Model save
Browse files- README.md +81 -0
- all_results.json +9 -0
- generation_config.json +6 -0
- train_results.json +9 -0
- trainer_state.json +891 -0
README.md
ADDED
|
@@ -0,0 +1,81 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
base_model: alignment-handbook/zephyr-7b-sft-full
|
| 4 |
+
tags:
|
| 5 |
+
- trl
|
| 6 |
+
- dpo
|
| 7 |
+
- generated_from_trainer
|
| 8 |
+
model-index:
|
| 9 |
+
- name: zephyr-7b-dpo-full-ultrabin-reward-scale-05
|
| 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 |
+
# zephyr-7b-dpo-full-ultrabin-reward-scale-05
|
| 17 |
+
|
| 18 |
+
This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on an unknown dataset.
|
| 19 |
+
It achieves the following results on the evaluation set:
|
| 20 |
+
- Loss: 0.5419
|
| 21 |
+
- Rewards/chosen: -2.0657
|
| 22 |
+
- Rewards/rejected: -3.5528
|
| 23 |
+
- Rewards/accuracies: 0.7812
|
| 24 |
+
- Rewards/margins: 1.4871
|
| 25 |
+
- Logps/rejected: -617.9430
|
| 26 |
+
- Logps/chosen: -469.2008
|
| 27 |
+
- Logits/rejected: 3.0322
|
| 28 |
+
- Logits/chosen: 2.1926
|
| 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: 5e-07
|
| 48 |
+
- train_batch_size: 8
|
| 49 |
+
- eval_batch_size: 8
|
| 50 |
+
- seed: 55
|
| 51 |
+
- distributed_type: multi-GPU
|
| 52 |
+
- num_devices: 8
|
| 53 |
+
- gradient_accumulation_steps: 2
|
| 54 |
+
- total_train_batch_size: 128
|
| 55 |
+
- total_eval_batch_size: 64
|
| 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 |
+
| 0.6746 | 0.1046 | 50 | 0.6514 | 0.0215 | -0.0844 | 0.6953 | 0.1059 | -271.1068 | -260.4849 | -2.5751 | -2.6121 |
|
| 66 |
+
| 0.5801 | 0.2092 | 100 | 0.5963 | -1.2413 | -2.0024 | 0.6914 | 0.7611 | -462.9021 | -386.7607 | 0.8478 | 0.5614 |
|
| 67 |
+
| 0.561 | 0.3138 | 150 | 0.5612 | -1.3516 | -2.3053 | 0.7422 | 0.9537 | -493.1910 | -397.7852 | 2.1227 | 1.6750 |
|
| 68 |
+
| 0.552 | 0.4184 | 200 | 0.5634 | -1.7910 | -3.0147 | 0.7539 | 1.2237 | -564.1274 | -441.7259 | 2.6771 | 2.0183 |
|
| 69 |
+
| 0.5367 | 0.5230 | 250 | 0.5404 | -1.6069 | -2.8715 | 0.7656 | 1.2646 | -549.8127 | -423.3247 | 2.8098 | 2.1736 |
|
| 70 |
+
| 0.5231 | 0.6276 | 300 | 0.5511 | -1.8243 | -3.2523 | 0.7656 | 1.4280 | -587.8877 | -445.0558 | 2.9864 | 2.2075 |
|
| 71 |
+
| 0.5092 | 0.7322 | 350 | 0.5402 | -1.9840 | -3.4024 | 0.7734 | 1.4184 | -602.9061 | -461.0307 | 2.8834 | 2.0946 |
|
| 72 |
+
| 0.5231 | 0.8368 | 400 | 0.5417 | -2.0950 | -3.5645 | 0.7812 | 1.4695 | -619.1116 | -472.1271 | 3.0542 | 2.2365 |
|
| 73 |
+
| 0.5232 | 0.9414 | 450 | 0.5419 | -2.0657 | -3.5528 | 0.7812 | 1.4871 | -617.9430 | -469.2008 | 3.0322 | 2.1926 |
|
| 74 |
+
|
| 75 |
+
|
| 76 |
+
### Framework versions
|
| 77 |
+
|
| 78 |
+
- Transformers 4.44.0.dev0
|
| 79 |
+
- Pytorch 2.1.2
|
| 80 |
+
- Datasets 2.20.0
|
| 81 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.5571172007955767,
|
| 5 |
+
"train_runtime": 12724.9946,
|
| 6 |
+
"train_samples": 61134,
|
| 7 |
+
"train_samples_per_second": 4.804,
|
| 8 |
+
"train_steps_per_second": 0.038
|
| 9 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 1,
|
| 4 |
+
"eos_token_id": 2,
|
| 5 |
+
"transformers_version": "4.44.0.dev0"
|
| 6 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.5571172007955767,
|
| 5 |
+
"train_runtime": 12724.9946,
|
| 6 |
+
"train_samples": 61134,
|
| 7 |
+
"train_samples_per_second": 4.804,
|
| 8 |
+
"train_steps_per_second": 0.038
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,891 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"eval_steps": 50,
|
| 6 |
+
"global_step": 478,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.02092050209205021,
|
| 13 |
+
"grad_norm": 5.176846146347138,
|
| 14 |
+
"learning_rate": 1.0416666666666667e-07,
|
| 15 |
+
"logits/chosen": -2.630842924118042,
|
| 16 |
+
"logits/rejected": -2.5769855976104736,
|
| 17 |
+
"logps/chosen": -288.64373779296875,
|
| 18 |
+
"logps/rejected": -275.88287353515625,
|
| 19 |
+
"loss": 0.6932,
|
| 20 |
+
"rewards/accuracies": 0.4937500059604645,
|
| 21 |
+
"rewards/chosen": 0.00047967396676540375,
|
| 22 |
+
"rewards/margins": 0.0007994862971827388,
|
| 23 |
+
"rewards/rejected": -0.0003198123595211655,
|
| 24 |
+
"step": 10
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.04184100418410042,
|
| 28 |
+
"grad_norm": 4.778746903378828,
|
| 29 |
+
"learning_rate": 2.0833333333333333e-07,
|
| 30 |
+
"logits/chosen": -2.6447551250457764,
|
| 31 |
+
"logits/rejected": -2.6132736206054688,
|
| 32 |
+
"logps/chosen": -293.56829833984375,
|
| 33 |
+
"logps/rejected": -259.22283935546875,
|
| 34 |
+
"loss": 0.6927,
|
| 35 |
+
"rewards/accuracies": 0.5687500238418579,
|
| 36 |
+
"rewards/chosen": 0.002531626494601369,
|
| 37 |
+
"rewards/margins": 0.0012841664720326662,
|
| 38 |
+
"rewards/rejected": 0.0012474602553993464,
|
| 39 |
+
"step": 20
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.06276150627615062,
|
| 43 |
+
"grad_norm": 4.484860702954145,
|
| 44 |
+
"learning_rate": 3.1249999999999997e-07,
|
| 45 |
+
"logits/chosen": -2.6660215854644775,
|
| 46 |
+
