Model save
Browse files- README.md +61 -0
- adapter_model.safetensors +1 -1
- all_results.json +8 -0
- runs/May20_00-57-21_gpu4-119-5/events.out.tfevents.1716130996.gpu4-119-5.4102717.0 +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-v1-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-v1-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 sha256:7f451daa77b9398369666a12197830386bfeb694c07204233488576127a51cac
|
| 3 |
size 335579632
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.3386969359874725,
|
| 4 |
+
"train_runtime": 4609.7316,
|
| 5 |
+
"train_samples": 20000,
|
| 6 |
+
"train_samples_per_second": 4.339,
|
| 7 |
+
"train_steps_per_second": 0.136
|
| 8 |
+
}
|
runs/May20_00-57-21_gpu4-119-5/events.out.tfevents.1716130996.gpu4-119-5.4102717.0
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:a34c0e9268669f697a8c3e959c9f727898f8549649e0b8a4a62dd48603f6f888
|
| 3 |
+
size 45062
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.3386969359874725,
|
| 4 |
+
"train_runtime": 4609.7316,
|
| 5 |
+
"train_samples": 20000,
|
| 6 |
+
"train_samples_per_second": 4.339,
|
| 7 |
+
"train_steps_per_second": 0.136
|
| 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": 0.36046722531318665,
|
| 15 |
+
"logits/rejected": 0.23892882466316223,
|
| 16 |
+
"logps/chosen": -419.6329345703125,
|
| 17 |
+
"logps/rejected": -437.8076171875,
|
| 18 |
+
"loss": 0.3567,
|
| 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": 0.25086185336112976,
|
| 29 |
+
"logits/rejected": 0.22883599996566772,
|
| 30 |
+
"logps/chosen": -438.00592041015625,
|
| 31 |
+
"logps/rejected": -430.2698059082031,
|
| 32 |
+
"loss": 0.3631,
|
| 33 |
+
"rewards/accuracies": 0.3333333432674408,
|
| 34 |
+
"rewards/chosen": -0.00057529954938218,
|
| 35 |
+
"rewards/margins": 4.002878631581552e-05,
|
| 36 |
+
"rewards/rejected": -0.0006153282593004405,
|
| 37 |
+
"step": 10
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.03,
|
| 41 |
+
"learning_rate": 1.5873015873015873e-06,
|
| 42 |
+
"logits/chosen": 0.19798453152179718,
|
| 43 |
+
"logits/rejected": 0.2455386370420456,
|
| 44 |
+
"logps/chosen": -401.9042663574219,
|
| 45 |
+
"logps/rejected": -393.48822021484375,
|
| 46 |
+
"loss": 0.3746,
|
| 47 |
+
"rewards/accuracies": 0.38749998807907104,
|
| 48 |
+
"rewards/chosen": -0.0006311516626738012,
|
| 49 |
+
"rewards/margins": 7.311334047699347e-05,
|
| 50 |
+
"rewards/rejected": -0.0007042650249786675,
|
| 51 |
+
"step": 20
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.05,
|
| 55 |
+
"learning_rate": 2.380952380952381e-06,
|
| 56 |
+
"logits/chosen": 0.16958989202976227,
|
| 57 |
+
"logits/rejected": 0.3218519687652588,
|
| 58 |
+
"logps/chosen": -421.8275451660156,
|
| 59 |
+
"logps/rejected": -406.69171142578125,
|
| 60 |
+
"loss": 0.3735,
|
| 61 |
+
"rewards/accuracies": 0.38749998807907104,
|
| 62 |
+
"rewards/chosen": -0.0018598968163132668,
|
| 63 |
+
"rewards/margins": 8.136472024489194e-05,
|
| 64 |
+
"rewards/rejected": -0.0019412615802139044,
|
| 65 |
+
"step": 30
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.06,
|
| 69 |
+
"learning_rate": 3.1746031746031746e-06,
|
| 70 |
+
"logits/chosen": 0.2768496870994568,
|
| 71 |
+
"logits/rejected": 0.2004733383655548,
|
| 72 |
+
"logps/chosen": -364.9704895019531,
|
| 73 |
+
"logps/rejected": -383.6319274902344,
|
| 74 |
+
"loss": 0.3722,
|
| 75 |
+
"rewards/accuracies": 0.41874998807907104,
|
| 76 |
+
"rewards/chosen": -0.0037128510884940624,
|
| 77 |
+
"rewards/margins": 0.00042822357499971986,
|
| 78 |
+
"rewards/rejected": -0.004141074605286121,
|
| 79 |
+
"step": 40
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.08,
|
| 83 |
+
"learning_rate": 3.968253968253968e-06,
|
| 84 |
+
"logits/chosen": 0.21148820221424103,
|
| 85 |
+
"logits/rejected": 0.1991022676229477,
|
| 86 |
+
"logps/chosen": -363.0907897949219,
|
| 87 |
+
"logps/rejected": -372.4688415527344,
|
| 88 |
+
"loss": 0.3683,
|
| 89 |
+
"rewards/accuracies": 0.38749998807907104,
|
| 90 |
+
"rewards/chosen": -0.008444704115390778,
|
| 91 |
+
"rewards/margins": 0.000421268108766526,
|
| 92 |
+
"rewards/rejected": -0.00886597204953432,
|
| 93 |
+
"step": 50
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.1,
|
| 97 |
+
"learning_rate": 4.761904761904762e-06,
|
| 98 |
+
"logits/chosen": 0.12682433426380157,
|
| 99 |
+
"logits/rejected": 0.10598279535770416,
|
| 100 |
+
"logps/chosen": -406.6046447753906,
|
| 101 |
+
"logps/rejected": -408.136474609375,
|
| 102 |
+
"loss": 0.3698,
|
| 103 |
+
"rewards/accuracies": 0.5,
|
| 104 |
+
"rewards/chosen": -0.013907397165894508,
|
| 105 |
+
"rewards/margins": 0.0012330651516094804,
|
