Model save
Browse files- README.md +90 -0
- all_results.json +9 -0
- generation_config.json +7 -0
- model.safetensors +1 -1
- train_results.json +9 -0
- trainer_state.json +595 -0
README.md
ADDED
|
@@ -0,0 +1,90 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
base_model: hZzy/qwen2.5-0.5b-sft-news-IFT
|
| 4 |
+
tags:
|
| 5 |
+
- trl
|
| 6 |
+
- expo
|
| 7 |
+
- generated_from_trainer
|
| 8 |
+
model-index:
|
| 9 |
+
- name: qwen2.5-0.5b-expo-L2EXPO-ES-0.001
|
| 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 |
+
[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="200" height="32"/>](https://wandb.ai/zhiyuzha-university-of-florida/huggingface/runs/1mxe21by)
|
| 17 |
+
# qwen2.5-0.5b-expo-L2EXPO-ES-0.001
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [hZzy/qwen2.5-0.5b-sft-news-IFT](https://huggingface.co/hZzy/qwen2.5-0.5b-sft-news-IFT) on an unknown dataset.
|
| 20 |
+
It achieves the following results on the evaluation set:
|
| 21 |
+
- Loss: 0.3955
|
| 22 |
+
- Logps: -657.3067
|
| 23 |
+
- Logits: -11.2559
|
| 24 |
+
- Objective: 0.3958
|
| 25 |
+
- Dpo Loss: 0.6701
|
| 26 |
+
- Regularize: 0.3958
|
| 27 |
+
- Ranking Simple: 0.6009
|
| 28 |
+
- Ranking Idealized: 0.9871
|
| 29 |
+
- Ranking Idealized Expo: 0.6320
|
| 30 |
+
- Wo Beta: 190.5371
|
| 31 |
+
|
| 32 |
+
## Model description
|
| 33 |
+
|
| 34 |
+
More information needed
|
| 35 |
+
|
| 36 |
+
## Intended uses & limitations
|
| 37 |
+
|
| 38 |
+
More information needed
|
| 39 |
+
|
| 40 |
+
## Training and evaluation data
|
| 41 |
+
|
| 42 |
+
More information needed
|
| 43 |
+
|
| 44 |
+
## Training procedure
|
| 45 |
+
|
| 46 |
+
### Training hyperparameters
|
| 47 |
+
|
| 48 |
+
The following hyperparameters were used during training:
|
| 49 |
+
- learning_rate: 5e-06
|
| 50 |
+
- train_batch_size: 4
|
| 51 |
+
- eval_batch_size: 4
|
| 52 |
+
- seed: 42
|
| 53 |
+
- distributed_type: multi-GPU
|
| 54 |
+
- num_devices: 3
|
| 55 |
+
- gradient_accumulation_steps: 12
|
| 56 |
+
- total_train_batch_size: 144
|
| 57 |
+
- total_eval_batch_size: 12
|
| 58 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 59 |
+
- lr_scheduler_type: cosine
|
| 60 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 61 |
+
- num_epochs: 5
|
| 62 |
+
|
| 63 |
+
### Training results
|
| 64 |
+
|
| 65 |
+
| Training Loss | Epoch | Step | Dpo Loss | Logits | Logps | Validation Loss | Objective | Ranking Idealized | Ranking Idealized Expo | Ranking Simple | Regularize | Wo Beta |
|
| 66 |
+
|:-------------:|:------:|:----:|:--------:|:--------:|:---------:|:---------------:|:---------:|:-----------------:|:----------------------:|:--------------:|:----------:|:--------:|
|
| 67 |
+
| 0.418 | 0.1417 | 50 | 0.6927 | -1.8105 | -107.4392 | 0.4150 | 0.4128 | 0.9871 | 0.6320 | 0.5352 | 0.4128 | 22.0269 |
|
| 68 |
+
| 0.416 | 0.2834 | 100 | 0.6896 | -2.0485 | -230.8855 | 0.4087 | 0.4081 | 0.9871 | 0.6320 | 0.5559 | 0.4081 | 52.9494 |
|
| 69 |
+
| 0.387 | 0.4251 | 150 | 0.6844 | -3.9840 | -343.5519 | 0.4032 | 0.4021 | 0.9871 | 0.6320 | 0.5766 | 0.4021 | 90.2215 |
|
| 70 |
+
| 0.3587 | 0.5668 | 200 | 0.6754 | -6.1681 | -390.3867 | 0.3917 | 0.3893 | 0.9871 | 0.6320 | 0.6004 | 0.3893 | 124.6577 |
|
| 71 |
+
| 0.3299 | 0.7085 | 250 | 0.6765 | -7.7444 | -474.0688 | 0.3968 | 0.3968 | 0.9871 | 0.6320 | 0.5958 | 0.3968 | 147.7626 |
|
| 72 |
+
| 0.294 | 0.8503 | 300 | 0.6728 | -8.8910 | -573.5075 | 0.3942 | 0.3931 | 0.9871 | 0.6320 | 0.6102 | 0.3931 | 160.3578 |
|
| 73 |
+
| 0.2753 | 0.9920 | 350 | 0.6731 | -9.9981 | -593.1101 | 0.3965 | 0.3960 | 0.9871 | 0.6320 | 0.5937 | 0.3960 | 171.5761 |
|
| 74 |
+
| 0.2316 | 1.1337 | 400 | 0.6718 | -9.6479 | -564.7661 | 0.3966 | 0.3956 | 0.9871 | 0.6320 | 0.5875 | 0.3956 | 171.6054 |
|
| 75 |
+
