Model save
Browse files- README.md +61 -0
- adapter_model.safetensors +1 -1
- all_results.json +8 -0
- runs/May17_18-40-26_gpu4-119-5/events.out.tfevents.1715935367.gpu4-119-5.1559685.0 +2 -2
- train_results.json +8 -0
- trainer_state.json +912 -0
README.md
ADDED
|
@@ -0,0 +1,61 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: peft
|
| 3 |
+
tags:
|
| 4 |
+
- trl
|
| 5 |
+
- dpo
|
| 6 |
+
- generated_from_trainer
|
| 7 |
+
base_model: DUAL-GPO-2/phi-2-gpo-v34-merged-i1
|
| 8 |
+
model-index:
|
| 9 |
+
- name: phi-2-gpo-v3-i2
|
| 10 |
+
results: []
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 14 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 15 |
+
|
| 16 |
+
# phi-2-gpo-v3-i2
|
| 17 |
+
|
| 18 |
+
This model is a fine-tuned version of [DUAL-GPO-2/phi-2-gpo-v34-merged-i1](https://huggingface.co/DUAL-GPO-2/phi-2-gpo-v34-merged-i1) on the None dataset.
|
| 19 |
+
|
| 20 |
+
## Model description
|
| 21 |
+
|
| 22 |
+
More information needed
|
| 23 |
+
|
| 24 |
+
## Intended uses & limitations
|
| 25 |
+
|
| 26 |
+
More information needed
|
| 27 |
+
|
| 28 |
+
## Training and evaluation data
|
| 29 |
+
|
| 30 |
+
More information needed
|
| 31 |
+
|
| 32 |
+
## Training procedure
|
| 33 |
+
|
| 34 |
+
### Training hyperparameters
|
| 35 |
+
|
| 36 |
+
The following hyperparameters were used during training:
|
| 37 |
+
- learning_rate: 5e-06
|
| 38 |
+
- train_batch_size: 4
|
| 39 |
+
- eval_batch_size: 4
|
| 40 |
+
- seed: 42
|
| 41 |
+
- distributed_type: multi-GPU
|
| 42 |
+
- num_devices: 3
|
| 43 |
+
- gradient_accumulation_steps: 4
|
| 44 |
+
- total_train_batch_size: 48
|
| 45 |
+
- total_eval_batch_size: 12
|
| 46 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 47 |
+
- lr_scheduler_type: cosine
|
| 48 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 49 |
+
- num_epochs: 1
|
| 50 |
+
|
| 51 |
+
### Training results
|
| 52 |
+
|
| 53 |
+
|
| 54 |
+
|
| 55 |
+
### Framework versions
|
| 56 |
+
|
| 57 |
+
- PEFT 0.7.1
|
| 58 |
+
- Transformers 4.36.2
|
| 59 |
+
- Pytorch 2.1.2+cu121
|
| 60 |
+
- Datasets 2.14.6
|
| 61 |
+
- Tokenizers 0.15.2
|
adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 167807296
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8fbbc9dd5fc67f90863c456dbe01c0dc61da9ff4043e900573dc72fb686a9dbd
|
| 3 |
size 167807296
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 1.7737385009765625,
|
| 4 |
+
"train_runtime": 8419.2459,
|
| 5 |
+
"train_samples": 30000,
|
| 6 |
+
"train_samples_per_second": 3.563,
|
| 7 |
+
"train_steps_per_second": 0.074
|
| 8 |
+
}
|
runs/May17_18-40-26_gpu4-119-5/events.out.tfevents.1715935367.gpu4-119-5.1559685.0
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:fe3c2fca28a53903cd4872a221a0bcaa442e0950f36cde00bfc7da224b580057
|
| 3 |
+
size 44993
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 1.7737385009765625,
|
| 4 |
+
"train_runtime": 8419.2459,
|
| 5 |
+
"train_samples": 30000,
|
| 6 |
+
"train_samples_per_second": 3.563,
|
| 7 |
+
"train_steps_per_second": 0.074
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,912 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 625,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0,
|
| 13 |
+
"learning_rate": 7.936507936507937e-08,
|
| 14 |
+
"logits/chosen": -0.16161203384399414,
|
| 15 |
+
"logits/rejected": -0.17215211689472198,
|
| 16 |
+
"logps/chosen": -349.4459228515625,
|
| 17 |
+
"logps/rejected": -405.1337890625,
|
| 18 |
+
"loss": 2.2717,
|
| 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": -0.16214627027511597,
|
| 29 |
+
"logits/rejected": -0.07371369749307632,
|
| 30 |
+
"logps/chosen": -404.864990234375,
|
| 31 |
+
"logps/rejected": -404.12548828125,
|
| 32 |
+
"loss": 2.1676,
|
| 33 |
+
"rewards/accuracies": 0.4166666567325592,
|
| 34 |
+
"rewards/chosen": 0.00014017699868418276,
|
| 35 |
+
"rewards/margins": -3.6880755942547694e-05,
|
| 36 |
+
"rewards/rejected": 0.00017705775098875165,
|
| 37 |
+
"step": 10
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.03,
|
| 41 |
+
"learning_rate": 1.5873015873015873e-06,
|
| 42 |
+
"logits/chosen": -0.14529487490653992,
|
| 43 |
+
"logits/rejected": -0.10662194341421127,
|
| 44 |
+
"logps/chosen": -417.0052185058594,
|
| 45 |
+
"logps/rejected": -372.1432189941406,
|
| 46 |
+
"loss": 2.2266,
|
| 47 |
+
"rewards/accuracies": 0.41874998807907104,
|
| 48 |
+
"rewards/chosen": -0.0004417522868607193,
|
| 49 |
+
"rewards/margins": -3.551782720023766e-05,
|
| 50 |
+
"rewards/rejected": -0.0004062344669364393,
|
| 51 |
+
"step": 20
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.05,
|
| 55 |
+
"learning_rate": 2.380952380952381e-06,
|
| 56 |
+
"logits/chosen": -0.18683600425720215,
|
| 57 |
+
"logits/rejected": -0.12256692349910736,
|
| 58 |
+
"logps/chosen": -439.61053466796875,
|
| 59 |
+
"logps/rejected": -411.0472106933594,
|
| 60 |
+
"loss": 2.2231,
|
| 61 |
+
"rewards/accuracies": 0.5562499761581421,
|