| 106 |
+
"rewards/rejected": -0.01514046173542738,
|
| 107 |
+
"step": 60
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.11,
|
| 111 |
+
"learning_rate": 4.998086282661188e-06,
|
| 112 |
+
"logits/chosen": 0.15495741367340088,
|
| 113 |
+
"logits/rejected": 0.2392895221710205,
|
| 114 |
+
"logps/chosen": -454.59564208984375,
|
| 115 |
+
"logps/rejected": -472.60150146484375,
|
| 116 |
+
"loss": 0.3647,
|
| 117 |
+
"rewards/accuracies": 0.4937500059604645,
|
| 118 |
+
"rewards/chosen": -0.01984778419137001,
|
| 119 |
+
"rewards/margins": 0.0015219683991745114,
|
| 120 |
+
"rewards/rejected": -0.021369751542806625,
|
| 121 |
+
"step": 70
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.13,
|
| 125 |
+
"learning_rate": 4.988720025682995e-06,
|
| 126 |
+
"logits/chosen": 0.10600455850362778,
|
| 127 |
+
"logits/rejected": 0.19427841901779175,
|
| 128 |
+
"logps/chosen": -456.883544921875,
|
| 129 |
+
"logps/rejected": -457.2091369628906,
|
| 130 |
+
"loss": 0.366,
|
| 131 |
+
"rewards/accuracies": 0.45625001192092896,
|
| 132 |
+
"rewards/chosen": -0.027602821588516235,
|
| 133 |
+
"rewards/margins": 0.0021790049504488707,
|
| 134 |
+
"rewards/rejected": -0.029781833291053772,
|
| 135 |
+
"step": 80
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.14,
|
| 139 |
+
"learning_rate": 4.9715789537359126e-06,
|
| 140 |
+
"logits/chosen": 0.19685187935829163,
|
| 141 |
+
"logits/rejected": 0.1898462027311325,
|
| 142 |
+
"logps/chosen": -429.89715576171875,
|
| 143 |
+
"logps/rejected": -405.14947509765625,
|
| 144 |
+
"loss": 0.374,
|
| 145 |
+
"rewards/accuracies": 0.5062500238418579,
|
| 146 |
+
"rewards/chosen": -0.02775459550321102,
|
| 147 |
+
"rewards/margins": 0.0036151439417153597,
|
| 148 |
+
"rewards/rejected": -0.03136973828077316,
|
| 149 |
+
"step": 90
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.16,
|
| 153 |
+
"learning_rate": 4.946716615897932e-06,
|
| 154 |
+
"logits/chosen": 0.17538510262966156,
|
| 155 |
+
"logits/rejected": 0.20797649025917053,
|
| 156 |
+
"logps/chosen": -436.5021057128906,
|
| 157 |
+
"logps/rejected": -441.6522521972656,
|
| 158 |
+
"loss": 0.3647,
|
| 159 |
+
"rewards/accuracies": 0.5,
|
| 160 |
+
"rewards/chosen": -0.030668649822473526,
|
| 161 |
+
"rewards/margins": 0.0047523751854896545,
|
| 162 |
+
"rewards/rejected": -0.03542102500796318,
|
| 163 |
+
"step": 100
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.18,
|
| 167 |
+
"learning_rate": 4.9142106826480114e-06,
|
| 168 |
+
"logits/chosen": 0.23190093040466309,
|
| 169 |
+
"logits/rejected": 0.2070162296295166,
|
| 170 |
+
"logps/chosen": -495.9847717285156,
|
| 171 |
+
"logps/rejected": -481.325927734375,
|
| 172 |
+
"loss": 0.36,
|
| 173 |
+
"rewards/accuracies": 0.5062500238418579,
|
| 174 |
+
"rewards/chosen": -0.030186587944626808,
|
| 175 |
+
"rewards/margins": 0.006671676877886057,
|
| 176 |
+
"rewards/rejected": -0.036858268082141876,
|
| 177 |
+
"step": 110
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.19,
|
| 181 |
+
"learning_rate": 4.874162703221823e-06,
|
| 182 |
+
"logits/chosen": 0.0589652881026268,
|
| 183 |
+
"logits/rejected": 0.25573208928108215,
|
| 184 |
+
"logps/chosen": -414.246826171875,
|
| 185 |
+
"logps/rejected": -430.9029235839844,
|
| 186 |
+
"loss": 0.3575,
|
| 187 |
+
"rewards/accuracies": 0.5,
|
| 188 |
+
"rewards/chosen": -0.04433692619204521,
|
| 189 |
+
"rewards/margins": 0.015341699123382568,
|
| 190 |
+
"rewards/rejected": -0.05967862531542778,
|
| 191 |
+
"step": 120
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.21,
|
| 195 |
+
"learning_rate": 4.826697788369752e-06,
|
| 196 |
+
"logits/chosen": 0.12831085920333862,
|
| 197 |
+
"logits/rejected": 0.1346360296010971,
|
| 198 |
+
"logps/chosen": -544.7901000976562,
|
| 199 |
+
"logps/rejected": -542.2559814453125,
|
| 200 |
+
"loss": 0.3611,
|
| 201 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 202 |
+
"rewards/chosen": -0.09338424354791641,
|
| 203 |
+
"rewards/margins": 0.020744359120726585,
|
| 204 |
+
"rewards/rejected": -0.11412860453128815,
|
| 205 |
+
"step": 130
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.22,
|
| 209 |
+
"learning_rate": 4.7719642195082224e-06,
|
| 210 |
+
"logits/chosen": 0.004035499412566423,
|
| 211 |
+
"logits/rejected": 0.12359336763620377,
|
| 212 |
+
"logps/chosen": -569.91748046875,
|
| 213 |
+
"logps/rejected": -576.3137817382812,
|
| 214 |
+
"loss": 0.3399,
|
| 215 |
+
"rewards/accuracies": 0.5,
|
| 216 |
+
"rewards/chosen": -0.15238870680332184,
|
| 217 |
+
"rewards/margins": 0.03546822816133499,
|
| 218 |
+
"rewards/rejected": -0.18785694241523743,
|
| 219 |
+
"step": 140
