Model save
Browse files- README.md +60 -0
- adapter_model.safetensors +1 -1
- all_results.json +8 -0
- runs/May15_01-01-24_gpu4-119-5/events.out.tfevents.1715698968.gpu4-119-5.2148799.0 +2 -2
- train_results.json +8 -0
- trainer_state.json +884 -0
README.md
ADDED
|
@@ -0,0 +1,60 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
library_name: peft
|
| 4 |
+
tags:
|
| 5 |
+
- trl
|
| 6 |
+
- dpo
|
| 7 |
+
- generated_from_trainer
|
| 8 |
+
base_model: DUAL-GPO/phi-2-gpo-new-i0
|
| 9 |
+
model-index:
|
| 10 |
+
- name: phi-2-gpo-v20-i1
|
| 11 |
+
results: []
|
| 12 |
+
---
|
| 13 |
+
|
| 14 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 15 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 16 |
+
|
| 17 |
+
# phi-2-gpo-v20-i1
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [DUAL-GPO/phi-2-gpo-new-i0](https://huggingface.co/DUAL-GPO/phi-2-gpo-new-i0) on the None dataset.
|
| 20 |
+
|
| 21 |
+
## Model description
|
| 22 |
+
|
| 23 |
+
More information needed
|
| 24 |
+
|
| 25 |
+
## Intended uses & limitations
|
| 26 |
+
|
| 27 |
+
More information needed
|
| 28 |
+
|
| 29 |
+
## Training and evaluation data
|
| 30 |
+
|
| 31 |
+
More information needed
|
| 32 |
+
|
| 33 |
+
## Training procedure
|
| 34 |
+
|
| 35 |
+
### Training hyperparameters
|
| 36 |
+
|
| 37 |
+
The following hyperparameters were used during training:
|
| 38 |
+
- learning_rate: 5e-06
|
| 39 |
+
- train_batch_size: 4
|
| 40 |
+
- eval_batch_size: 4
|
| 41 |
+
- seed: 42
|
| 42 |
+
- distributed_type: multi-GPU
|
| 43 |
+
- gradient_accumulation_steps: 4
|
| 44 |
+
- total_train_batch_size: 16
|
| 45 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 46 |
+
- lr_scheduler_type: cosine
|
| 47 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 48 |
+
- num_epochs: 1
|
| 49 |
+
|
| 50 |
+
### Training results
|
| 51 |
+
|
| 52 |
+
|
| 53 |
+
|
| 54 |
+
### Framework versions
|
| 55 |
+
|
| 56 |
+
- PEFT 0.7.1
|
| 57 |
+
- Transformers 4.36.2
|
| 58 |
+
- Pytorch 2.1.2+cu121
|
| 59 |
+
- Datasets 2.14.6
|
| 60 |
+
- 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 sha256:1427b2200d950b3c6817d647ae52086b1d13da4be617ca1f7da52d6446e39f72
|
| 3 |
size 167807296
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.3152459278702736,
|
| 4 |
+
"train_runtime": 5630.9992,
|
| 5 |
+
"train_samples": 9600,
|
| 6 |
+
"train_samples_per_second": 1.705,
|
| 7 |
+
"train_steps_per_second": 0.107
|
| 8 |
+
}
|
runs/May15_01-01-24_gpu4-119-5/events.out.tfevents.1715698968.gpu4-119-5.2148799.0
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 sha256:438278cee20ac92fc6ef15337a15b342e3fb7e6993ab9a8052c9976c782aa310
|
| 3 |
+
size 43712
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.3152459278702736,
|
| 4 |
+
"train_runtime": 5630.9992,
|
| 5 |
+
"train_samples": 9600,
|
| 6 |
+
"train_samples_per_second": 1.705,
|
| 7 |
+
"train_steps_per_second": 0.107
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,884 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 600,
|
| 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": 8.333333333333334e-08,
|
| 14 |
+
"logits/chosen": 0.1956433206796646,
|
| 15 |
+
"logits/rejected": 0.2183472067117691,
|
| 16 |
+
"logps/chosen": -380.8453369140625,
|
| 17 |
+
"logps/rejected": -356.37103271484375,
|
| 18 |
+
"loss": 0.409,
|
| 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": 8.333333333333333e-07,
|
| 28 |
+
"logits/chosen": 0.09696108102798462,
|
| 29 |
+
"logits/rejected": 0.21700751781463623,
|
| 30 |
+
"logps/chosen": -387.7613525390625,
|
| 31 |
+
"logps/rejected": -328.28912353515625,
|
| 32 |
+
"loss": 0.3709,
|
| 33 |
+
"rewards/accuracies": 0.4236111044883728,
|
| 34 |
+
"rewards/chosen": 2.4463388399453834e-05,
|
| 35 |
+
"rewards/margins": 3.930643651983701e-05,
|
| 36 |
+
"rewards/rejected": -1.4843050848867279e-05,
|
| 37 |
+
"step": 10
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.03,
|
| 41 |
+
"learning_rate": 1.6666666666666667e-06,
|
| 42 |
+
"logits/chosen": 0.13495060801506042,
|
| 43 |
+
"logits/rejected": 0.22846023738384247,
|
| 44 |
+
"logps/chosen": -338.4651184082031,
|
| 45 |
+
"logps/rejected": -336.63897705078125,
|
| 46 |
+
"loss": 0.3742,
|
| 47 |
+
"rewards/accuracies": 0.4312500059604645,
|
| 48 |
+
"rewards/chosen": 6.384013977367431e-05,
|
| 49 |
+
"rewards/margins": -0.00011528225149959326,
|
| 50 |
+
"rewards/rejected": 0.00017912239127326757,
|
| 51 |
+
"step": 20
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.05,
|
| 55 |
+
"learning_rate": 2.5e-06,
|
| 56 |
+
"logits/chosen": 0.1001209244132042,
|
| 57 |
+
"logits/rejected": 0.2776273190975189,
|
| 58 |
+
"logps/chosen": -342.38140869140625,
|
| 59 |
+
"logps/rejected": -354.812744140625,
|
| 60 |
+
"loss": 0.3626,
|
| 61 |
+