| 0.2205 | 1.2754 | 450 | 0.6725 | -10.9673 | -599.2516 | 0.3962 | 0.3983 | 0.9871 | 0.6320 | 0.5859 | 0.3983 | 182.4877 |
|
| 76 |
+
| 0.2058 | 1.4171 | 500 | 0.6741 | -9.6175 | -589.5045 | 0.4005 | 0.4029 | 0.9871 | 0.6320 | 0.5797 | 0.4029 | 188.1013 |
|
| 77 |
+
| 0.2027 | 1.5588 | 550 | 0.6730 | -10.3937 | -622.4691 | 0.3995 | 0.4000 | 0.9871 | 0.6320 | 0.5947 | 0.4000 | 185.8620 |
|
| 78 |
+
| 0.1897 | 1.7029 | 600 | 0.4028 | -755.1119| -11.5540 | 0.4023 | 0.6716 | 0.4023 | 0.5952 | 0.9871 | 0.6320 | 201.2357 |
|
| 79 |
+
| 0.1797 | 1.8446 | 650 | 0.3997 | -673.7770| -10.8193 | 0.3992 | 0.6730 | 0.3992 | 0.5942 | 0.9871 | 0.6320 | 188.3079 |
|
| 80 |
+
| 0.1689 | 1.9863 | 700 | 0.3985 | -653.8336| -11.0772 | 0.3970 | 0.6713 | 0.3970 | 0.5911 | 0.9871 | 0.6320 | 182.3852 |
|
| 81 |
+
| 0.1492 | 2.1280 | 750 | 0.3959 | -624.3672| -11.4717 | 0.3956 | 0.6708 | 0.3956 | 0.6025 | 0.9871 | 0.6320 | 182.7602 |
|
| 82 |
+
| 0.143 | 2.2697 | 800 | 0.3955 | -657.3067| -11.2559 | 0.3958 | 0.6701 | 0.3958 | 0.6009 | 0.9871 | 0.6320 | 190.5371 |
|
| 83 |
+
|
| 84 |
+
|
| 85 |
+
### Framework versions
|
| 86 |
+
|
| 87 |
+
- Transformers 4.42.0
|
| 88 |
+
- Pytorch 2.3.0+cu121
|
| 89 |
+
- Datasets 3.2.0
|
| 90 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.269721303731696,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.051899830102920534,
|
| 5 |
+
"train_runtime": 6898.2847,
|
| 6 |
+
"train_samples": 50802,
|
| 7 |
+
"train_samples_per_second": 36.822,
|
| 8 |
+
"train_steps_per_second": 0.255
|
| 9 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token_id": 151644,
|
| 3 |
+
"eos_token_id": 151645,
|
| 4 |
+
"max_new_tokens": 2048,
|
| 5 |
+
"pad_token_id": 151645,
|
| 6 |
+
"transformers_version": "4.42.0"
|
| 7 |
+
}
|
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1975192208
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a7d26852a019ce9adaca2772a35f45bc7a6d56eda126534dd9a5ddb1f6fed6a4
|
| 3 |
size 1975192208
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.269721303731696,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.051899830102920534,
|
| 5 |
+
"train_runtime": 6898.2847,
|
| 6 |
+
"train_samples": 50802,
|
| 7 |
+
"train_samples_per_second": 36.822,
|
| 8 |
+
"train_steps_per_second": 0.255
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,595 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.6102484464645386,
|
| 3 |
+
"best_model_checkpoint": "./qwen2.5-0.5b/qwen2.5-0.5b-expo-L2EXPO-ES-0.001/checkpoint-300",
|
| 4 |
+
"epoch": 2.269721303731696,
|
| 5 |
+
"eval_steps": 50,
|
| 6 |
+
"global_step": 800,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"dpo_loss": 0.6931471824645996,
|
| 13 |
+
"epoch": 0.002834199338686821,
|
| 14 |
+
"grad_norm": 0.3688501423173312,
|
| 15 |
+
"learning_rate": 2.840909090909091e-08,
|
| 16 |
+
"logits": -1.359458565711975,
|
| 17 |
+
"logps": -84.69721221923828,
|
| 18 |
+
"loss": 0.3913,
|
| 19 |
+
"objective": 0.3618059456348419,
|
| 20 |
+
"ranking_idealized": 1.0,
|
| 21 |
+
"ranking_idealized_expo": 0.6875,
|
| 22 |
+
"ranking_simple": 0.5833333134651184,
|
| 23 |
+
"regularize": 0.3618059456348419,
|
| 24 |
+
"step": 1,
|
| 25 |
+
"wo_beta": 14.830931663513184
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"dpo_loss": 0.6929696202278137,
|
| 29 |
+
"epoch": 0.14170996693434104,
|
| 30 |
+
"grad_norm": 0.6973089072493135,
|
| 31 |
+
"learning_rate": 1.4204545454545458e-06,
|
| 32 |
+
"logits": -1.5327931642532349,
|
| 33 |
+
"logps": -83.60688018798828,
|
| 34 |
+
"loss": 0.418,
|
| 35 |
+
"objective": 0.42159923911094666,
|
| 36 |
+
"ranking_idealized": 0.9876700639724731,
|
| 37 |
+
"ranking_idealized_expo": 0.6386054158210754,
|
| 38 |
+
"ranking_simple": 0.5280612111091614,
|
| 39 |
+
"regularize": 0.42159923911094666,
|
| 40 |
+
"step": 50,
|
| 41 |
+
"wo_beta": 16.323150634765625
