Qin Liu
commited on
Model save
Browse files- README.md +62 -0
- all_results.json +9 -0
- train_results.json +9 -0
- trainer_state.json +1767 -0
README.md
ADDED
|
@@ -0,0 +1,62 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: meta-llama/Meta-Llama-3-8B-Instruct
|
| 3 |
+
library_name: peft
|
| 4 |
+
license: llama3
|
| 5 |
+
tags:
|
| 6 |
+
- trl
|
| 7 |
+
- dpo
|
| 8 |
+
- generated_from_trainer
|
| 9 |
+
model-index:
|
| 10 |
+
- name: llama3-sudo-dpo-instruct-5epochs-0909
|
| 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 |
+
# llama3-sudo-dpo-instruct-5epochs-0909
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) on an unknown 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: 2
|
| 40 |
+
- eval_batch_size: 8
|
| 41 |
+
- seed: 42
|
| 42 |
+
- distributed_type: multi-GPU
|
| 43 |
+
- num_devices: 8
|
| 44 |
+
- gradient_accumulation_steps: 4
|
| 45 |
+
- total_train_batch_size: 64
|
| 46 |
+
- total_eval_batch_size: 64
|
| 47 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 48 |
+
- lr_scheduler_type: cosine
|
| 49 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 50 |
+
- num_epochs: 5
|
| 51 |
+
|
| 52 |
+
### Training results
|
| 53 |
+
|
| 54 |
+
|
| 55 |
+
|
| 56 |
+
### Framework versions
|
| 57 |
+
|
| 58 |
+
- PEFT 0.12.0
|
| 59 |
+
- Transformers 4.44.0
|
| 60 |
+
- Pytorch 2.1.2
|
| 61 |
+
- Datasets 2.20.0
|
| 62 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.04901049308554833,
|
| 5 |
+
"train_runtime": 10890.5088,
|
| 6 |
+
"train_samples": 14642,
|
| 7 |
+
"train_samples_per_second": 6.722,
|
| 8 |
+
"train_steps_per_second": 0.105
|
| 9 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.04901049308554833,
|
| 5 |
+
"train_runtime": 10890.5088,
|
| 6 |
+
"train_samples": 14642,
|
| 7 |
+
"train_samples_per_second": 6.722,
|
| 8 |
+
"train_steps_per_second": 0.105
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1767 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 5.0,
|
| 5 |
+
"eval_steps": 1000,
|
| 6 |
+
"global_step": 1145,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.004366812227074236,
|
| 13 |
+
"grad_norm": 0.47458936036527805,
|
| 14 |
+
"learning_rate": 4.347826086956522e-08,
|
| 15 |
+
"logits/chosen": -1.077685832977295,
|
| 16 |
+
"logits/rejected": -0.9262686967849731,
|
| 17 |
+
"logps/chosen": -270.19927978515625,
|
| 18 |
+
"logps/rejected": -284.0829772949219,
|
| 19 |
+
"loss": 0.6931,
|
| 20 |
+
"rewards/accuracies": 0.0,
|
| 21 |
+
"rewards/chosen": 0.0,
|
| 22 |
+
"rewards/margins": 0.0,
|
| 23 |
+
"rewards/rejected": 0.0,
|
| 24 |
+
"step": 1
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.043668122270742356,
|
| 28 |
+
"grad_norm": 0.523539374370341,
|
| 29 |
+
"learning_rate": 4.347826086956522e-07,
|
| 30 |
+
"logits/chosen": -0.9952558875083923,
|
| 31 |
+
"logits/rejected": -1.0596376657485962,
|
| 32 |
+
"logps/chosen": -300.6241760253906,
|
| 33 |
+
"logps/rejected": -255.2275390625,
|
| 34 |
+
"loss": 0.6929,
|
| 35 |
+
"rewards/accuracies": 0.4583333432674408,
|
| 36 |
+
"rewards/chosen": 0.0006611866410821676,
|
| 37 |
+
"rewards/margins": 0.0006012354861013591,
|
| 38 |
+
"rewards/rejected": 5.995116225676611e-05,
|
| 39 |
+
"step": 10
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.08733624454148471,
|
| 43 |
+
"grad_norm": 0.5365126766074364,
|
| 44 |
+
"learning_rate": 8.695652173913044e-07,
|
| 45 |
+
"logits/chosen": -1.1060526371002197,
|
| 46 |
+
"logits/rejected": -1.0136970281600952,
|
| 47 |
+
"logps/chosen": -282.16522216796875,
|
| 48 |
+
"logps/rejected": -316.65264892578125,
|
| 49 |
+
"loss": 0.6925,
|
| 50 |
+
"rewards/accuracies": 0.574999988079071,
|
| 51 |
+
"rewards/chosen": 0.0014794444432482123,
|
| 52 |
+
"rewards/margins": 0.0008522253483533859,
|
| 53 |
+
"rewards/rejected": 0.0006272192695178092,
|
| 54 |
+
"step": 20
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.13100436681222707,
|
| 58 |
+
"grad_norm": 0.580247902784444,
|
| 59 |
+
"learning_rate": 1.3043478260869566e-06,
|
| 60 |
+
"logits/chosen": -1.0645825862884521,
|
| 61 |
+
"logits/rejected": -1.1114238500595093,
|
| 62 |
+
"logps/chosen": -287.91937255859375,
|
| 63 |
+
"logps/rejected": -264.2998352050781,
|
| 64 |
+
"loss": 0.69,
|
| 65 |
+
"rewards/accuracies": 0.737500011920929,
|
| 66 |
+
"rewards/chosen": 0.00400827219709754,
|
| 67 |
+
"rewards/margins": 0.006094515323638916,
|
| 68 |
+
"rewards/rejected": -0.002086243825033307,
|
| 69 |
+
"step": 30
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.17467248908296942,
|
| 73 |
+
"grad_norm": 0.45391259957771296,
|
| 74 |
+
"learning_rate": 1.7391304347826088e-06,
|
| 75 |
+
"logits/chosen": -1.0716311931610107,
|
| 76 |
+
"logits/rejected": -1.1120277643203735,
|
| 77 |
+
"logps/chosen": -285.3992919921875,
|
| 78 |
+
"logps/rejected": -272.4431457519531,
|
| 79 |
+
"loss": 0.6833,
|
| 80 |
+
"rewards/accuracies": 0.862500011920929,
|
| 81 |
+
"rewards/chosen": 0.008399474434554577,
|
| 82 |
+
"rewards/margins": 0.01975000463426113,
|
| 83 |
+
"rewards/rejected": -0.011350530199706554,
|
| 84 |
+
"step": 40
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.2183406113537118,
|
| 88 |
+
"grad_norm": 0.46277317667378487,
|
| 89 |
+
"learning_rate": 2.173913043478261e-06,
|
| 90 |
+
"logits/chosen": -1.111750841140747,
|
| 91 |
+
"logits/rejected": -1.0212160348892212,
|
| 92 |
+
"logps/chosen": -264.515380859375,
|
| 93 |
+
"logps/rejected": -304.44281005859375,
|
| 94 |
+
"loss": 0.6728,
|
| 95 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 96 |
+
"rewards/chosen": 0.018858566880226135,
|
| 97 |
+
"rewards/margins": 0.04448834806680679,
|
| 98 |
+
"rewards/rejected": -0.02562977746129036,
|
| 99 |
+
"step": 50
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.26200873362445415,
|
| 103 |
+
"grad_norm": 0.538292399167113,
|
| 104 |
+
"learning_rate": 2.6086956521739132e-06,
|
| 105 |
+
"logits/chosen": -1.070939302444458,
|
| 106 |
+
"logits/rejected": -1.1072248220443726,
|
| 107 |
+
"logps/chosen": -285.0442810058594,
|
| 108 |
+
"logps/rejected": -262.81597900390625,
|
| 109 |
+
"loss": 0.654,
|
| 110 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 111 |
+
"rewards/chosen": 0.03614560887217522,
|
| 112 |
+
"rewards/margins": 0.09129359573125839,
|
| 113 |
+
"rewards/rejected": -0.05514797568321228,
|
| 114 |
+
"step": 60
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.3056768558951965,
|
| 118 |
+
"grad_norm": 0.8400838184943603,
|
| 119 |
+
"learning_rate": 3.043478260869566e-06,
|
| 120 |
+
"logits/chosen": -1.1011688709259033,
|
| 121 |
+
"logits/rejected": -1.1285476684570312,
|
| 122 |
+
"logps/chosen": -279.41290283203125,
|
| 123 |
+
"logps/rejected": -276.6390686035156,
|
| 124 |
+
"loss": 0.6152,
|
| 125 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 126 |
+
"rewards/chosen": 0.04447701945900917,
|
| 127 |
+
"rewards/margins": 0.16786250472068787,
|
| 128 |
+
"rewards/rejected": -0.123385488986969,
|
| 129 |
+
"step": 70
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.34934497816593885,
|
| 133 |
+
"grad_norm": 1.363956535309491,
|
| 134 |
+
"learning_rate": 3.4782608695652175e-06,
|
| 135 |
+
"logits/chosen": -1.1735438108444214,
|
| 136 |
+
"logits/rejected": -1.0770442485809326,
|
| 137 |
+
"logps/chosen": -256.97564697265625,
|
| 138 |
+
"logps/rejected": -330.7611083984375,
|
| 139 |
+
"loss": 0.514,
|
| 140 |
+
"rewards/accuracies": 1.0,
|
| 141 |
+
"rewards/chosen": 0.10425807535648346,
|
| 142 |
+
"rewards/margins": 0.4447970986366272,
|
| 143 |
+
"rewards/rejected": -0.34053897857666016,
|
| 144 |
+
"step": 80
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 0.3930131004366812,
|
| 148 |
+
"grad_norm": 1.0652559890564237,
|
| 149 |
+
"learning_rate": 3.91304347826087e-06,
|
| 150 |
+
"logits/chosen": -1.0946518182754517,
|
| 151 |
+
"logits/rejected": -1.0877329111099243,
|
| 152 |
+
"logps/chosen": -266.5739440917969,
|
| 153 |
+
"logps/rejected": -361.3409729003906,
|
| 154 |
+
"loss": 0.278,
|
| 155 |
+
"rewards/accuracies": 1.0,
|
| 156 |
+
"rewards/chosen": 0.2085898369550705,
|
| 157 |
+
"rewards/margins": 1.1870415210723877,
|
| 158 |
+
"rewards/rejected": -0.9784518480300903,
|
| 159 |
+
"step": 90
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.4366812227074236,
|
| 163 |
+
"grad_norm": 0.3433521660739356,
|
| 164 |
+
"learning_rate": 4.347826086956522e-06,
|
| 165 |
+
"logits/chosen": -1.055832862854004,
|
| 166 |
+
"logits/rejected": -0.9789533615112305,
|
| 167 |
+
"logps/chosen": -219.6240692138672,
|
| 168 |
+
"logps/rejected": -508.79132080078125,
|
| 169 |
+
"loss": 0.0735,
|
| 170 |
+
"rewards/accuracies": 1.0,
|
| 171 |
+
"rewards/chosen": 0.38825297355651855,
|
| 172 |
+
"rewards/margins": 2.7440783977508545,
|
| 173 |
+
"rewards/rejected": -2.355825662612915,
|
| 174 |
+
"step": 100
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 0.48034934497816595,
|
| 178 |
+
"grad_norm": 0.11711005718135792,
|
| 179 |
+
"learning_rate": 4.782608695652174e-06,
|
| 180 |
+
"logits/chosen": -0.9485265016555786,
|
| 181 |
+
"logits/rejected": -0.9038203358650208,
|
| 182 |
+
"logps/chosen": -269.6318664550781,
|
| 183 |
+
"logps/rejected": -716.6309814453125,
|
| 184 |
+
"loss": 0.018,
|
| 185 |
+
"rewards/accuracies": 1.0,
|
| 186 |
+
"rewards/chosen": 0.2308163344860077,
|
| 187 |
+
"rewards/margins": 4.760537147521973,