"rewards/accuracies": 0.4312500059604645,
|
| 62 |
+
"rewards/chosen": 3.7229918234515935e-05,
|
| 63 |
+
"rewards/margins": 9.480830340180546e-05,
|
| 64 |
+
"rewards/rejected": -5.757838880526833e-05,
|
| 65 |
+
"step": 30
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.07,
|
| 69 |
+
"learning_rate": 3.3333333333333333e-06,
|
| 70 |
+
"logits/chosen": 0.12598159909248352,
|
| 71 |
+
"logits/rejected": 0.21262428164482117,
|
| 72 |
+
"logps/chosen": -343.973876953125,
|
| 73 |
+
"logps/rejected": -333.87969970703125,
|
| 74 |
+
"loss": 0.3845,
|
| 75 |
+
"rewards/accuracies": 0.5687500238418579,
|
| 76 |
+
"rewards/chosen": 9.938174480339512e-05,
|
| 77 |
+
"rewards/margins": 0.00036102396552450955,
|
| 78 |
+
"rewards/rejected": -0.0002616422134451568,
|
| 79 |
+
"step": 40
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.08,
|
| 83 |
+
"learning_rate": 4.166666666666667e-06,
|
| 84 |
+
"logits/chosen": 0.1064462661743164,
|
| 85 |
+
"logits/rejected": 0.1775200366973877,
|
| 86 |
+
"logps/chosen": -351.95745849609375,
|
| 87 |
+
"logps/rejected": -328.10955810546875,
|
| 88 |
+
"loss": 0.3802,
|
| 89 |
+
"rewards/accuracies": 0.543749988079071,
|
| 90 |
+
"rewards/chosen": 0.00037883114418946207,
|
| 91 |
+
"rewards/margins": 0.0006485713529400527,
|
| 92 |
+
"rewards/rejected": -0.0002697401796467602,
|
| 93 |
+
"step": 50
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.1,
|
| 97 |
+
"learning_rate": 5e-06,
|
| 98 |
+
"logits/chosen": 0.10294230282306671,
|
| 99 |
+
"logits/rejected": 0.15821890532970428,
|
| 100 |
+
"logps/chosen": -332.59686279296875,
|
| 101 |
+
"logps/rejected": -317.3880310058594,
|
| 102 |
+
"loss": 0.3864,
|
| 103 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 104 |
+
"rewards/chosen": 0.0005501825944520533,
|
| 105 |
+
"rewards/margins": 0.0013122591190040112,
|
| 106 |
+
"rewards/rejected": -0.0007620764663442969,
|
| 107 |
+
"step": 60
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.12,
|
| 111 |
+
"learning_rate": 4.995770395678171e-06,
|
| 112 |
+
"logits/chosen": 0.12952658534049988,
|
| 113 |
+
"logits/rejected": 0.21082790195941925,
|
| 114 |
+
"logps/chosen": -337.88616943359375,
|
| 115 |
+
"logps/rejected": -337.0188903808594,
|
| 116 |
+
"loss": 0.3761,
|
| 117 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 118 |
+
"rewards/chosen": 0.0009980153990909457,
|
| 119 |
+
"rewards/margins": 0.002471885411068797,
|
| 120 |
+
"rewards/rejected": -0.0014738701283931732,
|
| 121 |
+
"step": 70
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.13,
|
| 125 |
+
"learning_rate": 4.983095894354858e-06,
|
| 126 |
+
"logits/chosen": 0.09408712387084961,
|
| 127 |
+
"logits/rejected": 0.1505356878042221,
|
| 128 |
+
"logps/chosen": -326.36724853515625,
|
| 129 |
+
"logps/rejected": -358.09759521484375,
|
| 130 |
+
"loss": 0.3757,
|
| 131 |
+
"rewards/accuracies": 0.668749988079071,
|
| 132 |
+
"rewards/chosen": 0.0015949418302625418,
|
| 133 |
+
"rewards/margins": 0.004127982072532177,
|
| 134 |
+
"rewards/rejected": -0.0025330407079309225,
|
| 135 |
+
"step": 80
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.15,
|
| 139 |
+
"learning_rate": 4.962019382530521e-06,
|
| 140 |
+
"logits/chosen": 0.11112242937088013,
|
| 141 |
+
"logits/rejected": 0.1587546467781067,
|
| 142 |
+
"logps/chosen": -337.84051513671875,
|
| 143 |
+
"logps/rejected": -348.18634033203125,
|
| 144 |
+
"loss": 0.3735,
|
| 145 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 146 |
+
"rewards/chosen": 0.0002811970771290362,
|
| 147 |
+
"rewards/margins": 0.005385980941355228,
|
| 148 |
+
"rewards/rejected": -0.0051047843880951405,
|
| 149 |
+
"step": 90
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.17,
|
| 153 |
+
"learning_rate": 4.93261217644956e-06,
|
| 154 |
+
"logits/chosen": 0.13481362164020538,
|
| 155 |
+
"logits/rejected": 0.19625934958457947,
|
| 156 |
+
"logps/chosen": -347.8192443847656,
|
| 157 |
+
"logps/rejected": -345.80511474609375,
|
| 158 |
+
"loss": 0.3628,
|
| 159 |
+
"rewards/accuracies": 0.637499988079071,
|
| 160 |
+
"rewards/chosen": -0.0003084948402829468,
|
| 161 |
+
"rewards/margins": 0.01107403077185154,
|
| 162 |
+
"rewards/rejected": -0.01138252578675747,
|
| 163 |
+
"step": 100
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.18,
|
| 167 |
+
"learning_rate": 4.894973780788722e-06,
|
| 168 |
+
"logits/chosen": 0.10719671100378036,
|
| 169 |
+
"logits/rejected": 0.2091417759656906,
|
| 170 |
+
"logps/chosen": -347.8533630371094,
|
| 171 |
+
"logps/rejected": -352.93115234375,
|
| 172 |
+
"loss": 0.3404,
|
| 173 |
+
"rewards/accuracies": 0.6875,
|
| 174 |
+
"rewards/chosen": -0.0032309088855981827,
|
| 175 |
+
"rewards/margins": 0.021253572776913643,
|
| 176 |
+
"rewards/rejected": -0.024484481662511826,
