metadata

base_model: meta-llama/Meta-Llama-3-8B-Instruct
library_name: peft
license: llama3
metrics:
  - accuracy
tags:
  - generated_from_trainer
model-index:
  - name: reward_model
    results: []

reward_model

This model is a fine-tuned version of meta-llama/Meta-Llama-3-8B-Instruct on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.7036
Accuracy: 0.5236

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 5e-05
train_batch_size: 16
eval_batch_size: 8
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_ratio: 0.02
num_epochs: 2

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.7293	0.08	128	0.7252	0.4850
0.7412	0.15	256	0.6925	0.5386
0.7182	0.23	384	0.6954	0.5327
0.6997	0.3	512	0.6941	0.5277
0.7547	0.38	640	0.6959	0.5279
0.7123	0.45	768	0.6993	0.5252
0.7281	0.53	896	0.6962	0.5275
0.7169	0.6	1024	0.6986	0.5156
0.7244	0.68	1152	0.6981	0.5125
0.7199	0.75	1280	0.7000	0.5060
0.7311	0.83	1408	0.6959	0.5140
0.7123	0.9	1536	0.6956	0.5154
0.7344	0.98	1664	0.6970	0.5100
0.7105	1.05	1792	0.6933	0.5219
0.6947	1.13	1920	0.6944	0.5259
0.7261	1.21	2048	0.6960	0.5256
0.6997	1.28	2176	0.6974	0.5188
0.7442	1.36	2304	0.6960	0.5163
0.7004	1.43	2432	0.6987	0.5286
0.7089	1.51	2560	0.6982	0.5288
0.7142	1.58	2688	0.7014	0.5154
0.7364	1.66	2816	0.6997	0.5202
0.6915	1.73	2944	0.7043	0.5200
0.7322	1.81	3072	0.7037	0.5229
0.7524	1.88	3200	0.7019	0.5219
0.7192	1.96	3328	0.7036	0.5236

Framework versions

PEFT 0.12.0
Transformers 4.36.0
Pytorch 2.2.0
Datasets 2.20.0
Tokenizers 0.15.2