|
--- |
|
license: llama3 |
|
base_model: meta-llama/Meta-Llama-3-8B-Instruct |
|
tags: |
|
- alignment-handbook |
|
- generated_from_trainer |
|
datasets: |
|
- trl-lib/kto-mix-14k |
|
- chaoweihuang/lf-response-llama3-f1_100_0.8-fg0.5 |
|
model-index: |
|
- name: kto-mix-14k-lf-response-llama3-f1_100_0.8-fg0.5-fgudw4.0-kto-fg |
|
results: [] |
|
--- |
|
|
|
# FactAlign-LLaMA-3-8B |
|
This model is aligned with our **FactAlign** framework for improved long-form factuality, from [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct). |
|
|
|
For more information, please refer to our paper: [FactAlign: Long-form Factuality Alignment of Large Language Models](https://huggingface.co/papers/2410.01691). |
|
|
|
|
|
## Model description |
|
|
|
More information needed |
|
|
|
## Intended uses & limitations |
|
|
|
More information needed |
|
|
|
## Training and evaluation data |
|
|
|
|
|
This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) on the trl-lib/kto-mix-14k and the chaoweihuang/lf-response-llama3-f1_100_0.8-fg0.5 datasets. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.4110 |
|
- Rewards/chosen: 1.7360 |
|
- Logps/chosen: -336.0412 |
|
- Rewards/rejected: -2.2628 |
|
- Logps/rejected: -406.1173 |
|
- Rewards/margins: 3.9987 |
|
- Kl: 0.0141 |
|
- Fg Rewards/chosen Sum: -1.5560 |
|
- Fg Logps/policy Chosen: -6.7332 |
|
- Fg Logps/reference Chosen: -6.0419 |
|
- Count/fg Chosen: 30.1832 |
|
- Fg Rewards/rejected Sum: -0.9033 |
|
- Fg Logps/policy Rejected: -8.6269 |
|
- Fg Logps/reference Rejected: -7.5807 |
|
- Count/fg Rejected: 6.9239 |
|
- Fg Logps/policy Kl: -14.7946 |
|
- Fg Logps/reference Kl: -11.4736 |
|
- Fg Kl: nan |
|
- Fg Loss: 0.7625 |
|
|
|
|
|
## Training procedure |
|
|
|
### Training hyperparameters |
|
|
|
The following hyperparameters were used during training: |
|
- learning_rate: 5e-07 |
|
- train_batch_size: 1 |
|
- eval_batch_size: 1 |
|
- seed: 42 |
|
- distributed_type: multi-GPU |
|
- num_devices: 4 |
|
- gradient_accumulation_steps: 4 |
|
- total_train_batch_size: 16 |
|
- total_eval_batch_size: 4 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- lr_scheduler_warmup_ratio: 0.1 |
|
- num_epochs: 1.0 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Logps/chosen | Rewards/rejected | Logps/rejected | Rewards/margins | Kl | Fg Rewards/chosen Sum | Fg Logps/policy Chosen | Fg Logps/reference Chosen | Count/fg Chosen | Fg Rewards/rejected Sum | Fg Logps/policy Rejected | Fg Logps/reference Rejected | Count/fg Rejected | Fg Logps/policy Kl | Fg Logps/reference Kl | Fg Kl | Fg Loss | |
|
|:-------------:|:------:|:----:|:---------------:|:--------------:|:------------:|:----------------:|:--------------:|:---------------:|:------:|:---------------------:|:----------------------:|:-------------------------:|:---------------:|:-----------------------:|:------------------------:|:---------------------------:|:-----------------:|:------------------:|:---------------------:|:-----:|:-------:| |
|
| 0.4478 | 0.4103 | 400 | 0.4325 | 1.3169 | -340.2313 | -1.7364 | -400.8539 | 3.0534 | 0.0280 | -1.3939 | -6.6287 | -6.0419 | 30.1832 | -0.6768 | -8.3632 | -7.5807 | 6.9239 | -13.6783 | -11.4736 | nan | 0.7654 | |
|
| 0.4043 | 0.8205 | 800 | 0.4110 | 1.7360 | -336.0412 | -2.2628 | -406.1173 | 3.9987 | 0.0141 | -1.5560 | -6.7332 | -6.0419 | 30.1832 | -0.9033 | -8.6269 | -7.5807 | 6.9239 | -14.7946 | -11.4736 | nan | 0.7625 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.41.1 |
|
- Pytorch 2.3.0+cu121 |
|
- Datasets 2.20.0 |
|
- Tokenizers 0.19.1 |
|
|