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 0.14170996693434104,
|
| 45 |
+
"eval_dpo_loss": 0.6926723122596741,
|
| 46 |
+
"eval_logits": -1.810510516166687,
|
| 47 |
+
"eval_logps": -107.43917083740234,
|
| 48 |
+
"eval_loss": 0.41496098041534424,
|
| 49 |
+
"eval_objective": 0.4128378927707672,
|
| 50 |
+
"eval_ranking_idealized": 0.9870600700378418,
|
| 51 |
+
"eval_ranking_idealized_expo": 0.6319875717163086,
|
| 52 |
+
"eval_ranking_simple": 0.5351966619491577,
|
| 53 |
+
"eval_regularize": 0.4128378927707672,
|
| 54 |
+
"eval_runtime": 308.8567,
|
| 55 |
+
"eval_samples_per_second": 18.747,
|
| 56 |
+
"eval_steps_per_second": 1.564,
|
| 57 |
+
"eval_wo_beta": 22.02687644958496,
|
| 58 |
+
"step": 50
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"dpo_loss": 0.6878826022148132,
|
| 62 |
+
"epoch": 0.2834199338686821,
|
| 63 |
+
"grad_norm": 2.816808446897818,
|
| 64 |
+
"learning_rate": 2.8409090909090916e-06,
|
| 65 |
+
"logits": -1.554953932762146,
|
| 66 |
+
"logps": -162.0962677001953,
|
| 67 |
+
"loss": 0.416,
|
| 68 |
+
"objective": 0.4231468141078949,
|
| 69 |
+
"ranking_idealized": 0.987500011920929,
|
| 70 |
+
"ranking_idealized_expo": 0.6287500262260437,
|
| 71 |
+
"ranking_simple": 0.597083330154419,
|
| 72 |
+
"regularize": 0.4231468141078949,
|
| 73 |
+
"step": 100,
|
| 74 |
+
"wo_beta": 36.895965576171875
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.2834199338686821,
|
| 78 |
+
"eval_dpo_loss": 0.6895796060562134,
|
| 79 |
+
"eval_logits": -2.0485212802886963,
|
| 80 |
+
"eval_logps": -230.8855438232422,
|
| 81 |
+
"eval_loss": 0.4087027907371521,
|
| 82 |
+
"eval_objective": 0.40811070799827576,
|
| 83 |
+
"eval_ranking_idealized": 0.9870600700378418,
|
| 84 |
+
"eval_ranking_idealized_expo": 0.6319875717163086,
|
| 85 |
+
"eval_ranking_simple": 0.5559006333351135,
|
| 86 |
+
"eval_regularize": 0.40811070799827576,
|
| 87 |
+
"eval_runtime": 308.2409,
|
| 88 |
+
"eval_samples_per_second": 18.784,
|
| 89 |
+
"eval_steps_per_second": 1.567,
|
| 90 |
+
"eval_wo_beta": 52.94944381713867,
|
| 91 |
+
"step": 100
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"dpo_loss": 0.6755279302597046,
|
| 95 |
+
"epoch": 0.42512990080302315,
|
| 96 |
+
"grad_norm": 4.687417348983457,
|
| 97 |
+
"learning_rate": 4.2613636363636365e-06,
|
| 98 |
+
"logits": -2.6038007736206055,
|
| 99 |
+
"logps": -284.96868896484375,
|
| 100 |
+
"loss": 0.387,
|
| 101 |
+
"objective": 0.39124202728271484,
|
| 102 |
+
"ranking_idealized": 0.9887499809265137,
|
| 103 |
+
"ranking_idealized_expo": 0.6299999952316284,
|
| 104 |
+
"ranking_simple": 0.6474999785423279,
|
| 105 |
+
"regularize": 0.39124202728271484,
|
| 106 |
+
"step": 150,
|
| 107 |
+
"wo_beta": 83.33438110351562
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.42512990080302315,
|
| 111 |
+
"eval_dpo_loss": 0.6843881607055664,
|
| 112 |
+
"eval_logits": -3.9840264320373535,
|
| 113 |
+
"eval_logps": -343.55194091796875,
|
| 114 |
+
"eval_loss": 0.40323564410209656,
|
| 115 |
+
"eval_objective": 0.40214404463768005,
|
| 116 |
+
"eval_ranking_idealized": 0.9870600700378418,
|
| 117 |
+
"eval_ranking_idealized_expo": 0.6319875717163086,
|
| 118 |
+
"eval_ranking_simple": 0.5766045451164246,
|
| 119 |
+
"eval_regularize": 0.40214404463768005,
|
| 120 |
+
"eval_runtime": 308.6069,
|
| 121 |
+
"eval_samples_per_second": 18.762,
|
| 122 |
+
"eval_steps_per_second": 1.565,
|
| 123 |
+
"eval_wo_beta": 90.2214584350586,
|
| 124 |
+
"step": 150
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"dpo_loss": 0.6573903560638428,
|
| 128 |
+
"epoch": 0.5668398677373642,
|
| 129 |
+
"grad_norm": 5.4307407287631735,
|
| 130 |
+
"learning_rate": 4.997168347957521e-06,
|
| 131 |
+
"logits": -5.123294353485107,
|
| 132 |
+
"logps": -384.2275390625,