|
| 177 |
+
"step": 110
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.2,
|
| 181 |
+
"learning_rate": 4.849231551964771e-06,
|
| 182 |
+
"logits/chosen": 0.13324591517448425,
|
| 183 |
+
"logits/rejected": 0.1466611921787262,
|
| 184 |
+
"logps/chosen": -377.8055114746094,
|
| 185 |
+
"logps/rejected": -370.2322082519531,
|
| 186 |
+
"loss": 0.3684,
|
| 187 |
+
"rewards/accuracies": 0.606249988079071,
|
| 188 |
+
"rewards/chosen": -0.022458815947175026,
|
| 189 |
+
"rewards/margins": 0.015764199197292328,
|
| 190 |
+
"rewards/rejected": -0.0382230170071125,
|
| 191 |
+
"step": 120
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.22,
|
| 195 |
+
"learning_rate": 4.7955402672006855e-06,
|
| 196 |
+
"logits/chosen": 0.11208464950323105,
|
| 197 |
+
"logits/rejected": 0.21028931438922882,
|
| 198 |
+
"logps/chosen": -390.74896240234375,
|
| 199 |
+
"logps/rejected": -424.40826416015625,
|
| 200 |
+
"loss": 0.3223,
|
| 201 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 202 |
+
"rewards/chosen": -0.043638426810503006,
|
| 203 |
+
"rewards/margins": 0.047044601291418076,
|
| 204 |
+
"rewards/rejected": -0.09068302810192108,
|
| 205 |
+
"step": 130
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.23,
|
| 209 |
+
"learning_rate": 4.734081600808531e-06,
|
| 210 |
+
"logits/chosen": 0.16790533065795898,
|
| 211 |
+
"logits/rejected": 0.20132648944854736,
|
| 212 |
+
"logps/chosen": -450.4337463378906,
|
| 213 |
+
"logps/rejected": -493.0623474121094,
|
| 214 |
+
"loss": 0.3234,
|
| 215 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 216 |
+
"rewards/chosen": -0.11827051639556885,
|
| 217 |
+
"rewards/margins": 0.043106190860271454,
|
| 218 |
+
"rewards/rejected": -0.1613767296075821,
|
| 219 |
+
"step": 140
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.25,
|
| 223 |
+
"learning_rate": 4.665063509461098e-06,
|
| 224 |
+
"logits/chosen": 0.1597878634929657,
|
| 225 |
+
"logits/rejected": 0.22755786776542664,
|
| 226 |
+
"logps/chosen": -539.1220703125,
|
| 227 |
+
"logps/rejected": -538.5697021484375,
|
| 228 |
+
"loss": 0.3578,
|
| 229 |
+
"rewards/accuracies": 0.65625,
|
| 230 |
+
"rewards/chosen": -0.17835457623004913,
|
| 231 |
+
"rewards/margins": 0.05051771551370621,
|
| 232 |
+
"rewards/rejected": -0.22887229919433594,
|
| 233 |
+
"step": 150
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.27,
|
| 237 |
+
"learning_rate": 4.588719528532342e-06,
|
| 238 |
+
"logits/chosen": 0.19429577887058258,
|
| 239 |
+
"logits/rejected": 0.2786257565021515,
|
| 240 |
+
"logps/chosen": -445.50531005859375,
|
| 241 |
+
"logps/rejected": -506.572998046875,
|
| 242 |
+
"loss": 0.3039,
|
| 243 |
+
"rewards/accuracies": 0.668749988079071,
|
| 244 |
+
"rewards/chosen": -0.1227174624800682,
|
| 245 |
+
"rewards/margins": 0.07187845557928085,
|
| 246 |
+
"rewards/rejected": -0.19459593296051025,
|
| 247 |
+
"step": 160
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.28,
|
| 251 |
+
"learning_rate": 4.50530798188761e-06,
|
| 252 |
+
"logits/chosen": 0.15851017832756042,
|
| 253 |
+
"logits/rejected": 0.28014975786209106,
|
| 254 |
+
"logps/chosen": -546.3292236328125,
|
| 255 |
+
"logps/rejected": -604.9276123046875,
|
| 256 |
+
"loss": 0.32,
|
| 257 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 258 |
+
"rewards/chosen": -0.15538683533668518,
|
| 259 |
+
"rewards/margins": 0.07926145941019058,
|
| 260 |
+
"rewards/rejected": -0.23464830219745636,
|
| 261 |
+
"step": 170
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.3,
|
| 265 |
+
"learning_rate": 4.415111107797445e-06,
|
| 266 |
+
"logits/chosen": 0.16010740399360657,
|
| 267 |
+
"logits/rejected": 0.2813720405101776,
|
| 268 |
+
"logps/chosen": -585.3643798828125,
|
| 269 |
+
"logps/rejected": -665.8712158203125,
|
| 270 |
+
"loss": 0.3048,
|
| 271 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 272 |
+
"rewards/chosen": -0.2374914586544037,
|
| 273 |
+
"rewards/margins": 0.11958248913288116,
|
| 274 |
+
"rewards/rejected": -0.35707396268844604,
|
| 275 |
+
"step": 180
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.32,
|
| 279 |
+
"learning_rate": 4.318434103932622e-06,
|
| 280 |
+
"logits/chosen": 0.20392921566963196,
|
| 281 |
+
"logits/rejected": 0.2997474670410156,
|
| 282 |
+
"logps/chosen": -593.7650146484375,
|
| 283 |
+
"logps/rejected": -700.028076171875,
|
| 284 |
+
"loss": 0.3191,
|
| 285 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 286 |
+
"rewards/chosen": -0.27146199345588684,
|
| 287 |
+
"rewards/margins": 0.11621884256601334,
|
| 288 |
+
"rewards/rejected": -0.3876808285713196,
|
| 289 |
+
"step": 190
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.33,
|
| 293 |
+
"learning_rate": 4.215604094671835e-06,
|
| 294 |
+
"logits/chosen": 0.18635989725589752,