|
| 188 |
+
"rewards/rejected": -4.529721260070801,
|
| 189 |
+
"step": 110
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.5240174672489083,
|
| 193 |
+
"grad_norm": 0.056145921913412625,
|
| 194 |
+
"learning_rate": 4.999709285361594e-06,
|
| 195 |
+
"logits/chosen": -0.9236040115356445,
|
| 196 |
+
"logits/rejected": -0.8355329632759094,
|
| 197 |
+
"logps/chosen": -271.9773254394531,
|
| 198 |
+
"logps/rejected": -864.2019653320312,
|
| 199 |
+
"loss": 0.0057,
|
| 200 |
+
"rewards/accuracies": 1.0,
|
| 201 |
+
"rewards/chosen": 0.2535458505153656,
|
| 202 |
+
"rewards/margins": 6.195461273193359,
|
| 203 |
+
"rewards/rejected": -5.941915035247803,
|
| 204 |
+
"step": 120
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 0.5676855895196506,
|
| 208 |
+
"grad_norm": 0.03098316075816059,
|
| 209 |
+
"learning_rate": 4.997383973910631e-06,
|
| 210 |
+
"logits/chosen": -0.8370451927185059,
|
| 211 |
+
"logits/rejected": -0.6872082948684692,
|
| 212 |
+
"logps/chosen": -263.49090576171875,
|
| 213 |
+
"logps/rejected": -959.0076293945312,
|
| 214 |
+
"loss": 0.0029,
|
| 215 |
+
"rewards/accuracies": 1.0,
|
| 216 |
+
"rewards/chosen": 0.2465527057647705,
|
| 217 |
+
"rewards/margins": 7.134703159332275,
|
| 218 |
+
"rewards/rejected": -6.888150215148926,
|
| 219 |
+
"step": 130
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.611353711790393,
|
| 223 |
+
"grad_norm": 0.01781004140188894,
|
| 224 |
+
"learning_rate": 4.9927355140895775e-06,
|
| 225 |
+
"logits/chosen": -0.8542898297309875,
|
| 226 |
+
"logits/rejected": -0.6182370185852051,
|
| 227 |
+
"logps/chosen": -256.8946533203125,
|
| 228 |
+
"logps/rejected": -1113.857666015625,
|
| 229 |
+
"loss": 0.002,
|
| 230 |
+
"rewards/accuracies": 1.0,
|
| 231 |
+
"rewards/chosen": 0.1456744223833084,
|
| 232 |
+
"rewards/margins": 8.21844482421875,
|
| 233 |
+
"rewards/rejected": -8.072771072387695,
|
| 234 |
+
"step": 140
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.6550218340611353,
|
| 238 |
+
"grad_norm": 0.0626793578117912,
|
| 239 |
+
"learning_rate": 4.985768230048011e-06,
|
| 240 |
+
"logits/chosen": -0.8041971325874329,
|
| 241 |
+
"logits/rejected": -0.645085334777832,
|
| 242 |
+
"logps/chosen": -274.4373779296875,
|
| 243 |
+
"logps/rejected": -1066.285400390625,
|
| 244 |
+
"loss": 0.0019,
|
| 245 |
+
"rewards/accuracies": 1.0,
|
| 246 |
+
"rewards/chosen": 0.16038499772548676,
|
| 247 |
+
"rewards/margins": 8.415651321411133,
|
| 248 |
+
"rewards/rejected": -8.255266189575195,
|
| 249 |
+
"step": 150
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.6986899563318777,
|
| 253 |
+
"grad_norm": 0.016666200992927045,
|
| 254 |
+
"learning_rate": 4.976488602981748e-06,
|
| 255 |
+
"logits/chosen": -0.8367046117782593,
|
| 256 |
+
"logits/rejected": -0.6429373025894165,
|
| 257 |
+
"logps/chosen": -258.5212707519531,
|
| 258 |
+
"logps/rejected": -1201.1590576171875,
|
| 259 |
+
"loss": 0.001,
|
| 260 |
+
"rewards/accuracies": 1.0,
|
| 261 |
+
"rewards/chosen": 0.2543124556541443,
|
| 262 |
+
"rewards/margins": 9.140569686889648,
|
| 263 |
+
"rewards/rejected": -8.88625717163086,
|
| 264 |
+
"step": 160
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 0.74235807860262,
|
| 268 |
+
"grad_norm": 0.07994746225935201,
|
| 269 |
+
"learning_rate": 4.9649052651038255e-06,
|
| 270 |
+
"logits/chosen": -0.7291176915168762,
|
| 271 |
+
"logits/rejected": -0.5780690908432007,
|
| 272 |
+
"logps/chosen": -279.2346496582031,
|
| 273 |
+
"logps/rejected": -1126.931396484375,
|
| 274 |
+
"loss": 0.0009,
|
| 275 |
+
"rewards/accuracies": 1.0,
|
| 276 |
+
"rewards/chosen": 0.16093984246253967,
|
| 277 |
+
"rewards/margins": 8.925050735473633,
|
| 278 |
+
"rewards/rejected": -8.764110565185547,
|
| 279 |
+
"step": 170
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.7860262008733624,
|
| 283 |
+
"grad_norm": 0.01171334168304518,
|
| 284 |
+
"learning_rate": 4.9510289916145295e-06,
|
| 285 |
+
"logits/chosen": -0.7453992962837219,
|
| 286 |
+
"logits/rejected": -0.5254510641098022,
|
| 287 |
+
"logps/chosen": -256.3668518066406,
|
| 288 |
+
"logps/rejected": -1209.734619140625,
|
| 289 |
+
"loss": 0.0006,
|
| 290 |
+
"rewards/accuracies": 1.0,
|
| 291 |
+
"rewards/chosen": 0.1778395026922226,
|
| 292 |
+
"rewards/margins": 9.462778091430664,
|
| 293 |
+
"rewards/rejected": -9.28493881225586,
|
| 294 |
+
"step": 180
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 0.8296943231441049,
|
| 298 |
+
"grad_norm": 0.017372174016347157,
|
| 299 |
+
"learning_rate": 4.934872690677953e-06,
|
| 300 |
+
"logits/chosen": -0.7389127612113953,
|
| 301 |
+
"logits/rejected": -0.5030776858329773,
|
| 302 |
+
"logps/chosen": -271.6781921386719,
|
| 303 |
+
"logps/rejected": -1233.863037109375,
|
| 304 |
+
"loss": 0.0007,
|
| 305 |
+
"rewards/accuracies": 1.0,
|
| 306 |
+
"rewards/chosen": 0.13182204961776733,
|
| 307 |
+
"rewards/margins": 9.499404907226562,
|
| 308 |
+
"rewards/rejected": -9.367582321166992,
|
| 309 |
+
"step": 190
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.8733624454148472,
|
| 313 |
+
"grad_norm": 0.005099638792861995,
|
| 314 |
+
"learning_rate": 4.9164513914144005e-06,
|
| 315 |
+
"logits/chosen": -0.6438565254211426,
|
| 316 |
+
"logits/rejected": -0.3725091516971588,
|
| 317 |
+
"logps/chosen": -258.71893310546875,
|
| 318 |
+
"logps/rejected": -1232.365478515625,
|
| 319 |
+
"loss": 0.0005,
|
| 320 |
+
"rewards/accuracies": 1.0,
|
| 321 |
+
"rewards/chosen": 0.11995329707860947,
|
| 322 |
+
"rewards/margins": 9.731719017028809,
|
| 323 |
+
"rewards/rejected": -9.611766815185547,
|
| 324 |
+
"step": 200
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.9170305676855895,
|
| 328 |
+
"grad_norm": 0.007449286013455516,
|
| 329 |
+
"learning_rate": 4.8957822299198045e-06,
|
| 330 |
+
"logits/chosen": -0.7947207689285278,
|
| 331 |
+
"logits/rejected": -0.5452377200126648,
|
| 332 |
+
"logps/chosen": -277.94647216796875,
|
| 333 |
+
"logps/rejected": -1344.936767578125,
|
| 334 |
+
"loss": 0.0006,
|
| 335 |
+
"rewards/accuracies": 1.0,
|
| 336 |
+
"rewards/chosen": 0.07659594714641571,
|
| 337 |
+
"rewards/margins": 10.560900688171387,
|
| 338 |
+
"rewards/rejected": -10.484306335449219,
|
| 339 |
+
"step": 210
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.9606986899563319,
|
| 343 |
+
"grad_norm": 0.025567537238137088,
|
| 344 |
+
"learning_rate": 4.872884433325169e-06,
|
| 345 |
+
"logits/chosen": -0.7050657272338867,
|
| 346 |
+
"logits/rejected": -0.49923476576805115,
|
| 347 |
+
"logps/chosen": -286.5049743652344,
|
| 348 |
+
"logps/rejected": -1234.7254638671875,
|
| 349 |
+
"loss": 0.0004,
|
| 350 |
+
"rewards/accuracies": 1.0,
|
| 351 |
+
"rewards/chosen": 0.08888520300388336,
|
| 352 |
+
"rewards/margins": 9.898573875427246,
|
| 353 |
+
"rewards/rejected": -9.809688568115234,
|
| 354 |
+
"step": 220
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 1.0043668122270741,
|
| 358 |
+
"grad_norm": 0.03935313476723805,
|
| 359 |
+
"learning_rate": 4.847779301910868e-06,
|
| 360 |
+
"logits/chosen": -0.7885675430297852,
|
| 361 |
+
"logits/rejected": -0.48259955644607544,
|
| 362 |
+
"logps/chosen": -235.9558563232422,
|
| 363 |
+
"logps/rejected": -1333.6160888671875,
|
| 364 |
+
"loss": 0.0005,
|
| 365 |
+
"rewards/accuracies": 1.0,
|
| 366 |
+
"rewards/chosen": 0.1478557288646698,
|
| 367 |
+
"rewards/margins": 10.591659545898438,
|
| 368 |
+
"rewards/rejected": -10.443804740905762,
|
| 369 |
+
"step": 230
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 1.0480349344978166,
|
| 373 |
+
"grad_norm": 0.019331140450817564,
|
| 374 |
+
"learning_rate": 4.820490189292415e-06,
|
| 375 |
+
"logits/chosen": -0.6384273767471313,
|
| 376 |
+
"logits/rejected": -0.32141706347465515,
|
| 377 |
+
"logps/chosen": -256.14813232421875,
|
| 378 |
+
"logps/rejected": -1328.8856201171875,
|
| 379 |
+
"loss": 0.0004,
|
| 380 |
+
"rewards/accuracies": 1.0,
|
| 381 |
+
"rewards/chosen": 0.03197110444307327,
|
| 382 |
+
"rewards/margins": 10.722203254699707,
|
| 383 |
+
"rewards/rejected": -10.690231323242188,
|
| 384 |
+
"step": 240
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 1.091703056768559,
|
| 388 |
+
"grad_norm": 0.001917220346998511,
|
| 389 |
+
"learning_rate": 4.791042480696179e-06,
|
| 390 |
+
"logits/chosen": -0.8021818399429321,
|
| 391 |
+
"logits/rejected": -0.5436561107635498,
|
| 392 |
+
"logps/chosen": -285.2548828125,
|
| 393 |
+
"logps/rejected": -1359.247802734375,
|
| 394 |
+
"loss": 0.0004,
|
| 395 |
+
"rewards/accuracies": 1.0,
|
| 396 |
+
"rewards/chosen": 0.09929420799016953,
|
| 397 |
+
"rewards/margins": 10.846672058105469,
|
| 398 |
+
"rewards/rejected": -10.747377395629883,
|
| 399 |
+
"step": 250
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 1.1353711790393013,
|
| 403 |
+
"grad_norm": 0.009816737966176867,
|
| 404 |
+
"learning_rate": 4.759463569345205e-06,
|
| 405 |
+
"logits/chosen": -0.7113679051399231,
|
| 406 |
+
"logits/rejected": -0.45295238494873047,
|
| 407 |
+
"logps/chosen": -268.17083740234375,
|
| 408 |
+
"logps/rejected": -1343.6666259765625,
|
| 409 |
+
"loss": 0.0004,
|
| 410 |
+
"rewards/accuracies": 1.0,
|
| 411 |
+
"rewards/chosen": 0.10034344345331192,
|
| 412 |
+
"rewards/margins": 10.877279281616211,
|
| 413 |
+
"rewards/rejected": -10.776936531066895,
|
| 414 |
+
"step": 260
