train_svamp_101112_1757596157

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: 0.4107
• Num Input Tokens Seen: 1348864

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: 2
• eval_batch_size: 2
• 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
0.6409	1.0	315	0.7571	67488
0.262	2.0	630	0.3623	134832
0.0962	3.0	945	0.2180	202352
0.0468	4.0	1260	0.1878	269776
0.0382	5.0	1575	0.2140	337328
0.0017	6.0	1890	0.3292	404608
0.0037	7.0	2205	0.3098	472144
0.005	8.0	2520	0.3992	539664
0.0	9.0	2835	0.3648	607136
0.0002	10.0	3150	0.3280	674496
0.0	11.0	3465	0.3562	741840
0.0001	12.0	3780	0.3841	809312
0.0	13.0	4095	0.3958	876784
0.0	14.0	4410	0.4013	944080
0.0	15.0	4725	0.4053	1011456
0.0	16.0	5040	0.4078	1078880
0.0	17.0	5355	0.4081	1146416
0.0	18.0	5670	0.4113	1213888
0.0	19.0	5985	0.4104	1281488
0.0	20.0	6300	0.4107	1348864

Framework versions

• PEFT 0.15.2
• Transformers 4.51.3
• Pytorch 2.8.0+cu128
• Datasets 3.6.0
• Tokenizers 0.21.1