train_rte_1753094154

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

• Loss: 0.0771
• Num Input Tokens Seen: 3481336

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
0.0643	0.5009	281	0.1055	176032
0.1141	1.0018	562	0.0771	349200
0.1187	1.5027	843	0.0927	524208
0.0694	2.0036	1124	0.1341	699264
0.0087	2.5045	1405	0.0975	873600
0.0004	3.0053	1686	0.1157	1048184
0.024	3.5062	1967	0.0987	1223864
0.0003	4.0071	2248	0.1363	1397624
0.0001	4.5080	2529	0.1388	1570936
0.0002	5.0089	2810	0.1386	1746384
0.0	5.5098	3091	0.1823	1922384
0.0	6.0107	3372	0.1837	2092320
0.0001	6.5116	3653	0.1998	2267520
0.0	7.0125	3934	0.2042	2441688
0.0	7.5134	4215	0.2252	2614936
0.0	8.0143	4496	0.2238	2790832
0.0	8.5152	4777	0.2269	2963888
0.0	9.0160	5058	0.2289	3137352
0.0	9.5169	5339	0.2281	3312648

Framework versions

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