Model save
Browse files- README.md +83 -0
- all_results.json +9 -0
- train_results.json +9 -0
- trainer_state.json +1852 -0
README.md
ADDED
|
@@ -0,0 +1,83 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: slm-research-vn/Qwen2-7B-Instruct-SPPO-Function-call-v2.8
|
| 3 |
+
library_name: peft
|
| 4 |
+
tags:
|
| 5 |
+
- trl
|
| 6 |
+
- dpo
|
| 7 |
+
- generated_from_trainer
|
| 8 |
+
model-index:
|
| 9 |
+
- name: Qwen2-7B-Instruct-SPPO-Function-call-v2.12
|
| 10 |
+
results: []
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 14 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 15 |
+
|
| 16 |
+
# Qwen2-7B-Instruct-SPPO-Function-call-v2.12
|
| 17 |
+
|
| 18 |
+
This model is a fine-tuned version of [slm-research-vn/Qwen2-7B-Instruct-SPPO-Function-call-v2.8](https://huggingface.co/slm-research-vn/Qwen2-7B-Instruct-SPPO-Function-call-v2.8) on the None dataset.
|
| 19 |
+
It achieves the following results on the evaluation set:
|
| 20 |
+
- Loss: 0.3328
|
| 21 |
+
- Rewards/chosen: 0.5530
|
| 22 |
+
- Rewards/rejected: -0.6974
|
| 23 |
+
- Rewards/accuracies: 0.9075
|
| 24 |
+
- Rewards/margins: 1.2504
|
| 25 |
+
- Logps/rejected: -278.6715
|
| 26 |
+
- Logps/chosen: -129.0585
|
| 27 |
+
- Logits/rejected: -0.5998
|
| 28 |
+
- Logits/chosen: -0.7754
|
| 29 |
+
|
| 30 |
+
## Model description
|
| 31 |
+
|
| 32 |
+
More information needed
|
| 33 |
+
|
| 34 |
+
## Intended uses & limitations
|
| 35 |
+
|
| 36 |
+
More information needed
|
| 37 |
+
|
| 38 |
+
## Training and evaluation data
|
| 39 |
+
|
| 40 |
+
More information needed
|
| 41 |
+
|
| 42 |
+
## Training procedure
|
| 43 |
+
|
| 44 |
+
### Training hyperparameters
|
| 45 |
+
|
| 46 |
+
The following hyperparameters were used during training:
|
| 47 |
+
- learning_rate: 1e-06
|
| 48 |
+
- train_batch_size: 1
|
| 49 |
+
- eval_batch_size: 1
|
| 50 |
+
- seed: 42
|
| 51 |
+
- distributed_type: multi-GPU
|
| 52 |
+
- num_devices: 8
|
| 53 |
+
- gradient_accumulation_steps: 4
|
| 54 |
+
- total_train_batch_size: 32
|
| 55 |
+
- total_eval_batch_size: 8
|
| 56 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 57 |
+
- lr_scheduler_type: cosine
|
| 58 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 59 |
+
- num_epochs: 1
|
| 60 |
+
|
| 61 |
+
### Training results
|
| 62 |
+
|
| 63 |
+
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|
| 64 |
+
|:-------------:|:------:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
|
| 65 |
+
| 0.6806 | 0.0916 | 100 | 0.6816 | 0.0303 | 0.0099 | 0.6445 | 0.0205 | -264.5260 | -139.5110 | -0.5879 | -0.7638 |
|
| 66 |
+
| 0.5704 | 0.1832 | 200 | 0.5993 | 0.3495 | 0.1473 | 0.8237 | 0.2023 | -261.7780 | -133.1277 | -0.5881 | -0.7638 |
|
| 67 |
+
| 0.5032 | 0.2749 | 300 | 0.5313 | 0.5795 | 0.1792 | 0.8526 | 0.4003 | -261.1383 | -128.5271 | -0.5893 | -0.7651 |
|
| 68 |
+
| 0.4548 | 0.3665 | 400 | 0.4727 | 0.6406 | 0.0523 | 0.8844 | 0.5884 | -263.6780 | -127.3051 | -0.5901 | -0.7660 |
|
| 69 |
+
| 0.3823 | 0.4581 | 500 | 0.4235 | 0.6412 | -0.1314 | 0.8931 | 0.7726 | -267.3507 | -127.2934 | -0.5914 | -0.7672 |
|
| 70 |
+
| 0.3513 | 0.5497 | 600 | 0.3843 | 0.6087 | -0.3415 | 0.9133 | 0.9502 | -271.5532 | -127.9448 | -0.5936 | -0.7693 |
|
| 71 |
+
| 0.3444 | 0.6413 | 700 | 0.3571 | 0.5871 | -0.5028 | 0.9104 | 1.0898 | -274.7784 | -128.3763 | -0.5965 | -0.7721 |
|
| 72 |
+
| 0.3486 | 0.7329 | 800 | 0.3427 | 0.5681 | -0.6155 | 0.9104 | 1.1836 | -277.0341 | -128.7559 | -0.5971 | -0.7725 |
|
| 73 |
+
| 0.3317 | 0.8246 | 900 | 0.3349 | 0.5586 | -0.6739 | 0.9133 | 1.2326 | -278.2013 | -128.9451 | -0.5993 | -0.7748 |
|
| 74 |
+
| 0.3077 | 0.9162 | 1000 | 0.3328 | 0.5530 | -0.6974 | 0.9075 | 1.2504 | -278.6715 | -129.0585 | -0.5998 | -0.7754 |
|
| 75 |
+
|
| 76 |
+
|
| 77 |
+
### Framework versions
|
| 78 |
+
|
| 79 |
+
- PEFT 0.12.0
|
| 80 |
+
- Transformers 4.44.0
|
| 81 |
+
- Pytorch 2.3.1+cu121
|
| 82 |
+
- Datasets 2.20.0
|
| 83 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.9995419147961521,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.4306906612229719,
|
| 5 |
+
"train_runtime": 8689.0202,
|
| 6 |
+
"train_samples": 34924,
|
| 7 |
+
"train_samples_per_second": 4.019,
|
| 8 |
+
"train_steps_per_second": 0.126
|
| 9 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.9995419147961521,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.4306906612229719,
|
| 5 |
+
"train_runtime": 8689.0202,
|
| 6 |
+
"train_samples": 34924,
|
| 7 |
+
"train_samples_per_second": 4.019,
|
| 8 |
+
"train_steps_per_second": 0.126
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1852 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9995419147961521,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 1091,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0009161704076958314,
|
| 13 |
+
"grad_norm": 0.8362945914268494,
|
| 14 |
+
"learning_rate": 9.09090909090909e-09,
|
| 15 |
+
"logits/chosen": -0.33099836111068726,
|
| 16 |
+
"logits/rejected": -0.6684360504150391,
|
| 17 |
+
"logps/chosen": -201.79928588867188,
|
| 18 |
+
"logps/rejected": -183.10525512695312,
|
| 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": 0.009161704076958314,
|
| 28 |
+
"grad_norm": 0.7959509491920471,
|
| 29 |
+
"learning_rate": 9.09090909090909e-08,
|
| 30 |
+
"logits/chosen": -0.8238513469696045,
|
| 31 |
+
"logits/rejected": -0.6410912275314331,
|
| 32 |
+
"logps/chosen": -101.17930603027344,
|
| 33 |
+
"logps/rejected": -201.2146759033203,
|
| 34 |
+
"loss": 0.693,
|
| 35 |
+
"rewards/accuracies": 0.5833333134651184,
|
| 36 |
+
"rewards/chosen": 0.0013685817830264568,
|
| 37 |
+
"rewards/margins": 0.003857892006635666,
|
| 38 |
+
"rewards/rejected": -0.002489310223609209,
|
| 39 |
+
"step": 10
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.01832340815391663,
|
| 43 |
+
"grad_norm": 0.791206419467926,
|
| 44 |
+
"learning_rate": 1.818181818181818e-07,
|
| 45 |
+
"logits/chosen": -0.696135401725769,
|
| 46 |
+
"logits/rejected": -0.6421085000038147,
|
| 47 |
+
"logps/chosen": -103.54312896728516,
|
| 48 |
+
"logps/rejected": -195.77432250976562,
|
| 49 |
+
"loss": 0.6932,
|
| 50 |
+
"rewards/accuracies": 0.5,
|
| 51 |
+
"rewards/chosen": -0.0055698854848742485,
|
| 52 |
+
"rewards/margins": -0.009762462228536606,
|
| 53 |
+
"rewards/rejected": 0.004192576743662357,
|
| 54 |
+
"step": 20
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.027485112230874943,
|
| 58 |
+
"grad_norm": 0.8699631690979004,
|
| 59 |
+
"learning_rate": 2.727272727272727e-07,
|
| 60 |
+
"logits/chosen": -0.7419141530990601,
|
| 61 |
+
"logits/rejected": -0.6500910520553589,
|
| 62 |
+
"logps/chosen": -137.67446899414062,
|
| 63 |
+
"logps/rejected": -262.87994384765625,
|
| 64 |
+
"loss": 0.6912,
|
| 65 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 66 |
+
"rewards/chosen": 0.00041480286745354533,
|
| 67 |
+
"rewards/margins": 0.004314146004617214,
|
| 68 |
+
"rewards/rejected": -0.0038993440102785826,
|
| 69 |
+
"step": 30
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.03664681630783326,
|
| 73 |
+
"grad_norm": 0.8166934847831726,
|
| 74 |
+
"learning_rate": 3.636363636363636e-07,
|
| 75 |
+
"logits/chosen": -0.7547220587730408,
|
| 76 |
+
"logits/rejected": -0.5034279823303223,
|
| 77 |
+
"logps/chosen": -94.3946762084961,
|
| 78 |
+
"logps/rejected": -208.92074584960938,
|
| 79 |
+
"loss": 0.6937,
|
| 80 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 81 |
+
"rewards/chosen": -0.002515423344448209,
|
| 82 |
+
"rewards/margins": -0.0004801487666554749,
|
| 83 |
+
"rewards/rejected": -0.002035274403169751,
|
| 84 |
+
"step": 40
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.04580852038479157,
|
| 88 |
+
"grad_norm": 0.8770909309387207,
|
| 89 |
+
"learning_rate": 4.545454545454545e-07,
|
| 90 |
+
"logits/chosen": -0.7312489151954651,
|
| 91 |
+
"logits/rejected": -0.521498441696167,
|
| 92 |
+
"logps/chosen": -132.73512268066406,
|
| 93 |
+
"logps/rejected": -239.8402862548828,
|
| 94 |
+
"loss": 0.6924,
|
| 95 |
+
"rewards/accuracies": 0.5,
|
| 96 |
+
"rewards/chosen": -7.614419155288488e-05,
|
| 97 |
+
"rewards/margins": 0.0043979184702038765,
|
| 98 |
+
"rewards/rejected": -0.004474063403904438,
|
| 99 |
+
"step": 50
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.054970224461749886,
|
| 103 |
+
"grad_norm": 0.8896933794021606,
|
| 104 |
+
"learning_rate": 5.454545454545454e-07,
|
| 105 |
+
"logits/chosen": -0.7862143516540527,
|
| 106 |
+
"logits/rejected": -0.5695743560791016,
|
| 107 |
+
"logps/chosen": -166.35006713867188,
|
| 108 |
+
"logps/rejected": -254.9455108642578,
|
| 109 |
+
"loss": 0.6945,
|
| 110 |
+
"rewards/accuracies": 0.625,
|
| 111 |
+
"rewards/chosen": 0.003317029681056738,
|
| 112 |
+
"rewards/margins": 0.011281955055892467,
|
| 113 |
+
"rewards/rejected": -0.007964925840497017,
|
| 114 |
+
"step": 60
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.0641319285387082,
|
| 118 |
+
"grad_norm": 1.035460114479065,
|
| 119 |
+
"learning_rate": 6.363636363636363e-07,
|
| 120 |
+
"logits/chosen": -0.8119982481002808,
|
| 121 |
+
"logits/rejected": -0.6650353670120239,
|
| 122 |
+
"logps/chosen": -91.4764404296875,
|
| 123 |
+
"logps/rejected": -195.1119842529297,
|
| 124 |
+
"loss": 0.6915,
|
| 125 |
+
"rewards/accuracies": 0.550000011920929,
|
| 126 |
+
"rewards/chosen": 0.005646915640681982,
|
| 127 |
+
"rewards/margins": 0.004837614018470049,
|
| 128 |
+
"rewards/rejected": 0.000809302378911525,
|
| 129 |
+
"step": 70
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.07329363261566652,
|
| 133 |
+
"grad_norm": 1.1378659009933472,
|
| 134 |
+
"learning_rate": 7.272727272727272e-07,
|
| 135 |
+
"logits/chosen": -0.8396841287612915,
|
| 136 |
+
"logits/rejected": -0.739294707775116,
|
| 137 |
+
"logps/chosen": -128.26870727539062,
|
| 138 |
+
"logps/rejected": -237.46743774414062,
|
| 139 |
+
"loss": 0.688,
|
| 140 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 141 |
+
"rewards/chosen": 0.007331106811761856,
|
| 142 |
+
"rewards/margins": 0.0071782395243644714,
|
| 143 |
+
"rewards/rejected": 0.0001528669090475887,
|
| 144 |
+
"step": 80
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 0.08245533669262482,
|
| 148 |
+
"grad_norm": 0.8795006275177002,
|
| 149 |
+
"learning_rate": 8.181818181818182e-07,
|
| 150 |
+
"logits/chosen": -0.7932317852973938,
|
| 151 |
+
