Model save
Browse files- README.md +78 -0
- all_results.json +9 -0
- generation_config.json +6 -0
- train_results.json +9 -0
- trainer_state.json +992 -0
README.md
ADDED
|
@@ -0,0 +1,78 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: transformers
|
| 3 |
+
license: apache-2.0
|
| 4 |
+
base_model: alignment-handbook/zephyr-7b-sft-full
|
| 5 |
+
tags:
|
| 6 |
+
- trl
|
| 7 |
+
- dpo
|
| 8 |
+
- generated_from_trainer
|
| 9 |
+
model-index:
|
| 10 |
+
- name: zephyr-7b-align-scan-7e-07-0.99-cosine-2.0
|
| 11 |
+
results: []
|
| 12 |
+
---
|
| 13 |
+
|
| 14 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 15 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 16 |
+
|
| 17 |
+
# zephyr-7b-align-scan-7e-07-0.99-cosine-2.0
|
| 18 |
+
|
| 19 |
+
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.
|
| 20 |
+
It achieves the following results on the evaluation set:
|
| 21 |
+
- Loss: 1.1824
|
| 22 |
+
- Rewards/chosen: 2.5463
|
| 23 |
+
- Rewards/rejected: 0.5342
|
| 24 |
+
- Rewards/accuracies: 0.3452
|
| 25 |
+
- Rewards/margins: 2.0120
|
| 26 |
+
- Logps/rejected: -80.5887
|
| 27 |
+
- Logps/chosen: -71.9193
|
| 28 |
+
- Logits/rejected: -2.6463
|
| 29 |
+
- Logits/chosen: -2.6632
|
| 30 |
+
|
| 31 |
+
## Model description
|
| 32 |
+
|
| 33 |
+
More information needed
|
| 34 |
+
|
| 35 |
+
## Intended uses & limitations
|
| 36 |
+
|
| 37 |
+
More information needed
|
| 38 |
+
|
| 39 |
+
## Training and evaluation data
|
| 40 |
+
|
| 41 |
+
More information needed
|
| 42 |
+
|
| 43 |
+
## Training procedure
|
| 44 |
+
|
| 45 |
+
### Training hyperparameters
|
| 46 |
+
|
| 47 |
+
The following hyperparameters were used during training:
|
| 48 |
+
- learning_rate: 7e-07
|
| 49 |
+
- train_batch_size: 8
|
| 50 |
+
- eval_batch_size: 8
|
| 51 |
+
- seed: 42
|
| 52 |
+
- distributed_type: multi-GPU
|
| 53 |
+
- num_devices: 4
|
| 54 |
+
- gradient_accumulation_steps: 2
|
| 55 |
+
- total_train_batch_size: 64
|
| 56 |
+
- total_eval_batch_size: 32
|
| 57 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 58 |
+
- lr_scheduler_type: cosine
|
| 59 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 60 |
+
- num_epochs: 2
|
| 61 |
+
|
| 62 |
+
### Training results
|
| 63 |
+
|
| 64 |
+
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|
| 65 |
+
|:-------------:|:------:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
|
| 66 |
+
| 0.9118 | 0.3484 | 100 | 0.8952 | 1.8796 | 1.1355 | 0.3353 | 0.7441 | -79.9814 | -72.5926 | -2.5564 | -2.5727 |
|
| 67 |
+
| 0.9553 | 0.6969 | 200 | 1.0700 | 2.5989 | 1.4006 | 0.3413 | 1.1983 | -79.7136 | -71.8661 | -2.5726 | -2.5893 |
|
| 68 |
+
| 0.4066 | 1.0453 | 300 | 1.0729 | 2.3164 | 0.9125 | 0.3433 | 1.4038 | -80.2066 | -72.1515 | -2.5962 | -2.6126 |
|
| 69 |
+
| 0.3805 | 1.3937 | 400 | 1.1546 | 2.9774 | 1.1937 | 0.3373 | 1.7837 | -79.9225 | -71.4837 | -2.6247 | -2.6413 |
|
| 70 |
+
| 0.3975 | 1.7422 | 500 | 1.1824 | 2.5463 | 0.5342 | 0.3452 | 2.0120 | -80.5887 | -71.9193 | -2.6463 | -2.6632 |
|
| 71 |
+
|
| 72 |
+
|
| 73 |
+
### Framework versions
|
| 74 |
+
|
| 75 |
+
- Transformers 4.44.2
|
| 76 |
+
- Pytorch 2.4.0
|
| 77 |
+
- Datasets 2.21.0
|
| 78 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.0,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.6661292189920406,
|
| 5 |
+
"train_runtime": 6507.4651,
|
| 6 |
+
"train_samples": 18340,
|
| 7 |
+
"train_samples_per_second": 5.637,
|
| 8 |
+
"train_steps_per_second": 0.088
|
| 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": "4.44.2"
|
| 6 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.0,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.6661292189920406,
|
| 5 |
+
"train_runtime": 6507.4651,
|
| 6 |
+
"train_samples": 18340,
|
| 7 |
+
"train_samples_per_second": 5.637,
|
| 8 |
+
"train_steps_per_second": 0.088
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,992 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.0,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 574,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.003484320557491289,
|
| 13 |
+
"grad_norm": 572.2640100744908,
|
| 14 |
+
"learning_rate": 1.2068965517241378e-08,
|
| 15 |
+
"logits/chosen": -2.5345611572265625,
|
| 16 |
+
"logits/rejected": -2.581700563430786,
|
| 17 |
+
"logps/chosen": -60.002105712890625,
|
| 18 |
+
"logps/rejected": -99.98374938964844,
|
| 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": 0.03484320557491289,
|
| 28 |
+
"grad_norm": 548.4940657606528,
|
| 29 |
+
"learning_rate": 1.206896551724138e-07,
|
| 30 |
+
"logits/chosen": -2.5637850761413574,
|
| 31 |
+
"logits/rejected": -2.562532424926758,
|
| 32 |
+
"logps/chosen": -59.66712188720703,
|
| 33 |
+
"logps/rejected": -73.37606811523438,
|
| 34 |
+
"loss": 0.7017,
|
| 35 |
+
"rewards/accuracies": 0.1805555522441864,
|
| 36 |
+
"rewards/chosen": -0.00738478871062398,
|
| 37 |
+
"rewards/margins": -0.01218735333532095,
|
| 38 |
+
"rewards/rejected": 0.00480256462469697,
|
| 39 |
+
"step": 10
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.06968641114982578,
|
| 43 |
+
"grad_norm": 669.195296069125,
|
| 44 |
+
"learning_rate": 2.413793103448276e-07,
|
| 45 |
+
"logits/chosen": -2.606231927871704,
|
| 46 |
+
"logits/rejected": -2.565000534057617,
|
| 47 |
+
"logps/chosen": -104.05134582519531,
|
| 48 |
+
"logps/rejected": -94.89227294921875,
|
| 49 |
+
"loss": 0.6861,
|
| 50 |
+
"rewards/accuracies": 0.3687500059604645,
|
| 51 |
+
"rewards/chosen": 0.07496137917041779,
|
| 52 |
+
"rewards/margins": 0.07313241064548492,
|
| 53 |
+
"rewards/rejected": 0.001828978187404573,