|
| 133 |
+
"loss": 0.3587,
|
| 134 |
+
"objective": 0.36179742217063904,
|
| 135 |
+
"ranking_idealized": 0.9891666769981384,
|
| 136 |
+
"ranking_idealized_expo": 0.6195833086967468,
|
| 137 |
+
"ranking_simple": 0.684166669845581,
|
| 138 |
+
"regularize": 0.36179742217063904,
|
| 139 |
+
"step": 200,
|
| 140 |
+
"wo_beta": 136.6976318359375
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.5668398677373642,
|
| 144 |
+
"eval_dpo_loss": 0.6754290461540222,
|
| 145 |
+
"eval_logits": -6.168106555938721,
|
| 146 |
+
"eval_logps": -390.38665771484375,
|
| 147 |
+
"eval_loss": 0.3916535973548889,
|
| 148 |
+
"eval_objective": 0.3892715573310852,
|
| 149 |
+
"eval_ranking_idealized": 0.9870600700378418,
|
| 150 |
+
"eval_ranking_idealized_expo": 0.6319875717163086,
|
| 151 |
+
"eval_ranking_simple": 0.6004140973091125,
|
| 152 |
+
"eval_regularize": 0.3892715573310852,
|
| 153 |
+
"eval_runtime": 309.2072,
|
| 154 |
+
"eval_samples_per_second": 18.725,
|
| 155 |
+
"eval_steps_per_second": 1.562,
|
| 156 |
+
"eval_wo_beta": 124.65765380859375,
|
| 157 |
+
"step": 200
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"dpo_loss": 0.6350498199462891,
|
| 161 |
+
"epoch": 0.7085498346717053,
|
| 162 |
+
"grad_norm": 6.529681856006604,
|
| 163 |
+
"learning_rate": 4.973122855144066e-06,
|
| 164 |
+
"logits": -6.972330093383789,
|
| 165 |
+
"logps": -475.43475341796875,
|
| 166 |
+
"loss": 0.3299,
|
| 167 |
+
"objective": 0.3326449394226074,
|
| 168 |
+
"ranking_idealized": 0.9837499856948853,
|
| 169 |
+
"ranking_idealized_expo": 0.6225000023841858,
|
| 170 |
+
"ranking_simple": 0.7291666865348816,
|
| 171 |
+
"regularize": 0.3326449394226074,
|
| 172 |
+
"step": 250,
|
| 173 |
+
"wo_beta": 194.3697967529297
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 0.7085498346717053,
|
| 177 |
+
"eval_dpo_loss": 0.6764808893203735,
|
| 178 |
+
"eval_logits": -7.7444071769714355,
|
| 179 |
+
"eval_logps": -474.0688171386719,
|
| 180 |
+
"eval_loss": 0.39678695797920227,
|
| 181 |
+
"eval_objective": 0.3968295454978943,
|
| 182 |
+
"eval_ranking_idealized": 0.9870600700378418,
|
| 183 |
+
"eval_ranking_idealized_expo": 0.6319875717163086,
|
| 184 |
+
"eval_ranking_simple": 0.5957556962966919,
|
| 185 |
+
"eval_regularize": 0.3968295454978943,
|
| 186 |
+
"eval_runtime": 308.4253,
|
| 187 |
+
"eval_samples_per_second": 18.773,
|
| 188 |
+
"eval_steps_per_second": 1.566,
|
| 189 |
+
"eval_wo_beta": 147.76255798339844,
|
| 190 |
+
"step": 250
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"dpo_loss": 0.6194267868995667,
|
| 194 |
+
"epoch": 0.8502598016060463,
|
| 195 |
+
"grad_norm": 7.819775688472429,
|
| 196 |
+
"learning_rate": 4.924776641419513e-06,
|
| 197 |
+
"logits": -8.305917739868164,
|
| 198 |
+
"logps": -537.9514770507812,
|
| 199 |
+
"loss": 0.294,
|
| 200 |
+
"objective": 0.29155203700065613,
|
| 201 |
+
"ranking_idealized": 0.9829166531562805,
|
| 202 |
+
"ranking_idealized_expo": 0.6075000166893005,
|
| 203 |
+
"ranking_simple": 0.7541666626930237,
|
| 204 |
+
"regularize": 0.29155203700065613,
|
| 205 |
+
"step": 300,
|
| 206 |
+
"wo_beta": 228.0120849609375
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.8502598016060463,
|
| 210 |
+
"eval_dpo_loss": 0.6728345155715942,
|
| 211 |
+
"eval_logits": -8.890972137451172,
|
| 212 |
+
"eval_logps": -573.5075073242188,
|
| 213 |
+
"eval_loss": 0.3941618800163269,
|
| 214 |
+
"eval_objective": 0.3930768370628357,
|
| 215 |
+
"eval_ranking_idealized": 0.9870600700378418,
|
| 216 |
+
"eval_ranking_idealized_expo": 0.6319875717163086,
|
| 217 |
+
"eval_ranking_simple": 0.6102484464645386,
|
| 218 |
+
"eval_regularize": 0.3930768370628357,
|
| 219 |
+
"eval_runtime": 308.0714,
|
| 220 |
+
"eval_samples_per_second": 18.794,
|
| 221 |
+
"eval_steps_per_second": 1.568,