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 1.1790393013100438,
|
| 418 |
+
"grad_norm": 0.003536564037828518,
|
| 419 |
+
"learning_rate": 4.725782830977145e-06,
|
| 420 |
+
"logits/chosen": -0.7203564047813416,
|
| 421 |
+
"logits/rejected": -0.4096035957336426,
|
| 422 |
+
"logps/chosen": -257.78900146484375,
|
| 423 |
+
"logps/rejected": -1433.7529296875,
|
| 424 |
+
"loss": 0.0002,
|
| 425 |
+
"rewards/accuracies": 1.0,
|
| 426 |
+
"rewards/chosen": 0.05666704103350639,
|
| 427 |
+
"rewards/margins": 11.476617813110352,
|
| 428 |
+
"rewards/rejected": -11.419949531555176,
|
| 429 |
+
"step": 270
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 1.222707423580786,
|
| 433 |
+
"grad_norm": 0.0031083495470266283,
|
| 434 |
+
"learning_rate": 4.690031596517984e-06,
|
| 435 |
+
"logits/chosen": -0.6830414533615112,
|
| 436 |
+
"logits/rejected": -0.3940260410308838,
|
| 437 |
+
"logps/chosen": -252.91641235351562,
|
| 438 |
+
"logps/rejected": -1419.6212158203125,
|
| 439 |
+
"loss": 0.0003,
|
| 440 |
+
"rewards/accuracies": 1.0,
|
| 441 |
+
"rewards/chosen": 0.10215254127979279,
|
| 442 |
+
"rewards/margins": 11.425537109375,
|
| 443 |
+
"rewards/rejected": -11.323384284973145,
|
| 444 |
+
"step": 280
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 1.2663755458515285,
|
| 448 |
+
"grad_norm": 0.0031010184167246683,
|
| 449 |
+
"learning_rate": 4.652243122936987e-06,
|
| 450 |
+
"logits/chosen": -0.7523640394210815,
|
| 451 |
+
"logits/rejected": -0.4601876735687256,
|
| 452 |
+
"logps/chosen": -267.94482421875,
|
| 453 |
+
"logps/rejected": -1480.731201171875,
|
| 454 |
+
"loss": 0.0003,
|
| 455 |
+
"rewards/accuracies": 1.0,
|
| 456 |
+
"rewards/chosen": 0.141164630651474,
|
| 457 |
+
"rewards/margins": 12.00882339477539,
|
| 458 |
+
"rewards/rejected": -11.867658615112305,
|
| 459 |
+
"step": 290
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 1.3100436681222707,
|
| 463 |
+
"grad_norm": 0.03893958758819322,
|
| 464 |
+
"learning_rate": 4.612452562309975e-06,
|
| 465 |
+
"logits/chosen": -0.6891266703605652,
|
| 466 |
+
"logits/rejected": -0.3029942512512207,
|
| 467 |
+
"logps/chosen": -241.3924560546875,
|
| 468 |
+
"logps/rejected": -1440.799072265625,
|
| 469 |
+
"loss": 0.0002,
|
| 470 |
+
"rewards/accuracies": 1.0,
|
| 471 |
+
"rewards/chosen": 0.11804157495498657,
|
| 472 |
+
"rewards/margins": 11.638472557067871,
|
| 473 |
+
"rewards/rejected": -11.520431518554688,
|
| 474 |
+
"step": 300
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 1.3537117903930131,
|
| 478 |
+
"grad_norm": 0.0020487177308325163,
|
| 479 |
+
"learning_rate": 4.570696929119717e-06,
|
| 480 |
+
"logits/chosen": -0.6615210771560669,
|
| 481 |
+
"logits/rejected": -0.40224489569664,
|
| 482 |
+
"logps/chosen": -264.6934814453125,
|
| 483 |
+
"logps/rejected": -1409.147216796875,
|
| 484 |
+
"loss": 0.0003,
|
| 485 |
+
"rewards/accuracies": 1.0,
|
| 486 |
+
"rewards/chosen": 0.0794425830245018,
|
| 487 |
+
"rewards/margins": 11.551889419555664,
|
| 488 |
+
"rewards/rejected": -11.47244644165039,
|
| 489 |
+
"step": 310
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 1.3973799126637554,
|
| 493 |
+
"grad_norm": 0.0017833767780479168,
|
| 494 |
+
"learning_rate": 4.527015065823841e-06,
|
| 495 |
+
"logits/chosen": -0.6528028249740601,
|
| 496 |
+
"logits/rejected": -0.4198477864265442,
|
| 497 |
+
"logps/chosen": -277.9957580566406,
|
| 498 |
+
"logps/rejected": -1393.375244140625,
|
| 499 |
+
"loss": 0.0001,
|
| 500 |
+
"rewards/accuracies": 1.0,
|
| 501 |
+
"rewards/chosen": 0.09100461006164551,
|
| 502 |
+
"rewards/margins": 11.455917358398438,
|
| 503 |
+
"rewards/rejected": -11.364913940429688,
|
| 504 |
+
"step": 320
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 1.4410480349344978,
|
| 508 |
+
"grad_norm": 0.007670342609009089,
|
| 509 |
+
"learning_rate": 4.481447606722309e-06,
|
| 510 |
+
"logits/chosen": -0.6851311326026917,
|
| 511 |
+
"logits/rejected": -0.44172611832618713,
|
| 512 |
+
"logps/chosen": -280.6171875,
|
| 513 |
+
"logps/rejected": -1444.6007080078125,
|
| 514 |
+
"loss": 0.0002,
|
| 515 |
+
"rewards/accuracies": 1.0,
|
| 516 |
+
"rewards/chosen": 0.1056002750992775,
|
| 517 |
+
"rewards/margins": 11.923059463500977,
|
| 518 |
+
"rewards/rejected": -11.817459106445312,
|
| 519 |
+
"step": 330
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 1.48471615720524,
|
| 523 |
+
"grad_norm": 0.004356336133031361,
|
| 524 |
+
"learning_rate": 4.434036940158062e-06,
|
| 525 |
+
"logits/chosen": -0.7159875631332397,
|
| 526 |
+
"logits/rejected": -0.26825979351997375,
|
| 527 |
+
"logps/chosen": -226.5546875,
|
| 528 |
+
"logps/rejected": -1620.3592529296875,
|
| 529 |
+
"loss": 0.0004,
|
| 530 |
+
"rewards/accuracies": 1.0,
|
| 531 |
+
"rewards/chosen": 0.12559999525547028,
|
| 532 |
+
"rewards/margins": 13.081733703613281,
|
| 533 |
+
"rewards/rejected": -12.956133842468262,
|
| 534 |
+
"step": 340
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 1.5283842794759825,
|
| 538 |
+
"grad_norm": 0.006924018476345942,
|
| 539 |
+
"learning_rate": 4.384827169085993e-06,
|
| 540 |
+
"logits/chosen": -0.722930908203125,
|
| 541 |
+
"logits/rejected": -0.54079270362854,
|
| 542 |
+
"logps/chosen": -305.4542541503906,
|
| 543 |
+
"logps/rejected": -1445.2802734375,
|
| 544 |
+
"loss": 0.0002,
|
| 545 |
+
"rewards/accuracies": 1.0,
|
| 546 |
+
"rewards/chosen": 0.036442067474126816,
|
| 547 |
+
"rewards/margins": 11.9239501953125,
|
| 548 |
+
"rewards/rejected": -11.887507438659668,
|
| 549 |
+
"step": 350
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 1.572052401746725,
|
| 553 |
+
"grad_norm": 0.00824202881851081,
|
| 554 |
+
"learning_rate": 4.333864070046938e-06,
|
| 555 |
+
"logits/chosen": -0.6452621221542358,
|
| 556 |
+
"logits/rejected": -0.2813284993171692,
|
| 557 |
+
"logps/chosen": -266.8562927246094,
|
| 558 |
+
"logps/rejected": -1508.1966552734375,
|
| 559 |
+
"loss": 0.0001,
|
| 560 |
+
"rewards/accuracies": 1.0,
|
| 561 |
+
"rewards/chosen": 0.023952916264533997,
|
| 562 |
+
"rewards/margins": 12.330470085144043,
|
| 563 |
+
"rewards/rejected": -12.306516647338867,
|
| 564 |
+
"step": 360
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 1.6157205240174672,
|
| 568 |
+
"grad_norm": 0.0028510754034206916,
|
| 569 |
+
"learning_rate": 4.28119505058483e-06,
|
| 570 |
+
"logits/chosen": -0.5809464454650879,
|
| 571 |
+
"logits/rejected": -0.29732316732406616,
|
| 572 |
+
"logps/chosen": -292.64154052734375,
|
| 573 |
+
"logps/rejected": -1430.4862060546875,
|
| 574 |
+
"loss": 0.0002,
|
| 575 |
+
"rewards/accuracies": 1.0,
|
| 576 |
+
"rewards/chosen": -0.08253034949302673,
|
| 577 |
+
"rewards/margins": 11.704535484313965,
|
| 578 |
+
"rewards/rejected": -11.787064552307129,
|
| 579 |
+
"step": 370
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 1.6593886462882095,
|
| 583 |
+
"grad_norm": 0.004071842447232605,
|
| 584 |
+
"learning_rate": 4.226869105146658e-06,
|
| 585 |
+
"logits/chosen": -0.6976742148399353,
|
| 586 |
+
"logits/rejected": -0.374077707529068,
|
| 587 |
+
"logps/chosen": -267.6322326660156,
|
| 588 |
+
"logps/rejected": -1627.9765625,
|
| 589 |
+
"loss": 0.0001,
|
| 590 |
+
"rewards/accuracies": 1.0,
|
| 591 |
+
"rewards/chosen": 0.027708619832992554,
|
| 592 |
+
"rewards/margins": 13.307157516479492,
|
| 593 |
+
"rewards/rejected": -13.279451370239258,
|
| 594 |
+
"step": 380
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 1.703056768558952,
|
| 598 |
+
"grad_norm": 0.0030757362239471637,
|
| 599 |
+
"learning_rate": 4.170936769506222e-06,
|
| 600 |
+
"logits/chosen": -0.6652247309684753,
|
| 601 |
+
"logits/rejected": -0.3628092110157013,
|
| 602 |
+
"logps/chosen": -288.57257080078125,
|
| 603 |
+
"logps/rejected": -1520.739990234375,
|
| 604 |
+
"loss": 0.0001,
|
| 605 |
+
"rewards/accuracies": 1.0,
|
| 606 |
+
"rewards/chosen": -0.052382566034793854,
|
| 607 |
+
"rewards/margins": 12.51613998413086,
|
| 608 |
+
"rewards/rejected": -12.568521499633789,
|
| 609 |
+
"step": 390
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 1.7467248908296944,
|
| 613 |
+
"grad_norm": 0.003598764645985237,
|
| 614 |
+
"learning_rate": 4.1134500737541026e-06,
|
| 615 |
+
"logits/chosen": -0.6975280046463013,
|
| 616 |
+
"logits/rejected": -0.3748314380645752,
|
| 617 |
+
"logps/chosen": -277.32843017578125,
|
| 618 |
+
"logps/rejected": -1550.5653076171875,
|
| 619 |
+
"loss": 0.0002,
|
| 620 |
+
"rewards/accuracies": 1.0,
|
| 621 |
+
"rewards/chosen": 0.06987209618091583,
|
| 622 |
+
"rewards/margins": 12.744722366333008,
|
| 623 |
+
"rewards/rejected": -12.674848556518555,
|
| 624 |
+
"step": 400
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 1.7903930131004366,
|
| 628 |
+
"grad_norm": 0.007268762647014337,
|
| 629 |
+
"learning_rate": 4.054462493897569e-06,
|
| 630 |
+
"logits/chosen": -0.7424625754356384,
|
| 631 |
+
"logits/rejected": -0.30923086404800415,
|
| 632 |
+
"logps/chosen": -234.93551635742188,
|
| 633 |
+
"logps/rejected": -1707.387451171875,
|
| 634 |
+
"loss": 0.0002,
|
| 635 |
+
"rewards/accuracies": 1.0,
|
| 636 |
+
"rewards/chosen": 0.10221578180789948,
|
| 637 |
+
"rewards/margins": 13.916183471679688,
|
| 638 |
+
"rewards/rejected": -13.81396770477295,
|
| 639 |
+
"step": 410
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 1.8340611353711789,
|
| 643 |
+