|
| 54 |
+
"step": 20
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.10452961672473868,
|
| 58 |
+
"grad_norm": 701.7098593476545,
|
| 59 |
+
"learning_rate": 3.620689655172414e-07,
|
| 60 |
+
"logits/chosen": -2.5953564643859863,
|
| 61 |
+
"logits/rejected": -2.575517177581787,
|
| 62 |
+
"logps/chosen": -82.11582946777344,
|
| 63 |
+
"logps/rejected": -91.40339660644531,
|
| 64 |
+
"loss": 0.6768,
|
| 65 |
+
"rewards/accuracies": 0.36250001192092896,
|
| 66 |
+
"rewards/chosen": 0.37693265080451965,
|
| 67 |
+
"rewards/margins": 0.24151858687400818,
|
| 68 |
+
"rewards/rejected": 0.1354140341281891,
|
| 69 |
+
"step": 30
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.13937282229965156,
|
| 73 |
+
"grad_norm": 569.2211096554255,
|
| 74 |
+
"learning_rate": 4.827586206896552e-07,
|
| 75 |
+
"logits/chosen": -2.49927020072937,
|
| 76 |
+
"logits/rejected": -2.497345209121704,
|
| 77 |
+
"logps/chosen": -77.96027374267578,
|
| 78 |
+
"logps/rejected": -73.55770111083984,
|
| 79 |
+
"loss": 0.6699,
|
| 80 |
+
"rewards/accuracies": 0.3187499940395355,
|
| 81 |
+
"rewards/chosen": -0.04143521189689636,
|
| 82 |
+
"rewards/margins": 0.5093806982040405,
|
| 83 |
+
"rewards/rejected": -0.5508158802986145,
|
| 84 |
+
"step": 40
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.17421602787456447,
|
| 88 |
+
"grad_norm": 446.25211605228685,
|
| 89 |
+
"learning_rate": 6.034482758620689e-07,
|
| 90 |
+
"logits/chosen": -2.5248429775238037,
|
| 91 |
+
"logits/rejected": -2.529026508331299,
|
| 92 |
+
"logps/chosen": -63.092262268066406,
|
| 93 |
+
"logps/rejected": -75.61325073242188,
|
| 94 |
+
"loss": 0.7196,
|
| 95 |
+
"rewards/accuracies": 0.30000001192092896,
|
| 96 |
+
"rewards/chosen": 0.838420033454895,
|
| 97 |
+
"rewards/margins": 0.35312479734420776,
|
| 98 |
+
"rewards/rejected": 0.48529529571533203,
|
| 99 |
+
"step": 50
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.20905923344947736,
|
| 103 |
+
"grad_norm": 431.8823983807605,
|
| 104 |
+
"learning_rate": 6.999740526496426e-07,
|
| 105 |
+
"logits/chosen": -2.494865894317627,
|
| 106 |
+
"logits/rejected": -2.4892258644104004,
|
| 107 |
+
"logps/chosen": -70.98899841308594,
|
| 108 |
+
"logps/rejected": -66.5857925415039,
|
| 109 |
+
"loss": 0.7101,
|
| 110 |
+
"rewards/accuracies": 0.32499998807907104,
|
| 111 |
+
"rewards/chosen": 2.1953351497650146,
|
| 112 |
+
"rewards/margins": 0.3907918632030487,
|
| 113 |
+
"rewards/rejected": 1.8045432567596436,
|
| 114 |
+
"step": 60
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.24390243902439024,
|
| 118 |
+
"grad_norm": 543.5545103197261,
|
| 119 |
+
"learning_rate": 6.990662992822431e-07,
|
| 120 |
+
"logits/chosen": -2.5150294303894043,
|
| 121 |
+
"logits/rejected": -2.510296583175659,
|
| 122 |
+
"logps/chosen": -61.68668746948242,
|
| 123 |
+
"logps/rejected": -66.4362564086914,
|
| 124 |
+
"loss": 0.7767,
|
| 125 |
+
"rewards/accuracies": 0.3062500059604645,
|
| 126 |
+
"rewards/chosen": 2.524014949798584,
|
| 127 |
+
"rewards/margins": 0.5535615086555481,
|
| 128 |
+
"rewards/rejected": 1.9704535007476807,
|
| 129 |
+
"step": 70
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.2787456445993031,
|
| 133 |
+
"grad_norm": 602.8845854238214,
|
| 134 |
+
"learning_rate": 6.96865023062192e-07,
|
| 135 |
+
"logits/chosen": -2.461080551147461,
|
| 136 |
+
"logits/rejected": -2.452082395553589,
|
| 137 |
+
"logps/chosen": -73.27593231201172,
|
| 138 |
+
"logps/rejected": -75.97522735595703,
|
| 139 |
+
"loss": 0.8282,
|
| 140 |
+
"rewards/accuracies": 0.3125,
|
| 141 |
+
"rewards/chosen": 2.365041732788086,
|
| 142 |
+
"rewards/margins": 0.8407068252563477,
|
| 143 |
+
"rewards/rejected": 1.5243349075317383,
|
| 144 |
+
"step": 80
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 0.313588850174216,
|
| 148 |
+
"grad_norm": 645.4166740504302,
|
| 149 |
+
"learning_rate": 6.93378381182268e-07,
|
| 150 |
+
"logits/chosen": -2.5054688453674316,
|
| 151 |
+
"logits/rejected": -2.5200397968292236,
|
| 152 |
+
"logps/chosen": -63.54323196411133,
|
| 153 |
+
"logps/rejected": -68.08236694335938,
|
| 154 |
+
"loss": 0.9178,
|
| 155 |
+
"rewards/accuracies": 0.29374998807907104,
|
| 156 |
+
"rewards/chosen": 2.4146523475646973,
|
| 157 |
+
"rewards/margins": 0.40329408645629883,
|
| 158 |
+
"rewards/rejected": 2.0113582611083984,
|
| 159 |
+
"step": 90
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.34843205574912894,
|
| 163 |
+
"grad_norm": 625.9456943514227,
|
| 164 |
+
"learning_rate": 6.886192939700987e-07,
|
| 165 |
+
"logits/chosen": -2.495720863342285,
|
| 166 |
+
"logits/rejected": -2.494643449783325,
|
| 167 |
+
"logps/chosen": -72.901611328125,
|
| 168 |
+
"logps/rejected": -79.84459686279297,
|
| 169 |
+
"loss": 0.9118,
|
| 170 |
+
"rewards/accuracies": 0.32499998807907104,
|
| 171 |
+
"rewards/chosen": 2.345865488052368,
|
| 172 |
+
"rewards/margins": 1.1508872509002686,
|
| 173 |
+
"rewards/rejected": 1.1949782371520996,
|
| 174 |
+
"step": 100
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 0.34843205574912894,
|
| 178 |
+
"eval_logits/chosen": -2.572685718536377,
|
| 179 |
+
"eval_logits/rejected": -2.556438684463501,
|
| 180 |
+
"eval_logps/chosen": -72.59263610839844,
|
| 181 |
+
"eval_logps/rejected": -79.98140716552734,
|
| 182 |
+
"eval_loss": 0.8951926827430725,
|
| 183 |
+
"eval_rewards/accuracies": 0.335317462682724,
|
| 184 |
+
"eval_rewards/chosen": 1.8796159029006958,
|
| 185 |
+
"eval_rewards/margins": 0.7441409230232239,
|
| 186 |
+
"eval_rewards/rejected": 1.1354750394821167,