| 62 |
+
"rewards/chosen": -0.002564679365605116,
|
| 63 |
+
"rewards/margins": 0.0009385704761371017,
|
| 64 |
+
"rewards/rejected": -0.0035032504238188267,
|
| 65 |
+
"step": 30
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.06,
|
| 69 |
+
"learning_rate": 3.1746031746031746e-06,
|
| 70 |
+
"logits/chosen": -0.15973249077796936,
|
| 71 |
+
"logits/rejected": -0.06972485780715942,
|
| 72 |
+
"logps/chosen": -432.1578674316406,
|
| 73 |
+
"logps/rejected": -439.65692138671875,
|
| 74 |
+
"loss": 2.1315,
|
| 75 |
+
"rewards/accuracies": 0.543749988079071,
|
| 76 |
+
"rewards/chosen": -0.008030252531170845,
|
| 77 |
+
"rewards/margins": 0.005707239266484976,
|
| 78 |
+
"rewards/rejected": -0.013737490400671959,
|
| 79 |
+
"step": 40
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.08,
|
| 83 |
+
"learning_rate": 3.968253968253968e-06,
|
| 84 |
+
"logits/chosen": -0.180991992354393,
|
| 85 |
+
"logits/rejected": -0.12702423334121704,
|
| 86 |
+
"logps/chosen": -445.7837829589844,
|
| 87 |
+
"logps/rejected": -477.53204345703125,
|
| 88 |
+
"loss": 2.1043,
|
| 89 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 90 |
+
"rewards/chosen": -0.021977921947836876,
|
| 91 |
+
"rewards/margins": 0.014845812693238258,
|
| 92 |
+
"rewards/rejected": -0.036823734641075134,
|
| 93 |
+
"step": 50
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.1,
|
| 97 |
+
"learning_rate": 4.761904761904762e-06,
|
| 98 |
+
"logits/chosen": -0.2151261866092682,
|
| 99 |
+
"logits/rejected": -0.11040729284286499,
|
| 100 |
+
"logps/chosen": -516.127685546875,
|
| 101 |
+
"logps/rejected": -513.0391845703125,
|
| 102 |
+
"loss": 2.0342,
|
| 103 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 104 |
+
"rewards/chosen": -0.07830885052680969,
|
| 105 |
+
"rewards/margins": 0.029339421540498734,
|
| 106 |
+
"rewards/rejected": -0.10764826834201813,
|
| 107 |
+
"step": 60
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.11,
|
| 111 |
+
"learning_rate": 4.998086282661188e-06,
|
| 112 |
+
"logits/chosen": -0.19878175854682922,
|
| 113 |
+
"logits/rejected": -0.1446671187877655,
|
| 114 |
+
"logps/chosen": -586.38037109375,
|
| 115 |
+
"logps/rejected": -702.6896362304688,
|
| 116 |
+
"loss": 1.9578,
|
| 117 |
+
"rewards/accuracies": 0.643750011920929,
|
| 118 |
+
"rewards/chosen": -0.18542954325675964,
|
| 119 |
+
"rewards/margins": 0.10293972492218018,
|
| 120 |
+
"rewards/rejected": -0.2883692979812622,
|
| 121 |
+
"step": 70
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.13,
|
| 125 |
+
"learning_rate": 4.988720025682995e-06,
|
| 126 |
+
"logits/chosen": -0.2566753923892975,
|
| 127 |
+
"logits/rejected": -0.1441725641489029,
|
| 128 |
+
"logps/chosen": -738.7784423828125,
|
| 129 |
+
"logps/rejected": -817.1397705078125,
|
| 130 |
+
"loss": 1.8572,
|
| 131 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 132 |
+
"rewards/chosen": -0.3142145276069641,
|
| 133 |
+
"rewards/margins": 0.1317647397518158,
|
| 134 |
+
"rewards/rejected": -0.44597920775413513,
|
| 135 |
+
"step": 80
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.14,
|
| 139 |
+
"learning_rate": 4.9715789537359126e-06,
|
| 140 |
+
"logits/chosen": -0.221205472946167,
|
| 141 |
+
"logits/rejected": -0.1643596887588501,
|
| 142 |
+
"logps/chosen": -786.4627685546875,
|
| 143 |
+
"logps/rejected": -940.6861572265625,
|
| 144 |
+
"loss": 1.7929,
|
| 145 |
+
"rewards/accuracies": 0.606249988079071,
|
| 146 |
+
"rewards/chosen": -0.36182770133018494,
|
| 147 |
+
"rewards/margins": 0.19319193065166473,
|
| 148 |
+
"rewards/rejected": -0.5550196766853333,
|
| 149 |
+
"step": 90
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.16,
|
| 153 |
+
"learning_rate": 4.946716615897932e-06,
|
| 154 |
+
"logits/chosen": -0.2160252034664154,
|
| 155 |
+
"logits/rejected": -0.1826382726430893,
|
| 156 |
+
"logps/chosen": -811.1300659179688,
|
| 157 |
+
"logps/rejected": -964.46337890625,
|
| 158 |
+
"loss": 1.8284,
|
| 159 |
+
"rewards/accuracies": 0.668749988079071,
|
| 160 |
+
"rewards/chosen": -0.4078206419944763,
|
| 161 |
+
"rewards/margins": 0.19059345126152039,
|
| 162 |
+
"rewards/rejected": -0.5984141826629639,
|
| 163 |
+
"step": 100
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.18,
|
| 167 |
+
"learning_rate": 4.9142106826480114e-06,
|
| 168 |
+
"logits/chosen": -0.22243019938468933,
|
| 169 |
+
"logits/rejected": -0.1664743721485138,
|
| 170 |
+
"logps/chosen": -861.86767578125,
|
| 171 |
+
"logps/rejected": -994.6505737304688,
|
| 172 |
+
"loss": 1.9512,
|
| 173 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 174 |
+
"rewards/chosen": -0.44844961166381836,
|
| 175 |
+
"rewards/margins": 0.15178686380386353,
|
| 176 |
+
"rewards/rejected": -0.6002364158630371,
|
| 177 |
+
"step": 110
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.19,
|
| 181 |
+
"learning_rate": 4.874162703221823e-06,
|
| 182 |
+
"logits/chosen": -0.2382921725511551,
|
| 183 |