|
| 295 |
+
"logits/rejected": 0.25986188650131226,
|
| 296 |
+
"logps/chosen": -575.0554809570312,
|
| 297 |
+
"logps/rejected": -709.1888427734375,
|
| 298 |
+
"loss": 0.2834,
|
| 299 |
+
"rewards/accuracies": 0.699999988079071,
|
| 300 |
+
"rewards/chosen": -0.22922906279563904,
|
| 301 |
+
"rewards/margins": 0.12650123238563538,
|
| 302 |
+
"rewards/rejected": -0.3557303547859192,
|
| 303 |
+
"step": 200
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.35,
|
| 307 |
+
"learning_rate": 4.106969024216348e-06,
|
| 308 |
+
"logits/chosen": 0.21329033374786377,
|
| 309 |
+
"logits/rejected": 0.35237234830856323,
|
| 310 |
+
"logps/chosen": -569.5208129882812,
|
| 311 |
+
"logps/rejected": -619.1517944335938,
|
| 312 |
+
"loss": 0.3111,
|
| 313 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 314 |
+
"rewards/chosen": -0.1869223415851593,
|
| 315 |
+
"rewards/margins": 0.08501409739255905,
|
| 316 |
+
"rewards/rejected": -0.27193647623062134,
|
| 317 |
+
"step": 210
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.37,
|
| 321 |
+
"learning_rate": 3.992896479256966e-06,
|
| 322 |
+
"logits/chosen": 0.22448810935020447,
|
| 323 |
+
"logits/rejected": 0.2712326645851135,
|
| 324 |
+
"logps/chosen": -574.156982421875,
|
| 325 |
+
"logps/rejected": -612.8103637695312,
|
| 326 |
+
"loss": 0.3201,
|
| 327 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 328 |
+
"rewards/chosen": -0.1946686953306198,
|
| 329 |
+
"rewards/margins": 0.07545152306556702,
|
| 330 |
+
"rewards/rejected": -0.270120233297348,
|
| 331 |
+
"step": 220
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.38,
|
| 335 |
+
"learning_rate": 3.8737724451770155e-06,
|
| 336 |
+
"logits/chosen": 0.21266920864582062,
|
| 337 |
+
"logits/rejected": 0.241852805018425,
|
| 338 |
+
"logps/chosen": -489.32073974609375,
|
| 339 |
+
"logps/rejected": -567.0352783203125,
|
| 340 |
+
"loss": 0.3088,
|
| 341 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 342 |
+
"rewards/chosen": -0.18188850581645966,
|
| 343 |
+
"rewards/margins": 0.09176473319530487,
|
| 344 |
+
"rewards/rejected": -0.2736532390117645,
|
| 345 |
+
"step": 230
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.4,
|
| 349 |
+
"learning_rate": 3.7500000000000005e-06,
|
| 350 |
+
"logits/chosen": 0.21173226833343506,
|
| 351 |
+
"logits/rejected": 0.27939146757125854,
|
| 352 |
+
"logps/chosen": -584.0433349609375,
|
| 353 |
+
"logps/rejected": -665.869873046875,
|
| 354 |
+
"loss": 0.2928,
|
| 355 |
+
"rewards/accuracies": 0.668749988079071,
|
| 356 |
+
"rewards/chosen": -0.22605593502521515,
|
| 357 |
+
"rewards/margins": 0.10531251132488251,
|
| 358 |
+
"rewards/rejected": -0.33136844635009766,
|
| 359 |
+
"step": 240
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.42,
|
| 363 |
+
"learning_rate": 3.621997950501156e-06,
|
| 364 |
+
"logits/chosen": 0.26176512241363525,
|
| 365 |
+
"logits/rejected": 0.28036120533943176,
|
| 366 |
+
"logps/chosen": -581.51318359375,
|
| 367 |
+
"logps/rejected": -680.9810791015625,
|
| 368 |
+
"loss": 0.3056,
|
| 369 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 370 |
+
"rewards/chosen": -0.2372002899646759,
|
| 371 |
+
"rewards/margins": 0.08836118876934052,
|
| 372 |
+
"rewards/rejected": -0.3255614936351776,
|
| 373 |
+
"step": 250
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.43,
|
| 377 |
+
"learning_rate": 3.4901994150978926e-06,
|
| 378 |
+
"logits/chosen": 0.18492260575294495,
|
| 379 |
+
"logits/rejected": 0.3097878098487854,
|
| 380 |
+
"logps/chosen": -620.8358154296875,
|
| 381 |
+
"logps/rejected": -666.6085815429688,
|
| 382 |
+
"loss": 0.3084,
|
| 383 |
+
"rewards/accuracies": 0.675000011920929,
|
| 384 |
+
"rewards/chosen": -0.24568219482898712,
|
| 385 |
+
"rewards/margins": 0.09036718308925629,
|
| 386 |
+
"rewards/rejected": -0.3360493779182434,
|
| 387 |
+
"step": 260
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.45,
|
| 391 |
+
"learning_rate": 3.3550503583141726e-06,
|
| 392 |
+
"logits/chosen": 0.22367191314697266,
|
| 393 |
+
"logits/rejected": 0.3482346534729004,
|
| 394 |
+
"logps/chosen": -595.0042114257812,
|
| 395 |
+
"logps/rejected": -694.3624877929688,
|
| 396 |
+
"loss": 0.294,
|
| 397 |
+
"rewards/accuracies": 0.65625,
|
| 398 |
+
"rewards/chosen": -0.23263101279735565,
|
| 399 |
+
"rewards/margins": 0.1014803871512413,
|
| 400 |
+
"rewards/rejected": -0.33411139249801636,
|
| 401 |
+
"step": 270
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.47,
|
| 405 |
+
"learning_rate": 3.217008081777726e-06,
|
| 406 |
+
"logits/chosen": 0.2801993191242218,
|
| 407 |
+
"logits/rejected": 0.31503865122795105,
|
| 408 |
+
"logps/chosen": -591.4317626953125,
|
| 409 |
+
"logps/rejected": -664.4125366210938,
|
| 410 |
+
"loss": 0.3276,