|
| 222 |
+
"eval_wo_beta": 160.35784912109375,
|
| 223 |
+
"step": 300
|
| 224 |
+
},
|
| 225 |
+
{
|
| 226 |
+
"dpo_loss": 0.6043005585670471,
|
| 227 |
+
"epoch": 0.9919697685403873,
|
| 228 |
+
"grad_norm": 8.598806793267013,
|
| 229 |
+
"learning_rate": 4.8526047530778175e-06,
|
| 230 |
+
"logits": -8.97115421295166,
|
| 231 |
+
"logps": -561.0111083984375,
|
| 232 |
+
"loss": 0.2753,
|
| 233 |
+
"objective": 0.2713463306427002,
|
| 234 |
+
"ranking_idealized": 0.9850000143051147,
|
| 235 |
+
"ranking_idealized_expo": 0.6341666579246521,
|
| 236 |
+
"ranking_simple": 0.7933333516120911,
|
| 237 |
+
"regularize": 0.2713463306427002,
|
| 238 |
+
"step": 350,
|
| 239 |
+
"wo_beta": 253.40798950195312
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 0.9919697685403873,
|
| 243 |
+
"eval_dpo_loss": 0.6731387972831726,
|
| 244 |
+
"eval_logits": -9.998077392578125,
|
| 245 |
+
"eval_logps": -593.110107421875,
|
| 246 |
+
"eval_loss": 0.3965202271938324,
|
| 247 |
+
"eval_objective": 0.3959551751613617,
|
| 248 |
+
"eval_ranking_idealized": 0.9870600700378418,
|
| 249 |
+
"eval_ranking_idealized_expo": 0.6319875717163086,
|
| 250 |
+
"eval_ranking_simple": 0.5936853289604187,
|
| 251 |
+
"eval_regularize": 0.3959551751613617,
|
| 252 |
+
"eval_runtime": 308.2686,
|
| 253 |
+
"eval_samples_per_second": 18.782,
|
| 254 |
+
"eval_steps_per_second": 1.567,
|
| 255 |
+
"eval_wo_beta": 171.5761260986328,
|
| 256 |
+
"step": 350
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"dpo_loss": 0.5814460515975952,
|
| 260 |
+
"epoch": 1.1336797354747283,
|
| 261 |
+
"grad_norm": 7.338503940970108,
|
| 262 |
+
"learning_rate": 4.757316345716554e-06,
|
| 263 |
+
"logits": -9.615409851074219,
|
| 264 |
+
"logps": -592.6610717773438,
|
| 265 |
+
"loss": 0.2316,
|
| 266 |
+
"objective": 0.23074908554553986,
|
| 267 |
+
"ranking_idealized": 0.9850000143051147,
|
| 268 |
+
"ranking_idealized_expo": 0.6329166889190674,
|
| 269 |
+
"ranking_simple": 0.8429166674613953,
|
| 270 |
+
"regularize": 0.23074908554553986,
|
| 271 |
+
"step": 400,
|
| 272 |
+
"wo_beta": 295.14337158203125
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 1.1336797354747283,
|
| 276 |
+
"eval_dpo_loss": 0.6717904210090637,
|
| 277 |
+
"eval_logits": -9.647905349731445,
|
| 278 |
+
"eval_logps": -564.7660522460938,
|
| 279 |
+
"eval_loss": 0.3965735137462616,
|
| 280 |
+
"eval_objective": 0.3956299126148224,
|
| 281 |
+
"eval_ranking_idealized": 0.9870600700378418,
|
| 282 |
+
"eval_ranking_idealized_expo": 0.6319875717163086,
|
| 283 |
+
"eval_ranking_simple": 0.5874741077423096,
|
| 284 |
+
"eval_regularize": 0.3956299126148224,
|
| 285 |
+
"eval_runtime": 308.3873,
|
| 286 |
+
"eval_samples_per_second": 18.775,
|
| 287 |
+
"eval_steps_per_second": 1.566,
|
| 288 |
+
"eval_wo_beta": 171.6053924560547,
|
| 289 |
+
"step": 400
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"dpo_loss": 0.57535320520401,
|
| 293 |
+
"epoch": 1.2753897024090695,
|
| 294 |
+
"grad_norm": 7.067175428131206,
|
| 295 |
+
"learning_rate": 4.639847716126855e-06,
|
| 296 |
+
"logits": -10.082599639892578,
|
| 297 |
+
"logps": -593.1123046875,
|
| 298 |
+
"loss": 0.2205,
|
| 299 |
+
"objective": 0.21684379875659943,
|
| 300 |
+
"ranking_idealized": 0.9895833134651184,
|
| 301 |
+
"ranking_idealized_expo": 0.6204166412353516,
|
| 302 |
+
"ranking_simple": 0.8416666388511658,
|
| 303 |
+
"regularize": 0.21684379875659943,
|
| 304 |
+
"step": 450,
|
| 305 |
+
"wo_beta": 313.67462158203125
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 1.2753897024090695,
|
| 309 |
+
"eval_dpo_loss": 0.6725399494171143,
|
| 310 |
+
"eval_logits": -10.967289924621582,
|
| 311 |
+
"eval_logps": -599.2516479492188,
|
| 312 |
+
"eval_loss": 0.3962310552597046,
|
| 313 |
+
"eval_objective": 0.39826443791389465,