+
"logits/rejected": -0.1875036060810089,
|
| 184 |
+
"logps/chosen": -862.8167114257812,
|
| 185 |
+
"logps/rejected": -996.0676879882812,
|
| 186 |
+
"loss": 1.8951,
|
| 187 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 188 |
+
"rewards/chosen": -0.43303823471069336,
|
| 189 |
+
"rewards/margins": 0.16045571863651276,
|
| 190 |
+
"rewards/rejected": -0.5934939980506897,
|
| 191 |
+
"step": 120
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.21,
|
| 195 |
+
"learning_rate": 4.826697788369752e-06,
|
| 196 |
+
"logits/chosen": -0.22683772444725037,
|
| 197 |
+
"logits/rejected": -0.19122302532196045,
|
| 198 |
+
"logps/chosen": -942.7552490234375,
|
| 199 |
+
"logps/rejected": -1131.6470947265625,
|
| 200 |
+
"loss": 1.7237,
|
| 201 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 202 |
+
"rewards/chosen": -0.4950883984565735,
|
| 203 |
+
"rewards/margins": 0.21038611233234406,
|
| 204 |
+
"rewards/rejected": -0.7054744958877563,
|
| 205 |
+
"step": 130
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.22,
|
| 209 |
+
"learning_rate": 4.7719642195082224e-06,
|
| 210 |
+
"logits/chosen": -0.1632569581270218,
|
| 211 |
+
"logits/rejected": -0.17590998113155365,
|
| 212 |
+
"logps/chosen": -867.0773315429688,
|
| 213 |
+
"logps/rejected": -1070.54296875,
|
| 214 |
+
"loss": 1.7874,
|
| 215 |
+
"rewards/accuracies": 0.643750011920929,
|
| 216 |
+
"rewards/chosen": -0.47380104660987854,
|
| 217 |
+
"rewards/margins": 0.20263484120368958,
|
| 218 |
+
"rewards/rejected": -0.6764358878135681,
|
| 219 |
+
"step": 140
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.24,
|
| 223 |
+
"learning_rate": 4.710132985485355e-06,
|
| 224 |
+
"logits/chosen": -0.17881760001182556,
|
| 225 |
+
"logits/rejected": -0.148959681391716,
|
| 226 |
+
"logps/chosen": -871.22216796875,
|
| 227 |
+
"logps/rejected": -1014.7432861328125,
|
| 228 |
+
"loss": 1.7998,
|
| 229 |
+
"rewards/accuracies": 0.581250011920929,
|
| 230 |
+
"rewards/chosen": -0.43568116426467896,
|
| 231 |
+
"rewards/margins": 0.19202783703804016,
|
| 232 |
+
"rewards/rejected": -0.6277089715003967,
|
| 233 |
+
"step": 150
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.26,
|
| 237 |
+
"learning_rate": 4.641397248408122e-06,
|
| 238 |
+
"logits/chosen": -0.16328132152557373,
|
| 239 |
+
"logits/rejected": -0.1337490677833557,
|
| 240 |
+
"logps/chosen": -840.59423828125,
|
| 241 |
+
"logps/rejected": -991.12841796875,
|
| 242 |
+
"loss": 1.7773,
|
| 243 |
+
"rewards/accuracies": 0.606249988079071,
|
| 244 |
+
"rewards/chosen": -0.40007978677749634,
|
| 245 |
+
"rewards/margins": 0.1872185319662094,
|
| 246 |
+
"rewards/rejected": -0.5872983336448669,
|
| 247 |
+
"step": 160
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.27,
|
| 251 |
+
"learning_rate": 4.5659717401997655e-06,
|
| 252 |
+
"logits/chosen": -0.1909216344356537,
|
| 253 |
+
"logits/rejected": -0.14504285156726837,
|
| 254 |
+
"logps/chosen": -804.4796142578125,
|
| 255 |
+
"logps/rejected": -962.5818481445312,
|
| 256 |
+
"loss": 1.9007,
|
| 257 |
+
"rewards/accuracies": 0.625,
|
| 258 |
+
"rewards/chosen": -0.4276674687862396,
|
| 259 |
+
"rewards/margins": 0.1687774360179901,
|
| 260 |
+
"rewards/rejected": -0.5964449644088745,
|
| 261 |
+
"step": 170
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.29,
|
| 265 |
+
"learning_rate": 4.4840920917726425e-06,
|
| 266 |
+
"logits/chosen": -0.18420439958572388,
|
| 267 |
+
"logits/rejected": -0.15532034635543823,
|
| 268 |
+
"logps/chosen": -728.2938232421875,
|
| 269 |
+
"logps/rejected": -969.2843017578125,
|
| 270 |
+
"loss": 1.8113,
|
| 271 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 272 |
+
"rewards/chosen": -0.3586859405040741,
|
| 273 |
+
"rewards/margins": 0.2266702651977539,
|
| 274 |
+
"rewards/rejected": -0.5853561758995056,
|
| 275 |
+
"step": 180
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.3,
|
| 279 |
+
"learning_rate": 4.396014096912182e-06,
|
| 280 |
+
"logits/chosen": -0.13352888822555542,
|
| 281 |
+
"logits/rejected": -0.11805830150842667,
|
| 282 |
+
"logps/chosen": -830.2320556640625,
|
| 283 |
+
"logps/rejected": -1047.818359375,
|
| 284 |
+
"loss": 1.7352,
|
| 285 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 286 |
+
"rewards/chosen": -0.4081611633300781,
|
| 287 |
+
"rewards/margins": 0.24260549247264862,
|
| 288 |
+
"rewards/rejected": -0.6507667303085327,
|
| 289 |
+
"step": 190
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.32,
|
| 293 |
+
"learning_rate": 4.302012913171584e-06,
|
| 294 |
+
"logits/chosen": -0.21690301597118378,
|
| 295 |
+
"logits/rejected": -0.19134214520454407,
|
| 296 |
+
"logps/chosen": -941.5948486328125,
|
| 297 |
+
"logps/rejected": -1166.3424072265625,
|
| 298 |
+
"loss": 1.8958,
|
| 299 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 300 |
+
"rewards/chosen": -0.5229096412658691,
|
| 301 |
+
"rewards/margins": 0.22313475608825684,
|
| 302 |
+