"grad_norm": 0.0008727719544739926,
|
| 644 |
+
"learning_rate": 3.994028902115439e-06,
|
| 645 |
+
"logits/chosen": -0.6766982674598694,
|
| 646 |
+
"logits/rejected": -0.22497108578681946,
|
| 647 |
+
"logps/chosen": -246.74490356445312,
|
| 648 |
+
"logps/rejected": -1593.43115234375,
|
| 649 |
+
"loss": 0.0002,
|
| 650 |
+
"rewards/accuracies": 1.0,
|
| 651 |
+
"rewards/chosen": -0.03030253015458584,
|
| 652 |
+
"rewards/margins": 12.960153579711914,
|
| 653 |
+
"rewards/rejected": -12.990457534790039,
|
| 654 |
+
"step": 420
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 1.8777292576419216,
|
| 658 |
+
"grad_norm": 0.0015384551125673842,
|
| 659 |
+
"learning_rate": 3.932205515714189e-06,
|
| 660 |
+
"logits/chosen": -0.7006328701972961,
|
| 661 |
+
"logits/rejected": -0.38360628485679626,
|
| 662 |
+
"logps/chosen": -298.765380859375,
|
| 663 |
+
"logps/rejected": -1612.6402587890625,
|
| 664 |
+
"loss": 0.0001,
|
| 665 |
+
"rewards/accuracies": 1.0,
|
| 666 |
+
"rewards/chosen": -0.06020375341176987,
|
| 667 |
+
"rewards/margins": 13.307812690734863,
|
| 668 |
+
"rewards/rejected": -13.368016242980957,
|
| 669 |
+
"step": 430
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 1.9213973799126638,
|
| 673 |
+
"grad_norm": 0.008179209300348753,
|
| 674 |
+
"learning_rate": 3.86904984483277e-06,
|
| 675 |
+
"logits/chosen": -0.7087584137916565,
|
| 676 |
+
"logits/rejected": -0.4246200919151306,
|
| 677 |
+
"logps/chosen": -304.3985900878906,
|
| 678 |
+
"logps/rejected": -1534.143798828125,
|
| 679 |
+
"loss": 0.0001,
|
| 680 |
+
"rewards/accuracies": 1.0,
|
| 681 |
+
"rewards/chosen": -0.08279000222682953,
|
| 682 |
+
"rewards/margins": 12.747483253479004,
|
| 683 |
+
"rewards/rejected": -12.830273628234863,
|
| 684 |
+
"step": 440
|
| 685 |
+
},
|
| 686 |
+
{
|
| 687 |
+
"epoch": 1.965065502183406,
|
| 688 |
+
"grad_norm": 0.0014160258405897827,
|
| 689 |
+
"learning_rate": 3.8046206389447916e-06,
|
| 690 |
+
"logits/chosen": -0.6982067823410034,
|
| 691 |
+
"logits/rejected": -0.3726228177547455,
|
| 692 |
+
"logps/chosen": -276.84039306640625,
|
| 693 |
+
"logps/rejected": -1593.4991455078125,
|
| 694 |
+
"loss": 0.0001,
|
| 695 |
+
"rewards/accuracies": 1.0,
|
| 696 |
+
"rewards/chosen": -0.04725092276930809,
|
| 697 |
+
"rewards/margins": 13.1595458984375,
|
| 698 |
+
"rewards/rejected": -13.20679759979248,
|
| 699 |
+
"step": 450
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 2.0087336244541483,
|
| 703 |
+
"grad_norm": 0.0017767298206001502,
|
| 704 |
+
"learning_rate": 3.738977832207839e-06,
|
| 705 |
+
"logits/chosen": -0.6970980763435364,
|
| 706 |
+
"logits/rejected": -0.3680294156074524,
|
| 707 |
+
"logps/chosen": -261.3289794921875,
|
| 708 |
+
"logps/rejected": -1555.7623291015625,
|
| 709 |
+
"loss": 0.0002,
|
| 710 |
+
"rewards/accuracies": 1.0,
|
| 711 |
+
"rewards/chosen": 0.0009517133003100753,
|
| 712 |
+
"rewards/margins": 12.788881301879883,
|
| 713 |
+
"rewards/rejected": -12.787931442260742,
|
| 714 |
+
"step": 460
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 2.052401746724891,
|
| 718 |
+
"grad_norm": 0.001535205426650719,
|
| 719 |
+
"learning_rate": 3.6721824877107588e-06,
|
| 720 |
+
"logits/chosen": -0.7333104014396667,
|
| 721 |
+
"logits/rejected": -0.45936185121536255,
|
| 722 |
+
"logps/chosen": -308.5245666503906,
|
| 723 |
+
"logps/rejected": -1585.037841796875,
|
| 724 |
+
"loss": 0.0001,
|
| 725 |
+
"rewards/accuracies": 1.0,
|
| 726 |
+
"rewards/chosen": -0.0660587027668953,
|
| 727 |
+
"rewards/margins": 13.10444450378418,
|
| 728 |
+
"rewards/rejected": -13.170503616333008,
|
| 729 |
+
"step": 470
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 2.096069868995633,
|
| 733 |
+
"grad_norm": 0.005158531675148358,
|
| 734 |
+
"learning_rate": 3.604296740670768e-06,
|
| 735 |
+
"logits/chosen": -0.7633829116821289,
|
| 736 |
+
"logits/rejected": -0.33872467279434204,
|
| 737 |
+
"logps/chosen": -260.2208557128906,
|
| 738 |
+
"logps/rejected": -1724.6669921875,
|
| 739 |
+
"loss": 0.0001,
|
| 740 |
+
"rewards/accuracies": 1.0,
|
| 741 |
+
"rewards/chosen": 0.010257990099489689,
|
| 742 |
+
"rewards/margins": 14.2053804397583,
|
| 743 |
+
"rewards/rejected": -14.195123672485352,
|
| 744 |
+
"step": 480
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 2.1397379912663754,
|
| 748 |
+
"grad_norm": 0.006727466994391036,
|
| 749 |
+
"learning_rate": 3.5353837406332464e-06,
|
| 750 |
+
"logits/chosen": -0.7862290143966675,
|
| 751 |
+
"logits/rejected": -0.45894408226013184,
|
| 752 |
+
"logps/chosen": -251.74417114257812,
|
| 753 |
+
"logps/rejected": -1686.0699462890625,
|
| 754 |
+
"loss": 0.0001,
|
| 755 |
+
"rewards/accuracies": 1.0,
|
| 756 |
+
"rewards/chosen": 0.0987875908613205,
|
| 757 |
+
"rewards/margins": 13.866643905639648,
|
| 758 |
+
"rewards/rejected": -13.767855644226074,
|
| 759 |
+
"step": 490
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 2.183406113537118,
|
| 763 |
+
"grad_norm": 0.005878784292188106,
|
| 764 |
+
"learning_rate": 3.4655075927279576e-06,
|
| 765 |
+
"logits/chosen": -0.7339245080947876,
|
| 766 |
+
"logits/rejected": -0.42568492889404297,
|
| 767 |
+
"logps/chosen": -266.9426574707031,
|
| 768 |
+
"logps/rejected": -1574.525634765625,
|
| 769 |
+
"loss": 0.0002,
|
| 770 |
+
"rewards/accuracies": 1.0,
|
| 771 |
+
"rewards/chosen": -0.030187349766492844,
|
| 772 |
+
"rewards/margins": 13.049545288085938,
|
| 773 |
+
"rewards/rejected": -13.079732894897461,
|
| 774 |
+
"step": 500
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 2.2270742358078603,
|
| 778 |
+
"grad_norm": 0.0017376152792194543,
|
| 779 |
+
"learning_rate": 3.3947332980363552e-06,
|
| 780 |
+
"logits/chosen": -0.5695234537124634,
|
| 781 |
+
"logits/rejected": -0.19781272113323212,
|
| 782 |
+
"logps/chosen": -270.3812561035156,
|
| 783 |
+
"logps/rejected": -1615.6109619140625,
|
| 784 |
+
"loss": 0.0001,
|
| 785 |
+
"rewards/accuracies": 1.0,
|
| 786 |
+
"rewards/chosen": -0.023705072700977325,
|
| 787 |
+
"rewards/margins": 13.492143630981445,
|
| 788 |
+
"rewards/rejected": -13.515848159790039,
|
| 789 |
+
"step": 510
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 2.2707423580786026,
|
| 793 |
+
"grad_norm": 0.0008884061380883483,
|
| 794 |
+
"learning_rate": 3.3231266931254546e-06,
|
| 795 |
+
"logits/chosen": -0.7268373370170593,
|
| 796 |
+
"logits/rejected": -0.4064566493034363,
|
| 797 |
+
"logps/chosen": -288.7190246582031,
|
| 798 |
+
"logps/rejected": -1659.5142822265625,
|
| 799 |
+
"loss": 0.0001,
|
| 800 |
+
"rewards/accuracies": 1.0,
|
| 801 |
+
"rewards/chosen": -0.004935880657285452,
|
| 802 |
+
"rewards/margins": 13.8540620803833,
|
| 803 |
+
"rewards/rejected": -13.858996391296387,
|
| 804 |
+
"step": 520
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 2.314410480349345,
|
| 808 |
+
"grad_norm": 0.0012752069272254601,
|
| 809 |
+
"learning_rate": 3.250754388804495e-06,
|
| 810 |
+
"logits/chosen": -0.7004576921463013,
|
| 811 |
+
"logits/rejected": -0.3160724639892578,
|
| 812 |
+
"logps/chosen": -260.39801025390625,
|
| 813 |
+
"logps/rejected": -1695.2933349609375,
|
| 814 |
+
"loss": 0.0001,
|
| 815 |
+
"rewards/accuracies": 1.0,
|
| 816 |
+
"rewards/chosen": -0.0040903836488723755,
|
| 817 |
+
"rewards/margins": 14.074589729309082,
|
| 818 |
+
"rewards/rejected": -14.078680038452148,
|
| 819 |
+
"step": 530
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 2.3580786026200875,
|
| 823 |
+
"grad_norm": 0.0010551393838541235,
|
| 824 |
+
"learning_rate": 3.1776837081613893e-06,
|
| 825 |
+
"logits/chosen": -0.6237174868583679,
|
| 826 |
+
"logits/rejected": -0.2775794267654419,
|
| 827 |
+
"logps/chosen": -275.13775634765625,
|
| 828 |
+
"logps/rejected": -1654.786376953125,
|
| 829 |
+
"loss": 0.0002,
|
| 830 |
+
"rewards/accuracies": 1.0,
|
| 831 |
+
"rewards/chosen": -0.028180230408906937,
|
| 832 |
+
"rewards/margins": 13.729413032531738,
|
| 833 |
+
"rewards/rejected": -13.757593154907227,
|
| 834 |
+
"step": 540
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"epoch": 2.4017467248908297,
|
| 838 |
+
"grad_norm": 0.010736168566735408,
|
| 839 |
+
"learning_rate": 3.1039826239365754e-06,
|
| 840 |
+
"logits/chosen": -0.6697626113891602,
|
| 841 |
+
"logits/rejected": -0.3747546076774597,
|
| 842 |
+
"logps/chosen": -265.6369323730469,
|
| 843 |
+
"logps/rejected": -1629.97998046875,
|
| 844 |
+
"loss": 0.0001,
|
| 845 |
+
"rewards/accuracies": 1.0,
|
| 846 |
+
"rewards/chosen": 0.04310118779540062,
|
| 847 |
+
"rewards/margins": 13.556437492370605,
|
| 848 |
+
"rewards/rejected": -13.513336181640625,
|
| 849 |
+
"step": 550
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 2.445414847161572,
|
| 853 |
+
"grad_norm": 0.001439648164211988,
|
| 854 |
+
"learning_rate": 3.0297196952925533e-06,
|
| 855 |
+
"logits/chosen": -0.6776083111763,
|
| 856 |
+
"logits/rejected": -0.4190545678138733,
|
| 857 |
+
"logps/chosen": -304.7749328613281,
|
| 858 |
+
"logps/rejected": -1601.60791015625,
|
| 859 |
+
"loss": 0.0001,
|
| 860 |
+
"rewards/accuracies": 1.0,
|
| 861 |
+
"rewards/chosen": -0.04750397801399231,
|
| 862 |
+
"rewards/margins": 13.419174194335938,
|
| 863 |
+
"rewards/rejected": -13.466680526733398,
|
| 864 |
+
"step": 560
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 2.489082969432314,
|
| 868 |
+
"grad_norm": 0.0013072815191190771,
|
| 869 |