|
| 314 |
+
"eval_ranking_idealized": 0.9870600700378418,
|
| 315 |
+
"eval_ranking_idealized_expo": 0.6319875717163086,
|
| 316 |
+
"eval_ranking_simple": 0.5859213471412659,
|
| 317 |
+
"eval_regularize": 0.39826443791389465,
|
| 318 |
+
"eval_runtime": 308.0799,
|
| 319 |
+
"eval_samples_per_second": 18.794,
|
| 320 |
+
"eval_steps_per_second": 1.568,
|
| 321 |
+
"eval_wo_beta": 182.48765563964844,
|
| 322 |
+
"step": 450
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"dpo_loss": 0.5704715251922607,
|
| 326 |
+
"epoch": 1.4170996693434104,
|
| 327 |
+
"grad_norm": 8.491423005159572,
|
| 328 |
+
"learning_rate": 4.501353102310901e-06,
|
| 329 |
+
"logits": -10.000563621520996,
|
| 330 |
+
"logps": -594.431396484375,
|
| 331 |
+
"loss": 0.2058,
|
| 332 |
+
"objective": 0.2110779583454132,
|
| 333 |
+
"ranking_idealized": 0.9854166507720947,
|
| 334 |
+
"ranking_idealized_expo": 0.6070833206176758,
|
| 335 |
+
"ranking_simple": 0.8395833373069763,
|
| 336 |
+
"regularize": 0.2110779583454132,
|
| 337 |
+
"step": 500,
|
| 338 |
+
"wo_beta": 325.75177001953125
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 1.4170996693434104,
|
| 342 |
+
"eval_dpo_loss": 0.6740813851356506,
|
| 343 |
+
"eval_logits": -9.617457389831543,
|
| 344 |
+
"eval_logps": -589.5045166015625,
|
| 345 |
+
"eval_loss": 0.4004902243614197,
|
| 346 |
+
"eval_objective": 0.402893602848053,
|
| 347 |
+
"eval_ranking_idealized": 0.9870600700378418,
|
| 348 |
+
"eval_ranking_idealized_expo": 0.6319875717163086,
|
| 349 |
+
"eval_ranking_simple": 0.5797101259231567,
|
| 350 |
+
"eval_regularize": 0.402893602848053,
|
| 351 |
+
"eval_runtime": 307.827,
|
| 352 |
+
"eval_samples_per_second": 18.809,
|
| 353 |
+
"eval_steps_per_second": 1.569,
|
| 354 |
+
"eval_wo_beta": 188.1012725830078,
|
| 355 |
+
"step": 500
|
| 356 |
+
},
|
| 357 |
+
{
|
| 358 |
+
"dpo_loss": 0.5671207904815674,
|
| 359 |
+
"epoch": 1.5588096362777515,
|
| 360 |
+
"grad_norm": 8.449525787360594,
|
| 361 |
+
"learning_rate": 4.34319334202531e-06,
|
| 362 |
+
"logits": -9.910933494567871,
|
| 363 |
+
"logps": -622.9572143554688,
|
| 364 |
+
"loss": 0.2027,
|
| 365 |
+
"objective": 0.20102910697460175,
|
| 366 |
+
"ranking_idealized": 0.9883333444595337,
|
| 367 |
+
"ranking_idealized_expo": 0.6195833086967468,
|
| 368 |
+
"ranking_simple": 0.8554166555404663,
|
| 369 |
+
"regularize": 0.20102910697460175,
|
| 370 |
+
"step": 550,
|
| 371 |
+
"wo_beta": 329.6236877441406
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 1.5588096362777515,
|
| 375 |
+
"eval_dpo_loss": 0.6730445623397827,
|
| 376 |
+
"eval_logits": -10.393749237060547,
|
| 377 |
+
"eval_logps": -622.4690551757812,
|
| 378 |
+
"eval_loss": 0.39951398968696594,
|
| 379 |
+
"eval_objective": 0.40004628896713257,
|
| 380 |
+
"eval_ranking_idealized": 0.9870600700378418,
|
| 381 |
+
"eval_ranking_idealized_expo": 0.6319875717163086,
|
| 382 |
+
"eval_ranking_simple": 0.5947204828262329,
|
| 383 |
+
"eval_regularize": 0.40004628896713257,
|
| 384 |
+
"eval_runtime": 308.3244,
|
| 385 |
+
"eval_samples_per_second": 18.779,
|
| 386 |
+
"eval_steps_per_second": 1.567,
|
| 387 |
+
"eval_wo_beta": 185.8619842529297,
|
| 388 |
+
"step": 550
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"dpo_loss": 0.5614171624183655,
|
| 392 |
+
"epoch": 1.7028814359943316,
|
| 393 |
+
"grad_norm": 8.320325832092646,
|
| 394 |
+
"learning_rate": 4.16692250129073e-06,
|
| 395 |
+
"logits": -10.030584335327148,
|
| 396 |
+
"logps": -637.830078125,
|
| 397 |
+
"loss": 0.1897,
|
| 398 |
+
"objective": 0.1895790994167328,
|
| 399 |
+
"ranking_idealized": 0.9850000143051147,
|
| 400 |
+
"ranking_idealized_expo": 0.6225000023841858,
|
| 401 |
+
"ranking_simple": 0.8608333468437195,
|
| 402 |
+
"regularize": 0.1895790994167328,
|
| 403 |