"rewards/rejected": -0.746044397354126,
|
| 303 |
+
"step": 200
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.34,
|
| 307 |
+
"learning_rate": 4.202382202273702e-06,
|
| 308 |
+
"logits/chosen": -0.2081947773694992,
|
| 309 |
+
"logits/rejected": -0.14332859218120575,
|
| 310 |
+
"logps/chosen": -888.3336791992188,
|
| 311 |
+
"logps/rejected": -1034.1588134765625,
|
| 312 |
+
"loss": 1.8719,
|
| 313 |
+
"rewards/accuracies": 0.637499988079071,
|
| 314 |
+
"rewards/chosen": -0.45206698775291443,
|
| 315 |
+
"rewards/margins": 0.20269028842449188,
|
| 316 |
+
"rewards/rejected": -0.6547572612762451,
|
| 317 |
+
"step": 210
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.35,
|
| 321 |
+
"learning_rate": 4.097433212705492e-06,
|
| 322 |
+
"logits/chosen": -0.15774083137512207,
|
| 323 |
+
"logits/rejected": -0.12937268614768982,
|
| 324 |
+
"logps/chosen": -831.2762451171875,
|
| 325 |
+
"logps/rejected": -972.7151489257812,
|
| 326 |
+
"loss": 1.7465,
|
| 327 |
+
"rewards/accuracies": 0.625,
|
| 328 |
+
"rewards/chosen": -0.4013233780860901,
|
| 329 |
+
"rewards/margins": 0.16350558400154114,
|
| 330 |
+
"rewards/rejected": -0.5648289918899536,
|
| 331 |
+
"step": 220
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.37,
|
| 335 |
+
"learning_rate": 3.987493807371033e-06,
|
| 336 |
+
"logits/chosen": -0.17701825499534607,
|
| 337 |
+
"logits/rejected": -0.11814224720001221,
|
| 338 |
+
"logps/chosen": -738.5556640625,
|
| 339 |
+
"logps/rejected": -925.3514404296875,
|
| 340 |
+
"loss": 1.7971,
|
| 341 |
+
"rewards/accuracies": 0.625,
|
| 342 |
+
"rewards/chosen": -0.33476728200912476,
|
| 343 |
+
"rewards/margins": 0.19879043102264404,
|
| 344 |
+
"rewards/rejected": -0.5335577130317688,
|
| 345 |
+
"step": 230
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.38,
|
| 349 |
+
"learning_rate": 3.872907439340758e-06,
|
| 350 |
+
"logits/chosen": -0.14647835493087769,
|
| 351 |
+
"logits/rejected": -0.13141624629497528,
|
| 352 |
+
"logps/chosen": -827.5966796875,
|
| 353 |
+
"logps/rejected": -1106.738037109375,
|
| 354 |
+
"loss": 1.6676,
|
| 355 |
+
"rewards/accuracies": 0.75,
|
| 356 |
+
"rewards/chosen": -0.3911557197570801,
|
| 357 |
+
"rewards/margins": 0.2823970317840576,
|
| 358 |
+
"rewards/rejected": -0.6735527515411377,
|
| 359 |
+
"step": 240
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.4,
|
| 363 |
+
"learning_rate": 3.75403207889666e-06,
|
| 364 |
+
"logits/chosen": -0.165244922041893,
|
| 365 |
+
"logits/rejected": -0.10319925844669342,
|
| 366 |
+
"logps/chosen": -897.0632934570312,
|
| 367 |
+
"logps/rejected": -1235.7066650390625,
|
| 368 |
+
"loss": 1.6872,
|
| 369 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 370 |
+
"rewards/chosen": -0.44384971261024475,
|
| 371 |
+
"rewards/margins": 0.32984811067581177,
|
| 372 |
+
"rewards/rejected": -0.7736977934837341,
|
| 373 |
+
"step": 250
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.42,
|
| 377 |
+
"learning_rate": 3.631239095225417e-06,
|
| 378 |
+
"logits/chosen": -0.18860653042793274,
|
| 379 |
+
"logits/rejected": -0.20599384605884552,
|
| 380 |
+
"logps/chosen": -895.7766723632812,
|
| 381 |
+
"logps/rejected": -1160.929931640625,
|
| 382 |
+
"loss": 1.6633,
|
| 383 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 384 |
+
"rewards/chosen": -0.486502081155777,
|
| 385 |
+
"rewards/margins": 0.2695665657520294,
|
| 386 |
+
"rewards/rejected": -0.7560686469078064,
|
| 387 |
+
"step": 260
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.43,
|
| 391 |
+
"learning_rate": 3.5049120962530608e-06,
|
| 392 |
+
"logits/chosen": -0.14848849177360535,
|
| 393 |
+
"logits/rejected": -0.1331581175327301,
|
| 394 |
+
"logps/chosen": -992.6770629882812,
|
| 395 |
+
"logps/rejected": -1140.1298828125,
|
| 396 |
+
"loss": 1.7584,
|
| 397 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 398 |
+
"rewards/chosen": -0.510170578956604,
|
| 399 |
+
"rewards/margins": 0.20152750611305237,
|
| 400 |
+
"rewards/rejected": -0.7116979956626892,
|
| 401 |
+
"step": 270
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.45,
|
| 405 |
+
"learning_rate": 3.3754457302455464e-06,
|
| 406 |
+
"logits/chosen": -0.13460226356983185,
|
| 407 |
+
"logits/rejected": -0.08309068530797958,
|
| 408 |
+
"logps/chosen": -913.01416015625,
|
| 409 |
+
"logps/rejected": -1093.90234375,
|
| 410 |
+
"loss": 1.7657,
|
| 411 |
+
"rewards/accuracies": 0.6875,
|
| 412 |
+
"rewards/chosen": -0.45719900727272034,
|
| 413 |
+
"rewards/margins": 0.2582068145275116,
|
| 414 |
+
"rewards/rejected": -0.7154058218002319,
|
| 415 |
+
"step": 280
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.46,
|
| 419 |
+
"learning_rate": 3.2432444529190714e-06,
|
| 420 |
+
"logits/chosen": -0.16204790771007538,
|
| 421 |
+
"logits/rejected": -0.15286168456077576,
|
| 422 |
+
"logps/chosen": -848.2911987304688,
|
| 423 |
+
"logps/rejected": -1192.6624755859375,
|
| 424 |
