train_copa_42_1767961234

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: 0.0402
• Num Input Tokens Seen: 274272

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: 2
• eval_batch_size: 2
• 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: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Input Tokens Seen
0.019	0.5	90	0.0754	13776
0.0464	1.0	180	0.0514	27424
0.151	1.5	270	0.0658	41040
0.0007	2.0	360	0.0402	54832
0.0009	2.5	450	0.0501	68592
0.2835	3.0	540	0.0404	82256
0.0001	3.5	630	0.0425	95888
0.221	4.0	720	0.0548	109728
0.0001	4.5	810	0.0553	123424
0.0131	5.0	900	0.0602	137120
0.0001	5.5	990	0.0592	150816
0.0001	6.0	1080	0.0659	164560
0.0003	6.5	1170	0.0705	178304
0.0001	7.0	1260	0.0755	192032
0.0011	7.5	1350	0.0738	205776
0.0002	8.0	1440	0.0745	219456
0.1218	8.5	1530	0.0819	233072
0.0	9.0	1620	0.0793	246880
0.0	9.5	1710	0.0770	260592
0.0001	10.0	1800	0.0792	274272

Framework versions

• PEFT 0.17.1
• Transformers 4.51.3
• Pytorch 2.9.1+cu128
• Datasets 4.0.0
• Tokenizers 0.21.4