"logits/rejected": -0.634912371635437,
|
| 152 |
+
"logps/chosen": -111.9941177368164,
|
| 153 |
+
"logps/rejected": -242.5846710205078,
|
| 154 |
+
"loss": 0.6868,
|
| 155 |
+
"rewards/accuracies": 0.550000011920929,
|
| 156 |
+
"rewards/chosen": 0.021863078698515892,
|
| 157 |
+
"rewards/margins": 0.009898947551846504,
|
| 158 |
+
"rewards/rejected": 0.011964133009314537,
|
| 159 |
+
"step": 90
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.09161704076958314,
|
| 163 |
+
"grad_norm": 0.9012777209281921,
|
| 164 |
+
"learning_rate": 9.09090909090909e-07,
|
| 165 |
+
"logits/chosen": -0.7805367708206177,
|
| 166 |
+
"logits/rejected": -0.49868839979171753,
|
| 167 |
+
"logps/chosen": -140.55088806152344,
|
| 168 |
+
"logps/rejected": -290.98150634765625,
|
| 169 |
+
"loss": 0.6806,
|
| 170 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 171 |
+
"rewards/chosen": 0.02105652168393135,
|
| 172 |
+
"rewards/margins": 0.013599460944533348,
|
| 173 |
+
"rewards/rejected": 0.00745706120505929,
|
| 174 |
+
"step": 100
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 0.09161704076958314,
|
| 178 |
+
"eval_logits/chosen": -0.7638165950775146,
|
| 179 |
+
"eval_logits/rejected": -0.5878660082817078,
|
| 180 |
+
"eval_logps/chosen": -139.51101684570312,
|
| 181 |
+
"eval_logps/rejected": -264.5260009765625,
|
| 182 |
+
"eval_loss": 0.6816309690475464,
|
| 183 |
+
"eval_rewards/accuracies": 0.6445086598396301,
|
| 184 |
+
"eval_rewards/chosen": 0.030344096943736076,
|
| 185 |
+
"eval_rewards/margins": 0.02048710361123085,
|
| 186 |
+
"eval_rewards/rejected": 0.009856995195150375,
|
| 187 |
+
"eval_runtime": 253.9092,
|
| 188 |
+
"eval_samples_per_second": 10.882,
|
| 189 |
+
"eval_steps_per_second": 1.363,
|
| 190 |
+
"step": 100
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 0.10077874484654145,
|
| 194 |
+
"grad_norm": 1.0326433181762695,
|
| 195 |
+
"learning_rate": 1e-06,
|
| 196 |
+
"logits/chosen": -0.8220171928405762,
|
| 197 |
+
"logits/rejected": -0.720903754234314,
|
| 198 |
+
"logps/chosen": -186.0214385986328,
|
| 199 |
+
"logps/rejected": -243.47702026367188,
|
| 200 |
+
"loss": 0.6773,
|
| 201 |
+
"rewards/accuracies": 0.699999988079071,
|
| 202 |
+
"rewards/chosen": 0.05213131755590439,
|
| 203 |
+
"rewards/margins": 0.03893345221877098,
|
| 204 |
+
"rewards/rejected": 0.013197864405810833,
|
| 205 |
+
"step": 110
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.10994044892349977,
|
| 209 |
+
"grad_norm": 1.0610991716384888,
|
| 210 |
+
"learning_rate": 9.997436315234263e-07,
|
| 211 |
+
"logits/chosen": -0.687503457069397,
|
| 212 |
+
"logits/rejected": -0.6214867830276489,
|
| 213 |
+
"logps/chosen": -119.43013000488281,
|
| 214 |
+
"logps/rejected": -170.53787231445312,
|
| 215 |
+
"loss": 0.6699,
|
| 216 |
+
"rewards/accuracies": 0.824999988079071,
|
| 217 |
+
"rewards/chosen": 0.09322745352983475,
|
| 218 |
+
"rewards/margins": 0.058328378945589066,
|
| 219 |
+
"rewards/rejected": 0.034899067133665085,
|
| 220 |
+
"step": 120
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.11910215300045808,
|
| 224 |
+
"grad_norm": 0.9824616312980652,
|
| 225 |
+
"learning_rate": 9.989747889928883e-07,
|
| 226 |
+
"logits/chosen": -0.7877914905548096,
|
| 227 |
+
"logits/rejected": -0.6358338594436646,
|
| 228 |
+
"logps/chosen": -128.0840606689453,
|
| 229 |
+
"logps/rejected": -254.0060272216797,
|
| 230 |
+
"loss": 0.656,
|
| 231 |
+
"rewards/accuracies": 0.699999988079071,
|
| 232 |
+
"rewards/chosen": 0.102009616792202,
|
| 233 |
+
"rewards/margins": 0.07630841434001923,
|
| 234 |
+
"rewards/rejected": 0.02570119872689247,
|
| 235 |
+
"step": 130
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 0.1282638570774164,
|
| 239 |
+
"grad_norm": 1.0044574737548828,
|
| 240 |
+
"learning_rate": 9.976942608363393e-07,
|
| 241 |
+
"logits/chosen": -0.8087643384933472,
|
| 242 |
+
"logits/rejected": -0.5704258680343628,
|
| 243 |
+
"logps/chosen": -112.33183288574219,
|
| 244 |
+
"logps/rejected": -218.6199951171875,
|
| 245 |
+
"loss": 0.6455,
|
| 246 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 247 |
+
"rewards/chosen": 0.18401500582695007,
|
| 248 |
+
"rewards/margins": 0.12449517101049423,
|
| 249 |
+
"rewards/rejected": 0.05951983854174614,
|
| 250 |
+
"step": 140
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 0.1374255611543747,
|
| 254 |
+
"grad_norm": 1.028503179550171,
|
| 255 |
+
"learning_rate": 9.9590336020199e-07,
|
| 256 |
+
"logits/chosen": -0.7007342576980591,
|
| 257 |
+
"logits/rejected": -0.5621221661567688,
|
| 258 |
+
"logps/chosen": -114.86798095703125,
|
| 259 |
+
"logps/rejected": -248.943603515625,
|
| 260 |
+
"loss": 0.6348,
|
| 261 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 262 |
+
"rewards/chosen": 0.24283933639526367,
|
| 263 |
+
"rewards/margins": 0.14451688528060913,
|
| 264 |
+
"rewards/rejected": 0.09832244366407394,
|
| 265 |
+
"step": 150
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 0.14658726523133303,
|
| 269 |
+
"grad_norm": 0.8501743674278259,
|
| 270 |
+
"learning_rate": 9.936039236117095e-07,
|
| 271 |
+
"logits/chosen": -0.7554874420166016,
|
| 272 |
+
"logits/rejected": -0.7505234479904175,
|
| 273 |
+
"logps/chosen": -147.8970184326172,
|
| 274 |
+
"logps/rejected": -204.843505859375,
|
| 275 |
+
"loss": 0.6207,
|
| 276 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 277 |
+
"rewards/chosen": 0.26698964834213257,
|
| 278 |
+
"rewards/margins": 0.1363282948732376,
|
| 279 |
+
"rewards/rejected": 0.13066136837005615,
|
| 280 |
+
"step": 160
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 0.15574896930829135,
|
| 284 |
+
"grad_norm": 0.8963977098464966,
|
| 285 |
+
"learning_rate": 9.907983090777206e-07,
|
| 286 |
+
"logits/chosen": -0.7809524536132812,
|
| 287 |
+
"logits/rejected": -0.6654596924781799,
|
| 288 |
+
"logps/chosen": -131.55320739746094,
|
| 289 |
+
"logps/rejected": -207.1016082763672,
|
| 290 |
+
"loss": 0.6215,
|
| 291 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 292 |
+
"rewards/chosen": 0.30690616369247437,
|
| 293 |
+
"rewards/margins": 0.15369756519794464,
|
| 294 |
+
"rewards/rejected": 0.15320861339569092,
|
| 295 |
+
"step": 170
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 0.16491067338524965,
|
| 299 |
+
"grad_norm": 0.9032484292984009,
|
| 300 |
+
"learning_rate": 9.874893936845187e-07,
|
| 301 |
+
"logits/chosen": -0.6498181819915771,
|
| 302 |
+
"logits/rejected": -0.5487117767333984,
|
| 303 |
+
"logps/chosen": -152.4180908203125,
|
| 304 |
+
"logps/rejected": -292.68438720703125,
|
| 305 |
+
"loss": 0.6068,
|
| 306 |
+
"rewards/accuracies": 0.824999988079071,
|
| 307 |
+
"rewards/chosen": 0.3112695813179016,
|
| 308 |
+
"rewards/margins": 0.18446235358715057,
|
| 309 |
+
"rewards/rejected": 0.12680718302726746,
|
| 310 |
+
"step": 180
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.17407237746220797,
|
| 314 |
+
"grad_norm": 0.7833600044250488,
|
| 315 |
+
"learning_rate": 9.836805706384983e-07,
|
| 316 |
+
"logits/chosen": -0.8308868408203125,
|
| 317 |
+
"logits/rejected": -0.66489177942276,
|
| 318 |
+
"logps/chosen": -97.64247131347656,
|
| 319 |
+
"logps/rejected": -185.3607635498047,
|
| 320 |
+
"loss": 0.5958,
|
| 321 |
+
"rewards/accuracies": 0.824999988079071,
|
| 322 |
+
"rewards/chosen": 0.47603702545166016,
|
| 323 |
+
"rewards/margins": 0.26404884457588196,
|
| 324 |
+
"rewards/rejected": 0.21198812127113342,
|
| 325 |
+
"step": 190
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.1832340815391663,
|
| 329 |
+
"grad_norm": 0.9265146255493164,
|
| 330 |
+
"learning_rate": 9.793757457883061e-07,
|
| 331 |
+
"logits/chosen": -0.8464080691337585,
|
| 332 |
+
"logits/rejected": -0.6489285826683044,
|
| 333 |
+
"logps/chosen": -81.71176147460938,
|
| 334 |
+
"logps/rejected": -186.20758056640625,
|
| 335 |
+
"loss": 0.5704,
|
| 336 |
+
"rewards/accuracies": 0.875,
|
| 337 |
+
"rewards/chosen": 0.5865594148635864,
|
| 338 |
+
"rewards/margins": 0.3980158567428589,
|
| 339 |
+
"rewards/rejected": 0.18854358792304993,
|
| 340 |
+
"step": 200
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 0.1832340815391663,
|
| 344 |
+
"eval_logits/chosen": -0.7638057470321655,
|
| 345 |
+
"eval_logits/rejected": -0.5881015062332153,
|
| 346 |
+
"eval_logps/chosen": -133.12771606445312,
|
| 347 |
+
"eval_logps/rejected": -261.77801513671875,
|
| 348 |
+
"eval_loss": 0.5992826819419861,
|
| 349 |
+
"eval_rewards/accuracies": 0.823699414730072,
|
| 350 |
+
"eval_rewards/chosen": 0.34951016306877136,
|
| 351 |
+
"eval_rewards/margins": 0.20225529372692108,
|
| 352 |
+
"eval_rewards/rejected": 0.14725488424301147,
|
| 353 |
+
"eval_runtime": 252.1498,
|
| 354 |
+
"eval_samples_per_second": 10.958,
|
| 355 |
+
"eval_steps_per_second": 1.372,
|
| 356 |
+
"step": 200
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.1923957856161246,
|
| 360 |
+
"grad_norm": 0.8853036165237427,
|
| 361 |
+
"learning_rate": 9.745793336194975e-07,
|
| 362 |
+
"logits/chosen": -0.674169659614563,
|
| 363 |
+
"logits/rejected": -0.5945926904678345,
|
| 364 |
+
"logps/chosen": -106.3974380493164,
|
| 365 |
+
"logps/rejected": -215.2452850341797,
|
| 366 |
+
"loss": 0.574,
|
| 367 |
+
"rewards/accuracies": 0.925000011920929,
|
| 368 |
+
"rewards/chosen": 0.6075299978256226,
|
| 369 |
+
"rewards/margins": 0.40666985511779785,
|
| 370 |
+
"rewards/rejected": 0.20086011290550232,
|
| 371 |
+
"step": 210
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 0.2015574896930829,
|
| 375 |
+
"grad_norm": 0.8835374116897583,
|
| 376 |
+
"learning_rate": 9.69296252727595e-07,
|
| 377 |
+
"logits/chosen": -0.6831656098365784,
|
| 378 |
+
"logits/rejected": -0.6844693422317505,
|
| 379 |
+
"logps/chosen": -99.75518035888672,
|
| 380 |
+
"logps/rejected": -196.4925994873047,
|
| 381 |
+
"loss": 0.5604,
|
| 382 |
+
"rewards/accuracies": 0.824999988079071,
|
| 383 |
+
"rewards/chosen": 0.5753337740898132,
|
| 384 |
+
"rewards/margins": 0.31991419196128845,
|
| 385 |
+
"rewards/rejected": 0.25541964173316956,
|
| 386 |
+
"step": 220
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.21071919377004122,
|
| 390 |
+
"grad_norm": 0.8146816492080688,
|
| 391 |
+
"learning_rate": 9.63531920774199e-07,
|
| 392 |
+
"logits/chosen": -0.8509872555732727,
|
| 393 |
+
"logits/rejected": -0.7722374796867371,
|
| 394 |
