train_svamp_456_1760637773

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

• Loss: 0.1674
• Num Input Tokens Seen: 1432752

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: 456
• 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
0.1886	1.0	158	0.1601	71728
0.0963	2.0	316	0.0951	143392
0.0606	3.0	474	0.1081	214928
0.0531	4.0	632	0.0753	286624
0.048	5.0	790	0.0818	358096
0.0354	6.0	948	0.0833	429680
0.0218	7.0	1106	0.0843	501168
0.0228	8.0	1264	0.0884	573152
0.0277	9.0	1422	0.0969	644672
0.0074	10.0	1580	0.1022	716272
0.0143	11.0	1738	0.1302	787952
0.0118	12.0	1896	0.1231	859584
0.0033	13.0	2054	0.1430	931280
0.0032	14.0	2212	0.1486	1002960
0.0003	15.0	2370	0.1587	1074528
0.0011	16.0	2528	0.1645	1146144
0.0005	17.0	2686	0.1668	1217760
0.0007	18.0	2844	0.1684	1289504
0.0005	19.0	3002	0.1684	1361040
0.0001	20.0	3160	0.1691	1432752

Framework versions

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