+
"step": 600,
|
| 404 |
+
"wo_beta": 343.5328369140625
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 1.7028814359943316,
|
| 408 |
+
"eval_dpo_loss": 0.6715614199638367,
|
| 409 |
+
"eval_logits": -11.553997993469238,
|
| 410 |
+
"eval_logps": -755.1119384765625,
|
| 411 |
+
"eval_loss": 0.4028206467628479,
|
| 412 |
+
"eval_objective": 0.4023354947566986,
|
| 413 |
+
"eval_ranking_idealized": 0.9870600700378418,
|
| 414 |
+
"eval_ranking_idealized_expo": 0.6319875717163086,
|
| 415 |
+
"eval_ranking_simple": 0.5952380895614624,
|
| 416 |
+
"eval_regularize": 0.4023354947566986,
|
| 417 |
+
"eval_runtime": 308.0753,
|
| 418 |
+
"eval_samples_per_second": 18.794,
|
| 419 |
+
"eval_steps_per_second": 1.568,
|
| 420 |
+
"eval_wo_beta": 201.23565673828125,
|
| 421 |
+
"step": 600
|
| 422 |
+
},
|
| 423 |
+
{
|
| 424 |
+
"dpo_loss": 0.5596800446510315,
|
| 425 |
+
"epoch": 1.8445914029286725,
|
| 426 |
+
"grad_norm": 7.934464503648664,
|
| 427 |
+
"learning_rate": 3.974272604254906e-06,
|
| 428 |
+
"logits": -10.258437156677246,
|
| 429 |
+
"logps": -652.069091796875,
|
| 430 |
+
"loss": 0.1797,
|
| 431 |
+
"objective": 0.17688363790512085,
|
| 432 |
+
"ranking_idealized": 0.9837499856948853,
|
| 433 |
+
"ranking_idealized_expo": 0.6324999928474426,
|
| 434 |
+
"ranking_simple": 0.8679166436195374,
|
| 435 |
+
"regularize": 0.17688363790512085,
|
| 436 |
+
"step": 650,
|
| 437 |
+
"wo_beta": 346.150634765625
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 1.8445914029286725,
|
| 441 |
+
"eval_dpo_loss": 0.6730169653892517,
|
| 442 |
+
"eval_logits": -10.819306373596191,
|
| 443 |
+
"eval_logps": -673.7770385742188,
|
| 444 |
+
"eval_loss": 0.39966142177581787,
|
| 445 |
+
"eval_objective": 0.3991987109184265,
|
| 446 |
+
"eval_ranking_idealized": 0.9870600700378418,
|
| 447 |
+
"eval_ranking_idealized_expo": 0.6319875717163086,
|
| 448 |
+
"eval_ranking_simple": 0.5942028760910034,
|
| 449 |
+
"eval_regularize": 0.3991987109184265,
|
| 450 |
+
"eval_runtime": 307.7337,
|
| 451 |
+
"eval_samples_per_second": 18.815,
|
| 452 |
+
"eval_steps_per_second": 1.57,
|
| 453 |
+
"eval_wo_beta": 188.3079071044922,
|
| 454 |
+
"step": 650
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"dpo_loss": 0.5620754957199097,
|
| 458 |
+
"epoch": 1.9863013698630136,
|
| 459 |
+
"grad_norm": 7.492993323139572,
|
| 460 |
+
"learning_rate": 3.767136614452458e-06,
|
| 461 |
+
"logits": -10.352383613586426,
|
| 462 |
+
"logps": -666.3911743164062,
|
| 463 |
+
"loss": 0.1689,
|
| 464 |
+
"objective": 0.17240169644355774,
|
| 465 |
+
"ranking_idealized": 0.9908333420753479,
|
| 466 |
+
"ranking_idealized_expo": 0.621666669845581,
|
| 467 |
+
"ranking_simple": 0.8704166412353516,
|
| 468 |
+
"regularize": 0.17240169644355774,
|
| 469 |
+
"step": 700,
|
| 470 |
+
"wo_beta": 337.1234130859375
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 1.9863013698630136,
|
| 474 |
+
"eval_dpo_loss": 0.6712956428527832,
|
| 475 |
+
"eval_logits": -11.077164649963379,
|
| 476 |
+
"eval_logps": -653.8335571289062,
|
| 477 |
+
"eval_loss": 0.3985289931297302,
|
| 478 |
+
"eval_objective": 0.3969685435295105,
|
| 479 |
+
"eval_ranking_idealized": 0.9870600700378418,
|
| 480 |
+
"eval_ranking_idealized_expo": 0.6319875717163086,
|
| 481 |
+
"eval_ranking_simple": 0.5910972952842712,
|
| 482 |
+
"eval_regularize": 0.3969685435295105,
|
| 483 |
+
"eval_runtime": 308.1698,
|
| 484 |
+
"eval_samples_per_second": 18.788,
|
| 485 |
+
"eval_steps_per_second": 1.567,
|
| 486 |
+
"eval_wo_beta": 182.3851776123047,
|
| 487 |
+
"step": 700
|
| 488 |
+
},
|
| 489 |
+
{
|
| 490 |
+
"dpo_loss": 0.5451232194900513,
|
| 491 |
+
"epoch": 2.1280113367973548,
|
| 492 |
+
"grad_norm": 6.5335714189025635,
|
| 493 |
+
"learning_rate": 3.547549834686222e-06,
|
| 494 |
+
