train_math_qa_789_1760637949

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

• Loss: 0.8025
• Num Input Tokens Seen: 77933776

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: 0.03
• train_batch_size: 4
• eval_batch_size: 4
• seed: 789
• 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
0.784	1.0	6714	0.8027	3898224
0.8239	2.0	13428	0.8025	7796616
0.7743	3.0	20142	0.8332	11688128
0.9516	4.0	26856	0.8071	15585640
0.7931	5.0	33570	0.8109	19481256
0.7964	6.0	40284	0.8084	23379928
0.8347	7.0	46998	0.8091	27274992
0.7829	8.0	53712	0.8297	31169464
0.8378	9.0	60426	0.8148	35061680
0.7461	10.0	67140	0.8145	38957336
0.8215	11.0	73854	0.8084	42854488
0.7992	12.0	80568	0.8065	46754376
0.845	13.0	87282	0.8068	50647376
0.8099	14.0	93996	0.8044	54543272
0.8187	15.0	100710	0.8040	58447368
0.7777	16.0	107424	0.8044	62343120
0.7904	17.0	114138	0.8037	66240072
0.812	18.0	120852	0.8044	70140040
0.819	19.0	127566	0.8039	74035080
0.8216	20.0	134280	0.8043	77933776

Framework versions

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