+
"logps/chosen": -84.59940338134766,
|
| 395 |
+
"logps/rejected": -179.6902618408203,
|
| 396 |
+
"loss": 0.5335,
|
| 397 |
+
"rewards/accuracies": 0.875,
|
| 398 |
+
"rewards/chosen": 0.8087629079818726,
|
| 399 |
+
"rewards/margins": 0.43924540281295776,
|
| 400 |
+
"rewards/rejected": 0.36951744556427,
|
| 401 |
+
"step": 230
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.21988089784699955,
|
| 405 |
+
"grad_norm": 0.7882716059684753,
|
| 406 |
+
"learning_rate": 9.572922489313142e-07,
|
| 407 |
+
"logits/chosen": -0.7912776470184326,
|
| 408 |
+
"logits/rejected": -0.7742729187011719,
|
| 409 |
+
"logps/chosen": -122.55216217041016,
|
| 410 |
+
"logps/rejected": -203.78738403320312,
|
| 411 |
+
"loss": 0.5186,
|
| 412 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 413 |
+
"rewards/chosen": 0.6593917608261108,
|
| 414 |
+
"rewards/margins": 0.20427684485912323,
|
| 415 |
+
"rewards/rejected": 0.4551149010658264,
|
| 416 |
+
"step": 240
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.22904260192395787,
|
| 420 |
+
"grad_norm": 0.778360903263092,
|
| 421 |
+
"learning_rate": 9.505836358195993e-07,
|
| 422 |
+
"logits/chosen": -0.8406602144241333,
|
| 423 |
+
"logits/rejected": -0.6831453442573547,
|
| 424 |
+
"logps/chosen": -95.79948425292969,
|
| 425 |
+
"logps/rejected": -225.69931030273438,
|
| 426 |
+
"loss": 0.5362,
|
| 427 |
+
"rewards/accuracies": 0.800000011920929,
|
| 428 |
+
"rewards/chosen": 0.8291288614273071,
|
| 429 |
+
"rewards/margins": 0.5095328092575073,
|
| 430 |
+
"rewards/rejected": 0.3195960223674774,
|
| 431 |
+
"step": 250
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 0.23820430600091616,
|
| 435 |
+
"grad_norm": 0.8873792290687561,
|
| 436 |
+
"learning_rate": 9.434129609467483e-07,
|
| 437 |
+
"logits/chosen": -0.6178407669067383,
|
| 438 |
+
"logits/rejected": -0.6204096674919128,
|
| 439 |
+
"logps/chosen": -194.87338256835938,
|
| 440 |
+
"logps/rejected": -260.9487609863281,
|
| 441 |
+
"loss": 0.5372,
|
| 442 |
+
"rewards/accuracies": 0.949999988079071,
|
| 443 |
+
"rewards/chosen": 0.5945974588394165,
|
| 444 |
+
"rewards/margins": 0.42120856046676636,
|
| 445 |
+
"rewards/rejected": 0.17338886857032776,
|
| 446 |
+
"step": 260
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.24736601007787448,
|
| 450 |
+
"grad_norm": 0.833759069442749,
|
| 451 |
+
"learning_rate": 9.357875776527333e-07,
|
| 452 |
+
"logits/chosen": -0.7258036732673645,
|
| 453 |
+
"logits/rejected": -0.6290136575698853,
|
| 454 |
+
"logps/chosen": -110.7699966430664,
|
| 455 |
+
"logps/rejected": -214.27734375,
|
| 456 |
+
"loss": 0.5062,
|
| 457 |
+
"rewards/accuracies": 0.925000011920929,
|
| 458 |
+
"rewards/chosen": 0.6229942440986633,
|
| 459 |
+
"rewards/margins": 0.48382797837257385,
|
| 460 |
+
"rewards/rejected": 0.13916628062725067,
|
| 461 |
+
"step": 270
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 0.2565277141548328,
|
| 465 |
+
"grad_norm": 0.7017952799797058,
|
| 466 |
+
"learning_rate": 9.27715305569148e-07,
|
| 467 |
+
"logits/chosen": -0.7022525072097778,
|
| 468 |
+
"logits/rejected": -0.5847650766372681,
|
| 469 |
+
"logps/chosen": -103.59033203125,
|
| 470 |
+
"logps/rejected": -211.4336700439453,
|
| 471 |
+
"loss": 0.5042,
|
| 472 |
+
"rewards/accuracies": 0.925000011920929,
|
| 473 |
+
"rewards/chosen": 0.7977592349052429,
|
| 474 |
+
"rewards/margins": 0.5717019438743591,
|
| 475 |
+
"rewards/rejected": 0.2260572463274002,
|
| 476 |
+
"step": 280
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 0.2656894182317911,
|
| 480 |
+
"grad_norm": 0.8911390900611877,
|
| 481 |
+
"learning_rate": 9.192044226003788e-07,
|
| 482 |
+
"logits/chosen": -0.7471307516098022,
|
| 483 |
+
"logits/rejected": -0.6816040277481079,
|
| 484 |
+
"logps/chosen": -121.2009048461914,
|
| 485 |
+
"logps/rejected": -197.65383911132812,
|
| 486 |
+
"loss": 0.5119,
|
| 487 |
+
"rewards/accuracies": 0.875,
|
| 488 |
+
"rewards/chosen": 0.9325343370437622,
|
| 489 |
+
"rewards/margins": 0.6139361262321472,
|
| 490 |
+
"rewards/rejected": 0.3185981810092926,
|
| 491 |
+
"step": 290
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 0.2748511223087494,
|
| 495 |
+
"grad_norm": 0.7208541631698608,
|
| 496 |
+
"learning_rate": 9.102636564348294e-07,
|
| 497 |
+
"logits/chosen": -0.7008062601089478,
|
| 498 |
+
"logits/rejected": -0.6801473498344421,
|
| 499 |
+
"logps/chosen": -114.77880859375,
|
| 500 |
+
"logps/rejected": -210.2012176513672,
|
| 501 |
+
"loss": 0.5032,
|
| 502 |
+
"rewards/accuracies": 0.800000011920929,
|
| 503 |
+
"rewards/chosen": 0.8345085978507996,
|
| 504 |
+
"rewards/margins": 0.5057986378669739,
|
| 505 |
+
"rewards/rejected": 0.3287099599838257,
|
| 506 |
+
"step": 300
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.2748511223087494,
|
| 510 |
+
"eval_logits/chosen": -0.7651150226593018,
|
| 511 |
+
"eval_logits/rejected": -0.5892514586448669,
|
| 512 |
+
"eval_logps/chosen": -128.52711486816406,
|
| 513 |
+
"eval_logps/rejected": -261.1383056640625,
|
| 514 |
+
"eval_loss": 0.5312727093696594,
|
| 515 |
+
"eval_rewards/accuracies": 0.852601170539856,
|
| 516 |
+
"eval_rewards/chosen": 0.5795396566390991,
|
| 517 |
+
"eval_rewards/margins": 0.40029817819595337,
|
| 518 |
+
"eval_rewards/rejected": 0.17924147844314575,
|
| 519 |
+
"eval_runtime": 252.4578,
|
| 520 |
+
"eval_samples_per_second": 10.944,
|
| 521 |
+
"eval_steps_per_second": 1.371,
|
| 522 |
+
"step": 300
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 0.28401282638570774,
|
| 526 |
+
"grad_norm": 0.8867717385292053,
|
| 527 |
+
"learning_rate": 9.009021755949051e-07,
|
| 528 |
+
"logits/chosen": -0.7086865901947021,
|
| 529 |
+
"logits/rejected": -0.7210379242897034,
|
| 530 |
+
"logps/chosen": -109.81558990478516,
|
| 531 |
+
"logps/rejected": -152.71084594726562,
|
| 532 |
+
"loss": 0.5059,
|
| 533 |
+
"rewards/accuracies": 0.875,
|
| 534 |
+
"rewards/chosen": 0.8528501391410828,
|
| 535 |
+
"rewards/margins": 0.4039060175418854,
|
| 536 |
+
"rewards/rejected": 0.4489441514015198,
|
| 537 |
+
"step": 310
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 0.29317453046266606,
|
| 541 |
+
"grad_norm": 0.9950509071350098,
|
| 542 |
+
"learning_rate": 8.911295800349314e-07,
|
| 543 |
+
"logits/chosen": -0.6988198757171631,
|
| 544 |
+
"logits/rejected": -0.6140611171722412,
|
| 545 |
+
"logps/chosen": -160.05844116210938,
|
| 546 |
+
"logps/rejected": -236.6884307861328,
|
| 547 |
+
"loss": 0.4987,
|
| 548 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 549 |
+
"rewards/chosen": 0.7523847818374634,
|
| 550 |
+
"rewards/margins": 0.4543587267398834,
|
| 551 |
+
"rewards/rejected": 0.2980259656906128,
|
| 552 |
+
"step": 320
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 0.3023362345396244,
|
| 556 |
+
"grad_norm": 0.9498128294944763,
|
| 557 |
+
"learning_rate": 8.809558912966519e-07,
|
| 558 |
+
"logits/chosen": -0.6651901006698608,
|
| 559 |
+
"logits/rejected": -0.6055413484573364,
|
| 560 |
+
"logps/chosen": -84.35199737548828,
|
| 561 |
+
"logps/rejected": -208.5164794921875,
|
| 562 |
+
"loss": 0.4501,
|
| 563 |
+
"rewards/accuracies": 0.925000011920929,
|
| 564 |
+
"rewards/chosen": 0.9873196482658386,
|
| 565 |
+
"rewards/margins": 0.7461606860160828,
|
| 566 |
+
"rewards/rejected": 0.24115900695323944,
|
| 567 |
+
"step": 330
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 0.3114979386165827,
|
| 571 |
+
"grad_norm": 0.7728273272514343,
|
| 572 |
+
"learning_rate": 8.703915422323984e-07,
|
| 573 |
+
"logits/chosen": -0.5462226867675781,
|
| 574 |
+
"logits/rejected": -0.46084174513816833,
|
| 575 |
+
"logps/chosen": -129.86944580078125,
|
| 576 |
+
"logps/rejected": -210.6329345703125,
|
| 577 |
+
"loss": 0.4827,
|
| 578 |
+
"rewards/accuracies": 0.824999988079071,
|
| 579 |
+
"rewards/chosen": 0.8226272463798523,
|
| 580 |
+
"rewards/margins": 0.4101462960243225,
|
| 581 |
+
"rewards/rejected": 0.4124809801578522,
|
| 582 |
+
"step": 340
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"epoch": 0.320659642693541,
|
| 586 |
+
"grad_norm": 0.6749230623245239,
|
| 587 |
+
"learning_rate": 8.594473663064734e-07,
|
| 588 |
+
"logits/chosen": -0.8189064264297485,
|
| 589 |
+
"logits/rejected": -0.6820628046989441,
|
| 590 |
+
"logps/chosen": -83.3464126586914,
|
| 591 |
+
"logps/rejected": -180.37059020996094,
|
| 592 |
+
"loss": 0.4865,
|
| 593 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 594 |
+
"rewards/chosen": 1.0249218940734863,
|
| 595 |
+
"rewards/margins": 0.6868400573730469,
|
| 596 |
+
"rewards/rejected": 0.3380817770957947,
|
| 597 |
+
"step": 350
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.3298213467704993,
|
| 601 |
+
"grad_norm": 0.917298436164856,
|
| 602 |
+
"learning_rate": 8.481345864857146e-07,
|
| 603 |
+
"logits/chosen": -0.7375888824462891,
|
| 604 |
+
"logits/rejected": -0.5009106993675232,
|
| 605 |
+
"logps/chosen": -119.41560363769531,
|
| 606 |
+
"logps/rejected": -241.78866577148438,
|
| 607 |
+
"loss": 0.4839,
|
| 608 |
+
"rewards/accuracies": 0.824999988079071,
|
| 609 |
+
"rewards/chosen": 0.7711573839187622,
|
| 610 |
+
"rewards/margins": 0.5962369441986084,
|
| 611 |
+
"rewards/rejected": 0.1749204695224762,
|
| 612 |
+
"step": 360
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 0.3389830508474576,
|
| 616 |
+
"grad_norm": 0.6750978827476501,
|
| 617 |
+
"learning_rate": 8.36464803730636e-07,
|
| 618 |
+
"logits/chosen": -0.8278859853744507,
|
| 619 |
+
"logits/rejected": -0.761436402797699,
|
| 620 |
+
"logps/chosen": -101.12157440185547,
|
| 621 |
+
"logps/rejected": -170.44569396972656,
|
| 622 |
+
"loss": 0.4589,
|
| 623 |
+
"rewards/accuracies": 0.875,
|
| 624 |
+
"rewards/chosen": 1.0102306604385376,
|
| 625 |
+
"rewards/margins": 0.6717247366905212,
|
| 626 |
+
"rewards/rejected": 0.33850589394569397,
|
| 627 |
+
"step": 370
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 0.34814475492441593,
|
| 631 |
+
"grad_norm": 0.8837040066719055,
|
| 632 |
+
"learning_rate": 8.244499850989451e-07,
|
| 633 |
+
"logits/chosen": -0.8623167276382446,
|
| 634 |
+
"logits/rejected": -0.699422299861908,
|
| 635 |
+
"logps/chosen": -77.54528045654297,
|
| 636 |
+
"logps/rejected": -172.49044799804688,
|
| 637 |
+
"loss": 0.456,
|
| 638 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 639 |
+