+
"learning_rate": 2.9549640040379043e-06,
|
| 870 |
+
"logits/chosen": -0.7228536605834961,
|
| 871 |
+
"logits/rejected": -0.4134379029273987,
|
| 872 |
+
"logps/chosen": -293.1953430175781,
|
| 873 |
+
"logps/rejected": -1718.9398193359375,
|
| 874 |
+
"loss": 0.0,
|
| 875 |
+
"rewards/accuracies": 1.0,
|
| 876 |
+
"rewards/chosen": -0.06104196235537529,
|
| 877 |
+
"rewards/margins": 14.203579902648926,
|
| 878 |
+
"rewards/rejected": -14.264620780944824,
|
| 879 |
+
"step": 570
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 2.532751091703057,
|
| 883 |
+
"grad_norm": 0.0008461963585448364,
|
| 884 |
+
"learning_rate": 2.8797850903651274e-06,
|
| 885 |
+
"logits/chosen": -0.6659802794456482,
|
| 886 |
+
"logits/rejected": -0.360958993434906,
|
| 887 |
+
"logps/chosen": -270.3370056152344,
|
| 888 |
+
"logps/rejected": -1665.4697265625,
|
| 889 |
+
"loss": 0.0001,
|
| 890 |
+
"rewards/accuracies": 1.0,
|
| 891 |
+
"rewards/chosen": -0.0008804827812127769,
|
| 892 |
+
"rewards/margins": 13.956453323364258,
|
| 893 |
+
"rewards/rejected": -13.9573335647583,
|
| 894 |
+
"step": 580
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"epoch": 2.576419213973799,
|
| 898 |
+
"grad_norm": 0.0023387701751764562,
|
| 899 |
+
"learning_rate": 2.804252888162079e-06,
|
| 900 |
+
"logits/chosen": -0.8161109685897827,
|
| 901 |
+
"logits/rejected": -0.4084259569644928,
|
| 902 |
+
"logps/chosen": -272.53216552734375,
|
| 903 |
+
"logps/rejected": -1786.5172119140625,
|
| 904 |
+
"loss": 0.0001,
|
| 905 |
+
"rewards/accuracies": 1.0,
|
| 906 |
+
"rewards/chosen": -0.03314363956451416,
|
| 907 |
+
"rewards/margins": 14.708500862121582,
|
| 908 |
+
"rewards/rejected": -14.741642951965332,
|
| 909 |
+
"step": 590
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 2.6200873362445414,
|
| 913 |
+
"grad_norm": 0.0018232538424911746,
|
| 914 |
+
"learning_rate": 2.7284376599571776e-06,
|
| 915 |
+
"logits/chosen": -0.7031590342521667,
|
| 916 |
+
"logits/rejected": -0.37507158517837524,
|
| 917 |
+
"logps/chosen": -274.2117919921875,
|
| 918 |
+
"logps/rejected": -1641.595458984375,
|
| 919 |
+
"loss": 0.0001,
|
| 920 |
+
"rewards/accuracies": 1.0,
|
| 921 |
+
"rewards/chosen": -0.04686348885297775,
|
| 922 |
+
"rewards/margins": 13.67021656036377,
|
| 923 |
+
"rewards/rejected": -13.717081069946289,
|
| 924 |
+
"step": 600
|
| 925 |
+
},
|
| 926 |
+
{
|
| 927 |
+
"epoch": 2.6637554585152836,
|
| 928 |
+
"grad_norm": 0.000990528476960238,
|
| 929 |
+
"learning_rate": 2.652409931558898e-06,
|
| 930 |
+
"logits/chosen": -0.7374690175056458,
|
| 931 |
+
"logits/rejected": -0.43666791915893555,
|
| 932 |
+
"logps/chosen": -283.0863342285156,
|
| 933 |
+
"logps/rejected": -1705.676513671875,
|
| 934 |
+
"loss": 0.0,
|
| 935 |
+
"rewards/accuracies": 1.0,
|
| 936 |
+
"rewards/chosen": -0.040116313844919205,
|
| 937 |
+
"rewards/margins": 14.263372421264648,
|
| 938 |
+
"rewards/rejected": -14.303491592407227,
|
| 939 |
+
"step": 610
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 2.7074235807860263,
|
| 943 |
+
"grad_norm": 0.0009319562615286963,
|
| 944 |
+
"learning_rate": 2.5762404264503538e-06,
|
| 945 |
+
"logits/chosen": -0.6617239713668823,
|
| 946 |
+
"logits/rejected": -0.28903594613075256,
|
| 947 |
+
"logps/chosen": -267.43609619140625,
|
| 948 |
+
"logps/rejected": -1748.5576171875,
|
| 949 |
+
"loss": 0.0001,
|
| 950 |
+
"rewards/accuracies": 1.0,
|
| 951 |
+
"rewards/chosen": 0.03135783225297928,
|
| 952 |
+
"rewards/margins": 14.624079704284668,
|
| 953 |
+
"rewards/rejected": -14.592721939086914,
|
| 954 |
+
"step": 620
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 2.7510917030567685,
|
| 958 |
+
"grad_norm": 0.013526422059541143,
|
| 959 |
+
"learning_rate": 2.5e-06,
|
| 960 |
+
"logits/chosen": -0.7105010151863098,
|
| 961 |
+
"logits/rejected": -0.4328455328941345,
|
| 962 |
+
"logps/chosen": -281.528564453125,
|
| 963 |
+
"logps/rejected": -1675.9830322265625,
|
| 964 |
+
"loss": 0.0001,
|
| 965 |
+
"rewards/accuracies": 1.0,
|
| 966 |
+
"rewards/chosen": -0.0196510162204504,
|
| 967 |
+
"rewards/margins": 13.972375869750977,
|
| 968 |
+
"rewards/rejected": -13.992025375366211,
|
| 969 |
+
"step": 630
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 2.7947598253275108,
|
| 973 |
+
"grad_norm": 0.0003686574185361761,
|
| 974 |
+
"learning_rate": 2.423759573549647e-06,
|
| 975 |
+
"logits/chosen": -0.6781324148178101,
|
| 976 |
+
"logits/rejected": -0.2962037920951843,
|
| 977 |
+
"logps/chosen": -263.18975830078125,
|
| 978 |
+
"logps/rejected": -1710.5943603515625,
|
| 979 |
+
"loss": 0.0,
|
| 980 |
+
"rewards/accuracies": 1.0,
|
| 981 |
+
"rewards/chosen": -0.022569647058844566,
|
| 982 |
+
"rewards/margins": 14.243792533874512,
|
| 983 |
+
"rewards/rejected": -14.266362190246582,
|
| 984 |
+
"step": 640
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 2.8384279475982535,
|
| 988 |
+
"grad_norm": 0.0009413555719320386,
|
| 989 |
+
"learning_rate": 2.3475900684411027e-06,
|
| 990 |
+
"logits/chosen": -0.6239091157913208,
|
| 991 |
+
"logits/rejected": -0.22716188430786133,
|
| 992 |
+
"logps/chosen": -257.69842529296875,
|
| 993 |
+
"logps/rejected": -1695.5068359375,
|
| 994 |
+
"loss": 0.0,
|
| 995 |
+
"rewards/accuracies": 1.0,
|
| 996 |
+
"rewards/chosen": -0.07734155654907227,
|
| 997 |
+
"rewards/margins": 14.082303047180176,
|
| 998 |
+
"rewards/rejected": -14.159643173217773,
|
| 999 |
+
"step": 650
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 2.8820960698689957,
|
| 1003 |
+
"grad_norm": 0.000759592485033725,
|
| 1004 |
+
"learning_rate": 2.2715623400428228e-06,
|
| 1005 |
+
"logits/chosen": -0.694677472114563,
|
| 1006 |
+
"logits/rejected": -0.3776795268058777,
|
| 1007 |
+
"logps/chosen": -282.89202880859375,
|
| 1008 |
+
"logps/rejected": -1710.0386962890625,
|
| 1009 |
+
"loss": 0.0001,
|
| 1010 |
+
"rewards/accuracies": 1.0,
|
| 1011 |
+
"rewards/chosen": 0.0032756985165178776,
|
| 1012 |
+
"rewards/margins": 14.365133285522461,
|
| 1013 |
+
"rewards/rejected": -14.361857414245605,
|
| 1014 |
+
"step": 660
|
| 1015 |
+
},
|
| 1016 |
+
{
|
| 1017 |
+
"epoch": 2.925764192139738,
|
| 1018 |
+
"grad_norm": 0.019784253772424126,
|
| 1019 |
+
"learning_rate": 2.1957471118379213e-06,
|
| 1020 |
+
"logits/chosen": -0.7745624780654907,
|
| 1021 |
+
"logits/rejected": -0.4288128912448883,
|
| 1022 |
+
"logps/chosen": -304.70574951171875,
|
| 1023 |
+
"logps/rejected": -1791.9283447265625,
|
| 1024 |
+
"loss": 0.0001,
|
| 1025 |
+
"rewards/accuracies": 1.0,
|
| 1026 |
+
"rewards/chosen": -0.05531958490610123,
|
| 1027 |
+
"rewards/margins": 14.97740364074707,
|
| 1028 |
+
"rewards/rejected": -15.032722473144531,
|
| 1029 |
+
"step": 670
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 2.96943231441048,
|
| 1033 |
+
"grad_norm": 0.004147931478685181,
|
| 1034 |
+
"learning_rate": 2.120214909634873e-06,
|
| 1035 |
+
"logits/chosen": -0.7153472304344177,
|
| 1036 |
+
"logits/rejected": -0.43083199858665466,
|
| 1037 |
+
"logps/chosen": -297.9048767089844,
|
| 1038 |
+
"logps/rejected": -1724.3980712890625,
|
| 1039 |
+
"loss": 0.0001,
|
| 1040 |
+
"rewards/accuracies": 1.0,
|
| 1041 |
+
"rewards/chosen": -0.03991527482867241,
|
| 1042 |
+
"rewards/margins": 14.473901748657227,
|
| 1043 |
+
"rewards/rejected": -14.513816833496094,
|
| 1044 |
+
"step": 680
|
| 1045 |
+
},
|
| 1046 |
+
{
|
| 1047 |
+
"epoch": 3.013100436681223,
|
| 1048 |
+
"grad_norm": 0.0013738234416668116,
|
| 1049 |
+
"learning_rate": 2.045035995962097e-06,
|
| 1050 |
+
"logits/chosen": -0.7798169255256653,
|
| 1051 |
+
"logits/rejected": -0.43747028708457947,
|
| 1052 |
+
"logps/chosen": -279.0494079589844,
|
| 1053 |
+
"logps/rejected": -1814.386962890625,
|
| 1054 |
+
"loss": 0.0001,
|
| 1055 |
+
"rewards/accuracies": 1.0,
|
| 1056 |
+
"rewards/chosen": -0.06170396879315376,
|
| 1057 |
+
"rewards/margins": 15.159332275390625,
|
| 1058 |
+
"rewards/rejected": -15.221035957336426,
|
| 1059 |
+
"step": 690
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 3.056768558951965,
|
| 1063 |
+
"grad_norm": 0.0031209059954659443,
|
| 1064 |
+
"learning_rate": 1.970280304707447e-06,
|
| 1065 |
+
"logits/chosen": -0.714453399181366,
|
| 1066 |
+
"logits/rejected": -0.36034974455833435,
|
| 1067 |
+
"logps/chosen": -269.9892272949219,
|
| 1068 |
+
"logps/rejected": -1740.939208984375,
|
| 1069 |
+
"loss": 0.0001,
|
| 1070 |
+
"rewards/accuracies": 1.0,
|
| 1071 |
+
"rewards/chosen": 0.002551588462665677,
|
| 1072 |
+
"rewards/margins": 14.63127326965332,
|
| 1073 |
+
"rewards/rejected": -14.62872314453125,
|
| 1074 |
+
"step": 700
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"epoch": 3.1004366812227073,
|
| 1078 |
+
"grad_norm": 0.00034210696214912053,
|
| 1079 |
+
"learning_rate": 1.8960173760634257e-06,
|
| 1080 |
+
"logits/chosen": -0.7932115793228149,
|
| 1081 |
+
"logits/rejected": -0.5189129114151001,
|
| 1082 |
+
"logps/chosen": -322.1393737792969,
|
| 1083 |
+
"logps/rejected": -1711.671875,
|
| 1084 |
+
"loss": 0.0001,
|
| 1085 |
+
"rewards/accuracies": 1.0,
|
| 1086 |
+
"rewards/chosen": -0.08139355480670929,
|
| 1087 |
+
"rewards/margins": 14.367878913879395,
|
| 1088 |
+
"rewards/rejected": -14.449272155761719,
|
| 1089 |
+
"step": 710
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 3.14410480349345,
|
| 1093 |
+
"grad_norm": 0.0013002362609120892,
