train_mrpc_42_1760637563

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

• Loss: 0.1365
• Num Input Tokens Seen: 6769320

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: 42
• 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
0.2667	1.0	826	0.1915	337344
0.1032	2.0	1652	0.1365	675368
0.0683	3.0	2478	0.1844	1014008
0.0465	4.0	3304	0.1419	1353224
0.0003	5.0	4130	0.2116	1692320
0.1108	6.0	4956	0.2352	2030488
0.0003	7.0	5782	0.2487	2368032
0.0	8.0	6608	0.2919	2706552
0.0006	9.0	7434	0.2465	3044984
0.0	10.0	8260	0.3647	3384288
0.0	11.0	9086	0.2725	3722552
0.0	12.0	9912	0.2662	4061216
0.0	13.0	10738	0.3514	4399064
0.0	14.0	11564	0.3864	4737648
0.0	15.0	12390	0.4192	5075144
0.0	16.0	13216	0.4385	5414032
0.0	17.0	14042	0.4487	5752528
0.0	18.0	14868	0.4642	6091472
0.0	19.0	15694	0.4621	6430144
0.0	20.0	16520	0.4683	6769320

Framework versions

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