|
| 411 |
+
"rewards/accuracies": 0.65625,
|
| 412 |
+
"rewards/chosen": -0.24832281470298767,
|
| 413 |
+
"rewards/margins": 0.0828218162059784,
|
| 414 |
+
"rewards/rejected": -0.33114463090896606,
|
| 415 |
+
"step": 280
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.48,
|
| 419 |
+
"learning_rate": 3.0765396768561005e-06,
|
| 420 |
+
"logits/chosen": 0.23998506367206573,
|
| 421 |
+
"logits/rejected": 0.24142727255821228,
|
| 422 |
+
"logps/chosen": -615.881103515625,
|
| 423 |
+
"logps/rejected": -747.1182250976562,
|
| 424 |
+
"loss": 0.2889,
|
| 425 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 426 |
+
"rewards/chosen": -0.2754364013671875,
|
| 427 |
+
"rewards/margins": 0.12126360088586807,
|
| 428 |
+
"rewards/rejected": -0.3966999650001526,
|
| 429 |
+
"step": 290
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.5,
|
| 433 |
+
"learning_rate": 2.9341204441673267e-06,
|
| 434 |
+
"logits/chosen": 0.26361778378486633,
|
| 435 |
+
"logits/rejected": 0.31506967544555664,
|
| 436 |
+
"logps/chosen": -607.7698974609375,
|
| 437 |
+
"logps/rejected": -754.2791137695312,
|
| 438 |
+
"loss": 0.3049,
|
| 439 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 440 |
+
"rewards/chosen": -0.2881506085395813,
|
| 441 |
+
"rewards/margins": 0.12216220051050186,
|
| 442 |
+
"rewards/rejected": -0.41031280159950256,
|
| 443 |
+
"step": 300
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.52,
|
| 447 |
+
"learning_rate": 2.7902322853130758e-06,
|
| 448 |
+
"logits/chosen": 0.2686136066913605,
|
| 449 |
+
"logits/rejected": 0.32013028860092163,
|
| 450 |
+
"logps/chosen": -627.7980346679688,
|
| 451 |
+
"logps/rejected": -742.2149047851562,
|
| 452 |
+
"loss": 0.3474,
|
| 453 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 454 |
+
"rewards/chosen": -0.2792850434780121,
|
| 455 |
+
"rewards/margins": 0.09415403008460999,
|
| 456 |
+
"rewards/rejected": -0.3734390139579773,
|
| 457 |
+
"step": 310
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.53,
|
| 461 |
+
"learning_rate": 2.6453620722761897e-06,
|
| 462 |
+
"logits/chosen": 0.15263831615447998,
|
| 463 |
+
"logits/rejected": 0.2327421009540558,
|
| 464 |
+
"logps/chosen": -628.375,
|
| 465 |
+
"logps/rejected": -704.6365966796875,
|
| 466 |
+
"loss": 0.2827,
|
| 467 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 468 |
+
"rewards/chosen": -0.2568105161190033,
|
| 469 |
+
"rewards/margins": 0.10418572276830673,
|
| 470 |
+
"rewards/rejected": -0.360996276140213,
|
| 471 |
+
"step": 320
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.55,
|
| 475 |
+
"learning_rate": 2.5e-06,
|
| 476 |
+
"logits/chosen": 0.16394071280956268,
|
| 477 |
+
"logits/rejected": 0.21737909317016602,
|
| 478 |
+
"logps/chosen": -613.1915893554688,
|
| 479 |
+
"logps/rejected": -711.7744750976562,
|
| 480 |
+
"loss": 0.3134,
|
| 481 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 482 |
+
"rewards/chosen": -0.2661852240562439,
|
| 483 |
+
"rewards/margins": 0.10473362356424332,
|
| 484 |
+
"rewards/rejected": -0.370918869972229,
|
| 485 |
+
"step": 330
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.57,
|
| 489 |
+
"learning_rate": 2.3546379277238107e-06,
|
| 490 |
+
"logits/chosen": 0.21221554279327393,
|
| 491 |
+
"logits/rejected": 0.24156931042671204,
|
| 492 |
+
"logps/chosen": -627.9307861328125,
|
| 493 |
+
"logps/rejected": -728.6505126953125,
|
| 494 |
+
"loss": 0.3131,
|
| 495 |
+
"rewards/accuracies": 0.643750011920929,
|
| 496 |
+
"rewards/chosen": -0.28867658972740173,
|
| 497 |
+
"rewards/margins": 0.09915556758642197,
|
| 498 |
+
"rewards/rejected": -0.3878321349620819,
|
| 499 |
+
"step": 340
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.58,
|
| 503 |
+
"learning_rate": 2.2097677146869242e-06,
|
| 504 |
+
"logits/chosen": 0.11871862411499023,
|
| 505 |
+
"logits/rejected": 0.22523121535778046,
|
| 506 |
+
"logps/chosen": -647.56298828125,
|
| 507 |
+
"logps/rejected": -767.3802490234375,
|
| 508 |
+
"loss": 0.2886,
|
| 509 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 510 |
+
"rewards/chosen": -0.30263960361480713,
|
| 511 |
+
"rewards/margins": 0.11332044750452042,
|
| 512 |
+
"rewards/rejected": -0.41596007347106934,
|
| 513 |
+
"step": 350
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.6,
|
| 517 |
+
"learning_rate": 2.0658795558326745e-06,
|
| 518 |
+
"logits/chosen": 0.21589894592761993,
|
| 519 |
+
"logits/rejected": 0.2351434975862503,
|
| 520 |
+
"logps/chosen": -660.6466064453125,
|
| 521 |
+
"logps/rejected": -783.96875,
|
| 522 |
+
"loss": 0.303,
|
| 523 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 524 |
+
"rewards/chosen": -0.3056688904762268,
|
| 525 |
+
"rewards/margins": 0.12556755542755127,
|
| 526 |
+
"rewards/rejected": -0.4312364161014557,
