train_mnli_1753094135

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

• Loss: 0.0828
• Num Input Tokens Seen: 347859920

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: 10.0

Training results

Training Loss	Epoch	Step	Validation Loss	Input Tokens Seen
0.1413	0.5	44179	0.1377	17403808
0.2235	1.0	88358	0.1067	34786008
0.0359	1.5	132537	0.0968	52165240
0.0608	2.0	176716	0.0922	69564424
0.023	2.5	220895	0.0894	86951080
0.0966	3.0	265074	0.0874	104352808
0.1134	3.5	309253	0.0866	121746504
0.1315	4.0	353432	0.0853	139123792
0.148	4.5	397611	0.0848	156526672
0.0366	5.0	441790	0.0841	173916408
0.0499	5.5	485969	0.0839	191309592
0.0782	6.0	530148	0.0836	208701328
0.0465	6.5	574327	0.0839	226098768
0.033	7.0	618506	0.0831	243493272
0.054	7.5	662685	0.0831	260881240
0.063	8.0	706864	0.0829	278276232
0.0511	8.5	751043	0.0829	295687496
0.0792	9.0	795222	0.0829	313062872
0.0915	9.5	839401	0.0829	330444056
0.0121	10.0	883580	0.0828	347859920

Framework versions

• PEFT 0.15.2
• Transformers 4.51.3
• Pytorch 2.7.1+cu126
• Datasets 3.6.0
• Tokenizers 0.21.1