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.24,
|
| 223 |
+
"learning_rate": 4.710132985485355e-06,
|
| 224 |
+
"logits/chosen": 0.09404455125331879,
|
| 225 |
+
"logits/rejected": -0.0020357524044811726,
|
| 226 |
+
"logps/chosen": -502.76776123046875,
|
| 227 |
+
"logps/rejected": -537.9739990234375,
|
| 228 |
+
"loss": 0.3444,
|
| 229 |
+
"rewards/accuracies": 0.4375,
|
| 230 |
+
"rewards/chosen": -0.1469426155090332,
|
| 231 |
+
"rewards/margins": 0.03976910561323166,
|
| 232 |
+
"rewards/rejected": -0.18671169877052307,
|
| 233 |
+
"step": 150
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.26,
|
| 237 |
+
"learning_rate": 4.641397248408122e-06,
|
| 238 |
+
"logits/chosen": 0.10355547815561295,
|
| 239 |
+
"logits/rejected": 0.07864655554294586,
|
| 240 |
+
"logps/chosen": -578.9359741210938,
|
| 241 |
+
"logps/rejected": -687.6909790039062,
|
| 242 |
+
"loss": 0.3059,
|
| 243 |
+
"rewards/accuracies": 0.59375,
|
| 244 |
+
"rewards/chosen": -0.16360266506671906,
|
| 245 |
+
"rewards/margins": 0.07588207721710205,
|
| 246 |
+
"rewards/rejected": -0.2394847422838211,
|
| 247 |
+
"step": 160
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.27,
|
| 251 |
+
"learning_rate": 4.5659717401997655e-06,
|
| 252 |
+
"logits/chosen": 0.032590705901384354,
|
| 253 |
+
"logits/rejected": 0.08898362517356873,
|
| 254 |
+
"logps/chosen": -600.9769287109375,
|
| 255 |
+
"logps/rejected": -670.7233276367188,
|
| 256 |
+
"loss": 0.3212,
|
| 257 |
+
"rewards/accuracies": 0.48124998807907104,
|
| 258 |
+
"rewards/chosen": -0.19255095720291138,
|
| 259 |
+
"rewards/margins": 0.08060728758573532,
|
| 260 |
+
"rewards/rejected": -0.2731582522392273,
|
| 261 |
+
"step": 170
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.29,
|
| 265 |
+
"learning_rate": 4.4840920917726425e-06,
|
| 266 |
+
"logits/chosen": 0.04442215710878372,
|
| 267 |
+
"logits/rejected": 0.042511895298957825,
|
| 268 |
+
"logps/chosen": -588.660888671875,
|
| 269 |
+
"logps/rejected": -651.74853515625,
|
| 270 |
+
"loss": 0.3528,
|
| 271 |
+
"rewards/accuracies": 0.42500001192092896,
|
| 272 |
+
"rewards/chosen": -0.20966176688671112,
|
| 273 |
+
"rewards/margins": 0.05220552161335945,
|
| 274 |
+
"rewards/rejected": -0.2618672847747803,
|
| 275 |
+
"step": 180
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.3,
|
| 279 |
+
"learning_rate": 4.396014096912182e-06,
|
| 280 |
+
"logits/chosen": 0.03010064922273159,
|
| 281 |
+
"logits/rejected": 0.041020795702934265,
|
| 282 |
+
"logps/chosen": -532.5487060546875,
|
| 283 |
+
"logps/rejected": -654.7836303710938,
|
| 284 |
+
"loss": 0.3144,
|
| 285 |
+
"rewards/accuracies": 0.518750011920929,
|
| 286 |
+
"rewards/chosen": -0.18499326705932617,
|
| 287 |
+
"rewards/margins": 0.09925045073032379,
|
| 288 |
+
"rewards/rejected": -0.28424373269081116,
|
| 289 |
+
"step": 190
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.32,
|
| 293 |
+
"learning_rate": 4.302012913171584e-06,
|
| 294 |
+
"logits/chosen": -0.03510774299502373,
|
| 295 |
+
"logits/rejected": 0.035278841853141785,
|
| 296 |
+
"logps/chosen": -640.3848876953125,
|
| 297 |
+
"logps/rejected": -709.2119140625,
|
| 298 |
+
"loss": 0.3341,
|
| 299 |
+
"rewards/accuracies": 0.48124998807907104,
|
| 300 |
+
"rewards/chosen": -0.21959897875785828,
|
| 301 |
+
"rewards/margins": 0.06892947107553482,
|
| 302 |
+
"rewards/rejected": -0.2885284423828125,
|
| 303 |
+
"step": 200
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.34,
|
| 307 |
+
"learning_rate": 4.202382202273702e-06,
|
| 308 |
+
"logits/chosen": 0.017812447622418404,
|
| 309 |
+
"logits/rejected": 0.06379534304141998,
|
| 310 |
+
"logps/chosen": -641.2015991210938,
|
| 311 |
+
"logps/rejected": -672.9808349609375,
|
| 312 |
+
"loss": 0.3513,
|
| 313 |
+
"rewards/accuracies": 0.4937500059604645,
|
| 314 |
+
"rewards/chosen": -0.19523058831691742,
|
| 315 |
+
"rewards/margins": 0.05737825483083725,
|
| 316 |
+
"rewards/rejected": -0.2526088356971741,
|
| 317 |
+
"step": 210
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.35,
|
| 321 |
+
"learning_rate": 4.097433212705492e-06,
|
| 322 |
+
"logits/chosen": 0.0030833049677312374,
|
| 323 |
+
"logits/rejected": 0.027691733092069626,
|
| 324 |
+
"logps/chosen": -645.5457763671875,
|
| 325 |
+
"logps/rejected": -751.8395385742188,
|
| 326 |
+
"loss": 0.3278,
|
| 327 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 328 |
+
"rewards/chosen": -0.1966964602470398,
|
| 329 |
+
"rewards/margins": 0.08199279010295868,
|
| 330 |
+
"rewards/rejected": -0.2786892056465149,
|
| 331 |
+
"step": 220
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.37,
|
| 335 |
+