|
| 187 |
+
"eval_runtime": 114.849,
|
| 188 |
+
"eval_samples_per_second": 17.414,
|
| 189 |
+
"eval_steps_per_second": 0.549,
|
| 190 |
+
"step": 100
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 0.3832752613240418,
|
| 194 |
+
"grad_norm": 633.8111608364293,
|
| 195 |
+
"learning_rate": 6.826053970097538e-07,
|
| 196 |
+
"logits/chosen": -2.5071868896484375,
|
| 197 |
+
"logits/rejected": -2.4732370376586914,
|
| 198 |
+
"logps/chosen": -72.26206970214844,
|
| 199 |
+
"logps/rejected": -62.43015670776367,
|
| 200 |
+
"loss": 0.934,
|
| 201 |
+
"rewards/accuracies": 0.2562499940395355,
|
| 202 |
+
"rewards/chosen": 0.8988760709762573,
|
| 203 |
+
"rewards/margins": 0.3242531418800354,
|
| 204 |
+
"rewards/rejected": 0.5746229887008667,
|
| 205 |
+
"step": 110
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.4181184668989547,
|
| 209 |
+
"grad_norm": 481.08142593629555,
|
| 210 |
+
"learning_rate": 6.753589757901721e-07,
|
| 211 |
+
"logits/chosen": -2.5381650924682617,
|
| 212 |
+
"logits/rejected": -2.50740385055542,
|
| 213 |
+
"logps/chosen": -76.40122985839844,
|
| 214 |
+
"logps/rejected": -66.65751647949219,
|
| 215 |
+
"loss": 0.857,
|
| 216 |
+
"rewards/accuracies": 0.29374998807907104,
|
| 217 |
+
"rewards/chosen": 1.0497384071350098,
|
| 218 |
+
"rewards/margins": 1.0358079671859741,
|
| 219 |
+
"rewards/rejected": 0.013930544257164001,
|
| 220 |
+
"step": 120
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.4529616724738676,
|
| 224 |
+
"grad_norm": 807.6631711479599,
|
| 225 |
+
"learning_rate": 6.669068831226014e-07,
|
| 226 |
+
"logits/chosen": -2.5767252445220947,
|
| 227 |
+
"logits/rejected": -2.557955265045166,
|
| 228 |
+
"logps/chosen": -83.18122863769531,
|
| 229 |
+
"logps/rejected": -88.09697723388672,
|
| 230 |
+
"loss": 1.0372,
|
| 231 |
+
"rewards/accuracies": 0.33125001192092896,
|
| 232 |
+
"rewards/chosen": 0.6925557255744934,
|
| 233 |
+
"rewards/margins": 1.275132417678833,
|
| 234 |
+
"rewards/rejected": -0.5825767517089844,
|
| 235 |
+
"step": 130
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 0.4878048780487805,
|
| 239 |
+
"grad_norm": 443.55201800181277,
|
| 240 |
+
"learning_rate": 6.572804396330676e-07,
|
| 241 |
+
"logits/chosen": -2.4845736026763916,
|
| 242 |
+
"logits/rejected": -2.4748897552490234,
|
| 243 |
+
"logps/chosen": -79.59736633300781,
|
| 244 |
+
"logps/rejected": -70.54890441894531,
|
| 245 |
+
"loss": 0.9077,
|
| 246 |
+
"rewards/accuracies": 0.33125001192092896,
|
| 247 |
+
"rewards/chosen": 1.5325819253921509,
|
| 248 |
+
"rewards/margins": 1.3193892240524292,
|
| 249 |
+
"rewards/rejected": 0.21319285035133362,
|
| 250 |
+
"step": 140
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 0.5226480836236934,
|
| 254 |
+
"grad_norm": 574.2676129428694,
|
| 255 |
+
"learning_rate": 6.465153176986211e-07,
|
| 256 |
+
"logits/chosen": -2.579017162322998,
|
| 257 |
+
"logits/rejected": -2.5382275581359863,
|
| 258 |
+
"logps/chosen": -77.61248016357422,
|
| 259 |
+
"logps/rejected": -78.82749938964844,
|
| 260 |
+
"loss": 1.0493,
|
| 261 |
+
"rewards/accuracies": 0.2750000059604645,
|
| 262 |
+
"rewards/chosen": 1.853137731552124,
|
| 263 |
+
"rewards/margins": 1.137075662612915,
|
| 264 |
+
"rewards/rejected": 0.7160621881484985,
|
| 265 |
+
"step": 150
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 0.5574912891986062,
|
| 269 |
+
"grad_norm": 575.7733117089178,
|
| 270 |
+
"learning_rate": 6.346514092574479e-07,
|
| 271 |
+
"logits/chosen": -2.573914051055908,
|
| 272 |
+
"logits/rejected": -2.5928311347961426,
|
| 273 |
+
"logps/chosen": -62.48719024658203,
|
| 274 |
+
"logps/rejected": -70.96321105957031,
|
| 275 |
+
"loss": 0.9487,
|
| 276 |
+
"rewards/accuracies": 0.3062500059604645,
|
| 277 |
+
"rewards/chosen": 2.017285108566284,
|
| 278 |
+
"rewards/margins": 1.0176817178726196,
|
| 279 |
+
"rewards/rejected": 0.9996035695075989,
|
| 280 |
+
"step": 160
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 0.5923344947735192,
|
| 284 |
+
"grad_norm": 636.1057741626599,
|
| 285 |
+
"learning_rate": 6.21732677982701e-07,
|
| 286 |
+
"logits/chosen": -2.5915422439575195,
|
| 287 |
+
"logits/rejected": -2.5760223865509033,
|
| 288 |
+
"logps/chosen": -67.14725494384766,
|
| 289 |
+
"logps/rejected": -75.51747131347656,
|
| 290 |
+
"loss": 0.9285,
|
| 291 |
+
"rewards/accuracies": 0.26875001192092896,
|
| 292 |
+
"rewards/chosen": 1.8019440174102783,
|
| 293 |
+
"rewards/margins": 1.0133386850357056,
|
| 294 |
+
"rewards/rejected": 0.7886053919792175,
|
| 295 |
+
"step": 170
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 0.627177700348432,
|
| 299 |
+
"grad_norm": 634.0125212661374,
|
| 300 |
+
"learning_rate": 6.078069963678453e-07,
|
| 301 |
+
"logits/chosen": -2.6139795780181885,
|
| 302 |
+
"logits/rejected": -2.60251522064209,
|
| 303 |
+
"logps/chosen": -89.45849609375,
|
| 304 |
+
"logps/rejected": -84.91694641113281,
|
| 305 |
+
"loss": 1.1004,
|
| 306 |
+
"rewards/accuracies": 0.3375000059604645,
|
| 307 |
+
"rewards/chosen": 2.3746800422668457,
|
| 308 |
+
"rewards/margins": 0.797105073928833,
|
| 309 |
+
"rewards/rejected": 1.577574610710144,
|
| 310 |
+
"step": 180
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.662020905923345,
|
| 314 |
+
"grad_norm": 494.843810287881,
|
| 315 |
+
"learning_rate": 5.929259683272219e-07,
|
| 316 |
+
"logits/chosen": -2.602254629135132,
|
| 317 |
+
"logits/rejected": -2.594606399536133,
|
| 318 |
+
"logps/chosen": -68.91534423828125,
|
| 319 |
+
"logps/rejected": -80.00617218017578,
|
| 320 |