"rewards/chosen": 1.0262943506240845,
|
| 640 |
+
"rewards/margins": 0.7692972421646118,
|
| 641 |
+
"rewards/rejected": 0.25699710845947266,
|
| 642 |
+
"step": 380
|
| 643 |
+
},
|
| 644 |
+
{
|
| 645 |
+
"epoch": 0.35730645900137425,
|
| 646 |
+
"grad_norm": 0.6525785326957703,
|
| 647 |
+
"learning_rate": 8.121024514736377e-07,
|
| 648 |
+
"logits/chosen": -0.7333749532699585,
|
| 649 |
+
"logits/rejected": -0.5934922099113464,
|
| 650 |
+
"logps/chosen": -71.39646911621094,
|
| 651 |
+
"logps/rejected": -198.8276824951172,
|
| 652 |
+
"loss": 0.425,
|
| 653 |
+
"rewards/accuracies": 0.925000011920929,
|
| 654 |
+
"rewards/chosen": 0.9156519770622253,
|
| 655 |
+
"rewards/margins": 0.7522441148757935,
|
| 656 |
+
"rewards/rejected": 0.16340789198875427,
|
| 657 |
+
"step": 390
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 0.3664681630783326,
|
| 661 |
+
"grad_norm": 0.795269250869751,
|
| 662 |
+
"learning_rate": 7.994348649282532e-07,
|
| 663 |
+
"logits/chosen": -0.7343798279762268,
|
| 664 |
+
"logits/rejected": -0.5222686529159546,
|
| 665 |
+
"logps/chosen": -116.50679779052734,
|
| 666 |
+
"logps/rejected": -223.85409545898438,
|
| 667 |
+
"loss": 0.4548,
|
| 668 |
+
"rewards/accuracies": 0.875,
|
| 669 |
+
"rewards/chosen": 0.9559024572372437,
|
| 670 |
+
"rewards/margins": 0.7852972745895386,
|
| 671 |
+
"rewards/rejected": 0.17060522735118866,
|
| 672 |
+
"step": 400
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 0.3664681630783326,
|
| 676 |
+
"eval_logits/chosen": -0.7659581303596497,
|
| 677 |
+
"eval_logits/rejected": -0.5900666117668152,
|
| 678 |
+
"eval_logps/chosen": -127.30506896972656,
|
| 679 |
+
"eval_logps/rejected": -263.6780090332031,
|
| 680 |
+
"eval_loss": 0.47267913818359375,
|
| 681 |
+
"eval_rewards/accuracies": 0.884393036365509,
|
| 682 |
+
"eval_rewards/chosen": 0.6406417489051819,
|
| 683 |
+
"eval_rewards/margins": 0.5883842706680298,
|
| 684 |
+
"eval_rewards/rejected": 0.05225752666592598,
|
| 685 |
+
"eval_runtime": 252.1281,
|
| 686 |
+
"eval_samples_per_second": 10.959,
|
| 687 |
+
"eval_steps_per_second": 1.372,
|
| 688 |
+
"step": 400
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.3756298671552909,
|
| 692 |
+
"grad_norm": 0.6601365208625793,
|
| 693 |
+
"learning_rate": 7.8646021574225e-07,
|
| 694 |
+
"logits/chosen": -0.6952235102653503,
|
| 695 |
+
"logits/rejected": -0.46782374382019043,
|
| 696 |
+
"logps/chosen": -104.0670394897461,
|
| 697 |
+
"logps/rejected": -239.7484130859375,
|
| 698 |
+
"loss": 0.4482,
|
| 699 |
+
"rewards/accuracies": 0.949999988079071,
|
| 700 |
+
"rewards/chosen": 0.9586408734321594,
|
| 701 |
+
"rewards/margins": 0.7240998148918152,
|
| 702 |
+
"rewards/rejected": 0.23454102873802185,
|
| 703 |
+
"step": 410
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 0.3847915712322492,
|
| 707 |
+
"grad_norm": 0.7361147999763489,
|
| 708 |
+
"learning_rate": 7.731918090798113e-07,
|
| 709 |
+
"logits/chosen": -0.8017680048942566,
|
| 710 |
+
"logits/rejected": -0.6514943838119507,
|
| 711 |
+
"logps/chosen": -97.84058380126953,
|
| 712 |
+
"logps/rejected": -181.79403686523438,
|
| 713 |
+
"loss": 0.4194,
|
| 714 |
+
"rewards/accuracies": 0.949999988079071,
|
| 715 |
+
"rewards/chosen": 1.1092697381973267,
|
| 716 |
+
"rewards/margins": 0.7227686047554016,
|
| 717 |
+
"rewards/rejected": 0.3865012228488922,
|
| 718 |
+
"step": 420
|
| 719 |
+
},
|
| 720 |
+
{
|
| 721 |
+
"epoch": 0.39395327530920754,
|
| 722 |
+
"grad_norm": 0.8337231278419495,
|
| 723 |
+
"learning_rate": 7.596432513457482e-07,
|
| 724 |
+
"logits/chosen": -0.7028661966323853,
|
| 725 |
+
"logits/rejected": -0.6256899833679199,
|
| 726 |
+
"logps/chosen": -96.13346862792969,
|
| 727 |
+
"logps/rejected": -165.75747680664062,
|
| 728 |
+
"loss": 0.426,
|
| 729 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 730 |
+
"rewards/chosen": 1.0767663717269897,
|
| 731 |
+
"rewards/margins": 0.7708722949028015,
|
| 732 |
+
"rewards/rejected": 0.30589404702186584,
|
| 733 |
+
"step": 430
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"epoch": 0.4031149793861658,
|
| 737 |
+
"grad_norm": 0.826393723487854,
|
| 738 |
+
"learning_rate": 7.458284362324842e-07,
|
| 739 |
+
"logits/chosen": -0.7026184797286987,
|
| 740 |
+
"logits/rejected": -0.5175060033798218,
|
| 741 |
+
"logps/chosen": -73.34812927246094,
|
| 742 |
+
"logps/rejected": -228.7108917236328,
|
| 743 |
+
"loss": 0.4312,
|
| 744 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 745 |
+
"rewards/chosen": 1.0083404779434204,
|
| 746 |
+
"rewards/margins": 0.9987894296646118,
|
| 747 |
+
"rewards/rejected": 0.009551113471388817,
|
| 748 |
+
"step": 440
|
| 749 |
+
},
|
| 750 |
+
{
|
| 751 |
+
"epoch": 0.4122766834631241,
|
| 752 |
+
"grad_norm": 0.7044612765312195,
|
| 753 |
+
"learning_rate": 7.317615304724387e-07,
|
| 754 |
+
"logits/chosen": -0.6982114911079407,
|
| 755 |
+
"logits/rejected": -0.6542345881462097,
|
| 756 |
+
"logps/chosen": -97.9085922241211,
|
| 757 |
+
"logps/rejected": -188.80284118652344,
|
| 758 |
+
"loss": 0.4191,
|
| 759 |
+
"rewards/accuracies": 0.925000011920929,
|
| 760 |
+
"rewards/chosen": 1.0013755559921265,
|
| 761 |
+
"rewards/margins": 0.9130918383598328,
|
| 762 |
+
"rewards/rejected": 0.08828376233577728,
|
| 763 |
+
"step": 450
|
| 764 |
+
},
|
| 765 |
+
{
|
| 766 |
+
"epoch": 0.42143838754008245,
|
| 767 |
+
"grad_norm": 0.8462466597557068,
|
| 768 |
+
"learning_rate": 7.174569593104108e-07,
|
| 769 |
+
"logits/chosen": -0.8011992573738098,
|
| 770 |
+
"logits/rejected": -0.7858909368515015,
|
| 771 |
+
"logps/chosen": -111.00840759277344,
|
| 772 |
+
"logps/rejected": -198.2548370361328,
|
| 773 |
+
"loss": 0.4307,
|
| 774 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 775 |
+
"rewards/chosen": 1.0858404636383057,
|
| 776 |
+
"rewards/margins": 0.7915989756584167,
|
| 777 |
+
"rewards/rejected": 0.29424166679382324,
|
| 778 |
+
"step": 460
|
| 779 |
+
},
|
| 780 |
+
{
|
| 781 |
+
"epoch": 0.43060009161704077,
|
| 782 |
+
"grad_norm": 1.0214639902114868,
|
| 783 |
+
"learning_rate": 7.029293917108677e-07,
|
| 784 |
+
"logits/chosen": -0.6210706233978271,
|
| 785 |
+
"logits/rejected": -0.5300668478012085,
|
| 786 |
+
"logps/chosen": -198.18032836914062,
|
| 787 |
+
"logps/rejected": -255.27877807617188,
|
| 788 |
+
"loss": 0.4096,
|
| 789 |
+
"rewards/accuracies": 0.875,
|
| 790 |
+
"rewards/chosen": 0.7058089971542358,
|
| 791 |
+
"rewards/margins": 0.7782974243164062,
|
| 792 |
+
"rewards/rejected": -0.0724884420633316,
|
| 793 |
+
"step": 470
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.4397617956939991,
|
| 797 |
+
"grad_norm": 0.8063747882843018,
|
| 798 |
+
"learning_rate": 6.881937253153051e-07,
|
| 799 |
+
"logits/chosen": -0.8666943311691284,
|
| 800 |
+
"logits/rejected": -0.8363106846809387,
|
| 801 |
+
"logps/chosen": -119.9945068359375,
|
| 802 |
+
"logps/rejected": -200.77627563476562,
|
| 803 |
+
"loss": 0.4004,
|
| 804 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 805 |
+
"rewards/chosen": 1.2585500478744507,
|
| 806 |
+
"rewards/margins": 1.000775694847107,
|
| 807 |
+
"rewards/rejected": 0.25777435302734375,
|
| 808 |
+
"step": 480
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 0.4489234997709574,
|
| 812 |
+
"grad_norm": 0.7443576455116272,
|
| 813 |
+
"learning_rate": 6.732650711651031e-07,
|
| 814 |
+
"logits/chosen": -0.7361257076263428,
|
| 815 |
+
"logits/rejected": -0.5734608769416809,
|
| 816 |
+
"logps/chosen": -127.10151672363281,
|
| 817 |
+
"logps/rejected": -238.904296875,
|
| 818 |
+
"loss": 0.4098,
|
| 819 |
+
"rewards/accuracies": 0.925000011920929,
|
| 820 |
+
"rewards/chosen": 0.6828312873840332,
|
| 821 |
+
"rewards/margins": 0.7480920553207397,
|
| 822 |
+
"rewards/rejected": -0.06526081264019012,
|
| 823 |
+
"step": 490
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 0.45808520384791573,
|
| 827 |
+
"grad_norm": 0.6595803499221802,
|
| 828 |
+
"learning_rate": 6.581587382055491e-07,
|
| 829 |
+
"logits/chosen": -0.8137266039848328,
|
| 830 |
+
"logits/rejected": -0.7571356296539307,
|
| 831 |
+
"logps/chosen": -104.3070297241211,
|
| 832 |
+
"logps/rejected": -201.35647583007812,
|
| 833 |
+
"loss": 0.3823,
|
| 834 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 835 |
+
"rewards/chosen": 1.131914496421814,
|
| 836 |
+
"rewards/margins": 1.0377442836761475,
|
| 837 |
+
"rewards/rejected": 0.09417015314102173,
|
| 838 |
+
"step": 500
|
| 839 |
+
},
|
| 840 |
+
{
|
| 841 |
+
"epoch": 0.45808520384791573,
|
| 842 |
+
"eval_logits/chosen": -0.7672035098075867,
|
| 843 |
+
"eval_logits/rejected": -0.591365396976471,
|
| 844 |
+
"eval_logps/chosen": -127.29341125488281,
|
| 845 |
+
"eval_logps/rejected": -267.3507080078125,
|
| 846 |
+
"eval_loss": 0.4234526753425598,
|
| 847 |
+
"eval_rewards/accuracies": 0.8930636048316956,
|
| 848 |
+
"eval_rewards/chosen": 0.6412254571914673,
|
| 849 |
+
"eval_rewards/margins": 0.7726045250892639,
|
| 850 |
+
"eval_rewards/rejected": -0.13137909770011902,
|
| 851 |
+
"eval_runtime": 252.2594,
|
| 852 |
+
"eval_samples_per_second": 10.953,
|
| 853 |
+
"eval_steps_per_second": 1.372,
|
| 854 |
+
"step": 500
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 0.467246907924874,
|
| 858 |
+
"grad_norm": 0.7072412371635437,
|
| 859 |
+
"learning_rate": 6.428902175869126e-07,
|
| 860 |
+
"logits/chosen": -0.7340083122253418,
|
| 861 |
+
"logits/rejected": -0.6606656908988953,
|
| 862 |
+
"logps/chosen": -126.58953857421875,
|
| 863 |
+
"logps/rejected": -193.69363403320312,
|
| 864 |
+
"loss": 0.3779,
|
| 865 |
+
"rewards/accuracies": 0.949999988079071,
|
| 866 |
+
"rewards/chosen": 1.1136077642440796,
|
| 867 |
+
"rewards/margins": 1.0119339227676392,
|
| 868 |
+
"rewards/rejected": 0.10167402029037476,
|
| 869 |
+
"step": 510
|
| 870 |
+
},
|
| 871 |
+
{
|
| 872 |
+
"epoch": 0.4764086120018323,
|
| 873 |
+
"grad_norm": 0.7274028658866882,
|
| 874 |
+
"learning_rate": 6.274751667786761e-07,
|
| 875 |
+
"logits/chosen": -0.6646409630775452,
|
| 876 |
+
"logits/rejected": -0.5588616728782654,
|
| 877 |
+
"logps/chosen": -177.66925048828125,
|
| 878 |
+
"logps/rejected": -301.41876220703125,
|
| 879 |
+
"loss": 0.4,
|
| 880 |
+
"rewards/accuracies": 0.949999988079071,
|
| 881 |
+
"rewards/chosen": 0.8406426310539246,
|
| 882 |
+
"rewards/margins": 0.9421254992485046,