"learning_rate": 3.987493807371033e-06,
|
| 336 |
+
"logits/chosen": 0.06291202455759048,
|
| 337 |
+
"logits/rejected": -0.08009441941976547,
|
| 338 |
+
"logps/chosen": -500.90496826171875,
|
| 339 |
+
"logps/rejected": -572.1060791015625,
|
| 340 |
+
"loss": 0.3343,
|
| 341 |
+
"rewards/accuracies": 0.5,
|
| 342 |
+
"rewards/chosen": -0.14460867643356323,
|
| 343 |
+
"rewards/margins": 0.05721248313784599,
|
| 344 |
+
"rewards/rejected": -0.20182117819786072,
|
| 345 |
+
"step": 230
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.38,
|
| 349 |
+
"learning_rate": 3.872907439340758e-06,
|
| 350 |
+
"logits/chosen": 0.04002127796411514,
|
| 351 |
+
"logits/rejected": -0.06345082819461823,
|
| 352 |
+
"logps/chosen": -543.0759887695312,
|
| 353 |
+
"logps/rejected": -591.0164184570312,
|
| 354 |
+
"loss": 0.3242,
|
| 355 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 356 |
+
"rewards/chosen": -0.1588135063648224,
|
| 357 |
+
"rewards/margins": 0.05112007260322571,
|
| 358 |
+
"rewards/rejected": -0.2099335640668869,
|
| 359 |
+
"step": 240
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.4,
|
| 363 |
+
"learning_rate": 3.75403207889666e-06,
|
| 364 |
+
"logits/chosen": 0.10061119496822357,
|
| 365 |
+
"logits/rejected": 0.04563102871179581,
|
| 366 |
+
"logps/chosen": -578.7415771484375,
|
| 367 |
+
"logps/rejected": -615.0494384765625,
|
| 368 |
+
"loss": 0.3244,
|
| 369 |
+
"rewards/accuracies": 0.4437499940395355,
|
| 370 |
+
"rewards/chosen": -0.1841258704662323,
|
| 371 |
+
"rewards/margins": 0.055299896746873856,
|
| 372 |
+
"rewards/rejected": -0.23942574858665466,
|
| 373 |
+
"step": 250
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.42,
|
| 377 |
+
"learning_rate": 3.631239095225417e-06,
|
| 378 |
+
"logits/chosen": 0.0002536162792239338,
|
| 379 |
+
"logits/rejected": 0.06318163126707077,
|
| 380 |
+
"logps/chosen": -650.6129150390625,
|
| 381 |
+
"logps/rejected": -664.8394775390625,
|
| 382 |
+
"loss": 0.3604,
|
| 383 |
+
"rewards/accuracies": 0.518750011920929,
|
| 384 |
+
"rewards/chosen": -0.1936233639717102,
|
| 385 |
+
"rewards/margins": 0.039541881531476974,
|
| 386 |
+
"rewards/rejected": -0.23316523432731628,
|
| 387 |
+
"step": 260
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.43,
|
| 391 |
+
"learning_rate": 3.5049120962530608e-06,
|
| 392 |
+
"logits/chosen": 0.04564657807350159,
|
| 393 |
+
"logits/rejected": -0.02100910246372223,
|
| 394 |
+
"logps/chosen": -597.6544799804688,
|
| 395 |
+
"logps/rejected": -677.6250610351562,
|
| 396 |
+
"loss": 0.3337,
|
| 397 |
+
"rewards/accuracies": 0.53125,
|
| 398 |
+
"rewards/chosen": -0.18195489048957825,
|
| 399 |
+
"rewards/margins": 0.07418446242809296,
|
| 400 |
+
"rewards/rejected": -0.2561393678188324,
|
| 401 |
+
"step": 270
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.45,
|
| 405 |
+
"learning_rate": 3.3754457302455464e-06,
|
| 406 |
+
"logits/chosen": -0.03575817123055458,
|
| 407 |
+
"logits/rejected": 0.11129350960254669,
|
| 408 |
+
"logps/chosen": -567.7127685546875,
|
| 409 |
+
"logps/rejected": -608.2735595703125,
|
| 410 |
+
"loss": 0.3427,
|
| 411 |
+
"rewards/accuracies": 0.4312500059604645,
|
| 412 |
+
"rewards/chosen": -0.18929126858711243,
|
| 413 |
+
"rewards/margins": 0.05334322527050972,
|
| 414 |
+
"rewards/rejected": -0.24263450503349304,
|
| 415 |
+
"step": 280
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.46,
|
| 419 |
+
"learning_rate": 3.2432444529190714e-06,
|
| 420 |
+
"logits/chosen": -0.05213651806116104,
|
| 421 |
+
"logits/rejected": 0.0017232149839401245,
|
| 422 |
+
"logps/chosen": -600.5894165039062,
|
| 423 |
+
"logps/rejected": -662.463623046875,
|
| 424 |
+
"loss": 0.3298,
|
| 425 |
+
"rewards/accuracies": 0.5062500238418579,
|
| 426 |
+
"rewards/chosen": -0.16451598703861237,
|
| 427 |
+
"rewards/margins": 0.06247172877192497,
|
| 428 |
+
"rewards/rejected": -0.22698768973350525,
|
| 429 |
+
"step": 290
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.48,
|
| 433 |
+
"learning_rate": 3.1087212639117057e-06,
|
| 434 |
+
"logits/chosen": 0.04510731250047684,
|
| 435 |
+
"logits/rejected": 0.045666079968214035,
|
| 436 |
+
"logps/chosen": -590.2169189453125,
|
| 437 |
+
"logps/rejected": -643.4014892578125,
|
| 438 |
+
"loss": 0.3287,
|
| 439 |
+
"rewards/accuracies": 0.4937500059604645,
|
| 440 |
+
"rewards/chosen": -0.18261924386024475,
|
| 441 |
+
"rewards/margins": 0.05223492905497551,
|
| 442 |
+
"rewards/rejected": -0.23485419154167175,
|
| 443 |
+
"step": 300
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.5,
|
| 447 |
+
"learning_rate": 2.9722964165636263e-06,
|
| 448 |
+
"logits/chosen": 0.006776309106498957,