"logits/rejected": -2.589404821395874,
|
| 47 |
+
"logps/chosen": -294.7344970703125,
|
| 48 |
+
"logps/rejected": -287.273193359375,
|
| 49 |
+
"loss": 0.6908,
|
| 50 |
+
"rewards/accuracies": 0.625,
|
| 51 |
+
"rewards/chosen": 0.014913314953446388,
|
| 52 |
+
"rewards/margins": 0.008078203536570072,
|
| 53 |
+
"rewards/rejected": 0.006835112813860178,
|
| 54 |
+
"step": 30
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.08368200836820083,
|
| 58 |
+
"grad_norm": 4.460846213560835,
|
| 59 |
+
"learning_rate": 4.1666666666666667e-07,
|
| 60 |
+
"logits/chosen": -2.6350624561309814,
|
| 61 |
+
"logits/rejected": -2.5527660846710205,
|
| 62 |
+
"logps/chosen": -270.5862731933594,
|
| 63 |
+
"logps/rejected": -240.20895385742188,
|
| 64 |
+
"loss": 0.685,
|
| 65 |
+
"rewards/accuracies": 0.625,
|
| 66 |
+
"rewards/chosen": 0.033098004758358,
|
| 67 |
+
"rewards/margins": 0.02473345957696438,
|
| 68 |
+
"rewards/rejected": 0.008364550769329071,
|
| 69 |
+
"step": 40
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.10460251046025104,
|
| 73 |
+
"grad_norm": 5.258394523703879,
|
| 74 |
+
"learning_rate": 4.999733114418725e-07,
|
| 75 |
+
"logits/chosen": -2.5787136554718018,
|
| 76 |
+
"logits/rejected": -2.5705668926239014,
|
| 77 |
+
"logps/chosen": -264.2970886230469,
|
| 78 |
+
"logps/rejected": -246.346435546875,
|
| 79 |
+
"loss": 0.6746,
|
| 80 |
+
"rewards/accuracies": 0.6875,
|
| 81 |
+
"rewards/chosen": 0.01130986213684082,
|
| 82 |
+
"rewards/margins": 0.07325177639722824,
|
| 83 |
+
"rewards/rejected": -0.06194191053509712,
|
| 84 |
+
"step": 50
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.10460251046025104,
|
| 88 |
+
"eval_logits/chosen": -2.6120545864105225,
|
| 89 |
+
"eval_logits/rejected": -2.5751192569732666,
|
| 90 |
+
"eval_logps/chosen": -260.48486328125,
|
| 91 |
+
"eval_logps/rejected": -271.1068115234375,
|
| 92 |
+
"eval_loss": 0.6513926982879639,
|
| 93 |
+
"eval_rewards/accuracies": 0.6953125,
|
| 94 |
+
"eval_rewards/chosen": 0.021450327709317207,
|
| 95 |
+
"eval_rewards/margins": 0.10589740425348282,
|
| 96 |
+
"eval_rewards/rejected": -0.08444707095623016,
|
| 97 |
+
"eval_runtime": 104.2349,
|
| 98 |
+
"eval_samples_per_second": 19.187,
|
| 99 |
+
"eval_steps_per_second": 0.307,
|
| 100 |
+
"step": 50
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.12552301255230125,
|
| 104 |
+
"grad_norm": 6.387285250582947,
|
| 105 |
+
"learning_rate": 4.990398100856366e-07,
|
| 106 |
+
"logits/chosen": -2.5387370586395264,
|
| 107 |
+
"logits/rejected": -2.4989449977874756,
|
| 108 |
+
"logps/chosen": -267.24676513671875,
|
| 109 |
+
"logps/rejected": -259.34765625,
|
| 110 |
+
"loss": 0.6603,
|
| 111 |
+
"rewards/accuracies": 0.675000011920929,
|
| 112 |
+
"rewards/chosen": 0.0057476600632071495,
|
| 113 |
+
"rewards/margins": 0.13733819127082825,
|
| 114 |
+
"rewards/rejected": -0.13159053027629852,
|
| 115 |
+
"step": 60
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.14644351464435146,
|
| 119 |
+
"grad_norm": 10.713757122261297,
|
| 120 |
+
"learning_rate": 4.967775735898179e-07,
|
| 121 |
+
"logits/chosen": -2.6051666736602783,
|
| 122 |
+
"logits/rejected": -2.5439701080322266,
|
| 123 |
+
"logps/chosen": -302.8467712402344,
|
| 124 |
+
"logps/rejected": -307.2948303222656,
|
| 125 |
+
"loss": 0.6438,
|
| 126 |
+
"rewards/accuracies": 0.643750011920929,
|
| 127 |
+
"rewards/chosen": -0.17766565084457397,
|
| 128 |
+
"rewards/margins": 0.19495461881160736,
|
| 129 |
+
"rewards/rejected": -0.37262025475502014,
|
| 130 |
+
"step": 70
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 0.16736401673640167,
|
| 134 |
+
"grad_norm": 13.46181824781145,
|
| 135 |
+
"learning_rate": 4.931986719649298e-07,
|
| 136 |
+
"logits/chosen": -1.8741085529327393,
|
| 137 |
+
"logits/rejected": -1.8762668371200562,
|
| 138 |
+
"logps/chosen": -299.6753845214844,
|
| 139 |
+
"logps/rejected": -331.6051330566406,
|
| 140 |
+
"loss": 0.6154,
|
| 141 |
+
"rewards/accuracies": 0.59375,
|
| 142 |
+
"rewards/chosen": -0.40526852011680603,
|
| 143 |
+
"rewards/margins": 0.28667524456977844,
|
| 144 |
+
"rewards/rejected": -0.6919438242912292,
|
| 145 |
+
"step": 80
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 0.18828451882845187,
|
| 149 |
+
"grad_norm": 11.81447104504551,
|
| 150 |
+
"learning_rate": 4.883222001996351e-07,
|
| 151 |
+
"logits/chosen": -1.0113297700881958,
|
| 152 |
+
"logits/rejected": -0.8674372434616089,
|
| 153 |
+
"logps/chosen": -323.6530456542969,
|
| 154 |
+
"logps/rejected": -368.2422790527344,
|
| 155 |
+
"loss": 0.6023,
|
| 156 |
+
"rewards/accuracies": 0.6875,
|
| 157 |
+
"rewards/chosen": -0.5255545377731323,
|
| 158 |
+
"rewards/margins": 0.5487722158432007,
|
| 159 |
+
"rewards/rejected": -1.074326753616333,
|
| 160 |
+
"step": 90
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 0.20920502092050208,
|
| 164 |
+
"grad_norm": 20.253194674698854,
|
| 165 |
+
"learning_rate": 4.821741763807186e-07,
|
| 166 |
+
"logits/chosen": -0.3053513169288635,
|
| 167 |
+
"logits/rejected": 0.3383195400238037,
