train_mrpc_101112_1760638017

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: 0.9237
• Num Input Tokens Seen: 6014512

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: 1e-05
• train_batch_size: 4
• eval_batch_size: 4
• 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
0.1553	2.0	1468	0.1782	601520
0.1411	4.0	2936	0.1616	1201552
0.1154	6.0	4404	0.2181	1803312
0.165	8.0	5872	0.2984	2405152
0.0167	10.0	7340	0.4961	3007040
0.237	12.0	8808	0.5478	3609392
0.0	14.0	10276	0.7023	4210592
0.0	16.0	11744	0.9020	4811824
0.0	18.0	13212	0.9190	5413152
0.0	20.0	14680	0.9237	6014512

Framework versions

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