|
| 883 |
+
"rewards/rejected": -0.10148291289806366,
|
| 884 |
+
"step": 520
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.48557031607879064,
|
| 888 |
+
"grad_norm": 0.8565487861633301,
|
| 889 |
+
"learning_rate": 6.119293935132075e-07,
|
| 890 |
+
"logits/chosen": -0.7664127349853516,
|
| 891 |
+
"logits/rejected": -0.6106132864952087,
|
| 892 |
+
"logps/chosen": -105.25531005859375,
|
| 893 |
+
"logps/rejected": -184.4231414794922,
|
| 894 |
+
"loss": 0.3778,
|
| 895 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 896 |
+
"rewards/chosen": 0.9561585187911987,
|
| 897 |
+
"rewards/margins": 0.908736526966095,
|
| 898 |
+
"rewards/rejected": 0.04742204025387764,
|
| 899 |
+
"step": 530
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 0.49473202015574896,
|
| 903 |
+
"grad_norm": 0.8583614826202393,
|
| 904 |
+
"learning_rate": 5.962688395753437e-07,
|
| 905 |
+
"logits/chosen": -0.9779294729232788,
|
| 906 |
+
"logits/rejected": -0.892353355884552,
|
| 907 |
+
"logps/chosen": -84.86984252929688,
|
| 908 |
+
"logps/rejected": -183.84295654296875,
|
| 909 |
+
"loss": 0.3823,
|
| 910 |
+
"rewards/accuracies": 0.949999988079071,
|
| 911 |
+
"rewards/chosen": 0.9522911310195923,
|
| 912 |
+
"rewards/margins": 0.9800866842269897,
|
| 913 |
+
"rewards/rejected": -0.027795499190688133,
|
| 914 |
+
"step": 540
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 0.5038937242327073,
|
| 918 |
+
"grad_norm": 0.9459156394004822,
|
| 919 |
+
"learning_rate": 5.80509564454506e-07,
|
| 920 |
+
"logits/chosen": -0.7560229897499084,
|
| 921 |
+
"logits/rejected": -0.6336179971694946,
|
| 922 |
+
"logps/chosen": -60.79301834106445,
|
| 923 |
+
"logps/rejected": -198.7034912109375,
|
| 924 |
+
"loss": 0.3682,
|
| 925 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 926 |
+
"rewards/chosen": 1.1121090650558472,
|
| 927 |
+
"rewards/margins": 0.9977458715438843,
|
| 928 |
+
"rewards/rejected": 0.11436320841312408,
|
| 929 |
+
"step": 550
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 0.5130554283096656,
|
| 933 |
+
"grad_norm": 0.7807260155677795,
|
| 934 |
+
"learning_rate": 5.646677288761132e-07,
|
| 935 |
+
"logits/chosen": -0.7362383604049683,
|
| 936 |
+
"logits/rejected": -0.6392463445663452,
|
| 937 |
+
"logps/chosen": -94.71185302734375,
|
| 938 |
+
"logps/rejected": -229.9622802734375,
|
| 939 |
+
"loss": 0.3765,
|
| 940 |
+
"rewards/accuracies": 0.949999988079071,
|
| 941 |
+
"rewards/chosen": 0.8846451640129089,
|
| 942 |
+
"rewards/margins": 1.0037345886230469,
|
| 943 |
+
"rewards/rejected": -0.11908946186304092,
|
| 944 |
+
"step": 560
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 0.5222171323866239,
|
| 948 |
+
"grad_norm": 0.9470974802970886,
|
| 949 |
+
"learning_rate": 5.487595782291784e-07,
|
| 950 |
+
"logits/chosen": -0.807681679725647,
|
| 951 |
+
"logits/rejected": -0.6749477982521057,
|
| 952 |
+
"logps/chosen": -118.90516662597656,
|
| 953 |
+
"logps/rejected": -185.31967163085938,
|
| 954 |
+
"loss": 0.3591,
|
| 955 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 956 |
+
"rewards/chosen": 1.0300986766815186,
|
| 957 |
+
"rewards/margins": 1.0025510787963867,
|
| 958 |
+
"rewards/rejected": 0.02754763886332512,
|
| 959 |
+
"step": 570
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 0.5313788364635822,
|
| 963 |
+
"grad_norm": 0.776768684387207,
|
| 964 |
+
"learning_rate": 5.328014259070878e-07,
|
| 965 |
+
"logits/chosen": -0.7403959035873413,
|
| 966 |
+
"logits/rejected": -0.5926883816719055,
|
| 967 |
+
"logps/chosen": -135.42613220214844,
|
| 968 |
+
"logps/rejected": -229.9550323486328,
|
| 969 |
+
"loss": 0.3776,
|
| 970 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 971 |
+
"rewards/chosen": 0.7646117806434631,
|
| 972 |
+
"rewards/margins": 0.9938881993293762,
|
| 973 |
+
"rewards/rejected": -0.22927632927894592,
|
| 974 |
+
"step": 580
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 0.5405405405405406,
|
| 978 |
+
"grad_norm": 0.890230119228363,
|
| 979 |
+
"learning_rate": 5.168096365786402e-07,
|
| 980 |
+
"logits/chosen": -0.8721317052841187,
|
| 981 |
+
"logits/rejected": -0.6174881458282471,
|
| 982 |
+
"logps/chosen": -113.02799987792969,
|
| 983 |
+
"logps/rejected": -223.82821655273438,
|
| 984 |
+
"loss": 0.3664,
|
| 985 |
+
"rewards/accuracies": 0.925000011920929,
|
| 986 |
+
"rewards/chosen": 0.9226465225219727,
|
| 987 |
+
"rewards/margins": 1.041733980178833,
|
| 988 |
+
"rewards/rejected": -0.11908741295337677,
|
| 989 |
+
"step": 590
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.5497022446174988,
|
| 993 |
+
"grad_norm": 0.7550140023231506,
|
| 994 |
+
"learning_rate": 5.008006094065069e-07,
|
| 995 |
+
"logits/chosen": -0.6947864294052124,
|
| 996 |
+
"logits/rejected": -0.6369723081588745,
|
| 997 |
+
"logps/chosen": -108.70173645019531,
|
| 998 |
+
"logps/rejected": -202.474365234375,
|
| 999 |
+
"loss": 0.3513,
|
| 1000 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1001 |
+
"rewards/chosen": 0.9735838174819946,
|
| 1002 |
+
"rewards/margins": 0.9605382680892944,
|
| 1003 |
+
"rewards/rejected": 0.013045459985733032,
|
| 1004 |
+
"step": 600
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 0.5497022446174988,
|
| 1008 |
+
"eval_logits/chosen": -0.7693285346031189,
|
| 1009 |
+
"eval_logits/rejected": -0.5936357975006104,
|
| 1010 |
+
"eval_logps/chosen": -127.94479370117188,
|
| 1011 |
+
"eval_logps/rejected": -271.55316162109375,
|
| 1012 |
+
"eval_loss": 0.3842833936214447,
|
| 1013 |
+
"eval_rewards/accuracies": 0.913294792175293,
|
| 1014 |
+
"eval_rewards/chosen": 0.608656644821167,
|
| 1015 |
+
"eval_rewards/margins": 0.9501577615737915,
|
| 1016 |
+
"eval_rewards/rejected": -0.3415011465549469,
|
| 1017 |
+
"eval_runtime": 252.1964,
|
| 1018 |
+
"eval_samples_per_second": 10.956,
|
| 1019 |
+
"eval_steps_per_second": 1.372,
|
| 1020 |
+
"step": 600
|
| 1021 |
+
},
|
| 1022 |
+
{
|
| 1023 |
+
"epoch": 0.5588639486944572,
|
| 1024 |
+
"grad_norm": 0.7903901934623718,
|
| 1025 |
+
"learning_rate": 4.847907612303182e-07,
|
| 1026 |
+
"logits/chosen": -0.7707523107528687,
|
| 1027 |
+
"logits/rejected": -0.6204859614372253,
|
| 1028 |
+
"logps/chosen": -131.9110870361328,
|
| 1029 |
+
"logps/rejected": -249.21353149414062,
|
| 1030 |
+
"loss": 0.3699,
|
| 1031 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 1032 |
+
"rewards/chosen": 0.5509034991264343,
|
| 1033 |
+
"rewards/margins": 0.738714337348938,
|
| 1034 |
+
"rewards/rejected": -0.18781085312366486,
|
| 1035 |
+
"step": 610
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 0.5680256527714155,
|
| 1039 |
+
"grad_norm": 0.7700456380844116,
|
| 1040 |
+
"learning_rate": 4.687965097316223e-07,
|
| 1041 |
+
"logits/chosen": -0.8288325071334839,
|
| 1042 |
+
"logits/rejected": -0.6177199482917786,
|
| 1043 |
+
"logps/chosen": -75.93597412109375,
|
| 1044 |
+
"logps/rejected": -225.97067260742188,
|
| 1045 |
+
"loss": 0.3304,
|
| 1046 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1047 |
+
"rewards/chosen": 1.001242756843567,
|
| 1048 |
+
"rewards/margins": 1.3316209316253662,
|
| 1049 |
+
"rewards/rejected": -0.33037835359573364,
|
| 1050 |
+
"step": 620
|
| 1051 |
+
},
|
| 1052 |
+
{
|
| 1053 |
+
"epoch": 0.5771873568483737,
|
| 1054 |
+
"grad_norm": 0.9662141799926758,
|
| 1055 |
+
"learning_rate": 4.5283425659798175e-07,
|
| 1056 |
+
"logits/chosen": -0.8291882276535034,
|
| 1057 |
+
"logits/rejected": -0.8248087167739868,
|
| 1058 |
+
"logps/chosen": -143.07180786132812,
|
| 1059 |
+
"logps/rejected": -264.81463623046875,
|
| 1060 |
+
"loss": 0.3757,
|
| 1061 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1062 |
+
"rewards/chosen": 0.9661464691162109,
|
| 1063 |
+
"rewards/margins": 1.1662724018096924,
|
| 1064 |
+
"rewards/rejected": -0.20012597739696503,
|
| 1065 |
+
"step": 630
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 0.5863490609253321,
|
| 1069 |
+
"grad_norm": 0.7089081406593323,
|
| 1070 |
+
"learning_rate": 4.3692037070347123e-07,
|
| 1071 |
+
"logits/chosen": -0.8086493611335754,
|
| 1072 |
+
"logits/rejected": -0.6815121173858643,
|
| 1073 |
+
"logps/chosen": -89.39459228515625,
|
| 1074 |
+
"logps/rejected": -196.5904541015625,
|
| 1075 |
+
"loss": 0.3598,
|
| 1076 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1077 |
+
"rewards/chosen": 0.9766321182250977,
|
| 1078 |
+
"rewards/margins": 1.233741044998169,
|
| 1079 |
+
"rewards/rejected": -0.2571088671684265,
|
| 1080 |
+
"step": 640
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 0.5955107650022904,
|
| 1084 |
+
"grad_norm": 0.6838498115539551,
|
| 1085 |
+
"learning_rate": 4.21071171322823e-07,
|
| 1086 |
+
"logits/chosen": -0.6819313764572144,
|
| 1087 |
+
"logits/rejected": -0.649137556552887,
|
| 1088 |
+
"logps/chosen": -202.49642944335938,
|
| 1089 |
+
"logps/rejected": -305.983154296875,
|
| 1090 |
+
"loss": 0.358,
|
| 1091 |
+
"rewards/accuracies": 0.875,
|
| 1092 |
+
"rewards/chosen": 0.5383878946304321,
|
| 1093 |
+
"rewards/margins": 0.8751532435417175,
|
| 1094 |
+
"rewards/rejected": -0.3367653489112854,
|
| 1095 |
+
"step": 650
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 0.6046724690792488,
|
| 1099 |
+
"grad_norm": 0.8088337779045105,
|
| 1100 |
+
"learning_rate": 4.0530291139643755e-07,
|
| 1101 |
+
"logits/chosen": -0.8444887399673462,
|
| 1102 |
+
"logits/rejected": -0.7773497700691223,
|
| 1103 |
+
"logps/chosen": -92.40638732910156,
|
| 1104 |
+
"logps/rejected": -210.07962036132812,
|
| 1105 |
+
"loss": 0.3479,
|
| 1106 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1107 |
+
"rewards/chosen": 1.1181201934814453,
|
| 1108 |
+
"rewards/margins": 1.1484359502792358,
|
| 1109 |
+
"rewards/rejected": -0.030315812677145004,
|
| 1110 |
+
"step": 660
|
| 1111 |
+
},
|
| 1112 |
+
{
|
| 1113 |
+
"epoch": 0.613834173156207,
|
| 1114 |
+
"grad_norm": 0.743868350982666,
|
| 1115 |
+
"learning_rate": 3.8963176086341727e-07,
|
| 1116 |
+
"logits/chosen": -0.7523199319839478,
|
| 1117 |
+
"logits/rejected": -0.545455813407898,
|
| 1118 |
+
"logps/chosen": -94.30660247802734,
|
| 1119 |
+
"logps/rejected": -229.30667114257812,
|
| 1120 |
+
"loss": 0.3394,
|
| 1121 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 1122 |
+
"rewards/chosen": 0.5909605026245117,
|
| 1123 |
+
"rewards/margins": 1.0719151496887207,