+
"loss": 1.6853,
|
| 425 |
+
"rewards/accuracies": 0.6875,
|
| 426 |
+
"rewards/chosen": -0.4555201530456543,
|
| 427 |
+
"rewards/margins": 0.324947327375412,
|
| 428 |
+
"rewards/rejected": -0.7804675102233887,
|
| 429 |
+
"step": 290
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.48,
|
| 433 |
+
"learning_rate": 3.1087212639117057e-06,
|
| 434 |
+
"logits/chosen": -0.145342618227005,
|
| 435 |
+
"logits/rejected": -0.15378527343273163,
|
| 436 |
+
"logps/chosen": -855.2142333984375,
|
| 437 |
+
"logps/rejected": -1193.8294677734375,
|
| 438 |
+
"loss": 1.7113,
|
| 439 |
+
"rewards/accuracies": 0.668749988079071,
|
| 440 |
+
"rewards/chosen": -0.4761961102485657,
|
| 441 |
+
"rewards/margins": 0.3075980544090271,
|
| 442 |
+
"rewards/rejected": -0.7837941646575928,
|
| 443 |
+
"step": 300
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.5,
|
| 447 |
+
"learning_rate": 2.9722964165636263e-06,
|
| 448 |
+
"logits/chosen": -0.12143807113170624,
|
| 449 |
+
"logits/rejected": -0.16259871423244476,
|
| 450 |
+
"logps/chosen": -939.7080078125,
|
| 451 |
+
"logps/rejected": -1251.651123046875,
|
| 452 |
+
"loss": 1.7493,
|
| 453 |
+
"rewards/accuracies": 0.71875,
|
| 454 |
+
"rewards/chosen": -0.4829345643520355,
|
| 455 |
+
"rewards/margins": 0.3475037217140198,
|
| 456 |
+
"rewards/rejected": -0.8304383158683777,
|
| 457 |
+
"step": 310
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.51,
|
| 461 |
+
"learning_rate": 2.8343961050366275e-06,
|
| 462 |
+
"logits/chosen": -0.14929132163524628,
|
| 463 |
+
"logits/rejected": -0.1324179470539093,
|
| 464 |
+
"logps/chosen": -932.6903076171875,
|
| 465 |
+
"logps/rejected": -1189.304443359375,
|
| 466 |
+
"loss": 1.7696,
|
| 467 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 468 |
+
"rewards/chosen": -0.5320444107055664,
|
| 469 |
+
"rewards/margins": 0.24077489972114563,
|
| 470 |
+
"rewards/rejected": -0.7728193998336792,
|
| 471 |
+
"step": 320
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.53,
|
| 475 |
+
"learning_rate": 2.695451132874385e-06,
|
| 476 |
+
"logits/chosen": -0.18022935092449188,
|
| 477 |
+
"logits/rejected": -0.11948423087596893,
|
| 478 |
+
"logps/chosen": -945.2275390625,
|
| 479 |
+
"logps/rejected": -1175.975341796875,
|
| 480 |
+
"loss": 1.6686,
|
| 481 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 482 |
+
"rewards/chosen": -0.5022936463356018,
|
| 483 |
+
"rewards/margins": 0.2758629322052002,
|
| 484 |
+
"rewards/rejected": -0.7781565189361572,
|
| 485 |
+
"step": 330
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.54,
|
| 489 |
+
"learning_rate": 2.5558955671628964e-06,
|
| 490 |
+
"logits/chosen": -0.12632670998573303,
|
| 491 |
+
"logits/rejected": -0.114730104804039,
|
| 492 |
+
"logps/chosen": -963.2265625,
|
| 493 |
+
"logps/rejected": -1251.21533203125,
|
| 494 |
+
"loss": 1.6963,
|
| 495 |
+
"rewards/accuracies": 0.675000011920929,
|
| 496 |
+
"rewards/chosen": -0.5192445516586304,
|
| 497 |
+
"rewards/margins": 0.2935452461242676,
|
| 498 |
+
"rewards/rejected": -0.812789797782898,
|
| 499 |
+
"step": 340
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.56,
|
| 503 |
+
"learning_rate": 2.4161653824955654e-06,
|
| 504 |
+
"logits/chosen": -0.11149231344461441,
|
| 505 |
+
"logits/rejected": -0.1491885930299759,
|
| 506 |
+
"logps/chosen": -913.9876098632812,
|
| 507 |
+
"logps/rejected": -1161.088623046875,
|
| 508 |
+
"loss": 1.6863,
|
| 509 |
+
"rewards/accuracies": 0.65625,
|
| 510 |
+
"rewards/chosen": -0.4950644075870514,
|
| 511 |
+
"rewards/margins": 0.2864512801170349,
|
| 512 |
+
"rewards/rejected": -0.7815156579017639,
|
| 513 |
+
"step": 350
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.58,
|
| 517 |
+
"learning_rate": 2.2766970989791697e-06,
|
| 518 |
+
"logits/chosen": -0.09586126357316971,
|
| 519 |
+
"logits/rejected": -0.06141890212893486,
|
| 520 |
+
"logps/chosen": -859.8944091796875,
|
| 521 |
+
"logps/rejected": -1119.4293212890625,
|
| 522 |
+
"loss": 1.7026,
|
| 523 |
+
"rewards/accuracies": 0.65625,
|
| 524 |
+
"rewards/chosen": -0.4382789134979248,
|
| 525 |
+
"rewards/margins": 0.2800602316856384,
|
| 526 |
+
"rewards/rejected": -0.718339204788208,
|
| 527 |
+
"step": 360
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.59,
|
| 531 |
+
"learning_rate": 2.1379264185356545e-06,
|
| 532 |
+
"logits/chosen": -0.0828716978430748,
|
| 533 |
+
"logits/rejected": -0.0706620067358017,
|
| 534 |
+
"logps/chosen": -909.8837890625,
|
| 535 |
+
"logps/rejected": -1223.965576171875,
|
| 536 |
+
"loss": 1.7231,
|
| 537 |
+
"rewards/accuracies": 0.71875,
|
| 538 |
+
"rewards/chosen": -0.47254592180252075,
|
| 539 |
+
"rewards/margins": 0.340139776468277,
|
| 540 |
+
"rewards/rejected": -0.8126856684684753,
|
| 541 |
+
"step": 370
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.61,
|
| 545 |
+
"learning_rate": 2.000286863759934e-06,
|
| 546 |
+
"logits/chosen": -0.12816865742206573,