|
| 527 |
+
"step": 360
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.62,
|
| 531 |
+
"learning_rate": 1.9234603231439e-06,
|
| 532 |
+
"logits/chosen": 0.18540482223033905,
|
| 533 |
+
"logits/rejected": 0.20940378308296204,
|
| 534 |
+
"logps/chosen": -629.5787963867188,
|
| 535 |
+
"logps/rejected": -739.3985595703125,
|
| 536 |
+
"loss": 0.2899,
|
| 537 |
+
"rewards/accuracies": 0.65625,
|
| 538 |
+
"rewards/chosen": -0.30902260541915894,
|
| 539 |
+
"rewards/margins": 0.10534413158893585,
|
| 540 |
+
"rewards/rejected": -0.41436678171157837,
|
| 541 |
+
"step": 370
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.63,
|
| 545 |
+
"learning_rate": 1.7829919182222752e-06,
|
| 546 |
+
"logits/chosen": 0.1524067372083664,
|
| 547 |
+
"logits/rejected": 0.23221668601036072,
|
| 548 |
+
"logps/chosen": -677.9874877929688,
|
| 549 |
+
"logps/rejected": -802.863037109375,
|
| 550 |
+
"loss": 0.2776,
|
| 551 |
+
"rewards/accuracies": 0.6875,
|
| 552 |
+
"rewards/chosen": -0.3159080147743225,
|
| 553 |
+
"rewards/margins": 0.14397332072257996,
|
| 554 |
+
"rewards/rejected": -0.4598813056945801,
|
| 555 |
+
"step": 380
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.65,
|
| 559 |
+
"learning_rate": 1.6449496416858285e-06,
|
| 560 |
+
"logits/chosen": 0.18434450030326843,
|
| 561 |
+
"logits/rejected": 0.24084284901618958,
|
| 562 |
+
"logps/chosen": -654.8961181640625,
|
| 563 |
+
"logps/rejected": -759.6239013671875,
|
| 564 |
+
"loss": 0.2896,
|
| 565 |
+
"rewards/accuracies": 0.637499988079071,
|
| 566 |
+
"rewards/chosen": -0.2936326861381531,
|
| 567 |
+
"rewards/margins": 0.12481866776943207,
|
| 568 |
+
"rewards/rejected": -0.41845136880874634,
|
| 569 |
+
"step": 390
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.67,
|
| 573 |
+
"learning_rate": 1.509800584902108e-06,
|
| 574 |
+
"logits/chosen": 0.1810900717973709,
|
| 575 |
+
"logits/rejected": 0.2414156198501587,
|
| 576 |
+
"logps/chosen": -614.9703979492188,
|
| 577 |
+
"logps/rejected": -757.6051025390625,
|
| 578 |
+
"loss": 0.326,
|
| 579 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 580 |
+
"rewards/chosen": -0.31767985224723816,
|
| 581 |
+
"rewards/margins": 0.10633231699466705,
|
| 582 |
+
"rewards/rejected": -0.4240121841430664,
|
| 583 |
+
"step": 400
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.68,
|
| 587 |
+
"learning_rate": 1.3780020494988447e-06,
|
| 588 |
+
"logits/chosen": 0.18032656610012054,
|
| 589 |
+
"logits/rejected": 0.19579799473285675,
|
| 590 |
+
"logps/chosen": -655.676513671875,
|
| 591 |
+
"logps/rejected": -761.1375122070312,
|
| 592 |
+
"loss": 0.3115,
|
| 593 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 594 |
+
"rewards/chosen": -0.3214338421821594,
|
| 595 |
+
"rewards/margins": 0.09715566784143448,
|
| 596 |
+
"rewards/rejected": -0.4185895025730133,
|
| 597 |
+
"step": 410
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.7,
|
| 601 |
+
"learning_rate": 1.2500000000000007e-06,
|
| 602 |
+
"logits/chosen": 0.15863251686096191,
|
| 603 |
+
"logits/rejected": 0.2124270498752594,
|
| 604 |
+
"logps/chosen": -599.48388671875,
|
| 605 |
+
"logps/rejected": -698.137939453125,
|
| 606 |
+
"loss": 0.3008,
|
| 607 |
+
"rewards/accuracies": 0.71875,
|
| 608 |
+
"rewards/chosen": -0.28964871168136597,
|
| 609 |
+
"rewards/margins": 0.10956676304340363,
|
| 610 |
+
"rewards/rejected": -0.3992155194282532,
|
| 611 |
+
"step": 420
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.72,
|
| 615 |
+
"learning_rate": 1.1262275548229852e-06,
|
| 616 |
+
"logits/chosen": 0.15977849066257477,
|
| 617 |
+
"logits/rejected": 0.2368887960910797,
|
| 618 |
+
"logps/chosen": -677.1764526367188,
|
| 619 |
+
"logps/rejected": -847.1414184570312,
|
| 620 |
+
"loss": 0.2538,
|
| 621 |
+
"rewards/accuracies": 0.731249988079071,
|
| 622 |
+
"rewards/chosen": -0.31619545817375183,
|
| 623 |
+
"rewards/margins": 0.16232061386108398,
|
| 624 |
+
"rewards/rejected": -0.478516161441803,
|
| 625 |
+
"step": 430
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.73,
|
| 629 |
+
"learning_rate": 1.0071035207430352e-06,
|
| 630 |
+
"logits/chosen": 0.1530456840991974,
|
| 631 |
+
"logits/rejected": 0.2543787360191345,
|
| 632 |
+
"logps/chosen": -682.7639770507812,
|
| 633 |
+
"logps/rejected": -803.2896728515625,
|
| 634 |
+
"loss": 0.2922,
|
| 635 |
+
"rewards/accuracies": 0.71875,
|
| 636 |
+
"rewards/chosen": -0.3221725821495056,
|
| 637 |
+
"rewards/margins": 0.1354692280292511,
|
| 638 |
+
"rewards/rejected": -0.4576418399810791,
|
| 639 |
+
"step": 440
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.75,
|
| 643 |
+
"learning_rate": 8.930309757836517e-07,
|
| 644 |
+
"logits/chosen": 0.14446057379245758,
|
| 645 |
+
"logits/rejected": 0.2153489887714386,