+
"loss": 0.9944,
|
| 321 |
+
"rewards/accuracies": 0.28125,
|
| 322 |
+
"rewards/chosen": 2.114778518676758,
|
| 323 |
+
"rewards/margins": 0.607982873916626,
|
| 324 |
+
"rewards/rejected": 1.5067954063415527,
|
| 325 |
+
"step": 190
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.6968641114982579,
|
| 329 |
+
"grad_norm": 826.9190409978122,
|
| 330 |
+
"learning_rate": 5.771447379692167e-07,
|
| 331 |
+
"logits/chosen": -2.625168800354004,
|
| 332 |
+
"logits/rejected": -2.6301324367523193,
|
| 333 |
+
"logps/chosen": -87.10128784179688,
|
| 334 |
+
"logps/rejected": -90.02182006835938,
|
| 335 |
+
"loss": 0.9553,
|
| 336 |
+
"rewards/accuracies": 0.3687500059604645,
|
| 337 |
+
"rewards/chosen": 3.519533157348633,
|
| 338 |
+
"rewards/margins": 1.5109440088272095,
|
| 339 |
+
"rewards/rejected": 2.008589029312134,
|
| 340 |
+
"step": 200
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 0.6968641114982579,
|
| 344 |
+
"eval_logits/chosen": -2.58929705619812,
|
| 345 |
+
"eval_logits/rejected": -2.5725698471069336,
|
| 346 |
+
"eval_logps/chosen": -71.86607360839844,
|
| 347 |
+
"eval_logps/rejected": -79.71355438232422,
|
| 348 |
+
"eval_loss": 1.070008635520935,
|
| 349 |
+
"eval_rewards/accuracies": 0.341269850730896,
|
| 350 |
+
"eval_rewards/chosen": 2.5989110469818115,
|
| 351 |
+
"eval_rewards/margins": 1.1982702016830444,
|
| 352 |
+
"eval_rewards/rejected": 1.4006409645080566,
|
| 353 |
+
"eval_runtime": 113.335,
|
| 354 |
+
"eval_samples_per_second": 17.647,
|
| 355 |
+
"eval_steps_per_second": 0.556,
|
| 356 |
+
"step": 200
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.7317073170731707,
|
| 360 |
+
"grad_norm": 847.0633281542522,
|
| 361 |
+
"learning_rate": 5.605217852506545e-07,
|
| 362 |
+
"logits/chosen": -2.6012160778045654,
|
| 363 |
+
"logits/rejected": -2.57619047164917,
|
| 364 |
+
"logps/chosen": -67.00978088378906,
|
| 365 |
+
"logps/rejected": -62.69084548950195,
|
| 366 |
+
"loss": 0.9355,
|
| 367 |
+
"rewards/accuracies": 0.3687500059604645,
|
| 368 |
+
"rewards/chosen": 2.286196231842041,
|
| 369 |
+
"rewards/margins": 1.579746961593628,
|
| 370 |
+
"rewards/rejected": 0.7064491510391235,
|
| 371 |
+
"step": 210
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 0.7665505226480837,
|
| 375 |
+
"grad_norm": 558.5058882421968,
|
| 376 |
+
"learning_rate": 5.43118709269656e-07,
|
| 377 |
+
"logits/chosen": -2.652629852294922,
|
| 378 |
+
"logits/rejected": -2.6338040828704834,
|
| 379 |
+
"logps/chosen": -71.07757568359375,
|
| 380 |
+
"logps/rejected": -69.96295928955078,
|
| 381 |
+
"loss": 1.0861,
|
| 382 |
+
"rewards/accuracies": 0.2562499940395355,
|
| 383 |
+
"rewards/chosen": 2.3567943572998047,
|
| 384 |
+
"rewards/margins": 0.6735709309577942,
|
| 385 |
+
"rewards/rejected": 1.6832237243652344,
|
| 386 |
+
"step": 220
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.8013937282229965,
|
| 390 |
+
"grad_norm": 874.1019734865492,
|
| 391 |
+
"learning_rate": 5.25e-07,
|
| 392 |
+
"logits/chosen": -2.675318956375122,
|
| 393 |
+
"logits/rejected": -2.6557412147521973,
|
| 394 |
+
"logps/chosen": -86.0287857055664,
|
| 395 |
+
"logps/rejected": -86.36713409423828,
|
| 396 |
+
"loss": 1.1663,
|
| 397 |
+
"rewards/accuracies": 0.3375000059604645,
|
| 398 |
+
"rewards/chosen": 3.9321389198303223,
|
| 399 |
+
"rewards/margins": 2.168241024017334,
|
| 400 |
+
"rewards/rejected": 1.7638976573944092,
|
| 401 |
+
"step": 230
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.8362369337979094,
|
| 405 |
+
"grad_norm": 612.8944076163772,
|
| 406 |
+
"learning_rate": 5.062327993128697e-07,
|
| 407 |
+
"logits/chosen": -2.6777350902557373,
|
| 408 |
+
"logits/rejected": -2.6439225673675537,
|
| 409 |
+
"logps/chosen": -82.02351379394531,
|
| 410 |
+
"logps/rejected": -76.44059753417969,
|
| 411 |
+
"loss": 1.1645,
|
| 412 |
+
"rewards/accuracies": 0.33125001192092896,
|
| 413 |
+
"rewards/chosen": 4.670910835266113,
|
| 414 |
+
"rewards/margins": 1.0123672485351562,
|
| 415 |
+
"rewards/rejected": 3.658543825149536,
|
| 416 |
+
"step": 240
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.8710801393728222,
|
| 420 |
+
"grad_norm": 661.3902849256391,
|
| 421 |
+
"learning_rate": 4.868866521715546e-07,
|
| 422 |
+
"logits/chosen": -2.6785645484924316,
|
| 423 |
+
"logits/rejected": -2.642632007598877,
|
| 424 |
+
"logps/chosen": -91.0850601196289,
|
| 425 |
+
"logps/rejected": -86.79383850097656,
|
| 426 |
+
"loss": 0.8849,
|
| 427 |
+
"rewards/accuracies": 0.3687500059604645,
|
| 428 |
+
"rewards/chosen": 4.320570468902588,
|
| 429 |
+
"rewards/margins": 1.1997525691986084,
|
| 430 |
+
"rewards/rejected": 3.1208178997039795,
|
| 431 |
+
"step": 250
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 0.9059233449477352,
|
| 435 |
+
"grad_norm": 559.1976202405948,
|
| 436 |
+
"learning_rate": 4.6703324892109645e-07,
|
| 437 |
+
"logits/chosen": -2.567991256713867,
|
| 438 |
+
"logits/rejected": -2.581616163253784,
|
| 439 |
+
"logps/chosen": -56.56006622314453,
|
| 440 |
+
"logps/rejected": -64.0355224609375,
|
| 441 |
+
"loss": 1.1214,
|
| 442 |
+
"rewards/accuracies": 0.2874999940395355,
|
| 443 |
+
"rewards/chosen": 2.46440052986145,
|
| 444 |
+
"rewards/margins": 0.8875689506530762,
|
| 445 |
+
"rewards/rejected": 1.5768316984176636,
|
| 446 |
+
"step": 260
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.9407665505226481,
|
| 450 |
+
"grad_norm": 923.7220287977552,
|
| 451 |
+
"learning_rate": 4.4674615962787004e-07,
|
| 452 |
+
"logits/chosen": -2.647984504699707,
|
| 453 |
+
