train_copa_1757340232

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: 0.1061
• Num Input Tokens Seen: 281408

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: 456
• 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: 10.0

Training results

Training Loss	Epoch	Step	Validation Loss	Input Tokens Seen
0.5955	0.5	45	0.6678	14016
0.4699	1.0	90	0.4780	28096
0.0988	1.5	135	0.1207	42144
0.0479	2.0	180	0.1130	56352
0.1124	2.5	225	0.1106	70432
0.2231	3.0	270	0.1080	84544
0.0057	3.5	315	0.1061	98688
0.0402	4.0	360	0.1065	112800
0.0513	4.5	405	0.1061	126816
0.075	5.0	450	0.1063	140800
0.0207	5.5	495	0.1073	154848
0.0287	6.0	540	0.1071	168736
0.0148	6.5	585	0.1100	182688
0.0102	7.0	630	0.1107	196896
0.0228	7.5	675	0.1079	211072
0.0943	8.0	720	0.1082	225056
0.121	8.5	765	0.1081	239168
0.0395	9.0	810	0.1083	253312
0.0126	9.5	855	0.1080	267392
0.0798	10.0	900	0.1069	281408

Framework versions

• PEFT 0.15.2
• Transformers 4.51.3
• Pytorch 2.8.0+cu128
• Datasets 3.6.0
• Tokenizers 0.21.1