|
| 646 |
+
"logps/chosen": -625.8594970703125,
|
| 647 |
+
"logps/rejected": -766.3084716796875,
|
| 648 |
+
"loss": 0.2742,
|
| 649 |
+
"rewards/accuracies": 0.6875,
|
| 650 |
+
"rewards/chosen": -0.26791518926620483,
|
| 651 |
+
"rewards/margins": 0.13798010349273682,
|
| 652 |
+
"rewards/rejected": -0.40589532256126404,
|
| 653 |
+
"step": 450
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.77,
|
| 657 |
+
"learning_rate": 7.843959053281663e-07,
|
| 658 |
+
"logits/chosen": 0.19729985296726227,
|
| 659 |
+
"logits/rejected": 0.2402934730052948,
|
| 660 |
+
"logps/chosen": -648.9825439453125,
|
| 661 |
+
"logps/rejected": -759.14501953125,
|
| 662 |
+
"loss": 0.2909,
|
| 663 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 664 |
+
"rewards/chosen": -0.2782381474971771,
|
| 665 |
+
"rewards/margins": 0.12825573980808258,
|
| 666 |
+
"rewards/rejected": -0.4064939022064209,
|
| 667 |
+
"step": 460
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.78,
|
| 671 |
+
"learning_rate": 6.815658960673782e-07,
|
| 672 |
+
"logits/chosen": 0.17204996943473816,
|
| 673 |
+
"logits/rejected": 0.2728912830352783,
|
| 674 |
+
"logps/chosen": -618.990966796875,
|
| 675 |
+
"logps/rejected": -784.3754272460938,
|
| 676 |
+
"loss": 0.2578,
|
| 677 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 678 |
+
"rewards/chosen": -0.2586398720741272,
|
| 679 |
+
"rewards/margins": 0.15261724591255188,
|
| 680 |
+
"rewards/rejected": -0.4112570881843567,
|
| 681 |
+
"step": 470
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.8,
|
| 685 |
+
"learning_rate": 5.848888922025553e-07,
|
| 686 |
+
"logits/chosen": 0.13143357634544373,
|
| 687 |
+
"logits/rejected": 0.21312256157398224,
|
| 688 |
+
"logps/chosen": -622.7134399414062,
|
| 689 |
+
"logps/rejected": -721.8486328125,
|
| 690 |
+
"loss": 0.3171,
|
| 691 |
+
"rewards/accuracies": 0.668749988079071,
|
| 692 |
+
"rewards/chosen": -0.30107492208480835,
|
| 693 |
+
"rewards/margins": 0.08757724612951279,
|
| 694 |
+
"rewards/rejected": -0.38865217566490173,
|
| 695 |
+
"step": 480
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.82,
|
| 699 |
+
"learning_rate": 4.946920181123904e-07,
|
| 700 |
+
"logits/chosen": 0.13198260962963104,
|
| 701 |
+
"logits/rejected": 0.17995597422122955,
|
| 702 |
+
"logps/chosen": -632.0765380859375,
|
| 703 |
+
"logps/rejected": -736.373779296875,
|
| 704 |
+
"loss": 0.2857,
|
| 705 |
+
"rewards/accuracies": 0.737500011920929,
|
| 706 |
+
"rewards/chosen": -0.28522834181785583,
|
| 707 |
+
"rewards/margins": 0.11633868515491486,
|
| 708 |
+
"rewards/rejected": -0.4015669822692871,
|
| 709 |
+
"step": 490
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.83,
|
| 713 |
+
"learning_rate": 4.1128047146765936e-07,
|
| 714 |
+
"logits/chosen": 0.1584668904542923,
|
| 715 |
+
"logits/rejected": 0.249081090092659,
|
| 716 |
+
"logps/chosen": -613.0228271484375,
|
| 717 |
+
"logps/rejected": -733.0048828125,
|
| 718 |
+
"loss": 0.2768,
|
| 719 |
+
"rewards/accuracies": 0.637499988079071,
|
| 720 |
+
"rewards/chosen": -0.26978355646133423,
|
| 721 |
+
"rewards/margins": 0.1299673467874527,
|
| 722 |
+
"rewards/rejected": -0.39975088834762573,
|
| 723 |
+
"step": 500
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.85,
|
| 727 |
+
"learning_rate": 3.3493649053890325e-07,
|
| 728 |
+
"logits/chosen": 0.13333144783973694,
|
| 729 |
+
"logits/rejected": 0.25184234976768494,
|
| 730 |
+
"logps/chosen": -686.0787963867188,
|
| 731 |
+
"logps/rejected": -783.8185424804688,
|
| 732 |
+
"loss": 0.3103,
|
| 733 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 734 |
+
"rewards/chosen": -0.3115057349205017,
|
| 735 |
+
"rewards/margins": 0.11691226065158844,
|
| 736 |
+
"rewards/rejected": -0.42841801047325134,
|
| 737 |
+
"step": 510
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.87,
|
| 741 |
+
"learning_rate": 2.6591839919146963e-07,
|
| 742 |
+
"logits/chosen": 0.1042710691690445,
|
| 743 |
+
"logits/rejected": 0.19632944464683533,
|
| 744 |
+
"logps/chosen": -641.197021484375,
|
| 745 |
+
"logps/rejected": -739.2056884765625,
|
| 746 |
+
"loss": 0.3089,
|
| 747 |
+
"rewards/accuracies": 0.625,
|
| 748 |
+
"rewards/chosen": -0.3029845356941223,
|
| 749 |
+
"rewards/margins": 0.11691125482320786,
|
| 750 |
+
"rewards/rejected": -0.41989579796791077,
|
| 751 |
+
"step": 520
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.88,
|
| 755 |
+
"learning_rate": 2.044597327993153e-07,
|
| 756 |
+
"logits/chosen": 0.13513818383216858,
|
| 757 |
+
"logits/rejected": 0.2279660403728485,
|
| 758 |
+
"logps/chosen": -638.3856201171875,
|
| 759 |
+
"logps/rejected": -733.2174072265625,
|
| 760 |
+
"loss": 0.3394,
|
| 761 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 762 |
+
