train_copa_789_1760637876

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: 0.2602
• Num Input Tokens Seen: 563744

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
0.6809	1.0	90	0.5334	28192
0.6512	2.0	180	0.4626	56192
0.608	3.0	270	0.4240	84192
0.5009	4.0	360	0.3563	112544
0.5363	5.0	450	0.3249	140960
0.2618	6.0	540	0.2948	169216
0.2119	7.0	630	0.2802	197248
0.621	8.0	720	0.2738	225440
0.235	9.0	810	0.2606	253632
0.3074	10.0	900	0.2653	281984
0.2951	11.0	990	0.2702	310176
0.4453	12.0	1080	0.2673	338400
0.3689	13.0	1170	0.2658	366688
0.3106	14.0	1260	0.2602	394752
0.2445	15.0	1350	0.2658	422912
0.261	16.0	1440	0.2753	451008
0.5208	17.0	1530	0.2668	479104
0.2563	18.0	1620	0.2768	507392
0.4139	19.0	1710	0.2729	535584
0.3828	20.0	1800	0.2692	563744

Framework versions

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