"logits/rejected": -2.646916151046753,
|
| 454 |
+
"logps/chosen": -66.90876770019531,
|
| 455 |
+
"logps/rejected": -81.64918518066406,
|
| 456 |
+
"loss": 1.0676,
|
| 457 |
+
"rewards/accuracies": 0.3125,
|
| 458 |
+
"rewards/chosen": 1.9516960382461548,
|
| 459 |
+
"rewards/margins": 1.0779374837875366,
|
| 460 |
+
"rewards/rejected": 0.8737583160400391,
|
| 461 |
+
"step": 270
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 0.975609756097561,
|
| 465 |
+
"grad_norm": 647.9814126875335,
|
| 466 |
+
"learning_rate": 4.2610056145354496e-07,
|
| 467 |
+
"logits/chosen": -2.538573741912842,
|
| 468 |
+
"logits/rejected": -2.5176804065704346,
|
| 469 |
+
"logps/chosen": -65.74501037597656,
|
| 470 |
+
"logps/rejected": -70.30267333984375,
|
| 471 |
+
"loss": 0.862,
|
| 472 |
+
"rewards/accuracies": 0.3187499940395355,
|
| 473 |
+
"rewards/chosen": 2.068479299545288,
|
| 474 |
+
"rewards/margins": 1.5420581102371216,
|
| 475 |
+
"rewards/rejected": 0.5264211893081665,
|
| 476 |
+
"step": 280
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 1.0104529616724738,
|
| 480 |
+
"grad_norm": 56.46750622929306,
|
| 481 |
+
"learning_rate": 4.051729600736907e-07,
|
| 482 |
+
"logits/chosen": -2.5648231506347656,
|
| 483 |
+
"logits/rejected": -2.536726474761963,
|
| 484 |
+
"logps/chosen": -68.04057312011719,
|
| 485 |
+
"logps/rejected": -65.41937255859375,
|
| 486 |
+
"loss": 0.7918,
|
| 487 |
+
"rewards/accuracies": 0.38749998807907104,
|
| 488 |
+
"rewards/chosen": 3.9868416786193848,
|
| 489 |
+
"rewards/margins": 4.487275123596191,
|
| 490 |
+
"rewards/rejected": -0.5004340410232544,
|
| 491 |
+
"step": 290
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 1.0452961672473868,
|
| 495 |
+
"grad_norm": 20.388910811138857,
|
| 496 |
+
"learning_rate": 3.8404090617335413e-07,
|
| 497 |
+
"logits/chosen": -2.5940048694610596,
|
| 498 |
+
"logits/rejected": -2.579204559326172,
|
| 499 |
+
"logps/chosen": -59.690216064453125,
|
| 500 |
+
"logps/rejected": -76.28305053710938,
|
| 501 |
+
"loss": 0.4066,
|
| 502 |
+
"rewards/accuracies": 0.41874998807907104,
|
| 503 |
+
"rewards/chosen": 6.790108680725098,
|
| 504 |
+
"rewards/margins": 13.831901550292969,
|
| 505 |
+
"rewards/rejected": -7.0417938232421875,
|
| 506 |
+
"step": 300
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 1.0452961672473868,
|
| 510 |
+
"eval_logits/chosen": -2.612644672393799,
|
| 511 |
+
"eval_logits/rejected": -2.5961754322052,
|
| 512 |
+
"eval_logps/chosen": -72.1514663696289,
|
| 513 |
+
"eval_logps/rejected": -80.20658111572266,
|
| 514 |
+
"eval_loss": 1.0729384422302246,
|
| 515 |
+
"eval_rewards/accuracies": 0.3432539701461792,
|
| 516 |
+
"eval_rewards/chosen": 2.3163692951202393,
|
| 517 |
+
"eval_rewards/margins": 1.4038203954696655,
|
| 518 |
+
"eval_rewards/rejected": 0.9125491380691528,
|
| 519 |
+
"eval_runtime": 113.475,
|
| 520 |
+
"eval_samples_per_second": 17.625,
|
| 521 |
+
"eval_steps_per_second": 0.555,
|
| 522 |
+
"step": 300
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 1.0801393728222997,
|
| 526 |
+
"grad_norm": 3.5419086603813197,
|
| 527 |
+
"learning_rate": 3.6278270807018065e-07,
|
| 528 |
+
"logits/chosen": -2.558955430984497,
|
| 529 |
+
"logits/rejected": -2.5596249103546143,
|
| 530 |
+
"logps/chosen": -59.938392639160156,
|
| 531 |
+
"logps/rejected": -84.96260833740234,
|
| 532 |
+
"loss": 0.416,
|
| 533 |
+
"rewards/accuracies": 0.4375,
|
| 534 |
+
"rewards/chosen": 7.621278285980225,
|
| 535 |
+
"rewards/margins": 17.239816665649414,
|
| 536 |
+
"rewards/rejected": -9.618535995483398,
|
| 537 |
+
"step": 310
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 1.1149825783972125,
|
| 541 |
+
"grad_norm": 201.12315966640935,
|
| 542 |
+
"learning_rate": 3.414771415300036e-07,
|
| 543 |
+
"logits/chosen": -2.5916085243225098,
|
| 544 |
+
"logits/rejected": -2.5778439044952393,
|
| 545 |
+
"logps/chosen": -66.45283508300781,
|
| 546 |
+
"logps/rejected": -87.47081756591797,
|
| 547 |
+
"loss": 0.3907,
|
| 548 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 549 |
+
"rewards/chosen": 7.431988716125488,
|
| 550 |
+
"rewards/margins": 16.59490203857422,
|
| 551 |
+
"rewards/rejected": -9.16291332244873,
|
| 552 |
+
"step": 320
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 1.1498257839721253,
|
| 556 |
+
"grad_norm": 403.8100928464834,
|
| 557 |
+
"learning_rate": 3.2020315785022746e-07,
|
| 558 |
+
"logits/chosen": -2.5842907428741455,
|
| 559 |
+
"logits/rejected": -2.5574951171875,
|
| 560 |
+
"logps/chosen": -75.70329284667969,
|
| 561 |
+
"logps/rejected": -86.71253967285156,
|
| 562 |
+
"loss": 0.3965,
|
| 563 |
+
"rewards/accuracies": 0.518750011920929,
|
| 564 |
+
"rewards/chosen": 8.334002494812012,
|
| 565 |
+
"rewards/margins": 16.338943481445312,
|
| 566 |
+
"rewards/rejected": -8.004940032958984,
|
| 567 |
+
"step": 330
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 1.1846689895470384,
|
| 571 |
+
"grad_norm": 184.77353798917835,
|
| 572 |
+
"learning_rate": 2.9903959129274836e-07,
|
| 573 |
+
"logits/chosen": -2.561433792114258,
|
| 574 |
+
"logits/rejected": -2.5655293464660645,
|
| 575 |
+
"logps/chosen": -74.55548095703125,
|
| 576 |
+
"logps/rejected": -104.60441589355469,
|
| 577 |
+
"loss": 0.4148,
|
| 578 |
+
"rewards/accuracies": 0.53125,
|
| 579 |
+
"rewards/chosen": 9.100366592407227,
|
| 580 |
+
"rewards/margins": 18.328720092773438,
|
| 581 |
+
"rewards/rejected": -9.228353500366211,
|
| 582 |
+
"step": 340
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"epoch": 1.2195121951219512,
|