|
| 547 |
+
"logits/rejected": -0.12074576318264008,
|
| 548 |
+
"logps/chosen": -866.0748291015625,
|
| 549 |
+
"logps/rejected": -1134.2685546875,
|
| 550 |
+
"loss": 1.6912,
|
| 551 |
+
"rewards/accuracies": 0.637499988079071,
|
| 552 |
+
"rewards/chosen": -0.45770812034606934,
|
| 553 |
+
"rewards/margins": 0.29924601316452026,
|
| 554 |
+
"rewards/rejected": -0.7569540739059448,
|
| 555 |
+
"step": 380
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.62,
|
| 559 |
+
"learning_rate": 1.8642084235859764e-06,
|
| 560 |
+
"logits/chosen": -0.11092492192983627,
|
| 561 |
+
"logits/rejected": -0.06964071094989777,
|
| 562 |
+
"logps/chosen": -876.1008911132812,
|
| 563 |
+
"logps/rejected": -1147.765625,
|
| 564 |
+
"loss": 1.7318,
|
| 565 |
+
"rewards/accuracies": 0.71875,
|
| 566 |
+
"rewards/chosen": -0.40335574746131897,
|
| 567 |
+
"rewards/margins": 0.2833205759525299,
|
| 568 |
+
"rewards/rejected": -0.6866763234138489,
|
| 569 |
+
"step": 390
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.64,
|
| 573 |
+
"learning_rate": 1.7301162099921013e-06,
|
| 574 |
+
"logits/chosen": -0.10882829129695892,
|
| 575 |
+
"logits/rejected": -0.06481163948774338,
|
| 576 |
+
"logps/chosen": -871.3518676757812,
|
| 577 |
+
"logps/rejected": -1078.458740234375,
|
| 578 |
+
"loss": 1.61,
|
| 579 |
+
"rewards/accuracies": 0.643750011920929,
|
| 580 |
+
"rewards/chosen": -0.44269245862960815,
|
| 581 |
+
"rewards/margins": 0.25641289353370667,
|
| 582 |
+
"rewards/rejected": -0.6991053819656372,
|
| 583 |
+
"step": 400
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.66,
|
| 587 |
+
"learning_rate": 1.5984291299420117e-06,
|
| 588 |
+
"logits/chosen": -0.08046077191829681,
|
| 589 |
+
"logits/rejected": -0.05565166473388672,
|
| 590 |
+
"logps/chosen": -783.994873046875,
|
| 591 |
+
"logps/rejected": -1022.2307739257812,
|
| 592 |
+
"loss": 1.8072,
|
| 593 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 594 |
+
"rewards/chosen": -0.4169149398803711,
|
| 595 |
+
"rewards/margins": 0.2162661999464035,
|
| 596 |
+
"rewards/rejected": -0.6331812143325806,
|
| 597 |
+
"step": 410
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.67,
|
| 601 |
+
"learning_rate": 1.4695585767104092e-06,
|
| 602 |
+
"logits/chosen": -0.11396467685699463,
|
| 603 |
+
"logits/rejected": -0.10157410800457001,
|
| 604 |
+
"logps/chosen": -899.9319458007812,
|
| 605 |
+
"logps/rejected": -1116.689697265625,
|
| 606 |
+
"loss": 1.7069,
|
| 607 |
+
"rewards/accuracies": 0.668749988079071,
|
| 608 |
+
"rewards/chosen": -0.46262964606285095,
|
| 609 |
+
"rewards/margins": 0.2319665402173996,
|
| 610 |
+
"rewards/rejected": -0.6945961713790894,
|
| 611 |
+
"step": 420
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.69,
|
| 615 |
+
"learning_rate": 1.3439071446815452e-06,
|
| 616 |
+
"logits/chosen": -0.11295236647129059,
|
| 617 |
+
"logits/rejected": -0.08282504230737686,
|
| 618 |
+
"logps/chosen": -887.32421875,
|
| 619 |
+
"logps/rejected": -1047.8634033203125,
|
| 620 |
+
"loss": 1.6503,
|
| 621 |
+
"rewards/accuracies": 0.65625,
|
| 622 |
+
"rewards/chosen": -0.43825840950012207,
|
| 623 |
+
"rewards/margins": 0.21870870888233185,
|
| 624 |
+
"rewards/rejected": -0.656967043876648,
|
| 625 |
+
"step": 430
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.7,
|
| 629 |
+
"learning_rate": 1.2218673716356919e-06,
|
| 630 |
+
"logits/chosen": -0.07290169596672058,
|
| 631 |
+
"logits/rejected": -0.06002408266067505,
|
| 632 |
+
"logps/chosen": -929.90380859375,
|
| 633 |
+
"logps/rejected": -1127.547119140625,
|
| 634 |
+
"loss": 1.7147,
|
| 635 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 636 |
+
"rewards/chosen": -0.4546731114387512,
|
| 637 |
+
"rewards/margins": 0.27696481347084045,
|
| 638 |
+
"rewards/rejected": -0.7316379547119141,
|
| 639 |
+
"step": 440
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.72,
|
| 643 |
+
"learning_rate": 1.103820512452661e-06,
|
| 644 |
+
"logits/chosen": -0.10286526381969452,
|
| 645 |
+
"logits/rejected": -0.071159228682518,
|
| 646 |
+
"logps/chosen": -874.0986328125,
|
| 647 |
+
"logps/rejected": -1238.300048828125,
|
| 648 |
+
"loss": 1.6778,
|
| 649 |
+
"rewards/accuracies": 0.699999988079071,
|
| 650 |
+
"rewards/chosen": -0.46896892786026,
|
| 651 |
+
"rewards/margins": 0.3574833273887634,
|
| 652 |
+
"rewards/rejected": -0.8264522552490234,
|
| 653 |
+
"step": 450
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.74,
|
| 657 |
+
"learning_rate": 9.901353480633468e-07,
|
| 658 |
+
"logits/chosen": -0.05727295204997063,
|
| 659 |
+
"logits/rejected": -0.08084852248430252,
|
| 660 |
+
"logps/chosen": -864.6730346679688,
|
| 661 |
+
"logps/rejected": -1181.377685546875,
|
| 662 |
+
"loss": 1.6027,
|
| 663 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 664 |
+
"rewards/chosen": -0.4679865837097168,
|
| 665 |
+
"rewards/margins": 0.3022565245628357,
|
| 666 |
+