|
| 168 |
+
"logps/chosen": -372.4469299316406,
|
| 169 |
+
"logps/rejected": -375.91619873046875,
|
| 170 |
+
"loss": 0.5801,
|
| 171 |
+
"rewards/accuracies": 0.737500011920929,
|
| 172 |
+
"rewards/chosen": -0.784656286239624,
|
| 173 |
+
"rewards/margins": 0.7117811441421509,
|
| 174 |
+
"rewards/rejected": -1.4964375495910645,
|
| 175 |
+
"step": 100
|
| 176 |
+
},
|
| 177 |
+
{
|
| 178 |
+
"epoch": 0.20920502092050208,
|
| 179 |
+
"eval_logits/chosen": 0.5614331960678101,
|
| 180 |
+
"eval_logits/rejected": 0.8477897644042969,
|
| 181 |
+
"eval_logps/chosen": -386.7606506347656,
|
| 182 |
+
"eval_logps/rejected": -462.9020690917969,
|
| 183 |
+
"eval_loss": 0.5962740182876587,
|
| 184 |
+
"eval_rewards/accuracies": 0.69140625,
|
| 185 |
+
"eval_rewards/chosen": -1.241307258605957,
|
| 186 |
+
"eval_rewards/margins": 0.7610923647880554,
|
| 187 |
+
"eval_rewards/rejected": -2.0023996829986572,
|
| 188 |
+
"eval_runtime": 102.9081,
|
| 189 |
+
"eval_samples_per_second": 19.435,
|
| 190 |
+
"eval_steps_per_second": 0.311,
|
| 191 |
+
"step": 100
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.2301255230125523,
|
| 195 |
+
"grad_norm": 14.600851352900357,
|
| 196 |
+
"learning_rate": 4.747874028753375e-07,
|
| 197 |
+
"logits/chosen": 0.512942910194397,
|
| 198 |
+
"logits/rejected": 0.9520395398139954,
|
| 199 |
+
"logps/chosen": -399.2005920410156,
|
| 200 |
+
"logps/rejected": -453.3448791503906,
|
| 201 |
+
"loss": 0.5789,
|
| 202 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 203 |
+
"rewards/chosen": -1.0446428060531616,
|
| 204 |
+
"rewards/margins": 0.7325394749641418,
|
| 205 |
+
"rewards/rejected": -1.7771823406219482,
|
| 206 |
+
"step": 110
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.2510460251046025,
|
| 210 |
+
"grad_norm": 18.76879266561917,
|
| 211 |
+
"learning_rate": 4.662012913161997e-07,
|
| 212 |
+
"logits/chosen": 0.7682480812072754,
|
| 213 |
+
"logits/rejected": 1.3250311613082886,
|
| 214 |
+
"logps/chosen": -378.8160095214844,
|
| 215 |
+
"logps/rejected": -432.3802795410156,
|
| 216 |
+
"loss": 0.5736,
|
| 217 |
+
"rewards/accuracies": 0.71875,
|
| 218 |
+
"rewards/chosen": -1.0223931074142456,
|
| 219 |
+
"rewards/margins": 0.9169514775276184,
|
| 220 |
+
"rewards/rejected": -1.9393446445465088,
|
| 221 |
+
"step": 120
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 0.2719665271966527,
|
| 225 |
+
"grad_norm": 15.069786999974797,
|
| 226 |
+
"learning_rate": 4.5646165232345103e-07,
|
| 227 |
+
"logits/chosen": 0.34132882952690125,
|
| 228 |
+
"logits/rejected": 0.8815134167671204,
|
| 229 |
+
"logps/chosen": -395.4140930175781,
|
| 230 |
+
"logps/rejected": -450.15667724609375,
|
| 231 |
+
"loss": 0.5599,
|
| 232 |
+
"rewards/accuracies": 0.706250011920929,
|
| 233 |
+
"rewards/chosen": -0.9794706106185913,
|
| 234 |
+
"rewards/margins": 0.8771657943725586,
|
| 235 |
+
"rewards/rejected": -1.856636643409729,
|
| 236 |
+
"step": 130
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.2928870292887029,
|
| 240 |
+
"grad_norm": 19.253522278201135,
|
| 241 |
+
"learning_rate": 4.456204510851956e-07,
|
| 242 |
+
"logits/chosen": 0.5900996327400208,
|
| 243 |
+
"logits/rejected": 1.196023941040039,
|
| 244 |
+
"logps/chosen": -394.09033203125,
|
| 245 |
+
"logps/rejected": -462.4150390625,
|
| 246 |
+
"loss": 0.5556,
|
| 247 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 248 |
+
"rewards/chosen": -0.9378790855407715,
|
| 249 |
+
"rewards/margins": 0.9935140609741211,
|
| 250 |
+
"rewards/rejected": -1.931393027305603,
|
| 251 |
+
"step": 140
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 0.3138075313807531,
|
| 255 |
+
"grad_norm": 17.705226251621124,
|
| 256 |
+
"learning_rate": 4.337355301007335e-07,
|
| 257 |
+
"logits/chosen": 1.6807399988174438,
|
| 258 |
+
"logits/rejected": 2.5443928241729736,
|
| 259 |
+
"logps/chosen": -434.1851501464844,
|
| 260 |
+
"logps/rejected": -498.25689697265625,
|
| 261 |
+
"loss": 0.561,
|
| 262 |
+
"rewards/accuracies": 0.71875,
|
| 263 |
+
"rewards/chosen": -1.4313938617706299,
|
| 264 |
+
"rewards/margins": 0.8576656579971313,
|
| 265 |
+
"rewards/rejected": -2.289059638977051,
|
| 266 |
+
"step": 150
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.3138075313807531,
|
| 270 |
+
"eval_logits/chosen": 1.6750099658966064,
|
| 271 |
+
"eval_logits/rejected": 2.122723340988159,
|
| 272 |
+
"eval_logps/chosen": -397.78521728515625,
|
| 273 |
+
"eval_logps/rejected": -493.1910095214844,
|
| 274 |
+
"eval_loss": 0.5612272620201111,
|
| 275 |
+
"eval_rewards/accuracies": 0.7421875,
|
| 276 |
+
"eval_rewards/chosen": -1.351552963256836,
|
| 277 |
+
"eval_rewards/margins": 0.9537361860275269,
|
| 278 |
+
"eval_rewards/rejected": -2.3052892684936523,
|
| 279 |
+
"eval_runtime": 102.8554,
|
| 280 |
+
"eval_samples_per_second": 19.445,
|
| 281 |
+
"eval_steps_per_second": 0.311,
|
| 282 |
+
"step": 150
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.33472803347280333,
|
| 286 |
+
"grad_norm": 17.923487490122945,