|
| 449 |
+
"logits/rejected": -0.025738414376974106,
|
| 450 |
+
"logps/chosen": -530.9435424804688,
|
| 451 |
+
"logps/rejected": -585.0616455078125,
|
| 452 |
+
"loss": 0.3287,
|
| 453 |
+
"rewards/accuracies": 0.46875,
|
| 454 |
+
"rewards/chosen": -0.15342922508716583,
|
| 455 |
+
"rewards/margins": 0.056588392704725266,
|
| 456 |
+
"rewards/rejected": -0.2100176066160202,
|
| 457 |
+
"step": 310
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.51,
|
| 461 |
+
"learning_rate": 2.8343961050366275e-06,
|
| 462 |
+
"logits/chosen": -0.02025730162858963,
|
| 463 |
+
"logits/rejected": 0.08972244709730148,
|
| 464 |
+
"logps/chosen": -529.8350830078125,
|
| 465 |
+
"logps/rejected": -604.9669799804688,
|
| 466 |
+
"loss": 0.3274,
|
| 467 |
+
"rewards/accuracies": 0.4937500059604645,
|
| 468 |
+
"rewards/chosen": -0.16149136424064636,
|
| 469 |
+
"rewards/margins": 0.06798367202281952,
|
| 470 |
+
"rewards/rejected": -0.22947505116462708,
|
| 471 |
+
"step": 320
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.53,
|
| 475 |
+
"learning_rate": 2.695451132874385e-06,
|
| 476 |
+
"logits/chosen": -0.0479627326130867,
|
| 477 |
+
"logits/rejected": 0.021157974377274513,
|
| 478 |
+
"logps/chosen": -564.2462158203125,
|
| 479 |
+
"logps/rejected": -604.8670654296875,
|
| 480 |
+
"loss": 0.3409,
|
| 481 |
+
"rewards/accuracies": 0.42500001192092896,
|
| 482 |
+
"rewards/chosen": -0.17655444145202637,
|
| 483 |
+
"rewards/margins": 0.06839374452829361,
|
| 484 |
+
"rewards/rejected": -0.24494817852973938,
|
| 485 |
+
"step": 330
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.54,
|
| 489 |
+
"learning_rate": 2.5558955671628964e-06,
|
| 490 |
+
"logits/chosen": -0.053826987743377686,
|
| 491 |
+
"logits/rejected": 0.036486681550741196,
|
| 492 |
+
"logps/chosen": -616.4885864257812,
|
| 493 |
+
"logps/rejected": -669.8521728515625,
|
| 494 |
+
"loss": 0.3326,
|
| 495 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 496 |
+
"rewards/chosen": -0.20758116245269775,
|
| 497 |
+
"rewards/margins": 0.050070859491825104,
|
| 498 |
+
"rewards/rejected": -0.25765201449394226,
|
| 499 |
+
"step": 340
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.56,
|
| 503 |
+
"learning_rate": 2.4161653824955654e-06,
|
| 504 |
+
"logits/chosen": -0.025303969159722328,
|
| 505 |
+
"logits/rejected": -0.017024224624037743,
|
| 506 |
+
"logps/chosen": -617.4375610351562,
|
| 507 |
+
"logps/rejected": -696.9207153320312,
|
| 508 |
+
"loss": 0.3223,
|
| 509 |
+
"rewards/accuracies": 0.518750011920929,
|
| 510 |
+
"rewards/chosen": -0.19740080833435059,
|
| 511 |
+
"rewards/margins": 0.08515138924121857,
|
| 512 |
+
"rewards/rejected": -0.28255218267440796,
|
| 513 |
+
"step": 350
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.58,
|
| 517 |
+
"learning_rate": 2.2766970989791697e-06,
|
| 518 |
+
"logits/chosen": -0.07398111373186111,
|
| 519 |
+
"logits/rejected": -0.03359592705965042,
|
| 520 |
+
"logps/chosen": -612.1632690429688,
|
| 521 |
+
"logps/rejected": -696.7496337890625,
|
| 522 |
+
"loss": 0.3282,
|
| 523 |
+
"rewards/accuracies": 0.5,
|
| 524 |
+
"rewards/chosen": -0.20693747699260712,
|
| 525 |
+
"rewards/margins": 0.08319561928510666,
|
| 526 |
+
"rewards/rejected": -0.2901330888271332,
|
| 527 |
+
"step": 360
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.59,
|
| 531 |
+
"learning_rate": 2.1379264185356545e-06,
|
| 532 |
+
"logits/chosen": -0.06404633074998856,
|
| 533 |
+
"logits/rejected": 0.025870006531476974,
|
| 534 |
+
"logps/chosen": -648.0364379882812,
|
| 535 |
+
"logps/rejected": -723.4362182617188,
|
| 536 |
+
"loss": 0.3191,
|
| 537 |
+
"rewards/accuracies": 0.5562499761581421,
|
| 538 |
+
"rewards/chosen": -0.225542813539505,
|
| 539 |
+
"rewards/margins": 0.09016820043325424,
|
| 540 |
+
"rewards/rejected": -0.31571102142333984,
|
| 541 |
+
"step": 370
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.61,
|
| 545 |
+
"learning_rate": 2.000286863759934e-06,
|
| 546 |
+
"logits/chosen": 0.06071531027555466,
|
| 547 |
+
"logits/rejected": 0.017028797417879105,
|
| 548 |
+
"logps/chosen": -611.7649536132812,
|
| 549 |
+
"logps/rejected": -710.9679565429688,
|
| 550 |
+
"loss": 0.3274,
|
| 551 |
+
"rewards/accuracies": 0.5562499761581421,
|
| 552 |
+
"rewards/chosen": -0.19213876128196716,
|
| 553 |
+
"rewards/margins": 0.08261444419622421,
|
| 554 |
+
"rewards/rejected": -0.27475324273109436,
|
| 555 |
+
"step": 380
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.62,
|
| 559 |
+
"learning_rate": 1.8642084235859764e-06,
|
| 560 |
+
"logits/chosen": 0.0644644945859909,
|
| 561 |
+
"logits/rejected": -0.021000124514102936,
