train_multirc_789_1770347547

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

• Loss: 0.1373
• Num Input Tokens Seen: 264395536

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: 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
0.3515	1.0	6130	0.1758	13229504
0.4644	2.0	12260	0.1555	26459312
0.0673	3.0	18390	0.1501	39686560
0.0434	4.0	24520	0.1467	52924864
0.0446	5.0	30650	0.1434	66146528
0.1041	6.0	36780	0.1441	79364192
0.0612	7.0	42910	0.1468	92568704
0.1859	8.0	49040	0.1373	105788704
0.0601	9.0	55170	0.1433	119004560
0.1029	10.0	61300	0.1435	132223184
0.1188	11.0	67430	0.1397	145445936
0.0692	12.0	73560	0.1439	158686320
0.0733	13.0	79690	0.1437	171910720
0.0066	14.0	85820	0.1449	185139040
0.0955	15.0	91950	0.1477	198344384
0.2482	16.0	98080	0.1450	211581728
0.0487	17.0	104210	0.1476	224787008
0.1336	18.0	110340	0.1471	238003072
0.1346	19.0	116470	0.1484	251207552
0.1536	20.0	122600	0.1481	264395536

Framework versions

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