"rewards/chosen": -0.31149208545684814,
|
| 763 |
+
"rewards/margins": 0.09216032922267914,
|
| 764 |
+
"rewards/rejected": -0.4036524295806885,
|
| 765 |
+
"step": 530
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.9,
|
| 769 |
+
"learning_rate": 1.507684480352292e-07,
|
| 770 |
+
"logits/chosen": 0.13924403488636017,
|
| 771 |
+
"logits/rejected": 0.19593167304992676,
|
| 772 |
+
"logps/chosen": -614.1495971679688,
|
| 773 |
+
"logps/rejected": -716.1968994140625,
|
| 774 |
+
"loss": 0.2755,
|
| 775 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 776 |
+
"rewards/chosen": -0.2909039855003357,
|
| 777 |
+
"rewards/margins": 0.11327805370092392,
|
| 778 |
+
"rewards/rejected": -0.4041820466518402,
|
| 779 |
+
"step": 540
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.92,
|
| 783 |
+
"learning_rate": 1.0502621921127776e-07,
|
| 784 |
+
"logits/chosen": 0.1402362883090973,
|
| 785 |
+
"logits/rejected": 0.2016625702381134,
|
| 786 |
+
"logps/chosen": -635.9324340820312,
|
| 787 |
+
"logps/rejected": -708.5713500976562,
|
| 788 |
+
"loss": 0.2907,
|
| 789 |
+
"rewards/accuracies": 0.6875,
|
| 790 |
+
"rewards/chosen": -0.30609840154647827,
|
| 791 |
+
"rewards/margins": 0.10378583520650864,
|
| 792 |
+
"rewards/rejected": -0.4098842144012451,
|
| 793 |
+
"step": 550
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.93,
|
| 797 |
+
"learning_rate": 6.738782355044048e-08,
|
| 798 |
+
"logits/chosen": 0.1220315471291542,
|
| 799 |
+
"logits/rejected": 0.1830691397190094,
|
| 800 |
+
"logps/chosen": -661.2696533203125,
|
| 801 |
+
"logps/rejected": -801.816162109375,
|
| 802 |
+
"loss": 0.2863,
|
| 803 |
+
"rewards/accuracies": 0.699999988079071,
|
| 804 |
+
"rewards/chosen": -0.3084556758403778,
|
| 805 |
+
"rewards/margins": 0.14121660590171814,
|
| 806 |
+
"rewards/rejected": -0.44967228174209595,
|
| 807 |
+
"step": 560
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.95,
|
| 811 |
+
"learning_rate": 3.798061746947995e-08,
|
| 812 |
+
"logits/chosen": 0.12625311315059662,
|
| 813 |
+
"logits/rejected": 0.19530022144317627,
|
| 814 |
+
"logps/chosen": -651.5577392578125,
|
| 815 |
+
"logps/rejected": -755.2259521484375,
|
| 816 |
+
"loss": 0.284,
|
| 817 |
+
"rewards/accuracies": 0.706250011920929,
|
| 818 |
+
"rewards/chosen": -0.2925582528114319,
|
| 819 |
+
"rewards/margins": 0.12503091990947723,
|
| 820 |
+
"rewards/rejected": -0.4175891876220703,
|
| 821 |
+
"step": 570
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.97,
|
| 825 |
+
"learning_rate": 1.6904105645142443e-08,
|
| 826 |
+
"logits/chosen": 0.16926367580890656,
|
| 827 |
+
"logits/rejected": 0.21377988159656525,
|
| 828 |
+
"logps/chosen": -626.6007080078125,
|
| 829 |
+
"logps/rejected": -761.6741943359375,
|
| 830 |
+
"loss": 0.2901,
|
| 831 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 832 |
+
"rewards/chosen": -0.30488914251327515,
|
| 833 |
+
"rewards/margins": 0.11805585771799088,
|
| 834 |
+
"rewards/rejected": -0.4229450225830078,
|
| 835 |
+
"step": 580
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.98,
|
| 839 |
+
"learning_rate": 4.229604321829561e-09,
|
| 840 |
+
"logits/chosen": 0.12586399912834167,
|
| 841 |
+
"logits/rejected": 0.20252175629138947,
|
| 842 |
+
"logps/chosen": -655.3035278320312,
|
| 843 |
+
"logps/rejected": -739.29833984375,
|
| 844 |
+
"loss": 0.3106,
|
| 845 |
+
"rewards/accuracies": 0.643750011920929,
|
| 846 |
+
"rewards/chosen": -0.2980835437774658,
|
| 847 |
+
"rewards/margins": 0.1191561371088028,
|
| 848 |
+
"rewards/rejected": -0.4172397255897522,
|
| 849 |
+
"step": 590
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 1.0,
|
| 853 |
+
"learning_rate": 0.0,
|
| 854 |
+
"logits/chosen": 0.159867063164711,
|
| 855 |
+
"logits/rejected": 0.22124703228473663,
|
| 856 |
+
"logps/chosen": -676.0281982421875,
|
| 857 |
+
"logps/rejected": -778.9498291015625,
|
| 858 |
+
"loss": 0.2705,
|
| 859 |
+
"rewards/accuracies": 0.762499988079071,
|
| 860 |
+
"rewards/chosen": -0.28896114230155945,
|
| 861 |
+
"rewards/margins": 0.14466610550880432,
|
| 862 |
+
"rewards/rejected": -0.43362727761268616,
|
| 863 |
+
"step": 600
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 1.0,
|
| 867 |
+
"step": 600,
|
| 868 |
+
"total_flos": 0.0,
|
| 869 |
+
"train_loss": 0.3152459278702736,
|
| 870 |
+
"train_runtime": 5630.9992,
|
| 871 |
+
"train_samples_per_second": 1.705,
|
| 872 |
+
"train_steps_per_second": 0.107
|
| 873 |
+
}
|
| 874 |
+
],
|
| 875 |
+
"logging_steps": 10,
|
| 876 |
+
"max_steps": 600,
|
| 877 |
+
"num_input_tokens_seen": 0,
|
| 878 |
+
"num_train_epochs": 1,
|
| 879 |
+
"save_steps": 100,
|
| 880 |
+
"total_flos": 0.0,
|
| 881 |
+
"train_batch_size": 4,
|
| 882 |
+
"trial_name": null,
|
| 883 |
+
"trial_params": null
|
| 884 |
+
}
|