|
| 562 |
+
"logps/chosen": -533.0925903320312,
|
| 563 |
+
"logps/rejected": -604.2486572265625,
|
| 564 |
+
"loss": 0.3359,
|
| 565 |
+
"rewards/accuracies": 0.5,
|
| 566 |
+
"rewards/chosen": -0.17501072585582733,
|
| 567 |
+
"rewards/margins": 0.06491357088088989,
|
| 568 |
+
"rewards/rejected": -0.23992428183555603,
|
| 569 |
+
"step": 390
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.64,
|
| 573 |
+
"learning_rate": 1.7301162099921013e-06,
|
| 574 |
+
"logits/chosen": -0.021198943257331848,
|
| 575 |
+
"logits/rejected": 0.045659858733415604,
|
| 576 |
+
"logps/chosen": -591.815185546875,
|
| 577 |
+
"logps/rejected": -630.9583129882812,
|
| 578 |
+
"loss": 0.328,
|
| 579 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 580 |
+
"rewards/chosen": -0.18797656893730164,
|
| 581 |
+
"rewards/margins": 0.05078692361712456,
|
| 582 |
+
"rewards/rejected": -0.23876352608203888,
|
| 583 |
+
"step": 400
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.66,
|
| 587 |
+
"learning_rate": 1.5984291299420117e-06,
|
| 588 |
+
"logits/chosen": -0.06265034526586533,
|
| 589 |
+
"logits/rejected": -0.03586529195308685,
|
| 590 |
+
"logps/chosen": -620.4393310546875,
|
| 591 |
+
"logps/rejected": -664.763916015625,
|
| 592 |
+
"loss": 0.3131,
|
| 593 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 594 |
+
"rewards/chosen": -0.20063380897045135,
|
| 595 |
+
"rewards/margins": 0.056611478328704834,
|
| 596 |
+
"rewards/rejected": -0.2572453022003174,
|
| 597 |
+
"step": 410
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.67,
|
| 601 |
+
"learning_rate": 1.4695585767104092e-06,
|
| 602 |
+
"logits/chosen": -0.02047324925661087,
|
| 603 |
+
"logits/rejected": -0.022855054587125778,
|
| 604 |
+
"logps/chosen": -573.5113525390625,
|
| 605 |
+
"logps/rejected": -660.650390625,
|
| 606 |
+
"loss": 0.315,
|
| 607 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 608 |
+
"rewards/chosen": -0.1739957630634308,
|
| 609 |
+
"rewards/margins": 0.0810604989528656,
|
| 610 |
+
"rewards/rejected": -0.2550562918186188,
|
| 611 |
+
"step": 420
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.69,
|
| 615 |
+
"learning_rate": 1.3439071446815452e-06,
|
| 616 |
+
"logits/chosen": 0.018320618197321892,
|
| 617 |
+
"logits/rejected": -0.02154999040067196,
|
| 618 |
+
"logps/chosen": -510.105224609375,
|
| 619 |
+
"logps/rejected": -601.5553588867188,
|
| 620 |
+
"loss": 0.333,
|
| 621 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 622 |
+
"rewards/chosen": -0.1630430966615677,
|
| 623 |
+
"rewards/margins": 0.07893416285514832,
|
| 624 |
+
"rewards/rejected": -0.241977259516716,
|
| 625 |
+
"step": 430
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.7,
|
| 629 |
+
"learning_rate": 1.2218673716356919e-06,
|
| 630 |
+
"logits/chosen": 0.02131899632513523,
|
| 631 |
+
"logits/rejected": 0.102879598736763,
|
| 632 |
+
"logps/chosen": -642.7557373046875,
|
| 633 |
+
"logps/rejected": -687.9267578125,
|
| 634 |
+
"loss": 0.3358,
|
| 635 |
+
"rewards/accuracies": 0.4937500059604645,
|
| 636 |
+
"rewards/chosen": -0.2222980558872223,
|
| 637 |
+
"rewards/margins": 0.061403900384902954,
|
| 638 |
+
"rewards/rejected": -0.28370195627212524,
|
| 639 |
+
"step": 440
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.72,
|
| 643 |
+
"learning_rate": 1.103820512452661e-06,
|
| 644 |
+
"logits/chosen": -0.02241530455648899,
|
| 645 |
+
"logits/rejected": -0.032763995230197906,
|
| 646 |
+
"logps/chosen": -534.8072509765625,
|
| 647 |
+
"logps/rejected": -587.8253173828125,
|
| 648 |
+
"loss": 0.3429,
|
| 649 |
+
"rewards/accuracies": 0.4375,
|
| 650 |
+
"rewards/chosen": -0.18417826294898987,
|
| 651 |
+
"rewards/margins": 0.06289814412593842,
|
| 652 |
+
"rewards/rejected": -0.2470763921737671,
|
| 653 |
+
"step": 450
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.74,
|
| 657 |
+
"learning_rate": 9.901353480633468e-07,
|
| 658 |
+
"logits/chosen": -0.029488349333405495,
|
| 659 |
+
"logits/rejected": -0.008380506187677383,
|
| 660 |
+
"logps/chosen": -640.1456298828125,
|
| 661 |
+
"logps/rejected": -714.1085205078125,
|
| 662 |
+
"loss": 0.3348,
|
| 663 |
+
"rewards/accuracies": 0.5062500238418579,
|
| 664 |
+
"rewards/chosen": -0.2099982500076294,
|
| 665 |
+
"rewards/margins": 0.06554147601127625,
|
| 666 |
+
"rewards/rejected": -0.27553972601890564,
|
| 667 |
+
"step": 460
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.75,
|
| 671 |
+
"learning_rate": 8.811670333701544e-07,
|
| 672 |
+
"logits/chosen": -0.09662221372127533,
|
| 673 |
+
"logits/rejected": -0.06438769400119781,
|
| 674 |
+
"logps/chosen": -583.837158203125,
|