"rewards/rejected": -0.7702431082725525,
|
| 667 |
+
"step": 460
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.75,
|
| 671 |
+
"learning_rate": 8.811670333701544e-07,
|
| 672 |
+
"logits/chosen": -0.05785631015896797,
|
| 673 |
+
"logits/rejected": -0.03484424576163292,
|
| 674 |
+
"logps/chosen": -947.6015625,
|
| 675 |
+
"logps/rejected": -1211.000732421875,
|
| 676 |
+
"loss": 1.5945,
|
| 677 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 678 |
+
"rewards/chosen": -0.4864828586578369,
|
| 679 |
+
"rewards/margins": 0.31723761558532715,
|
| 680 |
+
"rewards/rejected": -0.8037204742431641,
|
| 681 |
+
"step": 470
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.77,
|
| 685 |
+
"learning_rate": 7.772559877354341e-07,
|
| 686 |
+
"logits/chosen": -0.08603869378566742,
|
| 687 |
+
"logits/rejected": -0.0594349130988121,
|
| 688 |
+
"logps/chosen": -976.7135009765625,
|
| 689 |
+
"logps/rejected": -1181.75,
|
| 690 |
+
"loss": 1.7634,
|
| 691 |
+
"rewards/accuracies": 0.606249988079071,
|
| 692 |
+
"rewards/chosen": -0.5735828280448914,
|
| 693 |
+
"rewards/margins": 0.23867933452129364,
|
| 694 |
+
"rewards/rejected": -0.8122621774673462,
|
| 695 |
+
"step": 480
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.78,
|
| 699 |
+
"learning_rate": 6.787268315040604e-07,
|
| 700 |
+
"logits/chosen": -0.1035851389169693,
|
| 701 |
+
"logits/rejected": -0.10194051265716553,
|
| 702 |
+
"logps/chosen": -891.7512817382812,
|
| 703 |
+
"logps/rejected": -1107.476806640625,
|
| 704 |
+
"loss": 1.6117,
|
| 705 |
+
"rewards/accuracies": 0.675000011920929,
|
| 706 |
+
"rewards/chosen": -0.48722830414772034,
|
| 707 |
+
"rewards/margins": 0.2562137544155121,
|
| 708 |
+
"rewards/rejected": -0.7434421181678772,
|
| 709 |
+
"step": 490
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.8,
|
| 713 |
+
"learning_rate": 5.858873718824829e-07,
|
| 714 |
+
"logits/chosen": -0.06155969947576523,
|
| 715 |
+
"logits/rejected": -0.018813952803611755,
|
| 716 |
+
"logps/chosen": -938.9776611328125,
|
| 717 |
+
"logps/rejected": -1191.833740234375,
|
| 718 |
+
"loss": 1.6467,
|
| 719 |
+
"rewards/accuracies": 0.737500011920929,
|
| 720 |
+
"rewards/chosen": -0.5143369436264038,
|
| 721 |
+
"rewards/margins": 0.2735580801963806,
|
| 722 |
+
"rewards/rejected": -0.7878950834274292,
|
| 723 |
+
"step": 500
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.82,
|
| 727 |
+
"learning_rate": 4.990276413423817e-07,
|
| 728 |
+
"logits/chosen": -0.06636036932468414,
|
| 729 |
+
"logits/rejected": -0.10939071327447891,
|
| 730 |
+
"logps/chosen": -859.03515625,
|
| 731 |
+
"logps/rejected": -1074.6214599609375,
|
| 732 |
+
"loss": 1.7273,
|
| 733 |
+
"rewards/accuracies": 0.612500011920929,
|
| 734 |
+
"rewards/chosen": -0.480158269405365,
|
| 735 |
+
"rewards/margins": 0.23960845172405243,
|
| 736 |
+
"rewards/rejected": -0.7197667360305786,
|
| 737 |
+
"step": 510
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.83,
|
| 741 |
+
"learning_rate": 4.184189915529796e-07,
|
| 742 |
+
"logits/chosen": -0.049408238381147385,
|
| 743 |
+
"logits/rejected": -0.06695321202278137,
|
| 744 |
+
"logps/chosen": -991.7838134765625,
|
| 745 |
+
"logps/rejected": -1224.253173828125,
|
| 746 |
+
"loss": 1.6447,
|
| 747 |
+
"rewards/accuracies": 0.675000011920929,
|
| 748 |
+
"rewards/chosen": -0.529580295085907,
|
| 749 |
+
"rewards/margins": 0.23929505050182343,
|
| 750 |
+
"rewards/rejected": -0.7688754200935364,
|
| 751 |
+
"step": 520
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.85,
|
| 755 |
+
"learning_rate": 3.4431324567258176e-07,
|
| 756 |
+
"logits/chosen": -0.0758766159415245,
|
| 757 |
+
"logits/rejected": -0.06797490268945694,
|
| 758 |
+
"logps/chosen": -894.0300903320312,
|
| 759 |
+
"logps/rejected": -1302.058837890625,
|
| 760 |
+
"loss": 1.654,
|
| 761 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 762 |
+
"rewards/chosen": -0.4872868061065674,
|
| 763 |
+
"rewards/margins": 0.3845263421535492,
|
| 764 |
+
"rewards/rejected": -0.8718131184577942,
|
| 765 |
+
"step": 530
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.86,
|
| 769 |
+
"learning_rate": 2.769419116476052e-07,
|
| 770 |
+
"logits/chosen": -0.12563803791999817,
|
| 771 |
+
"logits/rejected": -0.09884506464004517,
|
| 772 |
+
"logps/chosen": -938.1819458007812,
|
| 773 |
+
"logps/rejected": -1275.197021484375,
|
| 774 |
+
"loss": 1.6659,
|
| 775 |
+
"rewards/accuracies": 0.699999988079071,
|
| 776 |
+
"rewards/chosen": -0.5264378786087036,
|
| 777 |
+
"rewards/margins": 0.3382587432861328,
|
| 778 |
+
"rewards/rejected": -0.8646965026855469,
|
| 779 |
+
"step": 540
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.88,
|
| 783 |
+
"learning_rate": 2.1651545897676512e-07,
|
| 784 |
+
"logits/chosen": -0.11725203692913055,
|
| 785 |
+
"logits/rejected": -0.11871466785669327,
|
| 786 |
+
"logps/chosen": -898.6276245117188,
|
| 787 |
+
"logps/rejected": -1201.9349365234375,