|
| 287 |
+
"learning_rate": 4.2087030056579986e-07,
|
| 288 |
+
"logits/chosen": 1.7796026468276978,
|
| 289 |
+
"logits/rejected": 2.404552936553955,
|
| 290 |
+
"logps/chosen": -391.76666259765625,
|
| 291 |
+
"logps/rejected": -457.2339782714844,
|
| 292 |
+
"loss": 0.5577,
|
| 293 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 294 |
+
"rewards/chosen": -1.2124366760253906,
|
| 295 |
+
"rewards/margins": 0.963293731212616,
|
| 296 |
+
"rewards/rejected": -2.1757304668426514,
|
| 297 |
+
"step": 160
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.35564853556485354,
|
| 301 |
+
"grad_norm": 22.93702368800413,
|
| 302 |
+
"learning_rate": 4.070934040463998e-07,
|
| 303 |
+
"logits/chosen": 2.2401363849639893,
|
| 304 |
+
"logits/rejected": 3.037454128265381,
|
| 305 |
+
"logps/chosen": -417.4873962402344,
|
| 306 |
+
"logps/rejected": -499.13262939453125,
|
| 307 |
+
"loss": 0.5345,
|
| 308 |
+
"rewards/accuracies": 0.71875,
|
| 309 |
+
"rewards/chosen": -1.4775826930999756,
|
| 310 |
+
"rewards/margins": 0.9278723001480103,
|
| 311 |
+
"rewards/rejected": -2.4054548740386963,
|
| 312 |
+
"step": 170
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 0.37656903765690375,
|
| 316 |
+
"grad_norm": 22.059850949763494,
|
| 317 |
+
"learning_rate": 3.9247834624635404e-07,
|
| 318 |
+
"logits/chosen": 2.308513879776001,
|
| 319 |
+
"logits/rejected": 3.129546642303467,
|
| 320 |
+
"logps/chosen": -479.5143127441406,
|
| 321 |
+
"logps/rejected": -558.447509765625,
|
| 322 |
+
"loss": 0.5486,
|
| 323 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 324 |
+
"rewards/chosen": -1.7520792484283447,
|
| 325 |
+
"rewards/margins": 1.1397490501403809,
|
| 326 |
+
"rewards/rejected": -2.8918280601501465,
|
| 327 |
+
"step": 180
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 0.39748953974895396,
|
| 331 |
+
"grad_norm": 21.54003656099098,
|
| 332 |
+
"learning_rate": 3.7710310482256523e-07,
|
| 333 |
+
"logits/chosen": 1.515995979309082,
|
| 334 |
+
"logits/rejected": 2.3186123371124268,
|
| 335 |
+
"logps/chosen": -405.1570129394531,
|
| 336 |
+
"logps/rejected": -478.882080078125,
|
| 337 |
+
"loss": 0.531,
|
| 338 |
+
"rewards/accuracies": 0.6875,
|
| 339 |
+
"rewards/chosen": -1.5137742757797241,
|
| 340 |
+
"rewards/margins": 0.8394115567207336,
|
| 341 |
+
"rewards/rejected": -2.3531858921051025,
|
| 342 |
+
"step": 190
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"epoch": 0.41841004184100417,
|
| 346 |
+
"grad_norm": 21.26002965858666,
|
| 347 |
+
"learning_rate": 3.610497133404795e-07,
|
| 348 |
+
"logits/chosen": 2.5295510292053223,
|
| 349 |
+
"logits/rejected": 3.3073439598083496,
|
| 350 |
+
"logps/chosen": -462.2782287597656,
|
| 351 |
+
"logps/rejected": -518.1025390625,
|
| 352 |
+
"loss": 0.552,
|
| 353 |
+
"rewards/accuracies": 0.6875,
|
| 354 |
+
"rewards/chosen": -2.019491672515869,
|
| 355 |
+
"rewards/margins": 0.858087420463562,
|
| 356 |
+
"rewards/rejected": -2.8775787353515625,
|
| 357 |
+
"step": 200
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 0.41841004184100417,
|
| 361 |
+
"eval_logits/chosen": 2.018301010131836,
|
| 362 |
+
"eval_logits/rejected": 2.6770527362823486,
|
| 363 |
+
"eval_logps/chosen": -441.72589111328125,
|
| 364 |
+
"eval_logps/rejected": -564.1273803710938,
|
| 365 |
+
"eval_loss": 0.563401997089386,
|
| 366 |
+
"eval_rewards/accuracies": 0.75390625,
|
| 367 |
+
"eval_rewards/chosen": -1.7909597158432007,
|
| 368 |
+
"eval_rewards/margins": 1.2236928939819336,
|
| 369 |
+
"eval_rewards/rejected": -3.0146522521972656,
|
| 370 |
+
"eval_runtime": 104.172,
|
| 371 |
+
"eval_samples_per_second": 19.199,
|
| 372 |
+
"eval_steps_per_second": 0.307,
|
| 373 |
+
"step": 200
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.4393305439330544,
|
| 377 |
+
"grad_norm": 21.451368111326985,
|
| 378 |
+
"learning_rate": 3.4440382358952115e-07,
|
| 379 |
+
"logits/chosen": 1.9598217010498047,
|
| 380 |
+
"logits/rejected": 2.9120290279388428,
|
| 381 |
+
"logps/chosen": -447.69476318359375,
|
| 382 |
+
"logps/rejected": -529.9307861328125,
|
| 383 |
+
"loss": 0.5413,
|
| 384 |
+
"rewards/accuracies": 0.731249988079071,
|
| 385 |
+
"rewards/chosen": -1.7580368518829346,
|
| 386 |
+
"rewards/margins": 0.9606950879096985,
|
| 387 |
+
"rewards/rejected": -2.7187318801879883,
|
| 388 |
+
"step": 210
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 0.4602510460251046,
|
| 392 |
+
"grad_norm": 20.160436671538378,
|
| 393 |
+
"learning_rate": 3.272542485937368e-07,
|
| 394 |
+
"logits/chosen": 1.119225025177002,
|
| 395 |
+
"logits/rejected": 1.9501537084579468,
|
| 396 |
+
"logps/chosen": -417.14422607421875,
|
| 397 |
+
"logps/rejected": -513.2694091796875,
|
| 398 |
+
"loss": 0.555,
|
| 399 |
+
"rewards/accuracies": 0.75,
|
| 400 |
+
"rewards/chosen": -1.3289070129394531,
|
| 401 |
+
"rewards/margins": 1.0918917655944824,
|
| 402 |
+
"rewards/rejected": -2.4207987785339355,
|
| 403 |
+
"step": 220
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"epoch": 0.4811715481171548,