| 675 |
+
"logps/rejected": -702.3072509765625,
|
| 676 |
+
"loss": 0.3233,
|
| 677 |
+
"rewards/accuracies": 0.550000011920929,
|
| 678 |
+
"rewards/chosen": -0.18826954066753387,
|
| 679 |
+
"rewards/margins": 0.0807666927576065,
|
| 680 |
+
"rewards/rejected": -0.2690362334251404,
|
| 681 |
+
"step": 470
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.77,
|
| 685 |
+
"learning_rate": 7.772559877354341e-07,
|
| 686 |
+
"logits/chosen": -0.057563841342926025,
|
| 687 |
+
"logits/rejected": 0.0010903537040576339,
|
| 688 |
+
"logps/chosen": -573.5906372070312,
|
| 689 |
+
"logps/rejected": -646.9957885742188,
|
| 690 |
+
"loss": 0.3245,
|
| 691 |
+
"rewards/accuracies": 0.4937500059604645,
|
| 692 |
+
"rewards/chosen": -0.17442554235458374,
|
| 693 |
+
"rewards/margins": 0.08631418645381927,
|
| 694 |
+
"rewards/rejected": -0.2607397139072418,
|
| 695 |
+
"step": 480
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.78,
|
| 699 |
+
"learning_rate": 6.787268315040604e-07,
|
| 700 |
+
"logits/chosen": -0.06349989026784897,
|
| 701 |
+
"logits/rejected": 0.024532323703169823,
|
| 702 |
+
"logps/chosen": -573.9180908203125,
|
| 703 |
+
"logps/rejected": -613.2721557617188,
|
| 704 |
+
"loss": 0.3313,
|
| 705 |
+
"rewards/accuracies": 0.4312500059604645,
|
| 706 |
+
"rewards/chosen": -0.18243160843849182,
|
| 707 |
+
"rewards/margins": 0.051405251026153564,
|
| 708 |
+
"rewards/rejected": -0.2338368445634842,
|
| 709 |
+
"step": 490
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.8,
|
| 713 |
+
"learning_rate": 5.858873718824829e-07,
|
| 714 |
+
"logits/chosen": -0.06668631732463837,
|
| 715 |
+
"logits/rejected": -0.06943139433860779,
|
| 716 |
+
"logps/chosen": -553.5023193359375,
|
| 717 |
+
"logps/rejected": -593.6824951171875,
|
| 718 |
+
"loss": 0.3319,
|
| 719 |
+
"rewards/accuracies": 0.4375,
|
| 720 |
+
"rewards/chosen": -0.17030836641788483,
|
| 721 |
+
"rewards/margins": 0.0499136857688427,
|
| 722 |
+
"rewards/rejected": -0.22022204101085663,
|
| 723 |
+
"step": 500
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.82,
|
| 727 |
+
"learning_rate": 4.990276413423817e-07,
|
| 728 |
+
"logits/chosen": -0.08017369359731674,
|
| 729 |
+
"logits/rejected": -0.02853136695921421,
|
| 730 |
+
"logps/chosen": -577.9718017578125,
|
| 731 |
+
"logps/rejected": -648.7777099609375,
|
| 732 |
+
"loss": 0.3326,
|
| 733 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 734 |
+
"rewards/chosen": -0.19647789001464844,
|
| 735 |
+
"rewards/margins": 0.058233268558979034,
|
| 736 |
+
"rewards/rejected": -0.2547111511230469,
|
| 737 |
+
"step": 510
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.83,
|
| 741 |
+
"learning_rate": 4.184189915529796e-07,
|
| 742 |
+
"logits/chosen": -0.1332094371318817,
|
| 743 |
+
"logits/rejected": 0.09619398415088654,
|
| 744 |
+
"logps/chosen": -579.1356201171875,
|
| 745 |
+
"logps/rejected": -609.83544921875,
|
| 746 |
+
"loss": 0.3251,
|
| 747 |
+
"rewards/accuracies": 0.5062500238418579,
|
| 748 |
+
"rewards/chosen": -0.16035783290863037,
|
| 749 |
+
"rewards/margins": 0.07735449075698853,
|
| 750 |
+
"rewards/rejected": -0.2377123385667801,
|
| 751 |
+
"step": 520
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.85,
|
| 755 |
+
"learning_rate": 3.4431324567258176e-07,
|
| 756 |
+
"logits/chosen": -0.11564228683710098,
|
| 757 |
+
"logits/rejected": -0.10405786335468292,
|
| 758 |
+
"logps/chosen": -530.2095947265625,
|
| 759 |
+
"logps/rejected": -617.2594604492188,
|
| 760 |
+
"loss": 0.3344,
|
| 761 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 762 |
+
"rewards/chosen": -0.17116107046604156,
|
| 763 |
+
"rewards/margins": 0.08385736495256424,
|
| 764 |
+
"rewards/rejected": -0.255018413066864,
|
| 765 |
+
"step": 530
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.86,
|
| 769 |
+
"learning_rate": 2.769419116476052e-07,
|
| 770 |
+
"logits/chosen": -0.05448366329073906,
|
| 771 |
+
"logits/rejected": 0.020873898640275,
|
| 772 |
+
"logps/chosen": -649.448486328125,
|
| 773 |
+
"logps/rejected": -724.3923950195312,
|
| 774 |
+
"loss": 0.3254,
|
| 775 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 776 |
+
"rewards/chosen": -0.20293286442756653,
|
| 777 |
+
"rewards/margins": 0.0811840295791626,
|
| 778 |
+
"rewards/rejected": -0.2841168940067291,
|
| 779 |
+
"step": 540
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.88,
|
| 783 |
+
"learning_rate": 2.1651545897676512e-07,
|
| 784 |
+
"logits/chosen": -0.09775115549564362,
|
| 785 |
+
"logits/rejected": 0.037994496524333954,
|
| 786 |
+
"logps/chosen": -631.9583740234375,
|
| 787 |
+
"logps/rejected": -665.892578125,
|
| 788 |