| 586 |
+
"grad_norm": 188.43635904114095,
|
| 587 |
+
"learning_rate": 2.7806486695056977e-07,
|
| 588 |
+
"logits/chosen": -2.5871026515960693,
|
| 589 |
+
"logits/rejected": -2.55499529838562,
|
| 590 |
+
"logps/chosen": -58.55460739135742,
|
| 591 |
+
"logps/rejected": -74.59812927246094,
|
| 592 |
+
"loss": 0.4111,
|
| 593 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 594 |
+
"rewards/chosen": 9.941993713378906,
|
| 595 |
+
"rewards/margins": 18.771812438964844,
|
| 596 |
+
"rewards/rejected": -8.829817771911621,
|
| 597 |
+
"step": 350
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 1.254355400696864,
|
| 601 |
+
"grad_norm": 154.26226580174057,
|
| 602 |
+
"learning_rate": 2.573567101306622e-07,
|
| 603 |
+
"logits/chosen": -2.5829975605010986,
|
| 604 |
+
"logits/rejected": -2.5526790618896484,
|
| 605 |
+
"logps/chosen": -62.894317626953125,
|
| 606 |
+
"logps/rejected": -71.01216888427734,
|
| 607 |
+
"loss": 0.4026,
|
| 608 |
+
"rewards/accuracies": 0.4437499940395355,
|
| 609 |
+
"rewards/chosen": 8.866216659545898,
|
| 610 |
+
"rewards/margins": 14.160728454589844,
|
| 611 |
+
"rewards/rejected": -5.294511795043945,
|
| 612 |
+
"step": 360
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 1.289198606271777,
|
| 616 |
+
"grad_norm": 16.651956659184666,
|
| 617 |
+
"learning_rate": 2.369918583299939e-07,
|
| 618 |
+
"logits/chosen": -2.5568833351135254,
|
| 619 |
+
"logits/rejected": -2.5733718872070312,
|
| 620 |
+
"logps/chosen": -62.53009033203125,
|
| 621 |
+
"logps/rejected": -82.81775665283203,
|
| 622 |
+
"loss": 0.4444,
|
| 623 |
+
"rewards/accuracies": 0.46875,
|
| 624 |
+
"rewards/chosen": 8.450610160827637,
|
| 625 |
+
"rewards/margins": 16.84958839416504,
|
| 626 |
+
"rewards/rejected": -8.398977279663086,
|
| 627 |
+
"step": 370
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 1.32404181184669,
|
| 631 |
+
"grad_norm": 210.98546985458898,
|
| 632 |
+
"learning_rate": 2.1704577687205507e-07,
|
| 633 |
+
"logits/chosen": -2.5712497234344482,
|
| 634 |
+
"logits/rejected": -2.567377805709839,
|
| 635 |
+
"logps/chosen": -76.60011291503906,
|
| 636 |
+
"logps/rejected": -98.51103210449219,
|
| 637 |
+
"loss": 0.4409,
|
| 638 |
+
"rewards/accuracies": 0.53125,
|
| 639 |
+
"rewards/chosen": 10.171670913696289,
|
| 640 |
+
"rewards/margins": 22.095985412597656,
|
| 641 |
+
"rewards/rejected": -11.924314498901367,
|
| 642 |
+
"step": 380
|
| 643 |
+
},
|
| 644 |
+
{
|
| 645 |
+
"epoch": 1.3588850174216027,
|
| 646 |
+
"grad_norm": 125.98956132873472,
|
| 647 |
+
"learning_rate": 1.975923792576331e-07,
|
| 648 |
+
"logits/chosen": -2.6548500061035156,
|
| 649 |
+
"logits/rejected": -2.6388492584228516,
|
| 650 |
+
"logps/chosen": -57.238861083984375,
|
| 651 |
+
"logps/rejected": -80.28170776367188,
|
| 652 |
+
"loss": 0.3853,
|
| 653 |
+
"rewards/accuracies": 0.4375,
|
| 654 |
+
"rewards/chosen": 9.658943176269531,
|
| 655 |
+
"rewards/margins": 18.589771270751953,
|
| 656 |
+
"rewards/rejected": -8.930827140808105,
|
| 657 |
+
"step": 390
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 1.3937282229965158,
|
| 661 |
+
"grad_norm": 30.8789509954886,
|
| 662 |
+
"learning_rate": 1.7870375326612014e-07,
|
| 663 |
+
"logits/chosen": -2.638012647628784,
|
| 664 |
+
"logits/rejected": -2.609297275543213,
|
| 665 |
+
"logps/chosen": -76.772705078125,
|
| 666 |
+
"logps/rejected": -106.2076644897461,
|
| 667 |
+
"loss": 0.3805,
|
| 668 |
+
"rewards/accuracies": 0.518750011920929,
|
| 669 |
+
"rewards/chosen": 9.280598640441895,
|
| 670 |
+
"rewards/margins": 17.006460189819336,
|
| 671 |
+
"rewards/rejected": -7.7258620262146,
|
| 672 |
+
"step": 400
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 1.3937282229965158,
|
| 676 |
+
"eval_logits/chosen": -2.6412570476531982,
|
| 677 |
+
"eval_logits/rejected": -2.6246657371520996,
|
| 678 |
+
"eval_logps/chosen": -71.48371124267578,
|
| 679 |
+
"eval_logps/rejected": -79.92253875732422,
|
| 680 |
+
"eval_loss": 1.1545917987823486,
|
| 681 |
+
"eval_rewards/accuracies": 0.3373015820980072,
|
| 682 |
+
"eval_rewards/chosen": 2.977447986602783,
|
| 683 |
+
"eval_rewards/margins": 1.783706545829773,
|
| 684 |
+
"eval_rewards/rejected": 1.1937415599822998,
|
| 685 |
+
"eval_runtime": 113.3185,
|
| 686 |
+
"eval_samples_per_second": 17.649,
|
| 687 |
+
"eval_steps_per_second": 0.556,
|
| 688 |
+
"step": 400
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 1.4285714285714286,
|
| 692 |
+
"grad_norm": 134.60970325056522,
|
| 693 |
+
"learning_rate": 1.604498938223354e-07,
|
| 694 |
+
"logits/chosen": -2.6223552227020264,
|
| 695 |
+
"logits/rejected": -2.611177444458008,
|
| 696 |
+
"logps/chosen": -71.12557220458984,
|
| 697 |
+
"logps/rejected": -88.45091247558594,
|
| 698 |
+
"loss": 0.6988,
|
| 699 |
+
"rewards/accuracies": 0.48124998807907104,
|
| 700 |
+
"rewards/chosen": 11.372858047485352,
|
| 701 |
+
"rewards/margins": 19.366708755493164,
|
| 702 |
+
"rewards/rejected": -7.993849754333496,
|
| 703 |
+
"step": 410
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 1.4634146341463414,
|
| 707 |
+
"grad_norm": 144.0362669522128,
|
| 708 |
+
"learning_rate": 1.4289844361876528e-07,
|
| 709 |
+
"logits/chosen": -2.665538787841797,
|
| 710 |
+
"logits/rejected": -2.665926933288574,
|
| 711 |
+
"logps/chosen": -67.73007202148438,
|
| 712 |
+
"logps/rejected": -91.9347915649414,
|
| 713 |
+
"loss": 0.394,
|
| 714 |
+
"rewards/accuracies": 0.42500001192092896,
|
| 715 |
+
"rewards/chosen": 8.27946662902832,
|
| 716 |
+
"rewards/margins": 15.272089958190918,
|
| 717 |