"logits": -11.081757545471191,
|
| 495 |
+
"logps": -661.833251953125,
|
| 496 |
+
"loss": 0.1492,
|
| 497 |
+
"objective": 0.1450423002243042,
|
| 498 |
+
"ranking_idealized": 0.9887499809265137,
|
| 499 |
+
"ranking_idealized_expo": 0.627916693687439,
|
| 500 |
+
"ranking_simple": 0.8899999856948853,
|
| 501 |
+
"regularize": 0.1450423002243042,
|
| 502 |
+
"step": 750,
|
| 503 |
+
"wo_beta": 375.9898986816406
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 2.1280113367973548,
|
| 507 |
+
"eval_dpo_loss": 0.6708096265792847,
|
| 508 |
+
"eval_logits": -11.47172737121582,
|
| 509 |
+
"eval_logps": -624.3671875,
|
| 510 |
+
"eval_loss": 0.39592552185058594,
|
| 511 |
+
"eval_objective": 0.3955978453159332,
|
| 512 |
+
"eval_ranking_idealized": 0.9870600700378418,
|
| 513 |
+
"eval_ranking_idealized_expo": 0.6319875717163086,
|
| 514 |
+
"eval_ranking_simple": 0.6024844646453857,
|
| 515 |
+
"eval_regularize": 0.3955978453159332,
|
| 516 |
+
"eval_runtime": 309.3114,
|
| 517 |
+
"eval_samples_per_second": 18.719,
|
| 518 |
+
"eval_steps_per_second": 1.562,
|
| 519 |
+
"eval_wo_beta": 182.7602081298828,
|
| 520 |
+
"step": 750
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"dpo_loss": 0.5449449419975281,
|
| 524 |
+
"epoch": 2.269721303731696,
|
| 525 |
+
"grad_norm": 6.765232021702658,
|
| 526 |
+
"learning_rate": 3.3176699082935546e-06,
|
| 527 |
+
"logits": -11.165037155151367,
|
| 528 |
+
"logps": -677.394775390625,
|
| 529 |
+
"loss": 0.143,
|
| 530 |
+
"objective": 0.14202022552490234,
|
| 531 |
+
"ranking_idealized": 0.98458331823349,
|
| 532 |
+
"ranking_idealized_expo": 0.6179166436195374,
|
| 533 |
+
"ranking_simple": 0.9054166674613953,
|
| 534 |
+
"regularize": 0.14202022552490234,
|
| 535 |
+
"step": 800,
|
| 536 |
+
"wo_beta": 368.1688537597656
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 2.269721303731696,
|
| 540 |
+
"eval_dpo_loss": 0.6701000332832336,
|
| 541 |
+
"eval_logits": -11.25588607788086,
|
| 542 |
+
"eval_logps": -657.3067016601562,
|
| 543 |
+
"eval_loss": 0.395544171333313,
|
| 544 |
+
"eval_objective": 0.395794540643692,
|
| 545 |
+
"eval_ranking_idealized": 0.9870600700378418,
|
| 546 |
+
"eval_ranking_idealized_expo": 0.6319875717163086,
|
| 547 |
+
"eval_ranking_simple": 0.600931704044342,
|
| 548 |
+
"eval_regularize": 0.395794540643692,
|
| 549 |
+
"eval_runtime": 308.0676,
|
| 550 |
+
"eval_samples_per_second": 18.795,
|
| 551 |
+
"eval_steps_per_second": 1.568,
|
| 552 |
+
"eval_wo_beta": 190.53713989257812,
|
| 553 |
+
"step": 800
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 2.269721303731696,
|
| 557 |
+
"step": 800,
|
| 558 |
+
"total_flos": 0.0,
|
| 559 |
+
"train_loss": 0.051899830102920534,
|
| 560 |
+
"train_runtime": 6898.2847,
|
| 561 |
+
"train_samples_per_second": 36.822,
|
| 562 |
+
"train_steps_per_second": 0.255
|
| 563 |
+
}
|
| 564 |
+
],
|
| 565 |
+
"logging_steps": 50,
|
| 566 |
+
"max_steps": 1760,
|
| 567 |
+
"num_input_tokens_seen": 0,
|
| 568 |
+
"num_train_epochs": 5,
|
| 569 |
+
"save_steps": 50,
|
| 570 |
+
"stateful_callbacks": {
|
| 571 |
+
"EarlyStoppingCallback": {
|
| 572 |
+
"args": {
|
| 573 |
+
"early_stopping_patience": 5,
|
| 574 |
+
"early_stopping_threshold": 0.0
|
| 575 |
+
},
|
| 576 |
+
"attributes": {
|
| 577 |
+
"early_stopping_patience_counter": 0
|
| 578 |
+
}
|
| 579 |
+
},
|
| 580 |
+
"TrainerControl": {
|
| 581 |
+
"args": {
|
| 582 |
+
"should_epoch_stop": false,
|
| 583 |
+
"should_evaluate": false,
|
| 584 |
+
"should_log": false,
|
| 585 |
+
"should_save": true,
|
| 586 |
+
"should_training_stop": true
|
| 587 |
+
},
|
| 588 |
+
"attributes": {}
|
| 589 |
+
}
|
| 590 |
+
},
|
| 591 |
+
"total_flos": 0.0,
|
| 592 |
+
"train_batch_size": 4,
|
| 593 |
+
"trial_name": null,
|
| 594 |
+
"trial_params": null
|
| 595 |
+
}
|