|
| 407 |
+
"grad_norm": 20.190647127947944,
|
| 408 |
+
"learning_rate": 3.096924887558854e-07,
|
| 409 |
+
"logits/chosen": 2.0996615886688232,
|
| 410 |
+
"logits/rejected": 3.000412940979004,
|
| 411 |
+
"logps/chosen": -453.54473876953125,
|
| 412 |
+
"logps/rejected": -568.1104125976562,
|
| 413 |
+
"loss": 0.5188,
|
| 414 |
+
"rewards/accuracies": 0.737500011920929,
|
| 415 |
+
"rewards/chosen": -1.7147760391235352,
|
| 416 |
+
"rewards/margins": 1.3763706684112549,
|
| 417 |
+
"rewards/rejected": -3.09114670753479,
|
| 418 |
+
"step": 230
|
| 419 |
+
},
|
| 420 |
+
{
|
| 421 |
+
"epoch": 0.502092050209205,
|
| 422 |
+
"grad_norm": 20.668039994089078,
|
| 423 |
+
"learning_rate": 2.9181224366319943e-07,
|
| 424 |
+
"logits/chosen": 3.0839335918426514,
|
| 425 |
+
"logits/rejected": 4.029969215393066,
|
| 426 |
+
"logps/chosen": -446.9358825683594,
|
| 427 |
+
"logps/rejected": -518.8782958984375,
|
| 428 |
+
"loss": 0.5411,
|
| 429 |
+
"rewards/accuracies": 0.699999988079071,
|
| 430 |
+
"rewards/chosen": -1.922851800918579,
|
| 431 |
+
"rewards/margins": 1.0858973264694214,
|
| 432 |
+
"rewards/rejected": -3.00874924659729,
|
| 433 |
+
"step": 240
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 0.5230125523012552,
|
| 437 |
+
"grad_norm": 21.91013779471945,
|
| 438 |
+
"learning_rate": 2.7370891215954565e-07,
|
| 439 |
+
"logits/chosen": 2.500542163848877,
|
| 440 |
+
"logits/rejected": 3.4184327125549316,
|
| 441 |
+
"logps/chosen": -480.619873046875,
|
| 442 |
+
"logps/rejected": -543.4830322265625,
|
| 443 |
+
"loss": 0.5367,
|
| 444 |
+
"rewards/accuracies": 0.75,
|
| 445 |
+
"rewards/chosen": -1.8631795644760132,
|
| 446 |
+
"rewards/margins": 1.1071960926055908,
|
| 447 |
+
"rewards/rejected": -2.9703755378723145,
|
| 448 |
+
"step": 250
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 0.5230125523012552,
|
| 452 |
+
"eval_logits/chosen": 2.173550844192505,
|
| 453 |
+
"eval_logits/rejected": 2.8097872734069824,
|
| 454 |
+
"eval_logps/chosen": -423.32470703125,
|
| 455 |
+
"eval_logps/rejected": -549.812744140625,
|
| 456 |
+
"eval_loss": 0.5404338836669922,
|
| 457 |
+
"eval_rewards/accuracies": 0.765625,
|
| 458 |
+
"eval_rewards/chosen": -1.6069477796554565,
|
| 459 |
+
"eval_rewards/margins": 1.2645587921142578,
|
| 460 |
+
"eval_rewards/rejected": -2.871506690979004,
|
| 461 |
+
"eval_runtime": 104.7771,
|
| 462 |
+
"eval_samples_per_second": 19.088,
|
| 463 |
+
"eval_steps_per_second": 0.305,
|
| 464 |
+
"step": 250
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.5439330543933054,
|
| 468 |
+
"grad_norm": 22.08171649508119,
|
| 469 |
+
"learning_rate": 2.55479083351317e-07,
|
| 470 |
+
"logits/chosen": 2.1905927658081055,
|
| 471 |
+
"logits/rejected": 3.057525873184204,
|
| 472 |
+
"logps/chosen": -468.181884765625,
|
| 473 |
+
"logps/rejected": -552.5390014648438,
|
| 474 |
+
"loss": 0.5252,
|
| 475 |
+
"rewards/accuracies": 0.793749988079071,
|
| 476 |
+
"rewards/chosen": -1.5366401672363281,
|
| 477 |
+
"rewards/margins": 1.3390202522277832,
|
| 478 |
+
"rewards/rejected": -2.8756604194641113,
|
| 479 |
+
"step": 260
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 0.5648535564853556,
|
| 483 |
+
"grad_norm": 18.888110325544,
|
| 484 |
+
"learning_rate": 2.3722002126275822e-07,
|
| 485 |
+
"logits/chosen": 2.343186140060425,
|
| 486 |
+
"logits/rejected": 3.283693790435791,
|
| 487 |
+
"logps/chosen": -480.1078186035156,
|
| 488 |
+
"logps/rejected": -560.9808349609375,
|
| 489 |
+
"loss": 0.5243,
|
| 490 |
+
"rewards/accuracies": 0.737500011920929,
|
| 491 |
+
"rewards/chosen": -1.8669660091400146,
|
| 492 |
+
"rewards/margins": 1.2250339984893799,
|
| 493 |
+
"rewards/rejected": -3.0920000076293945,
|
| 494 |
+
"step": 270
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.5857740585774058,
|
| 498 |
+
"grad_norm": 27.244776233826,
|
| 499 |
+
"learning_rate": 2.19029145890313e-07,
|
| 500 |
+
"logits/chosen": 2.048271656036377,
|
| 501 |
+
"logits/rejected": 2.768385887145996,
|
| 502 |
+
"logps/chosen": -433.47479248046875,
|
| 503 |
+
"logps/rejected": -546.029296875,
|
| 504 |
+
"loss": 0.5272,
|
| 505 |
+
"rewards/accuracies": 0.731249988079071,
|
| 506 |
+
"rewards/chosen": -1.7612041234970093,
|
| 507 |
+
"rewards/margins": 1.1764047145843506,
|
| 508 |
+
"rewards/rejected": -2.937608242034912,
|
| 509 |
+
"step": 280
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 0.606694560669456,
|
| 513 |
+
"grad_norm": 21.547725491962094,
|
| 514 |
+
"learning_rate": 2.0100351342479216e-07,
|
| 515 |
+
"logits/chosen": 2.24129056930542,
|
| 516 |
+
"logits/rejected": 2.8873372077941895,
|
| 517 |
+
"logps/chosen": -460.38690185546875,
|
| 518 |
+
"logps/rejected": -570.9521484375,
|
| 519 |
+
"loss": 0.523,
|
| 520 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 521 |
+
"rewards/chosen": -1.6904579401016235,
|
| 522 |
+
"rewards/margins": 1.2296130657196045,
|
| 523 |
+
"rewards/rejected": -2.9200711250305176,
|
| 524 |
+
"step": 290