|
| 1124 |
+
"rewards/rejected": -0.4809547960758209,
|
| 1125 |
+
"step": 670
|
| 1126 |
+
},
|
| 1127 |
+
{
|
| 1128 |
+
"epoch": 0.6229958772331654,
|
| 1129 |
+
"grad_norm": 0.7818168997764587,
|
| 1130 |
+
"learning_rate": 3.7407379007971506e-07,
|
| 1131 |
+
"logits/chosen": -0.6859625577926636,
|
| 1132 |
+
"logits/rejected": -0.5579516887664795,
|
| 1133 |
+
"logps/chosen": -143.53274536132812,
|
| 1134 |
+
"logps/rejected": -291.33453369140625,
|
| 1135 |
+
"loss": 0.3549,
|
| 1136 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1137 |
+
"rewards/chosen": 0.8891765475273132,
|
| 1138 |
+
"rewards/margins": 1.2517534494400024,
|
| 1139 |
+
"rewards/rejected": -0.36257684230804443,
|
| 1140 |
+
"step": 680
|
| 1141 |
+
},
|
| 1142 |
+
{
|
| 1143 |
+
"epoch": 0.6321575813101237,
|
| 1144 |
+
"grad_norm": 0.8612805604934692,
|
| 1145 |
+
"learning_rate": 3.586449533384048e-07,
|
| 1146 |
+
"logits/chosen": -0.6445218324661255,
|
| 1147 |
+
"logits/rejected": -0.5638427138328552,
|
| 1148 |
+
"logps/chosen": -84.69303894042969,
|
| 1149 |
+
"logps/rejected": -175.0728302001953,
|
| 1150 |
+
"loss": 0.3413,
|
| 1151 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1152 |
+
"rewards/chosen": 0.8365532159805298,
|
| 1153 |
+
"rewards/margins": 1.054081916809082,
|
| 1154 |
+
"rewards/rejected": -0.21752884984016418,
|
| 1155 |
+
"step": 690
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 0.641319285387082,
|
| 1159 |
+
"grad_norm": 0.6171064972877502,
|
| 1160 |
+
"learning_rate": 3.433610725089692e-07,
|
| 1161 |
+
"logits/chosen": -0.7035859823226929,
|
| 1162 |
+
"logits/rejected": -0.5256271362304688,
|
| 1163 |
+
"logps/chosen": -138.380615234375,
|
| 1164 |
+
"logps/rejected": -248.87451171875,
|
| 1165 |
+
"loss": 0.3444,
|
| 1166 |
+
"rewards/accuracies": 0.875,
|
| 1167 |
+
"rewards/chosen": 0.8877826929092407,
|
| 1168 |
+
"rewards/margins": 1.0520169734954834,
|
| 1169 |
+
"rewards/rejected": -0.16423439979553223,
|
| 1170 |
+
"step": 700
|
| 1171 |
+
},
|
| 1172 |
+
{
|
| 1173 |
+
"epoch": 0.641319285387082,
|
| 1174 |
+
"eval_logits/chosen": -0.7720704078674316,
|
| 1175 |
+
"eval_logits/rejected": -0.5964747667312622,
|
| 1176 |
+
"eval_logps/chosen": -128.3763427734375,
|
| 1177 |
+
"eval_logps/rejected": -274.77838134765625,
|
| 1178 |
+
"eval_loss": 0.35713571310043335,
|
| 1179 |
+
"eval_rewards/accuracies": 0.910404622554779,
|
| 1180 |
+
"eval_rewards/chosen": 0.5870785117149353,
|
| 1181 |
+
"eval_rewards/margins": 1.0898393392562866,
|
| 1182 |
+
"eval_rewards/rejected": -0.5027608871459961,
|
| 1183 |
+
"eval_runtime": 252.0905,
|
| 1184 |
+
"eval_samples_per_second": 10.96,
|
| 1185 |
+
"eval_steps_per_second": 1.373,
|
| 1186 |
+
"step": 700
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 0.6504809894640403,
|
| 1190 |
+
"grad_norm": 0.8962298631668091,
|
| 1191 |
+
"learning_rate": 3.2823782081238555e-07,
|
| 1192 |
+
"logits/chosen": -0.7447826266288757,
|
| 1193 |
+
"logits/rejected": -0.7338870167732239,
|
| 1194 |
+
"logps/chosen": -95.43080139160156,
|
| 1195 |
+
"logps/rejected": -192.2973175048828,
|
| 1196 |
+
"loss": 0.3405,
|
| 1197 |
+
"rewards/accuracies": 0.875,
|
| 1198 |
+
"rewards/chosen": 1.0037095546722412,
|
| 1199 |
+
"rewards/margins": 1.1646350622177124,
|
| 1200 |
+
"rewards/rejected": -0.16092538833618164,
|
| 1201 |
+
"step": 710
|
| 1202 |
+
},
|
| 1203 |
+
{
|
| 1204 |
+
"epoch": 0.6596426935409986,
|
| 1205 |
+
"grad_norm": 0.7918722033500671,
|
| 1206 |
+
"learning_rate": 3.132907067486471e-07,
|
| 1207 |
+
"logits/chosen": -0.7636629343032837,
|
| 1208 |
+
"logits/rejected": -0.6649468541145325,
|
| 1209 |
+
"logps/chosen": -114.23177337646484,
|
| 1210 |
+
"logps/rejected": -215.4693145751953,
|
| 1211 |
+
"loss": 0.3439,
|
| 1212 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 1213 |
+
"rewards/chosen": 0.6278459429740906,
|
| 1214 |
+
"rewards/margins": 0.8962922096252441,
|
| 1215 |
+
"rewards/rejected": -0.26844626665115356,
|
| 1216 |
+
"step": 720
|
| 1217 |
+
},
|
| 1218 |
+
{
|
| 1219 |
+
"epoch": 0.668804397617957,
|
| 1220 |
+
"grad_norm": 0.8416666388511658,
|
| 1221 |
+
"learning_rate": 2.985350581932005e-07,
|
| 1222 |
+
"logits/chosen": -0.776689350605011,
|
| 1223 |
+
"logits/rejected": -0.6382275819778442,
|
| 1224 |
+
"logps/chosen": -142.0974884033203,
|
| 1225 |
+
"logps/rejected": -211.6706085205078,
|
| 1226 |
+
"loss": 0.3482,
|
| 1227 |
+
"rewards/accuracies": 0.824999988079071,
|
| 1228 |
+
"rewards/chosen": 1.0184391736984253,
|
| 1229 |
+
"rewards/margins": 1.137699842453003,
|
| 1230 |
+
"rewards/rejected": -0.11926088482141495,
|
| 1231 |
+
"step": 730
|
| 1232 |
+
},
|
| 1233 |
+
{
|
| 1234 |
+
"epoch": 0.6779661016949152,
|
| 1235 |
+
"grad_norm": 0.8943964242935181,
|
| 1236 |
+
"learning_rate": 2.839860066786103e-07,
|
| 1237 |
+
"logits/chosen": -0.7382106781005859,
|
| 1238 |
+
"logits/rejected": -0.6326214075088501,
|
| 1239 |
+
"logps/chosen": -84.48265075683594,
|
| 1240 |
+
"logps/rejected": -175.4539031982422,
|
| 1241 |
+
"loss": 0.3261,
|
| 1242 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1243 |
+
"rewards/chosen": 1.1386158466339111,
|
| 1244 |
+
"rewards/margins": 1.2899872064590454,
|
| 1245 |
+
"rewards/rejected": -0.1513712853193283,
|
| 1246 |
+
"step": 740
|
| 1247 |
+
},
|
| 1248 |
+
{
|
| 1249 |
+
"epoch": 0.6871278057718736,
|
| 1250 |
+
"grad_norm": 0.8377578258514404,
|
| 1251 |
+
"learning_rate": 2.6965847187756553e-07,
|
| 1252 |
+
"logits/chosen": -0.7496690154075623,
|
| 1253 |
+
"logits/rejected": -0.6893197298049927,
|
| 1254 |
+
"logps/chosen": -96.63633728027344,
|
| 1255 |
+
"logps/rejected": -162.9729461669922,
|
| 1256 |
+
"loss": 0.3157,
|
| 1257 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1258 |
+
"rewards/chosen": 1.0216400623321533,
|
| 1259 |
+
"rewards/margins": 1.3754501342773438,
|
| 1260 |
+
"rewards/rejected": -0.3538101315498352,
|
| 1261 |
+
"step": 750
|
| 1262 |
+
},
|
| 1263 |
+
{
|
| 1264 |
+
"epoch": 0.6962895098488319,
|
| 1265 |
+
"grad_norm": 0.8681157827377319,
|
| 1266 |
+
"learning_rate": 2.5556714630314613e-07,
|
| 1267 |
+
"logits/chosen": -0.8700932264328003,
|
| 1268 |
+
"logits/rejected": -0.7129749059677124,
|
| 1269 |
+
"logps/chosen": -68.58832550048828,
|
| 1270 |
+
"logps/rejected": -177.4335479736328,
|
| 1271 |
+
"loss": 0.3173,
|
| 1272 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1273 |
+
"rewards/chosen": 1.1690860986709595,
|
| 1274 |
+
"rewards/margins": 1.593080759048462,
|
| 1275 |
+
"rewards/rejected": -0.42399463057518005,
|
| 1276 |
+
"step": 760
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 0.7054512139257902,
|
| 1280 |
+
"grad_norm": 0.7554333209991455,
|
| 1281 |
+
"learning_rate": 2.417264802420343e-07,
|
| 1282 |
+
"logits/chosen": -0.8102506399154663,
|
| 1283 |
+
"logits/rejected": -0.6963295936584473,
|
| 1284 |
+
"logps/chosen": -94.93528747558594,
|
| 1285 |
+
"logps/rejected": -227.2413787841797,
|
| 1286 |
+
"loss": 0.3225,
|
| 1287 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1288 |
+
"rewards/chosen": 0.9498472213745117,
|
| 1289 |
+
"rewards/margins": 1.3431434631347656,
|
| 1290 |
+
"rewards/rejected": -0.39329633116722107,
|
| 1291 |
+
"step": 770
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 0.7146129180027485,
|
| 1295 |
+
"grad_norm": 0.9492266178131104,
|
| 1296 |
+
"learning_rate": 2.2815066693612117e-07,
|
| 1297 |
+
"logits/chosen": -0.7795412540435791,
|
| 1298 |
+
"logits/rejected": -0.6747149229049683,
|
| 1299 |
+
"logps/chosen": -101.27651977539062,
|
| 1300 |
+
"logps/rejected": -204.02151489257812,
|
| 1301 |
+
"loss": 0.321,
|
| 1302 |
+
"rewards/accuracies": 0.875,
|
| 1303 |
+
"rewards/chosen": 1.01050865650177,
|
| 1304 |
+
"rewards/margins": 1.2811126708984375,
|
| 1305 |
+
"rewards/rejected": -0.2706039547920227,
|
| 1306 |
+
"step": 780
|
| 1307 |
+
},
|
| 1308 |
+
{
|
| 1309 |
+
"epoch": 0.7237746220797068,
|
| 1310 |
+
"grad_norm": 0.8731204867362976,
|
| 1311 |
+
"learning_rate": 2.1485362802770862e-07,
|
| 1312 |
+
"logits/chosen": -0.7779867053031921,
|
| 1313 |
+
"logits/rejected": -0.6321139335632324,
|
| 1314 |
+
"logps/chosen": -139.72247314453125,
|
| 1315 |
+
"logps/rejected": -309.1371154785156,
|
| 1316 |
+
"loss": 0.3098,
|
| 1317 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1318 |
+
"rewards/chosen": 0.7068808078765869,
|
| 1319 |
+
"rewards/margins": 1.2700955867767334,
|
| 1320 |
+
"rewards/rejected": -0.563214898109436,
|
| 1321 |
+
"step": 790
|
| 1322 |
+
},
|
| 1323 |
+
{
|
| 1324 |
+
"epoch": 0.7329363261566652,
|
| 1325 |
+
"grad_norm": 0.9384542107582092,
|
| 1326 |
+
"learning_rate": 2.018489992832283e-07,
|
| 1327 |
+
"logits/chosen": -0.7126377820968628,
|
| 1328 |
+
"logits/rejected": -0.5895640850067139,
|
| 1329 |
+
"logps/chosen": -134.41358947753906,
|
| 1330 |
+
"logps/rejected": -231.8523712158203,
|
| 1331 |
+
"loss": 0.3486,
|
| 1332 |
+
"rewards/accuracies": 0.824999988079071,
|
| 1333 |
+
"rewards/chosen": 0.8817362785339355,
|
| 1334 |
+
"rewards/margins": 1.268080234527588,
|
| 1335 |
+
"rewards/rejected": -0.3863441050052643,
|
| 1336 |
+
"step": 800
|
| 1337 |
+
},
|
| 1338 |
+
{
|
| 1339 |
+
"epoch": 0.7329363261566652,
|
| 1340 |
+
"eval_logits/chosen": -0.7724946141242981,
|
| 1341 |
+
"eval_logits/rejected": -0.5970612168312073,
|
| 1342 |
+
"eval_logps/chosen": -128.75588989257812,
|
| 1343 |
+
"eval_logps/rejected": -277.0340576171875,
|
| 1344 |
+
"eval_loss": 0.3426840603351593,
|
| 1345 |
+
"eval_rewards/accuracies": 0.910404622554779,
|
| 1346 |
+
"eval_rewards/chosen": 0.5681010484695435,
|
| 1347 |
+
"eval_rewards/margins": 1.1836459636688232,
|
| 1348 |
+
"eval_rewards/rejected": -0.6155449151992798,
|
| 1349 |
+
"eval_runtime": 251.9975,
|
| 1350 |
+
"eval_samples_per_second": 10.964,
|
| 1351 |
+
"eval_steps_per_second": 1.373,
|
| 1352 |
+
"step": 800
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 0.7420980302336234,
|
| 1356 |
+
"grad_norm": 0.9169598817825317,
|
| 1357 |
+
"learning_rate": 1.891501166101187e-07,
|
| 1358 |
+
"logits/chosen": -0.7657705545425415,
|
| 1359 |
+
"logits/rejected": -0.6736264228820801,
|
| 1360 |
+
"logps/chosen": -82.07408905029297,
|
| 1361 |
+
"logps/rejected": -185.07350158691406,
|
| 1362 |
+
"loss": 0.3199,
|