+
"rewards/rejected": -6.9926252365112305,
|
| 718 |
+
"step": 420
|
| 719 |
+
},
|
| 720 |
+
{
|
| 721 |
+
"epoch": 1.4982578397212545,
|
| 722 |
+
"grad_norm": 8.116844029360541,
|
| 723 |
+
"learning_rate": 1.2611444245438813e-07,
|
| 724 |
+
"logits/chosen": -2.629135847091675,
|
| 725 |
+
"logits/rejected": -2.6135053634643555,
|
| 726 |
+
"logps/chosen": -59.4090576171875,
|
| 727 |
+
"logps/rejected": -79.82469177246094,
|
| 728 |
+
"loss": 0.3924,
|
| 729 |
+
"rewards/accuracies": 0.46875,
|
| 730 |
+
"rewards/chosen": 9.160750389099121,
|
| 731 |
+
"rewards/margins": 16.42301368713379,
|
| 732 |
+
"rewards/rejected": -7.262263298034668,
|
| 733 |
+
"step": 430
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"epoch": 1.533101045296167,
|
| 737 |
+
"grad_norm": 20.390949955841545,
|
| 738 |
+
"learning_rate": 1.1016008621895228e-07,
|
| 739 |
+
"logits/chosen": -2.6179585456848145,
|
| 740 |
+
"logits/rejected": -2.6255526542663574,
|
| 741 |
+
"logps/chosen": -59.94586944580078,
|
| 742 |
+
"logps/rejected": -79.95707702636719,
|
| 743 |
+
"loss": 0.385,
|
| 744 |
+
"rewards/accuracies": 0.4312500059604645,
|
| 745 |
+
"rewards/chosen": 9.932461738586426,
|
| 746 |
+
"rewards/margins": 16.32512664794922,
|
| 747 |
+
"rewards/rejected": -6.392666339874268,
|
| 748 |
+
"step": 440
|
| 749 |
+
},
|
| 750 |
+
{
|
| 751 |
+
"epoch": 1.5679442508710801,
|
| 752 |
+
"grad_norm": 91.65774378764581,
|
| 753 |
+
"learning_rate": 9.509449641582943e-08,
|
| 754 |
+
"logits/chosen": -2.669267177581787,
|
| 755 |
+
"logits/rejected": -2.6314785480499268,
|
| 756 |
+
"logps/chosen": -78.78290557861328,
|
| 757 |
+
"logps/rejected": -96.71044158935547,
|
| 758 |
+
"loss": 0.4341,
|
| 759 |
+
"rewards/accuracies": 0.518750011920929,
|
| 760 |
+
"rewards/chosen": 11.938484191894531,
|
| 761 |
+
"rewards/margins": 22.57853889465332,
|
| 762 |
+
"rewards/rejected": -10.640054702758789,
|
| 763 |
+
"step": 450
|
| 764 |
+
},
|
| 765 |
+
{
|
| 766 |
+
"epoch": 1.6027874564459932,
|
| 767 |
+
"grad_norm": 203.85164100757706,
|
| 768 |
+
"learning_rate": 8.097350107751374e-08,
|
| 769 |
+
"logits/chosen": -2.643629550933838,
|
| 770 |
+
"logits/rejected": -2.621685743331909,
|
| 771 |
+
"logps/chosen": -65.54093933105469,
|
| 772 |
+
"logps/rejected": -86.33696746826172,
|
| 773 |
+
"loss": 0.4012,
|
| 774 |
+
"rewards/accuracies": 0.5,
|
| 775 |
+
"rewards/chosen": 10.532175064086914,
|
| 776 |
+
"rewards/margins": 19.771120071411133,
|
| 777 |
+
"rewards/rejected": -9.238944053649902,
|
| 778 |
+
"step": 460
|
| 779 |
+
},
|
| 780 |
+
{
|
| 781 |
+
"epoch": 1.6376306620209058,
|
| 782 |
+
"grad_norm": 0.18713678602063968,
|
| 783 |
+
"learning_rate": 6.784942788562304e-08,
|
| 784 |
+
"logits/chosen": -2.6330292224884033,
|
| 785 |
+
"logits/rejected": -2.6224162578582764,
|
| 786 |
+
"logps/chosen": -53.2932243347168,
|
| 787 |
+
"logps/rejected": -78.58303833007812,
|
| 788 |
+
"loss": 0.4188,
|
| 789 |
+
"rewards/accuracies": 0.4000000059604645,
|
| 790 |
+
"rewards/chosen": 7.929083347320557,
|
| 791 |
+
"rewards/margins": 14.772134780883789,
|
| 792 |
+
"rewards/rejected": -6.843050479888916,
|
| 793 |
+
"step": 470
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 1.6724738675958188,
|
| 797 |
+
"grad_norm": 0.6147060015446919,
|
| 798 |
+
"learning_rate": 5.5770910262027175e-08,
|
| 799 |
+
"logits/chosen": -2.6437172889709473,
|
| 800 |
+
"logits/rejected": -2.627516269683838,
|
| 801 |
+
"logps/chosen": -48.605804443359375,
|
| 802 |
+
"logps/rejected": -54.954551696777344,
|
| 803 |
+
"loss": 0.3885,
|
| 804 |
+
"rewards/accuracies": 0.36250001192092896,
|
| 805 |
+
"rewards/chosen": 5.401150703430176,
|
| 806 |
+
"rewards/margins": 8.344820022583008,
|
| 807 |
+
"rewards/rejected": -2.943669319152832,
|
| 808 |
+
"step": 480
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 1.7073170731707317,
|
| 812 |
+
"grad_norm": 77.39241567442491,
|
| 813 |
+
"learning_rate": 4.47827071496673e-08,
|
| 814 |
+
"logits/chosen": -2.630645275115967,
|
| 815 |
+
"logits/rejected": -2.618098735809326,
|
| 816 |
+
"logps/chosen": -65.59320831298828,
|
| 817 |
+
"logps/rejected": -76.5845718383789,
|
| 818 |
+
"loss": 0.4658,
|
| 819 |
+
"rewards/accuracies": 0.40625,
|
| 820 |
+
"rewards/chosen": 7.552302360534668,
|
| 821 |
+
"rewards/margins": 14.009483337402344,
|
| 822 |
+
"rewards/rejected": -6.457179069519043,
|
| 823 |
+
"step": 490
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 1.7421602787456445,
|
| 827 |
+
"grad_norm": 9.754129842794342,
|
| 828 |
+
"learning_rate": 3.492553715089692e-08,
|
| 829 |
+
"logits/chosen": -2.5552730560302734,
|
| 830 |
+
"logits/rejected": -2.546638011932373,
|
| 831 |
+
"logps/chosen": -67.24018096923828,
|
| 832 |
+
"logps/rejected": -92.03102111816406,
|
| 833 |
+
"loss": 0.3975,
|
| 834 |
+
"rewards/accuracies": 0.46875,
|
| 835 |
+
"rewards/chosen": 8.957259178161621,
|
| 836 |
+
"rewards/margins": 17.181909561157227,
|
| 837 |
+
"rewards/rejected": -8.224650382995605,
|
| 838 |
+
"step": 500
|
| 839 |
+
},
|
| 840 |
+
{
|
| 841 |
+
"epoch": 1.7421602787456445,
|
| 842 |
+
"eval_logits/chosen": -2.663222551345825,
|
| 843 |
+
"eval_logits/rejected": -2.6463263034820557,
|
| 844 |
+
"eval_logps/chosen": -71.91925048828125,
|
| 845 |
+
"eval_logps/rejected": -80.58873748779297,
|
| 846 |
+
"eval_loss": 1.182449460029602,
|
| 847 |
+
"eval_rewards/accuracies": 0.3452380895614624,
|
| 848 |
+
"eval_rewards/chosen": 2.54626727104187,
|
| 849 |
+
"eval_rewards/margins": 2.0120491981506348,
|
| 850 |
+