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 0.6276150627615062,
|
| 528 |
+
"grad_norm": 21.34579167577593,
|
| 529 |
+
"learning_rate": 1.8323929841460178e-07,
|
| 530 |
+
"logits/chosen": 2.917755603790283,
|
| 531 |
+
"logits/rejected": 3.5216128826141357,
|
| 532 |
+
"logps/chosen": -471.07965087890625,
|
| 533 |
+
"logps/rejected": -600.7293701171875,
|
| 534 |
+
"loss": 0.5231,
|
| 535 |
+
"rewards/accuracies": 0.737500011920929,
|
| 536 |
+
"rewards/chosen": -2.0175368785858154,
|
| 537 |
+
"rewards/margins": 1.2063273191452026,
|
| 538 |
+
"rewards/rejected": -3.2238643169403076,
|
| 539 |
+
"step": 300
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 0.6276150627615062,
|
| 543 |
+
"eval_logits/chosen": 2.2075393199920654,
|
| 544 |
+
"eval_logits/rejected": 2.986379384994507,
|
| 545 |
+
"eval_logps/chosen": -445.05584716796875,
|
| 546 |
+
"eval_logps/rejected": -587.8876953125,
|
| 547 |
+
"eval_loss": 0.5511458516120911,
|
| 548 |
+
"eval_rewards/accuracies": 0.765625,
|
| 549 |
+
"eval_rewards/chosen": -1.8242592811584473,
|
| 550 |
+
"eval_rewards/margins": 1.42799711227417,
|
| 551 |
+
"eval_rewards/rejected": -3.252256393432617,
|
| 552 |
+
"eval_runtime": 103.1559,
|
| 553 |
+
"eval_samples_per_second": 19.388,
|
| 554 |
+
"eval_steps_per_second": 0.31,
|
| 555 |
+
"step": 300
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.6485355648535565,
|
| 559 |
+
"grad_norm": 18.70711648745981,
|
| 560 |
+
"learning_rate": 1.6583128063291573e-07,
|
| 561 |
+
"logits/chosen": 1.9323720932006836,
|
| 562 |
+
"logits/rejected": 2.7933290004730225,
|
| 563 |
+
"logps/chosen": -479.4503479003906,
|
| 564 |
+
"logps/rejected": -588.09228515625,
|
| 565 |
+
"loss": 0.5303,
|
| 566 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 567 |
+
"rewards/chosen": -1.8548892736434937,
|
| 568 |
+
"rewards/margins": 1.2552098035812378,
|
| 569 |
+
"rewards/rejected": -3.1100995540618896,
|
| 570 |
+
"step": 310
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 0.6694560669456067,
|
| 574 |
+
"grad_norm": 24.633583083411633,
|
| 575 |
+
"learning_rate": 1.488723393865766e-07,
|
| 576 |
+
"logits/chosen": 1.4087620973587036,
|
| 577 |
+
"logits/rejected": 2.465496778488159,
|
| 578 |
+
"logps/chosen": -487.93646240234375,
|
| 579 |
+
"logps/rejected": -566.53466796875,
|
| 580 |
+
"loss": 0.5063,
|
| 581 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 582 |
+
"rewards/chosen": -1.9570667743682861,
|
| 583 |
+
"rewards/margins": 1.188623070716858,
|
| 584 |
+
"rewards/rejected": -3.1456902027130127,
|
| 585 |
+
"step": 320
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 0.6903765690376569,
|
| 589 |
+
"grad_norm": 20.927701473521733,
|
| 590 |
+
"learning_rate": 1.3245295796480788e-07,
|
| 591 |
+
"logits/chosen": 1.6705989837646484,
|
| 592 |
+
"logits/rejected": 2.787672758102417,
|
| 593 |
+
"logps/chosen": -478.430908203125,
|
| 594 |
+
"logps/rejected": -544.5813598632812,
|
| 595 |
+
"loss": 0.5178,
|
| 596 |
+
"rewards/accuracies": 0.6875,
|
| 597 |
+
"rewards/chosen": -1.948035478591919,
|
| 598 |
+
"rewards/margins": 1.0792946815490723,
|
| 599 |
+
"rewards/rejected": -3.027329921722412,
|
| 600 |
+
"step": 330
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"epoch": 0.7112970711297071,
|
| 604 |
+
"grad_norm": 21.58725862881672,
|
| 605 |
+
"learning_rate": 1.1666074087171627e-07,
|
| 606 |
+
"logits/chosen": 2.2645626068115234,
|
| 607 |
+
"logits/rejected": 3.0073537826538086,
|
| 608 |
+
"logps/chosen": -433.1025390625,
|
| 609 |
+
"logps/rejected": -569.9879150390625,
|
| 610 |
+
"loss": 0.5331,
|
| 611 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 612 |
+
"rewards/chosen": -1.8666340112686157,
|
| 613 |
+
"rewards/margins": 1.2376598119735718,
|
| 614 |
+
"rewards/rejected": -3.1042943000793457,
|
| 615 |
+
"step": 340
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 0.7322175732217573,
|
| 619 |
+
"grad_norm": 22.053584645202807,
|
| 620 |
+
"learning_rate": 1.0157994641835734e-07,
|
| 621 |
+
"logits/chosen": 2.3491084575653076,
|
| 622 |
+
"logits/rejected": 3.309586763381958,
|
| 623 |
+
"logps/chosen": -490.469970703125,
|
| 624 |
+
"logps/rejected": -593.7388916015625,
|
| 625 |
+
"loss": 0.5092,
|
| 626 |
+
"rewards/accuracies": 0.699999988079071,
|
| 627 |
+
"rewards/chosen": -2.0824320316314697,
|
| 628 |
+
"rewards/margins": 1.1687943935394287,
|
| 629 |
+
"rewards/rejected": -3.2512269020080566,
|
| 630 |
+
"step": 350
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 0.7322175732217573,
|
| 634 |
+
"eval_logits/chosen": 2.094608783721924,
|
| 635 |
+
"eval_logits/rejected": 2.88336443901062,
|
| 636 |
+
"eval_logps/chosen": -461.0306701660156,
|
| 637 |
+
"eval_logps/rejected": -602.9060668945312,
|
| 638 |
+
"eval_loss": 0.5401915311813354,
|
| 639 |
+
"eval_rewards/accuracies": 0.7734375,
|
| 640 |
+
"eval_rewards/chosen": -1.9840072393417358,
|
| 641 |
+
"eval_rewards/margins": 1.4184322357177734,
|
| 642 |
+
"eval_rewards/rejected": -3.402439594268799,
|
| 643 |
+
"eval_runtime": 104.739,