+
"loss": 0.3147,
|
| 789 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 790 |
+
"rewards/chosen": -0.192877396941185,
|
| 791 |
+
"rewards/margins": 0.06704525649547577,
|
| 792 |
+
"rewards/rejected": -0.25992268323898315,
|
| 793 |
+
"step": 550
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.9,
|
| 797 |
+
"learning_rate": 1.6322266119983222e-07,
|
| 798 |
+
"logits/chosen": -0.05889749526977539,
|
| 799 |
+
"logits/rejected": 0.006100955419242382,
|
| 800 |
+
"logps/chosen": -603.3699951171875,
|
| 801 |
+
"logps/rejected": -652.2211303710938,
|
| 802 |
+
"loss": 0.3269,
|
| 803 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 804 |
+
"rewards/chosen": -0.18066729605197906,
|
| 805 |
+
"rewards/margins": 0.0697588175535202,
|
| 806 |
+
"rewards/rejected": -0.25042611360549927,
|
| 807 |
+
"step": 560
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.91,
|
| 811 |
+
"learning_rate": 1.1723000616502167e-07,
|
| 812 |
+
"logits/chosen": -0.020711831748485565,
|
| 813 |
+
"logits/rejected": 0.08241745829582214,
|
| 814 |
+
"logps/chosen": -596.9910888671875,
|
| 815 |
+
"logps/rejected": -620.6817626953125,
|
| 816 |
+
"loss": 0.3278,
|
| 817 |
+
"rewards/accuracies": 0.46875,
|
| 818 |
+
"rewards/chosen": -0.21281632781028748,
|
| 819 |
+
"rewards/margins": 0.04477155953645706,
|
| 820 |
+
"rewards/rejected": -0.25758787989616394,
|
| 821 |
+
"step": 570
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.93,
|
| 825 |
+
"learning_rate": 7.868117591737585e-08,
|
| 826 |
+
"logits/chosen": -0.09641040116548538,
|
| 827 |
+
"logits/rejected": -0.06558734178543091,
|
| 828 |
+
"logps/chosen": -527.1932373046875,
|
| 829 |
+
"logps/rejected": -567.8707275390625,
|
| 830 |
+
"loss": 0.3403,
|
| 831 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 832 |
+
"rewards/chosen": -0.1879841685295105,
|
| 833 |
+
"rewards/margins": 0.048898279666900635,
|
| 834 |
+
"rewards/rejected": -0.23688244819641113,
|
| 835 |
+
"step": 580
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.94,
|
| 839 |
+
"learning_rate": 4.769659783295383e-08,
|
| 840 |
+
"logits/chosen": -0.08810693025588989,
|
| 841 |
+
"logits/rejected": -0.04142559692263603,
|
| 842 |
+
"logps/chosen": -554.7185668945312,
|
| 843 |
+
"logps/rejected": -658.844482421875,
|
| 844 |
+
"loss": 0.3395,
|
| 845 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 846 |
+
"rewards/chosen": -0.17377179861068726,
|
| 847 |
+
"rewards/margins": 0.08222335577011108,
|
| 848 |
+
"rewards/rejected": -0.25599512457847595,
|
| 849 |
+
"step": 590
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.96,
|
| 853 |
+
"learning_rate": 2.4373068401120358e-08,
|
| 854 |
+
"logits/chosen": -0.02737409994006157,
|
| 855 |
+
"logits/rejected": -0.07764561474323273,
|
| 856 |
+
"logps/chosen": -561.8909912109375,
|
| 857 |
+
"logps/rejected": -653.6951904296875,
|
| 858 |
+
"loss": 0.3241,
|
| 859 |
+
"rewards/accuracies": 0.4937500059604645,
|
| 860 |
+
"rewards/chosen": -0.18414655327796936,
|
| 861 |
+
"rewards/margins": 0.0771530419588089,
|
| 862 |
+
"rewards/rejected": -0.26129958033561707,
|
| 863 |
+
"step": 600
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.98,
|
| 867 |
+
"learning_rate": 8.78345083022425e-09,
|
| 868 |
+
"logits/chosen": -0.1056148037314415,
|
| 869 |
+
"logits/rejected": 0.01646628975868225,
|
| 870 |
+
"logps/chosen": -578.94970703125,
|
| 871 |
+
"logps/rejected": -632.1832275390625,
|
| 872 |
+
"loss": 0.3343,
|
| 873 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 874 |
+
"rewards/chosen": -0.18656322360038757,
|
| 875 |
+
"rewards/margins": 0.06646813452243805,
|
| 876 |
+
"rewards/rejected": -0.25303134322166443,
|
| 877 |
+
"step": 610
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.99,
|
| 881 |
+
"learning_rate": 9.764474213677654e-10,
|
| 882 |
+
"logits/chosen": -0.047748979181051254,
|
| 883 |
+
"logits/rejected": -0.0808296799659729,
|
| 884 |
+
"logps/chosen": -601.5391235351562,
|
| 885 |
+
"logps/rejected": -620.3516845703125,
|
| 886 |
+
"loss": 0.3484,
|
| 887 |
+
"rewards/accuracies": 0.41874998807907104,
|
| 888 |
+
"rewards/chosen": -0.20741529762744904,
|
| 889 |
+
"rewards/margins": 0.029957950115203857,
|
| 890 |
+
"rewards/rejected": -0.2373732626438141,
|
| 891 |
+
"step": 620
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 1.0,
|
| 895 |
+
"step": 625,
|
| 896 |
+
"total_flos": 0.0,
|
| 897 |
+
"train_loss": 0.3386969359874725,
|
| 898 |
+
"train_runtime": 4609.7316,
|
| 899 |
+
"train_samples_per_second": 4.339,
|
| 900 |
+
"train_steps_per_second": 0.136
|
| 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 |
+
}
|