| 1363 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1364 |
+
"rewards/chosen": 0.8503938913345337,
|
| 1365 |
+
"rewards/margins": 1.3659098148345947,
|
| 1366 |
+
"rewards/rejected": -0.515515923500061,
|
| 1367 |
+
"step": 810
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.7512597343105818,
|
| 1371 |
+
"grad_norm": 0.8475902080535889,
|
| 1372 |
+
"learning_rate": 1.767700023812e-07,
|
| 1373 |
+
"logits/chosen": -0.8032963871955872,
|
| 1374 |
+
"logits/rejected": -0.5944005846977234,
|
| 1375 |
+
"logps/chosen": -111.13800048828125,
|
| 1376 |
+
"logps/rejected": -270.47247314453125,
|
| 1377 |
+
"loss": 0.3154,
|
| 1378 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1379 |
+
"rewards/chosen": 0.8725293278694153,
|
| 1380 |
+
"rewards/margins": 1.7224490642547607,
|
| 1381 |
+
"rewards/rejected": -0.8499197959899902,
|
| 1382 |
+
"step": 820
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 0.7604214383875401,
|
| 1386 |
+
"grad_norm": 0.9381293058395386,
|
| 1387 |
+
"learning_rate": 1.6472135208057125e-07,
|
| 1388 |
+
"logits/chosen": -0.78323894739151,
|
| 1389 |
+
"logits/rejected": -0.6326896548271179,
|
| 1390 |
+
"logps/chosen": -106.88468170166016,
|
| 1391 |
+
"logps/rejected": -183.2010498046875,
|
| 1392 |
+
"loss": 0.3454,
|
| 1393 |
+
"rewards/accuracies": 0.875,
|
| 1394 |
+
"rewards/chosen": 0.8930851221084595,
|
| 1395 |
+
"rewards/margins": 1.1470720767974854,
|
| 1396 |
+
"rewards/rejected": -0.25398701429367065,
|
| 1397 |
+
"step": 830
|
| 1398 |
+
},
|
| 1399 |
+
{
|
| 1400 |
+
"epoch": 0.7695831424644984,
|
| 1401 |
+
"grad_norm": 0.8112012147903442,
|
| 1402 |
+
"learning_rate": 1.530165212847217e-07,
|
| 1403 |
+
"logits/chosen": -0.7700347900390625,
|
| 1404 |
+
"logits/rejected": -0.7001415491104126,
|
| 1405 |
+
"logps/chosen": -93.79769134521484,
|
| 1406 |
+
"logps/rejected": -191.95217895507812,
|
| 1407 |
+
"loss": 0.3337,
|
| 1408 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1409 |
+
"rewards/chosen": 1.0522786378860474,
|
| 1410 |
+
"rewards/margins": 1.627384901046753,
|
| 1411 |
+
"rewards/rejected": -0.5751065015792847,
|
| 1412 |
+
"step": 840
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 0.7787448465414567,
|
| 1416 |
+
"grad_norm": 0.674938976764679,
|
| 1417 |
+
"learning_rate": 1.4166751299221003e-07,
|
| 1418 |
+
"logits/chosen": -0.70830899477005,
|
| 1419 |
+
"logits/rejected": -0.6749675273895264,
|
| 1420 |
+
"logps/chosen": -106.66004943847656,
|
| 1421 |
+
"logps/rejected": -215.84115600585938,
|
| 1422 |
+
"loss": 0.3163,
|
| 1423 |
+
"rewards/accuracies": 0.875,
|
| 1424 |
+
"rewards/chosen": 1.1951297521591187,
|
| 1425 |
+
"rewards/margins": 1.7206119298934937,
|
| 1426 |
+
"rewards/rejected": -0.5254822969436646,
|
| 1427 |
+
"step": 850
|
| 1428 |
+
},
|
| 1429 |
+
{
|
| 1430 |
+
"epoch": 0.7879065506184151,
|
| 1431 |
+
"grad_norm": 0.7585360407829285,
|
| 1432 |
+
"learning_rate": 1.306859653149025e-07,
|
| 1433 |
+
"logits/chosen": -0.7158334851264954,
|
| 1434 |
+
"logits/rejected": -0.6363730430603027,
|
| 1435 |
+
"logps/chosen": -111.70841979980469,
|
| 1436 |
+
"logps/rejected": -218.56442260742188,
|
| 1437 |
+
"loss": 0.3239,
|
| 1438 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1439 |
+
"rewards/chosen": 0.8049896359443665,
|
| 1440 |
+
"rewards/margins": 1.498777985572815,
|
| 1441 |
+
"rewards/rejected": -0.6937884092330933,
|
| 1442 |
+
"step": 860
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 0.7970682546953733,
|
| 1446 |
+
"grad_norm": 0.8166142106056213,
|
| 1447 |
+
"learning_rate": 1.2008313954339305e-07,
|
| 1448 |
+
"logits/chosen": -0.5638805627822876,
|
| 1449 |
+
"logits/rejected": -0.42692404985427856,
|
| 1450 |
+
"logps/chosen": -148.00503540039062,
|
| 1451 |
+
"logps/rejected": -250.53756713867188,
|
| 1452 |
+
"loss": 0.3133,
|
| 1453 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1454 |
+
"rewards/chosen": 0.6730099320411682,
|
| 1455 |
+
"rewards/margins": 1.2878690958023071,
|
| 1456 |
+
"rewards/rejected": -0.6148591637611389,
|
| 1457 |
+
"step": 870
|
| 1458 |
+
},
|
| 1459 |
+
{
|
| 1460 |
+
"epoch": 0.8062299587723316,
|
| 1461 |
+
"grad_norm": 0.6894632577896118,
|
| 1462 |
+
"learning_rate": 1.098699085988432e-07,
|
| 1463 |
+
"logits/chosen": -0.6931108236312866,
|
| 1464 |
+
"logits/rejected": -0.6952135562896729,
|
| 1465 |
+
"logps/chosen": -102.37767791748047,
|
| 1466 |
+
"logps/rejected": -263.8074035644531,
|
| 1467 |
+
"loss": 0.3437,
|
| 1468 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1469 |
+
"rewards/chosen": 0.7925483584403992,
|
| 1470 |
+
"rewards/margins": 1.331838607788086,
|
| 1471 |
+
"rewards/rejected": -0.5392903089523315,
|
| 1472 |
+
"step": 880
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 0.81539166284929,
|
| 1476 |
+
"grad_norm": 0.8443871736526489,
|
| 1477 |
+
"learning_rate": 1.0005674588308566e-07,
|
| 1478 |
+
"logits/chosen": -0.8288379907608032,
|
| 1479 |
+
"logits/rejected": -0.7619950175285339,
|
| 1480 |
+
"logps/chosen": -87.10929870605469,
|
| 1481 |
+
"logps/rejected": -192.11410522460938,
|
| 1482 |
+
"loss": 0.2805,
|
| 1483 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1484 |
+
"rewards/chosen": 1.0762431621551514,
|
| 1485 |
+
"rewards/margins": 1.3870527744293213,
|
| 1486 |
+
"rewards/rejected": -0.31080955266952515,
|
| 1487 |
+
"step": 890
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 0.8245533669262483,
|
| 1491 |
+
"grad_norm": 1.089480996131897,
|
| 1492 |
+
"learning_rate": 9.065371453842358e-08,
|
| 1493 |
+
"logits/chosen": -0.7295564413070679,
|
| 1494 |
+
"logits/rejected": -0.66633141040802,
|
| 1495 |
+
"logps/chosen": -80.64302062988281,
|
| 1496 |
+
"logps/rejected": -198.11285400390625,
|
| 1497 |
+
"loss": 0.3317,
|
| 1498 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1499 |
+
"rewards/chosen": 0.9881649017333984,
|
| 1500 |
+
"rewards/margins": 1.6359241008758545,
|
| 1501 |
+
"rewards/rejected": -0.6477591395378113,
|
| 1502 |
+
"step": 900
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 0.8245533669262483,
|
| 1506 |
+
"eval_logits/chosen": -0.7748228907585144,
|
| 1507 |
+
"eval_logits/rejected": -0.5993096232414246,
|
| 1508 |
+
"eval_logps/chosen": -128.94505310058594,
|
| 1509 |
+
"eval_logps/rejected": -278.2012939453125,
|
| 1510 |
+
"eval_loss": 0.33487406373023987,
|
| 1511 |
+
"eval_rewards/accuracies": 0.913294792175293,
|
| 1512 |
+
"eval_rewards/chosen": 0.5586429238319397,
|
| 1513 |
+
"eval_rewards/margins": 1.2325509786605835,
|
| 1514 |
+
"eval_rewards/rejected": -0.6739080548286438,
|
| 1515 |
+
"eval_runtime": 251.978,
|
| 1516 |
+
"eval_samples_per_second": 10.965,
|
| 1517 |
+
"eval_steps_per_second": 1.373,
|
| 1518 |
+
"step": 900
|
| 1519 |
+
},
|
| 1520 |
+
{
|
| 1521 |
+
"epoch": 0.8337150710032066,
|
| 1522 |
+
"grad_norm": 0.7584892511367798,
|
| 1523 |
+
"learning_rate": 8.167045712814108e-08,
|
| 1524 |
+
"logits/chosen": -0.6891669034957886,
|
| 1525 |
+
"logits/rejected": -0.509884774684906,
|
| 1526 |
+
"logps/chosen": -120.35499572753906,
|
| 1527 |
+
"logps/rejected": -280.7765197753906,
|
| 1528 |
+
"loss": 0.3101,
|
| 1529 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1530 |
+
"rewards/chosen": 0.7489343881607056,
|
| 1531 |
+
"rewards/margins": 1.3420490026474,
|
| 1532 |
+
"rewards/rejected": -0.5931144952774048,
|
| 1533 |
+
"step": 910
|
| 1534 |
+
},
|
| 1535 |
+
{
|
| 1536 |
+
"epoch": 0.8428767750801649,
|
| 1537 |
+
"grad_norm": 0.6903491020202637,
|
| 1538 |
+
"learning_rate": 7.311618574830569e-08,
|
| 1539 |
+
"logits/chosen": -0.6809369921684265,
|
| 1540 |
+
"logits/rejected": -0.5410372614860535,
|
| 1541 |
+
"logps/chosen": -115.5818862915039,
|
| 1542 |
+
"logps/rejected": -269.8819885253906,
|
| 1543 |
+
"loss": 0.2988,
|
| 1544 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1545 |
+
"rewards/chosen": 1.0188477039337158,
|
| 1546 |
+
"rewards/margins": 1.640873908996582,
|
| 1547 |
+
"rewards/rejected": -0.6220260262489319,
|
| 1548 |
+
"step": 920
|
| 1549 |
+
},
|
| 1550 |
+
{
|
| 1551 |
+
"epoch": 0.8520384791571233,
|
| 1552 |
+
"grad_norm": 0.837735116481781,
|
| 1553 |
+
"learning_rate": 6.499967258100514e-08,
|
| 1554 |
+
"logits/chosen": -0.7608433365821838,
|
| 1555 |
+
"logits/rejected": -0.6938787698745728,
|
| 1556 |
+
"logps/chosen": -108.01826477050781,
|
| 1557 |
+
"logps/rejected": -226.8501434326172,
|
| 1558 |
+
"loss": 0.3083,
|
| 1559 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1560 |
+
"rewards/chosen": 0.8835248947143555,
|
| 1561 |
+
"rewards/margins": 1.5439478158950806,
|
| 1562 |
+
"rewards/rejected": -0.6604230403900146,
|
| 1563 |
+
"step": 930
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 0.8612001832340815,
|
| 1567 |
+
"grad_norm": 0.5987643599510193,
|
| 1568 |
+
"learning_rate": 5.732924089870245e-08,
|
| 1569 |
+
"logits/chosen": -0.6554332971572876,
|
| 1570 |
+
"logits/rejected": -0.46006011962890625,
|
| 1571 |
+
"logps/chosen": -141.659423828125,
|
| 1572 |
+
"logps/rejected": -257.38958740234375,
|
| 1573 |
+
"loss": 0.2997,
|
| 1574 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1575 |
+
"rewards/chosen": 0.7329229712486267,
|
| 1576 |
+
"rewards/margins": 1.5281693935394287,
|
| 1577 |
+
"rewards/rejected": -0.795246422290802,
|
| 1578 |
+
"step": 940
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 0.8703618873110398,
|
| 1582 |
+
"grad_norm": 0.7886483073234558,
|
| 1583 |
+
"learning_rate": 5.011275652893782e-08,
|
| 1584 |
+
"logits/chosen": -0.696278989315033,
|
| 1585 |
+
"logits/rejected": -0.6031322479248047,
|
| 1586 |
+
"logps/chosen": -92.44393157958984,
|
| 1587 |
+
"logps/rejected": -203.4006805419922,
|
| 1588 |
+
"loss": 0.2837,
|
| 1589 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1590 |
+
"rewards/chosen": 0.9351547360420227,
|
| 1591 |
+
"rewards/margins": 1.3968000411987305,
|
| 1592 |
+
"rewards/rejected": -0.46164530515670776,
|
| 1593 |
+
"step": 950
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 0.8795235913879982,
|
| 1597 |
+
"grad_norm": 0.8714445233345032,
|
| 1598 |
+
"learning_rate": 4.3357619788127634e-08,
|
| 1599 |
+
"logits/chosen": -0.7752223014831543,
|
| 1600 |
+
"logits/rejected": -0.7101688981056213,
|
| 1601 |
+
"logps/chosen": -135.9874725341797,
|
| 1602 |
+
"logps/rejected": -251.438232421875,
