train_copa_123_1760637647

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

• Loss: 0.0494
• Num Input Tokens Seen: 563328

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: 123
• 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
0.2088	1.0	90	0.1073	28096
0.1251	2.0	180	0.0544	56128
0.0004	3.0	270	0.0494	84352
0.0	4.0	360	0.0727	112576
0.0	5.0	450	0.0587	140832
0.0	6.0	540	0.0599	169056
0.0	7.0	630	0.0557	197344
0.0	8.0	720	0.0576	225536
0.0	9.0	810	0.0584	253696
0.0	10.0	900	0.0572	281856
0.0	11.0	990	0.0583	310080
0.0	12.0	1080	0.0612	338144
0.0	13.0	1170	0.0603	366336
0.0	14.0	1260	0.0599	394464
0.0	15.0	1350	0.0602	422592
0.0	16.0	1440	0.0573	450624
0.0	17.0	1530	0.0606	478720
0.0	18.0	1620	0.0617	507008
0.0	19.0	1710	0.0624	535136
0.0	20.0	1800	0.0631	563328

Framework versions

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