|
| 1094 |
+
"learning_rate": 1.8223162918386122e-06,
|
| 1095 |
+
"logits/chosen": -0.7495840787887573,
|
| 1096 |
+
"logits/rejected": -0.3989501893520355,
|
| 1097 |
+
"logps/chosen": -285.207275390625,
|
| 1098 |
+
"logps/rejected": -1773.8642578125,
|
| 1099 |
+
"loss": 0.0,
|
| 1100 |
+
"rewards/accuracies": 1.0,
|
| 1101 |
+
"rewards/chosen": -0.09717712551355362,
|
| 1102 |
+
"rewards/margins": 14.81560230255127,
|
| 1103 |
+
"rewards/rejected": -14.9127779006958,
|
| 1104 |
+
"step": 720
|
| 1105 |
+
},
|
| 1106 |
+
{
|
| 1107 |
+
"epoch": 3.1877729257641922,
|
| 1108 |
+
"grad_norm": 0.008009764157245521,
|
| 1109 |
+
"learning_rate": 1.7492456111955052e-06,
|
| 1110 |
+
"logits/chosen": -0.6571623086929321,
|
| 1111 |
+
"logits/rejected": -0.30920231342315674,
|
| 1112 |
+
"logps/chosen": -278.0492248535156,
|
| 1113 |
+
"logps/rejected": -1723.248046875,
|
| 1114 |
+
"loss": 0.0001,
|
| 1115 |
+
"rewards/accuracies": 1.0,
|
| 1116 |
+
"rewards/chosen": -0.0555524006485939,
|
| 1117 |
+
"rewards/margins": 14.493664741516113,
|
| 1118 |
+
"rewards/rejected": -14.549217224121094,
|
| 1119 |
+
"step": 730
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 3.2314410480349345,
|
| 1123 |
+
"grad_norm": 0.002618631891345944,
|
| 1124 |
+
"learning_rate": 1.6768733068745468e-06,
|
| 1125 |
+
"logits/chosen": -0.7488084435462952,
|
| 1126 |
+
"logits/rejected": -0.4106837809085846,
|
| 1127 |
+
"logps/chosen": -282.20538330078125,
|
| 1128 |
+
"logps/rejected": -1826.658203125,
|
| 1129 |
+
"loss": 0.0001,
|
| 1130 |
+
"rewards/accuracies": 1.0,
|
| 1131 |
+
"rewards/chosen": -0.08377159386873245,
|
| 1132 |
+
"rewards/margins": 15.273724555969238,
|
| 1133 |
+
"rewards/rejected": -15.35749626159668,
|
| 1134 |
+
"step": 740
|
| 1135 |
+
},
|
| 1136 |
+
{
|
| 1137 |
+
"epoch": 3.2751091703056767,
|
| 1138 |
+
"grad_norm": 0.00023217650976632204,
|
| 1139 |
+
"learning_rate": 1.6052667019636462e-06,
|
| 1140 |
+
"logits/chosen": -0.7862998247146606,
|
| 1141 |
+
"logits/rejected": -0.44631749391555786,
|
| 1142 |
+
"logps/chosen": -276.70867919921875,
|
| 1143 |
+
"logps/rejected": -1797.5888671875,
|
| 1144 |
+
"loss": 0.0001,
|
| 1145 |
+
"rewards/accuracies": 1.0,
|
| 1146 |
+
"rewards/chosen": -0.024743851274251938,
|
| 1147 |
+
"rewards/margins": 15.001612663269043,
|
| 1148 |
+
"rewards/rejected": -15.026357650756836,
|
| 1149 |
+
"step": 750
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 3.3187772925764194,
|
| 1153 |
+
"grad_norm": 0.01865811559978352,
|
| 1154 |
+
"learning_rate": 1.5344924072720434e-06,
|
| 1155 |
+
"logits/chosen": -0.7188819050788879,
|
| 1156 |
+
"logits/rejected": -0.4172704815864563,
|
| 1157 |
+
"logps/chosen": -285.13140869140625,
|
| 1158 |
+
"logps/rejected": -1677.7828369140625,
|
| 1159 |
+
"loss": 0.0001,
|
| 1160 |
+
"rewards/accuracies": 1.0,
|
| 1161 |
+
"rewards/chosen": -0.061395276337862015,
|
| 1162 |
+
"rewards/margins": 14.092081069946289,
|
| 1163 |
+
"rewards/rejected": -14.153475761413574,
|
| 1164 |
+
"step": 760
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 3.3624454148471616,
|
| 1168 |
+
"grad_norm": 0.00039610874442709104,
|
| 1169 |
+
"learning_rate": 1.4646162593667535e-06,
|
| 1170 |
+
"logits/chosen": -0.7479511499404907,
|
| 1171 |
+
"logits/rejected": -0.4791211187839508,
|
| 1172 |
+
"logps/chosen": -314.30169677734375,
|
| 1173 |
+
"logps/rejected": -1737.6015625,
|
| 1174 |
+
"loss": 0.0,
|
| 1175 |
+
"rewards/accuracies": 1.0,
|
| 1176 |
+
"rewards/chosen": -0.053110092878341675,
|
| 1177 |
+
"rewards/margins": 14.654457092285156,
|
| 1178 |
+
"rewards/rejected": -14.70756721496582,
|
| 1179 |
+
"step": 770
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 3.406113537117904,
|
| 1183 |
+
"grad_norm": 0.0020264522877018994,
|
| 1184 |
+
"learning_rate": 1.3957032593292319e-06,
|
| 1185 |
+
"logits/chosen": -0.6951030492782593,
|
| 1186 |
+
"logits/rejected": -0.34219890832901,
|
| 1187 |
+
"logps/chosen": -285.89373779296875,
|
| 1188 |
+
"logps/rejected": -1782.4566650390625,
|
| 1189 |
+
"loss": 0.0001,
|
| 1190 |
+
"rewards/accuracies": 1.0,
|
| 1191 |
+
"rewards/chosen": -0.08678509294986725,
|
| 1192 |
+
"rewards/margins": 14.822759628295898,
|
| 1193 |
+
"rewards/rejected": -14.9095458984375,
|
| 1194 |
+
"step": 780
|
| 1195 |
+
},
|
| 1196 |
+
{
|
| 1197 |
+
"epoch": 3.449781659388646,
|
| 1198 |
+
"grad_norm": 0.0018935186827857544,
|
| 1199 |
+
"learning_rate": 1.3278175122892416e-06,
|
| 1200 |
+
"logits/chosen": -0.6852272748947144,
|
| 1201 |
+
"logits/rejected": -0.33605360984802246,
|
| 1202 |
+
"logps/chosen": -259.1459045410156,
|
| 1203 |
+
"logps/rejected": -1725.716796875,
|
| 1204 |
+
"loss": 0.0001,
|
| 1205 |
+
"rewards/accuracies": 1.0,
|
| 1206 |
+
"rewards/chosen": -0.036944665014743805,
|
| 1207 |
+
"rewards/margins": 14.497613906860352,
|
| 1208 |
+
"rewards/rejected": -14.534558296203613,
|
| 1209 |
+
"step": 790
|
| 1210 |
+
},
|
| 1211 |
+
{
|
| 1212 |
+
"epoch": 3.493449781659389,
|
| 1213 |
+
"grad_norm": 0.015223829550391095,
|
| 1214 |
+
"learning_rate": 1.261022167792161e-06,
|
| 1215 |
+
"logits/chosen": -0.7397433519363403,
|
| 1216 |
+
"logits/rejected": -0.36608046293258667,
|
| 1217 |
+
"logps/chosen": -281.9190673828125,
|
| 1218 |
+
"logps/rejected": -1852.1806640625,
|
| 1219 |
+
"loss": 0.0001,
|
| 1220 |
+
"rewards/accuracies": 1.0,
|
| 1221 |
+
"rewards/chosen": -0.03298696130514145,
|
| 1222 |
+
"rewards/margins": 15.54675006866455,
|
| 1223 |
+
"rewards/rejected": -15.579736709594727,
|
| 1224 |
+
"step": 800
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 3.537117903930131,
|
| 1228 |
+
"grad_norm": 0.010378807533355724,
|
| 1229 |
+
"learning_rate": 1.195379361055209e-06,
|
| 1230 |
+
"logits/chosen": -0.6590791940689087,
|
| 1231 |
+
"logits/rejected": -0.28567662835121155,
|
| 1232 |
+
"logps/chosen": -273.66009521484375,
|
| 1233 |
+
"logps/rejected": -1854.084716796875,
|
| 1234 |
+
"loss": 0.0,
|
| 1235 |
+
"rewards/accuracies": 1.0,
|
| 1236 |
+
"rewards/chosen": -0.010397704318165779,
|
| 1237 |
+
"rewards/margins": 15.640172004699707,
|
| 1238 |
+
"rewards/rejected": -15.650568962097168,
|
| 1239 |
+
"step": 810
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 3.5807860262008733,
|
| 1243 |
+
"grad_norm": 0.016204955970958463,
|
| 1244 |
+
"learning_rate": 1.1309501551672303e-06,
|
| 1245 |
+
"logits/chosen": -0.5972493886947632,
|
| 1246 |
+
"logits/rejected": -0.2114742249250412,
|
| 1247 |
+
"logps/chosen": -269.32476806640625,
|
| 1248 |
+
"logps/rejected": -1736.829833984375,
|
| 1249 |
+
"loss": 0.0001,
|
| 1250 |
+
"rewards/accuracies": 1.0,
|
| 1251 |
+
"rewards/chosen": -0.052963800728321075,
|
| 1252 |
+
"rewards/margins": 14.625813484191895,
|
| 1253 |
+
"rewards/rejected": -14.678777694702148,
|
| 1254 |
+
"step": 820
|
| 1255 |
+
},
|
| 1256 |
+
{
|
| 1257 |
+
"epoch": 3.6244541484716155,
|
| 1258 |
+
"grad_norm": 0.0008894376854101312,
|
| 1259 |
+
"learning_rate": 1.0677944842858112e-06,
|
| 1260 |
+
"logits/chosen": -0.6675132513046265,
|
| 1261 |
+
"logits/rejected": -0.26750123500823975,
|
| 1262 |
+
"logps/chosen": -279.586181640625,
|
| 1263 |
+
"logps/rejected": -1758.112060546875,
|
| 1264 |
+
"loss": 0.0,
|
| 1265 |
+
"rewards/accuracies": 1.0,
|
| 1266 |
+
"rewards/chosen": -0.050345610827207565,
|
| 1267 |
+
"rewards/margins": 14.880023002624512,
|
| 1268 |
+
"rewards/rejected": -14.930368423461914,
|
| 1269 |
+
"step": 830
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 3.668122270742358,
|
| 1273 |
+
"grad_norm": 0.0004087774037557016,
|
| 1274 |
+
"learning_rate": 1.005971097884561e-06,
|
| 1275 |
+
"logits/chosen": -0.705937922000885,
|
| 1276 |
+
"logits/rejected": -0.3553905785083771,
|
| 1277 |
+
"logps/chosen": -300.83929443359375,
|
| 1278 |
+
"logps/rejected": -1823.3248291015625,
|
| 1279 |
+
"loss": 0.0001,
|
| 1280 |
+
"rewards/accuracies": 1.0,
|
| 1281 |
+
"rewards/chosen": -0.14001181721687317,
|
| 1282 |
+
"rewards/margins": 15.270605087280273,
|
| 1283 |
+
"rewards/rejected": -15.410616874694824,
|
| 1284 |
+
"step": 840
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 3.7117903930131004,
|
| 1288 |
+
"grad_norm": 0.0006949799978363079,
|
| 1289 |
+
"learning_rate": 9.455375061024319e-07,
|
| 1290 |
+
"logits/chosen": -0.7351874709129333,
|
| 1291 |
+
"logits/rejected": -0.3912436366081238,
|
| 1292 |
+
"logps/chosen": -282.2884826660156,
|
| 1293 |
+
"logps/rejected": -1754.314453125,
|
| 1294 |
+
"loss": 0.0001,
|
| 1295 |
+
"rewards/accuracies": 1.0,
|
| 1296 |
+
"rewards/chosen": -0.06751924753189087,
|
| 1297 |
+
"rewards/margins": 14.702392578125,
|
| 1298 |
+
"rewards/rejected": -14.769912719726562,
|
| 1299 |
+
"step": 850
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 3.7554585152838427,
|
| 1303 |
+
"grad_norm": 0.006463044951812143,
|
| 1304 |
+
"learning_rate": 8.86549926245898e-07,
|
| 1305 |
+
"logits/chosen": -0.6956661939620972,
|
| 1306 |
+
"logits/rejected": -0.3181700110435486,
|
| 1307 |
+
"logps/chosen": -287.1238708496094,
|
| 1308 |
+
"logps/rejected": -1798.494384765625,
|
| 1309 |
+
"loss": 0.0001,
|
| 1310 |
+
"rewards/accuracies": 1.0,
|
| 1311 |
+
"rewards/chosen": -0.10201337188482285,
|
| 1312 |
+
"rewards/margins": 14.971240043640137,
|
| 1313 |
+
"rewards/rejected": -15.073251724243164,
|
| 1314 |
+
"step": 860