|
| 1603 |
+
"loss": 0.3198,
|
| 1604 |
+
"rewards/accuracies": 0.875,
|
| 1605 |
+
"rewards/chosen": 0.7239378094673157,
|
| 1606 |
+
"rewards/margins": 1.1686517000198364,
|
| 1607 |
+
"rewards/rejected": -0.444713830947876,
|
| 1608 |
+
"step": 960
|
| 1609 |
+
},
|
| 1610 |
+
{
|
| 1611 |
+
"epoch": 0.8886852954649564,
|
| 1612 |
+
"grad_norm": 0.7179840207099915,
|
| 1613 |
+
"learning_rate": 3.707075789273306e-08,
|
| 1614 |
+
"logits/chosen": -0.7246268391609192,
|
| 1615 |
+
"logits/rejected": -0.726759135723114,
|
| 1616 |
+
"logps/chosen": -98.63682556152344,
|
| 1617 |
+
"logps/rejected": -221.41909790039062,
|
| 1618 |
+
"loss": 0.3284,
|
| 1619 |
+
"rewards/accuracies": 0.875,
|
| 1620 |
+
"rewards/chosen": 1.0240366458892822,
|
| 1621 |
+
"rewards/margins": 1.735687255859375,
|
| 1622 |
+
"rewards/rejected": -0.711650550365448,
|
| 1623 |
+
"step": 970
|
| 1624 |
+
},
|
| 1625 |
+
{
|
| 1626 |
+
"epoch": 0.8978469995419148,
|
| 1627 |
+
"grad_norm": 0.8668420314788818,
|
| 1628 |
+
"learning_rate": 3.125861785558015e-08,
|
| 1629 |
+
"logits/chosen": -0.8971655964851379,
|
| 1630 |
+
"logits/rejected": -0.7343107461929321,
|
| 1631 |
+
"logps/chosen": -121.60533142089844,
|
| 1632 |
+
"logps/rejected": -274.42510986328125,
|
| 1633 |
+
"loss": 0.3273,
|
| 1634 |
+
"rewards/accuracies": 0.875,
|
| 1635 |
+
"rewards/chosen": 0.9639977216720581,
|
| 1636 |
+
"rewards/margins": 1.4876940250396729,
|
| 1637 |
+
"rewards/rejected": -0.52369624376297,
|
| 1638 |
+
"step": 980
|
| 1639 |
+
},
|
| 1640 |
+
{
|
| 1641 |
+
"epoch": 0.9070087036188731,
|
| 1642 |
+
"grad_norm": 0.8045462965965271,
|
| 1643 |
+
"learning_rate": 2.592715987461702e-08,
|
| 1644 |
+
"logits/chosen": -0.707330584526062,
|
| 1645 |
+
"logits/rejected": -0.6783408522605896,
|
| 1646 |
+
"logps/chosen": -161.0232391357422,
|
| 1647 |
+
"logps/rejected": -251.8162384033203,
|
| 1648 |
+
"loss": 0.3285,
|
| 1649 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1650 |
+
"rewards/chosen": 0.64153653383255,
|
| 1651 |
+
"rewards/margins": 1.0937219858169556,
|
| 1652 |
+
"rewards/rejected": -0.45218539237976074,
|
| 1653 |
+
"step": 990
|
| 1654 |
+
},
|
| 1655 |
+
{
|
| 1656 |
+
"epoch": 0.9161704076958315,
|
| 1657 |
+
"grad_norm": 0.7876152992248535,
|
| 1658 |
+
"learning_rate": 2.108185122088546e-08,
|
| 1659 |
+
"logits/chosen": -0.8192548751831055,
|
| 1660 |
+
"logits/rejected": -0.7370203733444214,
|
| 1661 |
+
"logps/chosen": -109.0053939819336,
|
| 1662 |
+
"logps/rejected": -241.18704223632812,
|
| 1663 |
+
"loss": 0.3077,
|
| 1664 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1665 |
+
"rewards/chosen": 0.7523542642593384,
|
| 1666 |
+
"rewards/margins": 1.3361942768096924,
|
| 1667 |
+
"rewards/rejected": -0.5838399529457092,
|
| 1668 |
+
"step": 1000
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 0.9161704076958315,
|
| 1672 |
+
"eval_logits/chosen": -0.7753556370735168,
|
| 1673 |
+
"eval_logits/rejected": -0.5998021960258484,
|
| 1674 |
+
"eval_logps/chosen": -129.05845642089844,
|
| 1675 |
+
"eval_logps/rejected": -278.6715087890625,
|
| 1676 |
+
"eval_loss": 0.33283570408821106,
|
| 1677 |
+
"eval_rewards/accuracies": 0.9075144529342651,
|
| 1678 |
+
"eval_rewards/chosen": 0.5529729723930359,
|
| 1679 |
+
"eval_rewards/margins": 1.2503920793533325,
|
| 1680 |
+
"eval_rewards/rejected": -0.697419285774231,
|
| 1681 |
+
"eval_runtime": 411.7768,
|
| 1682 |
+
"eval_samples_per_second": 6.71,
|
| 1683 |
+
"eval_steps_per_second": 0.84,
|
| 1684 |
+
"step": 1000
|
| 1685 |
+
},
|
| 1686 |
+
{
|
| 1687 |
+
"epoch": 0.9253321117727897,
|
| 1688 |
+
"grad_norm": 0.9411417245864868,
|
| 1689 |
+
"learning_rate": 1.672766063197789e-08,
|
| 1690 |
+
"logits/chosen": -0.6959950923919678,
|
| 1691 |
+
"logits/rejected": -0.671855092048645,
|
| 1692 |
+
"logps/chosen": -124.23388671875,
|
| 1693 |
+
"logps/rejected": -235.0277862548828,
|
| 1694 |
+
"loss": 0.3114,
|
| 1695 |
+
"rewards/accuracies": 0.875,
|
| 1696 |
+
"rewards/chosen": 0.9731966257095337,
|
| 1697 |
+
"rewards/margins": 1.4448806047439575,
|
| 1698 |
+
"rewards/rejected": -0.4716837406158447,
|
| 1699 |
+
"step": 1010
|
| 1700 |
+
},
|
| 1701 |
+
{
|
| 1702 |
+
"epoch": 0.934493815849748,
|
| 1703 |
+
"grad_norm": 0.7529547214508057,
|
| 1704 |
+
"learning_rate": 1.286905321672621e-08,
|
| 1705 |
+
"logits/chosen": -0.7256031632423401,
|
| 1706 |
+
"logits/rejected": -0.6639115214347839,
|
| 1707 |
+
"logps/chosen": -74.08079528808594,
|
| 1708 |
+
"logps/rejected": -190.6580047607422,
|
| 1709 |
+
"loss": 0.2967,
|
| 1710 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1711 |
+
"rewards/chosen": 1.259799838066101,
|
| 1712 |
+
"rewards/margins": 1.864816665649414,
|
| 1713 |
+
"rewards/rejected": -0.605016827583313,
|
| 1714 |
+
"step": 1020
|
| 1715 |
+
},
|
| 1716 |
+
{
|
| 1717 |
+
"epoch": 0.9436555199267064,
|
| 1718 |
+
"grad_norm": 0.8753419518470764,
|
| 1719 |
+
"learning_rate": 9.509985876349491e-09,
|
| 1720 |
+
"logits/chosen": -0.6738966703414917,
|
| 1721 |
+
"logits/rejected": -0.5317025184631348,
|
| 1722 |
+
"logps/chosen": -84.36373138427734,
|
| 1723 |
+
"logps/rejected": -239.25881958007812,
|
| 1724 |
+
"loss": 0.2962,
|
| 1725 |
+
"rewards/accuracies": 1.0,
|
| 1726 |
+
"rewards/chosen": 0.9429308772087097,
|
| 1727 |
+
"rewards/margins": 1.350412368774414,
|
| 1728 |
+
"rewards/rejected": -0.40748143196105957,
|
| 1729 |
+
"step": 1030
|
| 1730 |
+
},
|
| 1731 |
+
{
|
| 1732 |
+
"epoch": 0.9528172240036646,
|
| 1733 |
+
"grad_norm": 0.8753317594528198,
|
| 1734 |
+
"learning_rate": 6.6539032467546885e-09,
|
| 1735 |
+
"logits/chosen": -0.7684177756309509,
|
| 1736 |
+
"logits/rejected": -0.7221516370773315,
|
| 1737 |
+
"logps/chosen": -195.75209045410156,
|
| 1738 |
+
"logps/rejected": -291.89117431640625,
|
| 1739 |
+
"loss": 0.2878,
|
| 1740 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1741 |
+
"rewards/chosen": 0.786228597164154,
|
| 1742 |
+
"rewards/margins": 1.300559163093567,
|
| 1743 |
+
"rewards/rejected": -0.5143305063247681,
|
| 1744 |
+
"step": 1040
|
| 1745 |
+
},
|
| 1746 |
+
{
|
| 1747 |
+
"epoch": 0.961978928080623,
|
| 1748 |
+
"grad_norm": 0.7298064827919006,
|
| 1749 |
+
"learning_rate": 4.303734166152706e-09,
|
| 1750 |
+
"logits/chosen": -0.6869091987609863,
|
| 1751 |
+
"logits/rejected": -0.7067757844924927,
|
| 1752 |
+
"logps/chosen": -120.041748046875,
|
| 1753 |
+
"logps/rejected": -194.99925231933594,
|
| 1754 |
+
"loss": 0.3014,
|
| 1755 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1756 |
+
"rewards/chosen": 1.13614022731781,
|
| 1757 |
+
"rewards/margins": 1.6430591344833374,
|
| 1758 |
+
"rewards/rejected": -0.5069188475608826,
|
| 1759 |
+
"step": 1050
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 0.9711406321575813,
|
| 1763 |
+
"grad_norm": 0.6257432103157043,
|
| 1764 |
+
"learning_rate": 2.4618886716110676e-09,
|
| 1765 |
+
"logits/chosen": -0.7349929213523865,
|
| 1766 |
+
"logits/rejected": -0.6031204462051392,
|
| 1767 |
+
"logps/chosen": -102.07683563232422,
|
| 1768 |
+
"logps/rejected": -239.6611785888672,
|
| 1769 |
+
"loss": 0.3347,
|
| 1770 |
+
"rewards/accuracies": 0.875,
|
| 1771 |
+
"rewards/chosen": 0.5327378511428833,
|
| 1772 |
+
"rewards/margins": 1.1345620155334473,
|
| 1773 |
+
"rewards/rejected": -0.6018242835998535,
|
| 1774 |
+
"step": 1060
|
| 1775 |
+
},
|
| 1776 |
+
{
|
| 1777 |
+
"epoch": 0.9803023362345397,
|
| 1778 |
+
"grad_norm": 0.6151149868965149,
|
| 1779 |
+
"learning_rate": 1.1302555276238579e-09,
|
| 1780 |
+
"logits/chosen": -0.8125348091125488,
|
| 1781 |
+
"logits/rejected": -0.6750482320785522,
|
| 1782 |
+
"logps/chosen": -64.71864318847656,
|
| 1783 |
+
"logps/rejected": -191.86572265625,
|
| 1784 |
+
"loss": 0.2674,
|
| 1785 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1786 |
+
"rewards/chosen": 1.005023717880249,
|
| 1787 |
+
"rewards/margins": 1.796847939491272,
|
| 1788 |
+
"rewards/rejected": -0.7918239831924438,
|
| 1789 |
+
"step": 1070
|
| 1790 |
+
},
|
| 1791 |
+
{
|
| 1792 |
+
"epoch": 0.9894640403114979,
|
| 1793 |
+
"grad_norm": 0.7649210691452026,
|
| 1794 |
+
"learning_rate": 3.102002892329536e-10,
|
| 1795 |
+
"logits/chosen": -0.6314386129379272,
|
| 1796 |
+
"logits/rejected": -0.5421333909034729,
|
| 1797 |
+
"logps/chosen": -96.51834869384766,
|
| 1798 |
+
"logps/rejected": -234.3397979736328,
|
| 1799 |
+
"loss": 0.2975,
|
| 1800 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1801 |
+
"rewards/chosen": 1.1401491165161133,
|
| 1802 |
+
"rewards/margins": 1.9231317043304443,
|
| 1803 |
+
"rewards/rejected": -0.7829826474189758,
|
| 1804 |
+
"step": 1080
|
| 1805 |
+
},
|
| 1806 |
+
{
|
| 1807 |
+
"epoch": 0.9986257443884563,
|
| 1808 |
+
"grad_norm": 0.7103332877159119,
|
| 1809 |
+
"learning_rate": 2.5639016871248366e-12,
|
| 1810 |
+
"logits/chosen": -0.7133598923683167,
|
| 1811 |
+
"logits/rejected": -0.5197003483772278,
|
| 1812 |
+
"logps/chosen": -163.91204833984375,
|
| 1813 |
+
"logps/rejected": -249.02682495117188,
|
| 1814 |
+
"loss": 0.3111,
|
| 1815 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 1816 |
+
"rewards/chosen": 0.7740644216537476,
|
| 1817 |
+
"rewards/margins": 1.3091305494308472,
|
| 1818 |
+
"rewards/rejected": -0.5350660681724548,
|
| 1819 |
+
"step": 1090
|
| 1820 |
+
},
|
| 1821 |
+
{
|
| 1822 |
+
"epoch": 0.9995419147961521,
|
| 1823 |
+
"step": 1091,
|
| 1824 |
+
"total_flos": 0.0,
|
| 1825 |
+
"train_loss": 0.4306906612229719,
|
| 1826 |
+
"train_runtime": 8689.0202,
|
| 1827 |
+
"train_samples_per_second": 4.019,
|
| 1828 |
+
"train_steps_per_second": 0.126
|
| 1829 |
+
}
|
| 1830 |
+
],
|
| 1831 |
+
"logging_steps": 10,
|
| 1832 |
+
"max_steps": 1091,
|
| 1833 |
+
"num_input_tokens_seen": 0,
|
| 1834 |
+
"num_train_epochs": 1,
|
| 1835 |
+
"save_steps": 100,
|
| 1836 |
+
"stateful_callbacks": {
|
| 1837 |
+
"TrainerControl": {
|
| 1838 |
+
"args": {
|
| 1839 |
+
"should_epoch_stop": false,
|
| 1840 |
+
"should_evaluate": false,
|
| 1841 |
+
"should_log": false,
|
| 1842 |
+
"should_save": true,
|
| 1843 |
+
"should_training_stop": true
|
| 1844 |
+
},
|
| 1845 |
+
"attributes": {}
|
| 1846 |
+
}
|
| 1847 |
+
},
|
| 1848 |
+
"total_flos": 0.0,
|
| 1849 |
+
"train_batch_size": 1,
|
| 1850 |
+
"trial_name": null,
|
| 1851 |
+
"trial_params": null
|
| 1852 |
+
}
|