|
| 644 |
+
"eval_samples_per_second": 19.095,
|
| 645 |
+
"eval_steps_per_second": 0.306,
|
| 646 |
+
"step": 350
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.7531380753138075,
|
| 650 |
+
"grad_norm": 18.40926814309719,
|
| 651 |
+
"learning_rate": 8.729103716819111e-08,
|
| 652 |
+
"logits/chosen": 2.376570701599121,
|
| 653 |
+
"logits/rejected": 3.0961055755615234,
|
| 654 |
+
"logps/chosen": -463.7430725097656,
|
| 655 |
+
"logps/rejected": -592.4307861328125,
|
| 656 |
+
"loss": 0.5261,
|
| 657 |
+
"rewards/accuracies": 0.737500011920929,
|
| 658 |
+
"rewards/chosen": -2.168846368789673,
|
| 659 |
+
"rewards/margins": 1.2989672422409058,
|
| 660 |
+
"rewards/rejected": -3.467813491821289,
|
| 661 |
+
"step": 360
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 0.7740585774058577,
|
| 665 |
+
"grad_norm": 20.93465319965544,
|
| 666 |
+
"learning_rate": 7.387025063449081e-08,
|
| 667 |
+
"logits/chosen": 1.8760063648223877,
|
| 668 |
+
"logits/rejected": 3.0332190990448,
|
| 669 |
+
"logps/chosen": -485.4361267089844,
|
| 670 |
+
"logps/rejected": -611.39599609375,
|
| 671 |
+
"loss": 0.512,
|
| 672 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 673 |
+
"rewards/chosen": -1.975916862487793,
|
| 674 |
+
"rewards/margins": 1.4137961864471436,
|
| 675 |
+
"rewards/rejected": -3.3897128105163574,
|
| 676 |
+
"step": 370
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"epoch": 0.7949790794979079,
|
| 680 |
+
"grad_norm": 21.53735300502108,
|
| 681 |
+
"learning_rate": 6.138919252022435e-08,
|
| 682 |
+
"logits/chosen": 2.2272870540618896,
|
| 683 |
+
"logits/rejected": 2.9403111934661865,
|
| 684 |
+
"logps/chosen": -529.391845703125,
|
| 685 |
+
"logps/rejected": -598.6041870117188,
|
| 686 |
+
"loss": 0.5243,
|
| 687 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 688 |
+
"rewards/chosen": -2.129279613494873,
|
| 689 |
+
"rewards/margins": 1.0328638553619385,
|
| 690 |
+
"rewards/rejected": -3.1621437072753906,
|
| 691 |
+
"step": 380
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 0.8158995815899581,
|
| 695 |
+
"grad_norm": 22.856314943850858,
|
| 696 |
+
"learning_rate": 4.991445467064689e-08,
|
| 697 |
+
"logits/chosen": 2.5150341987609863,
|
| 698 |
+
"logits/rejected": 3.433137893676758,
|
| 699 |
+
"logps/chosen": -523.1361694335938,
|
| 700 |
+
"logps/rejected": -647.6205444335938,
|
| 701 |
+
"loss": 0.5212,
|
| 702 |
+
"rewards/accuracies": 0.762499988079071,
|
| 703 |
+
"rewards/chosen": -2.1657605171203613,
|
| 704 |
+
"rewards/margins": 1.4086949825286865,
|
| 705 |
+
"rewards/rejected": -3.5744547843933105,
|
| 706 |
+
"step": 390
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 0.8368200836820083,
|
| 710 |
+
"grad_norm": 20.459584660793766,
|
| 711 |
+
"learning_rate": 3.9507259776993954e-08,
|
| 712 |
+
"logits/chosen": 2.373227596282959,
|
| 713 |
+
"logits/rejected": 3.0179686546325684,
|
| 714 |
+
"logps/chosen": -520.0594482421875,
|
| 715 |
+
"logps/rejected": -614.9906005859375,
|
| 716 |
+
"loss": 0.5231,
|
| 717 |
+
"rewards/accuracies": 0.71875,
|
| 718 |
+
"rewards/chosen": -2.2612462043762207,
|
| 719 |
+
"rewards/margins": 1.1028645038604736,
|
| 720 |
+
"rewards/rejected": -3.3641109466552734,
|
| 721 |
+
"step": 400
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 0.8368200836820083,
|
| 725 |
+
"eval_logits/chosen": 2.2364866733551025,
|
| 726 |
+
"eval_logits/rejected": 3.0541534423828125,
|
| 727 |
+
"eval_logps/chosen": -472.12713623046875,
|
| 728 |
+
"eval_logps/rejected": -619.1116333007812,
|
| 729 |
+
"eval_loss": 0.5417460799217224,
|
| 730 |
+
"eval_rewards/accuracies": 0.78125,
|
| 731 |
+
"eval_rewards/chosen": -2.0949723720550537,
|
| 732 |
+
"eval_rewards/margins": 1.4695227146148682,
|
| 733 |
+
"eval_rewards/rejected": -3.564495086669922,
|
| 734 |
+
"eval_runtime": 103.2544,
|
| 735 |
+
"eval_samples_per_second": 19.37,
|
| 736 |
+
"eval_steps_per_second": 0.31,
|
| 737 |
+
"step": 400
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.8577405857740585,
|
| 741 |
+
"grad_norm": 20.75509691260348,
|
| 742 |
+
"learning_rate": 3.022313472693447e-08,
|
| 743 |
+
"logits/chosen": 1.8815155029296875,
|
| 744 |
+
"logits/rejected": 2.752415657043457,
|
| 745 |
+
"logps/chosen": -530.4354248046875,
|
| 746 |
+
"logps/rejected": -645.202392578125,
|
| 747 |
+
"loss": 0.514,
|
| 748 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 749 |
+
"rewards/chosen": -2.134329319000244,
|
| 750 |
+
"rewards/margins": 1.3350379467010498,
|
| 751 |
+
"rewards/rejected": -3.469367504119873,
|
| 752 |
+
"step": 410
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 0.8786610878661087,
|
| 756 |
+
"grad_norm": 21.72917985290927,
|
| 757 |
+
"learning_rate": 2.2111614344599684e-08,
|
| 758 |
+
"logits/chosen": 2.2683472633361816,
|
| 759 |
+
"logits/rejected": 3.545788526535034,
|
| 760 |
+
"logps/chosen": -509.83563232421875,
|
| 761 |
+
"logps/rejected": -617.6029052734375,
|
| 762 |
+
"loss": 0.4959,
|
| 763 |
+
"rewards/accuracies": 0.75,