|
| 788 |
+
"loss": 1.747,
|
| 789 |
+
"rewards/accuracies": 0.668749988079071,
|
| 790 |
+
"rewards/chosen": -0.5061923861503601,
|
| 791 |
+
"rewards/margins": 0.3278709053993225,
|
| 792 |
+
"rewards/rejected": -0.8340633511543274,
|
| 793 |
+
"step": 550
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.9,
|
| 797 |
+
"learning_rate": 1.6322266119983222e-07,
|
| 798 |
+
"logits/chosen": -0.04045483097434044,
|
| 799 |
+
"logits/rejected": -0.05498255044221878,
|
| 800 |
+
"logps/chosen": -983.4290771484375,
|
| 801 |
+
"logps/rejected": -1337.211181640625,
|
| 802 |
+
"loss": 1.5444,
|
| 803 |
+
"rewards/accuracies": 0.731249988079071,
|
| 804 |
+
"rewards/chosen": -0.5375570058822632,
|
| 805 |
+
"rewards/margins": 0.3453003168106079,
|
| 806 |
+
"rewards/rejected": -0.8828572034835815,
|
| 807 |
+
"step": 560
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.91,
|
| 811 |
+
"learning_rate": 1.1723000616502167e-07,
|
| 812 |
+
"logits/chosen": -0.09810743480920792,
|
| 813 |
+
"logits/rejected": -0.09926000982522964,
|
| 814 |
+
"logps/chosen": -835.6412353515625,
|
| 815 |
+
"logps/rejected": -1187.4744873046875,
|
| 816 |
+
"loss": 1.794,
|
| 817 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 818 |
+
"rewards/chosen": -0.46745967864990234,
|
| 819 |
+
"rewards/margins": 0.3435118794441223,
|
| 820 |
+
"rewards/rejected": -0.8109715580940247,
|
| 821 |
+
"step": 570
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.93,
|
| 825 |
+
"learning_rate": 7.868117591737585e-08,
|
| 826 |
+
"logits/chosen": -0.09739083796739578,
|
| 827 |
+
"logits/rejected": -0.10297133773565292,
|
| 828 |
+
"logps/chosen": -885.7447509765625,
|
| 829 |
+
"logps/rejected": -1204.79443359375,
|
| 830 |
+
"loss": 1.6582,
|
| 831 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 832 |
+
"rewards/chosen": -0.5217578411102295,
|
| 833 |
+
"rewards/margins": 0.28710317611694336,
|
| 834 |
+
"rewards/rejected": -0.8088610768318176,
|
| 835 |
+
"step": 580
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.94,
|
| 839 |
+
"learning_rate": 4.769659783295383e-08,
|
| 840 |
+
"logits/chosen": -0.11744657903909683,
|
| 841 |
+
"logits/rejected": -0.07945755124092102,
|
| 842 |
+
"logps/chosen": -882.5567626953125,
|
| 843 |
+
"logps/rejected": -1246.40234375,
|
| 844 |
+
"loss": 1.6835,
|
| 845 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 846 |
+
"rewards/chosen": -0.487631618976593,
|
| 847 |
+
"rewards/margins": 0.3489754796028137,
|
| 848 |
+
"rewards/rejected": -0.8366070985794067,
|
| 849 |
+
"step": 590
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.96,
|
| 853 |
+
"learning_rate": 2.4373068401120358e-08,
|
| 854 |
+
"logits/chosen": -0.08680951595306396,
|
| 855 |
+
"logits/rejected": -0.07239434868097305,
|
| 856 |
+
"logps/chosen": -865.5779418945312,
|
| 857 |
+
"logps/rejected": -1108.098388671875,
|
| 858 |
+
"loss": 1.7686,
|
| 859 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 860 |
+
"rewards/chosen": -0.48516979813575745,
|
| 861 |
+
"rewards/margins": 0.2544567883014679,
|
| 862 |
+
"rewards/rejected": -0.7396265864372253,
|
| 863 |
+
"step": 600
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.98,
|
| 867 |
+
"learning_rate": 8.78345083022425e-09,
|
| 868 |
+
"logits/chosen": -0.09715618193149567,
|
| 869 |
+
"logits/rejected": -0.08691917359828949,
|
| 870 |
+
"logps/chosen": -904.9318237304688,
|
| 871 |
+
"logps/rejected": -1211.5076904296875,
|
| 872 |
+
"loss": 1.7077,
|
| 873 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 874 |
+
"rewards/chosen": -0.5081444978713989,
|
| 875 |
+
"rewards/margins": 0.2958123981952667,
|
| 876 |
+
"rewards/rejected": -0.8039568662643433,
|
| 877 |
+
"step": 610
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.99,
|
| 881 |
+
"learning_rate": 9.764474213677654e-10,
|
| 882 |
+
"logits/chosen": -0.05566658824682236,
|
| 883 |
+
"logits/rejected": -0.07036083936691284,
|
| 884 |
+
"logps/chosen": -910.61767578125,
|
| 885 |
+
"logps/rejected": -1149.9638671875,
|
| 886 |
+
"loss": 1.6468,
|
| 887 |
+
"rewards/accuracies": 0.65625,
|
| 888 |
+
"rewards/chosen": -0.5058663487434387,
|
| 889 |
+
"rewards/margins": 0.25643160939216614,
|
| 890 |
+
"rewards/rejected": -0.7622979879379272,
|
| 891 |
+
"step": 620
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 1.0,
|
| 895 |
+
"step": 625,
|
| 896 |
+
"total_flos": 0.0,
|
| 897 |
+
"train_loss": 1.7737385009765625,
|
| 898 |
+
"train_runtime": 8419.2459,
|
| 899 |
+
"train_samples_per_second": 3.563,
|
| 900 |
+
"train_steps_per_second": 0.074
|
| 901 |
+
}
|
| 902 |
+
],
|
| 903 |
+
"logging_steps": 10,
|
| 904 |
+
"max_steps": 625,
|
| 905 |
+
"num_input_tokens_seen": 0,
|
| 906 |
+
"num_train_epochs": 1,
|
| 907 |
+
"save_steps": 100,
|
| 908 |
+
"total_flos": 0.0,
|
| 909 |
+
"train_batch_size": 4,
|
| 910 |
+
"trial_name": null,
|
| 911 |
+
"trial_params": null
|
| 912 |
+
}
|