train_rte_1754502820

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: 0.1195
• 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
0.2025	0.5009	281	0.1984	176032
0.1763	1.0018	562	0.1349	349200
0.0961	1.5027	843	0.1245	524208
0.1243	2.0036	1124	0.1514	699264
0.0632	2.5045	1405	0.1449	873600
0.0335	3.0053	1686	0.1411	1048184
0.098	3.5062	1967	0.1195	1223864
0.0022	4.0071	2248	0.1401	1397624
0.0106	4.5080	2529	0.1345	1570936
0.077	5.0089	2810	0.1460	1746384
0.0712	5.5098	3091	0.1524	1922384
0.0046	6.0107	3372	0.1515	2092320
0.0674	6.5116	3653	0.1517	2267520
0.0127	7.0125	3934	0.1563	2441688
0.0054	7.5134	4215	0.1573	2614936
0.0794	8.0143	4496	0.1566	2790832
0.0063	8.5152	4777	0.1583	2963888
0.0007	9.0160	5058	0.1581	3137352
0.0782	9.5169	5339	0.1591	3312648

Framework versions

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