|
| 1315 |
+
},
|
| 1316 |
+
{
|
| 1317 |
+
"epoch": 3.7991266375545854,
|
| 1318 |
+
"grad_norm": 0.0007296140394687955,
|
| 1319 |
+
"learning_rate": 8.29063230493779e-07,
|
| 1320 |
+
"logits/chosen": -0.7505240440368652,
|
| 1321 |
+
"logits/rejected": -0.3770305812358856,
|
| 1322 |
+
"logps/chosen": -265.15679931640625,
|
| 1323 |
+
"logps/rejected": -1858.8929443359375,
|
| 1324 |
+
"loss": 0.0001,
|
| 1325 |
+
"rewards/accuracies": 1.0,
|
| 1326 |
+
"rewards/chosen": -0.02774529531598091,
|
| 1327 |
+
"rewards/margins": 15.584371566772461,
|
| 1328 |
+
"rewards/rejected": -15.6121187210083,
|
| 1329 |
+
"step": 870
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 3.8427947598253276,
|
| 1333 |
+
"grad_norm": 0.008816147256943409,
|
| 1334 |
+
"learning_rate": 7.731308948533431e-07,
|
| 1335 |
+
"logits/chosen": -0.7113201022148132,
|
| 1336 |
+
"logits/rejected": -0.4147794246673584,
|
| 1337 |
+
"logps/chosen": -302.5737609863281,
|
| 1338 |
+
"logps/rejected": -1753.0826416015625,
|
| 1339 |
+
"loss": 0.0,
|
| 1340 |
+
"rewards/accuracies": 1.0,
|
| 1341 |
+
"rewards/chosen": -0.04450942575931549,
|
| 1342 |
+
"rewards/margins": 14.832493782043457,
|
| 1343 |
+
"rewards/rejected": -14.87700366973877,
|
| 1344 |
+
"step": 880
|
| 1345 |
+
},
|
| 1346 |
+
{
|
| 1347 |
+
"epoch": 3.88646288209607,
|
| 1348 |
+
"grad_norm": 0.005372270249774767,
|
| 1349 |
+
"learning_rate": 7.188049494151703e-07,
|
| 1350 |
+
"logits/chosen": -0.6462150812149048,
|
| 1351 |
+
"logits/rejected": -0.2772805094718933,
|
| 1352 |
+
"logps/chosen": -277.2967834472656,
|
| 1353 |
+
"logps/rejected": -1723.9173583984375,
|
| 1354 |
+
"loss": 0.0001,
|
| 1355 |
+
"rewards/accuracies": 1.0,
|
| 1356 |
+
"rewards/chosen": -0.0981786921620369,
|
| 1357 |
+
"rewards/margins": 14.48997688293457,
|
| 1358 |
+
"rewards/rejected": -14.588155746459961,
|
| 1359 |
+
"step": 890
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"epoch": 3.930131004366812,
|
| 1363 |
+
"grad_norm": 0.0007509813802026183,
|
| 1364 |
+
"learning_rate": 6.661359299530626e-07,
|
| 1365 |
+
"logits/chosen": -0.6561316847801208,
|
| 1366 |
+
"logits/rejected": -0.47288408875465393,
|
| 1367 |
+
"logps/chosen": -322.491455078125,
|
| 1368 |
+
"logps/rejected": -1582.517333984375,
|
| 1369 |
+
"loss": 0.0001,
|
| 1370 |
+
"rewards/accuracies": 1.0,
|
| 1371 |
+
"rewards/chosen": -0.14367668330669403,
|
| 1372 |
+
"rewards/margins": 13.482531547546387,
|
| 1373 |
+
"rewards/rejected": -13.626208305358887,
|
| 1374 |
+
"step": 900
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 3.9737991266375547,
|
| 1378 |
+
"grad_norm": 0.002905923093028766,
|
| 1379 |
+
"learning_rate": 6.151728309140071e-07,
|
| 1380 |
+
"logits/chosen": -0.7203381061553955,
|
| 1381 |
+
"logits/rejected": -0.382064551115036,
|
| 1382 |
+
"logps/chosen": -285.7363586425781,
|
| 1383 |
+
"logps/rejected": -1801.4898681640625,
|
| 1384 |
+
"loss": 0.0001,
|
| 1385 |
+
"rewards/accuracies": 1.0,
|
| 1386 |
+
"rewards/chosen": -0.07807151228189468,
|
| 1387 |
+
"rewards/margins": 15.15675163269043,
|
| 1388 |
+
"rewards/rejected": -15.234823226928711,
|
| 1389 |
+
"step": 910
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 4.0174672489082965,
|
| 1393 |
+
"grad_norm": 0.0020680720730432436,
|
| 1394 |
+
"learning_rate": 5.659630598419391e-07,
|
| 1395 |
+
"logits/chosen": -0.724728524684906,
|
| 1396 |
+
"logits/rejected": -0.3629612326622009,
|
| 1397 |
+
"logps/chosen": -295.44073486328125,
|
| 1398 |
+
"logps/rejected": -1742.753173828125,
|
| 1399 |
+
"loss": 0.0001,
|
| 1400 |
+
"rewards/accuracies": 1.0,
|
| 1401 |
+
"rewards/chosen": -0.13527199625968933,
|
| 1402 |
+
"rewards/margins": 14.602720260620117,
|
| 1403 |
+
"rewards/rejected": -14.737993240356445,
|
| 1404 |
+
"step": 920
|
| 1405 |
+
},
|
| 1406 |
+
{
|
| 1407 |
+
"epoch": 4.06113537117904,
|
| 1408 |
+
"grad_norm": 0.012272686016895323,
|
| 1409 |
+
"learning_rate": 5.185523932776923e-07,
|
| 1410 |
+
"logits/chosen": -0.6944893598556519,
|
| 1411 |
+
"logits/rejected": -0.3898335099220276,
|
| 1412 |
+
"logps/chosen": -287.7585144042969,
|
| 1413 |
+
"logps/rejected": -1794.3408203125,
|
| 1414 |
+
"loss": 0.0001,
|
| 1415 |
+
"rewards/accuracies": 1.0,
|
| 1416 |
+
"rewards/chosen": -0.05893585830926895,
|
| 1417 |
+
"rewards/margins": 15.143773078918457,
|
| 1418 |
+
"rewards/rejected": -15.202710151672363,
|
| 1419 |
+
"step": 930
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"epoch": 4.104803493449782,
|
| 1423 |
+
"grad_norm": 0.0053926273432672945,
|
| 1424 |
+
"learning_rate": 4.7298493417616024e-07,
|
| 1425 |
+
"logits/chosen": -0.6781089305877686,
|
| 1426 |
+
"logits/rejected": -0.40882062911987305,
|
| 1427 |
+
"logps/chosen": -303.1540222167969,
|
| 1428 |
+
"logps/rejected": -1783.308837890625,
|
| 1429 |
+
"loss": 0.0001,
|
| 1430 |
+
"rewards/accuracies": 1.0,
|
| 1431 |
+
"rewards/chosen": -0.1440161019563675,
|
| 1432 |
+
"rewards/margins": 14.9929780960083,
|
| 1433 |
+
"rewards/rejected": -15.136993408203125,
|
| 1434 |
+
"step": 940
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 4.148471615720524,
|
| 1438 |
+
"grad_norm": 0.0010027396216351911,
|
| 1439 |
+
"learning_rate": 4.293030708802834e-07,
|
| 1440 |
+
"logits/chosen": -0.7281903028488159,
|
| 1441 |
+
"logits/rejected": -0.4005254805088043,
|
| 1442 |
+
"logps/chosen": -281.6249694824219,
|
| 1443 |
+
"logps/rejected": -1874.48828125,
|
| 1444 |
+
"loss": 0.0,
|
| 1445 |
+
"rewards/accuracies": 1.0,
|
| 1446 |
+
"rewards/chosen": -0.086191326379776,
|
| 1447 |
+
"rewards/margins": 15.799471855163574,
|
| 1448 |
+
"rewards/rejected": -15.885663032531738,
|
| 1449 |
+
"step": 950
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 4.192139737991266,
|
| 1453 |
+
"grad_norm": 0.0027350722737438755,
|
| 1454 |
+
"learning_rate": 3.875474376900254e-07,
|
| 1455 |
+
"logits/chosen": -0.6054420471191406,
|
| 1456 |
+
"logits/rejected": -0.2656559944152832,
|
| 1457 |
+
"logps/chosen": -283.4683532714844,
|
| 1458 |
+
"logps/rejected": -1737.546630859375,
|
| 1459 |
+
"loss": 0.0,
|
| 1460 |
+
"rewards/accuracies": 1.0,
|
| 1461 |
+
"rewards/chosen": -0.1006595641374588,
|
| 1462 |
+
"rewards/margins": 14.711895942687988,
|
| 1463 |
+
"rewards/rejected": -14.812556266784668,
|
| 1464 |
+
"step": 960
|
| 1465 |
+
},
|
| 1466 |
+
{
|
| 1467 |
+
"epoch": 4.235807860262009,
|
| 1468 |
+
"grad_norm": 0.006760599239543763,
|
| 1469 |
+
"learning_rate": 3.4775687706301437e-07,
|
| 1470 |
+
"logits/chosen": -0.7967843413352966,
|
| 1471 |
+
"logits/rejected": -0.42173296213150024,
|
| 1472 |
+
"logps/chosen": -270.96148681640625,
|
| 1473 |
+
"logps/rejected": -1896.1468505859375,
|
| 1474 |
+
"loss": 0.0,
|
| 1475 |
+
"rewards/accuracies": 1.0,
|
| 1476 |
+
"rewards/chosen": -0.02015690691769123,
|
| 1477 |
+
"rewards/margins": 15.979083061218262,
|
| 1478 |
+
"rewards/rejected": -15.999239921569824,
|
| 1479 |
+
"step": 970
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 4.279475982532751,
|
| 1483 |
+
"grad_norm": 0.0008520732124624599,
|
| 1484 |
+
"learning_rate": 3.0996840348201717e-07,
|
| 1485 |
+
"logits/chosen": -0.7633962631225586,
|
| 1486 |
+
"logits/rejected": -0.4977475106716156,
|
| 1487 |
+
"logps/chosen": -313.98822021484375,
|
| 1488 |
+
"logps/rejected": -1697.9468994140625,
|
| 1489 |
+
"loss": 0.0,
|
| 1490 |
+
"rewards/accuracies": 1.0,
|
| 1491 |
+
"rewards/chosen": -0.18659470975399017,
|
| 1492 |
+
"rewards/margins": 14.239301681518555,
|
| 1493 |
+
"rewards/rejected": -14.425895690917969,
|
| 1494 |
+
"step": 980
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 4.323144104803493,
|
| 1498 |
+
"grad_norm": 0.0013051411886959893,
|
| 1499 |
+
"learning_rate": 2.742171690228562e-07,
|
| 1500 |
+
"logits/chosen": -0.7120814323425293,
|
| 1501 |
+
"logits/rejected": -0.3730967342853546,
|
| 1502 |
+
"logps/chosen": -298.5450134277344,
|
| 1503 |
+
"logps/rejected": -1744.666015625,
|
| 1504 |
+
"loss": 0.0001,
|
| 1505 |
+
"rewards/accuracies": 1.0,
|
| 1506 |
+
"rewards/chosen": -0.15538612008094788,
|
| 1507 |
+
"rewards/margins": 14.598345756530762,
|
| 1508 |
+
"rewards/rejected": -14.75373363494873,
|
| 1509 |
+
"step": 990
|
| 1510 |
+
},
|
| 1511 |
+
{
|
| 1512 |
+
"epoch": 4.366812227074236,
|
| 1513 |
+
"grad_norm": 0.002698793985143946,
|
| 1514 |
+
"learning_rate": 2.405364306547955e-07,
|
| 1515 |
+
"logits/chosen": -0.7172932028770447,
|
| 1516 |
+
"logits/rejected": -0.47384780645370483,
|
| 1517 |
+
"logps/chosen": -322.2046813964844,
|
| 1518 |
+
"logps/rejected": -1694.6380615234375,
|
| 1519 |
+
"loss": 0.0,
|
| 1520 |
+
"rewards/accuracies": 1.0,
|
| 1521 |
+
"rewards/chosen": -0.10717038065195084,
|
| 1522 |
+
"rewards/margins": 14.218195915222168,
|
| 1523 |
+
"rewards/rejected": -14.325366020202637,
|
| 1524 |
+
"step": 1000
|
| 1525 |
+
},
|
| 1526 |
+
{
|
| 1527 |
+
"epoch": 4.4104803493449785,
|
| 1528 |
+
"grad_norm": 0.00023879898445148078,
|
| 1529 |
+
"learning_rate": 2.0895751930382125e-07,
|
| 1530 |
+
"logits/chosen": -0.6613215804100037,
|
| 1531 |
+
"logits/rejected": -0.33443623781204224,
|
| 1532 |
+
"logps/chosen": -295.16278076171875,
|
| 1533 |
+
"logps/rejected": -1768.1917724609375,
|
| 1534 |
+
"loss": 0.0,
|
| 1535 |
+
"rewards/accuracies": 1.0,
|
| 1536 |
+
"rewards/chosen": -0.03670508787035942,
|
| 1537 |