"eval_rewards/rejected": 0.5342182517051697,
|
| 851 |
+
"eval_runtime": 113.333,
|
| 852 |
+
"eval_samples_per_second": 17.647,
|
| 853 |
+
"eval_steps_per_second": 0.556,
|
| 854 |
+
"step": 500
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 1.7770034843205575,
|
| 858 |
+
"grad_norm": 9.598053942856874,
|
| 859 |
+
"learning_rate": 2.6235927637971816e-08,
|
| 860 |
+
"logits/chosen": -2.625732898712158,
|
| 861 |
+
"logits/rejected": -2.6054017543792725,
|
| 862 |
+
"logps/chosen": -61.70050048828125,
|
| 863 |
+
"logps/rejected": -77.1646728515625,
|
| 864 |
+
"loss": 0.3933,
|
| 865 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 866 |
+
"rewards/chosen": 8.965360641479492,
|
| 867 |
+
"rewards/margins": 16.470144271850586,
|
| 868 |
+
"rewards/rejected": -7.504785060882568,
|
| 869 |
+
"step": 510
|
| 870 |
+
},
|
| 871 |
+
{
|
| 872 |
+
"epoch": 1.8118466898954704,
|
| 873 |
+
"grad_norm": 19.555147588045187,
|
| 874 |
+
"learning_rate": 1.8746079394836706e-08,
|
| 875 |
+
"logits/chosen": -2.636565685272217,
|
| 876 |
+
"logits/rejected": -2.6276707649230957,
|
| 877 |
+
"logps/chosen": -65.6178207397461,
|
| 878 |
+
"logps/rejected": -84.69874572753906,
|
| 879 |
+
"loss": 0.4072,
|
| 880 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 881 |
+
"rewards/chosen": 8.566861152648926,
|
| 882 |
+
"rewards/margins": 15.977938652038574,
|
| 883 |
+
"rewards/rejected": -7.411079406738281,
|
| 884 |
+
"step": 520
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 1.8466898954703832,
|
| 888 |
+
"grad_norm": 45.12187365083394,
|
| 889 |
+
"learning_rate": 1.2483747291799507e-08,
|
| 890 |
+
"logits/chosen": -2.610792636871338,
|
| 891 |
+
"logits/rejected": -2.603881597518921,
|
| 892 |
+
"logps/chosen": -62.07392501831055,
|
| 893 |
+
"logps/rejected": -80.65681457519531,
|
| 894 |
+
"loss": 0.3989,
|
| 895 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 896 |
+
"rewards/chosen": 9.169586181640625,
|
| 897 |
+
"rewards/margins": 18.3698787689209,
|
| 898 |
+
"rewards/rejected": -9.200292587280273,
|
| 899 |
+
"step": 530
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 1.8815331010452963,
|
| 903 |
+
"grad_norm": 828.0040833812193,
|
| 904 |
+
"learning_rate": 7.472137435272619e-09,
|
| 905 |
+
"logits/chosen": -2.624800205230713,
|
| 906 |
+
"logits/rejected": -2.638892889022827,
|
| 907 |
+
"logps/chosen": -54.112327575683594,
|
| 908 |
+
"logps/rejected": -85.64537048339844,
|
| 909 |
+
"loss": 0.4706,
|
| 910 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 911 |
+
"rewards/chosen": 9.522268295288086,
|
| 912 |
+
"rewards/margins": 21.70328712463379,
|
| 913 |
+
"rewards/rejected": -12.181015968322754,
|
| 914 |
+
"step": 540
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 1.916376306620209,
|
| 918 |
+
"grad_norm": 6.099605163051641,
|
| 919 |
+
"learning_rate": 3.729821173711411e-09,
|
| 920 |
+
"logits/chosen": -2.6054625511169434,
|
| 921 |
+
"logits/rejected": -2.585512161254883,
|
| 922 |
+
"logps/chosen": -78.79656982421875,
|
| 923 |
+
"logps/rejected": -100.56778717041016,
|
| 924 |
+
"loss": 0.3741,
|
| 925 |
+
"rewards/accuracies": 0.5,
|
| 926 |
+
"rewards/chosen": 11.91783332824707,
|
| 927 |
+
"rewards/margins": 25.462675094604492,
|
| 928 |
+
"rewards/rejected": -13.544839859008789,
|
| 929 |
+
"step": 550
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 1.951219512195122,
|
| 933 |
+
"grad_norm": 239.71729089566762,
|
| 934 |
+
"learning_rate": 1.2706662784136513e-09,
|
| 935 |
+
"logits/chosen": -2.601435422897339,
|
| 936 |
+
"logits/rejected": -2.609788417816162,
|
| 937 |
+
"logps/chosen": -56.35918045043945,
|
| 938 |
+
"logps/rejected": -76.89906311035156,
|
| 939 |
+
"loss": 0.4008,
|
| 940 |
+
"rewards/accuracies": 0.4437499940395355,
|
| 941 |
+
"rewards/chosen": 7.287221431732178,
|
| 942 |
+
"rewards/margins": 13.824111938476562,
|
| 943 |
+
"rewards/rejected": -6.536891937255859,
|
| 944 |
+
"step": 560
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 1.986062717770035,
|
| 948 |
+
"grad_norm": 23.401979133014883,
|
| 949 |
+
"learning_rate": 1.0378555420122448e-10,
|
| 950 |
+
"logits/chosen": -2.6814956665039062,
|
| 951 |
+
"logits/rejected": -2.6513266563415527,
|
| 952 |
+
"logps/chosen": -57.64264678955078,
|
| 953 |
+
"logps/rejected": -68.03236389160156,
|
| 954 |
+
"loss": 0.3939,
|
| 955 |
+
"rewards/accuracies": 0.4375,
|
| 956 |
+
"rewards/chosen": 7.494875907897949,
|
| 957 |
+
"rewards/margins": 12.909930229187012,
|
| 958 |
+
"rewards/rejected": -5.4150543212890625,
|
| 959 |
+
"step": 570
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 2.0,
|
| 963 |
+
"step": 574,
|
| 964 |
+
"total_flos": 0.0,
|
| 965 |
+
"train_loss": 0.6661292189920406,
|
| 966 |
+
"train_runtime": 6507.4651,
|
| 967 |
+
"train_samples_per_second": 5.637,
|
| 968 |
+
"train_steps_per_second": 0.088
|
| 969 |
+
}
|
| 970 |
+
],
|
| 971 |
+
"logging_steps": 10,
|
| 972 |
+
"max_steps": 574,
|
| 973 |
+
"num_input_tokens_seen": 0,
|
| 974 |
+
"num_train_epochs": 2,
|
| 975 |
+
"save_steps": 100,
|
| 976 |
+
"stateful_callbacks": {
|
| 977 |
+
"TrainerControl": {
|
| 978 |
+
"args": {
|
| 979 |
+
"should_epoch_stop": false,
|
| 980 |
+
"should_evaluate": false,
|
| 981 |
+
"should_log": false,
|
| 982 |
+
"should_save": true,
|
| 983 |
+
"should_training_stop": true
|
| 984 |
+
},
|
| 985 |
+
"attributes": {}
|
| 986 |
+
}
|
| 987 |
+
},
|
| 988 |
+
"total_flos": 0.0,
|
| 989 |
+
"train_batch_size": 8,
|
| 990 |
+
"trial_name": null,
|
| 991 |
+
"trial_params": null
|
| 992 |
+
}
|