train_openbookqa_101112_1760638027

This model is a fine-tuned version of meta-llama/Meta-Llama-3-8B-Instruct on the openbookqa dataset. It achieves the following results on the evaluation set:

• Loss: 0.1929
• Num Input Tokens Seen: 8474968

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: 4
• eval_batch_size: 4
• seed: 101112
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Input Tokens Seen
0.34	1.0	1116	0.1929	424568
0.1778	2.0	2232	0.2455	848552
0.2616	3.0	3348	0.3402	1271776
0.1286	4.0	4464	0.3154	1694896
0.0001	5.0	5580	0.4146	2118456
0.0	6.0	6696	0.6757	2542752
0.0	7.0	7812	0.6089	2966512
0.0	8.0	8928	0.5241	3390048
0.0	9.0	10044	0.5494	3814704
0.0	10.0	11160	0.7155	4238440
0.0	11.0	12276	0.7168	4662136
0.0	12.0	13392	0.5492	5086336
0.0	13.0	14508	0.6572	5510768
0.0	14.0	15624	0.6176	5933936
0.0	15.0	16740	0.6405	6357536
0.0	16.0	17856	0.6836	6779872
0.0	17.0	18972	0.6918	7203216
0.0	18.0	20088	0.7063	7626944
0.0	19.0	21204	0.7114	8051216
0.0	20.0	22320	0.7171	8474968

Framework versions

• PEFT 0.17.1
• Transformers 4.51.3
• Pytorch 2.9.0+cu128
• Datasets 4.0.0
• Tokenizers 0.21.4