+
"rewards/margins": 14.837631225585938,
|
| 1538 |
+
"rewards/rejected": -14.874334335327148,
|
| 1539 |
+
"step": 1010
|
| 1540 |
+
},
|
| 1541 |
+
{
|
| 1542 |
+
"epoch": 4.454148471615721,
|
| 1543 |
+
"grad_norm": 0.00023568210161956636,
|
| 1544 |
+
"learning_rate": 1.7950981070758488e-07,
|
| 1545 |
+
"logits/chosen": -0.6994235515594482,
|
| 1546 |
+
"logits/rejected": -0.41796016693115234,
|
| 1547 |
+
"logps/chosen": -287.3175048828125,
|
| 1548 |
+
"logps/rejected": -1685.6611328125,
|
| 1549 |
+
"loss": 0.0001,
|
| 1550 |
+
"rewards/accuracies": 1.0,
|
| 1551 |
+
"rewards/chosen": -0.08301335573196411,
|
| 1552 |
+
"rewards/margins": 14.271142959594727,
|
| 1553 |
+
"rewards/rejected": -14.354156494140625,
|
| 1554 |
+
"step": 1020
|
| 1555 |
+
},
|
| 1556 |
+
{
|
| 1557 |
+
"epoch": 4.497816593886463,
|
| 1558 |
+
"grad_norm": 0.0073133867021571005,
|
| 1559 |
+
"learning_rate": 1.5222069808913303e-07,
|
| 1560 |
+
"logits/chosen": -0.6575132012367249,
|
| 1561 |
+
"logits/rejected": -0.3197494447231293,
|
| 1562 |
+
"logps/chosen": -300.7660217285156,
|
| 1563 |
+
"logps/rejected": -1779.563232421875,
|
| 1564 |
+
"loss": 0.0001,
|
| 1565 |
+
"rewards/accuracies": 1.0,
|
| 1566 |
+
"rewards/chosen": -0.06065019965171814,
|
| 1567 |
+
"rewards/margins": 15.048309326171875,
|
| 1568 |
+
"rewards/rejected": -15.10895824432373,
|
| 1569 |
+
"step": 1030
|
| 1570 |
+
},
|
| 1571 |
+
{
|
| 1572 |
+
"epoch": 4.541484716157205,
|
| 1573 |
+
"grad_norm": 0.00034831682457422486,
|
| 1574 |
+
"learning_rate": 1.271155666748311e-07,
|
| 1575 |
+
"logits/chosen": -0.7639699578285217,
|
| 1576 |
+
"logits/rejected": -0.43667006492614746,
|
| 1577 |
+
"logps/chosen": -308.33502197265625,
|
| 1578 |
+
"logps/rejected": -1791.947998046875,
|
| 1579 |
+
"loss": 0.0001,
|
| 1580 |
+
"rewards/accuracies": 1.0,
|
| 1581 |
+
"rewards/chosen": -0.1204371303319931,
|
| 1582 |
+
"rewards/margins": 15.054000854492188,
|
| 1583 |
+
"rewards/rejected": -15.1744384765625,
|
| 1584 |
+
"step": 1040
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 4.585152838427947,
|
| 1588 |
+
"grad_norm": 0.0003771930949580811,
|
| 1589 |
+
"learning_rate": 1.0421777008019663e-07,
|
| 1590 |
+
"logits/chosen": -0.7766512632369995,
|
| 1591 |
+
"logits/rejected": -0.3912786841392517,
|
| 1592 |
+
"logps/chosen": -293.3248596191406,
|
| 1593 |
+
"logps/rejected": -1899.7158203125,
|
| 1594 |
+
"loss": 0.0,
|
| 1595 |
+
"rewards/accuracies": 1.0,
|
| 1596 |
+
"rewards/chosen": -0.07415080815553665,
|
| 1597 |
+
"rewards/margins": 15.951011657714844,
|
| 1598 |
+
"rewards/rejected": -16.02515983581543,
|
| 1599 |
+
"step": 1050
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 4.62882096069869,
|
| 1603 |
+
"grad_norm": 0.0005156543022691973,
|
| 1604 |
+
"learning_rate": 8.354860858560021e-08,
|
| 1605 |
+
"logits/chosen": -0.7269026637077332,
|
| 1606 |
+
"logits/rejected": -0.3382445275783539,
|
| 1607 |
+
"logps/chosen": -295.33038330078125,
|
| 1608 |
+
"logps/rejected": -1880.168212890625,
|
| 1609 |
+
"loss": 0.0001,
|
| 1610 |
+
"rewards/accuracies": 1.0,
|
| 1611 |
+
"rewards/chosen": -0.10000383853912354,
|
| 1612 |
+
"rewards/margins": 15.843725204467773,
|
| 1613 |
+
"rewards/rejected": -15.94372844696045,
|
| 1614 |
+
"step": 1060
|
| 1615 |
+
},
|
| 1616 |
+
{
|
| 1617 |
+
"epoch": 4.672489082969433,
|
| 1618 |
+
"grad_norm": 0.0012206160223788651,
|
| 1619 |
+
"learning_rate": 6.512730932204698e-08,
|
| 1620 |
+
"logits/chosen": -0.653832197189331,
|
| 1621 |
+
"logits/rejected": -0.3418962359428406,
|
| 1622 |
+
"logps/chosen": -320.64703369140625,
|
| 1623 |
+
"logps/rejected": -1710.6214599609375,
|
| 1624 |
+
"loss": 0.0,
|
| 1625 |
+
"rewards/accuracies": 1.0,
|
| 1626 |
+
"rewards/chosen": -0.06000875309109688,
|
| 1627 |
+
"rewards/margins": 14.490182876586914,
|
| 1628 |
+
"rewards/rejected": -14.550191879272461,
|
| 1629 |
+
"step": 1070
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 4.716157205240175,
|
| 1633 |
+
"grad_norm": 0.002712454997690386,
|
| 1634 |
+
"learning_rate": 4.897100838547081e-08,
|
| 1635 |
+
"logits/chosen": -0.6822378039360046,
|
| 1636 |
+
"logits/rejected": -0.3716619312763214,
|
| 1637 |
+
"logps/chosen": -306.8667907714844,
|
| 1638 |
+
"logps/rejected": -1840.0302734375,
|
| 1639 |
+
"loss": 0.0001,
|
| 1640 |
+
"rewards/accuracies": 1.0,
|
| 1641 |
+
"rewards/chosen": -0.12169905006885529,
|
| 1642 |
+
"rewards/margins": 15.402539253234863,
|
| 1643 |
+
"rewards/rejected": -15.524238586425781,
|
| 1644 |
+
"step": 1080
|
| 1645 |
+
},
|
| 1646 |
+
{
|
| 1647 |
+
"epoch": 4.759825327510917,
|
| 1648 |
+
"grad_norm": 0.00027359144331405783,
|
| 1649 |
+
"learning_rate": 3.5094734896174985e-08,
|
| 1650 |
+
"logits/chosen": -0.7298842668533325,
|
| 1651 |
+
"logits/rejected": -0.33024948835372925,
|
| 1652 |
+
"logps/chosen": -271.23095703125,
|
| 1653 |
+
"logps/rejected": -1889.5804443359375,
|
| 1654 |
+
"loss": 0.0001,
|
| 1655 |
+
"rewards/accuracies": 1.0,
|
| 1656 |
+
"rewards/chosen": -0.09704066812992096,
|
| 1657 |
+
"rewards/margins": 15.853654861450195,
|
| 1658 |
+
"rewards/rejected": -15.950695991516113,
|
| 1659 |
+
"step": 1090
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 4.8034934497816595,
|
| 1663 |
+
"grad_norm": 0.0027199074701782094,
|
| 1664 |
+
"learning_rate": 2.351139701825267e-08,
|
| 1665 |
+
"logits/chosen": -0.6037005186080933,
|
| 1666 |
+
"logits/rejected": -0.26832693815231323,
|
| 1667 |
+
"logps/chosen": -272.974853515625,
|
| 1668 |
+
"logps/rejected": -1676.4384765625,
|
| 1669 |
+
"loss": 0.0001,
|
| 1670 |
+
"rewards/accuracies": 1.0,
|
| 1671 |
+
"rewards/chosen": -0.06281115859746933,
|
| 1672 |
+
"rewards/margins": 14.172335624694824,
|
| 1673 |
+
"rewards/rejected": -14.235147476196289,
|
| 1674 |
+
"step": 1100
|
| 1675 |
+
},
|
| 1676 |
+
{
|
| 1677 |
+
"epoch": 4.847161572052402,
|
| 1678 |
+
"grad_norm": 0.0007236973858890073,
|
| 1679 |
+
"learning_rate": 1.4231769951990326e-08,
|
| 1680 |
+
"logits/chosen": -0.6609109044075012,
|
| 1681 |
+
"logits/rejected": -0.33717867732048035,
|
| 1682 |
+
"logps/chosen": -295.76898193359375,
|
| 1683 |
+
"logps/rejected": -1793.028076171875,
|
| 1684 |
+
"loss": 0.0,
|
| 1685 |
+
"rewards/accuracies": 1.0,
|
| 1686 |
+
"rewards/chosen": -0.19633208215236664,
|
| 1687 |
+
"rewards/margins": 15.07104206085205,
|
| 1688 |
+
"rewards/rejected": -15.267372131347656,
|
| 1689 |
+
"step": 1110
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 4.890829694323144,
|
| 1693 |
+
"grad_norm": 0.0002222821524569821,
|
| 1694 |
+
"learning_rate": 7.264485910423447e-09,
|
| 1695 |
+
"logits/chosen": -0.6304144263267517,
|
| 1696 |
+
"logits/rejected": -0.2548811733722687,
|
| 1697 |
+
"logps/chosen": -274.11383056640625,
|
| 1698 |
+
"logps/rejected": -1818.213134765625,
|
| 1699 |
+
"loss": 0.0,
|
| 1700 |
+
"rewards/accuracies": 1.0,
|
| 1701 |
+
"rewards/chosen": -0.09864435344934464,
|
| 1702 |
+
"rewards/margins": 15.282160758972168,
|
| 1703 |
+
"rewards/rejected": -15.380805969238281,
|
| 1704 |
+
"step": 1120
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 4.934497816593886,
|
| 1708 |
+
"grad_norm": 0.0096819850101334,
|
| 1709 |
+
"learning_rate": 2.6160260893692833e-09,
|
| 1710 |
+
"logits/chosen": -0.6611170768737793,
|
| 1711 |
+
"logits/rejected": -0.354715496301651,
|
| 1712 |
+
"logps/chosen": -308.61688232421875,
|
| 1713 |
+
"logps/rejected": -1704.0927734375,
|
| 1714 |
+
"loss": 0.0001,
|
| 1715 |
+
"rewards/accuracies": 1.0,
|
| 1716 |
+
"rewards/chosen": -0.12653221189975739,
|
| 1717 |
+
"rewards/margins": 14.360208511352539,
|
| 1718 |
+
"rewards/rejected": -14.486740112304688,
|
| 1719 |
+
"step": 1130
|
| 1720 |
+
},
|
| 1721 |
+
{
|
| 1722 |
+
"epoch": 4.978165938864628,
|
| 1723 |
+
"grad_norm": 0.0002604769137792249,
|
| 1724 |
+
"learning_rate": 2.9071463840540936e-10,
|
| 1725 |
+
"logits/chosen": -0.7605892419815063,
|
| 1726 |
+
"logits/rejected": -0.5090384483337402,
|
| 1727 |
+
"logps/chosen": -336.0806579589844,
|
| 1728 |
+
"logps/rejected": -1813.3853759765625,
|
| 1729 |
+
"loss": 0.0,
|
| 1730 |
+
"rewards/accuracies": 1.0,
|
| 1731 |
+
"rewards/chosen": -0.10771320015192032,
|
| 1732 |
+
"rewards/margins": 15.362016677856445,
|
| 1733 |
+
"rewards/rejected": -15.469728469848633,
|
| 1734 |
+
"step": 1140
|
| 1735 |
+
},
|
| 1736 |
+
{
|
| 1737 |
+
"epoch": 5.0,
|
| 1738 |
+
"step": 1145,
|
| 1739 |
+
"total_flos": 0.0,
|
| 1740 |
+
"train_loss": 0.04901049308554833,
|
| 1741 |
+
"train_runtime": 10890.5088,
|
| 1742 |
+
"train_samples_per_second": 6.722,
|
| 1743 |
+
"train_steps_per_second": 0.105
|
| 1744 |
+
}
|
| 1745 |
+
],
|
| 1746 |
+
"logging_steps": 10,
|
| 1747 |
+
"max_steps": 1145,
|
| 1748 |
+
"num_input_tokens_seen": 0,
|
| 1749 |
+
"num_train_epochs": 5,
|
| 1750 |
+
"save_steps": 100,
|
| 1751 |
+
"stateful_callbacks": {
|
| 1752 |
+
"TrainerControl": {
|
| 1753 |
+
"args": {
|
| 1754 |
+
"should_epoch_stop": false,
|
| 1755 |
+
"should_evaluate": false,
|
| 1756 |
+
"should_log": false,
|
| 1757 |
+
"should_save": true,
|
| 1758 |
+
"should_training_stop": true
|
| 1759 |
+
},
|
| 1760 |
+
"attributes": {}
|
| 1761 |
+
}
|
| 1762 |
+
},
|
| 1763 |
+
"total_flos": 0.0,
|
| 1764 |
+
"train_batch_size": 2,
|
| 1765 |
+
"trial_name": null,
|
| 1766 |
+
"trial_params": null
|
| 1767 |
+
}
|