train_mnli_42_1767887022

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: 0.1062
• Num Input Tokens Seen: 312915968

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
0.0033	0.5	88358	0.1789	15653504
0.0058	1.0	176716	0.1181	31283664
0.0653	1.5	265074	0.1154	46924576
0.0036	2.0	353432	0.1220	62587160
0.0059	2.5	441790	0.1077	78223688
0.0076	3.0	530148	0.1124	93878760
0.0032	3.5	618506	0.1145	109512680
0.0038	4.0	706864	0.1124	125170816
0.0136	4.5	795222	0.1062	140807360
0.0013	5.0	883580	0.1107	156460936
0.004	5.5	971938	0.1080	172101320
0.0994	6.0	1060296	0.1091	187749624
0.5485	6.5	1148654	0.1159	203384056
0.0711	7.0	1237012	0.1158	219037384
0.0051	7.5	1325370	0.1166	234681720
0.4636	8.0	1413728	0.1132	250330736
0.0016	8.5	1502086	0.1186	265969568
0.0034	9.0	1590444	0.1195	281623656
0.1838	9.5	1678802	0.1205	297267640
0.3616	10.0	1767160	0.1212	312915968

Framework versions

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