train_math_qa_789_1760637950

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: 0.9884
• 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: 0.001
• 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
0.7695	1.0	6714	0.8082	3898224
0.8229	2.0	13428	0.8038	7796616
0.7923	3.0	20142	0.7987	11688128
0.8019	4.0	26856	0.7998	15585640
0.7967	5.0	33570	0.7998	19481256
0.7831	6.0	40284	0.7972	23379928
0.8094	7.0	46998	0.7836	27274992
0.7408	8.0	53712	0.7834	31169464
0.7353	9.0	60426	0.7733	35061680
0.7777	10.0	67140	0.7718	38957336
0.7495	11.0	73854	0.7632	42854488
0.6448	12.0	80568	0.7571	46754376
0.7825	13.0	87282	0.7566	50647376
0.8034	14.0	93996	0.7552	54543272
0.754	15.0	100710	0.7568	58447368
0.6479	16.0	107424	0.7532	62343120
0.6796	17.0	114138	0.7595	66240072
0.7363	18.0	120852	0.7608	70140040
0.8079	19.0	127566	0.7632	74035080
0.7444	20.0	134280	0.7616	77933776

Framework versions

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