|
| 764 |
+
"rewards/chosen": -2.1364758014678955,
|
| 765 |
+
"rewards/margins": 1.329615831375122,
|
| 766 |
+
"rewards/rejected": -3.4660911560058594,
|
| 767 |
+
"step": 420
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 0.899581589958159,
|
| 771 |
+
"grad_norm": 20.573623570397274,
|
| 772 |
+
"learning_rate": 1.521597710086439e-08,
|
| 773 |
+
"logits/chosen": 2.243511438369751,
|
| 774 |
+
"logits/rejected": 3.2753005027770996,
|
| 775 |
+
"logps/chosen": -517.85498046875,
|
| 776 |
+
"logps/rejected": -601.6973876953125,
|
| 777 |
+
"loss": 0.5284,
|
| 778 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 779 |
+
"rewards/chosen": -2.218738317489624,
|
| 780 |
+
"rewards/margins": 1.2048834562301636,
|
| 781 |
+
"rewards/rejected": -3.4236221313476562,
|
| 782 |
+
"step": 430
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 0.9205020920502092,
|
| 786 |
+
"grad_norm": 17.236420967349105,
|
| 787 |
+
"learning_rate": 9.57301420397924e-09,
|
| 788 |
+
"logits/chosen": 2.407790422439575,
|
| 789 |
+
"logits/rejected": 3.659886598587036,
|
| 790 |
+
"logps/chosen": -511.99688720703125,
|
| 791 |
+
"logps/rejected": -602.4998168945312,
|
| 792 |
+
"loss": 0.5239,
|
| 793 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 794 |
+
"rewards/chosen": -2.2972302436828613,
|
| 795 |
+
"rewards/margins": 1.267395257949829,
|
| 796 |
+
"rewards/rejected": -3.5646255016326904,
|
| 797 |
+
"step": 440
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"epoch": 0.9414225941422594,
|
| 801 |
+
"grad_norm": 18.669490237353997,
|
| 802 |
+
"learning_rate": 5.212833302556258e-09,
|
| 803 |
+
"logits/chosen": 2.2820138931274414,
|
| 804 |
+
"logits/rejected": 3.1481757164001465,
|
| 805 |
+
"logps/chosen": -492.22052001953125,
|
| 806 |
+
"logps/rejected": -632.2754516601562,
|
| 807 |
+
"loss": 0.5232,
|
| 808 |
+
"rewards/accuracies": 0.768750011920929,
|
| 809 |
+
"rewards/chosen": -2.074744462966919,
|
| 810 |
+
"rewards/margins": 1.4727141857147217,
|
| 811 |
+
"rewards/rejected": -3.5474586486816406,
|
| 812 |
+
"step": 450
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 0.9414225941422594,
|
| 816 |
+
"eval_logits/chosen": 2.1926114559173584,
|
| 817 |
+
"eval_logits/rejected": 3.032156467437744,
|
| 818 |
+
"eval_logps/chosen": -469.2007751464844,
|
| 819 |
+
"eval_logps/rejected": -617.9429931640625,
|
| 820 |
+
"eval_loss": 0.5418744087219238,
|
| 821 |
+
"eval_rewards/accuracies": 0.78125,
|
| 822 |
+
"eval_rewards/chosen": -2.0657081604003906,
|
| 823 |
+
"eval_rewards/margins": 1.487100601196289,
|
| 824 |
+
"eval_rewards/rejected": -3.552809000015259,
|
| 825 |
+
"eval_runtime": 103.486,
|
| 826 |
+
"eval_samples_per_second": 19.326,
|
| 827 |
+
"eval_steps_per_second": 0.309,
|
| 828 |
+
"step": 450
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.9623430962343096,
|
| 832 |
+
"grad_norm": 21.564710597734962,
|
| 833 |
+
"learning_rate": 2.158697848236607e-09,
|
| 834 |
+
"logits/chosen": 2.6190943717956543,
|
| 835 |
+
"logits/rejected": 3.548638105392456,
|
| 836 |
+
"logps/chosen": -499.64208984375,
|
| 837 |
+
"logps/rejected": -623.9893188476562,
|
| 838 |
+
"loss": 0.5095,
|
| 839 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 840 |
+
"rewards/chosen": -2.363645076751709,
|
| 841 |
+
"rewards/margins": 1.3193647861480713,
|
| 842 |
+
"rewards/rejected": -3.683009624481201,
|
| 843 |
+
"step": 460
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 0.9832635983263598,
|
| 847 |
+
"grad_norm": 17.345861357972396,
|
| 848 |
+
"learning_rate": 4.269029751107489e-10,
|
| 849 |
+
"logits/chosen": 2.4917562007904053,
|
| 850 |
+
"logits/rejected": 2.865358591079712,
|
| 851 |
+
"logps/chosen": -476.9666442871094,
|
| 852 |
+
"logps/rejected": -609.1448974609375,
|
| 853 |
+
"loss": 0.5372,
|
| 854 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 855 |
+
"rewards/chosen": -2.138411045074463,
|
| 856 |
+
"rewards/margins": 1.2895748615264893,
|
| 857 |
+
"rewards/rejected": -3.427985668182373,
|
| 858 |
+
"step": 470
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 1.0,
|
| 862 |
+
"step": 478,
|
| 863 |
+
"total_flos": 0.0,
|
| 864 |
+
"train_loss": 0.5571172007955767,
|
| 865 |
+
"train_runtime": 12724.9946,
|
| 866 |
+
"train_samples_per_second": 4.804,
|
| 867 |
+
"train_steps_per_second": 0.038
|
| 868 |
+
}
|
| 869 |
+
],
|
| 870 |
+
"logging_steps": 10,
|
| 871 |
+
"max_steps": 478,
|
| 872 |
+
"num_input_tokens_seen": 0,
|
| 873 |
+
"num_train_epochs": 1,
|
| 874 |
+
"save_steps": 100,
|
| 875 |
+
"stateful_callbacks": {
|
| 876 |
+
"TrainerControl": {
|
| 877 |
+
"args": {
|
| 878 |
+
"should_epoch_stop": false,
|
| 879 |
+
"should_evaluate": false,
|
| 880 |
+
"should_log": false,
|
| 881 |
+
"should_save": true,
|
| 882 |
+
"should_training_stop": true
|
| 883 |
+
},
|
| 884 |
+
"attributes": {}
|
| 885 |
+
}
|
| 886 |
+
},
|
| 887 |
+
"total_flos": 0.0,
|
| 888 |
+
"train_batch_size": 8,
|
| 889 |
+
"trial_name": null,
|
| 890 |
+
"trial_params": null
|
| 891 |
+
}
|