate": 3.4431324567258176e-07,
|
| 756 |
+
"logits/chosen": -0.1948605328798294,
|
| 757 |
+
"logits/rejected": -0.08446381241083145,
|
| 758 |
+
"logps/chosen": -440.0462951660156,
|
| 759 |
+
"logps/rejected": -479.0397033691406,
|
| 760 |
+
"loss": 0.1885,
|
| 761 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 762 |
+
"rewards/chosen": -0.045319609344005585,
|
| 763 |
+
"rewards/margins": 0.04320630431175232,
|
| 764 |
+
"rewards/rejected": -0.08852590620517731,
|
| 765 |
+
"step": 530
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.86,
|
| 769 |
+
"learning_rate": 2.769419116476052e-07,
|
| 770 |
+
"logits/chosen": -0.2041836529970169,
|
| 771 |
+
"logits/rejected": -0.06282895803451538,
|
| 772 |
+
"logps/chosen": -415.84576416015625,
|
| 773 |
+
"logps/rejected": -452.8704528808594,
|
| 774 |
+
"loss": 0.1829,
|
| 775 |
+
"rewards/accuracies": 0.4437499940395355,
|
| 776 |
+
"rewards/chosen": -0.050526998937129974,
|
| 777 |
+
"rewards/margins": 0.03672551363706589,
|
| 778 |
+
"rewards/rejected": -0.08725249767303467,
|
| 779 |
+
"step": 540
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.88,
|
| 783 |
+
"learning_rate": 2.1651545897676512e-07,
|
| 784 |
+
"logits/chosen": -0.056377578526735306,
|
| 785 |
+
"logits/rejected": -0.0602828674018383,
|
| 786 |
+
"logps/chosen": -482.2828674316406,
|
| 787 |
+
"logps/rejected": -510.7273864746094,
|
| 788 |
+
"loss": 0.194,
|
| 789 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 790 |
+
"rewards/chosen": -0.04934944957494736,
|
| 791 |
+
"rewards/margins": 0.04233521968126297,
|
| 792 |
+
"rewards/rejected": -0.09168466180562973,
|
| 793 |
+
"step": 550
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.9,
|
| 797 |
+
"learning_rate": 1.6322266119983222e-07,
|
| 798 |
+
"logits/chosen": -0.08606644719839096,
|
| 799 |
+
"logits/rejected": -0.05609310790896416,
|
| 800 |
+
"logps/chosen": -435.7982482910156,
|
| 801 |
+
"logps/rejected": -460.5846252441406,
|
| 802 |
+
"loss": 0.185,
|
| 803 |
+
"rewards/accuracies": 0.5,
|
| 804 |
+
"rewards/chosen": -0.037545233964920044,
|
| 805 |
+
"rewards/margins": 0.0525357648730278,
|
| 806 |
+
"rewards/rejected": -0.09008099883794785,
|
| 807 |
+
"step": 560
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.91,
|
| 811 |
+
"learning_rate": 1.1723000616502167e-07,
|
| 812 |
+
"logits/chosen": -0.21670150756835938,
|
| 813 |
+
"logits/rejected": -0.034546859562397,
|
| 814 |
+
"logps/chosen": -498.1796875,
|
| 815 |
+
"logps/rejected": -531.9613037109375,
|
| 816 |
+
"loss": 0.1779,
|
| 817 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 818 |
+
"rewards/chosen": -0.05661952495574951,
|
| 819 |
+
"rewards/margins": 0.05806880071759224,
|
| 820 |
+
"rewards/rejected": -0.11468833684921265,
|
| 821 |
+
"step": 570
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.93,
|
| 825 |
+
"learning_rate": 7.868117591737585e-08,
|
| 826 |
+
"logits/chosen": -0.15769873559474945,
|
| 827 |
+
"logits/rejected": -0.11322434991598129,
|
| 828 |
+
"logps/chosen": -501.90802001953125,
|
| 829 |
+
"logps/rejected": -510.9298400878906,
|
| 830 |
+
"loss": 0.1865,
|
| 831 |
+
"rewards/accuracies": 0.518750011920929,
|
| 832 |
+
"rewards/chosen": -0.0535421147942543,
|
| 833 |
+
"rewards/margins": 0.04890746995806694,
|
| 834 |
+
"rewards/rejected": -0.10244959592819214,
|
| 835 |
+
"step": 580
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.94,
|
| 839 |
+
"learning_rate": 4.769659783295383e-08,
|
| 840 |
+
"logits/chosen": -0.1612723022699356,
|
| 841 |
+
"logits/rejected": -0.09152115881443024,
|
| 842 |
+
"logps/chosen": -478.5709533691406,
|
| 843 |
+
"logps/rejected": -501.6288146972656,
|
| 844 |
+
"loss": 0.1926,
|
| 845 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 846 |
+
"rewards/chosen": -0.04939908906817436,
|
| 847 |
+
"rewards/margins": 0.04336509853601456,
|
| 848 |
+
"rewards/rejected": -0.09276418387889862,
|
| 849 |
+
"step": 590
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.96,
|
| 853 |
+
"learning_rate": 2.4373068401120358e-08,
|
| 854 |
+
"logits/chosen": -0.16256621479988098,
|
| 855 |
+
"logits/rejected": 0.01677759736776352,
|
| 856 |
+
"logps/chosen": -480.29571533203125,
|
| 857 |
+
"logps/rejected": -480.0000915527344,
|
| 858 |
+
"loss": 0.1777,
|
| 859 |
+
"rewards/accuracies": 0.5062500238418579,
|
| 860 |
+
"rewards/chosen": -0.059523385018110275,
|
| 861 |
+
"rewards/margins": 0.03656711056828499,
|
| 862 |
+
"rewards/rejected": -0.09609050303697586,
|
| 863 |
+
"step": 600
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.98,
|
| 867 |
+
"learning_rate": 8.78345083022425e-09,
|
| 868 |
+
"logits/chosen": -0.2046932876110077,
|
| 869 |
+
"logits/rejected": -0.010766489431262016,
|
| 870 |
+
"logps/chosen": -467.3860778808594,
|
| 871 |
+
"logps/rejected": -513.6978149414062,
|
| 872 |
+
"loss": 0.195,
|
| 873 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 874 |
+
"rewards/chosen": -0.0569658987224102,
|
| 875 |
+
"rewards/margins": 0.04467035457491875,
|
| 876 |
+
"rewards/rejected": -0.10163625329732895,
|
| 877 |
+
"step": 610
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.99,
|
| 881 |
+
"learning_rate": 9.764474213677654e-10,
|
| 882 |
+
"logits/chosen": -0.15736037492752075,
|
| 883 |
+
"logits/rejected": -0.16347061097621918,
|
| 884 |
+
"logps/chosen": -454.09942626953125,
|
| 885 |
+
"logps/rejected": -512.8070678710938,
|
| 886 |
+
"loss": 0.1987,
|
| 887 |
+
"rewards/accuracies": 0.48124998807907104,
|
| 888 |
+
"rewards/chosen": -0.06590863317251205,
|
| 889 |
+
"rewards/margins": 0.044044043868780136,
|
| 890 |
+
"rewards/rejected": -0.10995267331600189,
|
| 891 |
+
"step": 620
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 1.0,
|
| 895 |
+
"step": 625,
|
| 896 |
+
"total_flos": 0.0,
|
| 897 |
+
"train_loss": 0.1911162176847458,
|
| 898 |
+
"train_runtime": 4612.7069,
|
| 899 |
+
"train_samples_per_second": 4.336,
|
| 900 |
+
"train_steps_per_second": 0.135
|
| 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 |
+
}
|